article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
a dusty plasma with massive dust particles can experience significant correlation effects when the coulomb coupling parameter @xmath1 ( ( @xmath2 $ ] , where @xmath3 , @xmath4 , @xmath5 , @xmath6 are number of charges on dust particles , interparticle distance , dust particle temperature and plasma debye length respectively ) . past theoretical works @xcite suggest that correlation effects can significantly modify the collective modes of a dusty plasmas .
kaw and sen @xcite have studied the influence of correlation effect on daw using a generalized hydrodynamic model and have obtained a new dispersive corrections , overall reduction in frequency and phase velocity , existence of @xmath7 regime . +
they have also pointed out that in most laboratory plasmas such correlation induced effects can get masked out by collisional or ion streaming effects . to the best of our knowledge , such correlation induced effects have not been experimentally measured earlier . in the present work we report results on such a measurement and compare it to theoretical results .
the schematic of experimental setup is shown in fig.[fig : exp_setup](a ) and it is described in detail elsewere @xcite .
the discharge was struck between the vacuum vessel ( cathode ) and rod shaped anode by apply a dc voltage ( @xmath8 ) .
the inner side of the cathode is covered with a thin stainless steel foil to avoid micro - arcs .
graphite flakes expanded in thermal plasma were used as dust in present experiment .
these graphite particles are of very low density ( @xmath9 ) .
the system was pumped down to a base pressure ( @xmath10 ) of @xmath11 mbar using a rotary pump .
it was filled with argon and plasma was formed by applying a discharge voltage ( @xmath8 ) @xmath12 volts to the anode at @xmath13 mbar . then the pressure was reduced gradually to 0.086 mbar . it was observed that dust particles begin to accumulate near the anode region .
the anode region provides suitable electric field against the gravitational field for levitation of dust grains .
the radial electric field confines the dust particles . an ac signal ( using a signal generator and a power amplifier )
was superimposed with the discharge voltage to excite the waves ( see fig.[fig : exp_setup](b ) ) .
the amplitude of the ac signal was increased with increase in pressure to excite the daw .
the frequency of applied ac signal was varied from 0 to 2 hz very precisely .
+ the levitated dust particles were illuminated by green nd - yag diode laser light .
the laser light was spread into a sheet by a cylindrical lens and forward scattered light from the dust cloud was used to visualize the dust particles .
the scattered light from the dust particles was recorded using a ccd camera ( 25 fps ) and later stored into a computer using a frame grabber card . the ion density ( @xmath14 ) , electron temperature ( @xmath15 )
was measured using single a langmuier probe .
the dust temperature ( @xmath5 ) was calculated from the velocity of the particles by tracing single particle in different frames .
c ) daw of different wavelength ]
digitized data was arranged into several still images and stored as separate data records in the interval of 1/25 sec .
consecutive data records were then fourier analyzed using the technique described by smith _
et al _ @xcite , to obtain cross - power , phase and coherence ( @xmath16 ) spectra respectively .
the frequency of the waves was computed using phase spectrum and time delay between consecutive frames .
the experiment was carried out to study the dispersive characteristics of dust acoustic wave in two different pressures regimes viz , low collisional regime ( p=0.086 mbar ) and high collisional regime ( p=0.27 mbar ) for a strongly coupled dusty plasma . in both regimes ,
initially a stable dust cloud was formed in the sheath region near the cathode ( vessel wall ) .
typical stable dust cloud at above mentioned parameters is shown in the fig .
it clearly shows that there is no background oscillations before applying ac signal . as the ac voltage was modulated with the dc voltage , the plasma glow as well as the discharge current exhibit oscillation of same frequency .
the plasma boundary also varies with time , thus dust cloud as a whole also exhibits similar motion .
in addition to this dust cloud breaks up into series of crests and troughs ( see in fig 2(b ) and 2(c ) ) similar to pressure perturbation when the ac voltage approaches minimum value .
these perturbations have been referred in text as daw oscillations and are similar to those reported earlier @xcite .
the dust cloud oscillates specially due to the application of oscillating voltage .
a ccd camera was used to capture the images of dust cloud when it crossed the field of view .
it was observed that daw were excited just before the oscillating discharge voltage approaches minimum value .
the frequency and wave length of the externally driven dust acoustic wave were measured from the fft techniques of consecutive frames . for our experimental parameter
the phase velocity was quite low ( few mm / sec ) compare to other experimental groups @xcite .
the dispersion relation of the externally driven daw in both regime are shown in fig 3(a ) and 3(b ) .
the open circles denote the experimental points . to compare our experimental results we choose a theory where strong correlation effects were taken into account along with the collisional effects @xcite .
we have plotted ( with solid and dotted line ) the dispersion relation according to the above theory for our experimental dusty plasma parameters . the daw dispersion relation from the above theory
is given by , @xmath17 where @xmath18 , @xmath19 and @xmath20 are adiabatic index and coefficient of compressibility as mentioned in reference @xcite .
we have used experimentally measured @xmath14 , @xmath15 , @xmath4 and @xmath5 to calculate @xmath3 , @xmath21 , @xmath22 and @xmath23 was used as a free parameter to fit the theoretical dispersion curve @xcite .
the above dispersion relation is valid for hydrodynamic regime ( @xmath24 , where @xmath25 is relaxation time ) . +
as seen , there is a good agreement between the experimental and theoretical disperson curves for high and low collisional regime . in the strongly correlated regime @xmath26
, since @xmath20 changes sign for @xmath27 and @xmath28 in the weekly correlated regime as discussed in @xcite , the experimental dispersion relation shows negative slope ( @xmath29 ) which is consistent with the theoretical model results predicted by kaw et al .
it is clear from the figures , at low collision frequencies the dispersion curves for strongly and weakly correlated regime are very different where as at higher collisional frequencies the differences are not significant .
this is because the modifications of the dispersion relation due to dust neutral collisions are much stronger than correlation effects in higher pressure .
close match of theoretical prediction with experimental observation clearly suggest that correlation effect is significant in our case
. substantial reduction of peak value of dispersive curve in figure 3(b ) shows the effect of strong collision .
we have carried out an experiment on externally excited dust acoustic wave in a strongly coupled regime .
the measured dispersive characteristic shows substantial reduction in phase velocity and existence of @xmath0 regime .
the dispersion relation at low pressure shows the effect of correlation .
the correlation effects gets masked by collisional effect at high neutral pressure .
these observations appear with past theoretical predictions . 1 p. carini , g. kalman and k. i. golden , phys . rev .
a. * 26 * , 1686 , ( 1982 ) .
p. kairani and g. kalman , phys .
a * 105 * , 232 ( 1984 ) .
f. postogna and m. p. tosi , nuovo cimento b*55 * , 399 ( 1980 )
. k. i. golden , phys .
a * 35 * , 5278 ( 1987 ) .
m. a. bervosky , phys .
a * 166 * , 365 ( 1992 ) .
g. kalman , m. rosenbarg and h. e. dewitt , phys .
lett . , * 84 * , 6030 , 2000 .
p. k. kaw , a. sen , phys .
plasmas , * 5 * , 3552 ( 1998 ) .
j. pramanik , g. prasad , a. sen , p. k. kaw , phys .
lett . * 88 * , 175001 , ( 2002 ) .
d.e smith , e.j .
powers and g.s.caldwell , ieee transaction on plasma science , * ps-2 * , 261 ( 1974 ) .
a. barkan , r.l .
marlino , n. dangelo , phys .
plasmas , * 2 * , 3563 ( 1995 ) . c. thompson , a. barkan , n. dangelo , r.l .
marlino , phys .
plasmas * 7 * , 1374 ( 2000 ) .
fortov et al .
plasmas , * 7 * , 1374 , 2000 . j. pramanik , b. m veeresha , g. prasad , a.sen , p.k .
kaw , phys lett a , * 312 * , 84 , ( 2003 ) .
j.b piper , j. goree , phys rev .
, * 77*,3137 ( 1996 ) . | low frequency dust acoustic waves ( daw ) were excited in a laboratory argon dusty plasma by modulating the discharge voltage with a low frequency ac signal .
metallic graphite particles were used as dust grains and a digital fft technique was used to obtain dispersion characteristics .
the experimental dispersion relation shows the reduction of phase velocity and a regime where @xmath0 .
a comparison is made with existing theoretical model .
address = institute for plasma research , bhat , gandhinagar-382428 , india address = institute for plasma research , bhat , gandhinagar-382428 , india address = institute for plasma research , bhat , gandhinagar-382428 , india address = institute for plasma research , bhat , gandhinagar-382428 , india |
hundred years have passed since the discovery of cosmic radiation , and an accurate reconstruction of both the energy and mass of primary cosmic rays at high energies still remains a compelling need in contemporary astro - particle physics .
only recently the methods for detection of coherent mhz radiation from extensive air showers as well as the understanding of the underlying emission physics have been strongly improved , reaching important milestones @xcite .
a precise reconstruction of both the primary energy and the depth of the shower maximum ( x@xmath0 ) of air showers is a fundamental goal for modern radio detection , which aims to become competitive and complementary to the already well - established fluorescence and cherenkov detection techniques , limited , in contrast to the radio technique , to a low duty cycle @xcite . in the following ,
we present a method ( slope method ) , which obtains information on the energy and the depth of maximum of air showers from features of the lateral distribution of the radio signals : on the one hand , a defined distance from the shower axis exists where the reconstruction of the primary energy is affected least by shower - to - shower fluctuations .
the presence of such a characteristic distance was previously predicted with reas2 simulations @xcite .
here , we demonstrate on the basis of state - of - the - art coreas simulations @xcite that such a characteristic distance is still present when refractive index effects and radiation due to the variation of the number of charged particles are taken into account , and that measurements at this distance can thus be exploited for an energy determination . on the other hand ,
the slope of the radio lateral distribution is related to the geometrical distance between the observer and the radio source @xcite .
therefore , information on the depth of shower maximum ( x@xmath0 ) and consecutively on the type of primary particle initiating the shower , can be directly extracted from the radio lateral distribution measurement : for iron - initiated air showers , which start earlier and develop faster in the atmosphere compared to proton - initiated air showers , the radio source is typically further away from an observer at ground , and thus the slope of the lateral distribution is flatter in comparison .
recently , this sensitivity of the slope of the measured radio lateral distribution to the development of air showers has been proven experimentally with lopes measurements @xcite .
we strive to take a practical approach tailored to the data quality provided by the lopes experiment . in particular , we aim to develop an analysis strategy which can exploit the potential of lopes in light of its experimental limitations ( relatively low number of antennas , limited lateral distance range covered in a single measurement , relatively high background noise ) without becoming overly involved and complex . for experiments with higher data quality
, the methods presented here should thus be refined and further improved .
another independent method has been separately developed and applied on lopes measurements to extract information on x@xmath0 .
this method considers the shape of the radio wave front @xcite , but lower precision in the reconstruction of x@xmath0 compared to the slope method was achieved .
the slope and wavefront methods could in the future also be combined with each other and with further methods such as the determination of the pulse width , or spectral slope @xcite , respectively , or the determination of the position of the cherenkov ring @xcite .
coreas simulations @xcite , which include a realistic treatment of the refractive index of the atmosphere , are more sophisticated and more complete than simulations with the reas predecessor codes .
the remarkable agreement between coreas simulations and the measured slope of the radio lateral distribution @xcite makes studies based on coreas simulations even more promising than previous analyses .
an amplitude scale mismatch between the measured ( lopes ) and simulated ( coreas ) radio pulses still exists @xcite .
however , it does not affect the part of this analysis in which the slope of the radio lateral distribution , i.e. , the ratio of amplitudes at different distances , is considered .
it only has influence when absolute amplitude values are used , i.e. , in our case on the investigation of the energy reconstruction ( sec .
[ sec_energy ] ) , in which it will influence the calibration constants determined for the linear correlations between primary particle energy and radio amplitude .
taking the analysis described in @xcite as a guideline , we enhance and further improve the slope method . as a first step ,
we apply the slope method directly on pure coreas simulations , deriving important calibration parameters . in a second step ,
we apply the slope method on events measured with lopes to investigate the reconstruction of the total primary energy and the depth the shower maximum .
lopes @xcite is one of the pioneering experiments in the digital detection of radio emission from air showers . placed at karlsruhe institute of technology , germany , lopes benefits from the co - location with the particle detector experiment kascade - grande @xcite , which provides us with the reconstruction of fundamental air - shower parameters : arrival direction , shower core , and primary energy .
characteristic for the lopes analysis pipeline is the interferometric combination of the recorded electric field strength traces : according to the arrival direction of the air shower , the traces detected in each individual antenna are first shifted in time .
afterwards , a cross - correlation beam ( cc - beam ) and a power - beam , which gives the total power received in all the antennas , are computed @xcite .
the digital cross - correlation beam forming plays an important role also for the reconstruction of the lateral distribution of amplitudes .
indeed , both the cc - beam and the power - beam are crucial for the selection of events with a clear radio signal , and for the determination of the exact time of the radio pulse ( cc - beam time ) .
a hilbert envelope is applied to the up - sampled trace of each individual antenna , and the maximum instantaneous amplitude of the radio signal is defined by the local maximum of the hilbert envelope closest to the cc - beam time .
more details can be found in @xcite .
the events selected for the slope method analysis presented here have been measured with the lopes30 and lopespol setups @xcite .
the former consisted of 30 calibrated dipole antennas , all oriented in the east - west direction , while the latter used 15 east - west and 15 north - south - aligned antennas . for the purpose of this analysis , only data of the east - west oriented channels are taken into account , primarily due to the higher available statistics @xcite . the effective frequency band used for the analysis is 43 - 74 mhz .
the selected events have a primary energy between 10@xmath2 and 10@xmath3 ev , a zenith angle less then 40@xmath4 and a core position at a maximum distance of 90 m from the center of the kascade array - ( 0 m west - east , 0 m south - north ) in fig .
[ fig_map_selection ] .
these cuts are made to achieve high quality in the kascade air shower reconstruction .
a high signal - to - noise ratio and high coherence for the radio signal in the antennas is required as well by forcing the signal - to - noise ratio of the cc - beam , i.e. ( amplitude of the cross - correlation beam / rms of the cross - correlation beam @xmath5 ) , since not always all the 30 antennas participate in each event ] to be larger than 9 , and the fraction of correlated power in the antennas ( cc - beam / power - beam ) larger than 80 % .
further quality cuts demand a good fit for the lateral distribution function ( sec .
[ sec_ldf ] ) .
all cuts are chosen to achieve high standards in the kascade - grande data reconstruction and high quality radio measurements , yet simultaneously acquire good statistics .
205 events of almost 900 kascade - grande pre - selected showers remain after applying all cuts .
the coreas code is one of the latest developments of simulation codes reflecting an increasingly precise knowledge and an improved modelling of the radio emission from air showers @xcite . through the inclusion of a realistic treatment of the refractive index of the atmosphere ,
coreas simulations take into account cherenkov - like time - compression of the radio emission occurring for observers near the cherenkov angle .
this effect has been recognized to be the main reason for the previously measured flattening of the radio lateral distribution close to the shower axis @xcite .
the flattening effects depend on the geometrical distance of the shower maximum to the observer , and thus on both the zenith angle of the air shower and on the type of the primary particle .
a detailed comparison between coreas simulations and lopes measurements is presented in @xcite . for the following analysis , corsika @xcite and coreas
are used to simulate the particle cascade and the radio emission from the electromagnetic component of air showers , respectively .
qgsjet ii.03 @xcite and fluka @xcite are used as high energy and low energy interaction models , respectively .
specific criteria are applied to generate the coreas simulations to facilitate the final application of the slope method directly on lopes data . on the one hand ,
the specific individual events measured by lopes are simulated : the energy , the incoming direction and the core position of each lopes event , reconstructed from the measurements with the particle detector kascade(-grande ) @xcite , are used as input parameters for the simulations .
the simulated electric field is filtered with an ideal rectangular bandpass filter ( 43 - 74 mhz ) .
the geomagnetic field for the simulations is set to values valid for the lopes site ( 47 @xmath6 t magnetic field strength , 65@xmath4 inclination angle , negligible declination ) . with this procedure ,
the geometrical acceptance and important characteristics of the lopes experiment are taken into account .
each lopes event is simulated as both proton and iron primary . on the other hand ,
the observer positions in the simulations appropriately represent the antenna positions in the lopes array : for each individual event , the correct positions of the individual lopes antennas with respect to the shower core are taken into account .
for the selected events , the cores of the shower typically lie in between the antennas , thus the fit of the lateral distribution function ( ldf ) does not only involve observer positions in a specific azimuthal direction relative to the core . as a consequence ,
the charge excess effect which typically contributes to the total radio emission at a level of @xmath7 % @xcite in aera , and can be expected to be about half as important in lopes due to the stronger local magnetic field , is included in the following analysis .
we can thus judge the robustness of our method in light of the complex asymmetries present in the actual measured radio lateral distributions .
no pre - selection of typical showers ( showers with a typical x@xmath0 value ) is applied to the coreas simulations .
therefore , shower - to - shower fluctuations are included in the investigation for the data set as a whole .
the lateral distribution function ( ldf ) describes the measured ( or simulated ) electric field as a function of the observer position with respect to the shower axis . due to the superposition of geomagnetic and charge - excess radiation in radio emission from extensive air showers ,
the ldf can exhibit significant asymmetries @xcite .
ideally , therefore a two - dimensional ldf taking into account the observer azimuth angle in addition to the axis distance should be fitted to the measured radio data . in the case of lopes , however , the limited number of antennas sampling the true ldf led us to apply a simpler approach . instead of an asymmetric two - dimensional ldf
, we apply a one - dimensional ldf to our data , thereby effectively averaging out the asymmetries in the true ldf . depending on the coverage in the sampled antenna locations ,
a per - event amplitude shift can be introduced by this procedure .
this effect possibly leads to increased scatter in the fitted amplitudes , which is , however , implicitly accounted for in the following analysis .
another effect of the one - dimensional averaging could be an overall bias in the calibration of radio - amplitudes to primary particle energies determined by kascade . to quantify these two effects , we compared simulations including the charge - excess asymmetry with special simulations in which the charge - excess contribution has been removed .
the result of this dedicated study are : the use of a one - dimensional ldf increases the scatter observed in the amplitude - energy correlations by an amount that is negligible with respect to other uncertainties , in particular lopes experimental uncertainties . for the data set as a whole a systematic bias of 4.4% in the estimated average amplitudes arises , which we account for in the following energy correlation analysis by adopting an extra systematic uncertainty of 5% , well - below our global experimental calibration scale uncertainty of 35% .
the uncertainty of the measured electric field amplitudes we fit with our one - dimensional ldf is taken as the root sum of squared uncertainties due to the individual antenna calibration ( 5 % ) and due to the noise @xcite .
the distance of the antennas from the shower axis is calculated in the shower - plane coordinate system with a correlated uncertainty which depends on both the kascade precision ( @xmath8 4 m @xcite ) and the geometry reconstruction accuracy ( @xmath9 0.3@xmath4 ) .
in contrast , the coreas electric field at each observer position is assumed to be exact , thus no uncertainty is associated with the pulse amplitude .
the coreas simulations are able to reproduce the complexity of the measured radio ldf , including asymmetries and flattening due to cherenkov - like time compression effects . even at the distances probed by the lopes experiment ,
this complexity is clearly visible @xcite . since only events with a core position relatively close to the antenna array
are selected ( sec .
[ sec_evselection ] and fig .
[ fig_map_selection ] ) , the flattening of the radio lateral distribution close to the reconstructed core becomes an important aspect for our investigation . in previous lopes analyses ,
we have used an exponential function to fit data and simulations of lopes events @xcite . here , we prefer to use a one - dimensional gaussian function @xmath10 with @xmath11 [ @xmath6v / m / mhz ] , @xmath12 [ m ] , and @xmath13 [ m ] as free parameters to fit the lopes ldf . while the average lopes event can be fitted by either a one - dimensional exponential or a gaussian with similar quality , there are two advantages in using the gaussian . for one ,
there are some events for which the gaussian can fit the data well , while the exponential gives a considerably worse fit .
second , and more importantly , we develop our analysis method on the basis of coreas simulations , and for simulated events the gaussian fits considerably better than an exponential . we illustrate this with a comparison of a gaussian and an exponential fit to data and simulations in fig .
[ fig_lpldf_gfit ] .
we stress that the analysis presented here can be ( and has also been ) performed on the basis of an exponential ldf , and the results agree within the systematic uncertainties quoted here .
however , for the above reasons , we prefer to use a gaussian ldf .
besides the depth of the shower maximum ( x@xmath0 ) , the air - shower inclination influences the slope of the radio ldf .
this is related as well to a geometrical effect : near - vertical events constitute a radio source closer to the observer at ground level than inclined showers , and therefore are characterized by a steeper slope of the radio ldf .
five zenith angle bins are considered separately to reduce the dependence on the inclination and to focus on the effects connected to the depth of shower maximum of the showers .
these zenith angle bins are chosen such that they cover the same solid angle .
, simulated once as proton ( blue ) , once as iron ( light - red ) primaries .
the dotted line represents the distance d@xmath14 at which the rms - spread of the fitted normalized amplitudes is minimal.,scaledwidth=50.0% ]
we first investigate the slope method using the coreas simulations of the lopes events .
the procedure aims to compare the ldf of several events with different primary energies and arrival directions , thus normalizations for the radio amplitudes are required . with the well - founded assumption that the leading contribution to the radio emission from air showers is of geomagnetic origin ( @xmath15 85 % ) @xcite , the east - west channel amplitudes for each given event are corrected for @xmath16 , the east - west component of the lorentz vector ( @xmath17 ) , with @xmath18 the arrival direction and @xmath19 the geomagnetic field vector .
( due to detection threshold effects , there are no events in the lopes data set with a shower axis parallel to the magnetic field ; the smallest geomagnetic angle is 15@xmath20 , corresponding to a @xmath21 of 0.26 , whereas the mean @xmath21 is 0.7 . ) in addition , we normalize the electric field amplitudes by the monte carlo true energy of each event , which has been adopted as the energy reconstructed by kascade(-grande ) for the underlying lopes event .
the normalized lateral distribution function fits for the events in the first zenith angle bin ( [email protected]@xmath4 ) are presented here as an example ( fig .
[ fig_ldfe ] ) .
this figure already illustrates that the spread of the obtained radio ldf fits drastically varies as a function of the distance from the shower axis .
the spread is quantified in fig .
[ fig_rmsldfall ] , where the profile of the relative rms - spread for the ldf fits is shown for each of the zenith angle bins .
the spread reaches from @xmath22 % at distances of 70 - 100 m from the shower axis up to more than 50 % at distances larger than 300 m. simulations @xcite previously predicted the existence of a peculiar distance ( d@xmath14 ) where the radio amplitudes do not carry information on x@xmath0 and the influence of shower - to - shower fluctuations is minimal . also here
, we observe the minimum spread of the ldf fits at a peculiar distance d@xmath14 .
the value of d@xmath14 as well as the value of the minimum rms - spread vary with the zenith angle of the air shower and are summarized in table [ tab_@xmath23 region ] . for lopes events
, d@xmath14 lies in the range 70 - 100 m from the shower axis . [
cols="^,^,^,^,^",options="header " , ] the feature of the radio lateral distribution function which correlates best with the depth of shower maximum is the slope of the ldf fit .
we first `` calibrate '' the relation of this slope with the depth of shower maximum on the basis of coreas simulations .
afterwards , we apply the found relation to lopes measurements .
we define the ldf slope @xmath24 as the ratio of the radio amplitudes taken at the distance d@xmath14 and at the fixed distance of 200 m ( @xmath24=@xmath25/@xmath26 ) .
the latter value is chosen for the specific investigation on the lopes events , since 200 m is a relatively large distance but is still in the range of the measurements .
the x@xmath0 values for each event are taken directly from the monte carlo simulations , and are considered as the `` true '' values .
the ratio of the radio amplitudes and the depth of the shower maximum are plotted against each other in fig .
[ fig_xmaxfit ] , for zenith angles less than 19.4@xmath4 .
there is a clear correlation which we fit with eq .
[ xm_eq ] .
@xmath27 \label{xm_eq}\ ] ] this ( more precisely the inverse ) three - parameter function was already introduced in @xcite and successfully applied to previous analyses @xcite .
the function ( eq . [ xm_eq ] ) appropriately describes the correlation for the events simulated with coreas , and it is adequate for all of the zenith bins .
the @xmath28 , @xmath29 and @xmath30 parameters , including the statistical uncertainties of the fit , are extracted separately for each zenith angle bin ( table [ xmax_sel2 ] ) , and the corresponding fit results are shown in fig .
[ fig_xmaxfits ] .
this function has not been optimized for minimal correlation between fit parameters , which can lead to comparably good fits with more than one set of parameters , visible as fluctuations in the derived parameter values in table [ xmax_sel2 ] .
this is fine as long as the resulting fits describe the distribution of data points correctly , which they do .
however , in the future it would be worth investigating better fit functions with minimally correlated fit parameters , ideally parameters that can be associated directly with physical parameters of the underlying air shower . to improve the determination of the three fitting parameters and to better constrain the fit over a wide range of x@xmath0 values , increased ( doubled ) statistics for the proton- and
iron - simulated showers are used .
we performed the fit procedure for all zenith angle bins both without and with associating an uncertainty propagated from the ldf fits to the individual data points , and the results vary only insignificantly . propagating the resulting fit uncertainties into an uncertainty for the reconstructed x@xmath0 values
can not be achieved by gaussian error propagation due to the non - linear nature of the problem .
one option to proceed would be to employ an involved procedure for the estimation of the uncertainty of the reconstructed x@xmath0 values , e.g. by bootstrapping methods . as we want to keep our analysis practical and
are anyway limited by the lack of an independent x@xmath0 determination to compare the reconstructed values with , however , we choose instead to use the resulting spread in the distribution around the fit as an estimate for the uncertainty of the reconstructed x@xmath0 values . in other words ,
we calculate the typical x@xmath0 precision of the slope method as the spread of the difference between the true monte carlo x@xmath0 values and the x@xmath0 values reconstructed using the slope method on coreas simulations . fig .
[ fig_truevsrecxmax ] shows the corresponding distributions for proton showers ( rms - spread 58 g/@xmath1 ) and iron showers ( rms - spread 51 g/@xmath1 ) for the complete selection of events . in table
[ xmax_sel2 ] , the combined spread for proton and iron showers is denoted as @xmath31x@xmath0 for the individual zenith angle bins .
for the coreas simulated events with doubled statistics , and considering the complete zenith angle range , the rms - spread is in the range of 45 - 65 g/@xmath1 .
it is important to note that the results presented here apply to the specific situation of lopes , in particular the altitude of the lopes site , the specific observing frequency band used @xcite , the energy range probed , and the dimension of the antenna array .
these parameters all affect the amplitude ratio and consequently play a role in the determination of the x@xmath0 calibration curves . indeed , for inclined showers
, larger amplitude ratios may be obtained when the far - away distance used to calculate the ratio is taken well beyond the 200 m distance used for this analysis and imposed , here , by the dimension of the lopes array .
the three - parameters function ( eq . [ xm_eq ] ) is now applied to reconstruct the x@xmath0 values from lopes measurements , using the coreas - derived @xmath28 , @xmath29 and @xmath30 parameters listed in table [ xmax_sel2 ] for the appropriate zenith angle bin of each individual event . in fig .
[ fig_xmax ] , we show the resulting distribution of x@xmath32 values in the black histogram , with 633.2 @xmath33 94.6 g/@xmath1 as the mean and the standard deviation of the distribution .
a comparison with the coreas proton ( blue dotted line ) and iron ( red line ) prediction is shown as well : x@xmath34
= 678.5 @xmath33 76.9 g/@xmath1 and x@xmath35 = 604.2 @xmath33 56.4 g/@xmath1 .
these values are derived by applying eq .
[ xm_eq ] to the ( coreas ) simulated events , or , in other words these are the values on the fit in fig .
[ fig_xmaxfit ] of the corresponding simulated ldf slopes .
the standard deviation value for the lopes distribution of @xmath895 g/@xmath1 can be considered itself as an upper - limit for the lopes precision on x@xmath0 , although it does not explicitly account for systematic uncertainties , e.g. , scale uncertainties due to modelling the atmosphere or due to hadronic interaction models used for the shower simulations .
the reconstructed values for x@xmath36 are compatible with the expectations . compared to previous analyses @xcite ,
in which the simulations used did not include the refractive index of the atmosphere , a much improved result is now obtained .
we can definitively affirm that the systematic shift observed previously in @xcite was caused by the discrepancy between the ldf slopes predicted by older simulation codes and the measured ldf slopes .
in addition , we examine the energy dependence of the x@xmath0 distribution ( fig .
[ fig_xmaxlge ] ) , integrated over the whole zenith angle range : each cross represents the mean value and the standard deviation of the x@xmath0 distributions , for the slope method applied to lopes measurements and to coreas simulations .
additionally , the mean and the error of the mean ( standard deviation/@xmath37 ) for the true monte carlo x@xmath0 are shown in each energy bin .
good consistency with the expectations is achieved in all energy bins .
next , we directly compare the mean lopes x@xmath0 values reconstructed for the four energy bins with the results of other experiments ( fig .
[ fig_xmspectra ] ) . in the figure ,
the solid lines represent the monte carlo expectations for qgsjetii-03 for pure iron - simulated ( red ) and pure proton - simulated showers .
the uncertainty indicated in this figure on the lopes average x@xmath0 values is the statistical uncertainty , i.e. standard deviation/@xmath37 .
the intrinsic uncertainty of the slope method , as determined by the spread between the reconstructed and monte carlo true values , is indicated as a shaded band .
no additional analysis on the systematics is carried out . in summary ,
the results of the lopes x@xmath0 determination are self - consistent and plausible .
nevertheless , we point out that the lopes uncertainties are too large for any conclusion concerning the mass composition .
in the following discussion , willful and imposed limitations which had influence on the slope method analysis of the lopes simulated and measured events are pointed out . the upper limit on the observed energy precision ( @xmath8 20 - 25 % ) includes the convolution of both the lopes and kascade - grande ( particle detector ) uncertainties . the good compatibility with the kascade - grande statistical uncertainty ( approx . 20 % )
indicates that the combined uncertainty of the energy is strongly influenced by kascade - grande and that the energy resolution of lopes is likely significantly better than 20 - 25 % .
the total energy precision of the slope method suggested by the pure coreas simulations is approximately 13 % and could potentially be achieved by applying the method on low - noise radio data .
it is worth stressing once again that radio emission is purely sensitive to the electromagnetic component of air showers .
if combined with other detection techniques which are able to provide information on other components , in particular the muonic one , radio detection can help to disentangle the individual components , which in itself opens access to composition - sensitive information .
moreover , the radio signals are not critically affected by uncertainties in the hadronic interaction models used to understand the data . one restriction of the analysis presented here concerns the deliberate exclusion of the detector - simulation in the investigation .
a more realistic comparison with the measurements may include both simulations of detector properties as well as simulated continuous and transient noise typical of the specific experimental site .
also , we did not try to apply a more sophisticated normalization of the radio field strengths including a charge - excess component . since we fit a one - dimensional ldf with observers at various azimuth angles , the charge - excess asymmetry mostly averages out in the ldf fit .
we confirmed with a dedicated simulation study that the additional spread introduced by this simplified treatment is negligible and that there is a bias in the reconstructed @xmath38-parameters of less than 5% , which is much smaller than the overall calibration scale uncertainty of 35% and for which we account with an additional systematic uncertainty of 5% . for experiments with better data quality than lopes ,
however , methods should be investigated to try and correct individual measured amplitudes by the presumed charge - excess contribution or fit two - dimenaional ldf functions to improve the fit and thus possibly the energy determination .
this aspect was not investigated since the precision we achieve in the combined analysis anyway seems to be limited by the reference energy provided by kascade - grande and not by our method .
the qgsjet ii.03 interaction model was employed to generate the coreas simulations of the lopes events .
other interaction models are available and they differ both in their implementation and their predictions .
we used the slope method to show how air shower parameters such as energy and x@xmath0 can be directly derived from the features of the radio ldf . regarding the energy reconstruction ,
different interaction models predict different fractions of the energy going into the electromagnetic component and other components , respectively .
this might lead to more significant shifts between the radio amplitudes observed for different primaries of the same energy than illustrated in this analysis .
such shifts , however , are not a disadvantage of radio detection , since in fact radio emission with its sensitivity to the pure electromagnetic component of the air shower could be used to study such effects and test hadronic interaction models . regarding the shower evolution
, different interaction models will make different predictions for the x@xmath0 values of air showers initiated by specific primaries with given energies .
however , the obtained `` calibration '' relating a given value of x@xmath0 to a corresponding radio ldf slope should not be strongly affected by the choice of interaction model .
this is because the slope method exploits an intrinsic property of the radio emission : the relation between the geometrical distance of the radio source and the slope of the radio lateral distribution arising from the forward - beaming of the emission .
a slight influence on the calibration curve can be expected by variations of the atmosphere , which relate the geometrical distance to the depth of shower maximum in g/@xmath1 . within the precision achievable by lopes ,
however , these effects can be neglected .
the x@xmath0 resolution obtained in this analysis ( @xmath895 g/@xmath1 ) is an upper - limit attained with ( lopes ) radio data affected by strong environmental noise .
much better results may be achieved by using higher quality radio measurements and a denser antenna array .
such is the case for lofar @xcite , which has recently applied a method based on global fits of the two - dimensional radio ldf and obtained a sensitivity in x@xmath0 comparable with the fluorescence technique ( @xmath820 g/@xmath1 ) .
the underlying principle is the same , namely the exploitation of the geometrical connection between the source distance and the radio ldf described above . in the analysis presented here , we strove to exploit the potential of the lopes data while at the same time keeping the analysis practical . existing and upcoming more modern experiments such as aera @xcite , tunka - rex @xcite and lofar @xcite will have higher - quality data with lower background noise , a larger number of antennas per event and a larger range of lateral distances probed with each event .
the methods presented here will have to be refined to achieve the maximum precision these instruments will be able to deliver in energy and x@xmath0 resolution . in particular , a lateral distribution function accounting for the asymmetries induced by the charge - excess should be used .
the function used to fit the x@xmath0 values as a function of amplitude ratios should be re - investigated and correlations between fit parameters should be minimized .
finally , a more sophisticated determination of the reconstruction uncertainties , taking into account the non - gaussian error propagation should be applied . in spite of these limitations
, we have demonstrated that a determination of the energy and x@xmath0 from the radio lateral distribution is feasible with promising upper limits on the achievable precision .
the purpose of this analysis was the reconstruction of the energy and depth of shower maximum of individual cosmic rays from radio measurements performed with the lopes experiment .
we achieved this goal by exploiting characteristic features of the lateral distribution function of the radio emission , in particular its slope .
we have developed and applied a coreas - calibrated method to the measurements of the lopes experiment .
significant results emerged : * a characteristic distance from the shower axis , i.e. d@xmath14 ( 70100 m ) , is predicted by coreas simulations of the lopes events to be the appropriate place for the primary energy investigation with radio data .
its existence is also indicated in lopes measurements . *
a simple linear correlation between the radio amplitude measured at the distance d@xmath14 and the kascade - grande reconstructed energy was found .
a zenith - angle dependent deviation in the correlation parameters remains between coreas simulations and lopes data ; this discrepancy needs to be investigated in a dedicated analysis . *
an upper - limit for the lopes precision on the total energy reconstruction was determined from the combined lopes
kascade - grande energy uncertainty ( @xmath820 - 25% depending on zenith angle ) .
the intrinsic energy resolution of the radio measurements suggested by the simulations is approximately @xmath39 , and might even be improved by taking the askaryan effect explicitly into accout in the ldf fits .
however , any further improvment is beyond the testing power of lopes . * the slope of the lopes lateral distribution function
was exploited to reconstruct x@xmath0 from lopes data .
the resulting distributions are plausible and comparable with expectations from cosmic ray nuclei . * by itself , and for the specific situation of lopes , the slope method predicted a precision of x@xmath0 around 50 g/@xmath1 , depending on the zenith angle of the event . from the analysis of the lopes measurements , an upper limit of @xmath895 g/@xmath1 on the precision was found .
x@xmath0 is the principle indicator for cosmic ray composition , and we have demonstrated the possibility to reconstruct it with the slope method applied to lopes radio measurements .
the importance of this result is underlined by the fact that radio detection does not suffer from low duty cycles as is the case for optical detection techniques .
comparisons of the reconstructed x@xmath0 values between , at least , two different detection methods are essential but not possible in the framework of the lopes and kascade - grande experiments .
this opportunity is offered by the new generation of antenna arrays , such as aera @xcite ( at the pierre auger observatory ) and tunka - rex @xcite , where the radio reconstruction can be cross checked with the fluorescence and cherenkov experimental results , respectively .
lopes and kascade - grande have been supported by the german federal ministry of education and research .
kascade - grande is partly supported by the miur and inaf of italy , the polish ministry of science and higher education and by the romanian authority for scientific research uefiscdi , idei , grant 271/2011 .
this research has been supported by grant number vh - ng-413 of the helmholtz association . 00 , , , ( ) , .
, , , , , ( ) , . , , , , ( ) , , , , ( ) , , , , ( ) , , , , ( ) , , , ( ) , , , , , ( ) , , , , , , ( ) , , , , ( ) , , , , ( ) , , , , ( ) , , , , ( ) , , , , , ( ) , , , , ( ) , , , , ( ) , , , , ph.d .
thesis , rheinische friedrich - wilhelms - universitt bonn , germany , , .
, , , in : , volume , pp . .
, , , ( ) , , | lopes is a digital radio interferometer located at karlsruhe institute of technology ( kit ) , germany , which measures radio emission from extensive air showers at mhz frequencies in coincidence with kascade - grande . in this article
, we explore a method ( slope method ) which leverages the slope of the measured radio lateral distribution to reconstruct crucial attributes of primary cosmic rays .
first , we present an investigation of the method on the basis of pure simulations .
second , we directly apply the slope method to lopes measurements . applying the slope method to simulations ,
we obtain uncertainties on the reconstruction of energy and depth of shower maximum ( x@xmath0 ) of 13 % and 50 g/@xmath1 , respectively . applying it to lopes measurements , we are able to reconstruct energy and x@xmath0 of individual events with upper limits on the precision of 20 - 25 % for the primary energy and 95 g/@xmath1 for x@xmath0 , despite strong human - made noise at the lopes site . cosmic rays , extensive air showers , radio emission , lopes , x@xmath0,energy |
the motivation to study the metal - insulator transition in the s - d exchange , or vonsovsky - zener , model @xcite ( now it is frequently called kondo lattice model ) is given by previous successes obtained studying the metal - insulator transition ( mit ) in the hubbard model @xcite .
the s - d exchange model is complementary to the hubbard model . while in the hubbard model the interacting electrons responsible for magnetism and for electron conductivity
are located in the same band , in the s - d exchange model they are located in different bands . in one band
the electrons are delocalized or itinerant , therefore these electrons are called conduction electrons .
the other band contains electrons that are localized on lattice sites .
the type of the interaction is also different between the two models .
the hubbard model contains an on - site coulomb interaction modeled by a single energy parameter @xmath0 .
the s - d exchange model contains a spin - spin interaction between the localized and conduction electrons with a coupling constant ( s - d exchange parameter ) @xmath1 .
the s - d exchange model is used in the description of rare earths , in which 4f - electrons have a very localized nature @xcite .
the description of colossal magnetoresistance materials such as ( la , ca , sr)mno compounds sometimes is based on the s - d exchange model assuming that the @xmath2 electrons are rather localized and the @xmath3 electrons are itinerant @xcite .
whereas for relatively weakly correlated systems the crystal with half - filled energy bands is a metal , for strong enough correlation the system should be an insulator , with each electron sitting in a localized state associated with a given site .
this corresponds to the splitting of the energy bands into hubbard subbands separated ( for a large enough interaction parameter @xmath0 ) by a gap @xcite . in the s - d exchange model
the parameter @xmath0 is replaced by @xmath4 where @xmath5 is the spin of localized electrons .
the hubbard - iii approximation @xcite provides a qualitative description of the metal - insulator transition for both models .
this description however does not take into account all quantum effects ; for the case of s - d exchange model , it turns out to correspond to the classical spin limit @xmath6 @xcite . the dynamical mean field theory ( dmft )
@xcite was a real breakthrough in this respect dramatically improving the description of metallic phase near the transition .
whereas within the hubbard - iii approximation the main effect of approaching the mit is the growth of damping of electron states at the fermi energy , together with appearance of shoulders as the precursors of future hubbard bands ( two - peak structure ) , the dmft gives the _ three_-peak structure , with the `` kondo peak '' near the fermi energy , that is , the growth of effective mass instead of the growth of the damping .
the origin of this peak is the formation of a resonance due to electron scattering by _ quantum _ spins @xcite .
the dynamical mean - field theory is formally exact in the limit of infinite space dimensionality whereas the hubbard - iii approximation is formally exact ( within the s - d exchange model ) in the limit of infinite space dimensionality _ and _ infinite , that is , classical , spin @xcite .
in particular , in this limit there is no difference between the case of ferromagnetic ( @xmath7 ) and antiferromagnetic ( @xmath8 ) s - d exchange coupling ; the kondo resonance exists only in the latter case ( the detailed discussion of the kondo effect and related physics can be found in ref . ) .
the dmft is the best local approximation , that is , it corresponds to an optimal self - consistent choice of the momentum - independent electron self - energy . for real two- and three - dimensional systems
it is just an approximation , which is not quite controllable .
the dual fermion formalism @xcite allows to construct a diagrammatic expansion for the non - local correlation effects starting from the dmft green s function as a zeroth - order approximation .
the differences between the dual fermion results and the dmft results were investigated recently @xcite for the hubbard model on triangular lattice , and turned out to be very essential .
the present paper discusses the differences between dual fermions and dmft for the metal - insulator transition in the s - d exchange model .
the mit in the s - d exchange model in the classical limit @xmath6 has been studied in ref .. in this work we consider a _
s - d exchange model at a two - dimensional ( 2d ) square lattice , for @xmath9 where we expect quantum effects to be maximal .
we study the role of non - local correlation effects by comparison of the results of dmft and dual fermion calculations .
we show that the dmft strongly overestimates the critical value of the coupling constant for the metal - insulator transition @xmath10 .
we also show that in the s - d exchange model quantum effects are indeed relevant , since the results for positive and negative values of the coupling constant @xmath1 are essentially different .
the s - d exchange model contains electrons that are localized on lattice sites , and conduction electrons that can hop between different sites @xmath11 .
let us introduce their annihilation operators @xmath12 , @xmath13 is the spin projection .
the annihilation operators for the localized electrons will be designated @xmath14 .
the hamiltonian contains ( in order of appearance in eq . ) a hopping term , a hubbard term that is included for calculational purposes , the s - d exchange interaction , and the chemical potential @xmath15 term : @xmath16 @xmath17 is the number of conduction electrons and @xmath18 .
the spin - spin interaction term reads : @xmath19 the hubbard interaction is included to ( computationally ) ensure the half - filling of the lattice sites ( one localized electron per site ) . strictly speaking
, it is reached only in the limit @xmath20 .
we will do the calculations for a large but finite @xmath0 . with this hubbard - term and @xmath21 and @xmath22 half - filling
is achieved to sufficient extent .
our goal was to study the density of conduction - electron states ( dos ) of the s - d exchange model at different coupling constants @xmath1 .
the calculations were carried out at the matsubara axis ( imaginary frequencies ) and , to obtain the densities of states we performed an analytical continuation of the green s functions on the real axis , using the maximum entropy method @xcite .
the green s functions we obtained self - consistently in dmft versus the dual fermion approach . in short ,
both dmft and dual fermion utilize an effective - medium concept , so that they deal with the impurity model described by the action @xmath23 , where @xmath24 is the action of an isolated lattice site , and @xmath25 is a hybridization by the effective bath .
frequency - dependence of the hybridization allows to handle the effects of local fluctuations , that underlie the kondo physics . on the other hand
, this @xmath26-dependence makes the impurity problem non - hamiltonian and therefore requires a usage of numerical solvers to obtain its green function @xmath27 and two - particle vertex @xmath28 ( the later is required for dual fermions ) . as a numerical solver
, we used the interaction - expansion continuous - time quantum monte carlo procedure @xcite .
the dual fermion lattice green s function equals ( spin and orbital indices are omitted ) @xmath29 where @xmath30 is a dispersion law arising from the hopping @xmath31 , and @xmath32 is the self - energy function of the dual ensemble obtained in certain diagrammatic approximation .
vertices of dual diagrams are vertex parts of the impurity problem .
dmft can be represented as the zeroth - order dual fermion scheme with @xmath33 . in this approximation ,
the lines ( propagators ) in dual diagrams have a simple meaning , as they are equal to the non - local part of the green s function , @xmath34 ( for a general case , @xmath35 and @xmath36 are also exactly related , although the relationship is more complicated ) .
one can easily see that dmft describes purely local contribution to the lattice self - energy ( and is actually shown to correspond to the summation of all local diagrams ) , whereas dual fermion diagrams describe non - local effects .
more technical detail can be found in refs .
@xcite .
the hybridization function @xmath37 was obtained self - consistently from the condition that the local part of the dual propagator vanishes , @xmath38 .
this eliminates the first - order hartree dual diagram , and thus the leading correction to @xmath32 is given by the second - order ( see diagram ( b ) and formula ( 20 ) of ref.@xcite ) , @xmath39 ( summation over spins is also implied ) .
the last expression is an approximation used for the dual fermion calculations in the present paper .
the calculations were performed for the paramagnetic state of the hamiltonian ( [ h ] ) on a square lattice with the nearest - neighbor hopping parameter @xmath40 ( which means a half - bandwidth equal to 1 ) , we have chosen @xmath41 and the inverse temperature @xmath42 .
dual fermion diagrams were calculated at a @xmath43 slab in momentum space .
the final accurate calculations took about two weeks of cpu load per @xmath1-value in the dmft case and about 10% more in the dual fermion case .
the amount of self - consistency loops was roughly 10 - 20 while the number of monte carlo steps was roughly doubled with each loop in the last 6 loops until it reached @xmath44 .
the mit for the s - d exchange model is visible in fig .
[ jcrit ] . we have quoted values of the coupling constant just around the transition .
although the temperature is quite high and the hubbard-@xmath41 we still see a qualitatively correct picture of the densities of states , accurate enough to estimate a reasonable value of the coupling constant @xmath1 for the metal - insulator transition .
we consider the state to be conducting if there is a local maximum at the fermi level and insulating if there is a local minimum at the fermi level , because the temperature is so high .
the choice of @xmath41 in the quantum monte carlo calculations is justified because in our calculations @xmath45 and @xmath46 so that half - filling is achieved to sufficient extent . [ cols="<,<,<",options="header " , ] in table [ jcrit_table ] we compare several different calculational methods that were used to obtain a critical value of the coupling constant for the metal - insulator transition .
the difference in behavior of the dos between positive and negative values of the coupling constant is displayed in figures [ jsmall ] and [ jlarge ] in figure [ jsmall ] we display the results for @xmath47 calculated within both dmft and dual fermions .
the different colors represent positive and negative @xmath1-values .
stars correspond as before to the dmft - calculation while the dual fermion result is drawn with lines . in this picture
we look at the general shape of the curves .
we see that they are not the same for positive and negative @xmath1 and that incorporating non - local corrections through dual fermions changes the results : the difference is more pronounced in the dual fermion case .
in figure [ jlarge ] we display the results for @xmath48 , also calculated using dmft ( stars ) and dual fermions ( lines ) .
the different colors again represent the positive and negative @xmath1-values .
the general shape of the curves is much more different between positive and negative @xmath1 for @xmath48 than for @xmath47 .
in this work we justify two claims .
first , the dual fermion formalism description of the metal - insulator transition in the s - d exchange model to date predicts a critical value of the coupling constant @xmath49 .
this value differs from the dmft - result by a factor of more than 2 and corresponds nicely to the result obtained in the other work using a non - local correlations approach @xcite .
the different results between various calculational schemes are summarized in table [ jcrit_table ] .
second , there is a clear qualitative difference between densities of states of the s - d exchange model for antiferromagnetic and ferromagnetic s - d exchange coupling .
this shows that incorporating quantum effects such as the kondo effect is essential for a good description of the model . | the metal - insulator transition within the s - d exchange model is studied by the dynamical mean field theory and dual fermion approaches .
the latter takes into account nonlocal correlation effects which are shown to be essential .
in particular , the critical values of the s - d exchange coupling constant for these two methods turn out to be different by a factor of more than two ( for the case of square lattice and spin 1/2 localized electrons ) .
the calculations were performed using a continuous - time quantum monte carlo method .
the difference of the quantum spin-1/2 case and the classical - spin case is discussed . for the quantum case
the sign of s - d exchange coupling constant is relevant which demonstrates an importance of the kondo effect . |
one of the most surprising and exciting possibilities for new physics at future colliders is the production of microscopic black holes @xcite , which can occur in theories with large extra dimensions @xcite .
string theory suggests a scenario in which the standard model particles are confined to a submanifold with the familiar three spatial dimensions ( the ` brane ' ) , while gravity is free to propagate in the full @xmath0-dimensional space ( the ` bulk ' ) . in that case
the fundamental @xmath1-dimensional planck scale could be much lower than its effective 4-dimensional value , possibly of the order of tev .
then gravity would look weak on distance scales large compared to the size of the extra dimensions , but could become as strong as the other forces at short distances .
parton - parton collisions with centre - of - mass energies well above this scale could be treated as @xmath1-dimensional classical gravitational interactions .
numerical studies suggest that in such circumstances the probability of coalescence to form a microscopic black hole would be large @xcite
. the black hole would be a powerful source of hawking radiation , decaying rapidly into all kinds of standard model particles . from a study of the decay products
, it might be possible to deduce the number and size of the extra dimensions .
in this paper i shall review some recent work on the above scenario .
we shall see that , even if one accepts the basic assumption of tev - scale gravity in extra dimensions , there are many uncertainties and gaps in our understanding , which make reliable predictions impossible at present .
nevertheless one can build models and simulations that allow one to explore different options for filling in the gaps , and enable experimentalists to start thinking about how they might analyse black hole events .
i apologise for not mentioning all relevant topics and papers in this short review . for more complete discussion and references
, please see @xcite and chris harris s thesis @xcite .
on purely dimensional grounds we expect the parton - level cross section for black hole formation to be of the form @xmath2 where @xmath3 is the schwarzschild radius in @xmath1 dimensions , @xmath4^{\frac 1{n+1}}\ ] ] and @xmath5 is a `` formation factor '' of order unity .
notice that we assume here that the black hole mass @xmath6 is equal to the full parton centre - of - mass energy @xmath7 ; this point will be discussed further later .
we use the dimopoulos
landsberg @xcite definition of the planck mass , @xmath8^{-\frac{1}{n+2}}\ ] ] where @xmath9 is the @xmath1-dimensional newton constant . for illustrative purposes
, we shall usually take @xmath10 tev in this study .
the formation factor @xmath5 in equation [ fndef ] has been estimated in a variety of ways . a simple geometric argument @xcite goes as follows .
consider incoming partons that would pass each other at a separation ( impact parameter ) @xmath11 where @xmath12 is the horizon radius for a kerr black hole with the corresponding angular momentum @xmath13 and mass @xmath14 .
this is assumed to be the maximum impact parameter at which black hole formation could occur .
therefore @xmath15 but for a kerr black hole in @xmath1 dimensions @xmath16^{-\frac 1{n+1}}\ ] ] where @xmath17 and so we obtain @xmath18^{-\frac 2{n+1 } } \qquad\mbox{(``geometric'').}\ ] ] this formula , shown by the blue curve in figure [ fig : bhff ] , follows quite closely the numerical estimate of yoshino and nambu @xcite ( green ) .
on the other hand , the latter is only a lower bound on the cross section , obtained by finding a closed trapped surface on a particular slice of @xmath1-dimensional space - time @xcite ; such a surface must be shielded by an event horizon . in this way
one obtains a lower bound on the impact parameter for horizon formation , and hence on the cross section for black hole formation .
more recently , yoshino and rychkov @xcite have found a more optimal space - time slice , which leads to a larger lower bound on the formation factor , shown in red in figure [ fig : bhff ] .
, scaledwidth=65.0% ] adopting the yoshino - rychkov lower bound as an estimate of the black hole formation factor , and assuming again that @xmath19 , we obtain the parton - level cross section , shown as a function of the planck scale for a 5 tev black hole in figure [ fig : bhcs ] . for @xmath20 extra dimensions
, the @xmath21-dependence of the formation factor tends to cancel that of the schwarzschild radius , so that the cross section is not strongly dependent on the number of extra dimensions .
values of @xmath21 less than 3 are in any case strongly disfavoured on astrophysical grounds @xcite . ,
scaledwidth=65.0% ] , scaledwidth=65.0% ] to estimate the cross section for black hole production at a hadron collider , we must convolve the parton - level cross section with the parton distributions in the incident hadrons .
the resulting cross section for a @xmath22 collider at c.m .
energy 14 tev ( i.e. the lhc ) is shown in figure [ fig : bhlhc ] . at the lhc design luminosity of @xmath23 @xmath24s@xmath25 ,
this corresponds to more than one black hole per second with mass above 5 tev .
if we accept the yoshino - rychkov result as a reliable estimate of the black hole formation factor , we see from figures [ fig : bhcs ] and [ fig : bhlhc ] that a measurement of the cross section for a given range of black hole masses would fix the planck mass in a way that is substantially independent of the number of extra dimensions , at least in the astrophysically favoured region @xmath20 . of course , one expects some dramatic changes in the cross section and final state at partonic c.m .
energies around the planck scale , due to the onset of strong gravitational scattering . however , to predict those changes one would need a quantum theory of gravity , so deducing the planck mass from them would not be straightforward .
measurements well above the planck scale , on the other hand , can reasonably be interpreted in a classical approximation , as we are doing here .
the problem in any case is to make a reliable measurement of the black hole mass , or more correctly the partonic c.m .
energy for black hole formation .
since we do not observe the colliding partons , this can only be inferred from properties of the final state , which will be dominated by the decay of the black hole .
although the formation of a horizon in parton collisions well above the planck scale seems reliably established , the nature and fate of the object thus created is much less clear .
the usual working hypothesis has been that the evolution of the system has four phases : * _ balding phase _ : all ` hair ' ( characteristics other than mass , charge and angular momentum ) and multipole moments are lost through gravitational and hawking radiation , and the object becomes the multidimensional generalization of a kerr - newman black hole .
in fact any residual charge after this phase is probably negligible @xcite , so the kerr solution is assumed . * _ spin - down phase _ :
the kerr black hole loses angular momentum by hawking radiation and becomes a schwarzschild black hole . * _ schwarzschild phase _ :
the black hole loses mass through hawking radiation and its temperature rises until the mass and/or temperature reach the planck scale . *
_ planck phase _ : the object ( ` string ball ' ? ) is in the realm of quantum gravity and its fate can not be predicted .
it could decay into a few quanta with planck - scale energies @xcite , evaporate at the hagedorn temperature @xcite , or even form a new kind of stable relic object @xcite . at present , each of these decay phases is subject to great uncertainty .
the amount of gravitational radiation emitted in the balding phase is of major concern because this constitutes missing energy that would spoil the connection between collision energy and black hole mass @xcite . in higher dimensions
there are solutions to the einstein equation that are not simply generalizations of the four - dimensional solutions , such as black rings @xcite .
probably there are more complicated objects still to be discovered .
it is not clear whether such configurations would be able to spin down to the schwarzschild solution , or what their hawking radiation would look like . even assuming the generalized kerr solution , the amount and distribution of hawking radiation during spin - down is still under investigation @xcite . in the schwarzschild phase
too there are many points that require further clarification .
what fraction of the hawking radiation is emitted as detectable standard model particles on the brane , and how much escapes into the bulk ?
is the decay process too rapid for the relationship between black hole mass and temperature to remain valid throughout ?
do secondary decays significantly distort the hawking spectrum ? and in the final planck phase , what are the consequences of alternative models for the fate of the remnant object ?
many of these questions await theoretical answers , but some can be illuminated by numerical simulations . for the latter , we shall start with the hypotheses that practically all the energy of the parton collision goes into the black hole , that the decay is dominated by the schwarzschild phase , and that the hawking radiation consists entirely of standard model particles on the brane . with these assumptions , a detailed picture of the final state can be presented and the effects of some of the uncertainties can be investigated . with the above assumptions , the spectrum of particles emitted during black hole decay takes the form @xmath26 where as usual the @xmath27 applies to bosons and fermions , @xmath28 is the hawking temperature @xmath29 and @xmath30 is a @xmath1-dimensional _ grey - body factor _ @xcite . the latter takes account of the fact that the wave function of a particle created in the intense gravitational field near the horizon has to propagate through the curved space around the black hole in order for the particle to be observed . by detailed balance ,
the grey - body factor is equal to the absorption coefficient of the black hole for waves incident from infinity . for particles emitted at high energies
the wavenumber is large compared with the curvature , so grey - body effects are small and the spectrum is close to black - body .
but at low energy the emission is strongly modified in a way that depends on the spin and the number of extra dimensions .
figures [ grey0][grey1 ] , taken from @xcite , show the grey - body factors for scalars , fermions and gauge bosons as functions of the particle energy in units of the inverse horizon radius .
what is actually plotted here is the absorption cross section in units of @xmath31 , which tends to 4 at high energies .
we see that the main effect in extra dimensions is the suppression of low - energy gauge boson emission .
[ ] [ ] [ 2.5]@xmath32 [ ] [ ] [ 2.5]@xmath33 [ ] [ ] [ 2.5]@xmath32 [ ] [ ] [ 2.5]@xmath34 [ ] [ ] [ 2.5]@xmath32 [ ] [ ] [ 2.5]@xmath35 integrating the spectrum in equation 8 , including the grey - body factors , and multiplying by the number of standard model degrees of freedom for each spin , one obtains @xcite the total flux of particles of each type in black hole decay .
[ ] [ ] [ 2.5]@xmath21 [ ] [ ] [ 2.5]@xmath36 expressed in units of the inverse horizon radius , as shown in figure [ flux ] , the particle fluxes are independent of the black hole mass and planck scale . when the flux in these units exceeds unity , which we see is the case for @xmath20 , the time between emissions is less than the time for a light - signal to travel a distance equal to the horizon radius . in these circumstances
it is difficult to see how the emission can remain thermal .
however , we shall continue to make that assumption in the absence of any better understanding .
assuming that the schwarzschild phase of decay is dominant , and that the mass and temperature are related by equation [ temp ] throughout ( which we have just seen must be doubtful ) , the total energy flux can be integrated @xcite to find the time at which the entire mass of the black hole has been radiated away .
this measure of the lifetime , expressed in units of the inverse of the initial mass , is shown in figure [ life ] .
we see that the lifetime falls very steeply as a function of the number of dimensions , and indeed can be comparable with the inverse mass when @xmath37 , even for masses well above the planck scale .
when this is the case , the object formed can no longer really be said to have an independent existence as a black hole .
[ ] [ ] [ 2.5]@xmath21 [ ] [ ] [ 2.5]@xmath38 [ ] [ ] [ 1.75]@xmath39 [ ] [ ] [ 1.75]@xmath40
the simulation program charybdis @xcite generates black hole production and decay configurations assuming that all the partonic collision energy goes into the mass of the hole , and that the schwarzschild decay phase dominates . in equation [ fndef ] , so that the production cross section is somewhat smaller than it should be .
] the identities and momenta of the incoming partons and outgoing primary decay products are passed to the herwig @xcite event generator via the les houches interface @xcite .
herwig then handles all the qcd parton showering , hadronization and secondary decays . .main
charybdis parameters.[parameters ] [ cols="^,<,^,^",options="header " , ] table [ probtab ] shows the relative numbers of primary particles of different types emitted when the parameters ` timvar ` , ` grybdy ` , ` nbody ` and ` nbody ` have their default values , compared to the values obtained by integrating the theoretical spectra .
the minor discrepancies are due mainly to kinematic constraints , particle masses and charge conservation . turning from single - particle spectra to overall event characteristics ,
one interesting feature is that black hole decay is associated with large missing transverse energy , due to copious primary and secondary neutrino emission .
figure [ misset ] , from @xcite , shows a comparison with expected qcd and susy missing @xmath41 distributions generated using herwig @xcite .
we see that the black hole missing @xmath41 is typically larger even than that in supersymmetric processes , where missing energy is mostly carried off by a pair of neutralinos .
the effect is partly due to the larger mass of the black hole , relative to the assumed susy scale .
however , a large cross section with large missing energy would clearly be a good initial indicator of black hole production .
, scaledwidth=65.0% ] the large missing energy in black hole decay poses a problem for the reconstruction of the mass of the black hole from its decay products . in @xcite
we found that a cut on missing @xmath42 gev was necessary for a useful mass resolution of around 4% , i.e. @xmath43 gev at 5 tev , as illustrated in figure [ bhmass ] . at this low value of missing @xmath41
, qcd background has to be controlled by requiring at least 4 high-@xmath41 jets .
, title="fig:",scaledwidth=49.0% ] , title="fig:",scaledwidth=49.0% ]
in the event that black holes are produced at the lhc , the quantity of principal interest will be the number of extra dimensions , @xmath21 . given the sensitivity of the hawking temperature to @xmath21 in equation [ temp ] , it would seem that fitting the emission spectra for various black hole masses would allow one to extract this quantity @xcite .
however , we found in @xcite that the theoretical uncertainties , together with the distortion of the spectrum by secondary decays , would make it difficult to have confidence in such a measurement . for example , figure [ timvar ] shows the effect of possible time variation of the hawking temperature on a fit to the primary electron spectrum ( assuming this could be unfolded cleanly from the data ) . for events generated with
the temperature frozen ( left ) , the fit gives a result consistent with the input value @xmath44 , whereas re - thermalization between every emission ( right ) systematically shifts the fit to higher values , due to the higher average temperature .
since the true situation would presumably lie between these extremes , the true value of @xmath21 could not be extracted without a deeper understanding of the decay process .
vs @xmath6 from primary electron spectrum for @xmath44 .
( a ) ` timvar ` off ; ( b ) ` timvar ` on.[timvar],scaledwidth=85.0% ] figure [ cutoff ] illustrates another difficulty in fitting the emission spectrum as a function of black hole mass , this time due to kinematic effects at higher energies .
the limitation to emission energies less than half the total mass significantly truncates the spectrum generated according to equation [ spectrum ] at low masses and/or large values of @xmath21 . , showing the effect of the kinematic limit @xmath45 .
the colour code indicates the number of particles in each bin per 1000 events.[cutoff],title="fig : " ] + ( a ) @xmath44 .
, showing the effect of the kinematic limit @xmath45 .
the colour code indicates the number of particles in each bin per 1000 events.[cutoff],title="fig : " ] + ( b ) @xmath46 . in order to avoid the low - energy region of hawking emission , where secondary decays distort the spectrum , and the highest energies where the kinematic cutoff takes effect , we examined @xcite the region of high but not extreme energies , @xmath47 where @xmath48 is a few hundred gev .
particles in this region are also less sensitive to time variation of the temperature , since they tend to be emitted early in the evolution of the black hole .
this tendency is further enhanced by demanding that the highest - energy emission should have energy @xmath49 .
we therefore looked at the fraction @xmath50 of events satisfing this cut as a function of @xmath6 .
gev vs @xmath6 for different options , all with @xmath46 .
upper and lower bounds explained in the text are also shown.,title="fig : " ] + ( a ) default case .
gev vs @xmath6 for different options , all with @xmath46 .
upper and lower bounds explained in the text are also shown.,title="fig : " ] + ( b ) no time variation .
gev vs @xmath6 for different options , all with @xmath46 .
upper and lower bounds explained in the text are also shown.,title="fig : " ] + ( c ) kinematic cut on .
gev vs @xmath6 for different options , all with @xmath46 .
upper and lower bounds explained in the text are also shown.,title="fig : " ] + ( d ) 4-body remnant decay .
as shown in figure [ fig : klmodelinvariant ] , for @xmath51 gev this observable is indeed relatively insensitive to the uncertainties represented by the charybdis parameters ` timvar ` , ` kincut ` and ` nbody ` .
the upper and lower bounds were obtained by integrating the planck spectrum from @xmath52 up to infinity and up to @xmath53 , respectively , and include a @xmath43 gev systematic uncertainty in the black hole mass .
the value of @xmath50 always lies within , or at least is consistent with , the expected band .
therefore a fit to this observable should have less model dependence than analyses based , for example , on a fit to the full emission spectrum . , scaledwidth=65.0% ] in @xcite we combined a fit to this observable and a cross - section measurement with an assumed error of 20% , to make a joint determination of the planck mass and the number of extra dimensions .
as shown in figure [ kl_n_mpl ] , for @xmath10 tev and @xmath46 the method gives an unbiased estimate of these quantities , with @xmath54 errors of @xmath55 , @xmath56 ( strongly correlated ) .
if there are indeed extra dimensions large and numerous enough to reduce the fundamental planck scale to the tev range , black hole production at accelerators , with substantial cross sections , is a real possibility
. such a discovery would arguably be the most profound advance in fundamental physics since general relativity .
it would also constitute the first observation of hawking radiation .
however , as i have repeatedly emphasised , there remain many uncertainties about the precise nature , evolution and ultimate fate of the objects that would be formed .
the strategy adopted here has been to start with a crude but flexible simulation framework , and to look for observables that are less sensitive to some of these uncertainties , in order to stimulate thinking about how the fundamental parameters could be measured .
if real data on this process do become available , there will undoubtedly be an explosion of theoretical and experimental activity that will rapidly reduce the uncertainties .
many thanks to members of the cambridge susy working group , in particular the other authors of ref .
@xcite , for their collaboration and comments , and to steve giddings for helpful suggestions .
the hospitality of the cern theory group during part of this work is gratefully acknowledged .
this work was supported in part by the u.k . particle physics and astronomy research council .
99 s. b. giddings and s. d. thomas , `` high energy colliders as black hole factories : the end of short distance physics , '' phys .
d * 65 * , 056010 ( 2002 ) [ arxiv : hep - ph/0106219 ] . s. dimopoulos and g. landsberg , `` black holes at the lhc , '' phys .
lett . * 87 * , 161602 ( 2001 ) [ arxiv : hep - ph/0106295 ] . i. antoniadis , `` a possible new dimension at a few tev , '' phys .
b * 246 * , 377 ( 1990 ) ; n. arkani - hamed , s. dimopoulos and g. r. dvali , `` the hierarchy problem and new dimensions at a millimeter , '' phys .
b * 429 * , 263 ( 1998 ) [ arxiv : hep - ph/9803315 ] ; i. antoniadis , n. arkani - hamed , s. dimopoulos and g. r. dvali , `` new dimensions at a millimeter to a fermi and superstrings at a tev , '' phys .
b * 436 * , 257 ( 1998 ) [ arxiv : hep - ph/9804398 ] .
l. randall and r. sundrum , `` a large mass hierarchy from a small extra dimension , '' phys .
lett . * 83 * , 3370 ( 1999 ) [ arxiv : hep - ph/9905221 ] .
h. yoshino and y. nambu , `` black hole formation in the grazing collision of high - energy particles , '' phys .
d * 67 * , 024009 ( 2003 ) [ arxiv : gr - qc/0209003 ] .
d. m. eardley and s. b. giddings , `` classical black hole production in high - energy collisions , '' phys .
d * 66 * , 044011 ( 2002 ) [ arxiv : gr - qc/0201034 ] .
h. yoshino and v. s. rychkov , `` improved analysis of black hole formation in high - energy particle collisions , '' phys .
d * 71 * , 104028 ( 2005 ) [ arxiv : hep - th/0503171 ] . c. m. harris , m. j. palmer , m. a. parker , p. richardson , a. sabetfakhri and b. r. webber , `` exploring higher dimensional black holes at the large hadron collider , '' jhep * 0505 * , 053 ( 2005 ) [ arxiv : hep - ph/0411022 ] . c. m. harris , `` physics beyond the standard model : exotic leptons and black holes at future colliders , '' cambridge ph.d .
thesis , arxiv : hep - ph/0502005 .
d. ida , k. y. oda and s. c. park , `` rotating black holes at future colliders : greybody factors for brane fields , '' phys .
d * 67 * , 064025 ( 2003 ) [ erratum - ibid .
d * 69 * , 049901 ( 2004 ) ] [ arxiv : hep - th/0212108 ] .
s. hannestad and g. g. raffelt , `` stringent neutron - star limits on large extra dimensions , '' phys .
lett . * 88 * , 071301 ( 2002 ) [ arxiv : hep - ph/0110067 ] ; `` supernova and neutron - star limits on large extra dimensions reexamined , '' phys .
d * 67 * , 125008 ( 2003 ) [ erratum - ibid .
d * 69 * , 029901 ( 2004 ) ] [ arxiv : hep - ph/0304029 ] .
s. dimopoulos and r. emparan , `` string balls at the lhc and beyond , '' phys .
b * 526 * , 393 ( 2002 ) [ arxiv : hep - ph/0108060 ] .
b. koch , m. bleicher and s. hossenfelder , `` black hole remnants at the lhc , '' arxiv : hep - ph/0507138 .
l. a. anchordoqui , j. l. feng , h. goldberg and a. d. shapere , `` inelastic black hole production and large extra dimensions , '' phys .
b * 594 * , 363 ( 2004 ) [ arxiv : hep - ph/0311365 ] .
v. cardoso , e. berti and m. cavaglia , `` what we ( do nt ) know about black hole formation in high - energy collisions , '' class .
* 22 * , l61 ( 2005 ) [ arxiv : hep - ph/0505125 ]
. g. t. horowitz , `` higher dimensional generalizations of the kerr black hole , '' arxiv : gr - qc/0507080 . c. m. harris and p. kanti , `` hawking radiation from a ( 4+n)-dimensional rotating black hole , '' arxiv : hep - th/0503010 ; g. duffy , c. harris , p. kanti and e. winstanley , `` brane decay of a ( 4+n)-dimensional rotating black hole : spin-0 particles , '' jhep * 0509 * , 049 ( 2005 ) [ arxiv : hep - th/0507274 ] .
m. casals , p. kanti and e. winstanley , `` brane decay of a ( 4+n)-dimensional rotating black hole .
ii : spin-1 particles , '' arxiv : hep - th/0511163 .
d. ida , k. y. oda and s. c. park , `` rotating black holes at future colliders .
iii : determination of black hole evolution , '' arxiv : hep - th/0602188 .
p. kanti and j. march - russell , `` calculable corrections to brane black hole decay .
i : the scalar case , '' phys . rev .
d * 66 * , 024023 ( 2002 ) [ arxiv : hep - ph/0203223 ] ; `` ii : greybody factors for spin 1/2 and 1 , '' phys .
d * 67 * , 104019 ( 2003 ) [ arxiv : hep - ph/0212199 ] ; c. m. harris and p. kanti , `` hawking radiation from a ( 4+n)-dimensional black hole : exact results for the schwarzschild phase , '' jhep * 0310 * , 014 ( 2003 ) [ arxiv : hep - ph/0309054 ] . c. m. harris , p. richardson and b. r. webber , `` charybdis : a black hole event generator , '' jhep * 0308 * , 033 ( 2003 ) [ arxiv : hep - ph/0307305 ] . g. corcella _ et al .
_ , `` herwig 6 : an event generator for hadron emission reactions with interfering gluons ( including supersymmetric processes ) , '' jhep * 0101 * , 010 ( 2001 ) [ arxiv : hep - ph/0011363 ] ; `` herwig 6.5 release note , '' arxiv : hep - ph/0210213 .
e. boos _ et al .
_ , `` generic user process interface for event generators , '' arxiv : hep - ph/0109068 . s. moretti , k. odagiri , p. richardson , m. h. seymour and b. r. webber , `` implementation of supersymmetric processes in the herwig event generator , '' jhep * 0204 * , 028 ( 2002 ) [ arxiv : hep - ph/0204123 ] . | in theories with large extra dimensions and tev - scale gravity , black holes are copiously produced in particle collisions at energies well above the planck scale .
i briefly review some recent work on the phenomenology of this process , with emphasis on theoretical uncertainties and possible strategies for measuring the number of extra dimensions . |
as known to all , shannon s information theory deals mainly with the representation and transmission of information . in the development of both source and channel coding theorems , especially for their converses , fano s inequality serves as the key tool @xcite .
@xmath1 and @xmath2 are two random variables following @xmath3 @xmath4 and @xmath5 .
define @xmath6 , then @xmath7 where @xmath8 is the cardinality of @xmath1 and @xmath2 , @xmath9 is the binary entropy function @xmath10 $ ] for @xmath11 .
thus fano inequality can be further relaxed by @xmath12 .
usually , the left hand side of ( [ eq : fano ] ) is referred to as the equivocation , which is quantified by the conditional entropy .
particularly , it represents the uncertainty whether the restored / decoded message @xmath2 is the same as the original one , i.e. , @xmath1 . on the other hand ,
the right hand side implies the reliability of the source / channel coding in terms of a function of error probability @xmath13 .
it was shown in @xcite that vanishing equivocation implies vanishing error probability . however , vanishing error probability does not necessarily guarantee a vanishing equivocation , especially for some @xmath14 of countably infinite alphabet . in proving the converse of coding theorems ,
one wants to find the upper bound on the size of the codebook given arbitrary code length and error probability .
the following theorem is an immediate inference of fano s inequality , simple but useful .
[ th : fano_ixy ] @xcite suppose @xmath1 and @xmath2 are two random variables that take values on the same finite set with cardinality @xmath15 and at least one of them is equiprobable .
then the mutual information between them satisfies @xmath16 where @xmath6 and @xmath17 @xmath18 $ ] . as an inference of this result
, the following theorem gives an upper bound on the size of a code as a function of the average error probability .
[ th : fano_converse ] @xcite every @xmath19-code ( average probability of error ) for a random transformation @xmath20 satisfies @xmath21 where @xmath22 , @xmath10 $ ] .
although simple , these two theorems are insightful and easy to compute both in theory and numerically .
the classical coding theorems are mainly based on the asymptotic equipartition property ( aep ) and typical / joint - typical decoder @xcite .
however , applications of aep requires infinite long codewords , where the error probability goes either to 0 or 1 as the code length goes to infinity .
although these coding theorems provide fundamental limits for modern communications , research on the finite blocklength coding schemes are more important in engineering applications . given the block length , upper bounds on the achievable error probability and the achievable code size were obtained in @xcite .
most importantly , a tight approximation for the achievable maximal rate given the error probability and code length was presented . in this paper
, we consider the entropy of one random variable vector conditioned on another , and the corresponding probability of error in guessing one from the other , by proposing an extended fano s inequality .
the extended fano s equality has better performance by taking advantage of a more careful consideration on the error patterns .
it suits codings in the finite blocklength regime better and is useful in bounding the mutual information between random vectors , and the codebook size given the block length and average symbol error probability constraint . in the following part of this paper
, we present the extended fano s inequality in section [ sec : gener_fano ] first .
the lower bounds on the mutual information between two random variable vectors and a upper bound on the codebook size given the block length and error probability are given in section [ sec : converse ] .
an application of the the obtained result to the @xmath0-ary symmetric channels ( qsc ) are presented in section [ sec : app ] , which shows that the extended fano s inequality is tight for such channels . finally , we concluded the paper in section [ sec : conclusion ] . throughout the paper , vectors indicated by bold
although fano s inequality has been used widely in the past few years , it can be improved by treating the error events more carefully . in this section ,
a refinement of fano s inequality is presented , which is tighter and more applicable for finite blocklength coding design .
fano s inequality extension[th : gner_fano ] suppose that @xmath23 and @xmath24 @xmath25 are two @xmath26-dimension random vectors where @xmath27 and @xmath28 @xmath29 take values on the same finite set @xmath30 with cardinality @xmath31 .
then the conditional entropy satisfies @xmath32 where @xmath33 is the discrete entropy function .
@xmath34 is the error distribution , where the error probabilities are @xmath35 for @xmath36 .
@xmath37 is the generalized hamming distance , defined as the number of symbols in @xmath38 that are different from the corresponding symbol in @xmath39 .
define the error random variable as @xmath40 if @xmath41 for @xmath42 .
according to the chain rule of the joint entropy , @xmath43 can be expressed in the following ways , [ dr : fano_g ] @xmath44 particularly , in ( [ dr : fano_g].a ) , it is clear that @xmath45 .
then we have @xmath46 where ( a ) follows the fact that entropy increases if its condition is removed , i.e. , @xmath47 .
particularly , we have @xmath48 . according to its definition
, we have @xmath49 when considering @xmath50 , we know that there are @xmath51 disaccord symbol pairs between @xmath38 and @xmath39 . for each fixed @xmath52 , every symbol in @xmath38 which belongs to a disaccord pair
has @xmath53 possible choices except the one in @xmath54 .
thus @xmath55 has @xmath56 choices .
besides , there @xmath57 selections for the positions of error symbols for each given @xmath51 .
therefore , the total number of possible codeword @xmath38 is @xmath58 , which means @xmath59 particularly , note that @xmath60 since there is no uncertainty in determining @xmath38 from @xmath39 , if they are the same .
then , ( [ dr : condi_err ] ) can be written as @xmath61 by combining ( [ dr : hx_y ] ) and ( [ hey ] ) , the proof of the theorem is completed .
in fact , the error distribution @xmath62 is easy to calculate , especially for some special channels .
for example , the discrete @xmath0-ary symmetric channels is shown in fig [ fig : qbc ] . in this situation , @xmath63 .
-ary symmetric channel , width=240 ] it is clear that theorem [ th : gner_fano ] is a generalization of fano s inequality .
specifically , when the block length is 1 , i.e. , @xmath64 , we have @xmath65 , @xmath66 , and @xmath67 .
in this case , theorem [ th : gner_fano ] reduces to @xmath68 which is exactly the same as fano s inequality . as a variant of theorem [ th : gner_fano ] ,
the following theorem presents the conditional entropy in terms of relative entropy .
[ th : gner_fanorela ] suppose that @xmath23 and @xmath69 are two @xmath26-dimension random vectors where @xmath27 and @xmath28 @xmath29 take values on the same finite set @xmath30 with cardinality @xmath31 .
then the conditional entropy satisfies @xmath70 where @xmath71 is the discrete relative entropy function .
the error probabilities are @xmath35 for @xmath36 .
donate @xmath34 , and @xmath72 is a probability distribution with @xmath73 firstly , we know from the binomial theorem that @xmath74 is a probability distribution .
let @xmath75 and @xmath76 , we know that @xmath77 is a probability distribution where @xmath73 . according to theorem [ th : gner_fano ] , @xmath78 where ( a ) holds because @xmath79 for @xmath80 . by the definition of @xmath62 , it is clear that it reflects the error performance of the channel and is totaly determined by the channel itself . on the contrary
, @xmath77 is a distribution where each error patten is assumed to appear equiprobablely . in this situation ,
the probability that there are @xmath51 error symbols in the codeword is @xmath81 .
thus @xmath82 is the distance between the actual error pattern distribution and the uniform error pattern distribution .
particularly , if the channel is an error free one , i.e. , @xmath83 and @xmath84 for @xmath85 , we have @xmath86 . according to theorem [ th : gner_fanorela ] ,
we get @xmath87 , which is reasonable with the assumption on the channel . in this sense ,
theorem [ th : gner_fanorela ] is tight .
the fano s inequality has been playing an important role in the history of information theory because it built a close connection between conditional entropy and error probability . for extended fano s inequality given in theorem [
th : gner_fano ] , it is especially applicable in finite blocklength coding .
it also presents the relationship between conditional entropy and error probabilities , which are defined as follows .
_ block error probability _
@xmath88 is the average error probability of a block ( codeword ) , i.e. , @xmath89 .
then we have @xmath90 .
thus , we have @xmath91 for any @xmath92 .
_ symbol error probability _
@xmath93 is the average error probability of a symbol , i.e. , @xmath94 , which can be expressed by @xmath95 . in many communication systems , especially those using error correction channel codings ,
a block error does nt imply a system failure . on the contrary ,
the error can be corrected or part of the block can still be used with some performance degradation . in this case , the symbol error is more useful than the block error . particularly , a corollary following our result as shown below will answer this problem .
[ cor:1 ] suppose that @xmath23 and @xmath69 are two @xmath26-dimension random vectors where @xmath27 and @xmath28 take values on the same finite set @xmath30 with cardinality @xmath31 .
then the conditional entropy satisfies @xmath96 where @xmath97 is a probability distribution with @xmath98 .
according to theorem [ th : gner_fano ] , one has@xmath99 since the distribution @xmath100 can be expressed as @xmath101 , which is a binomial distribution with the symbol error probability of @xmath102 .
@xmath103 is a measure of the distance between the error probability distribution and the binomial distribution with parameter @xmath102 .
if one takes @xmath64 , corollary [ cor:1 ] will reduces to @xmath104 , which is a frequently used form of fano s inequality .
it is seen that the extended fano s inequality builds a natural connection between conditional entropy and symbol error and is especially applicable for finite length codings .
based on the proposed generalized fano s inequality , the following lower bounds on the mutual information between @xmath38 and @xmath39 can be obtained .
[ th : gner_fano_ixy ] suppose that @xmath23 and @xmath69 are two @xmath26-dimension random vectors that satisfy the following . 1 . @xmath27 and @xmath28 @xmath29 take values on the same finite set @xmath30 with cardinality @xmath31 .
either @xmath38 or @xmath39 is equiprobable .
the error probabilities are @xmath35 for @xmath36 .
donate the error distribution as @xmath34 .
then the mutual information between @xmath38 and @xmath39 satisfies @xmath105 where @xmath33 is the entropy function .
if @xmath38 is equiprobable , @xmath106 . on the other hand , the mutual information is given by @xmath107 .
together with theorem [ th : gner_fano ] , @xmath108 note that @xmath38 and @xmath39 are totally symmetric in ( [ dr : gener_ixy1 ] ) .
therefore , if @xmath39 is assumed to be equiprobable at the beginning of the proof , one can get the same result .
thus theorem [ th : gner_fano_ixy ] is proved . by using theorem [ th : gner_fanorela ] ,
the mutual information between @xmath38 and @xmath39 can be bounded by the following corollary .
[ cor:4 ] suppose that @xmath23 and @xmath69 are two @xmath26-dimension random vectors that satisfy the following . 1 .
@xmath27 and @xmath28 @xmath29 take values on the same finite set @xmath30 with cardinality @xmath31 .
either @xmath38 or @xmath39 is equiprobable .
3 . the error probabilities are @xmath35 and @xmath34 .
then the mutual information between @xmath38 and @xmath39 satisfies @xmath109 where @xmath71 is the discrete relative entropy function and @xmath72 is a probability distribution with @xmath73 the distribution @xmath77 means that the symbol in @xmath39 takes any value on @xmath110 with equal probability , regardless of what is sent in @xmath38 .
so it is a pure random distribution when @xmath1 and @xmath2 are independent from each other .
the most desirable coding scheme is that its error distribution @xmath62 is farthermost from @xmath77 , which also ensures a larger coding rate .
suppose @xmath30 is a finite alphabet with cardinality @xmath31 .
let s consider the input and output alphabets @xmath111 with @xmath112 and a channel to be a sequence of conditional probabilities @xcite @xmath113 .
we donate a codebook with @xmath15 codewords by @xmath114 .
a decoder is a random transformation @xmath115 where @xmath116 indicates that the decoder choose error .
if messages are equiprobable , the average error probability is defined as @xmath117 .
an codebook with @xmath15 codewords and a decoder whose average probability of error is smaller than @xmath118 are called an @xmath119-code . an upper bound on the size of a code as a function of the average probability of symbol error follows the corollary [ cor:1 ] .
[ th : conver_1 ] every @xmath119-code for a random transformation @xmath120 satisfies @xmath121 where @xmath34 is the error distribution with @xmath35 for @xmath36 , @xmath122 @xmath123 is a probability distribution with @xmath98 .
since the messages are equiprobable , we have @xmath124 . according corollary [ cor:1 ] , @xmath125 solving @xmath126 from ( [ dr : th_con1 ] ) , one can get ( [ rt : conver_1 ] ) , which completes the proof .
consider information transmission over a memoryless discrete @xmath0-ary symmetric channel with a channel code @xmath127 with crossover probability @xmath128 , as shown in fig .
[ fig : qbc ] . in this case , the probability of symbol error is @xmath129 .
then the error probabilities are @xmath130 and the block error probability is @xmath131 using the extended fano s inequality in theorem [ th : gner_fano ] , we have @xmath132 it is easy to see that the conditional entropy in theory is @xmath133 by corollary [ cor:4 ] , mutual information is lower bounded by@xmath134 . \end{split}\ ] ] while the capacity of the memoryless qsc is given by @xmath135 and the relative entropy @xmath136 can be derived as @xmath137\\ \end{split}\ ] ] for a given average symbol error probability constraint @xmath138 , the upper bound on the maximum codebook size given by theorem [ th : conver_1 ] is @xmath139 + n\epsilon\log(q-1 ) .
\end{split}\ ] ] on the other hand , by fano s inequality we have @xmath140 with @xmath88 given by ( [ eq : pb ] ) .
then the lower bound of the mutual information is @xmath141 finally , the upper bound on the codebook size is @xmath142 suppose the qsc parameter are @xmath143 and @xmath144 , we calculated the bounds on conditional entropy , mutual information and codebook size by our proposed results and fano s inequality .
firstly , the upper bound on the conditional entropy is presented in fig .
[ fig : hxy ] .
specially , @xmath145 is obtained by the extended fano s inequality ( [ sim : hexy ] ) , @xmath146 is calculated according to fano s inequality ( [ sim : hfxy ] ) and @xmath147 is the conditional entropy in theory ( [ sim : hxy ] ) .
it is clear that theorem [ th : gner_fano ] is tighter than fano s inequality .
particularly , we have @xmath148 for the qsc .
this is because the error distributions are the same for any @xmath52 .
so @xmath149 holds . besides
, the error pattern is uniformly distributed for a given @xmath51 , regardless of @xmath54 and @xmath150 holds .
therefore , the upper bound is tight .
however , for fano s equality , there are relaxations in both @xmath149 to @xmath151 and @xmath152 to @xmath153 . , width=259 ] , width=278 ] similarly
, the lower bound on mutual information @xmath154 given by ( [ sim : iexy ] ) coincides with @xmath155 in theory , given by ( [ sim : ixy ] ) and is better than that given by fano s inequality ( [ sim : ifxy ] ) .
when we use the upper bound on the codebook size in theorem [ th : conver_1 ] , it should be noted that it is presented as a function of symbol error probability @xmath93 .
in fact , @xmath88 is always larger than @xmath93
. therefore , we use the same fraction of them in the calculation of the bounds to make sense of the comparison , i.e. , @xmath156 for ( [ sim : mexy ] ) and @xmath157 for ( [ sim : mxy ] ) .
it is also seen from fig .
[ fig : ixy ] that our new developed result is tighter .
the performances of theorem [ th : gner_fano_ixy ] and theorem [ th : conver_1 ] versus the qsc parameter @xmath128 are shown in fig .
[ fig : epslong ] , where the block length is chosen as @xmath158 .
as shown , the mutual information bound is tight and our results are much better than fano s inequality . in the calculation of the upper bounds on codebook size , the selection of the error probability constraints are also chosen as @xmath159 and @xmath160 so that they are comparable . , @xmath161,width=278 ]
in this paper , we revisited fano s inequality and extended it to a general form . particularly , the relation between the conditional entropy and error probability of two random vectors was considered , other than that between two random variables .
this makes the developed results more suitable for source / channel codings in the finite blocklength regime . by investigating the block error pattern more detailedly ,
the conditional entropy of the original random vector given the received one is upper bounded more tightly by the extended fano s inequality .
furthermore , the extended fano s inequality is completely tight for some symmetric channels such the @xmath0-ary symmetric channels .
converse results are also presented in terms of lower bounds on the mutual information and a upper bound on the codebook size under the blocklength and symbol error constraints , which also have better performances .
this work was partially supported by inc research grant of chinese university of hongkong and the china major state basic research development program ( 973 program ) no .
2012cb316100(2 ) .
15 c. e. shannon,a mathematical theory of communication , " _ bell syst .
pp.623 - 656 , oct .
r. m. fano , _ class notes for transmission of information _ , course6.574 .
combridge , ma : mit .
r. w. raymond , _ information theory and network coding _ ,
berlin / newyork : springer . 2008 .
han and s. verdu , generalizing the fano inequality , " _ ieee trans .
inform . theory _
40 , no . 7 , pp .
1247 - 1250 , jul . 1994 . s. w. ho and s. verdu , on the interplay between conditional entropy and error probability " _ ieee trans .
inform . theory _ ,
vol.56 , no.12 , pp.5930 - 5942 , dec . 2010 .
d. l. tebbe and s. j. dwyer , iii , uncertainty and probability of error , " _ ieee trans .
inform . theory _
516 - 518 , may 1968 . | fano s inequality reveals the relation between the conditional entropy and the probability of error .
it has been the key tool in proving the converse of coding theorems in the past sixty years . in this paper , an extended fano s inequality
is proposed , which is tighter and more applicable for codings in the finite blocklength regime .
lower bounds on the mutual information and an upper bound on the codebook size are also given , which are shown to be tighter than the original fano s inequality .
especially , the extended fano s inequality is tight for some symmetric channels such as the @xmath0-ary symmetric channels ( qsc ) . fano s inequality , finite blocklength regime , channel coding , shannon theory . |
the infrared emission features ( iefs ) at 3.3 , 6.2 , 7.7 , 8.6 , 11.3 and 12.7 , and their associated continuum , have been the subject of intense study over the last decade .
the ief strengths correlate with the c / o ratio in planetary nebulae , suggesting a carbon - based carrier ( roche & aitken 1986 ; cohen et al .
1986 , 1989 ) .
duley & williams ( 1981 ) were the first to comment that some ief wavelengths are characteristic of the bending and stretching modes of various c h and c
c bonds in aromatic hydrocarbons .
there has since been a steady accumulation of evidence , both laboratory and observational , in support of this assertion .
the specific nature of the carrier(s ) , however , is yet to be determined .
the iefs have been attributed to polycyclic aromatic hydrocarbon ( pah ) molecules ( lger & puget 1984 ; allamandola , tielens , & barker 1985 ) and to more amorphous materials containing aromatic hydrocarbons ( sakata et al .
1984 , 1987 ; borghesi , bussoletti , & colangeli 1987 ; blanco , bussoletti , & colangeli 1988 ; duley 1988 ; papoular et al .
allamandola , tielens , & barker ( 1989 ) and bregman et al .
( 1989 ) attribute the broad spectral structures at 6 9 and 11 13 , which accompany the iefs , to pah clusters and amorphous carbon grains .
the origin of the 1 25 @xmath0 m continuum emission , like that of the associated ief emission , is still the subject of debate .
the continuum is observed in regions of very low radiation density , such as reflection nebulae and infrared cirrus , where the grain equilibrium temperature is too low to explain the emission as being thermal in nature ( sellgren 1984 ; draine & anderson 1985 ; guillois et al .
it has been attributed to single ultraviolet ( uv ) photon excitation of very small grains ( sellgren , werner , & dinerstein 1983 ; sellgren 1984 ) , to pah molecular fluorescence ( lger & puget 1984 ; allamandola et al .
1985 , 1989 ; puget & lger 1989 ) , and to hydrogenated amorphous carbon grain luminescence ( duley & williams 1988 ; duley 1988 ) .
the ief carriers are a ubiquitous component of the interstellar medium ( ism ) .
past studies have found iefs in _ emission _ near galactic regions , reflection nebulae , planetary nebulae , proto - planetary nebulae , and the diffuse ism of our own and other galaxies ( see reviews by aitken 1981 ; allamandola et al .
1989 ; puget & lger 1989 ; geballe 1997 ; and tokunaga 1997 ) .
the 6.2 @xmath0 m ief has recently been detected in _
absorption _ in the diffuse ism ( schutte et al .
1996 , 1998 ) .
the 3.3 @xmath0 m ief may have also been detected in _ absorption _ in molecular clouds ( sellgren , smith , & brooke 1994 ; sellgren et al . 1995 ; brooke , sellgren , & smith 1996 ; brooke , sellgren & geballe 1999 ) .
ief spectra have been recently classified into several groups ( geballe 1997 ; tokunaga 1997 ) .
class a spectra at 5 15 @xmath0 m display narrow iefs at 6.2 , 7.7 , 8.6 , 11.3 and 12.7 , as well as associated continuum emission .
the 7.7 ief is invariably the strongest of the class a narrow iefs ( geballe 1997 ) .
by contrast , class b spectra lack the 8.6 ief and have notably broader features at 8 and 11.5 ( kwok , volk , & hrivnak 1989 ) .
class b spectra sometimes also show a strong broad emission feature at 8.8 ( buss et al . 1993 ) . the overwhelming majority of ief spectra currently measured fall under the category of class a. sources with class a iefs are typically uv - rich sources , such as the diffuse ism , reflection nebulae , regions , young stars , planetary nebulae and galaxies hosting star formation .
by contrast , the half - dozen or so class b spectra found to date arise from f and g - type post - agb stars forming planetary nebulae ( i.e. proto - planetary nebulae ; geballe et al .
1992 ) .
well - developed theoretical work exists for the absorption and emission properties of pahs , inspired by the hypothesis that pahs are responsible for the iefs
. mixtures of pahs and other dust components have been used to model the uv and visible extinction curve and the infrared emission observed in the ism ( puget , lger , & boulanger 1985 ; chlewicki & laureijs 1988 ; dsert , boulanger & puget 1990 ; joblin , lger , & martin 1992 ; siebenmorgen & krgel 1992 ; schutte , tielens , & allamandola 1993 ; dwek et al .
1997 ; li & greenberg 1997 ; silva et al .
all of these models include a size distribution of pahs .
most of these size distributions include large pahs , expected in the ism but not yet measured in the laboratory , by extrapolation from laboratory data on smaller pahs .
interstellar pahs are predicted to be mainly positively charged and neutral in strong uv fields , and mainly neutral and negatively charged when the uv field is weak ( bakes & tielens 1994 , 1998 ; salama et al .
1996 ; dartois and dhendecourt 1997 ) . some interstellar models ( particularly dsert et al . 1990 ) include both neutral and ionized pahs .
these models for the absorption and emission from a size distribution of neutral and ionized pahs predict that the fraction of the total interstellar extinction curve due to pah absorption is much larger at uv wavelengths than at visible wavelengths .
this is vividly illustrated in figures 2 and 3 of dsert et al .
( 1990 ) , figures 1 , 2 , and 3 of joblin et al .
( 1992 ) , figures 1_b _ , 2_b _ and 4_b _ of siebenmorgen & krgel ( 1992 ) , figure 4 of dwek et al .
( 1997 ) , and figure 3 of silva et al .
if pahs contribute more to the total interstellar extinction at uv than visible wavelengths , then pahs should absorb and re - emit a larger fraction of the total absorption and re - emission by dust in uv - rich sources than in uv - poor sources .
reflection nebulae are valuable tools for the study of the class a interstellar iefs . the localized heating of an interstellar cloud by a nearby , optically visible , star allows us to study the iefs over varying excitation conditions by observing reflection nebulae illuminated by stars of different effective temperature , @xmath11 .
this localized heating also results in brighter infrared emission than would be observed in a similar cloud heated only by the diffuse interstellar radiation field .
sellgren , luan , & werner ( 1990 ) analyzed _ iras _ observations of reflection nebulae illuminated by stars with @xmath1 = 3,000 k
33,000 k. they observed that the ratio of nebular emission within the _ iras _ 12 @xmath0 m broadband filter , @xmath12(12 @xmath0 m ) , to the total bolometric far - infrared nebular emission , @xmath13(fir ) , was independent of @xmath1 , for @xmath11 = 5,000 k
33,000 k. the total infrared emission of a reflection nebula is due to starlight absorbed and re - radiated by the mixture of molecules and grains responsible for the total interstellar extinction curve .
their observation that @xmath12(12 @xmath0m)/@xmath13(fir ) is constant over a wide range of @xmath11 implies that the material emitting in the _ iras _
12 @xmath0 m band has an absorption curve with a wavelength dependence similar to that of the total interstellar absorption curve over both uv and visible wavelengths .
the sellgren et al .
( 1990 ) observations , finding @xmath14(12 @xmath0m)/@xmath13(fir ) to be independent of @xmath11 , appear to be in conflict with theoretical predictions that the _ iras _ 12 @xmath0 m band is dominated by pah emission and that pahs contribute more to the total interstellar extinction curve at uv than visible wavelengths ( dsert et al . 1990 ; joblin et al .
1992 ; siebenmorgen & krgel 1992 ; schutte et al .
1993 ; dwek et al .
1997 ; li & greenberg 1997 ; silva et al .
this is particularly puzzling as both dsert et al .
( 1990 ) and siebenmorgen & krgel ( 1992 ) explicitly fit a mixture of pahs and other dust components to the observed extinction curves and infrared emission of two reflection nebulae observed by sellgren et al .
( 1990 ) : ngc 2023 ( vdb 52 ; @xmath1 = 22,000 k ) and ngc 7023 ( vdb 139 ; @xmath1 = 17,000 k ) .
the sellgren et al .
( 1990 ) observations could be reconciled with theoretical predictions if other ism components besides pahs contribute to the _ iras _
12 @xmath0 m broad - band emission in reflection nebulae illuminated by cool stars .
infrared spectroscopy can test whether the fractional contribution of pahs to the _ iras _
12 @xmath0 m broad - band emission of reflection nebulae depends on @xmath11 .
sellgren , werner , & allamandola ( 1996 ) have searched for spectral differences in the near - infrared emission of reflection nebulae with @xmath1 = 3,600 k
33,000 k. they find that the equivalent width of the 3.3 @xmath0 m ief in reflection nebulae is independent of @xmath1 , over @xmath1 = 11,000 k 22,000 k , but they did not have the sensitivity to detect reflection nebulae illuminated by cooler stars .
this paper presents a spectroscopic imaging study to search for and detail interstellar ief emission toward a sample of reflection nebulae illuminated by stars with @xmath1 = 3,600 k 19,000 k ( table [ tab : stelpar ] ) .
we obtained our observations with the _ infrared space observatory _
( _ iso _ ) camera , isocam ( kessler et al .
1996 ; c. cesarsky et al . 1996 ) , using the circular variable filter ( cvf ) .
we have reported our initial result , the detection of iefs toward the uv - poor nebula vdb133 ( illuminated by two stars with @xmath1 = 6,800 k and 12,000 k ) , in uchida , sellgren , & werner ( 1998 ; hereafter paper i ) .
the goal of our study is to determine whether there are quantitative differences between the excitation of the different iefs and the continuum emission among nebulae with widely varying @xmath1 .
we selected reflection nebulae with previous _
iras _ detections of extended 12 emission ( sellgren et al .
the unprecedented sensitivity of the cryogenically cooled _
satellite to low surface brightness mid - infrared emission has allowed us to search for iefs toward the reflection nebulae illuminated by stars later than b - type , where the iefs are too faint for detection with ground - based or airborne telescopes at ambient temperature .
we obtained isocam multi - wavelength images of vdb101 , vdb111 , vdb133 and vdb135 between 1996 march19 and 1998 february 16 .
we used a set of 9 10 cvf wavelengths and 1 2 narrowband filters toward each nebula , in four sequences .
each sequence contains observations at three to four wavelengths , as listed in table [ tab : obswave ] .
we observed all four sequences for each nebula , except for vdb135 , where no sequence 3 observations were made .
while the filter observations within a given sequence were performed consecutively in time , the scheduling of the sequences themselves were not necessarily constrained so .
some sequences thus share common wavelength observations to provide a check for consistency with time .
all observations were made with a 6 pixel@xmath15 scale giving a total field of view of 32 with the 32@xmath932 pixel array .
the cvf spectral resolution was @xmath16 = @xmath17 = 40 and the narrowband filters at 4.5 ( lw1 ) and 15 ( lw9 ) had bandpasses of @xmath18 = 1 and 2 , respectively ( isocam team 1994 ) .
the observations were position - switched ( in a 2-cycle sky@xmath19source@xmath19sky@xmath19source sequence ) , with equal times on the source and sky to allow for the removal of the background and dark current contributions .
the source observations were centered on the illuminating star(s ) of each nebula and the sky positions were toward nearby regions deemed free , based on inspection of the _ iras _ 12 skyview images , of any significant emission above the zodiacal component and the local cirrus background .
table [ tab : obspos ] lists the positions we observed .
the basic integration time per exposure varied between 0.28 s and 10 s , depending on the wavelength and brightness of the central star .
the total number of exposures was set to attain a total on - source integration time per filter of @xmath20200s for all observations .
spectral images of vdb17 ( ngc 1333 ) , vdb59 ( ngc 2068 ) , vdb101 , vdb133 and vdb135 were also taken during the supplemental observing period of _
iso_. these involved full cvf wavelength scans at 5.14 9.44 ( cvf1 ) and 9.33 15.1 ( cvf2 ) , taken in both the increasing and decreasing wavelength directions , toward each of the source and sky positions .
the spectral resolution was @xmath16 = 40 .
the observations were spaced in wavelength to provide a factor of 1 2 oversampling in the vdb17 and vdb59 spectra , and 2 4 oversampling in those of the other sources .
the basic exposure time used was 2.1s , with 8 exposures per cvf step ( totalling 17.6s ) toward the bright sources vdb17 and vdb59 , and 11 exposures per cvf step ( totalling 23.1s ) toward vdb101 , vdb133 and vdb135 .
the pixel scale was 6 pixel@xmath15 for the vdb17 , vdb59 and vdb133 measurements , like that of the earlier position - switched observations , and 12 pixel@xmath15 for vdb101 and vdb135 . in this second set of observations ,
the `` source '' fields were offset from the central star so as not to contain the star , thus allowing us to use longer integration times without saturating the detector .
the vdb133 field was centered toward the peak ief emission in the region , as inferred from the earlier position - switched _ iso _ observations , and those of vdb17 and vdb59 were centered in the directions containing the strongest extended infrared emission in our unpublished near - infrared ground - based images of the regions .
the vdb101 and vdb135 fields were centered on the same positions observed by sellgren et al .
( 1990 ) in their _ iras _ observations .
the sky positions were chosen by examination of the 12 _ iras _ skyflux images .
table [ tab : obspos ] lists the source and sky positions .
the images were reduced , from the basic ( cisp ) data product , in the idl - based isocam data reduction environment known as `` ice '' , developed at the institut dastrophysique spatiale ( ias ) . by utilizing the basic ice modules we were able to more closely monitor and interact with the various stages of processing (
i.e. , the deglitching of cosmic ray hits from the data , detector transient response rectification and flat - fielding ) than with the standard ( more automated ) cam ia processing package .
indeed , special efforts were made in the restoration of the source fluxes ; the description of the method follows .
the isocam detector array experiences a response lag and a flux - dependent transient response ( or residual memory effect ) . in order to recover the actual source fluxes , @xmath21 ,
models of the detector behavior have been developed by several of the institutes involved with isocam .
we have adopted the transient model developed at the ias ( abergel et al .
1996 ) , @xmath22 which has proven from past experience to work best in rectifying low - level flux like that expected of the diffuse nebular emission . the measured flux , @xmath23 , at time , @xmath24 , is the sum of the instantaneous response of the detector , @xmath25 , and the integrated contribution of the exponentially decaying flux history of the detector ( abergel et al 1996 ) .
the detector is characterized by two parameters , @xmath26 and @xmath27 , the exact behavior of which is not well - known but which likely varies somewhat from pixel to pixel and perhaps with time , over the long term ( weeks to months ) .
the time constant for the exponential decay , @xmath28 , is a function both of the measured flux and of the detector parameter , @xmath27 . in the best of circumstances , the flux history of the detector prior to the start of one s own observations
is known and can then be used to rectify the first part of the flux time series .
since , however , the data prior to one s own are generally not available , a number of `` padding '' frames , all with an estimated constant flux value , are instead prepended .
if the detector response were perfect ( figure [ ts]_a _ ) , the flux history of a 2-cycle chop ( sky@xmath19source@xmath19sky@xmath19source ) sequence , for a single pixel or an average of pixels in the array , should appear as a series of square functions with the data points falling on two plateaus at a common minimum and two plateaus at a common maximum ; the entire pattern then repeats itself ( but likely with some general flux offset with respect to the first ) for each of the subsequent position - switched cycles taken with a different filter . figure [ ts]_b _ displays an observed flux time sequence in raw form .
the effects of the detector s finite response time and flux memory are seen to cause significant deviations from the ideal square pattern .
the individual source and sky segments of the flux sequence , which should otherwise have constant levels , are either temporally increasing or decreasing ( depending on the history of the detected signal ) because of the non - ideal detector behavior .
the position - switched data lend themselves well to the determination of the detector parameters . by using the ideal square - wave pattern as the target in a chi - square minimization procedure , with @xmath26 , @xmath27 , and
the `` padding '' constant as free parameters , the best values of these parameters that rectify the observed time sequences can be found .
indeed , a set of parameters was determined for each of the four observation sequences toward vdb133 from averages of those pixels containing the nebular emission .
the four parameter sets ranged between @xmath26 = 0.72 0.73 and @xmath27 = 954 1603 adu gain@xmath15 , with padding values of 9.2 13.8 adu gain@xmath15 , where adu gain@xmath15 are analog - to - digital units divided by the detector gain .
the specific parameter sets were used to rectify the four corresponding vdb133 observation sequences , and their average ( @xmath26 = 0.72 and @xmath27 = 1200 adu gain@xmath15 ) was used to rectify the data sequences of the other sources .
the parameter values determined by us provided some improvement over the results based on the values adopted by abergel et al .
( 1996 ) , @xmath26 = 0.63 and @xmath27 = 1200 adu gain@xmath15 , for observations toward @xmath29 oph .
figure [ ts]_c _ displays a rectified flux time sequence , after the observed flux time sequence ( fig .
[ ts]_b _ ) has been rectified using the appropriate parameter values .
in contrast , it is not possible to deduce the detector parameters from the full cvf scans with the method described above , since these scans involve sequential wavelength observations toward a given position and thus do not possess a predictable pattern , much less the regular `` square wave '' pattern of the position - switched sequences .
all full cvf scan sequences of our supplemental observations were thus rectified using the average of the parameters determined from the original vdb133 position - switched sequences : @xmath26 = 0.72 and @xmath27 = 1200 adu gain@xmath15 .
the source and sky images were rectified separately before being used to produce the final source@xmath30sky maps . in each case , the dark current image ( boulanger 1996 ) was subtracted from the maps to establish the absolute count levels required for the rectification procedure .
it was discovered , however , that the odd - even noise pattern ( isocam team 1994 ) remained in the images after a simple subtraction of the dark and that multiplication of the dark field ( with factors ranging over 1.03 1.13 ) was needed to eliminate this problem .
a single dark multiplication factor was used for all observations of a given source .
an uncertainty of @xmath200.03 adu gain@xmath15 in determining the dark factor value translates to an uncertainty of @xmath203 mjy sr@xmath15 in the individual source and sky map levels .
however , since the source image and the sky image of a given nebula each use the same dark field ( with the same dark multiplicative factor ) , any errors in absolute flux levels of the component fields do not propagate into the source@xmath30sky maps .
the level offset in the source and sky images may lead to differential errors by the rectification process that appear in the source@xmath30sky maps , however , on the order of 5% of the absolute sky level ( 4 ) .
figures [ iso1 ] , [ iso2 ] , and [ iso3 ] show our rectified , sky - subtracted spectra of reflection nebulae from position - switched observations and cvf scan observations .
each reflection nebula spectrum is a spatial _ average _ over a subsection of the image containing the localized ief emission .
the specifics of the spatial subregions used to produce the spectra are given in table [ tab : intreg ] .
for the position - switched data , where the image contained the illuminating star , the star and the immediately adjacent region containing a spurious arc - like reflection feature from the star ( typically pixels x , y = [ 620 , 1124 ] of the array ) were excluded from the spatial integration . the position - switched observations were the first performed in this study .
the 11 unique wavelength measurements were chosen to measure the strengths of the anticipated class a interstellar iefs ( 6.2 , 7.6 , 8.62 , 11.22 @xmath31 ) , their surrounding continua ( 4.5 , 5.8 , 9.5 , 10.5 , and 15.0 @xmath31 ) and broad spectral structures such as the 6 9 @xmath31 bump and the 11 13 @xmath31 plateau ( 7.0 , 8.4 , and 12.0 @xmath0 m ) .
the cvf scan observations , which provide a complete cvf spectrum over 5.14 15.1 @xmath0 m , were performed second . in nearly all the cvf scans , there is a strong curvature at the start of the upward and downward scans because of the detector s flux memory and the satellite having slewed across the bright central star . with such large flux deviations , especially with the case of vdb17 and vdb59 ,
the response at the start of each scan could not entirely be corrected by the aforementioned rectification procedures ( 3.1 ) . to avoid these transient artifacts , we discarded data shortward of 8.0 @xmath0 m for the upward scan and data longward of 12.0 @xmath0 m for the downward scan .
our final sky - subtracted spectra are thus constructed from the downward scan at 5.1 8.0 @xmath0 m , from an average of the upward and downward scans at 8.0 12.0 @xmath0 m , and from the upward scan at 12.0 15.1 @xmath0 m .
the difference between the up and down scans is plotted over a slightly wider range of wavelengths for each cvf - scanned spectrum , to estimate the uncertainty in our cvf scan spectra .
the 9 10 @xmath0 m segment of the cvf scan spectra also contain a prominent artifact resulting from the steeply declining transmission functions of the two cvf filter wheels near their wavelength overlap region : the long wavelength extreme of the cvf1 wheel ( 5.14 9.44 ) and the short wavelength extreme of the cvf2 wheel ( 9.33 16.52 ) . thus the 9 10 @xmath0 m spectral region , while shown in some spectra , has been excluded from detailed analysis .
the position - switched spectrum of vdb133 ( fig [ iso1]_a _ ) displays the distinct signature of class a interstellar iefs at 6.2 , 7.6 , 8.62 , and 11.22 @xmath0 m . the overwhelming majority of ief spectra observed fall within this class ( geballe 1997 ; tokunaga 1997 ) . the full cvf - scan spectrum ( fig .
[ iso1]_b _ , and fig .
1 of paper i ) taken during the supplemental observing period strongly confirms , both in location and relative intensities , the iefs found in the more coarsely sampled positioned - switched spectrum .
figure [ iso2 ] displays the cvf - scanned spectra toward the reflection nebulae vdb17 and vdb59 .
the spectra of both nebulae clearly display the classic 6.2 , 7.7 , 8.6 , 11.3 , and 12.7 iefs that characterize class a type interstellar ief spectra .
both these nebulae are illuminated by hot , uv - rich stars ( @xmath1 = 12,000 k and 19,000 k for vdb17 and vdb59 , respectively ) and are typical of those that display strong interstellar iefs .
figure [ iso3 ] presents 5 15 @xmath0 m spectra of vdb101 , vdb111 , and vdb135 obtained with isocam .
both the position - switched spectra and the fully - sampled cvf spectra of each source are shown , with the exception of vdb111 , toward which only position - switched observations were made .
the fully sampled spectra of vdb101 and vdb135 are generally consistent with their position - switched counterparts obtained toward slightly different positions ( table [ tab : intreg ] ) .
the source and background levels of vdb101 , vdb111 and vdb135 measured by isocam at 12 @xmath31are consistent with those of _
iras_. the reflection nebulae vdb101 , vdb111 , and vdb135 are all illuminated by cool , uv - poor stars ( @xmath1 = 5,000 k , 7,300 k , and 3,600 k , respectively ) .
these three nebulae have very low 12 @xmath31 _ iras _ fluxes , ranging between 0.3 and 0.8 mjy sr@xmath15 .
no iefs , or continuum emission , are apparent in any of the source@xmath30sky spectra above the noise level of the observations , @xmath201 mjy sr@xmath15 .
the task of detecting iefs toward vdb101 , vdb111 , and vdb135 , even with isocam operating under optimal conditions , was understood to be a difficult one .
the predicted sensitivity of isocam ( @xmath200.19 mjy sr@xmath15 with 200s integration time per filter in the position - switched observations ) should have been sufficient , in principle . in practice ,
our actual isocam sensitivity limit was @xmath201 mjy sr@xmath15 for our observations , due to the faintness of the three nebulae compared to the relatively high background emission from diffuse galactic emission along the same line of sight .
the larger than anticipated uncertainty is attributed primarily to the limitations of the rectification procedure to completely correct for the detector effects and the uncertainties in the dark - current subtraction procedure required for the rectification process ( discussed in 3 ) . to achieve
the expected sensitivity of 0.19 mjy sr@xmath15 on top of a typical background level of 20 mjy sr@xmath15 would require that our flux rectification procedure be accurate to @xmath201% .
comparison of the upwards and downwards cvf scans indicated that in practice the rectification is only accurate to @xmath205% .
in the following , we present a quantitative analysis of the iefs as a function of the uv radiation field . we explore the effect of both the hardness of the uv field ( @xmath1 of the illuminating star ) and the absolute intensity of the uv field ( @xmath8 ) .
we expand our analysis over a wider range of @xmath1 and @xmath8 by comparing our observed sample to published spectra of vdb106 ( @xmath29 oph ; boulanger et al .
1996 ) , and vdb139 ( ngc7023 ; d. cesarsky et al .
we also examine spectra at different distances from the illuminating star , for vdb 17 and vdb 59 .
this examines the effect of constant @xmath1 but varying @xmath8 .
the motivation for our _ iso _ observations of reflection nebulae is the observed lack of dependence on @xmath1 of @xmath14(12 @xmath0m)/@xmath13(fir ) in reflection nebulae ( sellgren et al .
1990 ; see 1 and table [ tab : stelpar ] ) .
this is puzzling because current dust models predict that the _ iras _
12 @xmath0 m band is dominated by pah emission and that pahs contribute more to the total interstellar extinction curve at uv than visible wavelengths ( dsert et al .
1990 ; joblin et al . 1992 ; siebenmorgen & krgel 1992 ; schutte et al .
1993 ; dwek et al .
1997 ; li & greenberg 1997 ; silva et al .
1998 ) . because we are interested in analyzing _
iso _ spectra as a function of uv strength , we need to quantify the uv strength in our sources .
we have parameterized the uv field in four ways .
our first method of characterizing the uv field is simply by @xmath11 for the illuminating star of each reflection nebula ( table [ tab : stelpar ] ) .
hotter stars emit a larger fraction of their total energy in the uv .
this is complicated , however , by the difference in @xmath1 for the two stars which illuminate vdb 133 ( @xmath1 = 6,800 k and 12,000 k ; the cooler star is more luminous ) .
another measure of the uv field is the ratio of the stellar flux emitted at uv wavelengths to the total stellar flux .
this simplifies the treatment of the uv field in vdb 133 , since this ratio equals the sum of the uv fluxes of the two stars divided by the sum of the total flux of the two stars .
the relationship between @xmath1 and the ratio of uv to total stellar flux is , however , defined by the longest wavelength included in the range of uv wavelengths .
we have explored two ranges of uv wavelengths , both inspired by previous theoretical , observational , and laboratory work on pahs and the iefs .
the low ratio of visible to uv absorption , for the distribution of pah sizes included in published models of interstellar absorption and emission , occurs because large molecules in general do not have electronic transitions beyond a cutoff wavelength , @xmath32 , proportional to the molecule s length ( platt 1956 ) .
values of @xmath33 = 300 500 nm have been estimated for neutral pahs containing roughly 40 carbon atoms ( crawford , tielens , & allamandola 1985 ; allamandola et al .
1989 ; lger et al . 1989 ; dsert et al . 1990 ; joblin et al .
1992 ; schutte et al .
ionized pahs absorb at longer wavelengths than their neutral counterparts ( salama et al .
1996 and references therein ) .
approximate formulae for the relation between @xmath33 and pah size have been published for neutral pahs ( dsert et al .
1990 ; schutte et al .
1993 ) and for ionized pahs ( dsert et al . 1990 ; salama et al .
1996 ) .
thus , as our second method of parameterizing the uv field , we define a uv wavelength range of 91 400 nm . in this case , the fraction of total stellar flux emitted at uv wavelengths is @xmath34/@xmath5 .
we chose @xmath33 = 400 nm for this definition because it corresponds to the value of @xmath35 , averaged over a distribution of pah sizes and charge , at which current dust models predict the contribution of pahs to the total interstellar extinction curve becomes negligible ( dsert et al .
1990 ; joblin et al . 1992 ; siebenmorgen & krgel 1992 ; dwek et al . 1997 ; silva et al .
note 400 nm falls just longward of the observational dividing line between uv and visible wavelengths .
our third method of parameterizing the uv field is to define a uv wavelength range of 91 240 nm .
many previously published theoretical and observational papers on the ism characterize the ambient radiation field by @xmath8 , the strength of the 91 240 nm uv radiation field .
examples of the prevalence of @xmath8 as a measure of the uv energy density include observational and theoretical studies of the 12 @xmath0 m _ iras _ emission in our own and other galaxies ( draine & anderson 1985 ; ryter , puget , & prault 1987 ; boulanger et al .
1988 ; dsert et al .
1990 ; helou , ryter , & soifer 1991 ; rowan - robinson 1992 ) , analysis of the observed strengths of the iefs from recent _ iso _ and _ irts _ spectra ( boulanger et al . 1996 , 1998a , 1998b ; onaka 1999 ) , and models exploring the ionization state of pahs and their effect on the heating and chemistry of different regions of the ism ( allamandola et al .
1989 ; bakes & tielens 1994 , 1998 ; spaans et al .
1994 ; wolfire et al .
1995 ; salama et al . 1996
; dartois & dhendecourt 1997 ) . for a uv range of 91 240 nm ,
the fraction of total stellar flux emitted at uv wavelengths is @xmath36/@xmath5 .
figure [ uvrat ] displays @xmath4/@xmath5 and @xmath36/@xmath5as a function of @xmath1 , determined using two different stellar models .
we use kurucz ( 1979 ) stellar atmospheric models for @xmath11 @xmath37 5,500k , with log(@xmath38 ) = 4.0 .
we find , however , that the results are not sensitive to the value of log(@xmath38 ) adopted .
since the kurucz models do not extend to cooler temperatures , we use a blackbody ( planck ) function to model stars with @xmath1 @xmath3 5,500k .
figure [ uvrat ] shows that a blackbody slightly overestimates @xmath6/@xmath5 and @xmath39/@xmath5 for the coolest stars for which we have kurucz models .
figure [ uvrat ] and table [ tab : stelpar ] show that @xmath6/@xmath5 drops by a factor of 10 , and @xmath36/@xmath5 falls by over a factor of 100 , as @xmath1 decreases from 21,000k to 5,000k .
our fourth method of parameterizing the uv field incident on a particular nebular position is by @xmath8 ( sometimes called @xmath40 or @xmath41 ) , the strength of the 91 240 nm uv radiation field in units of the local interstellar radiation field .
a value of @xmath42 = 1 corresponds to a uv radiation field of 1.6 @xmath9 10@xmath43 erg s@xmath15 @xmath44 at @xmath2 = 91 240 nm ( habing 1968 ) .
table [ tab : gfac ] presents the distance to each source and the projected angular separation between the illuminating stars and the observed nebular region used to derive @xmath42 .
a single value of @xmath8 suffices for vdb 106 , vdb 133 , and vdb 139 , where the observed spectrum was obtained over a narrow range of distances from the central star , and for vdb 101 , vdb 111 , and vdb 135 , which we did not detect with _ iso _ spectroscopy . for vdb 17 and
vdb 59 , we were able to obtain spectra at a range of values of stellar distance and thus @xmath8 .
table [ tab : gfac ] gives the value of @xmath8 at a typical nebular position for these two nebulae ; we discuss how the spectra of vdb 17 and vdb 59 depend on nebular position ( and thus @xmath8 ) in a later section .
we summarize the uv field for our observed sources and published comparison sources , as measured by @xmath1 , @xmath34/@xmath5 , @xmath45/@xmath5 , and @xmath8 , in tables [ tab : stelpar ] and [ tab : gfac ] .
the observed values of @xmath12(12 @xmath0m)/@xmath13(fir ) ( sellgren et al . 1990 ; see 1 and table [ tab : stelpar ] ) are independent of _ all _ these measures of the uv strength .
another reason to quantify the uv field in reflection nebulae is to search for spectroscopic differences between neutral and ionized pahs .
interstellar pahs are predicted to be mainly positively charged and neutral in reflection nebulae illuminated by b stars , and mainly neutral and negatively charged in the diffuse interstellar medium ( bakes & tielens 1994 , 1998 ; salama et al . 1996 ;
dartois and dhendecourt 1997 ) . predicted differences between neutral and ionized pahs include changes both in the central wavelengths and in the relative strengths of pah features ( de frees et al .
1993 ; szczepanski & vala 1993 ; hudgins , sandford , & allamandola 1994 ; pauzat , talbi , & ellinger 1995 , 1997 ; langhoff 1996 ; cook & saykally 1998 ; allamandola , hudgins , & sandford 1999 ; hudgins & allamandola 1999b ) .
the first ionization potential of pahs lies in the range 4 8 ev , with a tendency for the ionization potential to decrease with increasing pah size ( bakes & tielens 1994 ; salama et al .
hudgins & allamandola ( 1999a ) conclude from the wavelength spacing of the 6.2 and 7.7 @xmath0 m iefs observed in the ism that these iefs arise from ionized pahs with 50 80 carbon atoms .
a pah with 50 80 carbon atoms has a first ionization potential of 5.8 6.2 ev , corresponding to an ionization wavelength of 200 210 nm ( bakes & tielens 1994 ) . the value of @xmath8 ( with @xmath33 = 240 nm ) is thus a rough indicator of the intensity of stellar radiation capable of ionizing pahs at any given nebular position .
theory predicts that the mean ionization state of interstellar pahs is a function of @xmath8 ( bakes & tielens 1994 , 1998 ; salama et al .
1996 ; dartois and dhendecourt 1997 ) .
if the spectral features we observe in reflection nebulae are due to pahs , then we would expect to detect changes both in the central wavelengths and in the relative strengths of iefs between nebulae with different values of @xmath8 .
one goal of our observations is to test this prediction .
localized enhancements of ief emission , at 6.2 , 7.7 , 8.6 , 11.3 , and 12.7 @xmath0 m , were detected in the source@xmath30sky spectra in three of the six nebulae observed : vdb17 , vdb59 and vdb133 .
these three nebulae are illuminated by the hotter stars in our sample , with @xmath1 = 6,800 k 19,000 k. our @xmath201 mjy sr@xmath15 upper limits on the ief emission in vdb101 , vdb111 , and vdb135 are consistent with their 12 @xmath31 _ iras _ surface brightnesses of 0.3 0.8 mjy sr@xmath15 ( sellgren et al .
these faint nebulae have cooler illuminating stars , with @xmath1 = 3,600 k 7,300 k. figure [ lcomp ] displays the superposed spectra of vdb17 , vdb59 , and vdb133 , those nebulae we observe having ief emission in excess of the adjacent sky emission .
these spectra ( from figs .
[ iso1 ] and [ iso2 ] ) are spatial averages over the emission region in each reflection nebula , defined by visual inspection of the images .
each spectrum is scaled by its average intensity over the 12 @xmath0 m _ iras _ band ( @xmath46 m ) .
we choose this scaling to illustrate the ief peak and continuum levels ( in mjy sr@xmath15 ) , normalized to an _ iras _
12 @xmath0 m intensity of 1 mjy sr@xmath15 , for typical reflection nebulae .
such illustration is useful for planning for future mid - infrared spectroscopy with _
sirtf _ and _ sofia _ of sources detected by the _
iras _ all - sky survey .
the reflection nebulae spectra in figure [ lcomp ] , including their continua , broad emission bumps , and narrow iefs , are roughly similar in shape despite the widely varying range of conditions quantified in tables [ tab : stelpar ] and [ tab : gfac ] .
the normalized peak fluxes of the different iefs vary by less than a factor of two from source to source .
more importantly , these variations show no systematic dependence on @xmath1 . no evidence is seen among these three spectra for disappearance of iefs , appearance of new spectral features , shifts in ief wavelengths , or dramatic changes in the strengths of iefs
relative to the broad emission bumps or the continuum emission .
the quantitative fluxes of the narrow iefs , broad emission bumps , and continuum are quite sensitive to how the spectrum is divided into separate spectral components .
figure [ lfit ] displays the spectrum of vdb17 , to illustrate the range of possibilities for dividing the emission between the different spectral components .
figure [ lfit]_a _ shows method 1 , in which we simultaneously fit gaussians to the five _ narrow _ iefs ( at 6.2 , 7.7 , 8.6 , 11.3 and 12.7 ) and to two _ broad _ emission features at @xmath207 and 12 @xmath31 .
the latter two features represent the 6 9 @xmath0 m `` bump '' ( cohen et al . 1986 ; bregman et al . 1989 ) and the 11 13 @xmath0 m `` plateau '' ( cohen , tielens , & allamandola 1985 ) .
the _ continuum _ is approximated by a spline fit through the data points at @xmath205.4 , 10.3 , and 14.8 @xmath31 .
the central wavelengths , widths , and heights of the gaussians were all free parameters in the fit .
boulanger et al .
( 1998a ) have suggested that the iefs are best fit by lorentzians .
we used gaussians rather than lorentzians for the iefs because they provide a better spectral fit to the iefs observed in ngc 7023 at @xmath16 = 1000 ( sellgren et al .
1999 ) .
figure [ lfit]_b _ shows method 2 , in which we simultaneously fit gaussians to the five narrow iefs ( at 6.2 , 7.7 , 8.6 , 11.3 and 12.7 ) . the continuum is modeled by a spline fit to all spectral data points outside that of the narrow iefs ; we assume the broad emission features at 6 9 @xmath0 m and at 11 13 @xmath0 m are not distinct from the continuum but rather are simply undulations in the continuum level in this case . again
the central wavelengths , widths , and heights of the gaussians are free parameters in the fit .
figure [ lfit]_c _ shows method 3 , in which we modeled the continuum by a linear fit to the spectral data points at @xmath205.3 and 15.0 @xmath0 m , the wavelength extremes of the observations .
no broad emission features are assumed in this case .
the fluxes of the five narrow iefs were found by directly integrating over the continuum - subtracted spectra , the method most used in previous studies .
the integration limits adopted for the 6.2 , 7.7 , 8.6 , 11.3 and 12.7 lines were ( 6.0 6.6 @xmath0 m ) , ( 7.2 8.1 @xmath0 m ) , ( 8.1 9.9 @xmath0 m ) , ( 10.5 12.0 @xmath0 m ) , ( 12.0 13.5 @xmath0 m ) , respectively , as indicated in figure [ lfit]_c_. the limits of the 6.2 , 7.7 , 8.6 , and 11.3 @xmath31 iefs are those used by lu ( 1998 ) .
for the 12.7 ief , which lu does not include in his analysis , we integrate over a wavelength range similar in extent ( @xmath18 = 1.5 @xmath0 m ) to that used for the 11.3 line .
figure [ lfit ] illustrates the sensitivity of the estimated strengths of iefs , broad features , and continuum to whether method 1 , 2 , or 3 is used to fit the spectrum .
the gaussian used to fit the 6 9 @xmath0 m broad bump in method 1 is centered near 6.7 @xmath0 m , and as a result the narrow 7.7 @xmath0 m ief and especially the 8.6 @xmath0 m ief appear stronger .
the higher adopted continuum level at 6 9 @xmath0 m in method 2 is centered closer to 7.5 @xmath0 m , which results in a somewhat weaker 7.7 @xmath0 m ief and a much weaker 8.6 @xmath0 m ief .
the choice of a linear continuum ( fitted to points at 5.3 and 15.0 @xmath0 m ) in method 3 consistently produce the highest 6 9 @xmath0 m feature strengths of among the methods . the choice of broad integration limits for the 8.6 @xmath0 m feature ( 8.1 9.9 @xmath0 m ) result in strengths comparable to or exceeding that of the 6.2 @xmath0 m feature .
method 2 and method 3 represent two extremes for the treatment of the 6 9 @xmath0 m region .
previous studies of the relative strengths of different iefs vary in their choices of continuum placement and integration limits in wavelength ; the choice of continuum by cohen et al .
( 1986 , 1989 ) and lu ( 1998 ) in their analyses is similar to method 2 , while that of jourdain de muizon , dhendecourt , & geballe ( 1990 ) and zavagno , cox , & baluteau ( 1992 ) is more like that of methods 1 and 3 .
the sensitivity of our feature strengths to continuum placement , especially , demonstrates how much care must be exercised in intercomparing the results of different studies .
table [ tab : linerat ] presents our quantitative results for the strengths of the iefs , broad emission features , and continuum in vdb17 , vdb59 , vdb133 ( paper i ) , vdb139 ( ngc7023 ; d. cesarsky et al .
1996 ) , and vdb106 ( @xmath29 oph ; boulanger et al . 1996 ) derived by three different methods for defining the iefs and continuum . the methods by which the continuum and iefs are defined , and thus the feature ratios quoted , vary between studies ; the three methods employed here are representative of those typically used .
the individual feature strengths are given in units of integrated line flux ( ergs s@xmath15 @xmath44 sr@xmath15 ) divided by the total integrated 11.3 @xmath0 m line flux ( in ergs s@xmath15 @xmath44 sr@xmath15 ) .
table [ tab : linerat ] shows that the relative strengths we derive for the iefs , broad emission features , and continuum emission are very similar for all the sources we analyze , within a given method for decomposing the spectrum .
the relative ief strengths for any given source , however , are very sensitive to the method chosen .
figure [ lrat ] illustrates this with the 8.6 to 11.3 @xmath0 m ief ratio , @xmath47(8.6 @xmath0m)/@xmath47(11.3 @xmath0 m ) , plotted against @xmath1 .
we originally chose this ratio because it is predicted to be much stronger for ionized pahs than for neutral pahs ( de frees et al .
1993 ; szczepanski & vala 1993 ; hudgins et al . 1994
; pauzat et al .
1995 , 1997 ; langhoff 1996 ; cook & saykally 1998 ; allamandola et al . 1999 ; see 5.4.1 ) .
the value of @xmath1 ( table [ tab : stelpar ] ) is a measure of the hardness of the uv radiation field , which is independent of any adopted value of @xmath48 , and which should be similar for all nebular positions included in the spatially averaged spectrum of each source .
figure [ lrat ] shows , however , that systematic differences in the mean values of @xmath47(8.6 @xmath0m)/@xmath47(11.3 @xmath0 m ) , measured by methods 1 , 2 , and 3 , are much larger than the source - to - source scatter of @xmath47(8.6 @xmath0m)/@xmath47(11.3 @xmath0 m ) for any particular method .
much effort has gone into characterizing the mid - infrared emission of the ism using broadband filters , which provide limited spectral information but higher sensitivity than narrowband filters or spectrometers .
the entire sky has been surveyed in the _ iras _ 12 @xmath0 m filter , and many isocam studies have been conducted with the lw2 and lw3 filters , centered at 6.75 and 15 @xmath0 m respectively .
note that the isocam lw10 filter and the isophot 11.5 @xmath0 m filter are designed to be the same as the _ iras _ 12 @xmath0 m filter .
each of these broad filter bandpasses covers a mixture of iefs , broad emission features , and continuum emission from interstellar dust and molecules .
table [ tab : linerat ] therefore presents the expected _ iras _ 12 @xmath0 m source fluxes and isocam lw2 source fluxes found by convolving the source spectra with the 12 @xmath0 m _ iras _ band transmission profile and lw2 transmission profile ( fig .
[ lfit]_d _ ) , respectively .
we have also investigated what fraction of the flux within these broadband filters comes from each spectral component in the reflection nebulae we have observed .
table [ tab : lineband ] gives the relative contributions of the iefs , broad emission features , and continuum to the broadband fluxes measured with broadband filters in common use .
we derived these results by convolving the isocam cvf spectra with the transmission profile of the lw2 ( 6.75 @xmath0 m ) filter on isocam and the _ iras _ 12 @xmath31 filter .
again , we find the estimated percentage contributions of iefs , broad features , and continuum to the _ iras _ 12 @xmath0 m filter and the isocam lw2 filter are remarkably consistent from nebula to nebula , within a given method ( 1 , 2 , or 3 ) , but the absolute values for any nebula are quite sensitive to which method ( 1 , 2 , or 3 ) is used to fit the spectrum .
the choice of the continua via method 3 gives the maximal values of the narrow ief / total ratios , while those of method 2 represent the lower extreme .
for the _ iras _ 12 @xmath0 m filter , the narrow iefs contribute typically @xmath2020% ( method 1 ) , @xmath2040% ( method 2 ) , and @xmath2060% ( method 3 ) of the total flux depending on the method used to define the continuum .
all methods imply that 40 80% of the _ iras _
12 @xmath0 m flux is due to continuum emission and/or broad emission features , rather than the narrow iefs . the spectra for vdb17 and vdb59 plotted in figures [ iso2 ] and [ lcomp ] , and analyzed in figures [ lfit ] and [ lrat ] and table [ tab : linerat ] , have been integrated over a wide range of projected distance from the central star . in this next section
we explore how the spectral features within a single source image depend on position for vdb 17 and vdb 59 .
we have divided the spectrophotometric images of vdb 17 and 59 into radial bins , characterizing each bin by the incident uv starlight , @xmath8 .
we scale the values of @xmath8 in table [ tab : gfac ] by the inverse square of the projected angular separation of each pixel from the illuminating star .
figure [ rplot ] shows the average spectra of different radial bins within vdb17 and vdb59 .
the values of @xmath8 in these bins are equally spaced in log(@xmath8 ) ; they vary over @xmath8 = 20 4 @xmath9 10@xmath49 for vdb17 , and over @xmath8 = 200 6 @xmath9 10@xmath10 for vdb59 .
near - infrared long - slit spectroscopy of vdb 17 and 59 ( martini , sellgren , & depoy 1999 ) , as well as our unpublished ground - based near - infrared imaging and the _ iso _ images we analyze here , suggest that vdb 17 has a much simpler geometry than vdb 59 .
thus our use of the projected distance between a pixel and the central star to estimate @xmath8 for each pixel is likely to be a better approximation for vdb 17 than for vdb 59 .
figure [ rplot ] shows interesting variations in the spectra of vdb 17 as a function of nebular position ( or @xmath8 ) . at first glance , it appears that the 8.6 @xmath0 m ief steadily becomes less prominent in the spectra at low values of @xmath8 .
a more careful examination , however , shows that the _ peak _ intensity of the 8.6 @xmath0 m ief , relative to the peak intensity of the 7.7 @xmath0 m ief , is relatively constant with @xmath8 .
instead , it seems that the apparent weakening of the 8.6 @xmath0 m ief at low @xmath8 is rather due to a broadening of the overlapping 7.7 @xmath0 m ief .
figure [ fwvsgo ] quantifies the change in the full - width at half - maximum ( fwhm ) of the 7.7 @xmath0 m ief with @xmath8 in vdb 17 .
we adopted the method of boulanger et al .
( 1998a ) to measure the fwhms of different iefs , by fitting a linear baseline to continuum points near 5 and 15 @xmath0 m , and then fitting a blend of three lorentzians to the 6.2 , 7.7 , and 8.6 @xmath0 m iefs .
figure [ fwvsgo ] clearly shows that the 7.7 @xmath0 m ief is systematically broader at low values of @xmath8 in vdb 17 .
figure [ fwvsgo ] also suggests that perhaps the 6.2 @xmath0 m ief may also broaden with lower @xmath8 , but the evidence is not as compelling as it is for the 7.7 @xmath0 m ief .
no change in the fwhm of the 8.6 @xmath0 m ief is observed within the uncertainties .
the spectra of vdb 59 , in contrast to those of vdb 17 , do not show any strong systematic dependence on @xmath8 .
the width of the 7.7 @xmath0 m ief and the contrast of the 8.6 @xmath0 m ief against the red wing of the 7.7 @xmath0 m ief show little change in figure [ rplot ] . this may be due to the more complex geometry of vdb 59 , where emission from regions close to and far from the star appear to contribute to the same line of sight ( martini et al .
1999 ) . alternatively
, this may be because @xmath8 in vdb 59 does not extend to the low values ( @xmath8 = 20 60 ) where the most marked spectral changes are observed in vdb 17 .
the origin of the widths of the iefs is quite controversial and this complicates any explanation of why the 7.7 @xmath0 m ief width changes spatially within vdb17 .
an increase in pah temperature should broaden the iefs , but that would also result in a shift in the ief central wavelengths ( joblin et al .
we , however , do not observe any such shift in the 7.7 @xmath0 m ief central wavelength .
an increase in particle size might broaden the iefs ( boulanger et al .
a change in the mix of the ief emitters , all of which have slightly different wavelengths , could also lead to a change in the ief widths ( cook & saykally 1998 ; le coupanec et al .
1998 ; papoular 1999 ) .
other broadening mechanisms are discussed by the above authors ; whichever mechanism accounts for the broadening of the 7.7 @xmath0 m ief that we observe in vdb17 needs to be consistent with the much lower , or non - existent , broadening we observe in the 6.2 and 8.6 @xmath0 m iefs , and with the constancy in the central wavelengths of all the observed iefs .
the _ peak _ strengths of the 6.2 and 8.6 @xmath0 m iefs , relative to the peak strength of the 7.7 @xmath0 m ief , show little variation among spectra at different distances within vdb17 and vdb 59 , over @xmath8 = 20 6 @xmath9 10@xmath10 ( fig .
[ rplot ] ) .
the peak intensity of the 8.6 @xmath0 m ief of vdb17 , relative to the peak intensity of the 7.7 @xmath0 m ief , has values of 0.53 0.56 in figure [ rplot ] . a similar constancy in peak intensity ( relative to the peak 7.7 @xmath0 m ief intensity )
is observed for the 8.6 @xmath0 m ief within vdb 59 and for the 6.2 @xmath0 m ief within both vdb 17 and vdb 59 .
figure [ rplot ] also shows marginal evidence for a weak emission feature near 6.9 @xmath0 m appearing in the spectra of one or two of the most distant radial bins within vdb 17 .
a weak interstellar emission feature at 6.9 @xmath0 m was first discovered by bregman et al .
( 1983 ) and further analyzed by cohen et al .
this emission feature can be confused with [ ] emission at 7.0 @xmath0 m , and so can only be reliably detected at low spectral resolution in sources free of ionized hydrogen , such as reflection nebulae .
the 6.9 @xmath0 m feature is not detected in spectra of radial bins closer to the star in vdb 17 ( fig .
[ rplot ] ) .
the 6.9 @xmath0 m feature is also not detected in the spectra of any radial bin for vdb 59 ( fig .
[ rplot ] ) , or in the spatially integrated _ iso _ spectra of vdb 17 or vdb 59 ( fig .
[ iso2 ] ) , vdb 133 ( fig .
[ iso1 ] ; paper i ) , vdb 139 ( ngc 7023 ; d. cesarsky et al .
1996 ; laureijs et al . 1996 ) or
vdb 106 ( @xmath29 oph ; boulanger et al .
the 00 s(5 ) h@xmath50 line at 6.9 @xmath0 m is unlikely to be a contributor , as other h@xmath50 lines with similar predicted strengths , at 8.0 , 9.7 , and 12.3 @xmath0 m , are not detected , and because the 1 - 0 s(1 ) h@xmath50 emission in our unpublished narrowband images of vdb 17 is concentrated near the star at higher g@xmath51 values .
we are therefore cautious in claiming that we have detected 6.9 @xmath0 m emission in the outermost regions of vdb17 .
laboratory data and theoretical calculations for pahs show that the 8.6 to 11.3 @xmath0 m ief flux ratio , @xmath47(8.6
@xmath0m)/@xmath47(11.3 @xmath0 m ) , is much higher in ionized than neutral pahs ( de frees et al .
1993 ; szczepanski & vala 1993 ; hudgins et al .
1994 ; pauzat et al .
1995 , 1997 ; langhoff 1996 ; cook & saykally 1998 ; allamandola et al .
the 8.6 @xmath0 m and 11.3 @xmath0 m bands in pahs correspond to the in - plane bend and out - of - plane bend , respectively , of the same aromatic c h bond .
thus @xmath47(8.6 @xmath0m)/@xmath47(11.3 @xmath0 m ) should be a useful measure of pah ionization because it should be less sensitive to pah size or pah hydrogenation than the ratio of an aromatic c c vibration to an aromatic c h vibration .
the mean pah ionization is predicted to increase with @xmath8 ( bakes & tielens 1994 , 1998 ; salama et al .
1996 ; dartois and dhendecourt 1997 ) .
one test of the pah hypothesis , therefore , is to measure @xmath47(8.6 @xmath0m)/@xmath47(11.3 @xmath0 m ) as a function of @xmath8 .
joblin et al . (
1996 ) have observed @xmath47(8.6 @xmath0m)/@xmath47(11.3 @xmath0 m ) to increase closer to the star in the reflection nebula ngc 1333/svs-3 .
they interpret this as due to pahs becoming increasingly ionized near the b - type illuminating star .
figures [ lfit ] , [ lrat ] , and table [ tab : linerat ] , however , demonstrate how strongly the measured value of @xmath47(8.6 @xmath0m)/@xmath47(11.3 @xmath0 m ) in our data depends on the method used to separate the 8.6 @xmath0 m ief from other spectral components such as the 7.7 @xmath0 m ief and the 6 9 @xmath0 m broad emission bump . a local baseline ( method 2 ) to measure the strength of the 8.6 @xmath0 m ief would lead us to conclude that the 8.6 @xmath0 m ief virtually disappears at low values of @xmath8 in vdb 17 , as can be seen qualitatively in figure [ rplot ] .
the observed broadening of the 7.7 @xmath0 m ief at low @xmath8 in vdb 17 ( figure [ fwvsgo ] ) will , however , cause an increase in strength in the wing of the 7.7 @xmath0 m ief . because the red wing of the 7.7 @xmath0 m ief defines the local continuum for the 8.6 @xmath0 m ief in method 2 , any broadening of the 7.7 @xmath0 m ief ( at constant 7.7 @xmath0 m peak intensity ) naturally results in an apparent weakening in the 8.6 @xmath0 m ief strength measured by this method . when we treat the 7.7 and 8.6 @xmath0 m iefs as a blend of two symmetrical features lying above a much lower continuum level ( method 1 )
, then we find that the 8.6 @xmath0 m ief shows little variation with @xmath8 in vdb 17 .
we conclude that @xmath47(8.6 @xmath0m)/@xmath47(11.3 @xmath0 m ) is extremely dependent on how the 8.6 @xmath0 m ief is measured ( figs .
[ lfit ] and [ lrat ] ; table [ tab : linerat ] ) .
joblin et al . (
1996 ) use a local baseline ( method 2 ) in measuring the strength of the 8.6 @xmath0 m ief , because their ground - based spectra do not extend to short enough wavelengths to include the entire 7.7 @xmath0 m ief .
the increase in the 7.7 @xmath0 m ief width with decreasing @xmath8 we observe in vdb 17 ( figs .
[ rplot ] and [ fwvsgo ] ) leads us to urge caution in using @xmath47(8.6 @xmath0m)/@xmath47(11.3 @xmath0 m ) to quantify pah ionization .
the difficulty of quantifying the observed 8.6 @xmath0 m ief strength inspires us to search for a more robust way to test predictions of the pah ionization state as a function of @xmath8 .
laboratory data and theoretical calculations for pahs show that the sum of all the pah bands at 6 10 @xmath0 m , relative to the sum of all the pah bands at 10 14 @xmath0 m ,
is at least an order of magnitude higher in ionized pahs than in neutral pahs ( de frees et al .
1993 ; szczepanski & vala 1993 ; hudgins et al . 1994 ; pauzat et al .
1995 , 1997 ; langhoff 1996 ; cook & saykally 1998 ; allamandola et al .
we therefore have measured @xmath47(5.509.75 @xmath0m)/@xmath47(10.2514.0 @xmath0 m ) , the ratio of the integrated intensity at 5.50 9.75 @xmath0 m to the integrated intensity at 10.25 14.0 @xmath0 m , in _ iso _ spectra of reflection nebulae , as an alternate way to search for the spectroscopic signature of changing pah ionization state .
figure [ splrat ] presents our measurements of @xmath47(5.509.75 @xmath0m)/@xmath47(10.2514.0 @xmath0 m ) as a function of @xmath8 in reflection nebulae .
we subtracted a continuum from each spectrum before measuring @xmath47(5.509.75 @xmath0m)/@xmath47(10.2514.0
@xmath0 m ) .
the 5.50 9.75 @xmath0 m continuum was defined by a linear fit to data points near 5.3 and 10.0 @xmath0 m .
the 10.25 14.0 @xmath0 m continuum was defined by a linear fit to data points near 10.0 and 14.5 @xmath0 m .
we plot a single value of @xmath47(5.509.75 @xmath0m)/@xmath47(10.2514.0 @xmath0 m ) for vdb 106 ( @xmath29 oph ; boulanger et al .
1996 ) , vdb 133 ( fig .
[ iso1]_b _ and paper i ) , and vdb 139 ( ngc 7023 ; d. cesarsky et al .
1996 ) . for vdb 17 and vdb 59 , we plot @xmath47(5.509.75 @xmath0m)/@xmath47(10.2514.0 @xmath0 m ) at six values of @xmath8 , derived from the spectra of the radial bins shown in figure [ rplot ] .
the remarkable conclusion from figure [ splrat ] is that @xmath47(5.509.75 @xmath0m)/@xmath47(10.2514.0 @xmath0 m ) changes by no more than 40% among sources with @xmath8 = 20 6 @xmath9 10@xmath10 .
this observational ratio should not depend at all on how the observed interstellar spectra are divided into separate narrow iefs and broader emission components .
we emphasize that @xmath47(5.509.75 @xmath0m)/@xmath47(10.2514.0 @xmath0 m ) should be a more robust and sensitive measure of the pah ionization state than @xmath47(8.6 @xmath0m)/@xmath47(11.3 @xmath0 m ) .
our observation that the mid - infrared spectra of vdb 17 and vdb 59 are very similar for @xmath8 = 20 6 @xmath9 @xmath52 ( figs .
[ rplot ] and [ splrat ] ) are in strong agreement with other recent work .
boulanger et al .
( 1998b ) concludes that the ief flux ratios in _ iso _ spectra of the chamaeleon cloud , the @xmath29 oph cloud , ngc 7023 , and m17 show no systematic dependence on @xmath8 , over @xmath8 = 1 10@xmath53 .
onaka ( 1999 ) finds that the ief flux ratios in _ iso _ spectra of the carina nebula are independent of @xmath8 , over @xmath8 = 100 10@xmath53 .
we have obtained 5 15 @xmath0 m isocam cvf spectroscopy of six reflection nebulae , in order to test predictions of interstellar models for the infrared emission features .
our sample includes a range of uv field hardness ( @xmath1 = 3,600 k 19,000 k for the illuminating stars ; @xmath6/@xmath5= 0.01 0.87 ) and uv field intensity ( @xmath8 = 20 6 @xmath9 10@xmath10 ) .
our first goal was to search for spectroscopic differences as a function of @xmath1 for the illuminating stars .
detection of such spectroscopic differences could potentially reconcile 12 @xmath0 m broad - band _ iras _ observations showing that @xmath12(12 @xmath0m)/@xmath13(fir ) is independent of @xmath1 , for reflection nebulae illuminated by stars @xmath1 = 5,000 33,000 k ( sellgren et al . 1990 ) , with model predictions that the _ iras _ 12 @xmath0 m emission is dominated by pahs which contribute much more to the total interstellar extinction curve at uv wavelengths than at visible wavelengths ( dsert et al .
1990 ; joblin et al . 1992
; siebenmorgen & krgel 1992 ; dwek et al .
1997 ; silva et al . 1998 ) .
a second goal was to search for spectroscopic differences as a function of @xmath8 .
interstellar models predict that more pahs are positively charged when @xmath8 increases ( bakes & tielens 1994 , 1998 ; salama et al .
1996 ; dartois and dhendecourt 1997 ) .
laboratory data and theoretical calculations for pahs show that @xmath47(8.6 @xmath0m)/@xmath47(11.3 @xmath0 m ) and @xmath47(5.509.75 @xmath0m)/@xmath47(10.2514.0 @xmath0 m ) are both much higher in ionized than neutral pahs ( de frees et al .
1993 ; szczepanski & vala 1993 ; hudgins et al . 1994 ; pauzat et al .
1995 , 1997 ; langhoff 1996 ; cook & saykally 1998 ; allamandola et al .
1999 ) .
we detect _ interstellar _ iefs at 6.2 , 7.7 , 8.6 , 11.3 , and 12.7 @xmath0 m , broad emission features at 6 9 @xmath0 m and 11 13 @xmath0 m , and continuum emission at 5 15 @xmath0 m in the reflection nebulae vdb 17 , vdb 59 , and vdb 133 .
these nebulae were the brightest in our sample and illuminated by hotter stars ( @xmath4/@xmath5 = 0.22 0.87 ) .
the normalized peak fluxes of the different iefs vary by less than a factor of two among these three nebulae , with no systematic dependence on @xmath1 .
no evidence is seen among these spectra for disappearance of iefs , appearance of new spectral features , shifts in ief wavelengths , or dramatic changes in the strengths of iefs relative to the broad emission bumps or the continuum emission .
we do not detect any feature emission or continuum emission in vdb 101 , vdb 111 , and vdb 135 .
these nebulae were the faintest in our sample and were illuminated by cooler stars ( @xmath4/@xmath5 = 0.01 0.20 ) .
their upper limits are consistent with their low measured 12 @xmath0 m _ iras _ surface brightnesses .
we compare spatially averaged spectra of vdb 17 , vdb 59 , and vdb 133 to published spectra of vdb 106 ( @xmath29 oph ; boulanger et al .
1996 ) and vdb 139 ( ngc 7023 ; d. cesarsky et al .
qualitative comparison of spectra show all these spectra are very similar over a range of uv hardness ( @xmath1 = 6,800 22,000 ; @xmath54/@xmath5 = 0.22 0.92 ) and uv intensity ( @xmath8 = 40 1800 ) .
we find that quantifying the ief fluxes is quite sensitive to the method by which the spectra are separated into iefs , broad emission bumps , and continuum .
for instance , @xmath47(8.6 @xmath0m)/@xmath47(11.3 @xmath0 m ) shows no dependence on @xmath1 for any given method of defining the ief fluxes , but @xmath47(8.6
@xmath0m)/@xmath47(11.3 @xmath0 m ) shows strong systematic differences between the values derived by different methods .
we also quantify the contributions of iefs , broad emission bumps , and continuum to broad - band filters such as the _ iras _ 12 @xmath0 m filter and the isocam lw2 ( 6.75 @xmath0 m ) filter .
again , for any given method of distinguishing between these separate spectral components , the contributions of each component to the broad - band filters are independent of @xmath1 , but the absolute value of the contribution depends strongly on the method for how the spectrum is divided among iefs , broad emission bumps , and continuum .
we divide the spectra of vdb 17 and vdb 59 into radial bins , with the uv field of each bin characterized by @xmath8 .
we find that the _ peak _ intensities of the 6.2 and 8.6 @xmath0 m iefs , compared to the peak intensity of the 7.7 @xmath0 m ief , show no variation over @xmath8 = 20 6 @xmath9 10@xmath10 in vdb 17 and vdb 59 . in the most distant regions of vdb 17
, we observe that the 8.6 @xmath0 m ief appears weaker due to increased blending with the 7.7 @xmath0 m ief .
we quantify the fwhm of the 7.7 @xmath0 m feature , and find that it grows systematically broader at low @xmath8 in vdb 17 .
we conclude that the low contrast of the 8.6 @xmath0 m ief at low @xmath8 in vdb 17 is the result of the increased width of the 7.7 @xmath0 m ief .
this observed broadening of the 7.7 @xmath0 m ief , combined with the sensitivity of @xmath47(8.6 @xmath0m)/@xmath47(11.3 @xmath0 m ) to baseline placement for the 8.6 @xmath0 m ief , suggests that caution should be used in inferring pah ionization from @xmath47(8.6 @xmath0m)/@xmath47(11.3 @xmath0 m ) .
we suggest using @xmath47(5.509.75 @xmath0m)/@xmath47(10.2514.0 @xmath0 m ) as an alternate method of testing the predictions of pah ionization as a function of @xmath8 .
we measure this in single nebular positions in vdb 106 , vdb 133 , and vdb 139 and at different nebular positions within vdb 17 and vdb 59 .
we observe that @xmath47(5.509.75 @xmath0m)/@xmath47(10.2514.0 @xmath0 m ) decreases by no more than 40% over @xmath8 = 20 6 @xmath9 10@xmath10 .
our observational results are in strong agreement with boulanger et al .
( 1998b ) , who concludes that the ief flux ratios in _ iso _ spectra of the chamaeleon cloud , the @xmath29 oph cloud , ngc 7023 , and m17 show no systematic dependence on @xmath8 over @xmath8 = 1 10@xmath53 .
our results also strongly support onaka ( 1999 ) , who finds that the ief flux ratios in _ iso _ spectra of the carina nebula are independent of @xmath8 over @xmath8 = 100 10@xmath53 .
much is yet to be gleaned from this _ iso _ survey of reflection nebulae and especially the detection of iefs toward vdb133 . in a future paper
we will present a quantitative analysis of the ief energy budget using laboratory absorption profiles of various carrier candidates ( neutral pahs , ionized pahs , amorphous hydrocarbons , and coal tars ) .
we will also present follow - up 7.75 ( lw6 ) _ iso _ observations which cover an 8@xmath55 by 8@xmath55 mosaic around vdb133 , detailing the spatial distribution of the surrounding ief emission .
the relation of the ief carriers to the gas and other components of dust surrounding vdb133 will be explored by comparing the _ iso _ images to ground - based infrared and molecular line ( @xmath56co j=21 ) data that have also been obtained .
we would like to thank a. lger , jean - loup puget and c. moutou for hosting us at the institut dastrophysique spatiale .
we are grateful to w. reach , d. cesarsky and f. boulanger for providing their expertise and invaluable help in the reduction of our isocam data .
we appreciate the generosity of d. cesarsky and colleagues for providing us with their isocam spectrum of ngc 7023 , and f. boulanger and colleagues for providing us with their isocam spectrum of @xmath29 oph .
we acknowledge nasa support of the _ iso _ data analysis through nag 53366 and jpl contract 961562 , and also support from nato collaborative research grant 951347 .
this work was carried out in part at the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
abergel , a. , et al .
1996 , , 315 , l329 aitken , d. k. , 1981 , in iau symp .
96 , infrared astronomy , ed . c. g. wynn - williams & d. p. cruikshank ( dordrecht : reidel ) , 207 allamandola , l. j. , hudgins , d. m. , & sandford , s. a. 1999 , , 511 , l115 allamandola , l. j. , tielens , a. g. g. m. , & barker , j. r. 1985 , , 290 , l25 allamandola , l. j. , tielens , a. g. g. m. , & barker , j. r. 1989 , , 71 , 733 bakes , e. l. o. , & tielens , a. g. g. m. 1994 , , 427 , 822 bakes , e. l. o. , & tielens , a. g. g. m. 1998 , , 499 , 258 blanco , a. , bussoletti , e. , & colangeli , l. 1988 , , 334 , 875 borghesi , a. , bussoletti , e. , & colangeli , l. 1987 , , 314 , 422 boulanger , f. , 1996 , private communication boulanger , f. , beichman , c. , deert , f .- x .
, helou , g. , prault , m. , & ryter , c. 1988 , , 332 , 328 boulanger , f. , et al .
1996 , , 315 , l325 boulanger , f. , boisssel , p. , cesarsky , d. , & ryter , c. 1998a , , 339 , 194 boulanger , f. , et al . 1998b , in asp conf .
132 , star formation with the infrared space observatory , ed .
j. l. yun & r. liseau ( san francisco : asp ) , 15 bregman , j. d. , allamandola , l. j. , witteborn , f. c. , tielens , a. g. g. m. , geballe , t. r. 1989 , , 344 , 791 bregman , j. d. , dinerstein , h. l. , goebel , j. h. , lester , d. f. , witteborn , f. c. , & rank , d. m. 1983 , , 274 , 666 brooke , t. y. , sellgren , k. , & geballe , t. r. 1999 , , 517 , 883 brooke , t. y. , sellgren , k. , & smith , r. g. 1996 , , 459 , 209 buss , r. h. jr . , tielens , a. g. g. m. , cohen , m. , werner , m. w. , bregman , j. d. , & witteborn , f. c. 1993 , , 415 , 250 cesarsky , c. j. , et al .
1996 , , 315 , l32 cesarsky , d. , lequeux , j. , abergel , a. , prault , m. , palazzi , e. , madden , s. , tran , d. 1996 , , 315 , l309 chlewicki , g. & laureijs , r. j. 1988 , , 207 , l11 cohen , m. , allamandola , l. j. , tielens , a. g. g. m. , bregman , j. d. , simpson , j. p. , witteborn , f. c. , wooden , d. h. , rank , d. m. 1986 , , 302 , 737 cohen , m. , tielens , a. g. g. m. , & allamandola , l. j. 1985 , , 299 , l93 cohen , m. , tielens , a. g. g. m. , bregman , j. d. , witteborn , f. c. , rank , d. m. , allamandola , l. j. , wooden , d. h. , jourdain de muizon , m. 1989 , , 341 , 246 cook , d. j. , & saykally , r. j. 1998 , , 493 , 793 crawford , m. k. , tielens , a. g. g. m. , & allamandola , l. j. 1985 , , 293 , l45 dartois , e. , & dhendecourt , l. 1997 , , 323 , 534 de frees , d. j. , miller , m. d. , talbi , d. , pauzat , f. , & ellinger , y. 1993 , , 408 , 530 de geus , e. j. , de zeeuw , p. t. , & lub , j. 1989 , , 216 , 44 dsert , f .- x .
, boulanger , f. , & puget , j. l. 1990 , , 237 , 215 draine , b. t. , & anderson , n. 1985 , , 292 , 494 duley , w. w. 1988 , , 234 , 61 duley , w. w. , & williams , d. a. 1981 , , 196 , 269 duley , w. w. , & williams , d. a. 1988 , , 230 , 1p dwek , e. , et al .
1997 , , 475 , 565 garmany , c. d. , & stencel , r. e. 1992 , a&as , 94 , 211 geballe , t. r. 1997 , in asp conf .
122 , from stardust to planetesimals , ed .
y. j. pendleton & a. g. g. m. tielens ( san francisco : asp ) , 119 geballe , t. r. , tielens , a. g. g. m. , kwok , s. , & hrivnak , b. j. 1992 , , 387 , l89 guillois , o. , nenner , i. , papoular , r. , & reynaud , c. 1996 , , 464 , 810 habing , h. j. 1968 , , 19 , 421 helou , g. , ryter , c. , & soifer , b. t. 1991 , , 376 , 505 hudgins , d. m. , sandford , s. a. , & allamandola , l. j. 1994 , j. phys .
98 , 4243 hudgins , d. m. , & allamandola , l. j. 1999a , , 513 , l69 hudgins , d. m. , & allamandola , l. j. 1999b , , 516 , l41 humphreys , r. m. 1978 , , 38 , 309 isocam team 1994 , in `` isocam observer s manual '' joblin , c. , boissel , p. , lger , a. , dhendecourt , d. , & dfourneau , d. 1995 , , 299 , 835 joblin , c. , lger , a. , & martin , p. 1992
, , 393 , l79 joblin , c. , tielens , a. g. g. m. , geballe , t. r. , & wooden , d. h. 1996 , , 460 , l119 jourdain de muizon , m. , dhendecourt , & geballe , t. r. 1990 , , 227 , 526 kessler , m. f. , et al .
1996 , , 315 , l27 kurucz , r. l. 1979 , , 40 , 1 kwok , s. , volk , k. , & hrivnak , b. j. 1989 , , 345 , l51 langhoff , s. r. 1996 , j. phys .
, 100 , 2819 laureijs , r. j. , et al . 1996 , a&a , 315 , l317 le coupanec , p. , rouan , d. , & lger , a. 1998 , , 338 , 217 lger , a. , & puget , j. l. 1984 , , 137 , l5 lger , a. , verstraete , l. , dhendecourt , l. , dfourneau , d. , dutuit , o. , schmidt , w. , & lauer , j. c. 1989 , in iau symp .
135 , interstellar dust , ed .
l. j. allamandola & a. g. g. m. tielens ( dordrecht : kluwer ) , 173 li , a. , & greenberg , j. m. 1997 , , 323 , 566 lu , n. y. 1998 , , 498 , l65 martini , p. , sellgren , k. , & depoy , d. l. 1999 , apj , in press ( astro - ph/9905028 ) onaka , t. 1999 , adv .
research , in press papoular , r. , conard , j. , guiliano , m. , kister , j. , & mille , g. 1989 , , 217 , 204 papoular , r. 1999 , , 346 , 219 pauzat , f. , talbi , d. , & ellinger , y. 1995 , , 293 , 263 pauzat , f. , talbi , d. , & ellinger , y. 1997 , , 319 , 318 platt , j. r. 1956 , , 123 , 486 puget , j. l. , lger , a. 1989 , , 27 , 161 puget , j. l. , lger , a. & boulanger , f. 1985 , , 142 , l19 racine , r. 1968 , 73 , 233 roche , p. f. , & aitken , d. k. 1986 , , 221 , 63 rowan - robinson , m. 1992 , mnras , 258 , 787 ryter , c. , puget , j. l. , & prault , m. 1987 , , 186 , 312 sakata , a. , wada , s. , onaka , t. , & tokunaga , a. 1987 , , 320 , l63 sakata , a. , wada , s. , tanabe , t. , & onaka , t. 1984 , , 287 , l51 salama , f. , bakes , e. l. o. , allamandola , l. j. , & tielens , a. g. g. m. 1996 , , 458 , 621 schutte , w. a. , tielens , a. g. g. m. , & allamandola , l. j. 1993 , , 415 , 397 schutte , w. a. , et al .
1996 , , 315 , l333 schutte , w. a. , et al .
1998 , , 337 , 261 sellgren , k. 1984 , , 277 , 623 sellgren , k. 1990 , in dusty objects in the universe , ed .
e. bussoletti & a. a. vittone ( dordrecht : kluwer ) , 35 sellgren , k. 1994 , in the infrared cirrus and diffuse interstellar clouds , ed .
r. m. cutri & w. b. latter ( san francisco : asp ) , 243 sellgren , k. , brooke , t. y. , smith , r. g. , & geballe , t. r. 1995 , , 449 , l69 sellgren , k. , luan , l. , & werner , m. w. 1990 , , 359 , 384 sellgren , k. , smith , r. g. , & brooke , t. y. 1994 , , 433 , 179 sellgren , k. , werner , m. w. , & allamandola , l. j. 1996 , , 102 , 369 sellgren , k. , werner , m. w. , & dinerstein , h. l. 1983 , , 271 , l13 sellgren , k. , et al .
1999 , in preparation siebenmorgen , r. , & krgel , e. 1992 , , 259 , 614 silva , l. , granato , g. l. , bressan , a. , & danese , l. 1998 , , 509 , 103 spaans , m. , tielens , a. g. g. m. , van dishoeck , e. f. , & bakes , e. l. o. 1994 , , 437 , 270 strom , s. e. , grasdalen , g. l. , & strom , k. m. 1974 , , 191 , 111 strom , k. m. , strom , s. e. , carrasco , l. , vrba , f. j. 1975 , , 196 , 489 szczepanski , j. & vala , m. 1993 , , 414 , 646 tokunaga , a. 1997 , in asp conf .
124 , diffuse infrared radiation and the irts , ed .
h. okuda , t. matsumoto , & t. l. roellig ( san francisco : asp ) , 149 uchida , k. i. , sellgren , k. , & werner , m. 1998 , , 493 , l109 ( paper i ) whitcomb , s. e. , gatley , i. , hildebrand , r. h. , keene , j. , sellgren , k. & werner , m. w. 1981 , , 246 , 416 witt , a. n. , cottrell , m. j. 1980 , , 235 , 899 wolfire , m. g. , hollenbach , d. , mckee , c. f. , tielens , a. g. g. m. , & bakes , e. l. o. 1995 , , 443 , 152 zavagno , a. , cox , p. , & baluteau , j .-
1992 , , 259 , 241 1a & 5.8 , 6.2 , 7.0 , 9.5 & 101 , 133 , 135 1b & 4.5@xmath63 , 6.2 , 7.0 , 9.5 & 111 2 & 7.0 , 7.6 , 8.4 & 101 , 111 , 133 , 135 3 & 8.4 , 8.62 , 9.5 & 101 , 111 , 133 4 & 10.5 , 11.22 , 12.0 , 15.0@xmath64 & 101 , 111 , 133 , 135 & & & & vdb17 ( ngc1333 ) & @xmath5914 & @xmath3022 & 60 & 60 & 77.7 & le vdb59 ( ngc2068 ) & @xmath5990 & @xmath5928 & 72 & 78 & 5.8 & le vdb101 & @xmath5934 & @xmath59227 & 192 & 192 & 10.1 & er vdb133 & @xmath5985 & @xmath5943 & 60 & 48 & 83.8 & le vdb135 & @xmath30104 & @xmath59204 & 192 & 192 & 65.5 & er iefs & ( 6.2 @xmath31)/(11.3 @xmath31 ) & 1.4 & 1.7 & 1.0 & 1.1 & 1.2 & ( 7.7 @xmath31)/(11.3 @xmath31 ) & 3.8 & 4.4 & 4.8 & 3.4 & 4.1 & ( 8.6 @xmath31)/(11.3 @xmath31 ) & 2.0 & 2.3 & 2.9 & 1.8 & 2.0 & ( 12.7 @xmath31)/(11.3 @xmath31 ) & 0.18 & 0.30 & 0.29 & 0.21 & 0.21 iefs & ( 6.2 @xmath31)/(11.3 @xmath31 ) & 1.2 & 1.7 & 1.3 & 1.2 & 1.4 & ( 7.7 @xmath31)/(11.3 @xmath31 ) & 1.7 & 2.0 & 2.3 & 2.0 & 2.0 & ( 8.6 @xmath31)/(11.3 @xmath31 ) & 0.28 & 0.44 & 0.29 & 0.24 & 0.36 & ( 12.7 @xmath31)/(11.3 @xmath31 ) & 0.18 & 0.18 & 0.24 & 0.30 & 0.23 continuum & ( 5.2 15.1 @xmath31)/(11.3 @xmath31 ) & 14.0 & 11.0 & 15.0 & 13.0 & 15.0 & & & & & iefs & ( 6.2 @xmath31)/(11.3 @xmath31 ) & 1.4 & 1.8 & 1.0 & 1.2 & 1.0 & ( 7.7 @xmath31)/(11.3 @xmath31 ) & 2.3 & 3.0 & 2.4 & 2.5 & 2.1 & ( 8.6 @xmath31)/(11.3 @xmath31 ) & 1.4 & 1.7 & 1.7 & 1.3 & 1.3 & ( 12.7 @xmath31)/(11.3 @xmath31 ) & 0.87 & 0.50 & 0.50 & 0.63 & 0.60 continuum & ( 5 15 @xmath31)/(11.3 @xmath31 ) & 3.6 & 3.7 & 2.3 & 3.1 & 2.6 & & & & & method 1 ( 11.3 @xmath31 ) & & 1.3e@xmath65 & 2.8e@xmath65 & 1.5e@xmath66 & @xmath67 & 3.3e@xmath66 method 2 ( 11.3 @xmath31 ) & & 1.3e@xmath65 & 3.2e@xmath65 & 1.5e@xmath66 & @xmath67 & 3.3e@xmath66 method 3 ( 11.3 @xmath31 ) & & 2.2e@xmath65 & 4.0e@xmath65 & 3.2e@xmath66 & @xmath67 & 6.9e@xmath66 | we present 5 15 @xmath0 m imaging spectroscopy of the reflection nebulae vdb17 ( ngc 1333 ) , vdb59 ( ngc 2068 ) , vdb101 , vdb111 , vdb133 and vdb135 , obtained with the infrared camera and circular variable filter wheel on the _ infrared space observatory _ ( _ iso _ ) .
these nebulae are illuminated by stars with @xmath1 = 3,600 k 19,000 k , implying ultraviolet ( uv ; @xmath2 @xmath3 400 nm ) to total stellar flux ratios of @xmath4/@xmath5 = 0.01 0.87 .
we detect the infrared emission features ( iefs ) at 6.2 , 7.7 , 8.6 , 11.3 , and 12.7 , broad emission features at 6 9 @xmath0 m and 11 13 @xmath0 m , and 5 15 @xmath0 m continuum emission , from the interstellar medium in vdb17 , vdb59 , and vdb133 ( @xmath6/@xmath5 = 0.22 0.87 ) , and place upper limits on the emission from the interstellar medium in vdb101 , vdb111 , and vdb135 ( @xmath7/@xmath5 = 0.01 0.20 ) .
our goal is to test predictions of models attributing the iefs to polycyclic aromatic hydrocarbons ( pahs ) .
interstellar models predict pahs change from singly ionized to neutral as the uv intensity , @xmath8 , decreases . the ratio of pah emission at 6 10 @xmath0 m to pah emission at 10 14
@xmath0 m is expected to be ten times higher in ionized than neutral pahs .
we observe no spectroscopic differences with varying @xmath1 .
we analyze the spectra of vdb 17 and vdb 59 as a function of distance from the star , to see how the spectra depend on @xmath8 within each source . the only quantitative difference we find is a _ broadening _ of the 7.7 @xmath0 m ief at @xmath8 = 20 60 within vdb 17 .
we observe only a 40% change in the 6 10 @xmath0 m to 10 14 @xmath0 m flux ratio over @xmath8 = 20 to 6 @xmath9 10@xmath10 .
accepted for publication in the astrophysical journal , 5 october 1999 |
the discovery of neutrino oscillations @xcite marks a turning point in our understanding of nature and brings neutrino physics to the center of attention of the particle , nuclear and astrophysics communities .
the existence of small neutrino masses confirms theoretical expectations which date back to the early eighties .
they arise from the dimension - five operator @xmath2 where @xmath3 is the 1 higgs doublet and @xmath4 is a lepton doublet @xcite .
nothing is known from first principles about the mechanism that induces this operator , its associated mass scale or flavour structure .
its most popular realization is the seesaw mechanism @xcite which induces small neutrino masses from the exchange of heavy states , as expected in unified models .
in addition to the standard 1 doublet higgs multiplet whose vacuum expectation value ( vev ) generates gauge boson and charged fermion masses , the full seesaw model ( now called type - ii as opposed to the original terminology in @xcite ) , contains higgses transforming as 1 singlet and triplet , carrying 2 units of lepton number , and with vevs @xmath5 obeying @xmath6 with @xmath7 ( i=1,2,3 correspond to singlet , doublet and triplet , respectively ) . the resulting perturbative description of the seesaw is first given in the second paper in @xcite , while its _ effective _ model - independent low - energy description involves a 3@xmath86 charged current lepton mixing matrix which has 24 parameters .
these correspond to 12 mixing angles and 12 cp phases ( both dirac and majorana - type ) , given in @xcite ( first paper ) .
some of these parameters are involved in leptogenesis @xcite .
current neutrino oscillation data are well described by the _ simplest _ ( unitary approximation to the ) lepton mixing matrix neglecting cp violation . here
i focus mainly on the determination of neutrino mass and mixing parameters in neutrino oscillation studies , a currently thriving industry @xcite , with many new experiments underway or planned .
the interpretation of the data requires good solar and atmospheric neutrino flux calculations @xcite , neutrino cross sections and experimental response functions , and a careful description of matter effects @xcite in the sun and the earth . in the early eighties it was
also argued that , on quite _ general _ grounds , well beyond the details of the seesaw mechanism , massive neutrinos should be majorana particles @xcite , leading to l - violating processes such as . after summarizing the status of 3-neutrino oscillation parameters
i briefly discuss their impact on future searches @xcite .
the solar neutrino data includes the measured rates of the chlorine experiment at the homestake mine ( @xmath9 snu ) , the most up - to - date gallium results of sage ( @xmath10 snu ) and gallex / gno ( @xmath11 snu ) , as well as the 1496day super - k data in the form of 44 bins ( 8 energy bins , 6 of which are further divided into 7 zenith angle bins ) .
the sno data include the most recent data from the salt phase in the form of the neutral current ( nc ) , charged current ( cc ) and elastic scattering ( es ) fluxes , as well as the 2002 spectral day / night data ( 17 energy bins for each day and night period ) .
the analysis methods are described in @xcite and references therein .
we use a generalization of the pull approach for the @xmath12 calculation originally suggested in ref .
@xcite in which all systematic uncertainties such as those of the eight solar neutrino fluxes are included by introducing new parameters in the fit and adding a penalty function to the @xmath12 .
our generalized method is exact to all orders in the pulls and covers the case of correlated statistical errors @xcite as necessary to treat the sno
salt experiment .
this is particularly interesting as it allows us to include the standard solar model @xmath13b flux prediction as well as the sno nc measurement on the same footing , without pre - selecting a particular value , as implied by expanding around the predicted value : the fit itself chooses the best compromise between the sno nc data and the ssm prediction .
kamland detects reactor anti - neutrinos at the kamiokande site by the process @xmath14 , where the delayed coincidence of the prompt energy from the positron and a characteristic gamma from the neutron capture allows an efficient reduction of backgrounds .
most of the incident @xmath15 flux comes from nuclear plants at distances of @xmath16 km from the detector , far enough to probe the lma solution of the solar neutrino problem .
the neutrino energy is related to the prompt energy by @xmath17 , where @xmath18 is the neutron - proton mass difference and @xmath19 is the positron mass . for lower energies
there is a relevant contribution from geo - neutrino events to the signal @xcite . to avoid large uncertainties associated with the geo - neutrino flux an energy cut at 2.6 mev prompt
energy is applied for the oscillation analysis .
first kamland data corresponding to a 162 ton - year exposure gave 54 anti - neutrino events in the final sample , after all cuts , while @xmath20 events are predicted for no oscillations with @xmath21 background events .
the probability that the kamland result is consistent with the no disappearance hypothesis is less than 0.05% .
this gave the first evidence for the disappearance of neutrinos traveling to a detector from a power reactor and the first terrestrial confirmation of the solar neutrino anomaly . with a somewhat larger fiducial volume of the detector
an exposure corresponding to 766.3 ton - year ( including a reanalysis of the previous 2002 data ) has been given @xcite .
in total 258 events have been observed , versus @xmath22 reactor neutrino events expected in the case of no disappearance and @xmath23 background events .
this leads to a confidence level of 99.995% for @xmath24 disappearance , and the averaged survival probability is @xmath25 .
moreover evidence for spectral distortion is obtained @xcite .
it is convenient to bin the latest kamland data in @xmath26 , instead of the traditional bins of equal size in @xmath27 .
various systematic errors associated to the neutrino fluxes , backgrounds , reactor fuel composition and individual reactor powers , small matter effects , and improved @xmath15 flux parameterization are included @xcite .
kamland data are in beautiful agreement with the region implied by the lma solution to the solar neutrino problem , which in this way has been singled out as the only viable one in contrast to the previous variety of oscillation solutions @xcite .
however the stronger evidence for spectral distortion in the recent data leads to improved information on @xmath28 , substantially reducing the allowed region of oscillation parameters . from this point of view kamland
has played a fundamental role in the resolution of the solar neutrino problem .
assuming cpt one can directly compare the information obtained from solar neutrino experiments with the kamland reactor results . in fig .
[ fig : solkam - region ] we show the allowed regions from the combined analysis of solar and kamland data . -.2 cm regions allowed by solar and reactor data at 90% , 95% , 99% , and 3@xmath29 c.l . for 2 d.o.f
.. unshaded regions correspond to solar data only.,title="fig:",height=188 ] -.3 cm -.5
cm the zenith angle dependence of the @xmath30-like atmospheric neutrino data from the super - k experiment provided the first evidence for neutrino oscillations in 1998 , an effect confirmed also by other atmospheric neutrino experiments @xcite .
the dip in the @xmath31 distribution of the atmospheric @xmath32 survival probability seen in super - k gives a clearer signature for neutrino oscillations .
the analysis summarized below includes the most recent charged - current atmospheric neutrino data from super - k , with the @xmath33-like and @xmath30-like data samples of sub- and multi - gev contained events grouped into 10 zenith - angle bins , with 5 angular bins of stopping muons and 10 through - going bins of up - going muon events .
no information on @xmath34 appearance , multi - ring @xmath30 and neutral - current events is used since an efficient monte - carlo simulation of these data would require more details of the super - k experiment , in particular of the way the neutral - current signal is extracted from the data ( more details in refs .
in contrast to previous analyses using the bartol fluxes @xcite , here we use three dimensional atmospheric neutrino fluxes @xcite .
this way one obtains the regions of two - flavour @xmath35 oscillation parameters @xmath36 and @xmath37 shown by the hollow contours in fig .
[ fig : atm+k2k ] .
note that the @xmath37 values obtained with the three dimensional atmospheric neutrino fluxes are lower than obtained previously @xcite , in good agreement with the results of the super - k collaboration @xcite .
@xmath36@xmath37 regions allowed at 90% , 95% , 99% , and 3@xmath29 c.l . for 2 d.o.f .
( unshaded regions include atmospheric data only ) . , title="fig:",height=188 ] -.5 cm -.8 cm the kek to kamioka ( k2k ) long - baseline neutrino oscillation experiment @xcite tests @xmath32 disappearance in the same @xmath38 region as probed by atmospheric neutrinos .
the neutrino beam is produced by a 12 gev proton beam from the kek proton synchrotron , and has 98% muon neutrinos with 1.3 gev mean energy .
the beam is controlled by a near detector 300 m away from the proton target .
information on neutrino oscillations is obtained by comparing this near detector data with the @xmath32 content of the beam observed by the super - k detector at a distance of 250 km .
the k2k - i data sample gave 56 events in super - k , whereas @xmath39 were expected for no oscillations .
the probability that the observed flux is explained by a statistical fluctuation without neutrino oscillations is less than 1% @xcite .
k2k - ii started in fall 2002 , and released data at the neutrino2004 conference @xcite corresponding to @xmath40 protons on target , comparable to the k2k - i sample .
altogether k2k - i and k2k - ii give 108 events in super - k , to be compared with @xmath41 expected for no oscillations .
out of the 108 events 56 are so - called single - ring muon events .
this data sample contains mainly muon events from the quasi - elastic scattering @xmath42 , and the reconstructed energy is closely related to the true neutrino energy .
the k2k collaboration finds that the observed spectrum is consistent with the one expected for no oscillation only at a probability of 0.11% , whereas the spectrum predicted by the best fit oscillation parameters has a probability of 52% @xcite .
the re - analysis of k2k data given in @xcite uses the energy spectrum of the 56 single - ring muon events from k2k - i + k2k - ii ( unfortunately not the full k2k data sample of 108 events , for lack of information outside the k2k collaboration ) .
it is reasonable to fit the data divided into 15 bins in reconstructed neutrino energy .
one finds that the @xmath38 indicated by the @xmath32 disappearance in k2k agrees with atmospheric neutrino results , providing the first confirmation of oscillations with @xmath37 from a man - made neutrino source .
however currently k2k gives a weak limit on the mixing angle due to low statistics .
the shaded regions in fig .
[ fig : atm+k2k ] are the allowed ( @xmath36 , @xmath37 ) regions that follow from the combined analysis of k2k and super - k atmospheric neutrino data .
one sees that , although the determination of @xmath36 is completely dominated by atmospheric data , the k2k data start already to constrain the allowed region of @xmath37 . note also that , despite the downward shift of @xmath37 implied by the new atmospheric fluxes , the new result is statistically compatible both with the previous one in @xcite and with the value obtained by the super - k @xmath31 analysis @xcite .
note that the k2k constraint on @xmath37 from below is important for future long - baseline experiments , as such experiments are drastically affected if @xmath37 lies in the lower part of the 3@xmath29 range indicated by the atmospheric data alone .
the effective leptonic mixing matrix in various gauge theories of massive neutrinos such as seesaw models was first systematically studied in @xcite .
its simplest unitary form can be taken as -.3 cm @xmath43 where each factor contains an angle and a phase , @xmath44 this form holds exactly in radiative models of neutrino mass @xcite and approximately in the high - scale seesaw and models where supersymmetry is the origin of neutrino mass @xcite .
deviations from unitarity may be phenomenologically important @xcite in the inverse seesaw @xcite .
here we stick to the form above .
all three phases in @xmath45 are physical @xcite , one corresponds to the kobayashi - maskawa phase of the quarks ( dirac - phase ) and affects neutrino oscillations , while the two majorana phases show up in neutrinoless double beta decay and other lepton - number violating processes @xcite .
two of the three angles determine solar and atmospheric oscillations , @xmath46 and @xmath47 .
since current neutrino oscillation experiments are not sensitive to cp violation , we will neglect all phases ( future neutrino factories aim at probing the effects of the dirac phase @xcite ) . in this approximation
three - neutrino oscillations depend on the three mixing parameters @xmath48 and on the two mass - squared differences @xmath49 and @xmath50 characterizing solar and atmospheric neutrinos .
the hierarchy @xmath51 implies that one can set , to a good approximation , @xmath52 in the analysis of atmospheric and k2k data , and @xmath37 to infinity in the analysis of solar and kamland data .
the relevant neutrino oscillation data in a global three - neutrino analysis are those of sections [ sec : solar-+-kamland ] and [ sec : atmospheric-+-k2k ] together with the constraints from reactor experiments @xcite . the global three
neutrino oscillation results are summarized in fig .
[ fig : global ] and in tab .
[ tab : summary ] . in the upper panels
@xmath53 is shown as a function of the parameters @xmath54 , minimized with respect to the undisplayed parameters .
the lower panels show two - dimensional projections of the allowed regions in the five - dimensional parameter space .
the best fit values and the allowed 3@xmath29 ranges of the oscillation parameters from the global data are summarized in tab .
[ tab : summary ] .
this table gives the current status of the three
flavour neutrino oscillation parameters .
three neutrino regions allowed by the world s oscillation data at 90% , 95% , 99% , and 3@xmath29 c.l . for 2 d.o.f.,title="fig:",height=219 ]
-.5 cm ? = ?
.[tab : summary ] three neutrino oscillation parameters from @xcite . [ cols="<,^,^",options="header " , ] -.5 cm in a three neutrino scheme cp violation disappears when two neutrinos become degenerate @xcite or when one angle vanishes , @xmath55 .
genuine three
flavour effects are associated to the mass hierarchy parameter @xmath0 and the mixing angle @xmath56 . -.4 cm determination of @xmath57 and bound on @xmath1 from current data.,title="fig:",height=158 ] -.4 cm -.6 cm the left panel in fig . [ fig : alpha ] gives the parameter @xmath58 as determined from the global @xmath12 analysis of @xcite .
the figure also gives @xmath59 as a function of the parameter combination @xmath60 which , to leading order , determines the long baseline @xmath61 oscillation probability @xcite .
the last unknown angle in the three neutrino leptonic mixing matrix is @xmath56 , for which only an upper bound exists .
the left panel in fig .
[ fig : alpha ] gives @xmath59 as a function of @xmath1 for different data sample choices .
one finds that the new data from kamland have a surprisingly strong impact on this bound . before the kamland-2004 data the bound on @xmath1 from global data
was dominated by the chooz reactor experiment , together with the determination of @xmath62 from atmospheric data . however , with the kamland-2004 data the bound becomes comparable to the reactor bound , and contributes significantly to the final global bound 0.022 ( 0.047 ) at 90% c.l .
( 3@xmath29 ) for 1 d.o.f .
this improved @xmath1 bound follows from the strong spectral distortion found in the 2004 sample @xcite .
.2 cm upper bound on @xmath1 ( 1 d.o.f . ) from solar+reactor data versus @xmath37 . dashed ( solid )
curves correspond to 90% ( 3@xmath29 ) c.l .
bounds , thick ones include kamland-2004 data , thin ones do not .
light ( dark ) regions are excluded by chooz at 90% ( 3@xmath29 ) c.l .
the current @xmath37 best fit value is indicated by the horizontal line , hatched regions are excluded by atmospheric + k2k data at 3@xmath29.,title="fig:",height=170 ] -.4 cm -.4 cm note that , since the reactor bound on @xmath1 quickly deteriorates as @xmath37 decreases ( see fig . [
fig : t13-solar - chooz ] ) , the improvement is especially important at low @xmath37 values , as implied by the new three dimensional atmospheric fluxes @xcite . in fig .
[ fig : t13-solar - chooz ] we show the upper bound on @xmath1 as a function of @xmath37 from chooz data alone compared to the bound from an analysis including solar and reactor neutrino data .
one finds that , although for larger @xmath37 values the bound on @xmath1 is dominated by chooz , for @xmath63 the solar + kamland data start being relevant .
the bounds implied by the 2002 and 2004 kamland data are compared in fig .
[ fig : t13-solar - chooz ] .
in addition to reactor and accelerator neutrino oscillation searches , future studies of the day / night effect in large water cerenkov solar neutrino experiments like uno or hyper - k @xcite may give valuable information on @xmath1 @xcite .
neutrino oscillation data are sensitive only to mass differences , not to the absolute neutrino masses . nor do they have any bearing on the fundamental issue of whether neutrinos are dirac or majorana particles @xcite .
the significance of the decay is given by the fact that , in a gauge theory , irrespective of the mechanism that induces , it must also produce a majorana neutrino mass @xcite , as illustrated in fig .
[ fig : bbox ] .
now that oscillations have been confirmed we know that must be induced by the exchange of light majorana neutrinos .
the corresponding amplitude is sensitive both to the absolute scale of neutrino mass as well as the two majorana cp phases in the minimal 3-neutrino mixing matrix @xcite .
[ fig : nbbfut ] shows the estimated average mass parameter characterizing the neutrino exchange contribution to versus the lightest neutrino mass .
the upper ( lower ) panel corresponds to the cases of normal ( inverted ) neutrino mass spectra .
the calculation takes into account the current neutrino oscillation parameters from @xcite and the nuclear matrix elements of @xcite .
in contrast to the normal hierarchy , where a destructive interference of neutrino amplitudes is possible , the inverted neutrino mass hierarchy implies a `` lower '' bound for the amplitude .
quasi - degenerate neutrinos @xcite such as predicted in @xcite , give the largest amplitude , as can be seen by the rising diagonal bands on the right - hand side of the panels .
future experiments @xcite will provide an independent confirmation of the present hint @xcite and push the sensitivity to inverse hierarchy models .
complementary information on the absolute scale of neutrino mass comes from beta decays searches @xcite as well as cosmology @xcite . in conclusion
, we can say that , despite the great progress achieved recently we are still very far from a `` road map '' to the ultimate theory of neutrino properties .
we have no idea of the underlying neutrino mass generation mechanism , its characteristic scale or its flavor structure .
we have still a long way to go and need more data , especially a confirmation of the lsnd and neutrinoless double beta decay hints .
s. weinberg , phys .
* d22 * 1694 ( 1980 ) .
p. minkowski , phys .
b * 67 * ( 1977 ) 421 ; t. yanagida , _ kek lectures , 1979 _ ; m. gell - mann , p. ramond , r. slansky , cern print-80 - 0576 ; r. mohapatra , g. senjanovic , phys .
d * 23 * ( 1981 ) 16 ; for the general 1 description of the seesaw see ref .
@xcite j. schechter and j. w. f. valle , phys . rev . * d22 * , 2227 ( 1980 ) ; , 774 ( 1982 ) .
m. fukugita and t. yanagida , phys .
b * 174 * ( 1986 ) 45 .
m. goodman , neutrino oscillation industry web - page , http://neutrinooscillation.org/. j. n. bahcall and m. h. pinsonneault , phys .
lett . * 93 * , 121301 ( 2004 ) m. honda , t. kajita , k. kasahara and s. midorikawa , astro - ph/0404457 . s. p. mikheev and a. y. smirnov , sov .
. phys . * 42 * , 913 ( 1985 ) .
l. wolfenstein , phys . rev . *
d17 * , 2369 ( 1978 ) . l. wolfenstein , phys .
b * 107 * 77 ( 1981 ) ; j. schechter and j. w. f. valle , phys . rev . * d24 * , 1883 ( 1981 ) , err .
d25 , 283 ( 1982 ) ; see also aip conf . proc . * 687 * , 16 ( 2003 ) s. m. bilenky , a. faessler and f. simkovic , phys . rev .
* d70 * 033003 ( 2004 ) k. babu , e. ma and j. w. f. valle , phys . lett . *
b552 * 207 ( 2003 ) ; m. hirsch et al , phys .
d * 69 * 093006 ( 2004 ) c. albright , hep - ph/0407155 ; w. grimus et al , hep - ph/0408123 ; s. barr , i. dorsner , nucl .
b * 585 * 79 ( 2000 ) ; h. s. goh et al , phys .
b * 587 * 105 ( 2004 ) ; m. raidal , hep - ph/0404046 ; h. minakata , a. smirnov , hep - ph/0405088 ; j. ferrandis , s. pakvasa , phys .
b * 603 * , 184 ( 2004 ) ; g. altarelli , f. feruglio , new j. phys .
* 6 * ( 2004 ) 106 m. maltoni et al , phys
* d68 * , 113010 ( 2003 ) ; phys . rev . * d67 * , 013011 ( 2003 ) g. fogli et al , phys . rev .
* d66 * , 053010 ( 2002 ) a. b. balantekin and h. yuksel , phys . rev . * d68 * , 113002 ( 2003 ) g.
fiorentini et al , phys . lett . *
b558 * , 15 ( 2003 ) ; h. nunokawa , w. c. teves and r. zukanovich , jhep * 11 * , 020 ( 2003 ) m. gonzalez - garcia et al , phys . rev . *
d63 * , 033005 ( 2001 ) g. barr , t. k. gaisser and t. stanev , phys .
* d39 * , 3532 ( 1989 ) . http://eps2003.physik.rwth-aachen.de .
a. zee , phys . lett . *
b93 * , 389 ( 1980 ) .
k. s. babu , phys
b203 * , 132 ( 1988 ) .
m. hirsch and j. w. f. valle , new j. phys .
* 6 * , 76 ( 2004 ) [ hep - ph/0405015 ] j. bernabeu et al , phys
* b187 * , 303 ( 1987 ) ; g. branco , m. rebelo , j. w. f. valle , phys . lett . *
b225 * , 385 ( 1989 ) ; n. rius , j. w. f. valle , phys . lett . *
b246 * , 249 ( 1990 ) ; f. deppisch , j. w. f. valle , hep - ph/0406040 .
r. n. mohapatra and j. w. f. valle , phys .
* d34 * , 1642 ( 1986 ) .
j. schechter and j. w. f. valle , phys . rev . * d23 * , 1666 ( 1981 ) .
m. doi et al , phys . lett . *
b102 * , 323 ( 1981 ) ; s. bilenky et al , phys .
b * 94 * 49 ( 1980 ) k. dick , m. freund , m. lindner and a. romanino , nucl .
b562 * , 29 ( 1999 ) m. apollonio et al , phys . lett . *
b466 * , 415 ( 1999 ) ; f. boehm et al , phys . rev . * d64 * , 112001 ( 2001 ) m. freund , phys . rev . *
053003 ( 2001 ) e. k. akhmedov et al , hep - ph/0402175 .
c. yanagisawa , proc . of ahep2003 , published at jhep , prhep - ahep2003/062 , accessible from http://ific.uv.es / ahep/. e. k. akhmedov et al , jhep * 0405 * 057 ( 2004 ) ; m. blennow et al , phys
. rev . * d69
* , 073006 ( 2004 ) j. schechter and j. w. f. valle , phys . rev . * d25 * , 2951 ( 1982 ) d. caldwell , r. mohapatra , phys
. rev . * d48 * , 3259 ( 1993 ) ; a. ioannisian and j. w. f. valle , phys .
b * 332 * 93 ( 1994 ) ; a. s. joshipura , z. phys .
c * 64 * ( 1994 ) 31 ; s. antusch and s. f. king , nucl .
b * 705 * ( 2005 ) 239 h. v. klapdor et al , hep - ph/9910205 .
h. klapdor etal , phys .
b * 586 * ( 2004 ) 198 a. osipowicz , et al , hep - ex/0109033 . s. pastor , hep - ph/0306233 ; s. hannestad , hep - ph/0404239 ; g. fogli et al , hep - ph/0408045 | i discuss the theoretical background and the status of neutrino oscillation parameters from the current worlds global data sample and latest flux calculations .
i give their allowed ranges , best fit values and discuss the small parameters @xmath0 and @xmath1 , which characterize cp violation in neutrino oscillations .
i mention the significance of ( neutrinoless double beta decay ) and current expectations in view of oscillation results . |
the particle nature of the dark matter remains a mystery .
this is contrasted with the firmly established dark matter abundance , which has been measured using the cosmic microwave background and large scale structure observations @xcite .
a wide range of dark matter particle candidates have been proposed , including axions , the lightest supersymmetric particle , and sterile neutrinos .
recently , sterile neutrinos have gained renewed interest , since theoretical considerations have shown them to be a natural part of a minimally extended standard model , the @xmath0msm .
this @xmath0msm , which is the standard model extended with only 3 sterile neutrinos , manages to explain the masses of the active neutrinos @xcite , the baryon asymmetry of the universe @xcite , and the abundance of dark matter @xcite . arguably , one of the most appealing aspects of sterile neutrinos as dark matter candidates is , that they will very likely either be detected or rejected during the next couple of years .
this is because the sterile neutrinos decay virtually at rest into a photon and an active neutrino producing a sharp decay line at @xmath1 where @xmath2 is the mass of the sterile neutrinos @xcite .
this photon line can now be searched for in cosmic high energy data of the background radiation ( for recent studies see @xcite ) , towards massive structures like galaxy clusters @xcite , or towards dark matter structures with a very low fraction of baryons @xcite .
the sterile neutrino was originally introduced in order to alleviate discrepancies between the theory of cold dark matter structure formation and observations @xcite , which indicated that the dark matter particle should be warm .
traditionally `` warm '' means that the mass of the dark matter particle should be in the kev range , in order to suppress the formation of small scale structures .
while some of these problems have found other solutions , sterile neutrinos have found other uses , including an explanation for the pulsar peculiar velocities @xcite , synthesising early star formation @xcite , and , as mentioned above , as an explanation for the masses of the active neutrinos and the baryon asymmetry @xcite . also constraints on doube @xmath3 decays
have been derived @xcite .
we will here follow the idea of @xcite and look for a decay line towards a region with very little baryons .
specifically we will study x - ray `` blank '' regions in the sky with no known x - ray sources . by doing
so we will search for a signal from the dark matter halo of our own galaxy .
we will see that the signal is very low for photons with energies of the order 0.110 kev , which allows us to exclude part of the sterile neutrino parameter space .
we have analysed a 0.39 kev spectrum extracted from many combined _ chandra _ acis - s3 blank sky observations ( shown in [ fig : gaussian ] ) with ciao 3.3 , and fitted spectral models with the spectral fitting package xspec @xcite .
the line features at 1.74 kev , 2.12.2 kev , and 7.48 kev are si @xmath4 , the au @xmath5 complex , and ni @xmath4 respectively , originating from fluorescence of material in the telescope and focal plane ( i.e. they are seen in spectra obtained when acis - s3 is stowed and not pointing at the sky ) .
however , a decay line from dark matter could `` hide '' under these prominent lines . in this case
the decay line should be redshifted by a factor of @xmath6 where @xmath7 is the redshift of the dark matter emitter .
dark matter concentrations located at different distances should thus give rise to lines at different observed energies . in order to test for this , we extracted spectra from acis - s3 data of two galaxy clusters , a383 and a478 , at redshifts , @xmath8 and @xmath9 , respectively .
we optimized the ratio between expected dark matter signal and emission from hot intracluster gas by extracting spectra from the cluster outskirts .
the x - ray emission from hot intracluster gas is proportional to the gas density squared , whereas the emission of photons from dark matter is directly proportional to the dark matter density .
hence , in the cluster outskirts the gas surface brightness falls off with projected cluster - centric distance , @xmath10 as , @xmath11 ( assuming a typical @xmath3-model ) whereas the dark matter surface brightness is expected to fall off as @xmath12 ( assuming an isothermal sphere ) . specifically , we have extracted spectra from an outer radial bin , defined by optimizing the ratio of matter in the field of view ( proportional to the signal from decaying dark matter ) to the x - ray emission from the intracluster gas ( `` noise '' ) described by a @xmath3-profile .
the outer radius was chosen as the radius with the optimal ratio ( which is very close to the edge of the acis - s3 chip ) and the inner radius was chosen so the signal to noise ratio had a value of half the optimal value .
no obvious line features was detected at the corresponding redshifted energies .
we then proceeded to firmly constrain the flux from any dark matter decay line in the full energy interval spanned by the blank sky spectrum .
the blank sky spectrum was well fitted ( reduced @xmath13 for 540 d.o.f . ) by a composite model consisting of an exponential plus a power law plus four gaussians for the most prominent lines .
a hypothetical mono - energetic emission line in the spectrum was represented by a gaussian , centered at the line energy , and with a width , @xmath14 , given by the instrument spectral resolution : @xmath15 kev for @xmath16 kev and @xmath17 kev for @xmath18 kev .
( the velocity of the halo dark matter is producing negligible line broadening , @xmath19 ) .
for each energy in steps of 0.05 kev , we defined a gaussian with the instrumental width and maximum at the model value of the fit to the broad - band spectrum .
the flux of this gaussian was then calculated providing an upper limit of emission from decaying dark matter at the given energy .
this is a model independent and very conservative analysis , since any x - ray emitters within the field of view , and background , contributes to the extracted spectrum .
if we instead only allowed the decay signal from potential decaying dark matter to produce a bump above the `` base line '' spectrum , then we could improve the bounds by a large factor @xcite .
however , this approach would depend sensitively on the robustness of modelling background and emissions from sources contributing to the `` base line '' spectrum .
the amount of dark matter from our galaxy within the observed field of view is only a minute fraction of the total mass of the halo .
we use the model for the milky way presented in @xcite , namely an nfw dark matter profile with @xmath20 , virial radius @xmath21 kpc , concentration @xmath22 , and solar distance @xmath23 kpc .
we then integrate out through the milky way halo to find the expected signal .
changing the details of this model does not significantly affrect our results ( e.g. changing the inner density slope between @xmath24 and zero , or changing the outer slope between @xmath25 and @xmath26 , changes the predicted signal by less than a factor of 2 ) .
there is also an uncertainty , possibly as large as a factor of 2 , arising from using the value of @xmath20 .
the blank sky spectrum was extracted from the full angular opening of the _ chandra _ acis - s3 chip which is @xmath27 corresponding to a mass within the field of view of @xmath28 at a mean luminosity distance of @xmath29 kpc , derived from the model mentioned above . the luminosity from decaying sterile neutrinos is @xmath30 , where @xmath31 is the energy of the photons , @xmath32 is the number of sterile neutrinos particles in the field of view ( @xmath33 ) and @xmath34 is the decay rate .
the flux at a luminosity distance , @xmath35 , is : @xmath36 the observed flux , @xmath37 , at a given energy ( @xmath38 ) yields a conservative upper limit on the flux from decaying sterile neutrinos , so eq . [ eqn : flux ] can be rewritten as : @xmath39 from the blank sky flux , an upper limit on the decay rate has been derived as function of line energy and plotted in [ fig : decayrate ] .
the neutrinos can decay via various channels , and the lifetime of the sterile neutrinos is given by @xcite : @xmath40 @xmath41 is the mixing angle with the active neutrino , and @xmath42 takes into account the open decay channels so that for @xmath43 , where only the neutrino channel is open , @xmath44 .
the mixing angle is given by ( * ? ? ?
209 , 210 ) : @xmath45 where @xmath46 is the number of relativistic degrees of freedom at the temperature at which the sterile neutrinos are produced . for neutrino masses of @xmath47 kev ,
the neutrinos are produced near the qcd phase transition @xcite , making the distribution somewhat non - thermal @xcite .
the value of @xmath48 for such neutrinos is very conservatively considered to be between 10.75 and 20 , depending on the details of the qcd phase transition .
we will here use @xmath49 as a reference value .
a numerical calculation including the details of the qcd phase - transition @xcite has confirmed that the choice @xmath50 gives good agreement with the analytical results presented by @xcite .
s is a free parameter taking into account that additional entropy may be produced after the sterile neutrinos have been created @xcite .
this parameter was suggested to be in the range between 1 and 100 @xcite .
@xmath51 is a constant depending on which active neutrino the sterile neutrinos are assumed to mix with .
it takes the values @xmath52 for @xmath53 mixing with @xmath54 and @xmath55 for @xmath56 .
the sterile neutrinos are assumed to account for all dark matter , i.e. @xmath57 , and the hubble parameter at the present time , @xmath58 , given as @xmath59 .
@xcite derived the branching ratio for the radiative decay , @xmath60 , to be @xmath61 . now
the lifetime can be rewritten as : @xmath62 an upper limit on the decay rate is constrained from eq . [
eqn : maxdecay ] and @xmath63 the lower limit on the lifetime from the data and the model predictions are plotted in [ fig : lifetime ] for various values of @xmath64 and @xmath48 .
we see from [ fig : lifetime ] that the blank sky data ( solid red line ) is at least one order of magnitude less restrictive than the simplest sterile model predictions ( hatched region ) .
models with significant entropy production , @xmath65 , are excluded in the mass range @xmath66 .
for @xmath67 we exclude @xmath68 .
these constraints are more restrictive than the ones obtained for the coma cluster periphery @xcite for masses less than approximately @xmath69 kev .
msm prediction for @xmath70 and @xmath71 ( hatched ) and several variations of @xmath64 and @xmath48 ( black ) have been overplotted.,scaledwidth=49.0% ]
the simplest models with @xmath70 are bounded from below , @xmath72 , using ly@xmath73 observations @xcite , and possibly even @xmath74 kev according to a recent analysis @xcite .
these constraints are weakned when allowing for @xmath75 @xcite , since additional entropy dilutes the momentum of the sterile neutrinos .
upper limits on the sterile neutrino mass are derived from the flux of the diffuse photon background @xcite , and are of the order @xmath76 kev for @xmath70 @xcite .
these bounds are strengthened for @xmath75 , since the sterile neutrinos then will decay faster . a strict lower bound on the sterile neutrino mass arises since big bang nucleosynthesis only allows approximately @xmath77 additional relativistic neutrino species to be populated at @xmath78 mev temperatures .
this lower bound is approximately @xmath79 ev . the tremaine - gunn bound applied to dwarf speroidals give @xmath80 kev @xcite .
the strong claim of an upper mass limit presented in @xcite , under - estimated the flux from virgo cluster by 2 orders of magnitude , and is unreliable @xcite . comparing these existing bounds with our findings ,
we conclude that a broad band in neutrino parameter space is still open as shown in [ fig : m_s ] . taking the results of @xcite into account ,
then only masses above approximately @xmath81 kev are allowed . ignoring the results of @xcite ,
then another band opens in parameter space , for small masses and large entropy production .
parameter space with constraints from the _ chandra _ blank sky data ( red , fat , diagonal , this paper ) , ly-@xmath73 observations ( blue , diagonal , conservative @xcite , and green , vertical , ambitious @xcite ) , tremaine - gunn bound @xmath82 kev ( black , horizontal ) .
, scaledwidth=49.0% ]
we have used _ chandra _
blank sky observations , i.e. viewing out through the milky way halo , to search for a possible decay line from dark matter particles . no obvious decay line was identified .
the blank sky 0.39 kev x - ray flux is used to constrain the parameter space of sterile neutrinos , predicted to decay at rest to a detectable photon and an active neutrino .
we find that the entropy production must be limited , @xmath83 for the mass range , @xmath84 kev , and even @xmath85 for masses near @xmath86 kev .
this leaves a wide allowed space in parameter space , and the sterile neutrino is therefore still a viable dark matter candidate .
if all ly@xmath73 constraints are considered , then it appears that the sterile neutrino is either rather massive , or its momenta are diluted by the entropy production , rendering it a cold dark matter candidate .
if some of the ly@xmath73 constraints are relaxed , then there is a broad allowed band for smaller masses where the sterile neutrino is still a warm dark matter candidate .
our analysis is conservative and assumes that the entire blank sky signal at a given energy is arising from a decaying neutrino . by looking for decay lines above the expected blank sky emission ,
the results obtained here may be improved significantly .
shh is pleased to thank ben moore for inspiring discussions .
we thank alex kusenko for useful comments .
the dark cosmology centre is funded by the danish national research foundation .
shh is supported by the swiss national foundation .
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 .
+ _ note added : _ + after finishing this paper we became aware of an independent analysis , where the possibility of measuring decaying dark matter particles from our galaxy , using the blank sky observations with _ xmm - newton _ , was considered @xcite .
the conclusions reached in that paper are qualitatively similar to ours .
+ facilities : boyarsky , a. , neronov , a. , ruchayskiy , o. & shaposhnikov , m. 2005 arxiv : astro - ph/0512509 .
boyarsky , a. , neronov , a. , ruchayskiy , o. & shaposhnikov , m. 2006 , arxiv : astro - ph/0603368 .
boyarsky a. , neronov a. , ruchayskiy o. , shaposhnikov m. & tkachev i. 2006 , astro - ph/0603660 dodelson s. & widrow l. m. 1994 , phys .
72 , 17 [ arxiv : hep - ph/9303287 ] .
dolgov , a. d. , hansen , s. h. , raffelt , g. & semikoz , d. v. 2000 , nucl .
b590 , 562 dolgov , a. d. & hansen , s. h. 2002 astropart .
phys . , 16 , 339 dolgov , a.d .
2002 , phys .
370 , 333 fuller g. m. , kusenko a. , mocioiu i. & pascoli s. 2003 , phys .
d 68 , 103002 | we consider the possibility of constraining decaying dark matter by looking out through the milky way halo .
specifically we use _ chandra _ blank sky observations to constrain the parameter space of sterile neutrinos .
we find that a broad band in parameter space is still open , leaving the sterile neutrino as an excellent dark matter candidate . |
[ cols="<,<,<,<",options="header " , ] at the crystal surface of bi@xmath0se@xmath1 , the lattice structure is distorted along _
c_-axis due to the abrupt reduction of the interlayer van der waals force that binds the crystal together .
the surface distortion was calculated by density functional theory ( dft ) to be about 10% along @xmath7-axis , and depicted in ref . .
additionally , the observation of two - dimensional electron gas formed on bi@xmath0se@xmath1 surface is consistent with the subsurface van der waals gap expansion @xcite .
as the lattice is distorted , the frequencies of atomic vibration modes at the surface are usually modified to a smaller value than in the bulk at the brillouin zone center ( @xmath8 point ) @xcite .
if there is a gap in the phonon density - of - state ( dos ) and with large enough distortion , the surface phonon dos can be entirely separated from the bulk @xcite .
such modes are long lived and localized at the surface , where the dispersion can be quite different than the bulk . in bi@xmath0se@xmath1
, there are some residual phonon dos in the entire energy range @xcite , and the surface modes can decay into bulk phonon modes . if the frequency shift is small , where the dos of the surface mode is not entirely separated from the bulk , then strictly speaking a localized surface mode does not exist .
instead , one would only expect a `` surface resonance '' , which is only distinguishable from the bulk phonon by the slightly lower energy .
due to inversion symmetry breaking at the crystal interface , the surface resonance from the raman active @xmath3 and ir active @xmath4 phonons are both expected to appear in the @xmath9 symmetry ( @xmath6 group ) , corresponding to out - of - plane atomic motion .
this is consistent with the 4 @xmath9 modes we observed ( fig .
[ fig:3](a ) ) . from the energies of these @xmath9 modes
, we conclude that @xmath10 and @xmath11 are associated with the bulk phonon modes @xmath12 and @xmath13 , respectively .
the measured energy of the @xmath10 mode is somewhat different than the previously reported value of 68@xmath2 by time resolved arpes in ref . .
this may be partly due to surface quality variation .
arpes measured sample is usually cleaved in ultra high vacuum , whereas the surface in this study is cleaved in nitrogen environment .
this may also explain why this mode was never observed in the thin film samples ( fig .
[ fig:1 ] ) , where the sample is unavoidably exposed to air for a few minutes during the transfer between mbe chamber and raman cryostat .
the @xmath11 mode appears as a shoulder to the @xmath13 bulk phonon , requiring higher resolution to distinguish from the bulk mode , and therefore was overlooked in the previous raman study @xcite . in comparison ,
the surface modes @xmath14 and @xmath15 have higher intensity and are better resolved .
one possibility for this difference is that the bulk counterpart of these modes are the ir active @xmath16 and @xmath17 phonons , as the measured energy is close to the calculated values ( table [ table : mode ] ) .
since these bulk modes are raman inactive , we were able to better resolve the surface resonance .
another possibility is that the phonon dos is practically zero at these energies in the @xmath9 symmetry channel , and the surface vibration modes are truly localized .
distinguishing these two scenarios is in fact experimentally non - trivial , especially since the experimental values of the @xmath16 and @xmath17 bulk phonon energies are yet unknown .
nevertheless , both possibilities point to the surface origin of these two modes , which may provide us with information on the electron - phonon coupling at the ti surface .
in particular , the double peak line shape of the @xmath15 phonon mode was overlooked in previous raman studies and may be related to the 20mev `` kink '' in the topological surface state s energy dispersion curve reported by arpes measurements @xcite . on the other hand , while the bulk phonons show little resonance effect , the @xmath15 phonon displays antisymmetric line shape with near - infrared excitation ( fig .
[ fig:2 ] , inset ) .
this suggests that there are surface electronic dos that appear under resonance excitation with near - infrared wavelength @xcite , which interacts with the @xmath15 phonon and results in the observed fano line shape @xcite . since the in - plane symmetries
are mainly preserved as the dft calculated atomic surface distortion is purely out - of - plane @xcite , one would not expect any surface phonon with @xmath18 symmetry ( @xmath6 group ) for bi@xmath0se@xmath1 .
however , the in - plane bonding potential is also modified by having distortion along _ c_-axis , and therefore the phonon frequency at surface is still slightly different than the bulk value .
if the modification is tiny , the @xmath18 modes are expected to be weak and close to the bulk phonons . in fig .
[ fig:1](b ) and fig .
[ fig:3](b ) , we can see hints of two additional modes , labeled by @xmath19 and @xmath20 .
the energies of these modes are in fact close to the measured values of @xmath21 and @xmath22 bulk phonons @xcite , and are consistent with the previous raman study @xcite ( table [ table : mode ] ) .
however , the frequency of @xmath23 is slightly higher than @xmath21 , which is against the expectation from a surface resonance .
this may reflect the fact that this is an in - plane mode , orthogonal to the lattice distortion direction . or
, this may be indicative of non - trivial electron - phonon interaction with the surface states , and worth further studying .
in conclusion , we have done a systematic symmetry analysis on the raman spectra from high quality , freshly cleaved or grown _ ab _ surfaces of bi@xmath0se@xmath1 single crystal and thin films .
we observed 4 out - of - plane and possibly 2 in - plane surface vibrational modes .
the much larger intensity for the out - of - plane vibration modes is consistent with @xmath7-axis lattice distortion and van der waals gap expansion calculated for bi@xmath0se@xmath1 crystal surface @xcite .
the energies of the @xmath10 and @xmath11 modes are close to the bulk @xmath3 phonons . due to the smallness of calculated crystal surface distortion @xcite and the lack of phonon dos gap @xcite , these modes are likely not fully localized or separated from bulk phonon dos
. however , the @xmath14 and @xmath15 modes are much stronger and sharper compared to the other vibration modes , and may be candidates of localized surface phonons .
in particular , we found an anomalous double peak and fano lineshape for the @xmath15 mode , both in low doped bi@xmath0se@xmath1 single crystals .
the fano lineshape is usually indicative of electron - phonon coupling , important for understanding the relaxation and scattering of surface state excitations . here
, we found that the fano lineshape is dependent of the excitation energy , which may be able to explain the inconsistent surface electron - phonon coupling constant found in previous arpes studies @xcite .
g.b . and h .-
acknowledge support from nsf award dmr-1104884 .
h.k . also acknowledges partial support from the u.s .
doe , bes grant de - sc0005463 .
, n.k . and m.b .
are funded by gordon and betty moore foundation s epiqs initiative ( gbmf4418 ) and nsf(efma-1542798 ) .
s .- w.c . and x.w .
acknowledge support from nsf award dmref-1233349 . 67 ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1 '' '' @noop [ 0]secondoftwo sanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12 @startlink[1 ] @endlink[0 ] @bib@innerbibempty link:\doibase 10.1103/physrevlett.98.106803 [ * * , ( ) ] link:\doibase 10.1038/nphys1270 [ * * , ( ) ] link:\doibase 10.1038/nature08234 [ * * , ( ) ] link:\doibase 10.1038/nphys1274 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.103.246601 [ * * , ( ) ] http://dx.doi.org/10.1038/ncomms1131 [ * * , ( ) ] link:\doibase 10.1038/nphys2108 [ * * , ( ) ] link:\doibase 10.1146/annurev - conmatphys-062910 - 140432 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.100.096407 [ * * , ( ) ] link:\doibase 10.1103/physrevb.78.195424 [ * * , ( ) ] link:\doibase 10.1126/science.1167747 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.82.3045 [ * * , ( ) ] link:\doibase 10.1126/science.1187485 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.104.116401 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.83.1057 [ * * , ( ) ] link:\doibase 10.1126/science.1239834 [ * * , ( ) ] link:\doibase 10.1126/science.1248253 [ * * , ( ) ] link:\doibase 10.1103/physrevb.81.241301 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.108.187001 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.108.117601 [ * * , ( ) ] link:\doibase 10.1103/physrevb.88.075432 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1007/bf02746071 [ * * , ( ) ] link:\doibase 10.1103/physrev.105.540 [ * * , ( ) ] link:\doibase 10.1103/physrev.116.302 [ * * , ( ) ] inlink:\doibase 10.1007/978 - 3 - 642 - 75785 - 3_3 [ _ _ ] , ( , , ) pp
. http://dx.doi.org/10.1016/0039-6028(94)90684-x
[ * * , ( ) ] link:\doibase 10.1103/physrevlett.107.186102 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.108.185501 [ * * , ( ) ] link:\doibase 10.1103/physrevb.88.035402 [ * * , ( ) ] link:\doibase 10.1103/physrevb.83.241303 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.108.187001 [ * * , ( ) ] http://dx.doi.org/10.1038/srep02411 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.110.217601 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.115.257402 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.195118 [ * * , ( ) ] link:\doibase 10.1088/0957 - 4484/25/24/245701 [ * * , ( ) ] link:\doibase 10.1063/1.3573868 [ * * , ( ) ] in @noop _ _ , ( , ) pp . link:\doibase 10.1103/physrevlett.109.186403 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.109.116804 [ * * , ( ) ] link:\doibase 10.1016/s0254 - 0584(99)00145 - 5 [ * * , ( ) ] link:\doibase 10.1038/nnano.2011.214 [ * * , ( ) ] link:\doibase 10.1002/pssb.2220610218 [ * * , ( ) ] link:\doibase 10.1002/pssb.2220840226 [ * * , ( ) ] link:\doibase 10.1021/nl200773n [ * * , ( ) ] link:\doibase 10.1103/physrevb.81.125120 [ * * , ( ) ] link:\doibase 10.1002/pssb.2221330202 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/j.ssc.2013.03.010 [ * * , ( ) ] link:\doibase 10.1038/nphys2513 [ * * , ( ) ] link:\doibase 10.1016/s0025 - 5408(01)00746 - 2 [ * * , ( ) ] @noop _ _ , massachusetts institute of technology press research monograph ( , ) http://dx.doi.org/10.1063/1.3689759 [ * * , ( ) ] http://stacks.iop.org/0953-8984/27/i=5/a=052203 [ * * , ( ) ] http://dx.doi.org/10.1038/nphys2647 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/j.tsf.2014.01.082 [ * * , ( ) ] in @noop _ _ , ( , ) pp.@noop * * , ( ) link:\doibase 10.1142/s0217979291000237 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.73.500 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.113.157401 [ * * , ( ) ] link:\doibase 10.1103/physrevb.83.094301 [ * * , ( ) ] http://stacks.iop.org/0268-1242/27/i=12/a=124001 [ * * , ( ) ] http://dx.doi.org/10.1134/s0021364011140104 [ * * ( ) ] link:\doibase 10.1103/physrevlett.111.136802 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.116.036601 [ * * , ( ) ] in @noop _ _ , ( , ) pp . | we present polarization resolved raman scattering study of surface vibration modes in the topological insulator bi@xmath0se@xmath1 single crystal and thin films .
besides the 4 raman active bulk phonons , we observed 4 additional modes with much weaker intensity and slightly lower energy than the bulk counterparts . by symmetry analysis and comparison with theoretical calculation
, we assigned these additional modes to out - of - plane surface phonons , where the frequency is slightly modified from the bulk phonon due to out - of - plane lattice distortion near the crystal surface .
in particular , two of the surface modes at 60 and 173@xmath2 are associated with raman active @xmath3 bulk phonon modes , the other two at 136 and 158@xmath2 are associated with infrared active bulk phonons with @xmath4 symmetry . the latter become raman allowed due to restriction of crystalline symmetry from @xmath5 in the bulk to @xmath6 at the surface of bi@xmath0se@xmath1 . the 158@xmath2 surface phonon mode show a fano line shape , suggesting interaction with an electronic continuum at the crystal surface .
in addition , we observed two weak features at 67 and 126@xmath2 likely corresponding to in - plane surface vibrational modes . |
a great deal of interest has been recently devoted to the development and understanding of formal relationships known as fluctuations theorems .
these expressions generally constrain the probability distribution function of entropy production without relying on how close the system is to thermal equilibrium .
a number of complementary proposals have been put forward @xcite , and a considerable effort has been devoted to clarify the similarities among them .
experiments have been carried out to validate and use such expressions to extract useful information in different physical and biological systems @xcite .
a new fluctuation theorem for dissipative systems has been derived by seifert @xcite , which is based on the generalization of the concept of entropy production to the level of a single trajectory . the overall entropy production , namely the sum of the contributions from the environment and the one from the particle along its trajectory , is shown to obey an integral fluctuation relation ( and in some cases a detailed one ) for any initial conditions , arbitrary driving and over a generic finite time interval .
moreover , recent experiments have validated this theorem @xcite and the stochastic entropy production has been measured experimentally for the first time @xcite .
in this work , we consider a brownian particle moving on a periodic potential under the action of an external constant force .
such system allows us to carry out a detailed study of the total entropy production and assess carefully the implications of seifert s fluctuation theorem . after a summary of the main results of ref .
@xcite in section [ sect : theory ] , we particularize them for the system under study in section [ sect : model ] . in section [ sect : numerics ] , we discuss numerical results and compare them with analytic predictions where available .
we conclude with a discussion section .
in ref . @xcite the dynamics of one degree of freedom @xmath0 such as the position of a particle in a one dimensional system is studied .
it evolves according to a langevin equation , @xmath1 where @xmath2 is a gaussian white noise with autocorrelation @xmath3 , and @xmath4 is the diffusion constant so that in equilibrium @xmath5 .
we consider @xmath6 throughout this work .
the force @xmath7 splits into a term arising from a conservative potential , @xmath8 , plus a time dependent external source , @xmath9 , whose control parameter is @xmath10 .
the non - equilibrium gibbs entropy , @xmath11 can be interpreted as a mean over single trajectories as specified by the probability @xmath12 of finding the particle at position @xmath0 at time @xmath13 .
seifert uses the above expression to propose an entropy associated to each trajectory of the particle , @xmath14 the rate of change of entropy of the system is derived from the fokker - planck equation for the probability @xmath12 . the rate of heat dissipation in the medium and the increase of entropy in the medium can be related as @xmath15 .
therefore , it is possible to write down the rate of change of total entropy associated with a single trajectory or realization as @xmath16 where we have introduced the particle flux , which for the langevin dynamics we are considering is given by @xmath17 in the spirit of previous fluctuation theorems , which relate the weights of trajectories with their time reversed counterparts , seifert introduces the quantity @xmath18 , @xmath19 \equiv \ln \frac { p[x(t)|x_{0 } ] p_{0}(x_{0 } ) } { \tilde{p}[\tilde{x}(t)|\tilde{x}_{0 } ] p_{1}(\tilde{x}_{0 } ) } , \ ] ] which relates the weight of a trajectory with its time reversed analogue .
the ( arbitrary ) distributions of initial and final values are @xmath20 and @xmath21 respectively .
the tilde refers to the time reversal operation , which associates with a trajectory @xmath22 , a reversed one @xmath23 , where @xmath13 is the duration of the process . in this way
, @xmath24 retraces the reversed path of the stochastic trajectory @xmath22 , and it starts at @xmath25 , which corresponds to the final value of the stochastic trajectory @xmath22 .
the system is driven out of equilibrium by an external control parameter that changes with time according to a prescribed protocol @xmath10 , and @xmath26 $ ] and @xmath27 $ ] are the probabilities of the forward and backward paths respectively . while @xmath18 is in principle
defined for any normalized distribution @xmath28 , it only becomes the total entropy production ( and thus acquires a physical meaning ) when @xmath29 is the solution of the fokker - planck equation for the initial distribution @xmath30 .
then , the definition ( [ gibbs ] ) yields @xmath31 , the particle entropy change . the other terms appearing in eq .
( [ rr ] ) , related to the path probabilities , correspond to the medium entropy production @xmath32 / \tilde{p}[\tilde{x}(t)|\tilde{x}_{0}]$ ] , as shown in ref .
finally , we can identify @xmath18 with the total entropy production of the system , @xmath33 as we will show in short , the mathematical identity @xmath34 , together with eq .
( [ r+ ] ) , implies the following integral fluctuation theorem for the total entropy variation along a stochastic trajectory @xmath35 , @xmath36 it is absolutely general since it holds at any time along a trajectory generated from generic initial conditions and subject to arbitrary forces . in the subsequent sections we will analyze the implications and consequences of this relation for a simple example where a detailed analysis of @xmath18 will be performed .
let us consider a one dimensional stochastic system in which a particle is moving in a periodic potential @xmath8 and under an external force @xmath37 ( so that we restrict the general driving to the simplest case of constant driving in time and space , @xmath38 ) .
the corresponding langevin equation in the overdamped limit is @xmath39 where the friction coefficient has been absorbed in the time units for simplicity .
then the noise is gaussian , white , with zero mean and autocorrelation @xmath40 .
this system is a simplified version of models analyzed previously in refs .
@xcite to assess some aspects of the fluctuation theorem .
our aim is to derive an explicit expression for @xmath18 in the steady state . to this end , we first split @xmath18 into two contributions , @xmath41 } { \tilde{p}[\tilde{x}(t)|\tilde{x}_{0 } ] } + \ln \frac{p_{0}(x_{0 } ) } { p_{1}(\tilde{x}_{0})}.\ ] ] the first term on the rhs involves the weight of the forward and reversed paths . from the path integral
approach to stochastic systems , it is known that the probability to observe a given realization of the noise is @xmath42 \sim exp \left ( -\int_{0}^{t}dt ' \eta^{2}(t')/4 t \right).\ ] ] such a gaussian dependence can be used , together with the evolution equation ( eq . ( [ lang ] ) ) , to obtain an explicit expression for the transition probability @xmath26 $ ] in terms of the particle s position . if one takes into account the jacobian of the transformation , one arrives at @xcite @xmath43 } { \tilde{p}[\tilde{x}(t)|\tilde{x}_{0}]}=\int_{0}^{t } ( -v'(x)+f)\dot{x}d\tau /t,\ ] ] which in the present case reduces to a much simpler form , @xmath44 } { \tilde{p}[\tilde{x}(t)|\tilde{x}_{0 } ] } = ( f\delta x - \delta v)/t,\ ] ] where @xmath45 and @xmath46 .
it is worth mentioning that the final position @xmath47 corresponds to a generic observation time @xmath13 .
the second term on the rhs of eq . ( [ rsplit ] ) accounts for the ratio of the probabilities of initial and final positions .
we are interested in the steady state , and thus @xmath48 follows easily from the corresponding solution of the fokker - planck equation . for a system under an effective potential @xmath49 ,
being @xmath8 an arbitrary periodic function of period @xmath50 , one gets @xcite @xmath51 where @xmath52 ensures proper normalization .
since the system is in the steady state , @xmath53 , which implies @xmath54 and hence the second term of eq .
( [ rsplit ] ) can be obtained from eq .
( [ px ] ) . when inserting eqs .
( [ path ] ) , and ( [ px ] ) plus ( [ relationps ] ) into ( [ rsplit ] ) , a simple expression for @xmath18 is obtained , @xmath55 it shows that @xmath18 can be computed exactly by integrating the effective potential felt by the particle for every trajectory .
it is worth mentioning that this expression is only a function of the initial and final positions , which is a non trivial result
. this fact is of experimental relevance , since one does not need to keep detailed information of the complete trajectory to compute @xmath18 .
note also that the exchange of @xmath56 for @xmath57 reverses the sign of @xmath18 .
we will exploit this important symmetry subsequently .
further progress is feasible if a specific potential is chosen , so that the integrals determining @xmath18 in the previous equation can be worked out analytically . in particular , by picking the potential depicted in fig .
( [ potential].a ) , @xmath58 eq.([rm ] ) reduces to @xmath59 where @xmath60 and @xmath61 . for natural values of the parameter @xmath62
, the integrals can be done analytically , yielding to an explicit expression for @xmath18 as a function of the initial and final values of the position of the particle for a trajectory taking place during an interval @xmath13 . therefore , and without loss of generality , we consider @xmath63 .
we then arrive at the simple expression @xmath64 where @xmath65 the first term on the rhs of eq .
( [ rmain ] ) is directly related to the external work performed on the system by the external force @xmath37 and then dissipated .
the second term has information about the potential @xmath8 and of the external force too . in fig .
( [ potential].b ) we show the increase of @xmath18 as a function of the final position along a trajectory , @xmath57 , when @xmath66 for different values of the dimensionless driving force @xmath67 . when @xmath67 is increased , the relative contribution of the periodic potential to @xmath18 decreases , as can be seen directly from eq .
( [ rmain ] ) .
the interesting regime corresponds to the situations in which the external driving , the periodic potential and the thermal energy have comparable magnitudes .
we will concentrate on such regime from now on .
we perform numerical simulations of eq .
( [ lang ] ) with the potential defined in eq .
( [ pot ] ) to characterize the properties of @xmath18 .
such study is required because , even though @xmath18 depends only on the initial and final positions of the trajectory , the final position is a function of the noise and hence the probability distribution function of @xmath18 , @xmath68 , has to be obtained .
then its properties can be analyzed and the fluctuation theorem tested . from the definition of @xmath18
one can show that @xmath69p_{0}(x_{0 } ) e^{-r } = \nonumber \\
= \sum_{\tilde{x}(t),\tilde{x}_{0 } } \tilde{p}[\tilde{x}(t)|\tilde{x}_{0}]p_{1}(\tilde{x}_{0 } ) = 1 .
\label{eq : general_er}\end{aligned}\ ] ] this is a general result , independent of the specific form of the potential and external driving exerted on the particle . using the collection of values of @xmath70 and @xmath47 from the simulations and inserting them in expression ( [ rmain ] ) , we can perform the average @xmath71 , which can be written down as the sum @xmath72 where @xmath73 is the number of particle trajectories that are used to perform the average .
it can also be interpreted as the number of repetitions of the experiment with the same protocol .
we display in fig .
( [ e - r ] ) the average of @xmath74 as a function of the number of trajectories , when the time interval is @xmath75 and the external driving is @xmath76 . around this zone of the space of parameters
is where the relation is not trivial at the same time that it is numerically testable .
one can see that the average value predicted by eq .
( [ eq : general_er ] ) is rapidly attained .
the same must hold for any values of @xmath77 , @xmath67 and @xmath13 , as it has been verified for some of them ( data not shown ) .
although a noisy but progressive approach to unity is found as @xmath73 increases , jumps are observed , as depicted in the zoom of fig .
( [ e - r ] ) .
they are rare events that contribute significantly to the average because @xmath78 , and hence can be regarded as transient violations of the second law .
we will come back to this point in fig .
( [ transient ] ) when the evolution of @xmath18 in time for one single trajectory will be discussed . in order to gain more insight into the properties of @xmath18 and the implications behind the fluctuation theorem
, we analyze the probability distribution of the total entropy production , @xmath68 , as obtained from numerical simulations over many independent trajectories .
such a quantity can be measured experimentally , and a specific realization of a driven two level system with time - dependent rates has been already achieved @xcite . fig .
( [ p(r ) ] ) exhibits the probability distribution function @xmath68 at different times , for an external driving @xmath76 .
the distribution function is peaked and would reduce to a delta function at @xmath79 at all times for an equilibrium system . due to the increase of the mean of @xmath18 as a function of time , @xmath68 is shifted towards positive values of @xmath18 and spreads as time increases .
the broadening of the distribution can be attributed to the interaction of the particle with the underlying periodic potential together with thermal fluctuations .
it is interesting to note that @xmath68 develops a richer structure as time increases .
in particular the number of peaks increases because there is a higher chance to explore the periodic potential .
increasing @xmath67 would lead to a corresponding decrease in this structure in @xmath68 . for short time intervals
, there is a contribution to the distribution that comes from negative values of @xmath18 , in agreement with the jumps described in fig .
( [ e - r ] ) . as time increases the contributions to the negative tail decrease , being harder to observe .
precisely , the relation between negative and positive tails is very relevant , as we will discuss in short .
let s recall that @xmath18 can be expressed as a sum of two contributions , as written down in eq .
( [ rsplit ] ) .
the first term involves a ratio of conditional probabilities and is inextricably related to the heat @xmath80 interchanged with the heat bath @xcite , because @xmath81 reduces to eq .
( [ path ] ) when eq .
( [ lang ] ) is used . in turn
, @xmath80 can be decomposed in two significant contributions ; a first term proportional to the work produced by the external force , @xmath82 , and a second one , @xmath83 , which corresponds to the change in the internal potential energy of the system . the remaining contribution to @xmath18 , the second term on the rhs of eq .
( [ rsplit ] ) , is the entropy change of the particle @xmath84 along the stochastic trajectory . on physical grounds ,
it is fruitful to write @xmath85 so that three different physical sources contribute to the change in the overall entropy of the system .
the total entropy change @xmath18 has contributions from the medium and from the system itself , namely , the total heat dissipated @xmath80 ( which comes both from the external forcing and the change of internal energy ) and the entropy change associated to the change in the system s internal state , @xmath84 . as presented in fig .
( [ split ] ) , for the simple model under scrutiny it is possible to study each probability distribution function of the different contributions separately . obviously , although @xmath86 , @xmath87 .
one can see that the structure of the probability distribution function @xmath88 , whose contribution is directly the external force , has no internal structure and is simply characterized by a broad , slightly tilted , peak which moves to higher values of @xmath89 as time increases .
@xmath90 retains the structure of the peaks , which we attribute to the effect of the periodic potential .
@xcite for a study of the structure of the probability distribution function of heat exchanged in a simple temperature transducer . in fig .
( [ split].b ) , @xmath91 is shown .
it has a bound support and its shape does not change as time increases .
since the term associated to the external force takes increasing values , it is clear that the contributions from the periodic potential to the entropy change will have a restricted effect on the overall @xmath68 .
nevertheless , the potential is the responsible for the peaked structure in @xmath68 .
the contribution of @xmath84 is qualitatively analogous to that of @xmath92 .
a stronger fluctuation relation has also been shown to hold for the probability distribution function of @xmath18 itself @xcite , @xmath93 it constrains directly the properties of the probability distribution function of @xmath18 .
this stringent relation is closely related to the studies performed in refs .
@xcite , where similar fluctuation relations are derived for entropy , work and heat in different contexts .
the previous relation follows @xcite from the fact that for an arbitrary function @xmath94 , @xmath95p_{0}(x_{0 } ) \ ; g(r)e^{-r}= \nonumber \\
= \sum_{\tilde{x}(t),\tilde{x}_{0 } } \tilde{p}[\tilde{x}(t)|\tilde{x}_{0}]p_{1}(\tilde{x}_{0 } ) \ ; g(-r[\tilde{x}(t ) ] ) = \langle g(-r ) \rangle,\end{aligned}\ ] ] where the symmetric property of @xmath18 , namely @xmath96=-r[\tilde{x}(t)]$ ] , is used and necessary to prove the above equation .
this important condition must be verified in every case .
here it is satisfied because @xmath18 reverses sign when interchanging @xmath70 and @xmath47 .
then , for the particular case @xmath97-r)$ ] , one recovers eq .
( [ lnpr ] ) . in the inset of fig .
( [ strongrel ] ) we check the detailed fluctuation theorem by relating the negative and positive contributions to the probability distribution of @xmath18 , leading to an excellent agreement with the theoretical prediction .
as we will show in detail , @xmath98 increases linearly with time .
therefore , the probability distribution function of @xmath18 moves towards positive values as time goes by .
then , concerning the verification of the detailed fluctuation theorem , the longer the time of observation , the poorer the statistics at the negative tails , thus limiting the possibility of a meaningful tests .
experimental tests have been performed in ref .
@xcite to verify the detailed fluctuation relation .
one can further take profit of eq .
( [ demos ] ) by considering @xmath99 .
a set of relations involving average moments of powers of the total entropy change of the system are found , @xmath100 in fig .
( [ moments ] ) we display the lower order moments ( @xmath101 ) as a function of the number of independent trajectories , @xmath73 .
theoretical predictions are recovered .
the dependence on @xmath73 shows again the statistical nature of the fluctuation theorem .
note that the behavior is analogous to the one in fig .
( [ e - r ] ) .
seifert s integral fluctuation theorem is valid for arbitrary initial conditions . however , eq . ( [ rmain ] ) holds for @xmath18 only in the steady state .
therefore , when sampling from other different initial conditions , the relations verified so far will not be reproduced anymore .
( [ init ] ) , displays the histograms at @xmath75 corresponding to the dynamics of the system when three different ensembles of initial conditions for the position are considered : steady state , gaussian distributed and a delta function at zero position . the inset shows the ratio @xmath102 as a function of @xmath18 for all three initial distributions .
one can clearly see that only for the steady state initial ensemble the detailed fluctuation theorem holds .
it is worthy to stress this point to prevent confusions and errors when applying the predictions provided by the theorem in real experiments , in which one should have careful control on the preparation of the initial state of the system .
it would be interesting , although more tedious , to test the predictions derived from the fluctuation theorem away from the steady state , in which no simple expression such as eq .
( [ rmain ] ) is available . for dissipative systems in a steady state
, @xmath18 is closely related to the so called housekeeping heat @xmath103 .
this is the heat that is permanently dissipated while maintaining a nonequilibrium state at fixed external parameters .
it has been shown @xcite that @xmath104 , at the same time that @xmath105 the steady flux @xmath106 is related to the probability @xmath48 according to eq .
( [ eq : flux_diff ] ) , which makes it possible to express the housekeeping heat in terms of the external driving and the particle flux as @xmath107 .
since @xmath108 , the mean heat produced can be rewritten in the suggestive form @xmath109 which implies that the mean total entropy change can be simply expressed in terms of the mean work produced and dissipated by the driving force , @xmath110 , as @xmath111 we can derive an explicit expression for the mean velocity of the particle @xcite @xmath112 remember that @xmath113 . again , for the potential @xmath8 given by eq .
( [ pot ] ) with @xmath63 , one gets , after some integrations , @xmath114 note that the mean velocity could be recasted in terms of an effective mobility @xmath115 , which would depend on the applied force , so that @xmath116 , for small forces , @xmath117 which would correspond to the free diffusion of a brownian particle in the periodic potential , and for asymptotically large forces , @xmath115 reduces to the bare mobility , @xmath118 , in scaled units . finally , coming back to the mean entropy production , the following analytical expression is then encountered , @xmath119 where @xmath120 . in fig .
( [ meanr ] ) we compare the theoretical prediction for @xmath98 with numerical results , exploring the force @xmath67 and the time @xmath13 . theory and simulations are in excellent agreement . the linear dependence in time of the mean value of @xmath18 yields to a constant mean entropy production rate . although @xmath98 increases linearly in time , for a single trajectory @xmath18 will fluctuate .
as a result , for short times and for typical energies of the system comparable to the thermal energy , negative values of @xmath18 will be observable , corresponding to the so - called `` transient violations of the second law '' .
these violations contribute significantly to the averages at the same time that determine the tails of the distributions , which are crucial when examining the detailed fluctuation relation .
in fact , negative values of @xmath18 are directly related to figs .
( [ e - r ] ) , ( [ p(r ) ] ) and ( [ strongrel ] ) , and their corresponding subsections . in fig . ( [ transient ] ) we plot the time evolution of the entropy production @xmath18 of a single trajectory .
one can see that at short times @xmath18 is likely to become negative , indicating that up to that point the total entropy production associated to that stochastic trajectory is negative .
the inset shows that at long times @xmath18 is positive and increases . upon averaging over a number of independent trajectories , the linear increase predicted in the previous section
is recovered .
hence , the study of single trajectories highlights the stochastic nature of fluctuation theorems and the need to distinguish between mean response of the system and the properties of individual trajectories . as the period of observation increases
, it becomes rapidly unlikely that negative values of @xmath18 are observed .
this , together with the energy scale , sets the regime where examining `` transient violations '' becomes impractical .
we have considered a simple system in a nonequilibrium steady state and have studied in detail many of the predictions and consequences derived from the fluctuation theorem in ref .
in particular , by choosing an appropriate periodic potential , we have been able to find an explicit analytical and physically meaningful expression for the total entropy production along a stochastic trajectory . advantage has been taken from the fact that only initial and final positions for each trajectory are relevant , being a potential useful feature in future experiments .
numerical simulations of the dynamics of the langevin equation have been necessary to build the probability distribution functions of @xmath18 .
we have elucidated the relevance and properties of @xmath68 in the system .
moreover , all of the results based , related , derived and predicted from the fluctuation relation have been tested satisfactorily .
this work leads to a more profound understanding of the specific meaning and practical appliance of such abstract general results . at the same time
, it opens the door to study new and more complex cases such as interacting particles or the effect of temperature dependent substrates .
furthermore , it would be very interesting to analyze fluctuations observed in biological systems such as molecular motors under the perspective of fluctuation relations .
studies based on simple physical models are crucial to illustrate the key ingredients and foresee the consequences of these theoretical results on real experiments .
therefore , they bridge the fundamental advances in statistical mechanics with the active area of nano scale experiments on biological systems .
we thank udo seifert for useful discussions .
this research was supported by the ministerio de educacin y ciencia ( spain ) under the grant fpu - ap-2004 - 0770 ( a. g - m . ) and by dgicyt of the spanish government and distinci de la generalitat de catalunya ( spain ) ( i. p. ) . | we derive a simple closed analytical expression for the total entropy production along a single stochastic trajectory of a brownian particle diffusing on a periodic potential under an external constant force . by numerical simulations
we compute the probability distribution functions of the entropy and satisfactorily test many of the predictions based on seifert s integral fluctuation theorem .
the results presented for this simple model clearly illustrate the practical features and implications derived from such a result of nonequilibrium statistical mechanics . |
optimal stopping problems arise in various areas ranging from the classical sequential testing / change - point detection problems to applications in finance .
although all formulations reduce to the problem of maximizing / minimizing the expected payoff over a set of stopping times , the solution methods are mostly problem - specific ; they depend significantly on the underlying process , payoff function and time - horizon .
this paper pursues a common tool for the class of _ infinite - time horizon _ optimal stopping problems for _ spectrally negative lvy processes _ , or lvy processes with only negative jumps . by extending the classical continuous diffusion model to the lvy model
, one can achieve richer and more realistic models .
in mathematical finance , the continuity of paths is empirically rejected and can not explain , for example , the volatility smile and non - zero credit spreads for short - maturity corporate bonds .
these issues can often be alleviated by introducing jumps ; see , e.g. @xcite .
naturally , however , the optimal stopping problem becomes more challenging and can not enjoy a number of results obtained under the continuity of paths . in the case of one - dimensional continuous diffusion ,
a full characterization of the value function is known and some practical methods have been developed ( see e.g. @xcite ) .
most of these results rely heavily on the continuity assumption ; once jumps are involved , only problem - specific approaches are currently available . despite these differences
, there exists a common tool known as the _ scale function _ for both continuous diffusion and spectrally negative lvy processes . the scale function for
the former enables one to transform a problem of any arbitrary diffusion process to that of a standard brownian motion . for the latter ,
the scale function has been playing a central role in expressing fluctuation identities for a general spectrally negative lvy process ( see @xcite ) . by taking advantage of the potential measure that can be expressed using the scale function
, one can obtain the overshoot distribution at the first exit time , which is generally a big hurdle that typically makes the problem intractable .
the objective of this paper is to pursue , with the help of the scale function , a common technique for the class of optimal stopping problems for spectrally negative lvy processes .
focusing on the first time it down - crosses a fixed threshold , we express the corresponding expected payoff in terms of the scale function .
this semi - explicit form enables us to differentiate and take limits thanks to the smoothness and asymptotic properties of the scale function as obtained , for example , in @xcite . by differentiating the expected payoff with respect to the threshold level , we obtain the _ first - order condition _ as well as the candidate optimal level that makes it vanish .
we also obtain the _ continuous / smooth fit condition _ when the process is of bounded variation or when it contains a diffusion component .
these conditions are in fact equivalent and can be obtained generally under mild conditions .
the spectrally negative lvy model has been drawing much attention recently as a generalization of the classical black - scholes model in mathematical finance and also as a generalization of the cramr - lundberg model in insurance .
a number of authors have succeeded in extending the classical results to the spectrally negative lvy model by way of scale functions .
we refer the reader to @xcite for stochastic games , @xcite for the optimal dividend problem , @xcite for american and russian options , and @xcite for credit risk .
in particular , egami and yamazaki @xcite modeled and obtained the optimal timing of capital reinforcement . as an application of the results obtained in this paper
, we give a short proof of the mckean optimal stopping ( perpetual american put option ) problem with additional running rewards , as well as an extension and its analytical solution to @xcite .
the rest of the paper is organized as follows . in section [ sec : problem ] , we review the optimal stopping problem for spectrally negative lvy processes , and then express the expected value corresponding to the first down - crossing time in terms of the scale function
. in section [ sec : fit ] , we obtain the first - order condition as well as the continuous / smooth fit condition and show their equivalence . in section [ section_extension ] ,
we solve the mckean optimal stopping problem and an extension to @xcite .
we conclude the paper in section [ section_conclusion ] .
let @xmath0 be a probability space hosting a spectrally negative lvy process @xmath1 characterized uniquely by the _ laplace exponent _
@xmath2 = c \beta + \frac{1}{2}\sigma^2 \beta^2 + \int _ { ( 0,\infty)}(e^{-\beta z}-1+\beta z 1_{\{0
< z<1\}})\,\pi({{\rm d}}z ) , \quad { \beta \in \mathbb{r } } , \label{laplace_exp}\end{aligned}\ ] ] where @xmath3 , @xmath4 and @xmath5 is a measure on @xmath6 such that @xmath7 here @xmath8 is the conditional probability where @xmath9 and @xmath10 is its expectation .
it is well - known that @xmath11 is zero at the origin , convex on @xmath12 and has a right - continuous inverse : @xmath13 in particular , when @xmath14 we can rewrite @xmath15 where @xmath16 the process has paths of bounded variation if and only if @xmath17 and holds .
it is also assumed that @xmath18 is not a negative subordinator ( decreasing a.s . ) .
namely , we require @xmath19 to be strictly positive if @xmath20 and holds .
let @xmath21 be the filtration generated by @xmath18 and @xmath22 be a set of @xmath21-stopping times .
we shall consider a general optimal stopping problem of the form : @xmath23 , \quad x \in { \mathbb{r}}\label{general_problem}\end{aligned}\ ] ] for some discount factor @xmath24 and locally - bounded measurable functions @xmath25 which represent , respectively , the payoff received at a given stopping time @xmath26 and the running reward up to @xmath26 .
typically , its optimal stopping time is given by the _ first down - crossing time _ of the form @xmath27 let us denote the corresponding expected payoff by @xmath28 , \quad x , a \in { \mathbb{r } } , \ ] ] which can be decomposed into @xmath29 where , for every @xmath30 , @xmath31 , \\
\gamma_2(x;a ) & : = { \mathbb{e}}^x \left [ e^{-q \tau_a } ( g(x_{\tau_a})-g(a ) ) 1_{\{x_{\tau_a } < a , \ , \tau_a < \infty \}}\right ] , \\
\gamma_3(x;a ) & : = { \mathbb{e}}^{x}\left [ \int_0^{\tau_{a } } e^{-qt } h(x_t ) { { \rm d}}t \right ] . \end{split } \label{definition_gamma}\end{aligned}\ ] ] shortly below , we express each term via the scale function .
this paper does not consider the _
first up - crossing time _ defined by @xmath32 because , for the spectrally negative lvy case , the process always creeps upward ( @xmath33 a.s . on @xmath34 ) , and the expression of the expected value is much simplified .
we focus on a more interesting and challenging case where the optimal stopping time is conjectured to be a first down - crossing time .
associated with every spectrally negative lvy process , there exists a ( q-)scale function @xmath35 that is continuous and strictly increasing on @xmath36 and is uniquely determined by @xmath37 fix @xmath38 . if @xmath39 is the first time the process goes above @xmath40 and @xmath41 is the first time it goes below zero as a special case of , then we have @xmath42 = \frac { w^{(q)}(x ) } { w^{(q)}(a ) } \quad \textrm{and } \quad { \mathbb{e}}^x \left [ e^{-q \tau_0 } 1_{\left\ { \tau_a^+ > \tau_0 , \ , \tau_0 < \infty \right\}}\right ] = z^{(q)}(x ) - z^{(q)}(a ) \frac { w^{(q)}(x ) } { w^{(q)}(a)},\end{aligned}\ ] ] where @xmath43 here we have @xmath44.\ ] ] we also have @xmath45 = z^{(q)}(x ) - \frac q { \phi(q ) } w^{(q)}(x ) , \quad x > 0 . \label{laplace_tau_0}\end{aligned}\ ] ] in particular , @xmath46 is continuously differentiable on @xmath6 if @xmath5 does not have atoms and @xmath46 is twice - differentiable on @xmath6 if @xmath47 ; see , e.g. , @xcite . throughout this paper , we assume the former .
we assume that @xmath5 does not have atoms .
fix @xmath48 .
the scale function increases exponentially ; @xmath49 there exists a ( scaled ) version of the scale function @xmath50 that satisfies @xmath51 and @xmath52 moreover @xmath53 is increasing , and as is clear from , @xmath54 regarding its behavior in the neighborhood of zero , it is known that @xmath55 see lemmas 4.3 - 4.4 of @xcite . for a comprehensive account of the scale function ,
see @xcite .
see @xcite for numerical methods for computing the scale function .
we now express in terms of the scale function . for the rest of the paper , because @xmath48 , we must have @xmath56 . first , the following is immediate by .
[ lemma_gamma_1 ] for every @xmath30 , we have @xmath57.\end{aligned}\ ] ] for @xmath58 and @xmath59 , we use the potential measure written in terms of the scale function . by using theorem 1 of @xcite
( see also @xcite ) , we have , for every @xmath60 and @xmath61 , @xmath62 & = \int_{b \cap [ a,\infty ) } \left [ \frac { w^{(q)}(x - a ) w^{(q ) } ( a - y ) } { w^{(q)}(a - a ) } - 1_{\ { x \geq y \ } } w^{(q ) } ( x - y ) \right ] { { \rm d}}y.\end{aligned}\ ] ] by taking @xmath63 via the dominated convergence theorem , we can obtain @xmath64 in .
for the problem to be well - defined , we assume throughout the paper the following so that @xmath59 is finite . for a complete proof of lemma [ lemma_gamma_3 ] below , see @xcite .
[ assump_h ] we assume that @xmath65 .
[ lemma_gamma_3]for all @xmath30 , we have @xmath66 for @xmath58 , we first define , for every @xmath67 , @xmath68 [ lemma_lipschitz ] fix @xmath67 .
suppose 1 .
@xmath69 is @xmath70 in some neighborhood of @xmath71 and 2 .
@xmath69 satisfies @xmath72 then @xmath73
. see appendix [ proof_lemma_lipschitz ] . for every @xmath30 , we also define @xmath74 by - , @xmath75 and hence the finiteness of @xmath76 also implies that of @xmath77 for any @xmath30 . using these notations , lemma [ lemma_gamma_3 ] together with the compensation formula
shows the following .
[ lemma_gamma_2 ] if ( 1)-(2 ) of lemma [ lemma_lipschitz ] hold for a given @xmath67 , then @xmath78 let @xmath79 be the poisson random measure associated with @xmath80 and @xmath81 for all @xmath82 .
we also let @xmath83 and @xmath84 for any @xmath85 . by the compensation formula ( see , e.g. , @xcite ) , @xmath86 \\ & = { \mathbb{e}}^x \left [ \int_0^\infty \int_0^\infty n({{\rm d}}t , { { \rm d}}u ) e^ { -q t } ( g(x_{t-}-u)-g(a))_+1_{\{x_{t- } - u \leq a , \ ; \underline{x}_{t- } > a \}}\right ] \\ & = { \mathbb{e}}^x \left [ \int_0^\infty e^ { -q t } { { \rm d}}t \int_0^\infty \pi ( { { \rm d}}u ) ( g(x_{t-}-u)-g(a))_+ 1_{\{x_{t- } - u \leq a , \ ; \underline{x}_{t- } > a \}}\right ] \\ & = \int_0^\infty \pi ( { { \rm d}}u ) { \mathbb{e}}^x \left [ \int_0^\infty e^ { -q t } ( g(x_{t-}-u)-g(a))_+1_{\{x_{t- } - u \leq a , \ ; \underline{x}_{t- } > a \ } } { { \rm d}}t \right ] \\ & = \int_0^\infty \pi ( { { \rm d}}u ) { \mathbb{e}}^x \left [ \int_0^{\tau_a } e^ { -q t } ( g(x_{t-}-u)-g(a))_+1_{\{x_{t- } \leq a + u\ } } { { \rm d}}t \right].\end{aligned}\ ] ] by setting @xmath87 or equivalently @xmath88 in lemma [ lemma_gamma_3 ] , @xmath89 \\ & = w^{(q ) } ( x - a ) \int_0^u e^{- \phi(q ) y } ( g(y+a - u ) - g(a))_+{{\rm d}}y - \int_a^x w^{(q ) } ( x - y ) ( g(y - u)-g(a))_+ 1_{\{y \leq a+u\ } } { { \rm d}}y \\ & = w^{(q ) } ( x - a ) \int_0^u e^{- \phi(q ) y } ( g(y+a - u)-g(a))_+{{\rm d}}y - \int_0^{u \wedge ( x - a ) } w^{(q ) } ( x - z - a ) ( g(z+a - u)-g(a))_+ { { \rm d}}z.\end{aligned}\ ] ] by substituting this , we have @xmath86 \\ & = \int_0^\infty \pi ( { { \rm d}}u ) \left [ w^{(q ) } ( x - a ) \int_0^u e^{- \phi(q ) y } ( g(y+a - u)-g(a))_+ { { \rm d}}y \right . \\ & \qquad \left .
- \int_0^{u \wedge ( x - a ) } w^{(q ) } ( x - z - a ) ( g(z+a - u)-g(a))_+ { { \rm d}}z \right],\end{aligned}\ ] ] which is finite by lemma [ lemma_lipschitz ] and .
similarly , we can obtain @xmath90 $ ] and is immediate by subtraction . in view of above
, we can also write @xmath91
the most common way of choosing the candidate threshold level is the continuous and smooth fit principle .
define @xmath92 if these limits exist .
the continuous and smooth fit chooses @xmath71 such that @xmath93 and @xmath94 , respectively .
alternatively , one can differentiate @xmath95 with respect to @xmath71 and obtain the first - order condition . in this section
, we pursue the candidate threshold level @xmath96 in both ways .
we first obtain , for a general case , the first - derivative @xmath97 and @xmath71 that makes it vanish , and then the continuous fit condition for the case @xmath18 is of bounded variation and the smooth fit condition for the case @xmath18 has a diffusion component ( @xmath47 ) .
we further discuss the equivalence of these conditions and how to obtain optimal strategies .
we shall obtain @xmath98 for @xmath30 that satisfies ( 1)-(2 ) of lemma [ lemma_lipschitz ] .
let @xmath99 [ lemma_derivative_a ] for given @xmath30 , suppose ( 1)-(2 ) of lemma [ lemma_lipschitz ] hold and @xmath100 for some @xmath101 .
then , we have @xmath102 where @xmath103 because @xmath104 is increasing , @xmath105 is positive ( see also @xcite for an interpretation of @xmath105 as the resolvent measure of the ascending ladder height process of @xmath18 ) and hence @xmath106 if there exists @xmath96 such that @xmath107 then the stopping time @xmath108 naturally becomes a reasonable candidate for the optimal stopping time . in order to show proposition [ lemma_derivative_a ] above , we obtain the derivatives of @xmath109 for @xmath110 with respect to @xmath71 for any @xmath30 . by applying straightforward differentiation in lemma [ lemma_gamma_1 ] and because @xmath111 , @xmath112 + g(a ) \frac q { \phi(q ) } \theta^{(q)}(x - a ) .
\end{split } \label{derivative_gamma_1_a}\end{aligned}\ ] ] for @xmath58 , we first take the derivatives of with respect to @xmath71 .
[ lemma_derivative_a_individual ] fix @xmath30 .
under the assumptions in proposition [ lemma_derivative_a ] , @xmath113 { { \rm d}}y \\
= e^{-\phi(q ) a } \int_0^\infty \pi ( { { \rm d}}u ) \big [ g(a)-g(a - u ) - \frac { 1 - e^{-\phi(q ) u } } { \phi(q ) } g'(a ) \big],\end{gathered}\ ] ] and @xmath114 .
\label{derivative_a_individual_2}\end{aligned}\ ] ] see appendix [ proof_lemma_derivative_a_individual ] . by applying lemma [ lemma_derivative_a_individual ] in - , the derivative of @xmath58 with respect to @xmath71 is immediately obtained .
[ derivative_a_gamma_2 ] fix @xmath30 . under the assumptions in proposition [ lemma_derivative_a ] , @xmath115 for @xmath59 , as in the proof of lemma 4.4 of @xcite
, we have the following .
although the continuity of @xmath116 is assumed throughout in @xcite , it is not required in the following lemma ; this is clear from the proof of lemma 4.4 of @xcite .
[ derivative_a_gamma_3 ] for every @xmath30 , @xmath117 we are now ready to prove proposition [ lemma_derivative_a ] . by combining and lemmas [ derivative_a_gamma_2]-[derivative_a_gamma_3 ] , we obtain @xmath118 where @xmath119 by lemmas [ lemma_gamma_1 ] and modifying lemma [ lemma_gamma_2 ] , we can also write @xmath120 - { \mathbb{e}}^x\left [ e^{-q \tau_a } 1_{\{x_{\tau_a } < a , \ , \tau_a < \infty\}}\right ] = { \mathbb{e}}^x\left [ e^{-q \tau_a } 1_{\{x_{\tau_a } = a , \ , \tau_a < \infty \}}\right ] , \quad x > a.\end{aligned}\ ] ] a spectrally negative lvy process creeps downward if and only if there is a gaussian component , i.e. , @xmath121 for any @xmath30 if and only if @xmath47 ; see @xcite exercise 7.6 .
hence @xmath122 this proves the desired result for the case @xmath20 . for the case
@xmath47 , as in @xcite , we can also write @xmath123 and hence it also holds when @xmath47 as well .
we now pursue @xmath96 such that @xmath124 and @xmath125 for the cases 1 .
@xmath18 is of bounded variation , and 2 .
@xmath47 , respectively .
we exclude the case @xmath18 is of unbounded variation with @xmath20 ( in this case , @xmath126 by and hence the interchange of limits over integrals we conduct below may not be valid )
. however , this can be alleviated and the results hold generally for all spectrally negative lvy processes when @xmath69 is a constant in a neighborhood of @xmath96 .
examples include @xcite where @xmath127 on @xmath6 and @xcite where @xmath128 on @xmath129 $ ] and @xmath130 on @xmath6 ; see section [ section_extension ] . for continuous fit , we need to obtain @xmath131 if these limits exist .
define also @xmath132 , if it exists .
it is easy to see that @xmath133 the result for @xmath58 is immediate by the dominated convergence theorem thanks to lemma [ lemma_lipschitz ] and - . [ lemma_rho ] given ( 1)-(2 ) of lemma [ lemma_lipschitz ] for a given @xmath67 , we have 1 .
@xmath134 , 2 .
@xmath135 .
now lemma [ lemma_rho ] and show @xmath136 this together with shows the following .
[ lemma_continuous_fit ] fix @xmath67 and suppose ( 1)-(2 ) of lemma [ lemma_lipschitz ] hold . 1 .
if @xmath18 is of bounded variation , the continuous fit condition @xmath93 holds if and only if @xmath137 2 . if @xmath18 is of unbounded variation ( including the case @xmath20 ) , it is automatically satisfied . for the case
@xmath18 is of unbounded variation with @xmath47 , we shall pursue smooth fit condition at @xmath67 .
the following lemma says in this case that the derivative can go into the integral sign and we can further interchange the limit .
[ lemma_for_smooth_fit ] fix @xmath67 .
if @xmath47 and suppose ( 1)-(2 ) of lemma [ lemma_lipschitz ] hold , then @xmath138 { { \rm d}}z , \quad x > a , \label{varrho_derivative_g}\end{aligned}\ ] ] and @xmath139 see appendix [ proof_lemma_for_smooth_fit ] . in the case of unbounded variation with @xmath20
, it is expected that holds but does not .
this is because @xmath126 and the limit can not go into the integral .
we are now ready to obtain @xmath140 for @xmath110 .
[ lemma_smooth_fit ] fix @xmath67 .
suppose @xmath47 and ( 1)-(2 ) of lemma [ lemma_lipschitz ] hold .
then , 1 .
@xmath141 , 2 .
@xmath142 , 3 .
@xmath143 .
\(1 ) it is immediate by lemma [ lemma_gamma_1 ] .
( 2 ) by , @xmath144 by taking @xmath145 via , we have the claim .
( 3 ) using in particular @xmath146 , we have @xmath147 \\ & = w^{(q ) ' } ( 0 + ) \int_0^\infty e^{- \phi(q ) y } h(y+a ) { { \rm d}}y.\end{aligned}\ ] ] by the lemma above , we obtain @xmath148 or equivalently , by virtue of , the smooth fit condition at @xmath96 is equivalent to .
[ lemma_smooth_fit2 ] fix @xmath67 .
suppose @xmath47 and ( 1)-(2 ) of lemma [ lemma_lipschitz ] hold .
then , the smooth fit condition @xmath94 holds if and only if @xmath149 we summarize the results obtained in propositions [ lemma_continuous_fit]-[lemma_smooth_fit2 ] in table [ table : summary ] .
.summary of continuous- and smooth - fit conditions .
[ cols="^,^,^",options="header " , ] under specification [ assumption_g_h ] , there exists at most one @xmath150 that satisfies because @xmath151 + * verification of optimality : * we let @xmath96 be the unique root of @xmath152 if it exists and set it zero otherwise . as in the case of the mckean
optimal stopping problem , we only need to show ( iii ) in section [ subsection_obtaining ] because ( i ) holds by and lemma [ lemma_about_i ] and ( ii ) by proposition [ proposition_harmonic ] .
if @xmath150 , we have @xmath153 for every @xmath154 .
because @xmath155 for every @xmath156 , @xmath157 we shall show that this is negative . because @xmath150 , we must have @xmath158 and hence @xmath159 where the first inequality holds because @xmath69 and @xmath116 are increasing and @xmath160 , the second holds because @xmath69 is nonpositive , and the third holds because @xmath160 and @xmath69 is nonpositive .
this together with shows the result .
for the rest of the proof , we refer the reader to the proof of proposition 4.1 in @xcite . if @xmath150 , then @xmath108 is the optimal stopping time and the value function is given by @xmath161 for every @xmath156 .
if @xmath162 , then the value function is given by @xmath163 for every @xmath156 .
we have discussed the optimal stopping problem for spectrally negative lvy processes . by expressing the expected payoff via the scale function
, we achieved the first - order condition as well as the continuous / smooth fit condition and showed their equivalence .
the results obtained here can be applied to a wide range of optimal stopping problems for spectrally negative lvy processes . as examples
, we gave a short proof for the perpetual american option pricing problem and solved an extension to egami and yamazaki @xcite .
for future research , it would be interesting to pursue similar results for optimal stopping games .
typically , the equilibrium strategies are given by stopping times of threshold type as in @xcite .
similarly to the results obtained in this paper , the expected payoff admits expressions in terms of the scale function and hence the first - order condition and the continuous / smooth fit can be obtained analytically .
another direction is the extension to a general lvy process with both positive and negative jumps .
this can be obtained in terms of the wiener - hopf factor alternatively to the scale function .
finally , the results can be extended to a number of variants of optimal stopping such as optimal switching , impulse control and multiple stopping .
we choose @xmath172 that satisfies and fix @xmath173 . by the mean value theorem , there exists @xmath174 such that @xmath175 because , for every @xmath176 , we have @xmath177 the taylor expansion implies that , for every @xmath178 , @xmath179 @xmath180 hence uniformly in @xmath181 by @xmath182 on the other hand , by , @xmath183 which is finite by and how @xmath184 is chosen .
this allows us to apply the dominated convergence theorem , and we obtain @xmath185 the proof for the left - derivative is similar , and this completes the proof of . :
define @xmath186 { { \rm d}}y , \quad z \in { \mathbb{r}}\ ; \textrm{and } \ ; u > 0 . \label{q_tilde}\end{aligned}\ ] ] then , by , we have @xmath187 .
we use the same @xmath172 as in the proof of above and fix @xmath188 and @xmath189 such that @xmath190 \wedge \delta.\end{aligned}\ ] ] it is then clear that @xmath191 . we shall split @xmath192 and show that these two terms on the right - hand side are bounded in @xmath188 on @xmath193 . for every fixed @xmath194 ,
our assumptions imply that @xmath195 is @xmath70 on @xmath196 . by the mean value theorem
, there exists @xmath174 such that @xmath197 given @xmath198 at which @xmath69 is differentiable and also satisfying @xmath199 , differentiating obtains @xmath200 because @xmath201 ( or @xmath202 ) and @xmath69 is differentiable at @xmath203 , @xmath204 - g'(a+\xi ) \int_{a+\xi}^{u+a+\xi } w^{(q)}(x - y ) { { \rm d}}y \right| \\ & \leq w^{(q)}(x - a-\xi ) \left| g(a+\xi)-g(a+\xi - u ) - u g'(a+\xi ) \right| \\ & \qquad + \left| g'(a+\xi ) \int_{a+\xi}^{u+a+\xi } ( w^{(q)}(x - a-\xi)-w^{(q)}(x - y ) ) { { \rm d}}y \right| \\ & \leq w^{(q)}(x - a ) \left| g(a+\xi)-g(a+\xi - u ) - u g'(a+\xi ) \right|
u |g'(a+\xi)| \left| w^{(q)}(x - a-\xi)-w^{(q)}(x - u - a-\xi ) \right| \\ & \leq f_1(a;u , x ) + f_2(a;u , x)\end{aligned}\ ] ] where @xmath205 first , @xmath206 is finite because , for every @xmath207 , we have @xmath208 and @xmath209 which is @xmath5-integrable over @xmath210 by . on the other hand , by and because @xmath211 implies @xmath212 , we have @xmath213 and hence @xmath214 which is finite by .
we now obtain the bound for the second term of the right - hand side of . for every @xmath215 ( which implies @xmath216 ) , we have @xmath217 where @xmath218 , \\
b_2(a , c;u , x ) & : = \frac { |g(a+c ) - g(a)| } c \int_{a+c}^{(u+a ) \wedge x } w^{(q)}(x - y ) { { \rm d}}y.\end{aligned}\ ] ] for the former , we have @xmath219 here the first inequality holds because @xmath220 .
for the second inequality , it holds trivially when the maximum is attained for some @xmath221 . if it is attained at @xmath222 for some @xmath223 .
then , because @xmath224 ( thanks to @xmath225 ) and @xmath226 @xmath227 for the latter , by the @xmath70 property of @xmath69 in the neighborhood of @xmath71 , how @xmath184 is chosen and @xmath226 , we obtain @xmath228 therefore , @xmath229 which is bounded in @xmath188 on @xmath193 by and . it is known as in @xcite that @xmath47 guarantees that @xmath104 is twice continuously differentiable and hence @xmath231 is continuous on @xmath6 .
furthermore , implies @xmath232 and implies @xmath233 .
therefore , there exists @xmath234 such that @xmath235 because @xmath239 we have @xmath240 on the other hand , @xmath241 by , these bounds on @xmath242 and @xmath243 also bound , and hence by the dominated convergence theorem and because @xmath146 , @xmath244 the left - derivative can be obtained in the same way .
this proves .
choosing @xmath164 and @xmath253 as in the proof of lemma [ lemma_lipschitz ] , we obtain by @xmath254 because , by , @xmath255 \\ & \leq e^{\phi(q)(x - a ) } \big(lu + \frac { 1 - e^{-\phi(q ) u } } { \psi'(\phi(q ) ) } \big),\end{aligned}\ ] ] we have @xmath256 with @xmath257 which is finite thanks to and by applying the taylor expansion to @xmath258 to the first integral .
hence ( 1 ) is obtained .
the proof of the existence of @xmath259 that satisfies ( 2 ) is immediate because @xmath260 and @xmath261 by . | this paper is concerned with a class of infinite - time horizon optimal stopping problems for spectrally negative lvy processes .
focusing on strategies of threshold type , we write explicit expressions for the corresponding expected payoff via the scale function , and further pursue optimal candidate threshold levels .
we obtain and show the equivalence of the continuous / smooth fit condition and the first - order condition for maximization over threshold levels .
this together with problem - specific information about the payoff function can prove optimality over all stopping times .
as examples , we give a short proof of the mckean optimal stopping problem ( perpetual american put option ) and solve an extension to egami and yamazaki @xcite . |
the coronet cluster ( also known as cra ) is an obscured star - forming region , located at 170 pc ( knude & hog 1998 ) , associated to the herbig ae star r cra ( taylor & storey 1984 ) and a dense molecular core ( loren 1979 ) .
submillimeter , millimeter , ir , and optical surveys ( henning et al .
1994 ; de muizon et al . 1980 ; wilking et al . 1985
; marraco & rydgren 1981 ) found a few haebe and t tauri stars ( tts ) , suggesting a modest young stellar population of @xmath21 myr in age .
x - ray surveys using einstein , xmm - newton , rosat and chandra ( walter 1986 , 1997 ; neuhuser et al . 2000 ; hamaguchi et al . 2005 ; garmire & garmire 2003 ; forbrich & preibisch 2007 ) , less affected by extinction , identified more than 70 sources consistent with young stars , many of which are deeply embedded , all located within a region a bit smaller than 1 pc in diameter
. the presence of very embedded sources , including class 0 objects , suggested a very young region ( henning et al .
1994 ; chen et al . 1997 ; chini et al . 2003 ; groppi et al .
2004,2007 ; nisini et al . 2005
; nutter et al . 2005 ) . optical studies also revealed an important population of low - mass stars , characterized by strong h@xmath3 emission ( lpez - mart et al .
the variety of sources makes the small coronet cluster a very special place : containing class 0 , i , ii , and iii objects , it seems to be a relatively quiescent cluster , part of a larger molecular cloud where star formation has occurred only in the densest parts . spitzer irac / mips data and irs spectra revealed more than 100 low and very low - mass sources , mostly class i , class ii , and class iii objects , some of them highly extincted , constituting a very young and relatively rich cluster ( sicilia - aguilar et al .
2008 , from now on sa08 ; currie & sicilia - aguilar 2011 ) .
remarkably , the irs spectra of the disks around m - type cluster members ( which are the bulk of the cluster members ) reveal features that are usually accepted as signs of disk evolution : lack of silicate emission , spectral energy distributions ( sed ) consistent with flattened / settled disks , very low or absent near - ir excesses characteristic of optically thin inner disk / inner holes ( sa08 ) .
in addition , the estimated disk fraction is lower than expected for a young , embedded cluster , with only 2/3 of the stars being surrounded by protoplanetary disks .
moreover , recent membership searches ( lpez - mart et al .
2010 ) have found additional diskless low - mass members , which would reduce the disk fraction to about 50% , a number more consistent with 2 - 4 myr old regions than with a 1 - 2 myr old cluster ( sicilia - aguilar et al .
2006 ; hernndez et al .
2007 ; fedele et al .
2010 ) . here
we present new flames / vlt optical spectroscopy , deep ir imaging with hawk - i / vlt , and a 870@xmath0 m map taken with apex / laboca of the coronet cluster .
the data provide spectral types and accretion constrains for most of the remaining low - mass cluster members with moderate extinction , as well as a detailed image of the structures and substructures in the region , with associated class 0/i embedded sources .
the datasets and reduction techniques are discussed in section [ data ] .
spectral types , accretion , extinction maps , and the cluster structure and embedded members are presented in section [ analysis ] . in section [ age ]
we discuss the age of the coronet cluster based on the identification of class 0/i / ii / iii sources and cluster structure , and finally our conclusions are presented in section [ conclu ] .
a large number of members and potential members had been already observed with flames on the ut2/vlt ( see sa08 ) .
the remaining candidates were done as part of the eso program 083.c-0079 , following a similar scheme to our previous observations .
the data were taken in 5 nights ( 2009 may 24 , and 2009 june 3 , 17 , 18 , 19 ) , using flames together with the multifiber spectrograph giraffe and the medusa fibers , which allow to take up to 130 spectra , covering both objects and sky positions .
we observed one field centered near the cluster center with three different , intermediate - resolution ( r@xmath25600 - 8600 ) , gratings ( l682.2 , l773.4 , l881.7 ) .
the resulting spectra cover thus a large wavelength range from 6440to 9400 , which includes the main accretion features ( h@xmath3 , ca ii ) , the lithium i 6708 absorption line ( indicative of youth ) , and a large number of photospheric features and absorption bands , required for a spectral type classification of m - type stars .
we obtained 3@xmath42700 s exposures with each filter , which ensures good removal of cosmic rays .
the target list included the potential members detected via x - ray ( garmire & garmire 2003 ) that had not been observed in our previous run , together with a large number of sky positions ( more than 70 in each one of the filters ) to ensure a good background subtraction in this region with variable nebular emission .
in addition , we also assigned fibers to potential herbig haro ( hh ) objects in the cloud ( wang et al .
2004 ) and to the objects that , having being detected in our previous flames observations , presented low s / n which made difficult a spectral type classification ( especially , the disked stars g-85 and g-87 ) . in total , 12 sources were detected ( see table [ lines - tab ] ) .
the lack of detections among the remaining x - ray candidates confirms their status as extragalactic or highly embedded sources , as presumed in sa08 ( see sections [ hawki ] and [ protostars ] for more detailed information on individual sources ) .
the spectra were reduced using standard tasks within iraf .
the data were bias corrected and then extracted with the help of corresponding dome flats , flat fielded , and calibrated in wavelength using the available thar lamp .
individual spectra were then combined .
no flux calibration was performed , due to the inherent difficulties of calibrating absolute fluxes with a multifiber spectrograph and to the fact that flux calibration is irrelevant for our science case .
since the sky emission is largely variable throughout the field , the sky subtraction was performed in several steps .
first , we combined all the sky spectra for each filter and subtracted from all the objects .
we then examined each object to determine whether this general sky subtraction properly removed nebular and sky lines . for the objects located in the brightest parts of the nebula ( cra-432 , g-80 , g-85 , g-87 , g-88 , g-90 , g-32 , and g-48 ) , for which the standard sky subtraction proved insufficient , we then selected the brightest sky spectra in the region ( 29 in total ) and combined them in a new `` bright '' sky template , which provided an adequate removal of sky lines .
figure [ spec - fig ] shows the high- and moderate - s / n spectra , and figures [ halpha - fig ] and [ lines - fig ] provide more details on h@xmath3 emission and other accretion - related lines . using the ir camera hawk - i on the vlt
, we observed 11 [email protected] fields in the coronet cluster , as part of our eso program 083.c-0079 .
the location of the fields was determined by the presence of bright objects , given the minimum magnitude limits of hawk - i ( ks=8.6 mag , h=9.6 mag , j=9.3 mag ) .
due to these strong magnitude limitations , the central part of the cluster and most of the areas with strong star formation could not be observed with hawk - i .
observations took place on 6 different nights , 2009 april 28 and 29 , 2009 june 28 and 30 , 2009 july 12 and 16 , under clear weather and variable seeing conditions ( 0.5 - 1.0 " ) .
the data was reduced using iraf tasks within the _ ccdred _ package for bias correction and flat fielding . for each field and filter
, we took 6 dithered images with offsets of 30 " and ndit=20 , with total exposure times of 40s ( j ) , 34s ( h ) , and 58s ( k ) .
we used the iraf _ daofind _ package to identify detections , and _ apphot _ to perform aperture photometry , with 10 pix aperture and 20 - 25 pix sky annuli .
the photometry was calibrated field - by - field and chip - by - chip as relative photometry by comparing to 2mass objects present in the fields , using typically 20 - 120 stars for calibration .
typical calibration errors range from 1 to 5% , depending on the weather conditions .
therefore , signal to noise and background remain the main sources of uncertainty , especially for the fainter objects .
the data are complete in a large dynamical range ( j=11 - 18.5 mag , h=11 - 18.5 mag , k=10.5 - 18 mag ) .
table [ hawki - table ] lists the magnitudes of the previously known sources detected with hawk - i .
two of the x - ray sources in garmire & garmire ( 2003 ) that had not been detected with spitzer nor flames ( g-116 and g-119 ) are now detected with hawk - i .
their colors suggest that they are most likely extragalactic .
table [ hawki - table ] also includes a comment on the variability of the sources that are also detected in 2mass .
cra-4107 and cra-452 show some degree of variability ( 0.1 - 0.3 mag , depending on color ) which is probably due to rotation and the presence of spots .
the j magnitude of g-32 appears more than 1 magnitude fainter in our observations than in the 2mass catalog , but since the 2mass magnitude is close to the detection limit of the survey , the hawk - i data are more accurate .
the 870@xmath0 m data were obtained with apex using the laboca ( large apex bolometer camera ) bolometer array as part of the eso program 081.c-0204(a ) .
the data consist of 108 scans in raster map spiral mode , covering an approximate area 30 in diameter .
this results in an integration time of 9.2 hours on target .
the observations were carried out between 2008 june 21 and 2008 august 24 .
the opacity was determined with skydips performed at the beginning and the end of the night , varying between @xmath5=0.15 - 0.44 , with a median value 0.2 .
the flux calibration was estimated observing several sources per night , including neptune , uranus , and the secondary calibrators j1957 - 3845 and j1924 - 292 .
the map was created and calibrated following the pipeline reduction of laboca data within the boa package , especially developed for apex bolometer data .
boa allows to reduce skydips to derive the opacity , reduce the calibrators , and make the maps .
the 108 individual maps were reduced separately , and then coadded to obtain the final map . in each map ,
the data affected by high acceleration were flagged , as well as the dead and very noisy channels ( with rms larger than 5 times the median ) .
correlated noise was removed by comparing groups of channels .
finally , a linear baseline was fitted and subtracted from the data .
the final maps were obtained using the @xmath6 option , which is optimized for weak sources and also provided the best result for the extended emission in the region .
the final rms in the map center is @xmath20.010 jy , being higher near the brightest sources and the less - integrated edges of the map .
the resulting map is displayed in figure [ laboca - fig ] .
the second step was to identify and extract the detected structures in the map and to measure the fluxes .
the cluster contains a large , dense structure with several peaks as well as some elongated structures and isolated clumps ( see figure [ laboca - fig ] ) . the identification and extraction of clumps in the elongated and isolated structures
was done using the 2d clumpfind algorithm , developed by williams et al .
the 2d clumpfind works by contouring the data according to user - defined levels in order to determine the presence of structures . to efficiently detect the clumps in the different parts of the map , which have in general different sensitivity , we defined levels according to the local rms .
the most noisy parts of the maps were removed before running clumpfind .
the clump identification was repeated 5 times with various sets of levels , and only clumps detected in 3 or more trials are considered to be real .
several sources appear as a single clump or as a collection of adjacent clumps depending on the search parameters .
given the uncertainty on the source boundaries and fluxes , they are listed as a `` multiple source '' with a unique flux in table [ laboca - tab ] , with names that indicate that they are composed of more than one peaks or individual sources ( 7 + 10 + 24 , 8 + 22 , 9 + 14 , 13 + 25 , 17 + 20 + 26 , 18 + 21 ) .
the flux for each clump ( or clump collection ) was calculated with @xmath7 , also within the 2d clumpfind package .
the errors in the final flux include the local rms as well as the standard deviation from the flux measured by the slightly different levels imposed for the source detection .
the final clumps are listed in table [ laboca - tab ] . in general ,
small , compact structures are detected in a very robust way , while the limits and subdivisions of more extended structures are less well constrained .
several of the submillimeter sources are associated to known ir sources and/or nebular structures ( figure [ contours - fig ] ) .
we used the near - infrared colors of the sources detected in the hawk - i frames to calculate the dust extinction through the cra cloud .
in particular , we adopted the color - excess mapping method @xcite . in the following ,
we briefly describe the practical implementation of the method . for further details of the method
we refer to @xcite and @xcite .
the observed colors of stars shining through a dust cloud are related to the dust extinction via the equation : @xmath8 where @xmath9 is the color - excess , @xmath10 is extinction , @xmath11 is optical depth , and @xmath12 refers to the intrinsic color of stars , measured as the mean color of stars in a field that is free from extinction .
we calculated the mean colors using stars in a hawk - i frame close to ( @xmath13 , @xmath14 j2000 ) = ( 19:02:50 , -36:46:25 ) , estimated to contain very little extinction based on our earlier , lower - resolution extinction map of the cra region @xcite .
the mean colors in this field were @xmath15 mag and @xmath16 mag .
for the ratios of optical depths we adopted @xmath17 , @xmath18 , and @xmath19 @xcite . before calculating the extinction , we removed from the hawk - i detections the previously known members of the cra cloud , based on the identifications given by sa08 , forbrich & preibisch ( 2007 ) , and lpez - mart et al .
( 2005 , 2010 ) .
observations in three broadband filters ( @xmath20 ) result in measurements of two color excesses , namely @xmath21 and @xmath22 , towards each source .
these two measurements were combined to yield extinction values towards the sources following the formalism presented in @xcite .
the derived extinction values were then used to produce a regularly sampled map by computing the mean extinction in regular intervals ( map pixels ) of @xmath23 , using a gaussian with @xmath24 as a spatial weighting function . in this resolution , the maximum extinction in the resulting map reaches @xmath25 mag , with @xmath26 mag corresponding to 1 - 2 stars per pixel .
the error in the map depends on extinction , with 3-@xmath27 being about 0.8 mag at @xmath28 mag and 7 mag at @xmath29 mag .
figure [ extinction - fig ] shows the resulting extinction map , combined with the 2mass extinction map in the regions not included in our hawk - i imaging .
there is an excellent agreement between the 870@xmath0 m clumps and the regions of high extinction .
nevertheless , the extinction map reveals large parts of the molecular cloud that are not detected in the laboca survey due to their lower column density .
the most massive stars in the region are associated to the zones with the highest extinction and strongest submillimeter emission , as expected .
the distribution of low mass members with respect to the cloud is less clear , also due to the fact that the non - biased survey for low - mass members was done only in the central part of the cluster .
the flames spectra allow to determine spectral types for the objects with good s / n .
we followed a similar scheme as used in sa08 , using the spectral indices and bands from martn et al .
( 1996 ) , called pc1 , pc2 , pc3 and pc4 , and the r1 , r2 , r3 , tio 8465 indices in riddick et al .
the additional indices vo 2 and vo 7445 from riddick et al .
( 2007 ) were also used to refine the spectral types for objects m5 or later .
table [ index - table ] offers a list of the relevant indices and their valid spectral type ranges . a combination of spectral types and the available jhk photometry allows to estimate the extinction of the objects ( see below ) . for the objects with intermediate to high extinction ( a@xmath303 mag ) , we gave priority to the results derived from the indices pc1 , pc2 , r1 , r2 , r3 , tio 8475 , vo 2 , and vo 7445 , which are derived from flux rations in adjacent parts of the spectrum , being thus less sensitive to extinction .
for objects like cra-432 , that have nearly no signal in the bluer part of the spectrum ( l682.2 ) , we used only the indices involving the longer - wavelength parts . for every object ,
the spectral types derived from each index were compared , and those that largely deviated from the typical value were removed and not used ( deviating indices are usually affected by noise , atmospheric features , and/or poor s / n in one of the bands ) . in general , between 2 - 6 indices were used for the classification of each star , depending on the spectral type , extinction , and s / n in the three spectral bands .
the errors in the spectral types are calculated as the standard deviation of the values derived from each valid index .
the classification scheme failed for the object g-1 , which appears earlier than the rest .
it was thus classified by comparison to solar - type t tauri star templates obtained from a combination of hectospec / mmt observations in the cep ob2 region ( sicilia - aguilar et al . 2005 ) .
two of the objects , g-85 and g-87 , are highly extincted ( a@xmath3014 mag ) .
after a first tentative classification , we used the preliminary a@xmath31 value to de - redden the spectra ( using iraf task @xmath32 within the @xmath33 package ) before the final classification .
examination of two objects revealed that their spectral types are earlier than m2 , a range for which the mentioned indices are not valid .
we thus derived their spectral types using our solar - type t tauri templates as we did for g-1 ( see figure [ g8587-fig ] ) .
note that g-85 and g-87 had been previously observed in sa08 , but the new improved s / n ( due to better weather and seeing conditions ) now allows to refine their spectral types , confirming them to be earlier than previously thought by about 2 - 3 subtypes ( having g-85 and g-87 spectral types [email protected] and [email protected] , respectively ) .
the star cra-468 , also known as g-49 , had also previously been observed with flames ( sa08 ) .
the new spectrum is fully consistent with the previous one ( see figure [ halpha - fig ] ) , resulting in a similar h@xmath3 classification and spectral type ( [email protected] versus [email protected] ) .
the final spectral types are listed in table [ lines - tab ] .
the individual extinction values were derived from the jhk 2mass photometry , comparing to standard colors for stars and bd ( bessell et al .
1998 ; kirkpatrick et al . 1995 ) .
the extinction was derived from e(j - h ) and e(h - k ) , using the relations in cardelli et al .
( 1989 ) , a@xmath31=10.87 e(j - h ) and a@xmath31=5.95 e(j - k ) mag , valid for a standard galactic extinction law . for most sources ,
the agreement between a@xmath31 obtained from the two different colors suggests that most of the extinction is due to foreground and cloud material , with no evidence of anomalous extinction that could be caused by processed grains in the circumstellar environment , despite the presence of strong disk emission in some of the objects . only in the cases of cra-4108 and cra-468 we found large differences between the extinction values derived from different colors , which could indicate the presence of anomalous extinction / scattering . in the case of cra-468 ,
the spitzer data reveals it is located near substantial nebular material .
cra-4108 is out of our spitzer field .
the presence or lack of accretion was established by measuring the equivalent width and examining the profile and line width at 10% of the peak for the h@xmath3 line ( see figure [ halpha - fig ] ) .
the lines were measured using iraf tasks within the @xmath35 tool .
we adopted the classification of white & basri ( 2003 ) for the limiting ew of accreting and non - accreting stars , depending on their spectral type .
given that the line width is a more powerful tool than the ew to detect weak accretors ( sicilia - aguilar et al .
2006 ) , we consider an object to be accreting if its 10% line width exceeds 200 km / s ( natta et al . 2005 ) . in some cases , like g-1 ,
the broad profile is evident , although the ew is much smaller than the canonical limit for accretion in an m0.5 object , so we list it as an accretor .
the presence of accretion was also checked in other lines , like the ca ii ir triplet and the 6678 he i line .
the lack of ca ii emission in g-1 suggests an accretion rate below 10@xmath36 m@xmath37/yr ( muzerolle et al .
all the spectral lines measured are listed in table [ lines - tab ] .
the new accretion determination , together with our previous measurements in sa08 , confirms that most of the low - mass coronet members have very low accretion rates , with cra-432 and g-85 being the only exceptions . as we found in sa08 , transitional and highly evolved disks
( like g-87 ) do not show evidence of ongoing accretion ( within the 10@xmath38 - 10@xmath39 m@xmath37/yr limit imposed by the flames spectral resolution ) .
this differs from the observations of solar - type stars , for which usually half or more of the transition disks are found to be accreting ( sicilia - aguilar et al .
2006 , 2010 ; muzerolle et al .
2010 ; fang et al . 2009 ) .
all objects without ir excesses are consistent with no accretion . for the objects out of the spitzer fields ( cra-4108 and cra-452 )
, the spectroscopy reveals no evidence of accretion .
cra-4108 shows marginal emission that could belong to he i at 5875 , but the lack of ca ii lines in emission suggests that the object is not accreting ( see figure [ lines - fig ] ) .
the herbig haro ( hh ) object g-80 was reobserved in this program , confirming again its shock nature via [ n ii ] , [ s ii ] , and asymmetric h@xmath3 profile .
finally , three sources that were not detected with spitzer ( g-48 , g-88 , and g-90 ) are now marginally detected , with g-88 and g-90 displaying broad h@xmath3 emission that suggests they are very embedded class i sources .
their condition as embedded sources is also consistent with their location within one of the strongest submillimeter structures . as listed in table [ laboca - tab ] , several of the detected 870@xmath0 m sources are related to known class ii and embedded sources in the region .
assuming a distance of 170 pc , the 18.6 " beam results in a @xmath23000 au size .
condensations # 2 , 3 , 4 , 5 , and 27 are associated to the previously known sources irs 2 , iras 18595 - 3712 , vv cra , hd 176386b , and s cra , respectively .
in addition , # 3 and 4 also correspond to the millimeter emitting objects 23 and 24 from chini et al .
we combined the pre - existing data from spitzer ( see currie and sicilia - aguilar 2011 for the data on irs 2 , s cra , and hd 176386b and the reduction procedures applied to the rest of the sources ) , akari ( kataza et al . 2010 ) , and millimeter / submillimeter data ( gezari et al . 1999 ; chini et al .
2003 ; nutter et al .
2005 ) , to trace the seds of these objects .
table [ seds - table ] summarizes the available data for the seds shown in figure [ seds - fig ] .
the 870 @xmath0 m flux from irs 2 is in good agreement with previous submillimeter measurements of the source , and consistent with an embedded class i source . for s
cra , our data is also in very good agreement with previous submillimeter observations , and its sed is consistent with a massive protoplanetary disk .
both irs 2 and s cra appear as marginally spatially extended , which suggests the presence of envelopes or surrounding cloud material . in both cases ,
the estimated size of the envelopes would be of the order of @xmath25000 au ( assuming a distance of 170 pc ; knude & hog 1997 ) .
the emission associated to hd 176366b is extended , and appears related to the warm cloud material detected in the spitzer images . to the south of the cra region ,
two condensations , associated to the sources vv cra and iras 18595 - 3712 , are detected .
the emission associated to the binary vv cra is very compact and probably corresponds to the protoplanetary disk of the main source plus that of the disk and envelope of the embedded companion ( ratzka et al .
the submillimeter emission around iras 18598 - 3712 appears slightly extended to the east , suggesting the presence of a larger envelope .
the emission at longer wavelengths for this source is indeed consistent with a black body of t=70 k ( see figure [ seds - fig ] ) . the extinction map reveals substantial cloud material in this direction , which could also produce the extended emission near the source .
the sed of this object suggests the presence of a large amount of material , which could include the source envelope plus contamination from the surrounding cloud . regarding the rest of clumps detected in the laboca maps , they correspond to more complex regions that include several embedded sources and to denser nebular material in the region , which is confirmed by the extinction maps .
this is the case of clump # 1 , which includes the sources t cra , r cra , v710 , and irs 7e / w , among others ( see figure [ contours - fig ] right ) .
the peak of the emission appears to the south - west of irs 7 , which is consistent with the observations of chini et al.(2003 ) and nutter et al .
the embedded x - ray source g-17 is associated to clump # 12 .
although the size of the clump is too large for it to be considered as the same and single object , the presence of this x - ray source near the emission peak of clump # 12 indicates that g-17 is most likely a very embedded low - mass protostar , as it had been suggested by sa08 .
the star cra-465 appears superimposed to the clump # 9 , but the submillimeter emission is most likely due to background cloud material not related to the source , given both the size of the structure and the low extinction of cra-465 ( a@[email protected] ; sa08 ) . the strongest emission in the laboca map is thus associated to the largest concentration of intermediate - mass members , as it would be expected .
the most massive stars appear thus along the two arms ( running in the hd 176386 to vv cra direction , and in the hd 176386 to s cra directions , respectively ) of the molecular cloud , clearly depicted in the extinction maps ( figure [ extinction - fig ] ) .
it is not possible to establish a further correlation between the rest of structures with submillimeter emission and the distribution of cluster members in the region , since so far only the most central part of the cluster has been explored for membership in a systematic way .
the age of the coronet cluster has been subject to intense debate .
while most observations suggested an age of 1 - 3 myr ( meyer & wilking 2009 ) , the presence of a class 0 source and several embedded objects was rather consistent with a younger cluster , @xmath20.5 - 1 myr ( henning et al .
1994 ; chen et al . 1997 ; chini et al . 2003
; nisini et al . 2005 ; nutter et al . 2005 ; groppi et al .
2004 , 2007 ) .
nevertheless , both the disk frequency ( @xmath250 - 65% ) and the disk characteristics among the low - mass , m - type population appear closer to the values observed in older regions than to a 1 myr - old cluster ( sa08 ; lpez - mart et al . 2010 ; currie & sicilia - aguilar 2011 ) . on the other hand , the presence of substantial amounts of nebular material would be inconsistent with an age beyond 3 myr ( hartmann et al . 2001 ; ballesteros - paredes & hartmann 2007 ) .
finally , spectroscopic data also reveal a low accretion fraction , with all the transition and settled / dust depleted disks showing no evidence for accretion , which has been typically related to few - myr old regions ( muzerolle et al .
2010 ) .
few - myr age differences between cluster members are common in many regions , and many clusters include sequential star formation , which results in the presence of more than one star - forming event and thus two or more populations with different ages ( e.g. sicilia - aguilar et al .
2005 , 2006 ) . nevertheless , the coronet cluster is special because of it compactness : the members studied here are located within the densest region of the cra nebula , which is less than 1.3 pc across ( adopting a distance of 170 pc ; knude & hog 1997 ) . a strong correlation between the size of the region and
the age dispersion of the cluster members has been observed in our galaxy and in the magellanic clouds ( elmegreen 2000 ) : the age dispersion of cluster members is of the order of the crossing time in the region , which for the coronet cluster is smaller than 2 myr .
in fact , regions similar to the coronet cluster in size and stellar content ( e.g. the @xmath40 cha cluster ) are usually found to be strongly coeval ( lawson et al . 2009 ) .
we have thus estimated the age of the cluster by placing the known members in a color - magnitude diagram .
a color - magnitude diagram is preferred to a hr diagram since it involves less transformations of the original data that result in increased uncertainties .
the effect of the data uncertainties is also easier to control and determine in a color - magnitude diagram , since it does not involve propagation of non - independent errors .
our knowledge of spectral types allows to correct individually for the extinction , which is particularly important in regions with non - uniform cloud material .
figure [ colormagiso - fig ] displays the near - ir color - magnitude diagram h vs. j - h of the low - mass t tauri members of the coronet cluster , using their 2mass and hawk - i data , together with the isochrones from baraffe et al .
( 1998 ) for the distances of 130 ( only the 1 and 2 myr isochrones ) and 170 pc ( 1 , 2 , 5 , and 10 myr isochrones ) .
these two distances correspond to the estimates given by marraco & rydgren ( 1981 , 130 pc ) and the more recent one by knude & hog ( 1998 , 170 pc ) . given that age and distance
are degenerated , we consider both extreme values in order to ascertain the potential errors in the age determination .
table [ jhk - table ] summarizes the data of the sources used in the hr diagram .
the extinction of the objects has been corrected using their individual values in figure [ colormagiso - fig ] .
in addition to the photometric errors , we also display the uncertainties in the extinction , which are related to the uncertainties in the spectral types of the objects . due to the fact that there is little spread for the lower - mass objects in the j - h direction at these ages ,
it is hard to determine an age for the cluster from these objects . nevertheless , the location of the isochrone `` kink '' for the early m - type stars is more consistent with a distance of 170 pc .
in any case , for either a distance of 130 or 170 pc , and considering the limitations imposed by the small number of objects with early m - type in the cluster , the distribution in the color - magnitude diagram of the early m - type objects suggests that the population has an age in the range 0.5 - 2 myr . for the most plausible distance of 170 pc
, the age would be in the 0.5 - 1 myr range .
this is in agreement with the conclusions of chen et al .
( 1997 ) , based on the bolometric temperature and luminosity of young stellar objects ( yso ) , which stated that the cra region is younger than 2 myr and most likely , younger than taurus , even assuming a distance of 130 pc .
this age estimate is also consistent with the presence of class 0 and very embedded class i sources in the region .
nevertheless , the relatively low disk fraction found among t tauri stars ( 2/3 or lower ; sa08 , lpez - mart et al .
2010 ) is atypical for such a young region .
the `` evolved '' features of the disks ( lack of silicate emission as expected from disks with large grains , presence of disks with inner holes and small dust - depleted disks , seds consistent with strong dust settling ; sa08 , currie & sicilia - aguilar 2011 ) are hard to understand for a 1 myr - old cluster .
some of the objects ( cra-205 , cra-4108 , cra-4109 ) are located at larger distances from the center of the cloud , and may have thus a larger age difference with the main cluster .
nevertheless , the lack of spectral types and extinction ( for cra-205 and cra-4109 ) and the position of the source itself in a place where the isochrones are degenerated ( for cra-4108 ) do not let us test this hypothesis .
although the coronet is younger than ( or has a similar age to ) taurus , its median disk sed for m - type stars is well below the taurus median sed for the same type of objects ( sa08 ) . this may trace back to the formation history of the region and/or the evolution of disks around these low - mass objects in relatively quiescent and sparse regions .
it could also be explained if the cluster were unusually compact ( and actually not sparse ) and interactions would affect the disk structure and evolution .
further observations in a more extended area would be required to check for the presence of other populations .
we present new multiwavelength data for the coronet cluster , including optical spectroscopy obtained with the flames multifiber spectrograph , 870@xmath0 m mapping obtained with the laboca bolometer on the apex telescope , and near - ir imaging from hawk - i .
the optical data allows to derive spectral types , individual extinctions , and accretion status for the remaining x - ray members , completing the work started by sa08 .
the submillimeter data reveals the densest structures of the cloud , as well as the circumstellar disks and envelopes of some intermediate - mass and t tauri members .
the near - ir imaging provides photometry for some of the faintest members and allows to derive extinction maps , which trace the cloud material in low - density regions that are not detected in the submillimeter .
both extinction and submillimeter maps reveal a large cloud where star formation has only occurred along two arms that contain the densest parts of the molecular cloud , resulting in a quite compact cluster . by placing the extinction - corrected photometry of the known low - mass members in a near - ir color - magnitude diagram and comparing with the baraffe et al .
( 1998 ) isochrones , we estimate the age of the cluster to be below 2 myr , and probably within the 0.5 - 1 myr range .
this agrees with previous studies ( suggesting an age @xmath13 myr ; meyer & wilking 2009 ) as well as with the presence of substantial cloud material and highly embedded class i and class 0 objects , and it is also in agreement with the compactness of the region where most members are located .
but a young age is in a strong contrast with the relatively low disk fraction ( @xmath265 - 50% ; sa08 ; lpez - mart et al .
2010 ) and the evidence of disk evolution ( inner holes , lack of silicate features , strong settling , and dust depleted disks ; sa08 ; currie & sicilia - aguilar 2011 ) observed among the lower - mass and m - type cluster members .
further observations will be required to examine the presence of other low - mass members in the region and determine whether the region contains more than one population .
acknowledgments : we want to thank j. rodn and m. schmalzl for their help with the class and clumpfind routines , b. ziegler for his careful assistance in the eso service - mode observations , m. fang for his help with the figures , and the anonymous referee for his / her comments that helped to clarify this paper . a.s .-
a . acknowledges support by the deutsche forschungsgemeinschaft ( dfg ) grant number si-1486 - 1/1 .
this paper contains the results of the eso programs 083.c-0079 and 080.c-3024 .
this work makes use of data products from the two micron all sky survey , which is a joint project of the university of massachusetts and the infrared processing and analysis center / california institute of technology , funded by the national aeronautics and space administration and the national science foundation .
this research has made use of the simbad database , operated at cds , strasbourg , france .
lccccccccccccl cra-4108 & 19020968 - 3646345 & m3.5 ( @xmath341.0 ) & [email protected] & -7.6 & 0.52 & & & -0.2 & 0.31/0.85:/0.54 & 2.07/1.39 & 1.16/1.47 & 1.12/1.49 & na + cra-432 & 19005974 - 3647109 & m4.0 ( @xmath341.0 ) & [email protected] & -46 & & & & -2.8 : & -0.34/-0.59/-0.25 & 2.43/1.89 & 1.12/1.50 & 1.18/1.53 & a + cra-452 & 19004455 - 3702108 & m2.5 ( @xmath340.5 ) & [email protected] & -2.6 & 0.51 & & & & 0.23:/0.83:/0.80 & 1.45/0.82 & 0.83/1.15 & 0.78/1.03 & na + cra-468@xmath41 & 19014936 - 3700285 & m3.5 ( @xmath340.5 ) & [email protected] & -2.8 & 0.53 & & & & 0.5/0.7/0.7 & 1.77/1.07 & 0.99/1.28 & 1.02/1.26 & na + g-1 & 19022708 - 3658132 & m0.5 ( @xmath340.5 ) & [email protected] & -3.8 & 0.58 & & & & 0.4/1.5/1.1 & 1.28/0.62 & 0.62/0.77 & 0.59/0.79 & a + g-32 & 19015833 - 3700267 & & & -4 : & & & & & & & & & na + g-48 & 19:01:49.8 -36:52:43 & & & -9 : & & & & & & & & & uncertain + g-80 & 19:01:35.7 -37:02:32 & & & -275 : & & -25/-83 & -170:/-41 : & & & & & & hh + g-85 & 19013385 - 3657448 & m0.5 ( @xmath341.5 ) & 19@xmath342 & -31 & & & & & -2.6/-2.8/-2.1 & 2.27/1.06 & 0.72/0.98 & 0.76/0.98 & a + g-87 & 19013232 - 3658030 & m1.5 ( @xmath341.5 ) & 16@xmath342 & -3 : & & & & & 0.15/0.52/0.91 & 2.16/2.79 & 0.51/0.89 & /1.56 & na : + g-88 & 19:01:32.2 -36:51:20 & & & -13 & & & & & & & & & a : + g-90 & 19:01:31.4 -36:52:53 & & & -7.4 & & & & & & & & & a : + lcccccccl cra-4107 & 19:02:54.64 & -36:46:19.1 & 285.727661 & -36.771973 & [email protected] & [email protected] & [email protected] & variable ( 0.1 - 0.2 mag ) + cra-452 & 19:00:44.55 & -37:02:10.8 & 285.185638 & -37.036343 & [email protected] & & & variable ( @xmath20.3 mag ) + cra-465 & 19:01:53.74 & -37:00:33.9 & 285.473907 & -37.009418 & [email protected] & [email protected] & [email protected] & + cra-468 & 19:01:49.36 & -37:00:28.5 & 285.455658 & -37.007915 & [email protected] & [email protected] & & also called g-49 + g-14 & 19:02:12.01 & -37:03:09.3 & 285.550049 & -37.052582 & [email protected] & [email protected] & [email protected] & + g-30 & 19:02:00.12 & -37:02:22.0 & 285.500488 & -37.039444 & [email protected] & & & + g-32 & 19:01:58.33 & -37:00:26.7 & 285.493042 & -37.007416 & [email protected] & [email protected] & [email protected] & + g-116 & 19:01:15.40 & -36:52:36.0 & 285.314178 & -36.876667 & [email protected] & [email protected] & [email protected] & prob .
+ g-119 & 19:01:07.90 & -36:51:21.0 & 285.282928 & -36.855831 & [email protected] & [email protected] & [email protected] & prob .
+ lccccl 1 & 19:01:53.9 & -36:58:03 & 96@xmath348 & [email protected] & coa2 , ch10,11,12 , see fig .
[ contours - fig ] + 2 & 19:01:42.5 & -36:58:39 & 6@xmath341 & [email protected] & irs2,ch9 + 3 & 19:02:59.4 & -37:07:45 & 22@xmath347 & [email protected] & ch23 , iras 18595 - 3712 + 4 & 19:03:07.9 & -37:13:04 & 10@xmath342 & [email protected] & ch24 , vv cra + 5 & 19:01:39.5 & -36:53:29 & 35@xmath347 & [email protected] & hd-176386b , ch6,7,8 + 6 & 19:02:18.3 & -37:01:41 & 27@xmath3420 & [email protected] & ch21 + 7 + 10 + 24 & 19:03:06.4 & -37:14:53 & 32@xmath345 & [email protected] & coa4 + 8 + 22 & 19:00:55.4 & -36:55:27 & 29@xmath347 & [email protected] & + 9 + 14 & 19:01:57.0 & -37:01:05 & 20@xmath345 & [email protected] & ch14 + 11 & 19:02:11.4 & -36:56:22 & 15@xmath347 & [email protected] & g-17 + 12 & 19:02:13.7 & -36:57:26 & 7@xmath341 & [email protected] & ch19 + 13 + 25 & 19:02:11.4 & -37:00:47 & 16@xmath342 & [email protected] & ch18 : + 16 & 19:02:24.3 & -36:56:50 & 13@xmath344 & [email protected] & + 17 + 20 + 26 & 19:01:05.3 & -36:54:14 & 35@xmath3413 & [email protected] & ch2,3,4 + 18 + 21 & 19:01:13.7 & -36:54:23 & 27@xmath346 & [email protected] & ch5 + 19 & 19:02:33.5 & -37:02:08 & 13@xmath346 & [email protected] & + 23 & 19:02:18.3 & -37:04:43 & 13@xmath345 & [email protected] & + 27 & 19:01:09.1 & -36:57:26 & 6@xmath342 & [email protected] & s - cra , ch1 + lcclll pc1 &
7030 - 7050 & 7525 - 7550 & m3-m9 & -0.06 + 2.95 x & 1,2 + pc2 & 7540 - 7580 & 7030 - 7050 & m4-m8 & -0.63 + 3.89 x & 1,2 + pc3 & 8235 - 8265 & 7540 - 7580 & m3-m9 & -8.01 + 14.08 x-2.81 x@xmath42 & 2 + pc4 & 9190 - 9225 & 7540 - 7580 & m3-m9 & -0.94 + 4.66 x-0.52 x@xmath42 & 2 + r1 & 8025 - 8130 & 8015 - 8025 & m2.5-m8 & 2.8078
+ 21.085(x-1.044)-53.025(x-1.044)@xmath42 + 60.755(x-1.044)@xmath43 & 3 + r2 & 8415 - 8460 & 8460 - 8470 & m3-m8 & 2.9091 + 10.503(x-1.035)-14.105(x-1.035)@xmath42 + 8.5121(x-1.035)@xmath43 & 3 + r3 & ( 8025 - 8130)+(8415 - 8460 ) & ( 8015 - 8025)+(8460 - 8470 ) & m2.5-m8 & 2.8379 + 19.708(x-1.035)-47.679(x-1.035)@xmath42 + 52.531(x-1.035)@xmath43 & 3 + tio 8465 & 8405 - 8425 & 8455 - 8475 & m3-m8 & 3.2147 + 8.7311(x-1.085)-10.142(x-1.085)@xmath42 + 5.6765(x-1.085)@xmath43 & 3 + vo 2 & 7920 - 7960 & 8130 - 8150 & m3-m8 & 2.6102 - 7.9389(x-0.963)-8.3231(x-0.963)@xmath42 - 14.660(x-0.963)@xmath43 & 3 + vo 7445 & 0.5625(7350 - 7400)+0.4375(7510 - 7560 ) & 7420 - 7470 & m5-m8 & 5.0881 + 17.121(x-0.982)+13.078(x-0.982)@xmath42 & 3 + lcccccl 0.44/b & & & & 4.67 & 0.0335 & simbad + 0.55/v & & & & 4.38 & 0.0656 & simbad + 1.22/j & [email protected] & 0.00007 : & [email protected] & [email protected] & [email protected] & 2mass + 1.63/h & [email protected] & 0.0005 & [email protected] & [email protected] & [email protected] & 2mass + 2.19/k@xmath44 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] & 2mass + 3.6 & & [email protected] & [email protected] & [email protected] & [email protected] & irac / spitzer + 4.5 & & [email protected] & [email protected] & [email protected] & [email protected] & irac / spitzer + 5.8 & & [email protected] & [email protected] & @xmath45 & [email protected] & irac / spitzer + 8.0 & [email protected] & [email protected] & & @xmath45 & [email protected] & irac / spitzer + 9.0 & & & [email protected] & & [email protected] & akari + 12 & 13 : & & 31.9 : & & & iras + 18 & & [email protected] & [email protected] & & [email protected] & akari + 23.9 & & [email protected] & & & & + 25 & 50 : & 3.69 : & 69.1 : & & & iras + 60 & & 38.9 : & 131 : & & & iras + 65 & & & & & [email protected] & akari + 100 & & & 95.2 : & & & iras + 450 & 12@xmath341 & & & & [email protected] & nutter et al.(2005 ) + 850 & [email protected] & & & & [email protected] & nutter et al.(2005 ) + 870 & [email protected] & [email protected] & [email protected] & [email protected]@xmath46 & [email protected] & this work + 870 & & & & & [email protected] & gezari et al . ( 1999 ) + 1200 & [email protected] & [email protected] & & & [email protected] & chini et al . ( 2003 ) + lcccccc cra-4108 & 19020968 - 3646345 & [email protected] & [email protected] & [email protected] & m3.5 & 1.2@xmath34 1.1 + cra-4110 & 19011629 - 3656282 & [email protected] & [email protected] & [email protected] & m5 & 1.3@xmath34 0.4 + cra-4111 & 19012083 - 3703027 & [email protected] & [email protected] & [email protected] & m4.5 & 0.1@xmath34 0.5 + cra-432 & 19005974 - 3647109 & [email protected] & [email protected] & [email protected] & m4.0 & 2.7@xmath34 0.2 + cra-452 & 19004455 - 3702108 & [email protected] & [email protected] & [email protected] & m2.5 & 1.7@xmath34 0.2 + cra-453 & 19010460 - 3701292 & [email protected] & [email protected] & [email protected] & m4.5 & 0.4@xmath34 0.3 + cra-465 & 19015374 - 3700339 & [email protected] & [email protected] & [email protected] & m7.5 & 0.2@xmath34 1.0 + cra-466 & 19011893 - 3658282 & [email protected] & [email protected] & [email protected] & m2.0 & 8.1@xmath34 0.4 + cra-468 & 19014936 - 3700285 & [email protected] & [email protected] & [email protected] & m3.5 & 0.6@xmath34 0.7 + g-1 & 19022708 - 3658132 & [email protected] & [email protected] & [email protected] & m0.5 & 3.4@xmath34 0.2 + g-14 & 19021201 - 3703093 & [email protected] & [email protected] & [email protected] & m4.5 & 1.9@xmath34 0.1 + g-30 &
19020012 - 3702220 & [email protected] & [email protected] & [email protected] & m3.5 & 0.1@xmath34 0.5 + g-65 & 19014041 - 3651422 & [email protected] & [email protected] & [email protected] & m1.5 & 13.8@xmath34 0.5 + g-85 & 19013385 - 3657448 & [email protected] & [email protected] & [email protected] & m0.5 & 19@xmath34 2 + g-87 & 19013232 - 3658030 & [email protected] & [email protected] & [email protected] & m1.5 & 16@xmath34 2 + g-94 & 19012901 - 3701484 & [email protected] & [email protected] & [email protected] & m3.5 & 0.6@xmath34 0.1 + g-95 & 19012872 - 3659317 & [email protected] & [email protected] & [email protected] & m1.0 & [email protected] + g-102 & 19012562 - 3704535 & [email protected] & [email protected] & [email protected] & m5.0 & 0.7@xmath34 0.1 + m map obtained with apex / laboca .
the greyscale represent a logarithm scale between 0.030 - 1 jy ( approximately 3@xmath4rms ) .
the contours are in lineal scale , from 0.030 to 1 jy in steps of 0.050 jy .
the beam is displayed in the lower right corner .
the numbers correspond to the clumps identified by clumpfind 2d ( see also table [ laboca - tab ] ) .
right ascension and declination are in j2000.[laboca - fig ] ] | we present new optical spectroscopy with flames / vlt , near - ir imaging with hawk - i / vlt , and 870@xmath0 m mapping with apex / laboca of the coronet cluster .
the optical data allow to estimate spectral types , extinction and the presence of accretion in 6 more m - type members , in addition to the 12 that we had previously studied .
the submillimeter maps and near - ir data reveal the presence of nebular structures and high extinction regions , which are in some cases associated to known ir , optical , and x - ray sources .
most star formation is associated to two elongated structures crossing in the central part of the cluster . placing all the 18 objects with known spectral types and extinction in the hr diagram
suggests that the cluster is younger than previously thought ( @xmath12 myr , and probably @xmath20.5 - 1 myr ) .
the new age estimate is in agreement with the evolutionary status of the various protostars in the region and with its compactness ( @xmath11.3 pc across ) , but results in a conflict with the low disk and accretion fraction ( only 50 - 65% of low - mass stars appear to have protoplanetary disks , and most transitional and homologously depleted disks are consistent with no accretion ) and with the evolutionary features observed in the mid - ir spectra and spectral energy distributions of the disks . |
the spectral energy distribution ( sed ) of a galaxy contains valuable information about its physical properties , including the stellar , gas and dust content , the age and abundance distribution of the stellar populations resulting from the star formation history ( sfh ) , and their interaction with the interstellar medium ( ism ) .
the study of the sed therefore offers the most direct way to investigate galaxy formation and evolution , both through direct observations and corresponding theoretical modelling .
the different physical processes occurring in galaxies all leave their imprint on the global and detailed shape of the spectrum , each dominating at different wavelengths . therefore , by analysing and predicting the whole spectral range one can hope to deconvolve and interpret all the information contained in the sed in terms of the sfh and galaxy evolution in general .
stellar sources mainly emit in the uv / optical to nir spectral range , and the sed in this wavelength region is therefore heavily influenced by the star formation history of the galaxy .
the sed from a few @xmath2 m to the sub - mm ( the ir region ) is dominated by the interaction of dust grains with stellar radiation .
basically dust absorbs and scatters photons , mostly at wavelength @xmath12 1 @xmath2 m and thermally emit the absorbed energy in the ir .
the resulting sed is often substantially changed and in many relevant cases radically modified by the presence of dust .
a detailed modelling of the entire sed of galaxies is therefore very complex .
several different approaches have been proposed , depending on the purpose of the applications . some works ( e.g. devriendt , guiderdoni & sadat 1999 ; chary & elbaz 2001 ; dale et al .
2001 ; dale & helou 2002 ; galliano et al .
2003 ; lagache , dole & puget 2003 ; burgarella et al .
2005 , da cunha , charlot & elbaz 2008 ) have proposed semi - empirical treatments of the seds .
the aim of these works is in general to interpret large samples of data , requiring fast computing times making use of observationally or physically motivated seds .
there are then other codes , as hyperz ( bolzonella et al .
2000 ) or lephare ( arnouts et al .
1999 ) which usually perform a template - based sed fitting procedure based on a standard @xmath13 minimization method . usually here
dust effects are accounted for by computing the attenuation of the light from a source placed behind a foreground screen of dust and assuming different extinction curves .
other works are based , instead , on theoretical computations in order to have a more general applicability in terms of interpretative and predictive power .
several papers deal with the radiative transfer ( rt ) in spherical geometries , mainly aimed at modelling starburst galaxies ( e.g. rowan - robinson 1980 ; rowan - robinson & crawford 1989 ; efstathiou , rowan - robinson & siebenmorgen 2000 ; popescu et al .
2000 ; efstathiou & rowan - robinson 2003 ; takagi , arimoto & hanami 2003a ; takagi , vansevicius & arimoto 2003b ; siebenmorgen & krugel 2007 ; rowan - robinson 2012 ) .
early models of this kind did not include the evolution of stellar populations .
silva et al . (
1998 ) were the first to couple radiative transfer through a dusty ism and the spectral ( and chemical ) evolution of stellar populations .
modelling the emission from stars and dust consistently in order to get reliable estimates for the main galaxy physical parameters , like stellar mass ( @xmath14 ) , star formation rate ( sfr ) and the average extinction ( @xmath15 ) , involves solving the radiative transfer equation for idealised but realistic geometrical distributions of stars and dust , as well as taking advantage of the full sed coverage from uv to sub - mm .
the full sed allows the total luminosity of the galaxy to be robustly constrained without relying on extrapolations based on optical data alone .
while the shape of the optical - uv sed indicates some level of dust extinction , there may be stellar populations ( typically the newly born ) completely obscured by dust .
the only way to overcome this problem and quantify how much stellar light ( and therefore mass ) is missing from the optical , is to consider the dust - absorbed galaxy luminosity through the dust re - emitted spectrum in the ir , and model it together with the optical spectrum with a complete radiative transfer method . with total infrared luminosities between @xmath16 l@xmath5 and @xmath17 l@xmath5 ,
respectively , luminous and ultraluminous infrared galaxies [ ( u)lirgs hereafter ] are among the most luminous and complex extragalactic objects we can conceive , including all varieties of young and old stellar populations , dust absorption , scattering , grain thermal re - radiation , and agn emission ( lonsdale et al .
although they are quite rare in the local universe , they dominate the cosmic star formation rate and the fir background at z @xmath18 1 .
therefore they are suitable laboratories to study the main physical processes which drive galaxy formation and evolution .
observations of local ulirgs have shown that they are dominated by strong interactions and mergers and the fraction of mergers / interactions among them has been found to be strongly correlated with their ir luminosity , such that lower luminosity lirgs ( l@xmath19 @xmath20 @xmath21 l@xmath5 ) are ordinary disks while the highest luminosity ulirgs are advanced stage mergers .
high - z lirgs and ulirgs seem to be , instead , not equivalents of ulirgs in the nearby universe , but rather upscaled versions of normal galaxies ( elbaz et al .
2011 ; nordon et al . 2012 ; symeonidis et al .
2009 ; rujopakarn et al .
an implication is that high - z ulirgs could well have gradually evolving sfhs , and be not necessarily associated with merger events as they are at z@xmath00 .
this paper is the first of a series dedicated to a detailed physical investigation of the spectro - photometric properties of dusty high - redshift ( u)lirgs in an attempt to achieve a deeper understanding of these sources .
here we concentrate on a small sample of high-@xmath22 ( u)lirgs with the currently richest suite of photometric and spectroscopic data , available for the first time at redshift @xmath11 , combining deep _ herschel _ imaging with ultra - deep irs spectra from _
spitzer_. the data is analysed with a state - of - art chemo - spectro - photometric model ( grasil : silva et al.1998 ) including self - consistent treatment of dust absorption and reprocessing based on a full radiative transfer solution . in section
[ data ] we describe our data sample .
the physical modeling performed for this study is presented in section [ grasilandfadda ] .
the results concerning our sfh and sfr estimates are discussed in section [ sfrandsfh ] . while the predicted stellar masses and average extinctions for these high - z ( u)lirgs are presented in section [ massdet ] .
the conclusions can be found in section [ conclusions ] .
our cosmological model assumes h@xmath23=71 km @xmath24 mpc@xmath25 , @xmath26=0.73 , @xmath27=0.27 .
our high-@xmath22 ( u)lirgs have been selected from the sample of 48 ir - luminous galaxies in goods - s presented by fadda et al .
( 2010 , f10 hereafter ) .
it includes the faintest @xmath1 @xmath2 m sources observed with irs ( @xmath28 mjy ) in the two redshift bins ( @xmath29 and @xmath30 ) and samples the major contributors to the cosmic infrared background at the most active epochs . at these redshifts , the @xmath31
m fluxes translate to infrared luminosities roughly in the ranges for lirgs at @xmath10 and ulirgs at @xmath11 .
this sample is therefore crudely luminosity selected ( with the @xmath32 limits corresponding to different luminosity ranges for the z@xmath33 lirgs and z@xmath34 ulirgs ) and f10 did not apply any other selections .
these objects have been selected by f10 mostly using the photometric redshifts computed by caputi et al .
( 2006 ) since only 50% and 10% of them had spectroscopic - z in the z@xmath0 1 and 2 bins , respectively .
as emphasized by f10 , except for having higher dust obscuration , these galaxies do not have extremely deviant properties in the rest - frame uv / optical compared to galaxies selected at observed optical / near - ir band .
their observed optical / near - ir colors are very similar to those of extremely red galaxy ( erg ) populations selected by large area k - band surveys .
in this paper we concentrate on galaxies powered by star formation .
therefore we excluded from our present analysis objects previously classified by f10 as agn - dominated on the basis of several indicators such as broad and high ionization lines in optical spectra , lack of a @xmath35 m stellar bump in the sed , x - ray bright sources , low mid - ir @xmath36 m ew and optical morphology ( see also pozzi et al .
we also excluded the sources falling outside the multiwavelength music catalogue ( santini et al .
2009 ) , which is our reference for complementary optical / near - ir photometry . among the 31 ( u)lirgs fulfilling these requirements , 10 are at @xmath10 ( all in the nominal lirg regime ) and 21 , mostly ulirgs , are at @xmath11 .
the goods - s field was observed in 2010 with both _
( pilbrat et al . 2010 ) spire ( griffin et al . 2010 ) and pacs ( poglitsch et al .
2010 ) , giving a total of six bands from @xmath37 to @xmath38 @xmath2 m .
spire and pacs data are taken , respectively , from the herschel multi - tiered extragalactic survey ( hermes : oliver et al .
2012 ) and the pacs evolutionary probe ( pep : lutz et al .
2011 ) programs .
typical noise levels are @xmath39 mjy for pacs
@xmath40 @xmath2 m and @xmath41 mjy for spire @xmath42 @xmath2 m , including confusion .
[ cols="<,<,<,<,<",options="header " , ] higher sfr@xmath43/sfr@xmath44 values are dominated by mc emission .
an illustration of this is reported in fig .
[ outliers ] for the two pacs - selected sources of r11 : the one dominated by mc corresponds to the starred object in the upper figure , while that one dominated by cirrus emission falls in the lower part and shows minimal mc contribution .
all these aspects will be further discussed in a forthcoming paper extending the sample to all the outliers analysed in r11 and all the local ( u)lirgs of vega et al .
( 2008 ) . a bimodality in the star formation efficiency ( sfe = sfr / m@xmath45 ) and kennicutt - schmidt relation ( @xmath46 ) of local and high-@xmath22 mergers and non- or weakly - interacting star forming galaxies based on co measurements , has been recently suggested by genzel et al . (
2010 ) and daddi et al .
( 2010 ) . our upper and lower envelopes in fig .
3 , within which all our sources are contained , may correspond to these two sequences .
further investigation with larger and unbiased samples of galaxies will help in this sense .
the latter would be explained by the fact that , while locally galaxies with ir luminosities exceeding @xmath47 l@xmath5 are predominantly associated with merging events ( e.g. dasyra et al .
2006 ) , at high - z these high - ir luminosities may be achieved mostly via cold gas accretion ( powell et al . 2011 ) .
our study confirms that high - z ( u)lirgs are more likely up - scaled versions of normal galaxies rather than equivalents of local ( u)lirgs in terms of the mode of sf ( e.g. symeonidis et al . 2009 , rujopakarn et al . 2011 ) . with our current assumptions ( see [ grasilandfadda ] ) , all our mid - ir selected galaxies appear to include massive populations of old ( @xmath3 gyr ) stars and , at the same time , to host a moderate ongoing activity of sf ( sfr@xmath43 @xmath4 100 m@xmath5/yr , cf .
[ sfrchih ] ( right ) ) . the bulk of stars appear to have been formed a few gyr before the observation in essentially all cases .
average estimates can be inferred from fig [ sfrtm ] : for the z@xmath0 2 ( u)lirgs having the sfh shown in the central panel of fig . [ sfrtm ] , about 66 - 80% of the stellar mass has formed after @xmath0 1 gyr from the beginning of their star formation activity , proportionally higher at higher sf efficiency . for the z@xmath02 ( u)lirgs presenting , instead , more regular sfh , as those shown in the bottom panel of fig .
[ sfrchih ] , about 30 - 43% of stellar mass is formed after 1 gyr . finally for the lirgs , on average , @xmath0 28% of the stellar mass has already been formed after the first 1 gyr .
another key physical parameter of high - redshift galaxies is the stellar mass m@xmath48 .
our best - fit estimates appear in table [ tabulirg ] where we compare our best - fit m@xmath48 to those derived by f10 .
they fit bc03 models to the optical - to-8 @xmath2 m sed with the hyperz code ( bolzonella et al .
2000 ) allowing different sfhs , from a constant star formation to an exponentially declining sf with different e - folding times ( @xmath49-model ) , and assuming the calzetti et al .
( 2000 ) attenuation law to account for dust extinction effects .
the so called @xmath49-model is essentially an analytical approximation of the schmidt s law @xmath50 , with @xmath49=1/@xmath51 . as shown in tab .
[ tabulirg ] , grasil m@xmath48 are systematically larger than those based on hyperz .
this discrepancy is higher for the ulirgs at @xmath11 , for which it reaches a factor of @xmath0 @xmath52 in m@xmath48 ( median factor 2.5 ) , while it is lower for the @xmath10 lirgs whose median factor is @xmath53 .
michalowski et al .
( 2012 , m12 hereafter ) found the adopted specific sfh ( double component versus @xmath49-model ) to be the major factor affecting the derived stellar masses of high-@xmath22 and highly star forming smgs , accounting for a factor of @xmath0 2.5 in m@xmath48 .
we have investigated the possible origin of the larger differences in our m@xmath48 estimates by considering different models for the sfh , continuous vs starburst . in both cases we found similar m@xmath48 discrepancies with respect to the hyperz ones . as discussed in detail in m12
, the different evoutionary models may also contribute to a factor 2.5 difference in m@xmath48 estimates when a @xmath49-model is considered .
we tested this hypothesis by modeling our ( u)lirgs with similar sfhs ( tau - models ) , age , metallicity and @xmath54 plus calzetti attenuation law and by fitting the optical data alone as in the hyperz code . by adopting similar prescriptions we found our m@xmath48 to be in full agreement , ( i.e. within @xmath55 ) with the hyperz estimates .
taking into account the differences in the specific parameters ruling the sfh ( e.g. @xmath49 ) and on the ssp libraries used ( bc03 and grasil ssps rely on the same padova isochrones but the latter includes dusty envelopes around agb stars ) we can conclude that the different sfh ( i.e. different @xmath49 ) and ssps can not account for the larger discrepancies , we find among the more obscured ulirgs at @xmath11 ; for these the dominant factor is the dust exinction .
of course the effect of the different sfh on the stellar mass estimates becomes larger when comparing @xmath49-models with sfh characterized by recent ( last 50 myr ) burst of star formation , as shown by m12 .
all these galaxies are , however , normal star forming galaxies and the only five objects requiring the burst on top of the schimdt - type sf are all observed as late - sbs . using the grasil best - fits
, we estimate the average total extinction in the rest - frame @xmath56-band , @xmath54 , from the ratio of the extinguished ( @xmath57 ) to the unextinguished starlight ( @xmath58 ) : @xmath59 the average a@xmath60 values , reported in table [ tabulirg ] , are systematically higher for the @xmath11 than for the @xmath10 objects .
these differences are traceable to the different evolutionary stages at which the two classes of sources are observed .
our best - fit a@xmath60 are also systematically higher than those obtained with hyperz using the calzetti approximation in combination with the foreground screen of dust , expecially for the @xmath11 ulirgs for which the median a@xmath61 is @xmath62 and @xmath63 for the two cases , respectively . as shown in fig .
[ extinction ] ( top - left ) we find , for our ( u)lirgs , a tight correlation ( @xmath64 for @xmath11 ( u)lirgs and @xmath65 for @xmath10 lirgs ) between the difference in the m@xmath48 estimates based on grasil and hyperz solutions and the average extinction a@xmath60 , with the difference increasing as a function of the extinction a@xmath61 .
this clarifies why for ulirgs , with higher a@xmath61 on average , we see the largest stellar mass discrepancies .
the stellar mass missed by fitting optical data alone is hidden in dust . in concurrence with our results , wuyts et al .
( 2009 ) also found that the mass underestimate is more severe during the dusty , peak sf phase .
it is worth noting that the availability of the full wavelength coverage also plays a crucial role in constraining the models .
in fact the first basic constraint that is provided by the availability of uv - to - sub - mm data is the total l@xmath66 of the galaxy , which otherwise should be guessed .
this is easier when the optical contains most of the total energy .
the shape of the optical - uv sed indicates some level of dust , but there may be stellar populations totally obscured by dust ( typically the newly born ) , and even less young stellar populations may be partly extinguished . without a constraint on the optical - uv extinction , i.e. on
how much is re - emitted in the ir , it is difficult to quantify how much stellar light ( therefore mass ) is missing from the optical .
two examples are given in fig .
[ balance ] which compares the total l@xmath66 inferred from the unextinguished starlight component computed with hyperz ( in red ) to that one computed with grasil ( magenta long - dashed line ) and to the total l@xmath19 inferred from the grasil best - fits .
the two objects shown in this figure represent two extreme cases , the one on the left corresponding to a large discrepancy in stellar mass ( @xmath0 a factor of 4 ) and the other one on the right characterized by the lowest stellar mass discrepancy ( @xmath0 a factor of 1.17 ) . of course , due to energy balance , the energy of the pure stellar sed must be equal to that of the processed sed which fit the data .
therefore , the ir part of the sed is a further fundamental constraint when it contains an important fraction of the energy emitted by stellar populations .
figure [ extinction ] ( bottom ) shows that our stellar and dust masses are consistent with published estimates for high-@xmath22 smgs at comparable redshifts .
compared to local spirals and local ( u)lirgs our @xmath11 ( u)lirgs have higher dust - to - stellar mass ratios , by a factor of @xmath0 30 and 6 , respectively .
similar results were found by santini et al .
2010 for the high-@xmath22 smgs .
the m@xmath67/m@xmath48 of z@xmath33 lirgs are , instead , similar to those of local ( u)lirgs .
m@xmath48 and m@xmath67 of galaxies in this work appear also to be in agreement with the predictions of the calura et al . ( 2008 ) chemical evolution models for proto - ellipticals .
the distribution of our sources in this plot is also similar to that found by rowan - robinson et al .
2010 for the hermes lockman sample .
these specific aspects will be dealt with in detail in one of the forthcoming papers of the series .
the top - right panel of fig .
[ extinction ] illustrates the possible degeneracy in the model solutions , between the average extinction and stellar mass , color - coded by the value of @xmath13 , for a typical @xmath11 ulirg .
as we see , among the many solutions considered , acceptable best - fits , within @xmath68 + 30 , are clearly identified in the parameter space and not much degeneracy is apparent . in particular a@xmath60 and m@xmath48
seem to be well constrained within @xmath69 @xmath70 and @xmath71 @xmath72 dex .
these values refer to the case of fixed sfh , that is to say fixed @xmath73 and @xmath51 .
if we consider in addition to the many combinations of grasil parameters also the different combinations of parameters ruling the sfh , the scatter around the best - fit solution increases to @xmath69 @xmath74 and @xmath71 @xmath75 dex , well within the typical uncertainties for this kind of analysis .
this is likely to be a result of the continuous spectral coverage from uv to sub - mm for our sample .
another likely factor is the self - consistent radiative transfer modelling , accounting for the distribution of dust with respect to young and old stars , from which both the age - dependent dust attenuation and dust reprocessing are constrained .
the present work reports on a selection of results from a physical self - consistent study of ir luminous galaxies at @xmath76 and @xmath77 . for this study we combined far - infrared photometry from two related deep surveys , pep and hermes , carried out with the _ herschel space observatory _ , with deep photometry in the near - ir and mid - ir with the _ spitzer _ irac and mips imagers , as well as complete mid - infrared spectroscopic follow - up with the _ spitzer _ s infrared spectrograph irs .
our working sample includes 31 luminous and ultraluminous infrared galaxies selected in goods - s with @xmath1 @xmath2 m fluxes between 0.2 and 0.5 mjy . among them
10 are at z@xmath01 and 21 are at z@xmath02 .
they make the faintest high - redshift galaxies observed spectroscopically with irs , with an excellent suite of photometric data in all bands from uv to the millimeter , including our key far - ir and sub - mm photometry from _
herschel_. these data have prompted us to investigate the nature and the main physical properties - like stellar mass , bolometric luminosity , star - formation history , extinction , as well as the mass assembly history - for every object .
the novelty of our approach consists in modeling the data with a self - consistent physical code ( grasil ) , exploiting a state - of - the - art treatment of dust extinction and reprocessing .
we find that all of our galaxies require massive populations of old ( @xmath3 gyr ) stars and , at the same time , host a moderate ongoing activity of sf ( sfr@xmath78 100 m@xmath5/yr with sfr@xmath43 being the sfr averaged over the last 10 myr ) .
our detailed analysis appears to give important hints also on the past history of sf with the bulk of stars consistent with having been formed a few gyr before the observation in essentially all cases .
only five galaxies of the sample seem to require , on the basis of an application of the @xmath79-test to our @xmath13 analysis , a recent starburst superimposed on a low - level , secularly evolving star formation history . even in these five objects , however , the gas mass involved in the sb amounts to only a small fraction of the galactic mass , @xmath20 4% , and all of them are observed just at the end of the burst event ( i.e. late sbs ) .
we find substantial discrepancies between our results and those based on optical - only sed fitting for the same objects .
our physically consistent best - fit model solutions of the observed seds indicate higher extinctions ( by @xmath6a@xmath7 0.81 and 1.14 ) and higher stellar masses ( by @xmath6log(m@xmath8 0.16 and 0.36 dex ) for z@xmath01 and z@xmath02 ( u)lirgs , respectively . the stellar mass difference is larger for the most obscured objects and correlates with the total dust extinction .
this is in agreement with the results of wuyts et al .
( 2009 ) who also found that the mass underestimate is more severe during the dusty , peak sf phase .
we also find lower sfrs on average than those computed from the total @xmath9 using the kennicutt ( 1998 ) calibration due to the significant contribution to the dust heating by intermediate - age stellar populations through ` cirrus ' emission : only @xmath027% and @xmath034% of the total @xmath9 appears to be due to ongoing sf in mcs for @xmath10 and @xmath11 ( u)lirgs , respectively .
we have demonstrated with the present study the importance of a self - consistent treatment of dust extinction and reprocessing effects in luminous star - forming galaxies via radiative transfer modelling . in future works
many aspects will be further investigated , such as a more complete understanding of the degeneracies in our model solutions , including exploring the effects of also other imfs .
this is of particular relevance for a further refinement of the galaxy star - formation histories and the estimates of stellar mass functions and sfr functions , that have a significant impact on the studies of galaxy formation and evolution .
we acknowledge support from asi ( herschel science contract i/005/07/0 ) .
pacs has been developed by a consortium of institutes led by mpe ( germany ) and including uvie ( austria ) ; ku leuven , csl , imec ( belgium ) ; cea , lam(france);mpia ( germany ) ; inaf - ifsi / oaa / oap / oat , lens , sissa ( italy ) ; iac ( spain ) .
this development has been supported by the funding agencies bmvit ( austria ) ,
esa - prodex ( belgium ) , cea / cnes ( france ) , dlr ( germany ) , asi / inaf ( italy ) , and cicyt / mcyt ( spain ) .
spire has been developed by a consortium of institutes led by cardiff univ .
( uk ) and including : univ .
lethbridge ( canada ) ; naoc ( china ) ; cea , lam ( france ) ; ifsi , univ .
padua ( italy ) ; iac ( spain ) ; stockholm observatory ( sweden ) ; imperial college london , ral , ucl - mssl , ukatc , univ .
sussex ( uk ) ; and caltech , jpl , nhsc , univ .
colorado ( usa ) .
this development has been supported by national funding agencies : csa ( canada ) ; naoc ( china ) ; cea , cnes , cnrs ( france ) ; asi ( italy ) ; mcinn ( spain ) ; snsb ( sweden ) ; stfc , uksa ( uk ) ; and nasa ( usa ) .
arnouts , s. , cristiani , s. , moscardini , l. , et al . 1999 , , 310 , 540 bertelli , g. , bressan , a. , chiosi , c. , fagotto , f. , & nasi , e. 1994 , , 106 , 275 bolzonella , m. , miralles , j .-
, & pell , r. 2000 , , 363 , 476 bressan , a. , granato , g. l. , & silva , l. 1998 , , 332 , 135 bruzual , g. , & charlot , s. 2003 , , 344 , 1000 burgarella , d. , buat , v. , & iglesias - pramo , j. 2005 , , 360 , 1413 calura , f. , pipino , a. , & matteucci , f. 2008 , , 479 , 669 calzetti , d. , armus , l. , bohlin , r. c. , et al . 2000 , , 533 , 682 da cunha , e. , charlot , s. , & elbaz , d. 2008 , , 388 , 1595 chary , r. , & elbaz , d. 2001 , , 556 , 562 da cunha , e. , charmandaris , v. , daz - santos , t. , et al .
2010 , , 523 , a78 daddi , e. , dickinson , m. , morrison , g. , et al .
2007 , , 670 , 156 daddi , e. , elbaz , d. , walter , f. , et al . 2010 , , 714 , l118 dale , d. a. , helou , g. , contursi , a. , silbermann , n. a. , & kolhatkar , s. 2001 , , 549 , 215 dale , d. a. , & helou , g. 2002 , , 576 , 159 dasyra , k. m. , tacconi , l. j. , davies , r. i. , et al . 2006 , , 651 , 835 devriendt , j. e. g. , guiderdoni , b. , & sadat , r. 1999 , , 350 , 381 efstathiou , a. , rowan - robinson , m. , & siebenmorgen , r. 2000 , , 313 , 734 efstathiou , a. , & rowan - robinson , m. 2003 , , 343 , 322 elbaz , d. , hwang , h. s. , magnelli , b. , et al .
2010 , , 518 , l29 elbaz , d. , dickinson , m. , hwang , h. s. , et al .
2011 , , 533 , a119 fadda , d. , yan , l. , lagache , g. , et al .
2010 , , 719 , 425 galliano , f. , madden , s. c. , jones , a. p. , et al .
2003 , , 407 , 159 genzel , r. , tacconi , l. j. , gracia - carpio , j. , et al . 2010 , , 407 , 2091 griffin , m. j. , abergel , a. , abreu , a. , et al .
2010 , , 518 , l3 kartaltepe , j. s. , sanders , d. b. , scoville , n. z. , et al .
2007 , , 172 , 320 kennicutt , r. c. , jr .
1998 , , 498 , 541 lagache , g. , dole , h. , & puget , j .- l . 2003 , , 338 , 555 leitherer , c. , & heckman , t. m. 1995 , , 96 , 9 lo faro , b. , et al . 2012 , in preparation lonsdale ,
c. j. , farrah , d. , & smith , h. e. 2006 , astrophysics update 2 , 285 lutz , d. , poglitsch , a. , altieri , b. , et al .
2011 , , 532 , a90 magdis , g. e. , elbaz , d. , hwang , h. s. , pep team , & hermes team 2011 , galaxy evolution : infrared to millimeter wavelength perspective , 446 , 221 maraston , c. , pforr , j. , renzini , a. , et al .
2010 , , 407 , 830 michaowski , m. j. , dunlop , j. s. , cirasuolo , m. , et al .
2012 , , 541 , a85 nordon , r. , lutz , d. , shao , l. , et al .
2010 , , 518 , l24 nordon , r. , lutz , d. , genzel , r. , et al .
2012 , , 745 , 182 oliver , s. j. , bock , j. , altieri , b. , et al .
2012 , , 3269 pilbratt , g.l . , riedinger , j.r . , passvogel , t. et al . 2010 , , 518 , l1 poglitsch , a. , waelkens , c. , geis , n. , et al .
2010 , , 518 , l2 popescu , c. c. , misiriotis , a. , kylafis , n. d. , tuffs , r. j. , & fischera , j. 2000 , , 362 , 138 powell , l. c. , bournaud , f. , chapon , d. , et al .
2011 , iau symposium , 277 , 234 pozzi , f. , vignali , c. , gruppioni , c. , et al .
2012 , arxiv:1204.1152 rocca - volmerange , b. 1989 , , 236 , 47 rodighiero , g. , cimatti , a. , gruppioni , c. , et al .
2010 , , 518 , l25 rodighiero , g. , daddi , e. , baronchelli , i. , et al .
2011 , , 739 , l40 rowan - robinson , m. 1980 , , 44 , 403 rowan - robinson , m. , & crawford , j. 1989 , , 238 , 523 rowan - robinson , m. , roseboom , i. g. , vaccari , m. , et al . 2010 , , 409 , 2 rowan - robinson , m. 2012 , arxiv:1201.1559 rujopakarn , w. , rieke , g. h. , eisenstein , d. j. , & juneau , s. 2011 , , 726 , 93 santini , p. , fontana , a. , grazian , a. , et al .
2009 , vizier online data catalog , 350 , 40751 santini , p. , maiolino , r. , magnelli , b. , et al .
2010 , , 518 , l154 silva , l. , granato , g. l. , bressan , a. , & danese , l. 1998 , , 509 , 103 silva 1999 , phd thesis , sissa siebenmorgen , r. , & krgel , e. 2007 , , 461 , 445 silva , l. , schurer , a. , granato , g. l. , et al .
2011 , , 410 , 2043 symeonidis , m. , willner , s. p. , rigopoulou , d. , et al .
2008 , , 385 , 1015 symeonidis , m. , page , m. j. , seymour , n. , et al .
2009 , , 397 , 1728 takagi , t. , arimoto , n. , & hanami , h. 2003 , , 340 , 813 takagi , t. , vansevicius , v. , & arimoto , n. 2003 , , 55 , 385 vega , o. , clemens , m. s. , bressan , a. , et al .
2008 , , 484 , 631 wuyts , s. , franx , m. , cox , t. j. , et al .
2009 , , 696 , 348 wuyts , s. , frster schreiber , n. m. , lutz , d. , et al .
2011 , , 738 , 106 | we combine far - infrared photometry from _ herschel _ ( pep / hermes ) with deep mid - infrared spectroscopy from _
spitzer _ to investigate the nature and the mass assembly history of a sample of 31 luminous and ultraluminous infrared galaxies at z@xmath01 and 2 selected in goods - s with @xmath1 @xmath2 m fluxes between 0.2 and 0.5 mjy .
we model the data with a self - consistent physical model ( grasil ) which includes a state - of - the - art treatment of dust extinction and reprocessing .
we find that all of our galaxies appear to require massive populations of old ( @xmath3 gyr ) stars and , at the same time , to host a moderate ongoing activity of sf ( sfr @xmath4 100 m@xmath5/yr ) .
the bulk of the stars appear to have been formed a few gyr before the observation in essentially all cases .
only five galaxies of the sample require a recent starburst superimposed on a quiescent star formation history ( sfh ) .
we also find discrepancies between our results and those based on optical - only sed fitting for the same objects ; by fitting their observed spectral energy distributions with our physical model we find higher extinctions ( by @xmath6a@xmath7 0.81 and 1.14 ) and higher stellar masses ( by @xmath6log(m@xmath8 0.16 and 0.36 dex ) for z@xmath01 and z@xmath02 ( u)lirgs , respectively . the stellar mass difference is larger for the most dust obscured objects .
we also find lower sfrs than those computed from @xmath9 using the kennicutt relation due to the significant contribution to the dust heating by intermediate - age stellar populations through ` cirrus ' emission ( @xmath073% and @xmath066% of total @xmath9 for @xmath10 and @xmath11 ( u)lirgs , respectively ) . |
statistical thermal model has been very successful in describing the various particle ratios measured in heavy ion collisions , from the sps energies up to the highest rhic energies .
it is impressive that with a single chemical freeze - out temperature these models can describe all the different particle ratios up to the heaviest strange hyperons .
thermal parameters obtained from the fits to the various data sets seems to be extremely well behaved in a wide range of energy and different collisions systems and parameters such as temperature and baryo - chemical potential has been well parameterized @xcite as a function of energy . in many ways ,
this remarkable success of describing the data with a few simple statistical parameters can be viewed as an indication that indeed the particles in these collisions were produced in a chemically equilibrated system . however , if this is the case , there must be a critical system size or collision energy where the statistical thermal model fails to describe the observed data , where equilibration is not achieved due to the insufficient number of particle interactions . in this context
, it is important to perform a careful and systematic scan of the thermal model fits considering systems with different size and collision energies .
this is now possible with the large amount of data accumulated at rhic , and in particular , the high quality and high statistic data obtained by the star experiment with its large acceptance detectors and extended particle identification capability .
a detailed description of the star detector system and the different analysis techniques used to obtain the yields can be found elsewhere @xcite .
we present a study of the system size dependence of the hadronic freeze - out parameters by comparing the results from different event centrality classes of au+au and cu+cu collisions at @xmath0 gev measured by the star experiment . due to the fact that we are using data from the same detector system many of the systematic errors
were canceled and the relative error of the particle ratios used in the fit were reduced . for the statistical thermal fit , we used the thermus code from cleymans et al .
@xcite , that allows for fits to the data using a grand - canonical ensemble approach , with the inclusion of a strangeness saturation parameter @xmath1 and also a semi - canonical approach where the strangeness production was treated canonically while the light quark particles were still considered using a gand - canonical approach . in this study , we have used particle ratios that include the pions , protons , kaons , @xmath2 , @xmath3 , @xmath4 @xcite , the corresponding anti - particles and also the @xmath5 meson @xcite .
the pion and proton yields were corrected for the feed - down from the weak decays before the thermal fit .
variations in the thermal fit parameters due to uncertainties in the feed - down corrections were tested and found to add an extra uncertainty of less than @xmath6 in the final uncertainties of these parameters .
particle ratios from different event centrality classes were used to study the dependence of chemical freeze - out conditions with the system size .
figure [ label2 ] shows the results of the temperature and strangeness saturation parameter @xmath1 as a function of the mean number of participants . for comparison ,
data from p+p collisions was also fitted and included in these plots .
temperature , baryo - chemical potential ( @xmath7 ) and strangeness - chemical potential ( @xmath8 ) show little variation with the system size .
even the p+p data yields reasonable fit with the thermal model with a temperature equivalent to the au+au data , of approximately 160 mev .
however , the @xmath1 parameter shows a strong deviation from unity for smaller systems including peripheral au+au and cu+cu events and p+p .
this may be an indication that in these smaller systems , strangeness production is no longer well described with a simple grand - canonical approach .
3.1 in 3.1 in to evaluate the effect of a limited system size in the strange particle production , we used a semi - canonical ensemble approach , where only the conservation of the strange sector was imposed within a canonical volume and the light quarks were considered still in the grand - canonical approach .
figure [ label3]a shows the relative production of strange particles @xmath9 , @xmath2 , @xmath3 , @xmath4 and the @xmath5 with respect to the pions yields , as a function of the radius of the canonical volume obtained using the thermus code . the temperature and baryo - chemical potential used to calculate these curves were fixed to be the same values obtained from the fit to the data .
the solid points show the equivalent experimental particle ratios , for central au+au , peripheral au+au and p+p collisions .
as expected , the @xmath5 meson with its hidden strangeness shows no variation with the canonical volume , and the @xmath4 shows the highest sensitivity to the canonical volume .
the most peripheral data in au+au seems to yield results already consistent with values that are no longer subject to the reduction of strangeness production due to this canonical effect .
the thermal fit results of the temperature , @xmath7 , @xmath8 and @xmath1 seems to show no difference between the au+au and cu+cu data .
however , the relative yields per participant of strange particles and also of pions , measured in cu+cu collisions was observed to be higher than the equivalent yields measured in au+au collisions for the same number of @xmath10 @xcite .
this apparent discrepancy seems not to be reflected in the particle ratios , thus , no difference in the thermal parameters were observed between the results from cu+cu and au+au data .
the volume of the system at chemical freeze - out was calculated using the thermal parameters obtained from the fits to the particle ratios and also using the absolute yields of pions .
figure [ label3]b shows the relative volume per participant of the system for au+au and cu+cu data , normalized by the results of p+p . in this plot , we can observe that the cu+cu data yields a higher freeze - out volume than the equivalent au+au data .
3.1 in 3.1 in
most of the results of statistical thermal fits applied to data from rhic experiments correspond to particle ratios of @xmath11 at mid - rapidity , and not the integrated yield of particles in all rapidity range . to study the variation of the thermal fit parameters in the different rapidity range , we have used two different parameterizations of the rapidity distribution for the different particles to extrapolate the particle yields at the forward regions and
obtain the thermal fit parameters . in a first attempt
, we used a gaussian shape to describe the various particle distributions , adjusting the yield and width of the distribution using the available data from star @xcite and brahms @xcite experiment .
particle ratios were built by extracting the corresponding @xmath11 for the different rapidity regions and also using the total integrated yield . in a second approach ,
we have used the well known hijing monte carlo simulation code @xcite to generate the particle rapidity distributions .
the yields of the rapidity distributions generated by hijing were scaled to fit the data points from star at mid rapidity . once normalized , the shape of the rapidity distributions from hijing seems to describe well the available data from the brahms experiment in the forward rapidity region . in these fits ,
only particle ratios using protons , pions and kaons , and its corresponding antiparticles were used .
we noted that the main effect caused by the reduced number of particles in the thermal fit is the increase of the relative uncertainties in the final parameters and also , a reduction of the @xmath1 parameter , from 1.0 to approximately 0.8 .
figure [ label4]a shows an example of the rapidity distribution obtained using the gaussian extrapolation and the hijing shape for the ratio between anti - protons and pions . in this example , it is clear that there are some considerable difference between the gaussian parameterizations and the hijing prediction .
the solid star symbol corresponds to the star mid - rapidity measurement and the solid circles correspond to the data from the brahms experiment .
figure [ label4]b shows the temperature obtained from the thermal fit to this extrapolated data , and also the result of the fit to the star and brahms data .
both data , and the gaussian parameterization show a constant temperature up to higher values of rapidity .
the result from the thermal fit to the hijing parameterization show an increase of the temperature for the forward region with values higher than 2 units of rapidity . 3.1 in 3.1 in figure [ label5]a and [ label5]b show the result of the baryo - chemical potential @xmath7 and strangeness saturation parameter @xmath1 as a function of rapidity .
the baryo - chemical potential shows an increase with rapidity reflecting the increase of the net baryon density of protons over antiprotons .
the @xmath1 parameter seems to show a constant value up to higher values of rapidity and when using the data and the gaussian parameterization , but the results obtained using the hijing extrapolations show a strong decrease of the strangeness equilibration in the forward direction . despite the variations in the different parameterizations ,
all results obtained here seems to show that in the rhic data , thermal characteristics of the chemical freeze - out conditions are quite constant up to approximately two units of rapidity . 3.1 in 3.1 in
results from a systematic study using rhic data for different centrality classes and different rapidity regions show that chemical freeze - out parameters can fit well the observed particle ratios .
the thermal parameter such as temperature and baryo - chemical potential show negligible variation with the system size , and for event classes with @xmath12 the data is well described using a gand - canonical approach with the strangeness saturation parameter @xmath1 consistent with unity .
smaller systems , such as cu+cu and p+p , can still be well described but with a lower value of @xmath1 . using a strangeness canonical ensemble approach instead of the grand - canonical ensemble
, we were also able to fit the data .
the size of the canonical correlation volume for the au+au data is already consistent with a completely saturated strangeness production , where the values of the strange particle production no longer varies with the correlation volume .
fit to the p+p data resulted in a much smaller correlation volume , where canonical suppression is still a non - negligible effect into the production of strange particles .
we also showed a first attempt to parameterize the rapidity dependence of the thermal fits to the rhic data , where we observed that the temperature and the @xmath1 parameter is constant up to higher values of rapidity , while the baryo - chemical potential shows an increase with rapidity .
we wish to thank fundao de amparo a pesquisa do estado de so paulo , fapesp , brazil for the support to participate in the sqm2008 conference . | we present the results of a systematic test applying statistical thermal model fits in a consistent way for different particle ratios , and different system sizes using the various particle yields measured in the star experiment .
comparison between central and peripheral au+au and cu+cu collisions with data from p+p collisions provides an interesting tool to verify the dependence with the system size .
we also present a study of the rapidity dependence of the thermal fit parameters using available data from rhic in the forward rapidity regions and also using different parameterization for the rapidity distribution of different particles . |
coronal mass ejections ( cmes ) are a large - scale solar activity , releasing a vast amount of plasma and magnetic field into the interplanetary space and may cause severe disturbances of the space environment @xcite . in the classical eruptive flare model @xcite , magnetic reconnection is regarded as the fundamental energy releasing mechanism . in this kind of models ,
the originally anchored magnetic field lines are stretched upward by an erupting structure , and the field lines underneath are pushed toward the center forming a current sheet ( cs ) . the reconnection is induced in the cs , creating new closed field lines ( as seen in a 2-d projection ) at both tips of the cs .
magnetic reconnection is also believed to produce energetic particles .
these particles stream down along the newly reconnected field lines and produce the well observed flare ribbons and post - flare loops , as well as enhancing soft x - ray and hard x - ray emissions @xcite . above the reconnection region , poloidal magnetic fluxes are injected into the erupting structure , forming and/or strengthening a flux rope .
the enhanced poloidal flux in a flux rope would increase the upward lorentz self - force and thus further accelerate the flux rope outward forming a cme @xcite . from observations , @xcite found that a flare - associated cme usually undergoes three distinct phases of evolution : the slow - rising initiation phase , the impulsive acceleration phase , and the near - constant propagation phase .
the impulsive acceleration phase of the cme coincides in time very well with the rise phase of the flare ( also see , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
this temporal coincidence strongly implies that the same solar eruption could produce both the flare and cme phenomena , which are coupled together through magnetic reconnection .
the flux rope , a set of magnetic field lines winding around an axis , is a key feature in classical eruptive flare models and many cme observations .
helical magnetic field structures , the so - called magnetic clouds , have been frequently found in in - situ solar wind observations @xcite . the internal helical structure of cmes can be directly observed by coronagraphs ( e.g. , * ? ? ?
moreover , coronagraph images show that a cme usually consists of a three - part structure : the bright front or leading edge ( le ) , the enclosed dark cavity and the inner bright core @xcite .
the existence of a flux rope could explain the dark cavity structure of a cme ( e.g. , * ? ? ?
* ; * ? ? ?
flux ropes might be preceded by sigmoid structures in the corona before the eruption @xcite .
recent observations showed that the sigmoid structures , which originally consisted of two opposite j - shaped loops , could be transformed into a flux rope @xcite .
@xcite and @xcite studied the instability of the flux rope and proposed that the kink and/or torus instability can trigger the cme .
@xcite made a more detailed study and proposed the partial torus instability of the flux rope .
through nonlinear force - free field models , several authors were able to reconstruct a flux rope configuration in the corona from observed photospheric magnetic fields ( e.g. , * ? ? ? * ; * ? ? ?
* ; * ? ? ?
considerable efforts have also been made in numerical mhd simulations to study the formation and dynamic behavior of flux ropes @xcite . while it is widely accepted
that the flux rope is either formed during the cme eruption or exists prior to the eruption , the direct observation of the flux rope structure during the eruption process has been rare ( a plausible case was reported by * ? ? ?
* ) . in this letter ,
we report the observation of a cme event by _ solar dynamic observatory _ ( _ sdo _ ) , _ solar terrestrial relations observatory _ ( _ stereo _ ) , and _ solar and heliospheric observatory _ ( _ soho _ ) , which clearly showed a flux rope structure during the impulsive acceleration phase of the cme . the observed structure is also found to be consistent with the classical eruptive flare model described earlier . in section 2 , we present the observations and results , which are followed by a summary and discussions in section 3 .
the atmospheric imaging assembly ( aia ) on board _ sdo _ ( the first observatory of nasa s living with a star program ) is designed to image the multi - layered solar atmosphere , including photosphere , chromosphere , transition region , and corona in ten narrow uv and euv passbands with unprecedented cadence ( 12 seconds ) , high spatial resolution ( 0.6@xmath1 pixel size ) and large field of view ( 1.3@xmath2 ) .
aia s capacity of high cadence and multiple temperature provides the opportunity of observing the formation of the flux rope in the low corona during the impulsive acceleration phase of a cme .
in addition , the sun earth connection coronal and heliospheric investigation ( secchi ) on board _ stereo _ provides the direct observations of cmes in the coronagraphs .
the large angle spectroscopic coronagraph ( lasco ) on board _ soho _ provides the cme images from one additional perspective .
on 2010 november 3 , a c4.9 class soft x - ray flare occurred at the east limb of the sun , which started at @xmath012:07 ut and peaked at @xmath012:21 ut . with _
stereo_-b coronal images , which showed the flare on the disk , we find that the flare was located at @xmath0s20e96 from the perspective of the earth , and thus a fraction of the soft x - ray emission from the flare was occulted . through inspecting lasco / c2 images , it is confirmed that the flare was associated with a limb cme . in aia 131
images , a blob of plasma first appeared above the solar limb at @xmath012:13 ut ( figure [ f1](a)(c ) ) .
the circular blob was only visible as enhanced emission in the 94 and 131 images , but not in any other aia passbands , which indicate that the plasma in the blob was very hot with temperatures as high as @xmath07 to 11 mk ( see * ? ? ?
* for response functions of aia channels ) . at @xmath012:14 ut ,
the blob of hot plasma started to push its overlying magnetic field upward and form a sharp edge ( figure [ f1](d ) ) .
the overlying field lines seem to be stretched up continuously . immediately below the blob
, there appeared a y - type magnetic configuration , in which a bright thin line extended downward .
this kind of structure is largely consistent with the classical eruptive flare model @xcite , in which a cs forms below the rising flux rope , and magnetic reconnection induced in the cs converts the stretched surrounding magnetic field into the new poloidal flux of the flux rope .
the observed plasma blob is likely to be the flux rope as proposed in the classical theory ( more evidence will be given later ) .
the observations may indicate that the flux rope was undergoing the forming and growing phase ( figure [ f1](b)(f ) ) .
if this is the case , it might be the first time that the formation of flux rope is directly observed in the low corona .
the flux rope can be seen , not only because of the superb aia imaging capacity , but also the favorable orientation of the flux rope ; it is likely that the flux rope was oriented edge on , i.e. , the axis of the flux rope lied more or less along the line of sight .
nevertheless , @xmath3-b euvi 195 images did not reveal a definite orientation of the eruption structure , since a post - eruption loop arcade did not show up for this event in which the flare had a short duration in soft x - ray
. the newly added poloidal flux will increase the outward lorentz self - force of the flux rope , resulting in faster acceleration of the structure . in response to this outflow of magnetic flux
, the inflow speed of the anti - parallel ambient magnetic field toward the dissipating cs is expected to be strengthened , resulting in a faster magnetic reconnection rate , which further increases the poloidal flux injection rate .
thus , this is a positive feedback process between magnetic reconnection and flux rope formation / escaping .
the observations are consistent with this kind of process .
as the flux rope continually rose , the overlying fields seemed to be stretched as shown in figure [ f2](a)(c ) .
further , it is seen that the stretched magnetic field lines curved toward the dissipation region at 12:15:45 ut ( figure [ f2](a ) ) .
the curved magnetic field lines may arrive at the reconnection region and then participate in the reconnection ( 12:16:33 ut , figure [ f2](c ) ) . on the other hand
, we note that the curving process only lasted for @xmath02 minutes .
the curving of other field lines might be invisible own to a higher temperature ( @xmath411mk ) .
it is also worth noting that this very hot plasma present in the inflow of the region is difficult to be explained by the classical eruptive model .
there might be heating taking place in the region before the reconnection .
in addition to the inflow from two sides of the cs , we also observe the shrinkage of magnetic field lines underneath the cs , indicating the ongoing process of magnetic reconnection . at the lower y - type point of the cs ,
the reconnected field lines were originally cusp - shaped .
however , thanks to the magnetic tension force , they then shrank and became loop - shaped @xcite . after the cs forming , probably as early as at 12:14:45 , the reconnection in the cs formed the post - flare loops under the cs .
as the reconnection was ongoing , the y - type and reversed y - type configuration appeared and they was connected by a short and thin feature ( e.g. , figure [ f1](f ) ) . in order to illustrate the shrinkage , we plot the reversed y - type structure imaged in 131 in figure [ f2](d)(f ) .
the cusp - shaped field lines appeared near the crossing of the reversed y - type configuration ( figure [ f2](d ) ) ; they may be just formed by the reconnection of the two antiparallel legs of the stretched field lines .
subsequently , the field line with the cusp - shape rapidly shrank ( figure [ f2](e ) ) and then changed to the semi - circular post - flare loops ( figure [ f2](f ) ) .
the details of the shrinkage process can be better seen in the attached movie . in order to compare the emission produced in different aia passbands , we plot the temporal profiles of the intensity integrated over the eruption region ( the region shown in figure [ f1 ] ) for 131 , 94 , 211 , and 171 respectively in figure [ f3 ] , along with the @xmath5 soft x - ray 18 profile .
it is found that the profile of the 131 emission was similar to that of the soft x - ray , but with a systematic delay of @xmath02 minutes .
the emission profile of 94 was also similar to that of 131 but with a delay of @xmath07 minutes .
the delay is probably caused by the cooling from high to low temperature of the plasma located in and surrounding the reconnection region @xcite .
on the other hand , the integrated intensity of 211 and 171 started to decrease at 12:10 ut , probably caused by the reconnection heating moving the plasma out of the temperature range sensitive to these passbands .
in this section , we discuss the multi - component and multi - temperature nature of the cme structure in the cme formation phase or impulsive phase .
in addition to the flux rope , the formation of cme le was also observed .
it appeared that the top of the accelerated flux rope was compressing the surrounding plasma and producing the compression front with enhanced plasma density .
the le was formed from the successive stacking of the compression front as the flux rope was expanding and rising , but not the le does not represent the real rising of the same stretched field lines .
therefore , the le velocity of the cme is the apparent velocity of the compression wave running ahead of the cme flux rope , while the flux rope acts as the driver of the compression wave . by tracking the le , the blob top , and the blob center
, we plot their height - time profiles in figure [ f3 ] ( b ) , from which one can find that the average velocities of the le , blob top and blob center in the fov of aia are @xmath01200 km / s , 630 km / s , 500 km / s , respectively . both the le and cavity later appeared in the fov of lasco / c2 ( figure [ f2](g ) ) .
another important finding from the observations is that the cme has multiple structural components well organized by the temperature .
the appearance of the eruption at @xmath012:15 ut is shown in six different passbands ( 131 , 94 , 211 , 193 , 171 , and 304 ) in figure [ f4 ] .
note that the images are difference images by subtracting the corresponding base images at @xmath012:00 ut .
it is evident that the center part of the cme is of high temperature ( as high as @xmath07 to 11 mk ) ( figure [ f4](a ) and ( b ) ) ( see also * ? ? ?
the temperature of the le may distribute from @xmath00.05 to 2.5 mk , since the le is visible at 211 , 193 , 171 , and 304 ( figure [ f4](c)(f ) ) . from the composite images with three passbands ( 171 , 211 , and 131 ) ( attached movie )
, one can clearly identify that the cme has a multi - temperature structure ( a hot core and a cool le ) .
the multi - temperature structure of the cme also sheds light on the origin of the dimming region in the impulsive phase .
it is usually believed that the dimming is caused by the loss of mass in the low corona ( e.g. , * ? ? ?
* ) . however , the multi - temperature observations suggest that the high temperature in the center part of the cme also play a role in producing the observed dimming .
hot plasma induces the emission at higher temperature and is less visible or invisible in the passbands of lower temperature .
as shown in figure [ f4 ] , the shape of the dimming region ( at 211 and 171 ) was almost consistent with the shape of the cme hot core ( at 131 ) .
it reveals that the dimming may result from the absence of the low temperature emission , which holds at least during the early formation phase of the cme .
note that , the 193 also includes partial emission of the hot plasma so that the dimming present at 193 was blended by a few hot emission from the cme center part ( also see the 195 difference image in figure [ f2](i ) ) .
we summarize our observations of the formation of the flux rope and the multi - temperature structure of the cme in a schematic model shown in figure [ f5 ] .
the black lines denote the magnetic field configuration .
the different features relevant to the magnetic reconnection , including the flux rope , le , curving , and shrinkage , are indicated by the arrows .
the multi - temperature components of the cme structure are shown in different background colors ; the texts indicate the effective observing passbands for different features .
it is most likely that the flux rope structure was forming in the impulsive acceleration phase of the cme .
the formation process was highly consistent with the classical eruptive flare model @xcite .
the ejection of a blob of hot plasma , which is the possible pre - flux - rope structure , moved upward and stretched the overlying restraining magnetic field lines . underneath the blob
, the stretched field lines were pushed toward the line - shaped cs and reconnected in the cs .
for this particular event , it is largely the magnetic reconnection that produced the bulk of the flux rope structure through injecting the poloidal magnetic flux .
the ejection of the magnetic flux rope caused a decrease of magnetic pressure at both sides of the cs and thus led to an inflow toward the cs @xcite .
the inflow curved the legs of the newly stretched field lines toward the reconnection region and made the reconnection continuing @xcite . under the lower tip of the cs ,
the reconnected magnetic field lines has initially a cusp shape and then shrank to become the semi - circular post - flare loops @xcite . in the upper tip of the cs
, the newly formed poloidal magnetic field lines were added into the flux rope , carrying with them the heat generated in the reconnection region .
therefore , the temperature in the center part of the cme was higher than the surrounding part .
this resulted in the multi - temperature structure of the cme .
it is worthy noting that the counterpart of the flux rope feature we are reporting here may have been noticed in earlier observations , e.g , the plasmoid above soft x - ray loops @xcite .
@xcite derived the temperature of the plasmoid at @xmath010 mk , which was similar to the response temperature of 131 .
furthermore , we believe that the blob of the hot plasma , which later formed the full - fledged flux rope through magnetic reconnection , was originated from either a sheared core field in the low corona @xcite or a weakly twisted flux rope existing prior to the eruption @xcite . we thank the anonymous referee for his / her valuable comments that helped to improve the paper .
sdo is a mission of nasa s living with a star program .
x.c . , and m.d.d .
are supported by nsfc under grants 10673004 , 10828306 , and 10933003 and nkbrsf under grant 2011cb811402 .
is also supported by the scholarship granted by the china scholarship council ( csc ) under file no .
j.z . , is supported by nsf grant atm-0748003 and nasa grant nng05gg19 g .
soft x - ray 18 flux and 131 ( @xmath011 mk ) , 94 ( @xmath07 mk ) , 211 ( @xmath02 mk ) , and 171 ( @xmath00.6 mk ) integral intensity in the fov of figure 1 , which are normalized by their maxima .
two vertical lines show the onset and peak times of the flare in @xmath5 x - ray .
( b ) the height - time profiles of the le , the blob top , and blob center in the impulsive phase.,scaledwidth=48.0% ] | magnetic flux rope is believed to be an important structural component of coronal mass ejections ( cmes ) .
while there exist much observational evidence of the flux rope after the eruption , e.g. , as seen in remote - sensing coronagraph images or in - situ solar wind data , the direct observation of flux ropes during cme impulsive phase has been rare . in this letter , we present an unambiguous observation of a flux rope still in the formation phase in the low corona .
the cme of interest occurred above the east limb on 2010 november 03 with footpoints partially blocked .
the flux rope was seen as a bright blob of hot plasma in aia 131 passband ( peak temperature @xmath011 mk ) rising from the core of the source active region , rapidly moving outward and stretching upward the surrounding background magnetic field .
the stretched magnetic field seemed to curve - in behind the core , similar to the classical magnetic reconnection scenario in eruptive flares . on the other hand
, the flux rope appeared as a dark cavity in aia 211 passpand ( 2.0 mk ) and 171 passband ( 0.6 mk ) ; in these relatively cool temperature bands , a bright rim clearly enclosed the dark cavity .
the bright rim likely represents the pile - up of the surrounding coronal plasma compressed by the expanding flux rope .
the composite structure seen in aia multiple temperature bands is very similar to that in the corresponding coronagraph images , which consists of a bright leading edge and a dark cavity , commonly believed to be a flux rope . online - only material : color figures , animations |
circumstellar disks play a central role in both star and planet formation .
the disks are thought to be the conduit through which sun - like stars assemble the bulk of their mass @xcite .
they are also the birthplace of planets . despite years of active research ( see reviews by @xcite for early work ) , we still do not have answers to some of the most basic questions , such as `` when and how does a disk appear in the process of star formation ? ''
observationally , it is difficult to separate out the protostellar disk from the surrounding envelope during the early , embedded phase of star formation with the current generation of interferometers ( e.g. , @xcite ) , although the situation will improve drastically in a few years , when alma comes online @xcite .
@xcite inferred the existence of a large ( hundreds of au in radius ) , massive ( more than @xmath4 ) disk around the protostar serpens firs 1 , based mostly on carma dust continuum observations . without a detailed knowledge of the central stellar mass and gas kinematics on small ( arcsec ) scales , it is difficult to determine whether the inferred structure is a rotationally supported disk , a magnetically induced `` pseudodisk , '' or simply the inner part of an envelope that is dynamically more complicated than the @xcite s ( @xcite ) rotating , collapsing sphere adopted by @xcite for the protostellar envelope . theoretically , disk formation in dense cores of molecular clouds magnetized to the observed level ( corresponding to a mean dimensionless mass - to - flux ratio @xmath5 a few ; @xcite ) is controlled by magnetic braking , which has been difficult to quantify until recently .
detailed calculations have now shown that the magnetic braking is apparently so efficient as to inhibit the formation of rotationally supported disks _ during the main protostellar accretion phase _ for a moderate level of core magnetization , as long as the magnetic field is frozen in the matter ( i.e. , the ideal mhd limit ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; see , however , @xcite for a different view , and also @xcite and @xcite for the evolution of rotating , magnetized cores prior to the main accretion phase ) . in order for a rotationally supported disk to exist around a rapidly accreting protostar
, the magnetic braking must be weakened one way or another .
the most obvious possibility is through non - ideal mhd effects , which are expected to be important in lightly ionized star - forming cores .
however , we have recently shown that ambipolar diffusion , the best studied non - ideal mhd effect in star formation , does not sufficiently weaken the braking to allow rotationally supported disks to form for realistic levels of core magnetization and ionization ; in some cases , the magnetic braking is even enhanced ( @xcite , @xcite ; see @xcite , @xcite and @xcite , who considered the effects of ambipolar diffusion on magnetic braking and core fragmentation during the earlier phase of prestellar core evolution ) .
this motivates us to examine whether two of the remaining non - ideal mhd effects , ohmic dissipation and the hall effect , can weaken the coupling between the magnetic field and the bulk neutral matter ( and thus the strength of magnetic braking ) enough to enable the formation of rotationally supported disks .
this paper concentrates on the ohmic dissipation .
1ex the resistivity @xmath6 is related to the electrical conductivity @xmath7 through @xmath8 for lightly ionized molecular gas in dense cores , the electric conductivity is usually dominated by the contribution from electrons , although the grain contribution can dominate at high densities when large amounts of small grains are present ( e.g. , @xcite ) . to obtain a rough estimate for the conductivity , we will concentrate on the contribution from the electrons , which is given by @xmath9 where @xmath10 and @xmath11 .
we have adopted a value @xmath12 for a @xmath13 gas from @xcite .
this value is somewhat smaller than that used in @xcite and @xcite , which is @xmath14 , but somewhat larger than that of @xcite , which is @xmath15 .
it increases slowly with temperature , roughly as @xmath16 , yielding a lower resistivity at a higher temperature . since our calculation
is assumed to be isothermal , the temperature dependence is not used . substituting equation ( [ conductivity ] ) into equation ( [ resistivity ] ) yields @xmath17
the fractional ionization @xmath18 can be computed from detailed chemical networks . at the relatively high densities that are relevant for disk formation , it depends on the grain size distribution .
for the standard @xcite distribution , @xcite computed the ionization fraction over a large range of number density ( their figure 1 ) . over the critical range of density @xmath19@xmath20 where the transition between the rapidly infalling pseudodisk and a potential keplerian disk is expected to occur ,
the fraction can be fitted roughly by @xmath21 which corresponds to a resistivity of @xmath22 the fit of fractional ionization deviates substantially from the computed values at low densities .
however , the deviation is expected to be inconsequential because the low density envelope is sufficiently ionized to be close to the ideal mhd limit ( ignoring ambipolar diffusion , which is investigated separately elsewhere , see e.g. , @xcite and @xcite ) according to either equation ( [ ionization ] ) or the computed values .
if only relatively large grains of size ( of order @xmath23 ) exist , the ionization fraction would be higher ( by as much as three orders of magnitude , see figure 1 of @xcite , or compare figures 1 and 3 of @xcite ) , making the resistivity smaller .
equation ( [ standard ] ) should therefore be viewed as a rough indication of the plausible upper range of the `` classic '' ( as opposed to `` anomalous '' ) resistivity , especially in view of the fact that the actual temperature in the high density region of disk formation is expected to be higher than the @xmath13 adopted in our calculation , which would lower the resistivity .
it may be enhanced , however , through anomalous processes ( e.g. , @xcite ) .
the rest of the paper is organized as follows . in 2 , we describe the setup of the disk formation problem , focusing on the protostellar mass accretion phase , when a central object has formed . for illustrative purposes
, we will concentrate on a particular set of core parameters for which a @xmath3-scale disk is formed at a representative time @xmath24 ( or about @xmath25 ) in the absence of magnetic braking , and a magnetic field that is strong enough to suppress the formation of a rotationally supported disk in the ideal mhd limit .
we consider in 3 whether the illustrative classic resistivity given in equation ( [ standard ] ) can weaken the magnetic braking enough to enable disk formation , and find that the answer is no . in 4 , we explore the question of how large the resistivity must be in order for the @xmath3-scale rotationally supported disk to reappear in our model problem .
it turns out to be of order @xmath2 , a few orders of magnitude larger than the classic value expected on the @xmath3-scale , as anticipated by @xcite based on the concept of the so - called ohm sphere , although our value is one to two orders of magnitudes smaller than theirs , depending on the field strength .
the numerical results and their astrophysical implications are discussed in the last section , 6 .
our primary goal is to illustrate the effects of ohmic dissipation on magnetic braking and disk formation . for this purpose ,
it is desirable to have as small a numerical magnetic diffusivity as possible , which demands a high spatial resolution .
high resolution , however , puts a stringent limit on the computation time step @xmath26 , which is proportional to the square of the ( smallest ) grid size for the explicit method that we use to treat the ohmic dissipation .
although subcycling can be used to alleviate the time step problem substantially , a large number of cycles are still needed to reach a reasonable protostellar accretion time , say , @xmath27 , making the treatment of self - gravity difficult . for this initial exploration of the effects of ohmic dissipation
, we have decided to set up a simplified disk formation problem , with a rotating , non - self - gravitating envelope falling onto a central object of fixed mass @xmath28 .
it will turn out that a large `` anomalous '' resistivity is needed to save the disk , an elaborate disk formation model is probably not warranted at this early stage of the investigation . specifically
, we solve the equations of resistive mhd , @xmath29 @xmath30 @xmath31 @xmath32 under the assumption of axisymmetry , and adopt a spherical polar coordinate system @xmath33 , with a stellar object of @xmath34 at the origin such that the gravitational potential is @xmath35 , neglecting the self - gravity of the gas .
self - gravity will be included in a follow - up work , where we will evolve the magnetized core from its formation to collapse in a self - consistent manner .
for this initial study , we aim to capture the essence of the problem of magnetic suppression of disk formation in the simplest possible way , by adopting uniform distributions for the initial gas density and magnetic field , and to study the effects of a new ingredient , resistivity , in this simplest problem . at time @xmath36
, we fill the computation domain between @xmath37 and @xmath38 with a uniform density @xmath39 ( corresponding to @xmath40 ) , so that the total envelope mass is @xmath41 . for simplicity , we assume that the gas stays isothermal , with an isothermal sound speed @xmath42 ( corresponding to a temperature of about @xmath13 ) . for the initial rotation , we adopt the following prescription : @xmath43 where we choose @xmath44 , @xmath45 is the cylindrical radius , and @xmath46 .
the softening of the rotational profile inside the cylindrical radius @xmath47 is to prevent the angular speed from becoming singular near the rotation axis .
since the inner radius of our computational domain is @xmath48 , we are concerned with the formation or suppression of relatively large disks ( of tens of au or more ) only .
our initial rotation is relatively fast : @xcite estimated observationally that the ratio of rotational to gravitational energies for dense @xmath49 cores is typically @xmath50 , which would imply @xmath51 for our setup .
our choice of a faster rotational speed is conservative in that it is harder to remove a larger initial angular momentum through magnetic braking ( see also @xcite ) .
the calculations are done with a new version of zeus , dubbed `` zeustw '' .
zeustw , written in idiomatic fortran95 , is based on the `` zeus36 '' code @xcite , itself derived from zeus3d ( lca version 3.4.2 : @xcite ) ; zeus36 is parallel by domain decomposition , and utilizes dynamic memory for its field and grid arrays .
a small memory pool of temporary arrays replaces the worker arrays of zeus3d , increasing programming flexibility at essentially zero runtime cost .
zeustw adds to zeus36 the ability to solve many non - ideal mhd problems in an explicit form , covering the ohmic , hall , and ambipolar diffusion terms .
to treat the ohmic term relevant for this paper , we used a resistivity algorithm based on @xcite , which includes subcycling .
we tested the code by diffusing an initial gaussian profile in cartesian geometry , in one , two , and three dimensions .
we also tested slightly different forms of the algorithm ( such as changing the operator splitting by calculating the current density @xmath52 before or after the magnetic field @xmath53 is updated ) , and different subcycling prescriptions ( such as varying the maximum number of ohmic subcycles from 1 to 100 ) . the code passed all of these tests . .
logarithm of density ( colormap ) ; radial infall sonic transition @xmath54 ( white line ) ; poloidal velocity field ( arrows , scale can be estimated from the radial infall component at the white line ) ; and level of keplerian support @xmath55 ( dark gray dashes at 90% and solid line at 100% ) , where @xmath56 . ]
we adopt the standard hydro outflow boundary conditions at both the inner and outer ( radial ) boundaries of the computation domain ( these boundary conditions impose continuity of material outflowing the grid , helping to reduce artificial reflections of waves at boundaries ) .
we use @xmath57 non - uniformly spaced grid points in the radial direction , corresponding to a smallest radial grid size @xmath58 ( or @xmath59 ) and a ratio of adjacent zone sizes of @xmath60 .
we use 180 uniform zones in the @xmath61 direction , corresponding to a minimum polar grid size of @xmath62 , comparable to the smallest radial grid size . with this setup
, we obtain a rotationally supported disk of radius @xmath63 and mass @xmath64 at a time @xmath24 in the absence of magnetic braking ; the time is comparable to the duration of the deeply embedded class 0 phase ( @xcite , see , however , @xcite who concluded that the duration is longer ) .
a snapshot of the disk is shown in figure [ hydro ] .
the hydrodynamic disk is resolved in the radial direction by more than 200 zones .
we have experimented with coarser grids ( e.g. , a total of @xmath65 radial zones and @xmath66 angular zones ) , finding qualitatively similar results , which indicates that the resolution employed is sufficient .
an example of the lower resolution runs is shown in [ sec : er ] below .
the disk size at @xmath24 can be understood roughly as follows .
we first estimate the radius of the infall region @xmath67 at time @xmath68 , and the specific angular momentum @xmath69 of the material at that radius . assuming that pressure gradients are negligible compared to the central gravity
, we have the infall speed at any given radius to be near the free fall value @xmath70 .
the infall radius is then roughly @xmath71 , which yields @xmath72 at @xmath24 , much larger than @xmath47 ( the characteristic radius for rotation initial profile change ) .
therefore , the equatorial material at radius @xmath67 has a specific angular momentum of @xmath73 .
conservation of angular momentum around a star of mass @xmath74 tells us that the specific angular momentum corresponds to a centrifugal radius of @xmath75 , which is about a factor of 2 larger than our numerically obtained disk radius @xmath76 .
we consider the agreement satisfactory , given the crudeness of the infall estimate , and the fact that most material on the surface of the infall sphere at radius @xmath67 has a specific angular momentum @xmath69 less than the above estimate ( valid only for the equatorial region ) , and the centrifugal radius is sensitive to @xmath69 ( as @xmath77 ) .
but for the ideal mhd model .
the field lines are drawn as thin solid lines ( 20 contour lines of magnetic flux , with flux levels quadratically spaced from zero at the axis up to the flux value at the outer equator point of each magnetized figure ) .
the field component @xmath78 is @xmath79 at the center of the figure .
no rotationally supported structure forms in this simulation , although the flow structure is strongly affected by numerical reconnection of field lines . ] as mentioned in 1 , previous studies have shown that a moderate magnetic field corresponding to a dimensionless mass - to - flux ratio of a few can completely suppress the formation of rotationally supported disks in the ideal mhd limit ( at least when the rotation and magnetic axes are aligned ; see @xcite ) .
even though the mass - to - flux ratio is ill - defined in our current model setup that does not include the self - gravity of the envelope gas , we find that the disk formation can still be suppressed if the magnetic field is strong enough .
an example is shown in figure [ idealmhd ] , where the initially uniform mass distribution in the computational domain of the hydro model shown in figure [ hydro ] is threaded with a uniform magnetic field @xmath80 along the rotation axis .
the field strength is in the range inferred by @xcite for a sample of dark cloud cores , after correcting for projection effects . for comparison
, we also considered a weaker field of @xmath81 and obtained qualitatively similar results .
for mhd calculations , we adopt the standard outflow boundary condition for the magnetic field at the outer radial boundary , and a torque - free ( i.e. , @xmath82 ) outflow boundary condition at the inner radial boundary , as in @xcite .
the flow structure is strongly affected by numerical reconnection of the field lines .
the reconnection is an unavoidable consequence of the dragging of the field lines into a highly pinched , split - monopole type configuration by gravitational collapse in the ideal mhd limit @xcite .
the build up of magnetic flux in the central region as a result of continuous mass accretion forces the oppositely directed field lines above and below the equator closer and closer together , eventually triggering ( numerical ) reconnection .
it is present in other ideal mhd calculations of magnetized core collapse , such as @xcite . because of
the ( unavoidable ) numerical artifacts , it is difficult to quantify the exact strength of the magnetic braking . nevertheless , it is clear from figure [ idealmhd ] that a coherent disk does not exist ; it is replaced by a set of dense blobs , which are not rotationally supported ( and may break up asymmetrically in 3d ) .
the prominent dense blobs and sheets at radii @xmath83 and smaller are all rotating well below their local keplerian speed ( especially the innermost ones ) , as can be seen from the thick dashed and solid dark gray lines at radii greater than @xmath84 in figure [ idealmhd ] , which mark the location of @xmath85 and @xmath86 , respectively ( @xmath86 is the local keplerian speed ) .
the blobs and sheets are supported to a large extent by the magnetic tension force from the pinched field lines .
we now address the question of whether the illustrative classic resistivity given in equation ( [ standard ] ) can weaken the magnetic braking enough to enable the formation of rotationally supported disks or not . for this purpose ,
we have implemented into the mhd code zeustw an explicit treatment of resistivity , following the work of @xcite and @xcite ( @xcite ; see also @xcite ) . in these treatments ,
the resistive term of the induction equation is included through operator splitting , respecting the constrained transport condition that keeps the field divergence null to machine round - off .
numerical stability of this term limits the timestep @xmath87 to a value @xmath88 , a quadratic requirement that can be much more stringent than the time @xmath89 required for treating the hydrodynamics and the ideal mhd terms .
fortunately , the resistive term is computationally inexpensive compared to the rest of the induction equation , and so it can be efficiently subcycled .
we have checked that subcycling speeds up the calculations typically by a substantial factor and does not change the results significantly .
we have also tried small changes in the operator splitting scheme , such as altering the order of the ideal mhd and the resistivity terms , or adding both terms together .
these changes had no substantial effect on the results , and therefore we chose the scheme that was most convenient for efficient subcycling , which is one identical to that used in @xcite .
figure [ classic ] displays a snapshot of the simulation at the representative time @xmath27 .
a dense flattened structure develops in the inner equatorial region .
it is , however , fragmented as in the ideal mhd case , indicating that the classic resistivity is too small to completely suppress the numerical reconnection of field lines .
the dense blob at radius @xmath76 , in particular , is similar to the outermost blob in figure [ idealmhd ] , and is due to an early reconnection event .
the artificial reconnection events make it difficult to carry out detailed analysis . nevertheless , it is easy to show that the dynamics of the inner equatorial region is far from that of a rotationally supported disk .
first , the equatorial material inside @xmath76 is often counter - rotating , with a rotation speed that sometimes decreases toward the center ( rather than increasing as would be expected for a keplerian disk : see figure [ classic_vrvphi ] ) .
second , the radial motion in the region is mostly supersonic ( @xmath90 ) .
the rapid infall is an indication that the dense , flattened structure is a magnetically produced pseudodisk @xcite , rather than a rotationally supported keplerian disk .
there is one region of supersonic _ outflow _ at @xmath91 .
this region , and other outflow events that appeared during the simulation , are related to episodic reconnection events , as in the ideal mhd model . in any case
, classic resistivity does not appear capable of weakening magnetic braking enough to enable the formation of a rotationally supported disk .
enhanced resistivity is needed if ohmic dissipation is to save the disk . and [ idealmhd ]
but for the mhd model with the classic resistivity given in equation ( [ standard ] ) . a dense , flattened structure exists near the equator , but it is a highly dynamic magnetically produced pseudodisk rather than a rotationally supported disk .
the flow structure is affected by numerical reconnection of field lines . ]
( solid line ) and rotation @xmath92 ( dashed ) for the mhd model with classic resistivity , showing the lack of a rotationally supported disk .
note the counter - rotation inside @xmath83 , with rotation speed often decreasing ( rather than increasing ) toward the origin .
the spikes of positive radial velocity at @xmath93 are related to episodic reconnection events . ]
-1.5ex p7.5exp5.5exp3exp3exl & @xmath94(@xmath95 ) & @xmath96 & disk & notes + hd & 0 & no & yes & hydro simulation + imhd & 35.4 & no & no & ideal mhd + cr & 35.4 & yes & no & classic resistivity given by eq .
( [ standard ] ) + er18 & 35.4 & yes & no & enhanced resistivity @xmath97 + er18.5 & 35.4 & yes & no & enhanced resistivity @xmath98 + er19 & 35.4 & yes & no & enhanced resistivity @xmath99 + er19.5 & 35.4 & yes & yes & enhanced resistivity @xmath100 + er20 & 35.4 & yes & yes & enhanced resistivity @xmath101 + imhdw & 10.6 & no & no & weaker field , ideal mhd + er18w & 10.6 & yes & no & weaker field , @xmath97 + er18.5w & 10.6 & yes & yes & weaker field , @xmath98 + er19w & 10.6 & yes & yes & weaker field , @xmath99 + er19.5w & 10.6 & yes & yes & weaker field , @xmath100 + er20w & 10.6 & yes & yes & weaker field , @xmath101 + and [ idealmhd ] but for model er18 with an enhanced resistivity @xmath102 . ] and [ idealmhd ] but for model er19 with an enhanced resistivity @xmath103 . ] and
[ idealmhd ] but for model er20 with an enhanced resistivity @xmath104 . ] , with reduced space resolution . ]
it is well known that anomalous resistivity orders of magnitude larger than the classic value is required to explain the 11-year sunspot cycle ( @xcite , 9.3 ) .
large anomalous resistivity may also exist in star - forming molecular cloud cores at high densities @xcite , although its magnitude is uncertain . here
we consider the simplest case of a spatially uniform resistivity of a range of values and for two different magnetic strength @xmath105 and @xmath106 .
the models are summarized , together with the ones discussed earlier , in table [ table : models ] .
we will concentrate on three representative , stronger field ( @xmath107 ) models with @xmath97 ( model er18 ) , @xmath108 ( er19 ) , and @xmath109 ( er20 ) , respectively .
the weaker field models yield qualitatively similar results , and will be discussed briefly toward the end of the section . figures [ er18][er20 ] show snapshots of the three models with enhanced resistivity , and figures [ vr_all][bz_all ] display their infall and rotation speeds and vertical field strength on the equator .
additionally , figure [ er20r ] shows the effect on model er20 of reducing numerical resolution by a factor of @xmath110 in both @xmath67 and @xmath61 .
compared to the models without resistivity ( model hd shown in figure [ hydro ] ) and with the classic resistivity ( model cr , figure [ classic ] ) , the mass distribution and velocity field are much smoother , indicating that the enhanced resistivity in these models has largely eliminated the numerical reconnection .
the suppression facilitates quantitative analysis of the simulations .
the mass distributions in the two lower resistivity models ( models er18 and er19 ) appear similar .
both have a dense equatorial region sandwiched from above and below by a more diffuse infalling envelope .
the infall in the inner part of the envelope ( within a few hundred aus ) appears to be guided by the field lines .
it strikes the surfaces of the dense equatorial structure at a supersonic speed , forming two `` accretion shocks . ''
these are reminiscent of the classic accretion shocks around keplerian disks formed in non - magnetic simulations ( e.g. , @xcite ; see figure [ hydro ] ) .
their production is due , however , mainly to the magnetic field rather than rotation .
indeed , the same shocks form even in the absence of any rotation . for this reason , we will term these shocks `` the magnetically induced accretion shocks . ''
they surround a magnetically induced `` pseudodisk '' @xcite rather than a rotationally supported keplerian disk .
, solid line ) , er19 ( @xmath103 , dashed ) , and er20 ( @xmath104 , dash - dotted line ) at the representative time @xmath24 . the free - fall speed is plotted as the lower dotted line for comparison . ]
but for the equatorial rotation speed .
the keplerian speed is plotted as the upper dotted line for comparison . ]
but for the vertical magnetic field component @xmath111 on the equator .
] figure [ vr_all ] shows that the dense equatorial structure in models er18 and er19 contains a pseudodisk rather than a rotationally supported disk . in both cases , the infall is supersonic ( @xmath112 ) within @xmath113 of the central object , although there is significant difference in the infall speed within @xmath114 : the equatorial material in the less resistive model er18 falls inward much more slowly than that in the more resistive model er19 .
the prominent deceleration of infalling material around @xmath114 in model er18 is reminiscent of the ambipolar diffusion - induced c - shock ( or c - transition ) first discussed in @xcite ( @xcite ; see also @xcite ; @xcite ; @xcite ; @xcite and especially figure 3 of @xcite ) .
the physical origin is also similar . just as ambipolar diffusion , the ohmic dissipation enables the field lines to diffuse relative to the material that plunges into the central object .
the net effect is the formation of a strong ( poloidal ) field region at small radii , where the inward advection of the field lines by infalling material is balanced by the resistivity - induced outward magnetic diffusion .
a relatively small resistivity such as that in model er18 allows for relatively larger magnetic gradients and lorentz forces , which decelerate the infalling material more strongly .
when the resistivity is small , the magnetic field remains a significant barrier to protostellar accretion .
the magnetic barrier weakens as the resistivity increases . in model er19 , which is ten times more resistive than model er18 , the ohmic dissipation - induced deceleration has all but disappeared .
this is because the higher resistivity can support only a smaller current , making it hard to effectively `` trap '' the magnetic flux that is dragged to the small radii by the collapsing flow .
nevertheless , the infall speed is considerably smaller ( by about a factor of two or so ) than the free - fall speed , which is also plotted in figure [ vr_all ] for comparison .
the sub - free - fall motion comes about because the equatorial infall is retarded by a combination of centrifugal and magnetic forces , although the magnetic forces appear to be the more important contributor , since the rotation speed shown in figure [ vphi_all ] is about 1/3 of the keplerian speed or less ( yielding a centrifugal force an order of magnitude weaker than the gravity ) .
note from figure [ vphi_all ] , for the least resistive model er18 , the equatorial rotation speed within @xmath114 of the origin decreases , rather than increases , toward the center . obviously , the decrease is due to magnetic braking , which is apparently so efficient as to produce a region of counter - rotation at @xmath115 .
the counter rotation is reminiscent of the case of magnetized core collapse in the presence of ambipolar diffusion with a relatively high cosmic ray ionization rate @xmath116 studied by @xcite ( @xcite ; see the bottom panel of their figure 4 ) . in both cases ,
the efficient braking is due to the strong poloidal field trapped by infall at the small radii .
it removes essentially all of the angular momentum of the equatorial material that crosses the inner boundary of the computation domain . in the more resistive model er19 , the equatorial rotation speed inside @xmath114 is initially kept nearly constant , before increasing rapidly .
the peak rotation speed remains well below the local keplerian speed , indicating that a rotationally supported disk is not formed .
we conclude that the dense , flattened , equatorial structure produced in both models er18 and er19 feature dense pseudodisks , but no rotationally supported dense disks .
these pseudodisks are thinner than the keplerian disk shown in figure [ hydro ] for the hydro case because of vertical compression by pinched field lines , and smaller column densities ( due to the fast infall rate inside the pseudodisk ) , corresponding to reduced scale height for a given gas density .
the absence of a rotationally supported disk in these two cases ( which have resistivities much larger than the classical value ) supports our results in 2.2 and 3 that disk formation is suppressed by magnetic braking in the ideal mhd limit and in the presence of the classical resistivity , even though these results were affected by numerical reconnection .
the most resistive model er20 differs drastically from the lower resistivity cases .
it produces a much denser disk of radius @xmath117 at the representative time @xmath24 ( see figure [ er20 ] ) , which turns out to be rotationally supported .
the strongest evidence for rotational support comes from figure [ vphi_all ] , which shows that the rotation speed at the equator is nearly identical to the keplerian speed inside @xmath117 .
the support is also evident in figure [ vr_all ] , which shows that the infall has nearly stopped at small radii ( the spike near the inner edge of the computation grid is likely related to the outflow boundary condition , which is not ideal for a rotationally supported disk ; it was present in the hydro case as well ) .
the thick dashed and solid dark gray lines in figure [ er20 ] , which mark the location of @xmath85 and @xmath86 , respectively , also make the case .
the disk has a mass @xmath118 at @xmath24 , corresponding to an average accretion rate ( from the envelope to the disk ) @xmath119 . as in the pure hydro case ( 2.1 ) ,
little of the disk material passes through the disk onto the central object in our axisymmetric simulations .
gravitational torques may be needed to drive the further evolution of such disks ( in three dimensions ) .
another feature of model er20 is that the poloidal field lines pass through the equatorial region with only a moderate bending .
it is of course a consequence of the large resistivity , which allows the field lines to diffuse outward easily relative to the infalling material .
this is especially true inside the inner , keplerian disk , where the combination of slow radial motion and large resistivity means that the magnetic field must be nearly vertical and uniform ( see the dot - dashed line in figure [ bz_all ] ) .
a consequence of the nearly uniform field is that there is little vertical compression by the magnetic field , yielding a keplerian disk that is thicker than the magnetically compressed pseudodisks in the less resistive models ( see figures [ er18 ] and [ er19 ] ) . outside
the keplerian disk lies a supersonically infalling pseudodisk .
it is linked to the surrounding envelope by magnetic field lines ; they are separated by the magnetically induced accretion shocks .
upstream of the shock , the low density envelope collapses more or less along the field lines , indicating that flux - freezing is approximately valid , because the field lines there are relatively straight and there is little current to be dissipated by the ( large ) resistivity .
downstream of the shock , the dense pseudodisk collapses horizontally , nearly perpendicular to the ( poloidal ) field lines .
at the relatively late time shown in figure [ er20 ] , the field lines inside the pseudodisk are nearly held fixed in space , with the inward advection due to infalling material balanced by the outward resistivity - driven diffusion . the role of resistivity in enabling disk formation
can be seen clearly in figure [ bz_all ] , where the vertical component of the magnetic field on the equator , @xmath78 , is plotted as a function of time for all three models with different enhanced resistivities .
the equatorial infall in the pseudodisk tends to compress the poloidal field lines , leading to a sharp increase in the poloidal field strength at small radii .
resistivity , on the other hand , tends to smooth out the field distribution .
the net effect is that the poloidal field strength at small radii depends strongly on the magnitude of the resistivity , being larger for smaller resistivities .
since most of the magnetic braking occurs at relatively small radii , where the infalling material attempts to spin up the fastest to conserve angular momentum , the stronger poloidal fields in the two less resistive cases ( models er18 and er19 ) produce a stronger braking , which is enough to suppress the formation of a rotationally supported disk .
the most resistive case ( model er20 ) has the weakest poloidal field at small radii ( where magnetic braking matters most in the model ) which , together with a large slippage between the field lines and matter in the azimuthal direction ( which makes it hard to generate a toroidal field ) , makes the magnetic braking ineffective and disk formation possible . that the disk in model er20 forms as a result of enhanced resistivity rather than numerical diffusion is supported by the fact that a nearly identical disk is formed in the lower resolution
run ( see figure [ er20r ] ) .
one can roughly estimate the size of the resistivity - enabled disk as follows . in order for the formed rotationally supported disk to survive against magnetic braking , the disk rotation must not twist up the poloidal magnetic field that threads the disk too strongly .
this is satisfied when the magnetic diffusion speed becomes comparable to the keplerian speed , which leads to the condition @xmath120 , where @xmath121 is the vertical scale length over which the toroidal magnetic field varies . if we take @xmath121 to be the disk pressure scale height ( ignoring possible magnetic compression of the disk ) , and use the standard result for a thin - disk @xmath122 ( where @xmath67 is the radius and @xmath123 is the isothermal sound speed ) , then @xmath124 or @xmath125 ( for @xmath104 and @xmath126 )
this is about a factor of two larger than the disk radius that we found numerically in figure [ er20 ] .
we should caution the reader that the above estimate did not take into account the detailed magnetic geometry and braking efficiency , which depends on the field strength ( see below ) . for a weak enough field ,
a disk can form independent of the magnitude of the resistivity . besides the three representative models discussed above , we have carried out two additional simulations with @xmath98 ( model er18.5 ) and @xmath127 ( model er19.5 ) . as one would expect based on the results for models er18 and er19 , no rotationally supported disk forms in model er18.5 .
there is a small ( @xmath128 in radius ) rotationally supported disk in the more resistive model er19.5 at the fiducial time @xmath24 .
the disk shrinks with time , however .
by @xmath129 , its radius decreases to @xmath130 ; it may disappear altogether at a later time .
the trend indicates that @xmath131 is probably close to the critical resistivity @xmath132 needed for disk formation . the value for the critical resistivity @xmath132 depends on the initial field strength @xmath94 .
this is to be expected since , in the limit of infinitely weak field , rotationally supported disks can form without any ( enhanced ) resistivity at all ( i.e. , @xmath133 ) . for a moderately magnetized dense core of @xmath134 , we find that the transition between the formation of a rotationally support disk and its suppression occurs between @xmath97 ( model er18w ) and @xmath135 ( model er18.5w ) . this value of @xmath94 is probably close to the lower limit to the field strength in dense cores , judging from the fact that the median field strength for the more diffuse , cold neutral medium ( cnm ) of atomic gas is @xmath136 @xcite and that the directly measured line - of - sight component of the magnetic fields in a number of dense cores is of this order or higher @xcite ; the full strength of the core magnetic field is likely significantly higher .
it is therefore reasonable to expect the critical value @xmath132 for disk formation in cloud cores magnetized to a realistic level to lie somewhat between @xmath137 and @xmath138 . in what follows
, we will take @xmath139 as the characteristic critical value , with the understanding that it depends somewhat on the field strength , and can be uncertain by a factor of @xmath140 in either direction .
our most important result is that classical resistivity alone appears unable to weaken magnetic braking enough to allow a rotationally supported disk to form for a realistic level of core magnetization , and that an enhanced resistivity of order @xmath139 is needed .
a large , enhanced resistivity ( of order @xmath141 ) was advocated by @xcite to solve the so - called `` magnetic flux problem '' in star formation .
they introduced the concept of the ohm sphere , with a radius : @xmath142 where @xmath143 is the central stellar mass .
it can be obtained by setting the characteristic magnetic diffusion speed @xmath144 equal to the free - fall speed @xmath145 . in the limit of a weak magnetic field
( so that the magnetic forces do not affect the accretion dynamics ) and zero rotation , @xcite showed that the magnetic field becomes more or less uniform inside the ohm sphere ; the sphere may be thus regarded as the region of complete magnetic decoupling .
if a rotationally supported disk were to form inside the ohm sphere , it would be protected against strong braking by the decoupling .
if the ohm sphere must be larger than the rotationally supported disk , then to form a disk of @xmath3 in radius around a half - solar mass star , the resistivity must be @xmath146 .
our actual , numerically obtained , value of resistivity for disk formation is one to two orders of magnitude smaller , depending on the magnetic field strength .
the discrepancy indicates that the simple considerations leading to the ohm sphere need to be modified .
one such modification is that the ( radial ) magnetic diffusion speed @xmath147 is expected to be larger than @xmath148 , because the field lines bend sharply inside the pseudodisk , with a field variation scale @xmath149 much less than @xmath67 ( see figs .
[ er18 ] and [ er19 ] ) .
the other is that the infall speed can be significantly below the free - fall value ( see fig .
[ vr_all ] ) .
both modifications make it easier for the field lines to diffuse out , lowering the critical value for disk formation compared with the simple estimate based on the concept of ohm sphere . nevertheless , the critical resistivity for disk formation is still a few orders of magnitude higher than the classical value .
how such an enhanced resistivity may arise is unclear ( see @xcite ) .
one possibility is turbulent resistivity , which has been discussed , for example , in the context of magnetized accretion disks ( e.g. , @xcite ; @xcite ) .
there are at least two possible origins for turbulence on the disk formation scale ( of @xmath150@xmath113 ) .
first , the dense cores that collapse into stars are observed to have non - thermal motions on the tens of thousands of au scale , which can be interpreted as a subsonic turbulence of order half the sound speed @xcite .
it is likely that the turbulent motions would continue down to smaller scales as the core collapses .
second , the collapsing flow can be significantly retarded by magnetic forces ( see , e.g. , the infall speed for model er18 in figure [ vr_all ] )
. the magnetic support of the collapsing material against gravity may lead to interchange instability in 3d , which could develop into a turbulence , and might itself lead to an effective turbulent resistivity . whether turbulent resistivity can enable the formation of rotationally supported disks or not requires high - resolution 3d simulations , perhaps along the line of @xcite .
another possibility is anomalous resistivity associated with , for example , current - driven instabilities @xcite .
whether these and other enhancements of resistivity can enable disk formation remains to be determined . if not , other means of enabling disk formation , such as the hall effect and outflow stripping of the protostellar envelope , need to be explored .
although our idealized calculations indicate that a considerable enhancement in resistivity is needed if ohmic dissipation alone is to enable the formation of rotationally supported disks during the protostellar accretion phase of low - mass star formation , the exact value for the required enhancement is uncertain , and can be better determined with future refinements .
one refinement is to include the self - gravity , which would also allow us to follow the formation and evolution of the dense magnetized core to the point of point - mass formation .
the prestellar core evolution would provide a more self - consistent initial condition for the disk formation calculation during the protostellar accretion phase .
an integrated treatment of both core formation and protostellar accretion may enable us to determine the distribution of the matter surrounding the dense core , which can aid the core material in removing the angular momentum from the forming disk , especially if the core is highly flattened along the field lines ( @xcite ; @xcite ; see e.g. , @xcite for statistical evidence of core flattening ) .
another refinement , which we plan to do , is to study the combined effect of resistivity and ambipolar diffusion on disk formation . in addition
, it would be desirable to find a better way to quantify the numerical reconnection , originated physically from the strong collapse - induced magnetic pinching , that affects the ideal mhd and low resistivity simulations .
we conclude that how protostellar disks form around accreting protostars remains a mystery .
we acknowledge support by the theoretical institute for advanced research in astrophysics ( tiara ) , and by the national science council of taiwan through grant nsc97 - 2112-m-001 - 018-my3 . this work is supported in part by nasa atp ( nng06gj33 g ) and origins grants .
we acknowledge useful conversations with roger blandford , arieh knigl , chris mckee , jon mckinney , nir shaviv , and ron taam . | disk formation in magnetized cloud cores is hindered by magnetic braking .
previous work has shown that for realistic levels of core magnetization , the magnetic field suppresses the formation of rotationally supported disks during the protostellar mass accretion phase of low - mass star formation both in the ideal mhd limit and in the presence of ambipolar diffusion for typical rates of cosmic ray ionization .
additional effects , such as ohmic dissipation , the hall effect , and protostellar outflow , are needed to weaken the magnetic braking and enable the formation of persistent , rotationally supported , protostellar disks . in this paper , we first demonstrate that the classic microscopic resistivity is not large enough to enable disk formation by itself .
we then experiment with a set of enhanced values for the resistivity in the range @xmath0@xmath1 .
we find that a value of order @xmath2 is needed to enable the formation of a @xmath3-scale keplerian disk ; the value depends somewhat on the degree of core magnetization .
the required resistivity is a few orders of magnitude larger than the classic microscopic values . whether it can be achieved naturally during protostellar collapse
remains to be determined . |
band - gaps corresponding to ranges of frequencies at which waves can not propagate earned the interest of the scientific community .
these band - gaps ( bgs ) or stop - bands can be utilized for various applications such as filtering elastic waves , enhancing performance of ultrasonic transducers , supplying a vibrationless ambient when needed for industrial and biomedical applications , and more .
several techniques are available for investigating wave propagation in composites , and the effective medium methods . as in this work
we are interested in the appearance of bgs in infinite periodic laminates , we find it convenient to use the transfer matrix method , along with the bloch - floquet theorem .
specifically , we consider waves superposed on a pre - deformed state due to bias electric field acting on an infinite periodic laminate with a repeating sequence of of dielectric elastomers ( des ) layers .
the motivation for the choice of des as the composite constituents stems from their ability to undergo large deformations , and to change their mechanical and dielectric properties when subjected to electrostatic fields @xcite , thus effecting the way in which electroelastic waves propagate in the matter . as will be shown in the sequel , _ by properly tuning the bias electrostatic field different ranges of frequencies can be filtered out .
_ the structure of this paper is as follows .
following the works of and , the background required for describing the static and dynamic responses of elastic dielectric laminates is revisited in section 2 .
particularly , the governing equations for small fields superimposed on large deformations of electroactive elastomers are summarized . in this section
we also introduce an adequate extension of the transfer matrix method and the bloch - floquet theorem required for treating the propagation of incremental electroelastic waves superimposed on finitely deformed de laminates .
section 3 deals with the response an infinite periodic laminate with alternating layers , whose behaviors are governed by the _ incompressible dielectric neo - hookean _ model ( dh ) , to a fixed electric displacement field normal to the layers plane .
subsequently , the response to small harmonic perturbations superimposed on the finitely deformed laminate is analyzed .
the end result is given in terms of a dispersion relation relating the incremental fields frequency and phase velocity to the bloch - parameter .
several examples are considered in section 4 , where the influence of the morphology of the laminate , the contrast between the phases properties , and the bias electric field on the dispersion relation is examined .
finally , realizable tunable isolators are studied by considering combinations of commercially available des such as vhb 4910 , fluorosilicone 730 and elastosil rt-625 .
consider a heterogeneous body occupying a volume region @xmath2 made out of @xmath3 perfectly bonded different homogeneous phases .
the external boundary of the body @xmath4 separates it from the surrounding space @xmath5 , assumed to be vacuum .
each phase occupies a volume region @xmath6 and its boundary is @xmath7 .
let @xmath8 describe a continuous and twice - differentiable mapping of a material point @xmath9 from the reference configuration of the body to its current configuration @xmath10 with boundary @xmath11 by @xmath12 .
the corresponding velocity and acceleration are denoted , respectively , by @xmath13 and @xmath14 and the deformation gradient is @xmath15 vectors in the neighborhood of @xmath9 are transformed into vectors in the current configuration via @xmath16 .
the volume change of a referential volume element @xmath17 is given by @xmath18 , where @xmath19 is the corresponding volume element in the current configuration , and @xmath20 due to material impenetrability .
the conservation of mass implies that @xmath21 , where @xmath22 and @xmath23 are the material mass densities in the reference and the deformed configurations , respectively .
an area element @xmath24 with the unit normal @xmath25 in the reference configuration is transformed to a deformed area @xmath26 with the unit normal @xmath27 in the current configuration according to nanson s formula @xmath28 .
the right and left cauchy - green strain tensors are @xmath29 and @xmath30 . the electric field in the current configuration , denoted by @xmath31 , is given in terms of a gradient of a scalar field , termed the electrostatic potential .
in free space an induced electric displacement field @xmath32 is related to the electric field via the vacuum permittivity @xmath33 , such that @xmath34 . in dielectric bodies the connection between the fields
is specified by an adequate constitutive relation . in terms of a total stress tensor @xmath35 the equations of motion read @xmath36 where in order to satisfy the balance of angular momentum symmetry of @xmath35 is required .
the terminology total stress is used to indicate that @xmath35 accounts for both mechanical and electrical contributions .
thus , the traction @xmath37 on a deformed area element with a unit normal @xmath38 is given by @xmath39 when ideal dielectrics are considered no free body charge is present , and gauss s law takes the form @xmath40 faraday s law is @xmath41 when a quasi - electrostatic approximation is considered .
a full description of the system should consider interactions with fields outside the body , henceforth denoted by a star superscript . specifically , these are @xmath42,\label{eq : maxwell stress}\end{aligned}\ ] ] where @xmath43 is the identity tensor . in vacuum
the governing eqs . - for @xmath44 and @xmath45 are equivalent to laplace s equation for the electrostatic potential . as a consequence ,
@xmath46 known as the maxwell stress , is divergence - free . across the outer boundary @xmath11 ,
the electric jump conditions are @xmath47 where @xmath48 is the surface charge density . in order to formulate the jump in the stress we postulate a separation of the traction into a sum of a prescribed mechanical traction @xmath49 , and an electric traction @xmath50 induced by the external stress such that @xmath51 .
hence , the stress boundary condition is @xmath52 the jump conditions across a charge - free internal boundary between two adjacent phases @xmath53 and @xmath54 are @xmath55 where @xmath56 denotes the jump of fields between the two phases .
a lagrangian formulation of the governing equations and jump conditions is feasible by a _
pull - back _ of the eulerian fields @xmath57 for the total first piola - kirchhoff stress , lagrangian electric displacement and electric field , respectively .
thus , the governing eqs . , and transform , respectively , to @xmath58 the corresponding referential jump conditions across the external boundary @xmath4 are @xmath59 where @xmath60 and @xmath61 . across the referential interfaces eqs .
become @xmath62 following , the constitutive relation is expressed in terms of an _ augmented energy density function _ ( aedf ) @xmath63 with the independent variables @xmath64 and @xmath65 , such that the total first piola - kirchhoff stress and the lagrangian electric field are derived via @xmath66 the first of eq .
( [ eq : constitutive law ] ) should be modified when considering incompressible materials , as the kinematic constraint yields an additional workless part of stress .
the latter is accounted by introducing a lagrange multiplier @xmath67 which can determined only from the equilibrium equations together with the boundary conditions .
thus , the total first piola - kirchhoff is @xmath68 based on the framework developed in , we superimpose an infinitesimal time - dependent elastic and electric displacements @xmath69 and @xmath70 , on the pre - deformed configuration
. herein , and in the sequel , we use a superposed dot to denote incremental quantities .
let the eulerian quantities @xmath71 denote the _ push - forwards _ of increments in the first piola - kirchhoff stress , the lagrangian electric displacement and electric fields , respectively , namely @xmath72 in terms of these variables , the incremental governing equations read @xmath73 for an incompressible material the incremental constraint is @xmath74 where @xmath75 is the displacement gradient .
linearization of the incompressible material constitutive equations in the increments yields @xmath76 where @xmath77 .
the quantities @xmath78 and @xmath79 are the push - forwards of the referential reciprocal dielectric tensor , electroelastic coupling tensor , and elasticity tensor , respectively . in components
form @xmath80 where@xmath81the incremental outer fields are given by @xmath82.\label{eq : inc maxwell stress}\end{aligned}\ ] ] here again @xmath83 and @xmath84 are to satisfy eqs
. - , where subsequently @xmath85 .
the push - forward of the increments of eqs . gives the following incremental jump conditions across the external boundary in the current configuration @xmath86\normal & = & \check{\traction}_{m},\label{eq : in stress jump}\\ \left[\edlincpush-\inc{\outer{\ed}}-\left(\divg\inc{\sys}\right)\outer{\ed}+\disgrad\outer{\ed}\right]\cdot\normal & = & -\check{w}_{e},\label{eq : in charge jump}\\ \left[\mathrm{\efincpush}-\inc{\outer{\ef}}-\trans{\disgrad}\outer{\ef}\right]\times\normal & = & \mathbf{0},\label{eq : inc elec jump}\end{aligned}\ ] ] where @xmath87 and @xmath88 .
similarly , the push forward of the increments of eqs . results in the following jumps across the internal boundaries @xmath89 several techniques are available in the literature for tackling the problem of wave propagation in periodic composites . in layered structures
it is convenient to utilize the transfer - matrix method in conjunction with the bloch - floquet theorem @xcite .
herein we formulate an adequate adjustment for treating the propagation of incremental electroelastic waves superimposed on a finitely deformed multiphase de laminates .
-layers unit - cells . ]
consider a finitely deformed infinite de - laminate made out of @xmath53-phase periodic unit - cells subjected to incremental perturbations harmonic in time and normal to the layers plane .
a schematic illustration of the laminate is shown in fig .
let @xmath90 , @xmath91 , denote an incremental state vector in the @xmath92-phase of the @xmath93 cell consisting quantities which are continuous across the interface between the two adjacent layers .
in view of the linearity of the incremental problem it can be shown that the state vector at the top of the @xmath92-phase @xmath94 and the state vector at the bottom of the @xmath92-phase @xmath95 ( see fig .
[ tmm ] ) are related via a unique non - singular transfer matrix @xmath96 such that @xmath97 in terms of the state vectors , the continuity conditions across the interface are @xmath98 utilizing eqs . - recursively the relation between @xmath99 and @xmath100 is @xmath101 where @xmath102 is the total transfer matrix .
at one hand the continuity conditions across the interface between two successive cells @xmath103 and @xmath104 are @xmath105 at the other , the bloch - floquet s theorem states that the state - vectors of the same phase in adjacent cells are identical up to a phase shift in the form @xmath106 where @xmath107 is the bloch - floquet wavenumber in first irreducible brillouin zone , representing the smallest region where wave propagation is unique @xcite .
substitution of eq . in , followed by employing the latter with eq .
, yields the eigenvalue problem @xmath108 where @xmath109 is an identity matrix with the dimension of the total transfer matrix .
the solution of eq . provides the dispersion relation describing the manner by which waves propagate in the laminate .
henceforth we restrict our attention to isotropic phases .
consequently , the phases aedfs can be written in terms of the six invariants @xmath110 specifically , we consider phases which are characterized by _
incompressible dielectric neo - hookean _
( dh ) model @xmath111 where @xmath112 and @xmath113 denote the phases shear moduli and dielectric constants , respectively , and @xmath114 being the phases relative dielectric constant . the corresponding total stress is @xmath115 and the current displacement and electric field are related via the isotropic linear relation @xmath116 calculation of the associated constitutive tensors @xmath117 yields , in components form , @xmath118 where @xmath119 are the components of the kronecker delta .
consider a cartesian coordinate system with unit vectors @xmath120 and @xmath121 along @xmath122 and @xmath123 , respectively .
let a de laminate be composed of periodic two - phase unit - cells with thickness @xmath124 along the the @xmath125 direction , infinite along the @xmath126 and @xmath123 directions as illustrated in fig .
[ laminate ] . for convenience
we denote the phases @xmath127 and @xmath128 .
the sum of the phases thicknesses equals the unit - cell thickness @xmath129 , such that @xmath130 and @xmath131 are the volume fractions of phase @xmath127 and phase @xmath128 , respectively .
we assume that the composite is allowed to expand freely in the @xmath126 and @xmath123 directions .
of interest is the composite response when subjected to an overall electric displacement field @xmath132 , where @xmath133 . in each phase
we assume displacement fields compatible with the homogeneous uni - modular ( on account of incompressibility ) diagonal deformation gradients @xmath134}=\mathrm{diag}\left[\phs{\lambda_{1}},\phs{\lambda_{2}},\left(\phs{\lambda_{1}}\phs{\lambda_{2}}\right)^{-1}\right],\label{eq : defromation gradient p f , m}\ ] ] where henceforth @xmath135 . in virtue of the perfect bonding between the phases ,
the stretch ratios in the @xmath126 and @xmath123 directions in the phases are the same , i.e. , @xmath136 eqs .
imply that @xmath137 , and henceforth we define @xmath138 . from the symmetry of the problem in the @xmath139
it follows that @xmath140 .
the corresponding stresses in each phase are @xmath141 and @xmath128 phases , with initial thicknesses @xmath142 and @xmath143 , respectively . the electric bias field is along the @xmath125 direction . ]
we assume mechanical traction - free boundary conditions at infinity . utilizing the jump conditions in the first of ( [ eq : jumps ] ) and , along with the traction - free boundary conditions we have that @xmath144 from the latter together with the second of eq .
( [ eq : stress components bias ] ) the expressions for the pressure in the phases are @xmath145 substitution of eq . into the first of eq .
( [ eq : stress components bias ] ) leads to a relation between the applied lagrangian electric displacement field @xmath146 and the resultant stretch ratio , namely @xmath147 where @xmath148 and @xmath149
. the latter appropriately reduces to the corresponding relation for a homogeneous specimen . in terms of the dimensionless quantities @xmath150 , @xmath151 , and @xmath152 , denoting the shear contrast , permittivity contrast , and normalized lagrangian electric displacement field , respectively ,
the stretch ratio is @xmath153 for the above choice of constitutive relation , we also have that eq .
reduces to @xmath154 the response of the laminate to a harmonic excitation superimposed on the aforementioned finite deformation is addressed next .
let @xmath155 and @xmath156 denote , respectively , the components of the incremental displacement and the incremental electric potential in the phases , such that @xmath157 . the fields sought are to satisfy the incremental equations of motion along with gauss equation ( [ eq : inc eom gauss farady current conf ] ) and the incompressibility constraint . following @xcite , we consider fields that are functions of @xmath125 and time alone , independent of @xmath126 and @xmath123 , compatible with the simple - thickness mode @xcite .
consequently , in each phase the governing equations ( [ eq : inc eom gauss farady current conf ] ) and simplify to @xmath158 and @xmath159 respectively , where @xmath160 , @xmath161 .
substitution of eq . into eq .
reveals that @xmath162 , hence the incremental electric potential is linear in @xmath125 , say @xmath163 , where @xmath164 and @xmath165 are integration constants .
more significantly , it implies that the incremental electrical and mechanical fields are not coupled , as eq .
remains a function of @xmath166 alone , independent of @xmath167 , together with eq . which remains a function of @xmath167 alone , independent of @xmath166 .
states that along the thickness the incremental displacement is spatially constant . together with eqs .
and the solution for the incremental problem is @xmath168 where @xmath169 , @xmath170 , @xmath171 is the angular excitation frequency , and @xmath172 are integration constants . without loss of generality
the constants @xmath173 can be set to be zero .
the motion described by the solution eq . corresponds to displacement gradients which retain only shear components . for completeness
, the solution for the incremental pressure emerges from eq . , which reduces to @xmath174 and implies that @xmath175 , where @xmath176 are integration constants .
the incremental state vector of the @xmath93 cell takes the form @xmath177 the components of transfer matrix @xmath178 relating the top @xmath179 and bottom @xmath180 state vectors are given in the appendix .
note that the continuity conditions for @xmath181 is identically satisfied when we set @xmath173 to zero .
the corresponding solution of the eigenvalue problem eq .
resulting from the bloch - floquet theorem can be expressed in terms of the trigonometric expression @xmath182 where @xmath183 , \label{eq : dispersion sol of eta}\ ] ] with @xmath184 and @xmath185 .
is an extension of the solution given in to the class of finitely deformed infinite periodic de laminates .
the dependency on the bias field becomes evident when @xmath186 is rephrased in terms of the electric bias field , the referential geometry and the frequency , namely @xmath187 .
\label{eq : eta of d}\end{aligned}\ ] ] in virtue of the uncoupled nature of the incremental problem , the incremental electric quantities indeed do not appear in eq . .
we conclude this section noting that waves of certain frequencies can not propagate when the pertained solution of @xmath188 is complex .
these cases will be examined in the following section .
as functions of the normalized frequency @xmath189 , for ( a ) @xmath190 , ( b ) @xmath191 and ( c ) @xmath192 .
the red , blue and black curves correspond to @xmath193 and 0.8 , respectively . ] at @xmath194 , for ( a ) @xmath190 and ( b ) @xmath191 and ( c ) @xmath192 . ] with the help of a few examples we explore the dispersion relation characterizing the dynamic response of the composite .
we examine its dependency on the phases volume fractions and the contrast between the phases shear and dielectric moduli .
in particular we are interested in the influence of the electric bias field on the dynamic response of the composite . to this end
we make the following choice for the properties of phase @xmath128 @xmath195 , & \mt{\mu}=200\left[kpa\right ] , & \mt{\relper}=2,\end{aligned}\ ] ] and determine dispersion diagrams for various choices of phase @xmath127 properties and volume fractions at different values of the dimensionless bias field @xmath196 .
in all the forthcoming examples @xmath197 is assumed .
[ dispersion of alpha ] shows variations of the bloch wavenumber @xmath198 as functions of the normalized frequency @xmath189 where @xmath199 , @xmath200 , for a representative value of the bias field @xmath194
[ dispersion of alpha]a , b and c correspond to @xmath201 and 20 , respectively .
the red , blue and black curves correspond to @xmath202 and 0.8 , respectively .
the period of the dispersion curves becomes smaller with an increase of the softer phase volume fraction .
an inverse effect is revealed when the contrast is enhanced , that is longer periods for higher values of @xmath203 and @xmath204 .
[ bg of vm ] displays the regions of prohibited normalized frequencies @xmath205 as functions of the matrix volume fraction at a fixed value of the bias field @xmath194 .
[ bg of vm]a , b and c correspond to @xmath190 , @xmath206 and @xmath207 , respectively .
herein and henceforth the band - gaps are the blue regions in the plots .
we observe how an increase of the softer phase volume fraction results in appearance of additional thinner bands .
conversely , when the contrasts are increased , that is higher values of @xmath203 and @xmath204 , there are fewer thicker bands .
the latter is evident in fig .
[ bg of alpha ] which displays the prohibited frequencies as functions of @xmath203 for @xmath208 and @xmath194 , with different values of @xmath204 . specifically , fig .
[ bg of alpha]a corresponds to @xmath209 and fig .
[ bg of alpha]b corresponds to @xmath210 .
similar trends are observed when the dielectric contrast is investigated .
[ bg of b ] displays the prohibited frequencies @xmath205 as functions of @xmath204 for @xmath208 , with different values of @xmath203 .
[ bg of b]a corresponds to @xmath211 and fig .
[ bg of b]b to @xmath212 .
[ bg of alpha ] and [ bg of b ] reveal that whenever the phases have non - monotonous ratios of shear moduli and dielectric contrasts , i.e. @xmath213 and @xmath214 or @xmath215 and @xmath214 , the band gaps become thiner and denser . at @xmath194 and @xmath208 , for ( a ) @xmath209 and ( b ) @xmath210
] at @xmath194 and @xmath208 , for ( a ) @xmath211 and ( b ) @xmath212 . ]
[ bg of d ] illustrates the influence of the bias field by examining the variations of the bgs as functions of the normalized parameter @xmath196 .
values of @xmath216 and @xmath190 were chosen in fig .
[ bg of d ] a , @xmath208 and @xmath191 in fig . [ bg of d ] b , and @xmath217 and @xmath192 in fig .
[ bg of d ] c. the influence of the electric field is evident as the bgs are shifted toward lower frequencies when the field is increased .
this phenomena suggests that _ control of bgs and filtering of desired frequencies are feasible by activating de laminates with adequate bias electric fields . _ for ( a ) @xmath218 , ( b ) @xmath219 and ( c ) @xmath220 .
] finally , we consider realizable laminates made out of commercially available des .
we examine three materials : vhb-4910 by 3 m , elastosil rt-625 by wacker , and fluorosilicone 730 by dow corning .
approximate physical properties of these elastomers , as reported in the literature , are summarized in table 1 .
we investigate different combinations of the two - phase laminate and assume that @xmath221 .
.physical properties of commercially available des . [ cols="^,^,^,^ " , ] as functions of the normalized frequency @xmath189 for combinations of vhb-4910 with ( a ) elastosil rt-625 , and with ( b ) fluorosilicone 730 . ] for combinations of vhb-4910 with ( a ) elastosil rt-625 , and ( b ) fluorosilicone 730 . ]
[ dispersion ] displays the dispersion relation in terms of bloch wavenumber @xmath198 as a function of the normalized frequency @xmath222 .
[ dispersion]a corresponds to a combination of vhb-4910 and elastosil rt-625 , while fig .
[ dispersion]b to vhb-4910 and fluorosilicone 730 . in terms of the dimensionless parameters in fig .
[ dispersion]a @xmath223 and @xmath224 , and in fig .
[ dispersion]b @xmath225 and @xmath226 .
the blue , black and red curves correspond to @xmath227 and 1.5 , respectively .
[ bg ] shows the prohibited frequencies as functions of the bias field in terms of the dimensionless bias field @xmath196 .
again , fig . [ bg]a corresponds to the combination of vhb-4910 and elastosil rt-625 , and fig.[bg]b c to vhb-4910 and fluorosilicone 730 .
wider bands are depicted in fig .
[ bg]b in comparison with fig .
[ bg]a , since the latter corresponds to a laminate with higher shear contrast between the phases .
motivated by the ability of des to undergo large deformations and change their mechanical and electrical properties when electrostatically excited , the feasibility of inducing and controlling bgs by the electrical bias field is explored . to this end
we first considered the static response of an infinite periodic laminate to a bias electric displacement field .
subsequently , the propagation of small waves superposed on the pre - strained composite is addressed .
applications of the transfer matrix method along with the bloch - floquet theorem resulted in a dispersion relation between the waves frequencies , the phase velocities , and the bloch - parameter . numerical analysis of these relations was conducted for specific two - phase laminates in order to examine how the stop - bands , associated with complex bloch - parameter , vary as functions of the geometrical , mechanical and electric properties of the laminate .
most importantly , this analysis reveals the influence of the electrostatic bias fields on the bgs .
our findings demonstrate how when the concentration of the softer phase in increased , the bands become thinner and additional band gaps the appear .
an inverse effect is observed when the contrast between the phases properties is increased ( larger values of @xmath203 and @xmath204 ) .
the primary conclusion of our analysis is depicted in fig .
[ bg of d ] , demonstrating how variations in the bias electric displacement field lead to modifications in the width and shifts in the range of the prohibited frequencies .
we note that for lower values of @xmath203 and @xmath204 the effect of the bias electric field becomes more evident in the sense that the shift in the bgs range is larger . from practical viewpoint point , while stopping larger bands of frequencies is feasible by choosing a higher contrast , lower contrasts will allow to actively adjust the desired band at a higher precision .
the analysis was concluded with two examples in which we chose phases properties to be identical to those of commercially available des .
the results show how tunable stop - bands are achievable by properly adjusting the electrostatic bias field .
the components of the transfer matrices @xmath228 appearing in section 3 : @xmath229 with all the remaining components being zero .
debotton , g. , tevet - deree , l. and socolsky , e. a. ( 2007 ) .
electroactive heterogeneous polymers : analysis and applications to laminated composites , _ mechanics of advanced materials and structures _ * 14 * : 1322 .
kornbluh , r. and pelrine , r. ( 2008 ) .
high - performance acrylic and silicone elastomers , _ in _
f. carpi , d. de rossi , r. kornbluh , r. pelrine and p. sommer - larsen ( eds ) , _ dielectric elastomers as electromechanical transducers _ , elsevier , oxford , uk , chapter 4 , pp .
3342 .
ponte castaeda , p. and siboni , m. h. ( 2011 ) .
a finite - strain constitutive theory for electro - active polymer composites via homogenization , _ int .
j. nonlinear mech .
_ p. ( 10.1016/j.ijnonlinmec.2011.06.012 ) .
shmuel , g. , gei , m. and debotton , g. ( 2011 ) .
the rayleigh - lamb wave propagation in dielectric elastomer layers subjected to large deformations , _ int .
j. nonlinear mech .
_ p. ( doi:10.1016/j.ijnonlinmec.2011.06.013 ) .
wang , y. and auld , b. a. ( 1986 ) .
numerical analysis of bloch theory for acoustic wave propagation in one - dimensional periodic composites , _ applications of ferroelectrics . 1986 sixth ieee international symposium on _ , pp . | the thickness vibrations of a finitely deformed infinite periodic laminate made out of two layers of dielectric elastomers is studied .
the laminate is pre - stretched by inducing a bias electric field perpendicular the the layers .
incremental time - harmonic fields superimposed on the initial finite deformation are considered next . utilizing the bloch - floquet theorem along with the transfer matrix method we determine the dispersion relation which relates the incremental fields frequency and the phase velocity .
ranges of frequencies at which waves can not propagate are identified whenever the bloch - parameter is complex . these _ _ band - gap__s depend on the phases properties , their volume fraction , and most importantly on the electric bias field .
our analysis reveals how these band - gaps can be shifted and their width can be modified by changing the bias electric field .
this implies that by controlling the electrostatic bias field desired frequencies can be filtered out .
representative examples of laminates with different combinations of commercially available dielectric elastomers are examined .
_ keywords _ : dielectric elastomers ; wave propagation ; finite deformations ; thickness vibrations ; non - linear electroelasticity , band - gap , composite , bloch - floquet analysis .
# 1f_#1 # 1c_#1 # 1_#1 # 1b_#1 # 1f_#1 ^ -1 # 1_#1 # 1d_#1 # 1d_#1 # 1e_#1 # 1#1^ # 1 # 1#2 # 1(#1 ) # 1i_#1 # 1 # 1_#1 # 1_#1 # 1_#1 # 1h_#1 # 1#1^ # 1#1^(p ) # 1_#1 # 1_#1 # 1_#1 # 1#1^(p ) # 1_#1 # 1_#1 # 1#1^(a ) # 1#1^(b ) # 1#1^(p ) @xmath0department of mechanical engineering ben - gurion university beer - sheva , 84105 israel 0.2 in @xmath1department of biomedical engineering ben - gurion university beer - sheva , 84105 israel |
in 1993 shuryak and verbaarschot realized @xcite that a certain chiral random matrix theory ( rmt ) is equivalent to the low - energy limit of the qcd partition function @xcite and thus rmt , a tool that had shown merit in such diverse areas of physics as atomic physics , solid state physics , and nuclear physics , entered the realm of quantum chromodynamics ( see @xcite for an extensive review of the many applications of rmt and @xcite for a review of the application of rmt to qcd ) .
there is an amount of arbitrariness in the rmt of qcd since the potential @xmath3 that enters in the rmt partition function is arbitrary .
it was shown @xcite that when @xmath4 then rmt and qcd are equivalent in the mesoscopic scaling region , also called the ( double-)microscopic limit .
thus , if we want rmt to provide us with information about qcd in the microscopic limit we have to consider spectral correlators that are independent of @xmath3 in this limit .
if this is the case the correlators are said to be universal .
clearly , the question of universality is crucial if one wants to extract information about qcd from rmt .
many papers have been written about the connection between rmt and 4-dimensional qcd ( qcd@xmath5 ) with an arbitrary number of flavors as well as 3-dimensional qcd ( qcd@xmath0 ) with an even number of flavors ( see references in @xcite ) . however , for an odd number @xmath6 of flavors the situation is less understood and only treated in @xcite .
the reasons for this lack of balance are many .
the rmt partition function is @xmath7 where we integrate over @xmath8 hermitian matrices @xmath9 with the haar measure .
first of all , the effective field theory partition function , equivalent to ( [ zrmt ] ) when @xmath10 , depends on whether @xmath11 is even or odd @xcite : @xmath12 \ \ \ \ ( n{\textrm { even } } ) \\
\label{oddpartfctnodd } { \mathcal{z}}^{(2\alpha+1)}_n & = & \int { \textrm{d } } u \sinh \ ! \left [ { \textrm{tr } } ( n \mathcal{m } u \gamma_5 u^{\dag } ) \right]\ \ \ \ \ ( n{\textrm { odd}})\end{aligned}\ ] ] both integrals are over su@xmath13 , @xmath14 is the mass matrix , and @xmath15 .
the @xmath11-dependence is most peculiar .
furthermore , the rmt spectral density @xmath16 obeys the relation @xcite @xmath17 which implies that for odd @xmath11 the spectral density is an odd function and thus not positive definite . finally , the standard method for calculating spectral information in random matrix theories , the method of orthogonal polynomials , is not available for an odd number of flavors due to the fact that it is impossible to define orthogonal polynomials with respect to an odd measure on an interval symmetric about the origin .
in this paper we concentrate on the case of even @xmath11 and an odd number of flavors @xmath6 . in section [ sec : symmetries ] we propose a parity transformation for an odd number of flavors that leaves the qcd@xmath0-lagrangean invariant .
this is an essential point since the rmt of qcd@xmath0 has parity symmetry as an input , and the conclusion is that an odd number of masses have to be zero for the lagrangean to be invariant ( at the classical level ) . in section [ sec : universality ]
we introduce pseudo - orthogonal polynomials and calculate the spectral correlation functions in the microscopic limit for even @xmath11 .
it turns out that for an odd number of flavors the correlators can be expressed using the kernel from the chiral unitary ensemble , which is equivalent to four - dimensional qcd , when we equate the number of chiral fermions to @xmath18 and by analytic continuation work in the sector of topological charge @xmath2 . in section
@xmath19 we use pseudo - orthogonal polynomials to express the rmt kernel in terms of field theory finite - volume partition functions . by looking at the result from @xcite for the finite - volume partition function
we arrive at the conclusion that when one mass is zero , as is necessary for symmetry reasons , the @xmath11-odd field theory partition function is zero .
the lesson is that the odd-@xmath11 partition function should not be regarded as defining a theory in itself , but only as a function that is necessary in order to extract spectral information in the even-@xmath11 theory .
this is consistent with the fact that the spectral density does not make sense when @xmath11 is odd due to ( [ rho ] ) .
our starting point is the qcd@xmath0-lagrangean with @xmath20 flavors @xmath21 : @xmath22 where @xmath23 is the gluonic field strength and @xmath24 are the eigenvalues of the mass matrix . in three dimensions
we do not have chirality at our disposal because @xmath25 .
the lowest dimensional representation of the @xmath26-matrices is the pauli - matrices and the corresponding fields are two - spinors .
this has the disadvantage that the mass terms are parity @xmath27 is a rotation .
we choose ( somewhat arbitrarily ) to let @xmath28 denote inversion around the 1-axis . ]
if all masses are zero then @xmath29 is invariant under @xmath28 since the dirac - term in the lagrangean respects the parity transformation , but in general the masses are non - zero and @xmath29 is not invariant . for reasons that will become apparent
later we distinguish between an even and an odd number of flavors .
in @xcite it is argued that for an even number @xmath30 of flavors we can achieve parity invariance by using a four - dimensional representation of the @xmath26-matrices . the representation employed ( minkowski space ) is @xmath31 we see that for a four - spinor @xmath32 the mass term is @xmath33 the last expression is identical to @xmath34 in a two - dimensional representation of the @xmath26-matrices with @xmath35 . if the mass matrix @xmath36 is chosen to be @xmath37 it follows that the lagrangean ( [ l_qcd ] ) is invariant under parity when the transformation is defined as @xmath38 for an odd number @xmath6 of flavors we write the lagrangean as @xmath39 the @xmath26-matrices are represented by the pauli - matrices : @xmath40 above we used the grouping of the spinors @xmath41 into four - spinors and the @xmath42-matrices ( [ representation ] ) as @xmath26-matrices to obtain a parity invariant system when assigning masses of opposite sign to the constituents of the four - spinors . here
we use that if @xmath43 is a representation of the dirac algebra then @xmath44 is an inequivalent representation of the algebra @xcite .
this allows us to use negative masses also for an odd number of flavors , where we can not group the spinors as in ( [ lmass ] ) .
as far as @xmath45 and @xmath46 are concerned , we use the transformation defined in ( [ parity ] ) . for the remaining two - spinor @xmath47
we employ the transformation @xmath48 since a mass term changes sign under this transformation the overall symmetry is retained only when @xmath49 .
this is also the case if we had used the representation ( [ lonegamma ] ) with opposite signs on all the matrices .
provided that we want to study an invariant lagrangean we are therefore confined to consider an _
odd _ number of massless fermions for odd @xmath20 .
notice , that if we want to have an optimal amount of discrete symmetry in the theory , with as many non - zero mass terms as possible , there is no alternative to the choice @xmath50 ( with @xmath49 ) of the mass matrix .
we have an odd number of two - spinors that have to transform among each other .
this means that the transformed fields are permutations of the original fields ( multiplied by a pauli matrix such that mass terms change signs ) .
unless the fields swap positions two and two their masses have to be zero in order to have invariance , and when two fields interchange positions the @xmath51 assignment of the masses make sure that the mass terms are invariant .
this automatically leaves one field that has to transform into itself , thus giving rise to a symmetry breaking mass term unless @xmath49 .
having discussed the choice of mass matrix we now turn to the rmt of odd - flavored qcd@xmath0 .
the qcd@xmath0 rmt partition function for an odd number of flavors in the eigenvalue representation of the unitary ensemble is @xcite @xmath52 where @xmath53 is the vandermonde determinant .
notice , that it is precisely the choice @xmath49 , which we now make , that makes the partition function real .
with @xmath54 the @xmath55-point correlation function is given by @xmath56 the orthodox way to calculate these correlators would be to find orthogonal polynomials @xmath57 with respect to the weight function @xmath58 introduce the kernel @xmath59 where @xmath60 is the norm of @xmath61 , and then use the expression @xmath62 to determine the @xmath55-point correlation functions .
this method can not be used here since @xmath63 is odd and the integration interval is even , and thus orthogonal polynomials are not defined
. inspired by @xcite we will now introduce a method based on what we will call pseudo - orthogonal polynomials that can be applied to any odd measure on an even interval as long as @xmath11 is even . consider for simplicity the massless case where @xmath64 .
for the chiral unitary ensemble ( chue , with @xmath1 ) the weight function for @xmath65 massless fermions in the sector of topological charge @xmath66 is @xmath67 and we denote the associated orthogonal polynomials @xmath68 with norm @xmath69 . now put @xmath70 , @xmath2 , and define the pseudo - orthogonal polynomials by @xmath71 these new polynomials are pseudo - orthogonal with respect to the weight function @xmath64 on the real line in the sense that @xmath72 where the brackets indicate integration .
the proof is straightforward .
the chue - measure for @xmath65 massive flavors is @xmath73 the measure is valid in the sector of vanishing topological charge . in the sector of charge @xmath74
we obtain the correct measure by adding @xmath75 massive flavors and then letting the masses vanish .
we see that in the sector where @xmath2 and @xmath70 the measure is @xmath76 simple calculations shows that the pseudo - orthogonal polynomials with respect to ( [ uemassive ] ) are given by the polynomials orthogonal with respect to ( [ chuemassive ] ) by the definitions ( [ def1 ] ) and ( [ def2 ] ) .
because the qcd@xmath0 masses are pairwise assigned with opposite sign we may identify the @xmath77th chiral mass with the positive mass of the @xmath77th qcd@xmath0 fermion .
notice that in the matrix representation of chue half - integer values of @xmath66 are not permitted since the matrices integrated over are @xmath78 @xcite . in the eigenvalue representation , however , we can without problems make an analytic continuation such that @xmath2 makes sense , at least from a formal point of view .
we are now ready to calculate the microscopic spectral correlators of odd - flavored qcd@xmath0 , _ i.e. _ , the correlation functions taken in the limit @xmath79 , @xmath80 , @xmath81 where @xmath82 are kept constant .
we always assume that @xmath11 is even .
@xmath83 defined in ( [ vandermonde ] ) can be expressed as the determinant of an @xmath8 matrix : @xmath84 consider the @xmath8 matrix @xmath85 the crucial observation is that @xmath86 is related to @xmath83 .
performing on @xmath87 the row operations @xmath88 followed by @xmath89 we end up with @xmath90 where we have discarded a factor coming from the row operations .
the determinant of an @xmath8 matrix @xmath91 is defined as @xmath92 from which it follows that @xmath93 , or @xmath94 now notice that @xmath95 looking at ( [ vdmlike ] ) we see that we are in a position where we can substitute the entries in @xmath96 with the pseudo - orthogonal polynomials ( properly normalized ) . because of pseudo - orthogonality we may then write , as in the case of orthogonal polynomials , @xmath97 where the @xmath98 s are pseudo - orthogonal . with @xmath99 we obtain @xmath100 we have discarded irrelevant overall factors .
the introduction of pseudo - orthogonal polynomials have led us to an expression for the kernel that involves orthogonal polynomials for which the microscopic limit is universal @xcite : @xmath101 by the christoffel - darboux formula the sum over the polynomials is proportional to @xmath102/(\lambda^2-{\lambda'}^2)$ ] and since @xmath103 the @xmath79 limit corresponds to the @xmath104 limit .
now compare ( [ qcd3kernel ] ) with the expression for the chue kernel @xcite @xmath105 in the sector of zero topological charge .
to go to the sector of charge @xmath66 we can , by flavor - topology duality , take the @xmath106 kernel with @xmath107 massive flavors and then let @xmath75 masses vanish .
the sector with @xmath2 is then by continuation given by ( [ chuekernel ] ) multiplied by a factor @xmath108 .
we arrive at the following master formula for odd - flavored qcd@xmath0 : @xmath109 the formula implies that all properties of the chue that are universal similarly are universal in odd - flavored qcd@xmath0 and especially that all microscopic correlators are universal .
we see that in the case of the microscopic spectral density , where we have to evaluate @xmath110 , the odd - flavored qcd@xmath0 kernel and the chue kernel agree up to a proportionality constant .
they do not , however , agree for higher - order correlators .
the chue kernel is given by @xcite @xmath111 + 1 } \sqrt{\zeta_1\zeta_2}}{(\zeta_1 ^ 2-\zeta_2 ^ 2 ) \prod_f \sqrt{(\zeta_1 ^ 2+\mu_f^2)(\zeta_2 ^ 2+\mu_f^2 ) } } \frac{\det \mathcal{b}}{\det \mathcal{a } } ~,\ ] ] where @xmath112 is a normalization constant , and where @xmath113 is an @xmath114 matrix and @xmath115 is an @xmath116 matrix defined as @xmath117 in qcd@xmath0 the hitherto only known microscopic spectral correlator is the massless microscopic spectral density @xcite @xmath118\ ] ] which coincides with the massless chue microscopic spectral density @xmath119\ ] ] precisely when @xmath70 and @xmath2 . from ( [ kernelrelation ] )
it directly follows that the microscopic spectral density obeys a series of decoupling relations @xcite : @xmath120 these relations tell us that when we make two quarks with equal masses ( up to a sign ) infinitely heavy the spectral density of the new system becomes that of the system consisting of the remaining quarks .
this is illustrated in figure [ densities ] which shows the massless microscopic spectral density ( msd ) for one and three massless flavors , and for one massless and two massive flavors .
the massive msd is calculated using the master formula ( [ kernelrelation ] ) .
when the masses vanish the massive msd will become that of three massless fermions while the act of making the masses infinitely heavy will make the massive msd become that of one massless flavor .
the newly acquired universality of the microscopic spectral correlators will now allow us to express the odd - flavored rmt kernel for even @xmath11 in terms of the field theory finite - volume partition functions of qcd@xmath0 , as it was done in @xcite for the other ensembles .
having done this we go on to simplify the expression for the finite - volume partition functions considerably when we take the conditions for parity invariance into account . the main ingredient in the proof is a result from @xcite that in our case says that @xmath121 } \sqrt{\lambda \lambda ' } \prod_{f=1}^{\alpha } \sqrt{(\lambda^2+m^2_f)({\lambda'}^2+m^2_f ) } \nonumber \\ \times \int { \textrm{d}}^{{(n-1)}^2 } m e^{-n{\textrm{tr}}v(m^2 ) } \det(m ) \prod_{f=1}^{\alpha } \det(m^2+m_f^2 ) \det(\lambda - m ) \det(\lambda'-m ) ~.\end{aligned}\ ] ] the integral appearing in the second line is the rmt partition function in the case where the square hermitian matrices are of odd dimensions , and with two extra imaginary masses @xmath122 and @xmath123 .
the proof goes as in @xcite : @xmath124 where @xmath125 is the rmt measure .
rewriting in terms of eigenvalue variables we get @xmath126 where @xmath127 is given by ( [ uemassive ] ) .
now absorb the factors of @xmath128 and @xmath129 into the vandermonde determinants : @xmath130 these new vandermonde determinants are @xmath8 , @xmath11 even , and thus we have the method of pseudo - orthogonal polynomials at our disposal .
@xmath131 where the @xmath132 are the pseudo - orthogonal polynomials corresponding to @xmath127 .
note that the proof does not carry through when @xmath11 is odd .
now , because of universality the rmt partition function @xmath133 is equal ( up to a constant ) to the field theory finite - volume partition function @xmath134 and thus @xmath135 in formula ( [ 2oddkernel ] ) @xmath14 now has the microscopic masses as entries , and the partition functions have been labelled with @xmath11 and @xmath136 implicitly through @xmath137 .
notice that we have already taken the double - microscopic limit - the label is important because the partition functions depend on @xmath11 in a crucial manner .
we have @xcite @xmath138 ~.\ ] ] in the expression @xmath139 is a vandermonde determinant of the masses and @xmath140 is an @xmath141 matrix of the form ( @xmath142 , @xmath143 @xmath144 formula ( [ z_n ] ) can be simplified considerably when @xmath145 .
take the matrix @xmath146 and interchange @xmath147 such that the rows @xmath148 to @xmath149 become positioned right below the @xmath150-dependent row . because the interchange of two rows amounts to an overall sign change of the determinant we see that @xmath151 because of the @xmath51 symmetry of the masses and the block structure of ( [ d1])-([d2 ] ) .
thus @xmath152 \nonumber \\ & = & \left\ { \begin{array}{ll } 2 \det\mathbf{d}(\mu=0,\{\mu_k\ } ) & { \textrm{if $ n$ is even } } \\ 0 & { \textrm{if $ n$ is odd } } ~. \end{array } \right . \end{aligned}\
] ] besides the simplification this expression tells us that we should not look upon the odd-@xmath11 partition function as defining a theory in itself , but only as a function that enters in the expressions of the even-@xmath11 theory .
note that when @xmath2 the chue kernel ( [ chiralkernel ] ) becomes complex and thus the normalization coefficient @xmath153}\ ] ] becomes complex also .
we have explicitly checked that ( [ 2oddkernel ] ) gives the expected results in a few cases , and in each case the normalization constants had to be chosen complex .
finally , we mention that the decoupling relations ( [ decoupl ] ) can be explicitly derived from ( [ 2oddkernel ] ) .
there are two major points to take away from this section .
first , the fact that for even @xmath11 we are able to rewrite the rmt kernel in terms of finite - volume partition functions while the proof can not be applied to the case of odd @xmath11 .
second , the fact the finite - volume partition function for odd @xmath11 vanishes when the only input is the mass assignment that guarantees parity invariance , a crucial point when the rmt is constructed . from this
we conclude that when extracting spectral information about odd - flavored qcd@xmath0 from rmt we should look upon the even-@xmath11 theory as defining the theory while the odd-@xmath11 finite - volume partition function should be looked upon merely as a function that enters in the expression for the even-@xmath11 kernel .
we have considered the rmt formulation of qcd@xmath0 with an odd number of flavors in the mesoscopic scaling region and we have calculated the microscopic kernel from which all microscopic spectral correlation functions can be found .
we have shown the universality of the kernel by the discovery of a deep connection between the kernel of odd - flavored qcd@xmath0 with @xmath6 flavors and qcd@xmath5 ( @xmath1 ) with @xmath70 flavors in the sector of topological charge @xmath2 , brought about by the introduction of pseudo - orthogonal polynomials .
the result has the consequence that all properties of the four - dimensional theory that are universal similarly are universal in three dimensions .
we end by noting that odd - flavored qcd@xmath154 now finally has caught up with its older and more mature sisters qcd@xmath5 and even - flavored qcd@xmath0 as far as universality is concerned .
shuryak and j.j.m .
verbaarschot : random matrix theory and spectral sum rules for the dirac operator in qcd , nucl .
phys * a560 * ( 1993 ) 306 . h. leutwyler and a. smilga : spectrum of dirac operator and role of winding number in qcd , phys .
rev * d46 * ( 1992 ) 5607 .
t. guhr , a. m@xmath155ller - groeling , and h.a .
weidenm@xmath155ller : random matrix theories in quantum physics : common concepts , phys . rep .
* 299 * ( 1998 ) 189 .
j. verbaarschot : universal fluctuations in dirac spectra , hep - th/9709032 .
+ j. verbaarschot : universal behavior in dirac spectra , hep - th/9710114 .
verbaarschot : characterization of universal behavior in qcd dirac spectra , + hep - th/9807070 .
j.j.m . verbaarschot and i. zahed : random matrix theory and three - dimensional qcd , phys .
* 73 * ( 1994 ) 2288 .
g. akemann and p.h .
damgaard : microscopic spectra of dirac operators and finite - volume partition functions , nucl .
phys * b528 * ( 1998 ) 411 .
r. jackiw and s. templeton : how super - renormalizable interactions cure their infrared divergences , phys . rev .
* d23 * ( 1981 ) 2291 .
pisarski : chiral - symmetry breaking in three - dimensional electrodynamics , phys . rev .
* d29 * ( 1984 ) 2423 .
redlich : gauge noninvariance and parity nonconservation of three - dimensional fermions , phys . rev .
* 52 * ( 1984 ) 18
redlich : parity violation and gauge noninvariance of the effective gauge field action in three dimensions , phys . rev .
* d29 * ( 1984 ) 2366 .
sohnius : introducing supersymmetry , phys .
* 128 * ( 1985 ) 39 , appendix a. g. akemann , p.h .
damgaard , u. magnea , and s. nishigaki : universality of random matrices in the microscopic limit and the dirac operator spectrum , nucl . phys .
* b487 * ( 1997 ) 721 .
damgaard and s.m .
nishigaki : universal spectral correlators and massive dirac operators , nucl
* b518 * ( 1998 ) 495 .
damgaard : dirac operator spectra from finite - volume partition functions , phys .
* b424 * ( 1998 ) 322 .
zinn - justin : universality of correlation functions of hermitian random matrices in an external field , commun .
phys . * 194 * ( 1998 ) 631 . | we consider qcd@xmath0 with an odd number of flavors in the mesoscopic scaling region where the field theory finite - volume partition function is equivalent to a random matrix theory partition function .
we argue that the theory is parity invariant at the classical level if an odd number of masses are zero . by introducing so - called pseudo - orthogonal polynomials we are able to relate the kernel to the kernel of the chiral unitary ensemble with @xmath1 in the sector of topological charge @xmath2 .
we prove universality and are able to write the kernel in the microscopic limit in terms of field theory finite - volume partition functions . |
we denote here by @xmath0 an @xmath1-dimensional compact connected khler manifold . the motivations for the following shortest , but non - transparent , definition of specialness will be explained below .
see @xcite where this notion was introduced for more details .
a compact khler manifold @xmath0 is said to be * special * if , for any @xmath2 , any rank - one coherent subsheaf @xmath3 and any positive integer @xmath4 , the natural meromorphic map @xmath5 has an image of dimension at most @xmath6 if @xmath7 and is always at most equal to @xmath8 by a classical result of f. bogomolov . ] . here
@xmath9 denotes the dual of the complex vector space of sections of @xmath10 which take values in @xmath11 at the generic point of @xmath0 .
in particular , @xmath0 has then no surjective meromorphic map @xmath12 onto a manifold @xmath13 of general type and dimension @xmath2 , since otherwise @xmath14 would contradict the bound @xmath6 above . on the non - algebraic side ,
if @xmath0 has no map onto a positive dimensional - projective manifold , it is obviously special too . an equivalent geometric definition of specialness of @xmath0 actually requires that @xmath0 satisfies the more restrictive condition of having no meromorphic map onto an orbifold of general type ( see [ var speciale ] for the precise definition and the relevant notions concerning orbifolds ) .
specialness is preserved by bimeromorphic maps and finite lale covers , this last assertion being surprisingly difficult to show .
special manifolds generalise in higher dimensions the rational and elliptic curves , which are obviously exactly the special curves .
the next fundamental examples of special manifolds are indeed those which are either rationally connected or whith zero canonical dimension ( _ i.e. _ with @xmath15 , see example [ ex varietes speciales ] below ) .
however the class of special manifolds is much larger than the union of these two classes , since one shows by classification that a compact khler surface @xmath0 is special if and only if @xmath16 and @xmath17 is almost abelian
. in particular , ruled elliptic surfaces are special and surfaces with @xmath18 are special if and only if they do not map onto any hyperbolic curve after some finite tale cover .
no such simple characterisation is true when @xmath19 .
the central role of special manifolds in classification theory comes from the fact that , as shown in @xcite , any compact khler manifold is canonically and functorially decomposed by its _
core fibration _
@xmath20 into its special part ( the fibres of @xmath21 ) and its general type part ( the orbifold base @xmath22 , which is its usual base @xmath23 together with a ramification divisor @xmath24 on @xmath23 encoding the multiple fibres of @xmath21 ) .
special manifolds and general type orbifolds are thus the _ two _ antithetical building blocks from which arbitrary compact khler manifolds are built in _ one single step_. in contrast to general type manifolds , for which no classification scheme seems to be known or even expected , special manifolds are conjectured to have many fundamental properties in common with rational and elliptic curves .
conjecturally indeed , an orbifold version of the @xmath25 conjecture implies that any special manifold is canonically and functorially decomposed , by means of orbifold versions of the rational quotient and of the iitaka - moishezon fibration , as a tower of fibrations whose orbifold fibres have either @xmath15 , or @xmath26 ( a weak version of rational - connectedness , see @xcite ) .
let us stress that the orbifold considerations are essential here ( as in the lmmp , but for different reasons ) , and that apparently there is no possibility to work in the category of varieties without additional structure .
this tower decomposition permits to lift ( conditionally ) to special manifolds properties which are expected to be common to manifolds which are either rationally connected or with @xmath15 and naturally leads to the following conjectures : [ conj abelianite ] 1 .
abelianity conjecture : a special compact khler manifold has an almost abelian fundamental group ( i.e : @xmath17 has an abelian subgroup of finite index ) .
+ 2 . a compact khler manifold ( resp .
a projective manifold defined over a number field ) has an identically vanishing kobayashi pseudometric ( resp .
is potentially dense ) if and only if it is special ( this last statement is inspired by lang s conjectures ) .
for example , rationally connected manifold are simply connected @xcite and compact khler manifolds @xmath0 with @xmath27 have an almost abelian fundamental group by @xcite and @xcite . in this article we prove the abelianity conjecture above for threefolds ( as said above , this is known for surfaces by classification , see proposition [ surface speciale ] ) .
[ main ] let @xmath0 be a compact khler threefold .
if @xmath0 is special , its fundamental group is almost abelian
. this immediately implies , among several other things , a precise solution of shafarevich s conjecture in this case ( and for all special manifolds if the abelianity conjecture holds ) : if @xmath0 is a special compact k " ahler threefold with universal cover @xmath28 , let @xmath29 be any finite tale cover of @xmath0 with abelian torsionfree fundamental group and albanese variety @xmath30 of dimension @xmath31.then @xmath32 is holomorphically convex , the universal cover @xmath33 of @xmath30 being stein , since isomorphic to @xmath34 .
we shall prove the theorem in this section only in the cases where no orbifold structure is needed , that is except when either @xmath0 is projective and @xmath35 , or when @xmath36 .
the treatment of these two residual cases needs the consideration of two - dimensional projective special orbifolds and will be the subject of the subsequent sections .
because of the lack of a minimal model program in the k " ahler non - projective case , we need to treat it differently from the projective case .
so assume first that @xmath0 is projective .
we work according to the value of @xmath37 .
+ @xmath38 by miyaoka s theorem , @xmath0 is uniruled .
let @xmath39 be its rational quotient ( also known as its mrc fibration )
. then @xmath40 is special too , with @xmath41 , and @xmath42 , since the fibres of @xmath43 are rationally connected . since @xmath44 is almost abelian , so is @xmath17 .
@xmath45 if @xmath27 , the theorem is true , by @xcite and @xcite .
one reduces to this case by the minimal model program and @xcite ( see ( * ? ? ?
* ( 4.17.3 ) ) ) .
the details are as follows : there exists a terminal model @xmath29 birational to @xmath0 such that @xmath46 is torsion , hence trivial after finite tale ( in codimension @xmath47 cover .
then @xmath29 has only @xmath48 singularities .
it is thus smoothable in the projective category by @xcite .
the conclusion follows , since @xmath17 drops by specialisation .
@xmath49 let @xmath50 be the moishezon - iitaka fibration . because its base is a curve
, there exists a finite tale cover of @xmath0 ( still written @xmath51 such that @xmath52 has no multiple fibre and @xmath52 has one or two multiple fibres .
this case is easily treated similarly . ] .
then @xmath17 is an extension of @xmath53 @xmath54 being a rational or elliptic curve , since special ) by a quotient of @xmath55 , @xmath56 being the generic smooth fibre of @xmath52 , which is a surface with @xmath15 .
because these two groups are almost abelian , @xmath17 is _ a priori _ only polycyclic .
the following result however implies the conclusion .
[ tfbspec ] let @xmath57 be a fibration without multiple fibres in codimension @xmath58 on @xmath13 from a compact k " ahler manifold onto a manifold @xmath13 .
assume that @xmath13 and the generic fibre @xmath56 of @xmath59 both have an almost abelian fundamental group .
then @xmath0 also has an almost abelian fundamental group .
the proof of this theorem rests on two deep results of hodge theory : deligne s strictness theorem for morphisms of mhs , and hain s functorial mhs on the malev completion of @xmath17 when @xmath0 is compact k " ahler .
@xmath60 when @xmath0 is projective , this is the only remaining case .
notice that when @xmath18 , we could remove the multiple fibres of @xmath52 by making a suitable finite tale cover of @xmath0 because @xmath13 was a curve .
when @xmath35 this is no longer possible in general .
this is the reason why the notion of orbifold base , which virtually removes the multiple fibres , is introduced in this case , and why the geometry of such orbifolds needs to be developed .
but once this is done , and the corresponding properties established , the proof is entirely parallel .
+ we now deal with the case when @xmath0 is _ not _ projective .
we work this time according to the algebraic dimension @xmath61 of @xmath0 .
+ @xmath62 the irregularity of @xmath0 can take the values @xmath63 or @xmath64 ( since @xmath65 would imply @xmath66 .
the assumption @xmath67 leads to the finiteness of @xmath68 according to ( * ? ? ?
recall that the albanese map is connected and surjective when @xmath69 .
thus , if @xmath70 , then @xmath0 is birational to its albanese variety , and @xmath17 is abelian .
when @xmath71 , the albanese fibration @xmath72 has no multiple fibre in codimension one by ( * ? ?
moreover , the general fibre of @xmath73 is elliptic or rational , by ( * ? ? ?
13.8 ) . from theorem [ tfbspec ] , we conclude that @xmath68 is almost abelian .
@xmath74 the algebraic reduction @xmath75 is then a fibration onto a curve @xmath76 with general fibre special ( * ? ? ?
* th . 2.39 ) ; the fundamental groups of the base and of the fibre are then almost abelian ( see theorem [ surface speciale ] ) . as in the projective case with @xmath18
, we may assume that there are no multiple fibres .
then a final application of theorem [ tfbspec ] implies the result .
+ we are thus left with the following two cases : 1 .
@xmath0 is projective and @xmath35 , 2 .
@xmath0 is a compact khler threefold with @xmath36 .
let us briefly explain how the conclusion is obtained there .
see next sections for the notions which are being used . in both cases
we have a fibration ( which may be assumed to be neat after suitable modifications of @xmath0 and @xmath77 ) @xmath78 on a smooth projective surface @xmath77 , with generic fibres elliptic curves .
this defines a smooth ( hence klt ) orbifold base @xmath79 . because @xmath0 is special , so is @xmath79 . as above , for such a fibration
, @xmath17 is now an extension of @xmath80 by @xmath81 .
we show below : [ groupe orbisurface spciale ] the fundamental group of a special geometric orbifold @xmath82 of dimension 2 is almost abelian of even rank at most @xmath83 .
the conclusion now follows from the orbifold version of theorem [ tfbspec ] above .
+ let us stress that our proof of theorem [ groupe orbisurface spciale ] depends in an essential way on the fact that _ integral klt _ pairs are locally uniformisable by smooth germs of surfaces , because we are using the existence of ricci - flat orbifold metrics when @xmath84 .
this is the main reason why we can not extend the above theorem [ groupe orbisurface spciale ] to any higher dimension .
otherwise , using the abundance conjecture , it seems that one could derive the abelianity conjecture in the projective case for klt orbifolds in any dimension using inductively on the dimension essentially the same arguments as the ones presented below .
the orbifolds considered here are spaces with local smooth uniformisations under the action of some finite groups . to deal with fixed point in codimension one ,
the structure is enriched with a @xmath85-divisor of a specific form .
the definition below is taken from @xcite .
[ defi - orbi ] an orbifold is a pair @xmath86 where @xmath0 is a normal variety and @xmath87 a @xmath85-weil divisor of the following form : @xmath88 where the @xmath89 are integers , and the pair @xmath86 is locally uniformizable in the following sense : @xmath0 is covered by the domains @xmath90 of finite maps @xmath91 satisfying the following properties : a. @xmath92 is open in @xmath0 , b. @xmath93 is a galois cover whose branching divisor is exactly @xmath94 .
the support of @xmath87 is then : @xmath95
. * terminological remark . * + the orbifolds we consider are compatible with all situations in which this term is used : they are particular cases of the ones in @xcite and the special manifolds defined below are actually the _ classical _ special ones in _ ibid .
_ they are smooth deligne - mumford stacks too and also klt pairs of the lmmp .
because they are locally smoothly uniformised , we can attach to them fundamental groups and differential - geometric notions such as metrics and differential forms .
[ fibr canonique orbi ] the canonical divisor of such a pair is the ( weil ) @xmath85-divisor @xmath96 .
[ exemple snc ] let @xmath0 be a smooth variety , and @xmath95 be a normal crossing divisor ; the choice of multiplicities @xmath97 on each component of the divisor defines a canonical orbifold structure on @xmath86 . since @xmath98 is locally given by the equation @xmath99 in suitable coordinates @xmath100 , the map @xmath101 gives a local uniformization .
these orbifolds are said to be smooth and integral in @xcite .
in particular , an orbifold curve is simply a smooth curve with a finite set of points marked with integral multiplicities at least @xmath102 .
since we consider only integral and finite multiplicities here , we define : [ defi orbi geo ] a geometric orbifold is an orbifold @xmath86 with @xmath0 smooth and @xmath98 of normal crossings .
[ defi fibration nette ] let @xmath103 be a fibration ( surjective morphism with connected fibres ) between compact complex manifolds .
neat _ model of @xmath59 consists in a commutative diagram : @xmath104_f \ar[r]^u & x_0\ar[d]^{f_0 } \\
y\ar[r]_v & y_0 } \ ] ] where 1 .
@xmath0 and @xmath13 are smooth , 2
. @xmath105 and @xmath106 are bimeromorphic morphisms , 3 .
the locus of singular fibres of @xmath59 is a normal crossing divisor of @xmath13 , 4 .
every @xmath59-exceptional divisor is @xmath105-exceptional too .
such models actually exist birationally for any fibration ; this can be proven using hironaka desingularization and raynaud flattening s theorems ( * ? ? ?
* lemma 1.3 ) .
notice that any fibration is neat when @xmath13 is a curve .
+ given a neat fibration @xmath107 , we can naturally associate a @xmath85-divisor @xmath108 on the base @xmath13 of @xmath59 . this divisor will be supported on the singular locus of @xmath59 and will thus be normal crossings : the pair @xmath109 will be a geometric orbifold .
the construction goes as follow ( see also @xcite ) : if @xmath110 is any component of the singular locus of @xmath59 , its pull - back can be written @xmath111 where @xmath112 is @xmath59-exceptional and @xmath113 is mapped surjectively onto @xmath110 .
the multiplicity of @xmath59 along @xmath110 is defined by : @xmath114 [ base orbi stable ] the pair @xmath109 , where : @xmath115 is called the _ orbifold base _ of the fibration @xmath116 ( notations as above ) . + a fibration @xmath116
is said to be of general type if the canonical divisor of the orbifold base @xmath109 is _ big _ : @xmath117 [ var speciale ] a compact khler manifold @xmath0 is said to be ( classically ) special if it does not admit any fibration of general type .
[ ex varietes speciales ] the main examples of special manifolds are given by the following classes
@xcite : 1 . rationally connected manifolds , 2 .
compact khler manifolds @xmath0 with @xmath118 .
this is a consequence of the additivity of canonical dimensions in general type fibrations ( ( * ? ? ?
4.2 ) , an orbifold version of viehweg s theorem ) .
3 . the special curves are thus just the rational or elliptic ones .
4 . compact khler manifolds of algebraic dimension zero and more generally fibres of algebraic reductions .
conjecturally in an orbifold version of iitaka s @xmath25-conjecture , special manifolds can be reconstructed as a tower ( in a suitable sense ) of fibrations with fibres belonging to the classes above .
this reduces abelianity conjecture [ conj abelianite ] to the case of orbifolds with either @xmath15 or @xmath26 ( see @xcite ) . in dimension
@xmath102 only , we still have a simple topological characterisation of specialness : [ surface speciale ] a compact khler surface @xmath0 is special if and only if : @xmath119 and if @xmath68 is almost abelian .
we shall need also the notion of orbifold base of a fibration @xmath57 also when @xmath0 is equipped with an orbifold divisor @xmath120 , at least when @xmath13 is a curve and @xmath0 is a surface ( for the general case , see the definition in ( * ? ? ?
* def . 4.2 ) ) .
let @xmath121 be a geometric orbifold of dimension 2 and @xmath122 a fibration onto a curve .
let us define the multiplicity of a point @xmath123 ( relatively to @xmath59 and @xmath120 ) by the following formula : @xmath124 where @xmath125 .
the orbifold base is then the pair @xmath126 where : @xmath127 we say that @xmath59 is of general type if : @xmath128 the @xmath102-dimensional orbifold @xmath121 is then said to be special if @xmath121 does not admit any general type fibration onto a curve and if @xmath129 .
in this paragraph , we define fundamental groups of orbifolds and study the morphisms induced at the level of fundamental groups by classical orbifold morphisms .
[ groupe fondamental ] let @xmath86 be an orbifold ; its fundamental group @xmath130 is the quotient of the group @xmath131 by the normal subgroup generated by the loops @xmath132 where @xmath133 is a small loop around the component @xmath134 of multiplicity @xmath135 ( where @xmath136 denote the smooth locus of @xmath0 ) .
this definition is derived from the local models .
if @xmath137 where @xmath138 is a finite subgroup of @xmath139 and @xmath87 is the branching divisor of the projection @xmath140 , we recover the group @xmath138 as an orbifold fundamental group associated with @xmath86 and called the _ root stack _ by abramovich - vistoli .
] : @xmath141 [ exemple groupe courbe ] the case of orbifold curves is quite classical .
the structure of the fundamental group of such an orbifold curve @xmath142 ( we just keep in mind the deformation invariant , that is : the multiplicities , not the marked points ) is determined by the sign of its canonical bundle : @xmath143 when this quantity is positive ( resp .
zero , resp .
negative ) , the fundamental group of @xmath144 is commensurable to the fundamental group of an hyperbolic curve ( resp .
commensurable to @xmath145 , resp .
finite and explicitely known ) .
when dealing with fibrations having multiple fibres , we need to consider orbifold fundamental groups .
a fibration @xmath107 with general fibre @xmath56 gives rise to a natural sequence of fundamental groups : @xmath146 although @xmath147 is surjective since @xmath59 has connected fibres , this sequence is in general not exact in the middle .
multiple fibres are responsible for this failure , remedied by the orbifold fundamental group .
we will need the case where @xmath0 has an orbifold structure and so a slightly more general version . for the definition ( of a neat fibration in the orbifold setting ) , we refer to ( * ? ? ?
[ suite exacte ]
let @xmath148 be a neat fibration .
if @xmath56 denote a general fibre of @xmath59 and @xmath149 the restriction of the orbifold structure to @xmath56 , the sequence @xmath150 is exact .
[ remarque fibration nette ] we omit the detailled definition of neatness required for the preceding statement because in the sequel we will only use it in the two following quite simple situations : a. when @xmath120 is empty , this neatness assumption has been already encountered ( _ cf . _ definition [ defi fibration nette ] ) .
the content of the previous proposition is then that @xmath68 sits then in the middle of a short exact sequence ( with the fundamental group of the orbifold base on the right hand ) .
b. when @xmath13 is a curve : the fibration @xmath148 is always neat . to complete this study of the behaviour of fundamental groups in fibrations ,
we prove that nilpotency conditions on the fundamental groups are preserved in fibrations between khler orbifolds .
this remarkable fact ( obviously false even for submersions between complex manifolds as the iwasawa manifold shows ) is a consequence of deep results in mixed hodge theory .
[ fibration presque ablien ] let @xmath151 be a neat fibration ( see remark [ remarque fibration nette ] ) between smooth compact khler orbifolds . if the groups @xmath152 and @xmath153 are almost abelian , then @xmath154 is also nilpotent of nilpotency class at most @xmath155 . ] , then @xmath154 is almost abelian as well .
it is a reduction to the case when @xmath156 .
since @xmath59 is assumed to be neat , the sequence @xmath157 is exact . since @xmath152 and @xmath153 are almost abelian groups ( of finite type ) , @xmath158 is then an almost polycyclic group ; in particular , the group @xmath138 is linear . a famous result of selberg asserts that @xmath138 has then a finite index subgroup @xmath159 which is torsion free .
consider @xmath29 the orbifold cover of @xmath121 associated with @xmath160 ; the latter being torsion free , @xmath29 is a normal variety with no orbifold structure ( _ i.e. _ @xmath161 ) .
since @xmath0 has only quotient singularities , its fundamental group is isomorphic to the one of @xmath162 , a desingularisation of @xmath29 ( * ? ? ?
7.5 ) . to conclude , we consider ( a neat model of ) the stein factorization of @xmath163 indeed , taking further blow - up of @xmath162
, we can complete the picture : @xmath164\ar[rrd]_{\tilde{f } } & x'\ar[r ] & x\ar[r]^{f } & y\\ & & \tilde{y}\ar[ru]_{g } & } \ ] ] with @xmath165 a neat fibration and @xmath166 generically finite .
it is then easy to see that we have an exact subsequence ( with vertical maps having finite index images ) : @xmath167 \ar@{^{(}->}[d ] & \pi_1(\tilde{x } ) \ar[r]^{\tilde{f } _ * } \ar@{^{(}->}[d ] & \pi_1(\tilde{y},\delta^*(\tilde{f}))\ar[r ] \ar@{^{(}->}[d ] & 1\\ \pi_1(x_y,\delta_{x_y } ) \ar[r ] & \pi_1(x,\delta_x)\ar[r]^{f _ * } & \pi_1(y,\delta_y)\ar[r ] & 1 } \ ] ] so we are reduced to the same situation with @xmath156 and we can apply ( * ? ? ?
7.6 ) . as for of compact khler manifolds ,
when the first chern class is zero or positive , the now usual differential - geometric methods can then be applied to construct orbifold khler metrics with ricci - curvature of the corresponding sign ( using the local smooth uniformisations ) .
[ groupe orbi c1 positif ] let @xmath168 be an orbifold with @xmath0 compact k " ahler .
if the first chern class of @xmath169 is non negative , the group @xmath170 is almost abelian ; more precisely : 1 . if @xmath171 , then @xmath130 is almost abelian of even rank of finite type is the maximum rank of an abelian subgroup of finite index of @xmath138 . ] bounded by @xmath172 .
2 . if @xmath173 , then @xmath130 is finite . to begin with ,
let us recall some basic facts on differential calculus on orbifolds .
the smooth functions ( differential forms , hermitian metrics,@xmath174 ) on an orbifold @xmath86 are the smooth functions on @xmath175 which can be smoothly extended ( as the usual objects , after taking inverse images ) in local uniformizations .
for instance , if @xmath86 is a geometric orbifold , a khler metric has the following form in coordinates charts adapted to @xmath87 : @xmath176 local uniformizations can also be used to compute integrals of forms of maximal degree .
if @xmath177 is such a local cover and @xmath178 an orbifold top - form on @xmath0 , define @xmath179 this local computation can then easily be globalized using partitions of unity .
the canonical divisor has been defined in [ fibr canonique orbi ] as a @xmath85-divisor .
it should be noted here that ( @xmath0 being compact and the uniformizations being finite ) some integral multiple of this divisor defines a line bundle on @xmath0 and it can be used to compute the first chern class of @xmath0 . as a byproduct , the ricci form of any orbifold volume form is a ( 1,1)-orbifold form whose cohomology class coincides with @xmath180 . + as in the manifold case , every invariant form whose cohomology class coincides with @xmath180 is the ricci curvature of an orbifold khler metric .
this fact has been already noticed several times in the literature ( see for instance @xcite when @xmath181 , and @xcite when @xmath182 , and the references therein ) .
for sake of completeness , let us recall the statement .
[ orbi - calabi ] let @xmath183be a khler orbifold whose underlying space is compact and let us fix @xmath184 an orbifold khler metric .
for any representative ( _ i.e. _ smooth invariant @xmath185 form ) @xmath178 of @xmath180 , there exists a unique orbifold khler metric @xmath186 in the khler class @xmath187 $ ] such that : @xmath188 we reproduce the arguments given in ( * ? ? ? * th .
this is a simple adaptation of the proof exposed in @xcite .
since @xmath189 and @xmath178 define the same class , the orbifold @xmath190-lemma provides us a smooth ( orbifold ) function @xmath59 such that @xmath191 .
we normalize @xmath59 ( by addition of a suitable constant ) in such a way that @xmath192 the problem is then reduced to solving the monge - ampre equation @xmath193 to do so , we applied the continuity method as in @xcite and consider the set @xmath194 of @xmath195 $ ] for which the equation @xmath196 has a solution where @xmath197 when @xmath198 , @xmath199 is an obvious solution for @xmath200 .
the set @xmath194 is open by linearization of the problem and implicit functions theorem ( applied in local uniformizations ) . to see that @xmath194 is closed we need a priori estimates on the solution of @xmath201 which are established using two types of arguments : maximum principle and nash - moser iteration ( integral inequalities ) .
these can be used in the orbifold setting as well : maximum principle is applied in local uniformizations and integral inequalities are also valid in this formalism ( see above ) .
we proceed as in the manifold case to deduce theorem [ groupe orbi c1 positif ] from theorem [ orbi - calabi ] .
the positivity assumptions can be translated in the following way : in the fano case @xmath202 there exists a positive @xmath185 form @xmath178 representing @xmath180 ; in the ricci flat case @xmath203 we choose @xmath204
. we can now apply theorem [ orbi - calabi ] : @xmath183has an orbifold khler metric @xmath186 with prescribed ricci curvature : @xmath188 in the fano case , by compactness of @xmath0 , we get : @xmath205 , where @xmath206 .
the following orbifold version of myer s theorem applied to the orbifold universal cover of @xmath169 implies the finiteness of the orbifold fundamental group : [ orbi - myer ] let @xmath207 be a complete riemannian orbifold ( of dimension @xmath1 ) . if the ricci curvature satisfies the inequality @xmath208 for some @xmath209 , the underlying space @xmath13 of @xmath210 is then compact with diameter bounded above : @xmath211 when @xmath181 , the following orbifold splitting theorem [ orbi decomposition ] applied also to the orbifold universal cover of @xmath169 also implies the claim .
[ orbi decomposition ] let @xmath207 be a compact riemannian orbifold ( of dimension @xmath1 ) . if the ricci curvature is everywhere non negative , the orbifold universal cover @xmath212 admits a ( metric ) splitting @xmath213 where @xmath4 is a compact orbifold . the orbifold fundamental group of @xmath210 is also an extension of a crystallographic group by a finite group and is in particular almost abelian . from the arguments given in @xcite we deduce that the rank of the almost group @xmath130 is even and bouded by @xmath214.@xmath215
in this short paragraph , we gather from @xcite some well known facts on the mmp for log pairs in dimension 2 . starting with a pair @xmath86 where @xmath0 is a smooth surface and @xmath216 a @xmath85-weil effective divisor
, we can perform a sequence of divisorial contractions which ends with a birational model of the initial pair and whose geometry is simplest , according to the sign of the canonical bundle . in this process , however , @xmath217-negative curves are being contracted , and the resulting surface is no longer smooth in general .
the relevant preserved category of singularities is then described as follows .
[ def klt ] a pair @xmath86 where @xmath0 is a @xmath85-factorial normal variety is said to have only klt singularities if a. @xmath218 b. for any ( or equivalently , one ) log - resolution @xmath219 of @xmath86 , in the following equality , we have @xmath220 , for all @xmath221 : @xmath222 @xmath223 being the strict transform of @xmath87 , and @xmath224 the exceptional divisors of @xmath59 .
a smooth ( integral ) orbifold is of course a @xmath225 pair .
these singularities are preserved in the lmmp : [ lmmp surface ] let @xmath86 be a klt surface
. there exists a birational morphism @xmath226 such that the resulting pair @xmath227 is still klt and satisfies ( exactly ) one of the following properties : 1 .
@xmath228 is nef , 2 .
@xmath77 admits a fibration @xmath229 onto a ( smooth ) curve @xmath230 , the general fibre of @xmath231 being a smooth @xmath228-negative rational curve , 3 .
@xmath82 is del pezzo : @xmath232 and @xmath233 is ample .
when @xmath234 , it is well know that klt singularities coincide with quotient singularities in dimension 2 ( * ? ? ?
this property still holds for @xmath235 with integral multiplicities ( _ i.e. _ coefficients @xmath236 of the form @xmath237 ) .
[ klt2=quotient ] let @xmath86 be a pair where @xmath0 is a surface and @xmath87 has integral multiplicities .
the following conditions are equivalent near each point of @xmath0 : 1 .
@xmath86 is klt , 2 .
@xmath86 has a finite local fundamental group , 3 .
@xmath86 is locally presented as a quotient : @xmath238 , where @xmath138 is a finite group acting linearly on @xmath239 and @xmath87 is the unique @xmath85-divisor on @xmath0 such that @xmath240 .
this statement is found and used in several places @xcite , but never with a complete accessible proof : the appendix of @xcite consists in a list of the possible cases , and refers to the thesis @xcite for the proof .
for this reason , we give a proof of theorem [ klt2=quotient ] at the end of the present text ( _ cf .
_ appendix ) .
we now relate the fundamental groups of a geometric orbifold @xmath86 of dimension 2 and of its minimal model @xmath241 as in theorem [ lmmp surface ] .
[ groupe modele minimal ] let @xmath86 be a geometric orbifold of dimension 2 and @xmath242 its minimal model .
there is a natural surjective morphism of groups : @xmath243 in particular , if @xmath244 is almost abelian so is @xmath130 .
let us call @xmath245 the union of the curves contracted by @xmath59 ; @xmath59 being an isomorphism away from @xmath245 we have natural maps , @xmath246 being the smooth locus of @xmath77 : @xmath247 ( note that @xmath98 and @xmath245 may have common components ) .
at the level of fundamental groups , we get natural morphisms : @xmath248 to get the morphism @xmath249 we need only to remark that the loops around the components of @xmath250 are sent onto loops around corresponding components in @xmath87 with the same multiplicities , since @xmath59 is an isomorphism between @xmath246 and @xmath251 . [ attention ] in general , the kernel of this morphism can however be very big as shown by the following example .
let @xmath230 be an elliptic curve , and let @xmath0 be the blow - up of @xmath252 at a point @xmath253 .
let @xmath254 the fibre of the first projection through this point and let @xmath138 be its strict transform in @xmath0 .
if @xmath255 is an integer , the minimal model of @xmath256 is @xmath257 and the exceptional divisor @xmath245 of the blow - up is the only @xmath217-negative curve on @xmath0 .
since the orbifold base of the first projection is @xmath258 , the fundamental group of @xmath259 has ( a finite index subgroup which have ) a surjective morphism onto a non - abelian free group . on the other hand
the orbifold base of @xmath260 , @xmath166 being the composition of the first projection and the blow - up , is merely @xmath230 since the fibre over @xmath261 has @xmath245 as component and inherits from it the multiplicity 1 .
it is then easy to see that @xmath86 is special and that its fundamental group is isomorphic to the fundamental group of @xmath0 and thus abelian .
[ remarque groupe fondamental mmp ] the direction of the arrow @xmath249 in proposition [ groupe modele minimal ] does not look to be functorial .
so we explain its construction differently .
the map @xmath262 is in general not a ` classical ' orbifold morphism ( in the sense of @xcite ) and thus does not induce a functorial morphism of groups .
the map @xmath59 induces an orbifold morphism , and thus a morphism of groups , only when the multiplicities on the exceptional divisors of @xmath59 are sufficiently divisible ( for example by the order of the local fundamental group of @xmath259 at the point under consideration ) . in our example [ attention ] above , equipping @xmath245 with a multiplicity divisible by @xmath263 is the right condition .
let thus @xmath264 be an orbifold divisor on @xmath0 such that @xmath265 , and such that @xmath266 and @xmath267 are orbifold morphisms ( this means for @xmath268 that the multiplicity of any component of @xmath87 divides the corresponding multiplicity in @xmath269 .
we then get two functorial group morphisms : @xmath270 , which is an isomorphism , and @xmath271 , which is surjective ( as above ) .
our initial @xmath249 was nothing but @xmath272 . to conclude this section
, we shall use the preceding proposition to prove abelianity conjecture for geometric orbifolds of dimension 2 .
+ let @xmath259 the minimal model of the pair @xmath121 .
we argue according to the values of @xmath273 .
if @xmath274 , then @xmath0 admits a fibration @xmath59 onto a smooth curve @xmath230 whose general fiber @xmath275 satisfies @xmath276 ; the general orbifold fibre @xmath277 is then orbifold - elliptic and its fundamental group is almost abelian ( example [ exemple groupe courbe ] ) . adding to @xmath230 the orbifold divisor @xmath278 of classical multiplicities , we get an exact sequence @xmath279 the orbifold curve @xmath280 is special since so is @xmath86 and theorem [ fibration presque ablien ] shows that @xmath130 is almost abelian .
when @xmath259 admits a structure of mori fibre space over a ( special ) curve ( in particular @xmath281 ) , we still have a fibration on @xmath0 whose fibres are special ( in fact rational ) and we can proceed as in the lines above . there are two cases left : 1 .
@xmath282 , 2 . or @xmath259 is log del pezzo . in the first case ,
@xmath283 being _ nef _ it is semi - ample ( log abundance for surfaces , see @xcite ) and then torsion . in particular , the orbifold @xmath259 is ricci - flat . in the second case , @xmath259 is fano . in both cases
, we can apply theorem [ groupe orbi c1 positif ] to conclude that the fundamental group of @xmath259 is almost abelian .
finally , proposition [ groupe modele minimal ] can be applied ( it is only here we need it ) to show that @xmath154 is almost abelian as well .
in this appendix we give a proof of the classification of klt singularities for integral pairs ( theorem [ klt2=quotient ] ) .
the method is first to treat the singular case ( using the method of ( * ? ? ?
4.7 ) ) and then to reduce the smooth case to the preceding one using orbifold tale covers .
let then @xmath86 , @xmath284 be a germ of klt pair .
[ lfq ] let @xmath285 be a connected exceptional curve on a smooth complex surface .
let @xmath286 and @xmath287 , with @xmath288 .
assume that @xmath289 for any @xmath290 .
then either @xmath291 or @xmath292 for any @xmath290 .
assume that @xmath0 is singular , and that @xmath235 .
let @xmath293 be the minimal resolution of the germ @xmath0 , @xmath294 being the strict transform of @xmath87 in @xmath29 and @xmath295 the exceptional divisor .
the extended dual graph of @xmath296 is then one of the following ones below , in which the ( non - compact ) components of @xmath297 are indicated by black dots .
moreover , all intersections are transversal and all white dots are smooth rational curves
. these are @xmath298-curves , except possibly in the first case ( see remark [ rklt+ ] below for additional constraints in the first case ) .
we assume the knowledge of the classification of duval ( _ i.e. _ canonical ) and of klt germs of @xmath102-dimensional singularities , and more precisely the fact that the dual graphs of their minimal resolutions are given either by dynkin diagrams of @xmath298-curves of type @xmath300 , @xmath301 , @xmath302 ( for @xmath303 ) or by hirzebruch - jung chains with transversal intersections ( ( * ? ? ?
4.7 ) , the method of proof we shall now adapt ) .
assume first that there are ( at least ) two components @xmath308 and @xmath309 in @xmath297 , of multiplicities @xmath310 .
consider a shortest chain @xmath311 of components of @xmath245 joining them , such that @xmath308 ( resp .
@xmath312 meets @xmath313 ( resp .
@xmath314 , but no other @xmath315 .
an easy computation gives : @xmath316 consider now the curve @xmath317 , for some @xmath318 to be chosen latter .
we assume first that @xmath319 .
the intersection numbers are then given by : @xmath320 we choose now the value of @xmath321 according to the self - intersection of the @xmath322 : we thus have @xmath289 for any @xmath290 , and equality for some @xmath290 .
from lemma [ lfq ] we get @xmath291 and @xmath328 since @xmath86 is klt . using @xmath329
it easily implies @xmath330 because @xmath331 we get @xmath332 and we are thus in the first case .
this last case ( which is not normal crossings ) is excluded by two blow - ups which make the total transform of @xmath87 a normal crossings divisor .
we indeed get the following extended dual graph ( in which the @xmath342 appear as white circles in the order @xmath343 , while the left ( resp .
upper ) black dot is the strict transform of @xmath308 ( resp .
@xmath312 : assume next that our dual graph contains a fork ( of white vertices ) .
we know that @xmath346 in this case ( @xmath347 ) .
let @xmath348 be a shortest subgraph connecting the component @xmath308 to the fork @xmath349 , with end points the last two vertex .
the sequence is so labelled that @xmath308 meets @xmath313 only , and that @xmath350 is a chain . consider the curve @xmath351 a computation similar to the preceding one ( but simpler since now @xmath346 ) shows that @xmath352 .
thus @xmath291 and we are in the second case . the only case left is no fork and only one component in @xmath297 . this is the last diagram ( with transversal intersection by the klt and integral conditions , as above ) since the degenerate case where @xmath353 in the preceding case would lead to the existence of a fork by connectedness of @xmath245 .
[ rklt+ ] in the first case , the computation gives additionally : @xmath354 and also @xmath355 .
the arguments of ( * ? ? ?
* iii ( 5.1 ) ) show that the singularity of @xmath0 is cyclic of type @xmath356 with @xmath357 and @xmath358 with @xmath359 .
if @xmath234 , it is ( * ? ? ?
4.18 ) . otherwise ,
if the singularity is of type @xmath300 , an easy computation shows ( as in ( * ? ? ?
iv , 13 - 14 ) ) that the local fundamental group has order @xmath360 . if the singularity is of type @xmath301 , we reduce it to the preceding case by a double cover .
[ loe ] let @xmath361 be a finite map between normal germs of varieties .
assume that @xmath362 and @xmath363 are @xmath85-cartier the @xmath85-cartier assumption is superfluous . ] , and that @xmath364 .
then @xmath362 is klt if and only if @xmath363 is klt .
[ roe ] in this preceding case the local fundamental group of @xmath372 has index @xmath366 in the local fundamental group of @xmath86 .
moreover from the ramification formula we get : @xmath373 since @xmath374 .
it also follows that @xmath86 is klt if and only if so is @xmath372 .
+ we shall apply this remark only when @xmath375 , and @xmath376 for @xmath377 . in this case @xmath378 and the component @xmath308 ` disappears ' .
assume that @xmath308 is a cusp of equation @xmath379 ( @xmath380 coprime ) and of multiplicity @xmath381 which we write symbolically as @xmath382 .
then @xmath29 is the singularity of equation @xmath383 .
thus if @xmath86 is klt we have @xmath384 .
and so @xmath385 is up to order either @xmath386 or @xmath387 , @xmath388 .
moreover if the support of @xmath87 is reducible ( if @xmath377 ) then @xmath29 must be an @xmath300 singularity and the support of @xmath297 must have at most two components .
thus @xmath389 is , up to order , @xmath390 and @xmath391 .
if @xmath308 is not _ a priori _ assumed to be a cusp , but just be given by a parametrisation @xmath392 with @xmath380 coprime , then since the cover of degree @xmath381 ramified exactly over @xmath308 will be a klt singularity , and so will have the usual normal form in suitable coordinates .
we thus see _ a posteriori _ that @xmath308 was indeed a cusp in suitable coordinates . \2 .
assume that the support of @xmath87 has two smooth components @xmath393 tangent at order @xmath394 at the origin .
they have ( in this order ) , in suitable coordinates , equations : @xmath395 and @xmath396 .
consider the map : @xmath397 given by : @xmath398 . thus using the above remark [ roe
] , we see that making this ( orbifold tale ) cover @xmath166 we are lead to the case ( again klt ) where @xmath399 with @xmath400 having equation @xmath401 . in this case , letting @xmath366 be the @xmath402 of @xmath403 , the germ @xmath400 splits in @xmath366 irreducible components which are cusps of type @xmath404 and multiplicity @xmath405 with @xmath406 and are thus smooth if and only if either @xmath407 or @xmath408 or @xmath409 .
we consider next the remaining case in which the germ @xmath0 is smooth .
write @xmath410 . for @xmath411 ,
let @xmath412 be the multiplicity of the germ @xmath370 at the origin .
thus @xmath370 is smooth if and only if @xmath413 .
we say that an irreducible germ of curve in @xmath414 has a @xmath415-cusp at the origin if its equation in suitable coordinates is @xmath416 ( with @xmath8 and @xmath417 coprime ) ; in this case the multiplicity is @xmath418 .
1 . @xmath420 . then @xmath391 . if @xmath421 then @xmath422 ( _ i.e. _ either @xmath423 with @xmath424 arbitrary or @xmath425 with @xmath426 ) .
@xmath427 for some @xmath428 .
then either @xmath429 and @xmath370 has a @xmath430-cusp at the origin with @xmath431 ; or @xmath432 ( and assume @xmath433 ) .
there are two subcases : a. @xmath308 has a @xmath430-cusp with @xmath434 and @xmath309 is smooth and has intersection multiplicity @xmath102 with @xmath308 ; @xmath405 and @xmath417 ( odd ) are arbitrary .
b. @xmath308 has a @xmath435-cusp with @xmath434 , @xmath309 ( smooth ) has intersection multiplicity @xmath64 with @xmath308 and @xmath436 .
3 . @xmath437 for some @xmath428 .
then @xmath429 and @xmath308 is a @xmath438-cusp or @xmath439-cusp with @xmath434 .
let us make a blow - up @xmath440 at the origin with exceptional divisor @xmath313 and denote @xmath441 and @xmath442 .
let @xmath443 be the strict transform of @xmath87 in @xmath444 .
then @xmath445 where @xmath446
. the klt condition implies @xmath447 that is @xmath448 .
since @xmath449 , this implies @xmath450 .
thus @xmath391 and we get the following list of possible values : if @xmath456 and @xmath429 the tale orbifold cover ( see [ eoe ] above ) of degree @xmath381 ramified along the @xmath415-cusp @xmath308 leads to the singularity @xmath383 with the zero orbifold divisor , which is klt if and only if the claimed inequality @xmath384 holds . if @xmath457 , we thus get the cases described in cases 2 and 3 of lemma [ lbu1 ] . we are left with the case @xmath458 .
thus @xmath308 is a @xmath430-cusp of multiplicity @xmath381 and @xmath309 smooth of multiplicity @xmath405 with @xmath417 odd .
we distinguish two cases according to whether the intersection multiplicity of @xmath308 and @xmath309 is @xmath102 or more . in the first case ,
the orbifold tale cover of @xmath86 ramified to order @xmath405 along @xmath309 leads to the orbifold divisor @xmath297 supported on the locus of equation @xmath459 and with multiplicity @xmath381 . from @xmath372 still being klt we derive @xmath460 and it yields @xmath434 since @xmath461 .
if the intersection multiplicity of @xmath308 and @xmath309 is @xmath64 or more , the orbifold tale cover of @xmath86 ramified to order @xmath405 along @xmath309 leads to the orbifold divisor @xmath297 supported on the locus of equation @xmath462 , and with multiplicity @xmath381 ( still klt ) . if @xmath417 and @xmath405 are coprime we get the inequality @xmath463 whose the only possible solution is @xmath464 and @xmath465 .
we now show that @xmath417 and @xmath405 are coprime , which will complete the proof .
otherwise let @xmath466 be their @xmath402 .
then @xmath297 consists of @xmath366 components of multiplicity @xmath381 and equations @xmath467 with @xmath468 and @xmath469 .
these components need to be smooth since @xmath466 .
thus either @xmath470 or @xmath471 . since @xmath417 is odd it is not divisible by @xmath472 and so @xmath473 divides @xmath405 .
because @xmath297 is supported by @xmath474 smooth components , we finally get @xmath475 components having pairwise tangency of order @xmath476 at the origin .
this contradicts lemma [ lbu ] below .
[ lbu ] let @xmath86 be a two - dimensional smooth germ , the support of @xmath87 consisting of @xmath477 smooth germs having two - by - two contact order @xmath478 at the origin and each of multiplicity @xmath369 ( @xmath479 )
. then @xmath86 is klt if and only if @xmath480 the solutions @xmath481 are : we get a log - resolution of this pair in performing @xmath494 successive suitable point blow - ups with successive exceptional divisors @xmath495 .
if @xmath496 is the resulting composition one checks that in this process @xmath497 while @xmath498 where @xmath297 is the strict transform of @xmath87 .
we thus see that the initial pair is klt if and only if the coefficient of @xmath499 in @xmath500 is strictly greater than @xmath501 , _ i.e. _ if @xmath480 the solutions are then easily determined . ( 120,20 ) ( 1,1)(1,1)14(1,15)(1,-1)14(8,1)(0,1)14 ( 0,5)@xmath381(7,16)@xmath405(13,5)@xmath502 ( 24,1)(31,6)(31,15)(31,15)(31,6)(38,1)(31,15 ) ( 28,17)@xmath415(27,1)@xmath263 ( 48,1)(58,8)(48,15)(58,1)(48,8)(58,15)(53,8)(48,8)(1,0)10 ( 54,9)@xmath503(46,8)@xmath504(46,16)@xmath263(57,16)@xmath1 ( 68,1)(75,6)(75,15)(75,15)(75,6)(82,1)(75,15)(68,15)(1,0)14 ( 72,17)@xmath430(83,15)@xmath263(71,1)@xmath102 ( 92,1)(99,6)(99,15)(99,15)(99,6)(106,1)(99,15)(99,1)(0,1)14 ( 96,17)@xmath435(100,1)@xmath102(105,3)@xmath102 to prove proposition [ orbi lisse ] we are thus left with the case where @xmath0 is smooth and @xmath509 , the @xmath370 being smooth at the origin . we know
that @xmath391 .
when @xmath429 any @xmath381 leads to the klt situation . when @xmath432 the possible situations are described in the lemma [ lbu ] .
we thus assume that @xmath421 and that not all three components are normal crossings .
we shall then show that ( after reordering ) @xmath308 and @xmath309 are tangent to order @xmath510 while @xmath511 is transversal to them and that @xmath512 .
so that one has either @xmath464 , @xmath502 and @xmath494 arbitrary or @xmath513 and @xmath514 .
assume indeed that @xmath308 and @xmath309 have contact order @xmath510 .
we first remark that @xmath511 is transversal to them .
assume not , then after possible reordering of the components , we may assume that the order of contact of @xmath511 with @xmath308 and @xmath309 is at least @xmath494 .
in this situation , the proof of the above lemma [ lbu ] still applies to show that @xmath515 , which is impossible since @xmath516 .
the conclusion now simply follows from considering the orbifold tale cover ramified to order @xmath502 along @xmath511 , which replaces @xmath87 by @xmath297 , supported by two smooth components of multiplicities @xmath381 and @xmath405 and having contact order @xmath517 .
this implies by lemma [ lbu ] that @xmath512 since the inequality @xmath518 is equivalent to @xmath519 when @xmath432 .
the solutions are the classical ones .
wolf p. barth , klaus hulek , chris a. m. peters , and antonius van de ven , _ compact complex surfaces _
, second ed .
, ergebnisse der mathematik und ihrer grenzgebiete .
3 . folge .
a series of modern surveys in mathematics [ results in mathematics and related areas .
3rd series .
a series of modern surveys in mathematics ] , vol . 4 , springer - verlag , berlin , 2004 .
jnos kollr and shigefumi mori , _ birational geometry of algebraic varieties _ , cambridge tracts in mathematics , vol .
134 , cambridge university press , cambridge , 1998 , with the collaboration of c. h. clemens and a. corti , translated from the 1998 japanese original .
k. ueno , _ classification theory of algebraic varieties and compact complex spaces _ , springer - verlag , berlin , 1975 , notes written in collaboration with p. cherenack , lecture notes in mathematics , vol . 439 . | using orbifold metrics of the appropriately signed ricci curvature on orbifolds with negative or numerically trivial canonical bundle and the two - dimensional log minimal model program , we prove that the fundamental group of special compact khler threefolds is almost abelian .
this property was conjectured in all dimensions in @xcite , and also for orbifolds in @xcite , where the notion of specialness was introduced .
we briefly recall below the definition , basic properties , and the role of special manifolds in classification theory . _
dedicated to v. shokurov . _ |
let @xmath3 be a real analytic surface without border ( compact or not ) , and @xmath1 a real analytic vector field on it .
an open period annulus of @xmath1 is an union of period orbits of @xmath1 which is bi - analytic to the standard open annulus @xmath4 , the image of each circle @xmath5 being a periodic orbit of @xmath1 .
let @xmath2 , @xmath6 ) be an analytic family of analytic vector fields on @xmath3 , and let @xmath0 be an open period annulus of @xmath1 .
the _ cyclicity _
@xmath7 of @xmath0 with respect to the deformation @xmath2 is the maximal number of limit cycles of @xmath2 which tend to @xmath0 as @xmath8 tends to zero , see definition [ cyclicity ] bellow .
clearly the vector field @xmath1 has an analytic first integral @xmath9 in the period annulus @xmath0 which has no critical points . in
what follows we shall suppose that the open period annulus @xmath0 is relatively compact ( i.e. its closure @xmath10 is compact ) .
we shall say that @xmath1 is a hamiltonian vector field provided that it has a first integral with isolated critical points in a complex neighborhood of @xmath0 .
we shall say that @xmath1 is a generic darbouxian vector field provided that all singular points of @xmath1 in a neighborhood of @xmath11 are orbitally analytically equivalent to linear saddles @xmath12 with @xmath13 . in the case
when @xmath1 is a generic darbouxian vector field , as we shall see in the next section , it can be covered by a planar darbouxian vector field with a first integral of the `` darboux type '' @xmath14 for some analytic functions @xmath15 in a complex neighborhood of @xmath0 .
the main result of the paper are the following [ main1 ] the cyclicity @xmath7 of the open period annulus @xmath0 of a hamiltonian vector field @xmath1 is finite .
[ main2 ] the cyclicity @xmath7 of the open period annulus @xmath0 of a generic darbouxian vector field @xmath1 is finite .
the above theorems are a particular case of the roussarie s conjecture @xcite which claims that the cyclicity @xmath16 of every compact invariant set @xmath17 of @xmath1 is finite . indeed , as @xmath0 is relatively compact , then @xmath18 .
the finite cyclicity of the open period annulus without the assumptions of theorems [ main1 ] and [ main2 ] is an open question . to prove the finite cyclicity we note first that it suffices to show the finite cyclicity of a given one - parameter deformation @xmath19 .
this argument is based on the hironaka s desingularization theorem , see @xcite .
consider the first return map associated to @xmath0 and @xmath19 @xmath20 the cyclicity of the open period annulus @xmath0 is finite if and only if the poincar - pontryagin function @xmath21 has a finite number of zeros in @xmath22 .
it has been shown in @xcite that @xmath21 allows an integral representation as a linear combination of iterated path integrals along the ovals of @xmath0 of length at most @xmath23 .
the finite cyclicity follows then from the non - accumulation of zeros of such iterated integrals at @xmath24 and @xmath25 .
the proof of this fact will be different in the hamiltonian and in the generic darbouxian case .
in the hamiltonian case we observe that @xmath21 satisfies a fuchsian equation @xcite .
we prove in section [ proof1 ] that the associated monodromy representation is quasi - unipotent , which implies the desired property .
in the darbouxian case the above argument does not apply ( there is no fuchsian equation satisfied by @xmath21 ) .
we prove the non - oscillation property of an iterated integral by making use of its mellin transformation , along the lines of @xcite .
it seems to be difficult to remove the genericity assumption in the darbouxian case ( without this the hamiltonian case is a sub - case of the darbouxian one ) .
the paper is organized as follows . in section [ ppm ]
we recall the definition of cyclicity and the reduction of multi - parameter to one - parameter deformations . in section [ plane ]
we reduce the case of a vector field on a surface to the case of a plane vector field .
theorem [ main1 ] is proved in section [ proof1 ] according to the scheme @xmath26 @xmath27 @xmath28 theorem [ main2 ] is proved in section [ proof2 ] of the paper .
* acknowledgments . *
part of this paper was written while the second author was visiting the university of toulouse ( france ) .
he is obliged for the hospitality .
[ cyclicity ] let @xmath2 be a family of analytic real vector fields on a surface @xmath3 , depending analytically on a parameter @xmath29 , and let @xmath30 be a compact invariant set of @xmath31 .
we say that the pair @xmath32 ) has cyclicity @xmath33 with respect to the deformation @xmath2 , provided that @xmath34 is the smallest integer having the property : there exists @xmath35 and a neighborhood @xmath36 of @xmath37 , such that for every @xmath8 , such that @xmath38 , the vector field @xmath2 has no more than n limit cycles contained in @xmath36 .
if @xmath39 is an invariant set of @xmath31 ( possibly non - compact ) , then the cyclicity of the pair @xmath40 with respect to the deformation @xmath2 is @xmath41 the above definition implies that when @xmath42 is an open invariant set , then its cyclicity @xmath43 is the maximal number of limit cycles which tend to @xmath39 as @xmath8 tends to @xmath24 . to simplify the notation , and if there is no danger of confusion , we shall write @xmath44 on the place of @xmath45 let @xmath46 . one can easily see that @xmath47 has finite number of isolated positive zeros for each @xmath48 , and this number tends to infinity as @xmath49 . below
we construct a germ @xmath19 of a vector field having a monodromic planar singular point at the origin , with a return map @xmath50 .
since isolated singular points of the return map correspond to limit cycles , we see that the vector field @xmath19 has a finite number of limit cycles for each @xmath48 , and this number tends to infinity as @xmath48 tends to zero .
so the cyclicity of the open period annulus @xmath51 is infinity .
note that , however , the vector field @xmath19 is not analytic at the origin . here is a construction : on the strip @xmath52\times \r$ ] consider the equivalence relation @xmath53 .
let @xmath54 be the corresponding projection , and define @xmath55 .
one can check that for @xmath56 small enough thus defined @xmath57 is a blow - up of a smooth vector field @xmath19 defined near the origin , and the return map of @xmath19 is as prescribed by construction .
let @xmath58 be a cross - section of the period annulus @xmath0 which can be identified to the interval @xmath22 .
choose a local parameter @xmath59 on @xmath60 .
let @xmath61 be the first return map and @xmath62 the displacement function of @xmath2 .
for every closed interval @xmath63\subset \delta$ ] there exists @xmath35 such that the displacement function @xmath64 is well defined and analytic in @xmath65 . for every fixed @xmath8
there is a one - to - one correspondance between isolated zeros of @xmath64 and limit cycles of the vector field @xmath2 .
let @xmath66 and let us expand @xmath67 we define the bautin ideal @xmath68 of @xmath2 to be the ideal generated by the germs @xmath69 of @xmath70 in the local ring @xmath71 of analytic germs of functions at @xmath72 .
this ideal is noetherian .
let @xmath73 be a minimal system of its generators , where @xmath74 , and @xmath75 is the maximal ideal of the local ring @xmath71 .
let @xmath76 be analytic functions representing the generators of the bautin ideal in a neighborhood of the origin in @xmath77 .
[ bautin ] the bautin ideal does not depend on the point @xmath78 . for every @xmath63 \subset \delta$ ]
there is an open neighborhood @xmath79 of @xmath63\times \{0\}$ ] in @xmath80 and analytic functions @xmath81 in @xmath79 , such that @xmath82 the real vector space generated by the functions @xmath83 $ ] is of dimension @xmath84 .
suppose that the bautin ideal is principal and generated by @xmath85 .
then @xmath86 where @xmath87 .
the maximal number of the isolated zeros of @xmath88 on a closed interval @xmath63\subset ( 0,1)$ ] for sufficiently small @xmath89 is bounded by the number of the zeros of @xmath90 , counted with multiplicity , on @xmath63 $ ] .
this follows from the weierstrass preparation theorem , properly applied , see @xcite . therefore to prove the finite cyclicity of @xmath0 it is enough to show that @xmath90 has a finite number of zeros on @xmath22 .
consider a germ of analytic curve @xmath91 , @xmath92 , as well the analytic one - parameter family of vector fields @xmath93 .
the bautin ideal is principal , @xmath94 , and @xmath95 where the dots stay for terms containing @xmath96 , @xmath97 .
@xmath21 is the so called @xmath23th order higher poincar - pontryagin - melnikov function associated to the one - parameter deformation @xmath93 of the vector field @xmath1 .
therefore , if the cyclicity of the open period annulus is infinite , then @xmath98 has an infinite number of zeros on the interval @xmath22 of course , in general the bautin ideal is not principal .
however , by making use of the hironaka s theorem , we can always principalize it . more precisely , after several blow up s of the origin of the parameter space , we can replace the bautin ideal by an ideal sheaf which is principal , see @xcite for details .
this proves the following [ ppm ] if the cyclicity @xmath44 of the open period annulus @xmath0 is infinite , then there exists a one parameter deformation @xmath99 , such that the corresponding higher order poincar - pontryagin - melnikov function @xmath98 has an infinite number of zeros on the interval @xmath22 . in the next two sections we shall prove the non - oscillation property of @xmath21 in the hamiltonian and the darbouxian case ( under the restrictions stated in theorem [ main2 ] ) .
let @xmath1 be a real analytic vector field on a real analytic surface @xmath3 .
let @xmath0 be an open period annulus of @xmath1 with compact closure .
let the map @xmath100 be a bi - analytic isomorphism , such that @xmath101 is a closed orbit of @xmath1 .
we assume that @xmath1 is either hamiltonian or generalized darbouxian in some neighborhood of the closure @xmath11 of @xmath0 .
theorems [ main1 ] and [ main2 ] claims that cyclicity of @xmath0 in any family of analytic deformation @xmath102 of @xmath1 is finite .
this paragraph is devoted to the reduction of this general situation to the case of a vector field @xmath1 on @xmath103 of hamiltonian or darboux type near its polycycle
. then theorem [ main1 ] and theorem [ main2 ] follow from theorem [ main1 ] and theorem [ main2 ] below .
first , note that it is enough to prove finite cyclicity of @xmath104 only .
indeed , finite cyclicity of @xmath105\right)$ ] follows from gabrielov s theorem , and finite cyclicity of @xmath106 can be reduced to the above by replacing @xmath107 by @xmath108 .
consider the hausdorf limit @xmath109 .
it is a connected union of several fixed points @xmath110 of @xmath1 ( not necessarily pairwise different ) and orbits @xmath111 of @xmath1 such that @xmath112 exits from @xmath70 and enters @xmath113 ( where @xmath114 denotes @xmath115 ) . from now on
we consider only a sufficiently small neighborhood @xmath79 of @xmath17 .
we assume that @xmath116 , and denote this intersection again by @xmath0 .
we consider first the darbouxian case .
note that @xmath17 can not consist of just one singular point of @xmath1 by assumption about linearizability of singular points of @xmath1 in this case .
[ lem : covering ] assume that theorem [ main2 ] holds if @xmath79 is orientable and all @xmath70 are different .
then theorem [ main2 ] holds in full generality .
assume that for some real analytic surface @xmath117 there is an analytic mapping @xmath118 which is a finite covering on @xmath0 .
then the cyclicity of @xmath0 for @xmath2 is the same as cyclicity of @xmath119 for the lifting @xmath2 to @xmath117 .
the claim of the lemma follows from this principle applied to two types of coverings below .
first , taking a double covering of @xmath79 as @xmath117 , we can assume that @xmath79 is orientable .
second , let @xmath79 be represented as a union of neighborhoods @xmath120 of @xmath70 together with neighborhoods @xmath121 of @xmath112 .
glue @xmath117 as @xmath122 , where @xmath123 are bianalytically equivalent to @xmath120 and disjoint , and @xmath124 are bianalytically equivalent to @xmath121 , with natural glueing of @xmath123 to @xmath124 , of @xmath124 to @xmath125 and of of @xmath126 to @xmath127 . in other words
, @xmath118 is one - to - one away from @xmath70 and @xmath128-to - one in a neighborhood of @xmath70 if @xmath70 appears @xmath128 times in the list @xmath129 .
evidently , @xmath130 is one - to - one on @xmath0 , so is bianalytic .
= 0.25 we will now define a first integral @xmath131 of @xmath1 in @xmath79 .
take any non - singular point @xmath132 , and let @xmath131 be a local first integral of @xmath1 in a neighborhood @xmath133 of @xmath134 such that @xmath135 and @xmath136 .
since @xmath79 is orientable , @xmath0 lies from one side of @xmath17 , and we can assume that intersection of @xmath133 with each cycle @xmath137 is connected .
this allows to extend @xmath131 to a first integral of @xmath1 defined on @xmath138 . changing sign of @xmath131 if necessary
, we can assume that @xmath139 on @xmath140 .
we define @xmath141 by continuity .
[ lem : h extended ] extension of @xmath131 to @xmath138 by flow of @xmath1 can be extended to a multivalued holomorphic function defined in a neighborhood of @xmath17 in a complexification of @xmath79 .
first , @xmath131 is analytic in some neighborhood of @xmath142 , as it is an analytic function extended by analytic flow of @xmath1 .
choose local linearizing coordinates @xmath143 near @xmath144 in such a way that @xmath145 .
by assumption , @xmath146 is the local first integral of @xmath1 near @xmath144 .
therefore @xmath147 , and , restricting to a transversal @xmath148 , one can see that @xmath9 is analytic and invertible .
therefore @xmath131 can be extended to a neighborhood of @xmath144 .
moreover , @xmath149 is an analytic local first integral near the point @xmath150 of @xmath151 .
therefore it can be extended to a neighborhood of @xmath151 ( here we use that @xmath79 is orientable , so @xmath151 is different from @xmath142 ) , and , as above , to a neighborhood of @xmath152 ( here we use that @xmath153 ) , and so on .
note that from the above construction follows that near each @xmath112 the first integral @xmath131 is equal , up to an invertible function , to @xmath154 , where @xmath155 is a local equation of @xmath156 .
also , near any singular point of @xmath17 the first integral @xmath131 is equal , up to an invertible function , to @xmath157 .
[ cor : dlogh is merom ] the one - form @xmath158 is meromorphic one - form in @xmath79 with logarithmic singularities only .
assume that @xmath159 .
one can easily construct a @xmath160 isomorphism of a sufficiently small neighborhood @xmath79 of @xmath17 with a neighborhood of a regular @xmath161-gone in @xmath162 in such a way that the image of @xmath138 will lie inside the @xmath161-gone and image of @xmath17 coincides with the @xmath161-gone .
due to @xcite , some neighborhood @xmath163 of @xmath79 in its complexification is a stein manifold .
this implies that this isomorphism can be chosen bianalytic .
similarly , for @xmath164 one can map bianalytically a neighborhood of @xmath79 to a union of two arcs @xmath165 , which , for the rest of the paper , will be called `` regular 2-gone '' .
we transfer everything to plane using this isomorphism and will denote the images on plane of the previously defined objects by the same letters .
the first integral @xmath131 takes the form @xmath166 , where @xmath15 are analytic functions in @xmath79 with @xmath167 , @xmath168 is an analytic functions non - vanishing in its neighborhood @xmath79 and @xmath169 .
note that @xmath139 in the part of @xmath79 lying inside the @xmath161-gone .
further we assume that @xmath170 , so @xmath171 ( one can achieve this by e.g. taking @xmath172 instead of @xmath173 ) . the family @xmath2 becomes a family of planar analytic vector fields defined in a neighborhood @xmath79 of a regular @xmath161-gone @xmath174 , and @xmath1 has a first integral @xmath131 of darboux type in @xmath79 .
let @xmath175 be a one - parametric deformation of @xmath1 as in proposition [ ppm ] .
define meromorphic forms @xmath176 , @xmath177 as @xmath178 according to ( * ? ? ?
* theorem 2.1 ) , @xmath21 can be represented as a linear combination of iterated integrals over @xmath179 of forms which are combinations of gauss - manin derivatives of @xmath180 .
recall that the gauss - manin derivative of a form @xmath181 is defined as a form @xmath182 such that @xmath183 . in general ,
@xmath182 can not be uniquely defined from this equation , though its restrictions to @xmath179 are defined unambiguously . however , since @xmath163 is stein , in our situation one can choose a meromorphic in @xmath79 representative of @xmath182 , with poles on @xmath184 only ( where @xmath184 is the union of lines containing sides of @xmath17 ) .
therefore theorem [ main2 ] follows from the following claim [ main2 ] let @xmath14 be as above , and let @xmath185 be the connected component of its level set lying inside @xmath17 .
zeros of polynomials in iterated integrals @xmath186 corresponding to meromorphic one - forms @xmath187 with poles in @xmath184 can not accumulate to @xmath24 . from the above discussion
it is clear that theorem [ main1 ] follows on its turn from the following [ main1 ] let @xmath188 where @xmath131 is a real analytic function with isolated singularities in some complex neighborhood of the closed period annulus @xmath189 , where @xmath185 is the connected component of the level set of @xmath131 lying inside @xmath17 .
zeros of the first non - vanishing poincar - pontryagin function @xmath21 , corresponding to a one - parameter analytic deformation @xmath19 of @xmath1 can not accumulate to @xmath24 .
here shall prove theorem [ main1 ] .
this follows from the following two results [ th2 ] the poincar - pontryagin function @xmath21 satisfies a linear differential equation of a fuchs type in a suitable complex neighborhood of @xmath190 .
[ th3 ] the monodromy operator of the above fuchs equation corresponding to a loop encircling the origin in @xmath191 is quasi - unipotent .
let us recall that an endomorphism is called unipotent , if all its eigenvalues are equal to @xmath25 , and quasi - unipotent if all of them are roots of the unity .
the above theorems imply that the poincar - pontryagin - melnikov function has a representation in a neighborhood of @xmath192 @xmath193 where @xmath194 , @xmath195 , and @xmath196 are functions analytic in a neighborhood of @xmath192 .
this shows that the zeros of @xmath197 do not accumulate to @xmath24 . of course
, similar arguments hold in a neighborhood of @xmath198 , so @xmath21 has a finite number of zeros on @xmath22 .
this completes the proof of theorem [ main2 ] in the hamiltonian case . to the end of the section
we prove theorem [ th3 ] .
the open real surface @xmath3 is analytic and hence possesses a canonical complexification .
similarly , any analytic family of analytic vector fields @xmath2 is extended to a complex family of vector fields , depending on a complex parameter . in this section , by abuse of notation , the base field will be @xmath191 .
a real object and its complexification will be denoted by the same letter .
let @xmath199 be an open complex neighborhood of @xmath200 in which the complexified vector field @xmath1 has an analytic first integral @xmath9 with isolated critical points .
the restriction of @xmath9 on the interval @xmath22 ( after identifying @xmath0 to @xmath4 ) is a local variable with finite limits at @xmath24 and @xmath25 therefore we may suppose that @xmath201 , @xmath202 , and the restriction of @xmath9 to @xmath22 is the canonical local variable on @xmath203 .
the function @xmath9 defines a locally trivial milnor fibration in a neighborhood of every isolated critical point .
there exists a complex neighborhood @xmath79 of @xmath200 in which @xmath204 has only isolated critical points .
moreover the compactness of @xmath200 implies that there exists a complex neighborhood @xmath205 of the origin , homeomorphic to a disc , such that the fibration @xmath206 is locally trivial , and the fibers @xmath207 are open riemann surfaces homotopy equivalent to a bouquet of a finite number of circles .
consider a one - parameter analytic deformation @xmath19 of the vector field @xmath1 .
as @xmath9 is a first integral of @xmath1 , then there exists an unique symplectic two - form @xmath176 , such that @xmath208 indeed , if in local coordinates @xmath209 then @xmath210 implies @xmath211 , where @xmath8 is analytic in @xmath79 and non - vanishing in @xmath0 .
it follows that @xmath212 define a unique meromorphic one - form @xmath213 by the formula @xmath214 the one form @xmath215 is meromorphic in @xmath79 , depends analytically on @xmath216 , and @xmath217 .
its pole divisor does not depend on @xmath216 as in the local variables above it is defined by @xmath218 .
therefore @xmath219 where @xmath220 are given meromorphic one - forms in @xmath79 with a common pole divisor which does not intersect the period annulus @xmath0 . in the complement of the singular locus of @xmath19 the vector field @xmath19 and the one form @xmath221 define the same foliation , and therefore define the same first return map associated to @xmath0 .
denote this map by @xmath222 , where @xmath223 is the restriction of @xmath9 to a cross - section of the period annulus @xmath0 ( this does not depend on the choice of the cross - section ) .
we have @xmath224 on each leaf of the foliation defined by @xmath19 we have @xmath225 which implies @xmath226 where @xmath227 is the family of periodic orbits ( with appropriate orientation ) of @xmath1 , @xmath228 , @xcite .
thus the first poincar - pontryagin - melnikov function is an abelian integral and its monodromy representation is straightforward .
namely , the meromorphic one - form @xmath229 restricts to a meromorphic one - form on the fibers of the minlor fibration ( [ milnor ] ) .
we may also suppose that @xmath230 has a finite number of poles @xmath231 ( after choosing appropriately the domain @xmath79 ) .
denote @xmath232 the milnor fibration ( [ milnor ] ) induces a representation @xmath233 which implies the monodromy representation of @xmath234 .
suppose first that @xmath229 is analytic in @xmath79 .
it is well known that the operator of the classical monodromy of an isolated critical point of an analytic function is quasi - unipotent , e.g. @xcite .
therefore the representation in @xmath235 of a small loop about @xmath24 in @xmath236 is quasi - unipotent .
more generally , let @xmath229 be meromorphic one - form with a finite number of poles on the fibers @xmath237 .
a monodromy operator permutes the poles and hence an appropriate power of it leaves the poles fixed . therefore this operator is quasi - unipotent too and theorem [ th2 ] is proved in the case @xmath238 .
of course , it is well known that an abelian integral has a finite number of zeros @xcite .
let @xmath21 be the first non - zero poincar - pontryagin - melnikov function .
its `` universal '' monodromy representation was constructed in @xcite .
for convenience of the reader we reproduce it here .
recall first that @xmath239 depends on the _ free _ homotopy class of of the loop @xmath240 in @xmath241 ( * ? ? ?
* proposition 1 ) and that this property does not hold true for the first return map @xmath222 ( which depends on the homotopy class of @xmath240 in @xmath242 ) .
let @xmath243 be the fundamental group of @xmath244 .
it is a finitely generated free group .
let @xmath245 be the orbit of the loop @xmath240 under the action of @xmath246 induced by ( [ milnor ] ) .
the set @xmath247 generates a normal subgroup of @xmath204 which we denote by @xmath248 .
the commutator subgroup @xmath249 is the normal sub - group of @xmath204 generated by commutators @xmath250 .
the milnor fibration ( [ milnor ] ) induces a representation @xmath251 according to ( * ? ? ?
* theorem 1 ) , the monodromy representation of @xmath21 is a sub - representation of the monodromy representation dual to ( [ fmilnor ] ) . unfortunately the free abelian group @xmath252 is not necessarily of finite dimension . to obtain a finite - dimensional representation we use the fundamental fact that @xmath21 has an integral representation as an iterated path integral of length @xmath23 ( * ? ?
* theorem 2.1 ) . to use this ,
define by induction @xmath253 , @xmath254 .
we will later consider the associated graded group @xmath255 it is well know that an iterated integral of length @xmath23 along a loop contained in @xmath256 vanishes identically .
therefore , to study the monodromy representation of @xmath21 , we shall truncate with respect to @xmath256 and obtain a finite - dimensional representation .
namely , for every subgroup @xmath257 we denote @xmath258 the representation ( [ fmilnor ] ) induces a homomorphism @xmath259 and the monodromy representation of @xmath21 is a sub - representation of the representation dual to ( [ monodromy2 ] ) @xcite .
the abelian group @xmath260 is , however , finitely generated . indeed
the lower central series of @xmath261 is @xmath262 and hence @xmath263 is a finitely generated nilpotent group .
each sub - group of such a group is finitely generated too , e.g. @xcite .
the central result of this section is the following theorem , from which theorem [ th3 ] follows immediately [ th4 ] the monodromy representation ( [ monodromy2 ] ) is quasi - unipotent .
indeed , @xmath21 satisfies a fuchsian equation on @xmath264 , whose monodromy representation is a sub - representation of the representation dual to ( [ monodromy2 ] ) ( * ? ? ?
* theorem 1.1 ) and ( * ? ? ? * theorem 1 ) . to prove proposition [ th4 ] we recall first some basic facts from the theory of free groups , e.g. serre @xcite , hall @xcite .
the graded group @xmath265 ( [ graded ] ) associated to the free finitely generated group @xmath204 is a lie algebra with a bracket induced by the commutator @xmath266 on @xmath204 .
the milnor fibration ( [ milnor ] ) induces a representation @xmath267 where @xmath268 is the group of lie algebra automorphisms of @xmath269 .
let @xmath270 be a generator of @xmath271 .
it induces automorphisms @xmath272 and @xmath273 .
we note that @xmath274 and hence @xmath275 is quasi - unipotent .
[ quasi ] let @xmath276 be such that @xmath275 is quasi - unipotent .
then for every @xmath277 the automorphism @xmath278 is quasi - unipotent .
the proof is by induction .
let @xmath279 be the free generators of @xmath204 and consider the free lie algebra @xmath280 on @xmath281 .
it is a lie sub - algebra of the associative non - commutative algebra of polynomials in the variables @xmath282 with a lie bracket @xmath283=xy - yx$ ] .
the canonical map @xmath284 $ ] induces an isomorphism of lie algebras @xmath285 , ( * ? ? ? * theorem 6.1 ) .
let @xmath286 be the graded piece of degree @xmath23 .
we shall show that @xmath287 is quasi - unipotent .
the proof is by induction .
suppose that the restriction of @xmath288 on @xmath289 is quasi - unipotent , i.e. for some @xmath290 , the restriction of @xmath291 on @xmath292 is @xmath24 .
the operator @xmath293 is a linear automorphism , but not a lie algebra automorphism .
the identity @xmath294 & = & ( l_*^p - id)(xy - yx)= l_*^p \,x l_*^py- l_*^p \,y l_*^px -xy + yx \\ & = & [ var _ * x , var _ * y]+ [ var _ * x , y]+ [ x , var _ * y]\end{aligned}\ ] ] shows that the restriction of @xmath295 on @xmath296 vanishes identically .
therefore the automorphism @xmath288 restricted to @xmath296 or @xmath297 is quasi - unipotent .
the case @xmath298 is similar .
proposition [ quasi ] is proved.@xmath299 according to the above proposition for every @xmath300 there are integers @xmath301 , such that the polynomial @xmath302 annihilates @xmath278 . proposition [ th4 ] will follow on its hand from the following [ prop3 ] the polynomial @xmath303 annihilates @xmath304 .
* let @xmath305 .
it induces an automorphism of the abelian groups @xmath306 denoted , by abuse of notation , by @xmath288 .
we denote by @xmath307 the corresponding homomorphisms .
it follows from the definitions that the diagram ( [ diagram ] ) of abelian groups , is commutative ( the vertical arrows are induced by the canonical projections ) .
therefore if an equivalence classe @xmath308 \in g/(g , f)$ ] can be represented by a closed loop @xmath309 , then @xmath310 $ ] can be represented by a closed loop in @xmath311 .
therefore for every @xmath308 \in g/(g , f)$ ] , the equivalence class @xmath312 can be represented by a closed loop in @xmath256 . in other words
@xmath312 indices the zero automorphism of @xmath313.@xmath299 @xmath314^{p_i(l _ * ) } & f_i/(f_i , f ) \\
g \cap f_i / ( g \cap f_i , f ) \ar[r]^{p_i(l _ * ) } \ar[u]_{\pi_2 } \ar[d ] _ { \pi_1 } & g \cap f_i / ( g \cap f_i , f ) \ar[u]_{\pi_2 } \ar[d ] _ { \pi_1 } \\
g/(g , f ) \ar[r]^{p_i(l _ * ) } & g/(g , f ) \\}\ ] ]
in this section we prove theorem [ main2 ] .
first , we consider _ elementary iterated integrals _ - the iterated integrals over the piece of the cycle lying near the saddles .
we give a representation of the mellin transform of the elementary iterated integral as a converging multiple series .
this representation provides an asymptotic series for the elementary iterated integral , with some explicit estimate of the error , see theorem [ thm : elemint ] below .
the general iterated integral of length @xmath23 turns out to be a polynomial ( depending on @xmath1 and @xmath23 only ) in elementary iterated integrals , by lemma [ lem : mellin of iterated ] .
we give analogue of the estimates of theorem [ thm : elemint ] for such polynomials .
this allows to prove a quasianalyticity property : if the asymptotic series corresponding to the iterated integral is zero , then the integral itself is zero .
this implies theorem [ main2 ] since the zeros of the partial sums of the asymptotic series do not accumulate to @xmath24 , see corollary [ cor : quasian elemint ] .
[ [ iterated - integral - as - a - polynomial - in - elementary - iterated - integrals . ] ] iterated integral as a polynomial in elementary iterated integrals .
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ let @xmath315 , @xmath316 $ ] , be a parameterization of the cycle @xmath317 ( we fix some @xmath318 for a moment ) .
as in @xcite , the cycle of integration can be split into several pieces @xmath319 , those lying near the sides of the polycycle , and those near the vertices .
we assume that the vector field can be linearized in the charts containing these pieces , and call these pieces elementary .
let @xmath320 be the parameterization of the ends of these pieces .
consider connected components of the complement of @xmath60 to the union of hyperplanes @xmath323 .
each connected component can be defined as @xmath324 i.e. is a product @xmath325 of several simplices of smaller dimension of the form @xmath326 .
therefore , by fubini theorem , integral of @xmath327 over this connected component is equal to the product of integrals @xmath328 , i.e. to the product of iterated integrals @xmath329 .
there are two types of elementary pieces : those lying in charts covering sides of the polycycle , and those lying in charts covering saddles . similarly to @xcite , the elementary iterated integrals corresponding to the pieces of the first type are just meromorphic functions of the parameter on the transversal , i.e. of @xmath330 .
we give description of iterated integrals in terms of their mellin transforms .
recall that the mellin transform of a function @xmath333 on the interval @xmath334 $ ] is defined as @xmath335 . to describe the mellin transform of the elementary iterated integrals over @xmath336
let us introduce a generalized compensator .
we denote in this section by @xmath270 the length of the elementary iterated integral . for @xmath337 and @xmath338
we define @xmath339 as @xmath340 we call @xmath341 a generalized compensator . particular case of @xmath342 corresponds to the ecalle - roussarie compensator .
generalized compensator is a finite linear combination of monomials of type @xmath343 , for @xmath344 .
[ lem : mellin elemint ] after some rescaling of @xmath107 the mellin transform of an elementary iterated integral is given by the following formula : @xmath346 where @xmath347 is an upper bound for the order of poles of @xmath220 .
moreover , @xmath348 . in the linearizing
coordinates the first integral is written as @xmath349 .
the mellin transform of the iterated integral can be computed explicitly for monomial forms @xmath350 : @xmath351 similar formula holds for @xmath352 .
after rescaling of @xmath131 we can assume that the linearizing chart covers the bidisk @xmath353 .
then the coefficients of the forms @xmath220 are meromorphic in the bidisk , with poles on @xmath354 of order at most @xmath347 .
so @xmath220 can be represented as a convergent power series @xmath355 with coefficients @xmath356 decreasing as @xmath357 .
therefore the elementary iterated integral is a converging sum of elementary iterated integrals of monomial forms , with coefficients being products of @xmath356 , @xmath358 and @xmath359 . from ( [ eq : elem monomial int ] ) one gets upper bounds for the elementary iterated integrals of monomial forms , which guarantees that one can perform mellin transform termwise , and we get the required formula . as in @xcite
, one can check that the inverse mellin transform of mellin transforms of elementary iterated integrals can be defined as @xmath360 where @xmath347 is sufficiently big .
indeed , @xmath361 on @xmath0 , so ( [ eq : mellin elemint ] ) converges uniformly on this contour , so one can integrate the series ( [ eq : elem monomial int ] ) termwise .
however , for each term ( [ eq : invmellin ] ) does define the inverse mellin transform , as each term is just a rational function in @xmath362 .
[ lem : upper bound elemint ] let @xmath364 be an elementary iterated integral , and let @xmath365 be defined as in [ lem : mellin elemint ] .
for any @xmath366 denote by @xmath367 the minimal distance from @xmath3 to the poles of @xmath368 .
inverse mellin transform of @xmath372 is a linear combination of monomials of the type @xmath373 , where @xmath374 and @xmath375 .
collecting similar terms in the expression for the elementary iterated integral @xmath376 together , we get a formal series @xmath377 of such terms , possibly divergent : @xmath378 [ thm : elemint ] @xmath377 is an asymptotic series of @xmath376
. moreover , for each @xmath379 there exists @xmath380 $ ] such that the partial sums @xmath381 of @xmath377 satisfy the following : latexmath:[\[\label{eq : tail elemint } poles of @xmath368 are of the form @xmath383 . since @xmath384 , there are @xmath385 poles on the interval @xmath386 $ ] , @xmath379 .
therefore on each interval @xmath387 one can find a point @xmath388 such that @xmath389 . for each @xmath390
let us split the contour of integration @xmath391 into two parts : boundary of @xmath392 and boundary of @xmath393 .
computing residues , we see that @xmath394 is a partial sum @xmath395 of @xmath377 as defined above
. therefore @xmath396 . by lemma [ lem : upper bound elemint ] ,
@xmath397 on @xmath398 , and ( [ eq : tail elemint ] ) follows .
let @xmath400 be an element in @xmath399 , where @xmath401 $ ] and @xmath402 are elementary integrals .
substitution of convergent series from ( [ eq : elemint sum ] ) instead of @xmath402 gives a representation of @xmath9 as a converging multiple sum of products ( of length at most @xmath23 ) of generalized compensators .
collecting similar terms , we obtain a formal series @xmath403 similar to ( [ eq : formal elemint ] ) , probably divergent .
[ [ mellin - transform - of - a - product - of - several - generalized - compensators ] ] mellin transform of a product of several generalized compensators ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ this is a rational function of @xmath362 .
we want to show that it depends polynomially on @xmath421 .
let @xmath422 denotes the set of function @xmath423 with the condition @xmath424 , and define @xmath425 .
let @xmath434 be the ring of rational functions in @xmath362 vanishing at infinity , and define convolution @xmath435 for @xmath436 by extending the rule @xmath437 by linearity and continuity to the whole @xmath434 ( in particular , @xmath438 ) .
thus defined convolution is mellin - dual to the usual product .
therefore @xmath439 .
decomposing each factor into simple fractions @xmath440 and opening brackets , we see that @xmath441 reducing to a common denominator , we see that @xmath442 is a rational function in @xmath443 , with denominator dividing @xmath444 .
we claim that the factors @xmath445 do not enter denominator of @xmath446 .
indeed , presence of such factor would mean that @xmath442 becomes unbounded as @xmath447 tends to @xmath448 for each @xmath366 , which is not true : for any tuple @xmath432 and every sufficiently big @xmath449 the function @xmath442 is locally bounded near @xmath450 .
let @xmath451 be a product of several elementary iterated integrals , and let order of @xmath452 be @xmath453 . then using representation ( [ eq : elemint sum ] ) for @xmath452 and opening brackets
, we see that @xmath454 where @xmath455 .
let us estimate from above the terms @xmath458 from ( [ eq : monomial sum ] ) .
by lemma [ lem : univpol ] it is equal to @xmath459 , where @xmath460 is defined by @xmath461 as in ( [ eq : compensator ] ) .
this means that @xmath462 for some linear map @xmath463 .
therefore @xmath464 is a polynomial in @xmath465 , and @xmath466 let @xmath376 now be a polynomial in several elementary iterated integrals , @xmath473 .
the set of poles of the mellin transform @xmath368 of @xmath376 is the union of sets of poles of mellin transforms of each monomial of @xmath413 , so the number of poles of @xmath368 on an interval @xmath386 $ ] counted with multiplicities grows as some power of @xmath84 .
this means that for each @xmath390 one can find @xmath474 such that the distance @xmath475 from @xmath84 to the set of poles of @xmath368 will be bigger than @xmath476 for some @xmath477 .
then splitting the contour of integration of the inverse mellin transform as in theorem [ thm : elemint ] , we conclude from lemma [ lem : upper bound monomial ] that @xmath478 on the @xmath479 for some fixed @xmath480 , and the claim follows . 99 v. i. arnold , arnold s problems , springer , 2004 .
l. gavrilov , cyclicity of period annuli and principalization of bautin ideals , _ ergodic th . and
systems _ , 2008 , to appear .
l. gavrilov , i.d .
iliev , the displacement map associated to polynomial unfoldings of planar hamiltonian vector fields , american j. of math . , 127 ( 2005 ) 1153 - 1190 .
l. gavrilov , higher order poincare - pontryagin functions and iterated path integrals , ann .
toulouse math .
( 6 ) 14 ( 2005 ) , no .
663 - 682 .
h. grauert : on levi s problem and the imbedding of real - analytic manifolds .
ann . math .
68 ( 1958 ) 460 - 472 .
m. hall,_the theory of groups _
, ams chelsea publishing , 1976 .
n. katz , nilpotent connections and the monodromy theorem , publ .
i.h.e.s . ,
* 39 * ( 1970 ) 175 - 232 .
khovanskii , real analytic manifolds with the property of finiteness , and complex abelian integrals , funktsional .
i prilozhen .
* 18 * ( 1984 ) , no .
2 , 4050 .
d. novikov , on limit cycles appearing by polynomial perturbation of darbouxian integrable systems , to appear in gafa .
pontryagin , ber autoschwingungssysteme , die den hamiltonischen nahe liegen , _ phys
. z. sowjetunion _ * 6 * ( 1934 ) , 2528 ; on dynamics systems close to hamiltonian systems , zh . eksp . teor . fiz . * 4 * ( 1934 ) 234 - 238 , in russian .
r. roussarie , melnikov functions and bautin ideal .
theory dyn .
syst . 2 ( 2001 ) , no .
1 , 6778 .
r. roussarie , _ bifurcation of planar vector fields and hilbert s sixteenth problem _ ,
progress in mathematics , vol .
164 , birkhuser verlag , basel ( 1998 ) .
r. roussarie , cyclicit finie des lacets et des points cuspidaux , nonlinearity 2 ( 1989 ) , no .
1 , 73117 . j .-
serre , _ lie algebras and lie groups_,*1500 * _ lecture notes in mathematics _ , springer - verlag , berlin , 2006 .
varchenko , estimation of the number of zeros of an abelian integral depending on a parameter , and limit cycles .
i prilozhen . 18 ( 1984 ) , no .
2 , 1425 . | let @xmath0 be an open , relatively compact period annulus of real analytic vector field @xmath1 on an analytic surface .
we prove that the maximal number of limit cycles which bifurcate from @xmath0 under a given multi - parameter analytic deformation @xmath2 of @xmath1 is finite , provided that @xmath1 is either hamiltonian , or generic darbouxian vector field . |
weak magnetic fields threading conducting fluid media can play vital dynamical roles , even when traditional criteria , such as relative magnetic and gas pressures , suggest the fields are entirely negligible .
perhaps the best example of this is the destabilizing influence of a vanishingly small magnetic field crossing a keplerian accretion disk ( @xcite ) , where the mere presence of the field seems fundamentally to alter the local flow properties . other examples abound , however , that could be particularly important in astrophysics . among them , we would include weak fields penetrating turbulent or otherwise strongly unstable flows , where such fields significantly alter evolution and transport properties , ( @xcite ; @xcite ; @xcite ; @xcite ) .
sheared motion is a critical common element in many of these flows , and the consequent stretching of a weak , but large - scale field can lead to a locally enhanced role for the field .
the flows will also frequently lead to current sheets and associated magnetic field topologies unstable to reconnection , and that is central to the nonlinear evolution of the systems ( @xcite ; @xcite ) . through these processes
the fields can also have more global consequences .
study of the nonlinear evolution of the classical kelvin - helmholtz ( kh ) instability could be particularly useful as a well - defined example of strongly sheared flows . further , since kh unstable boundary layers are probably common , the behavior of the instability is important for its own sake .
although the kh instability is fairly well studied in ordinary hydrodynamics ( @xcite ) , comparable study has been much slower in magnetohydrodynamics ( mhd ) .
that is because the magnetic field substantially complicates the physics itself and also because computational methods and resources needed for such studies are only recently up to the task .
the linear analysis of the mhd kh instability is relatively straightforward and was long ago carried out for a number of simple flow and field configurations ( @xcite ; @xcite ) . generally , and especially if the velocity change is not supersonic , the ordinary fluid shear layer is unstable to perturbations with wave vectors in the plane of the shear layer and with wavelengths greater than the thickness of the layer ( @xcite ) .
when there is a field component projecting onto the flow field , magnetic tension provides a stabilizing influence .
a simple vortex sheet is stabilized against linear perturbations whenever the magnetic field strength is sufficient that @xmath4 , where @xmath5 is the velocity difference between the two layers , @xmath6 is the alfvn speed , @xmath7 is the perturbation wave vector , and @xmath8 is the direction of the magnetic field ( @xcite ) .
@xcite ( paper i ) and @xcite ( mbr ) recently presented complementary nonlinear analyses of the mhd kh instability in mildly compressible flows based on two - dimensional numerical simulations carried out with new ( and different ) riemann - solver - based mhd codes . while not the first numerical studies of the mhd kh instability , they represented big improvements over previous calculations in both numerical resolution and extent to which flow evolution was followed towards asymptotic states ( readers are referred to paper i for additional , earlier citations ) .
considering perturbed `` 2d '' flows that were uniform except for a thin , smooth velocity transition layer , those two papers emphasized the qualitatively different behaviors in the nonlinear evolution of unstable flows depending on how close the field strength is to its critical strength for stabilization . for fields only slightly below the critical value , enhancements in the tension of the field through linear growth can stabilize the flow before it develops distinctly nonlinear characteristics . for weaker fields
, however , the initial evolution of the instability is very similar to that for the ordinary kh instability .
that results in the formation of eddies , and hence to substantial stretching of the magnetic field lines as well as reconnection .
paper i emphasized the remarkable fact that in a case with a field 2.5 times weaker than critical , reconnection can lead to self - organization in the flow and fairly rapid relaxation to a quasi - steady laminar and marginally stable flow .
mbr presented summaries of simulations extending to somewhat weaker fields showing evidence for similar behaviors .
neither paper i nor mbr , however , explored the problem in sufficient depth to establish the conditions necessary for the previously mentioned self - organization .
in addition , it is very important in this situation to understand how the magnetic fields behave when they are `` very '' weak ( a concept whose definition needs clarification , in fact ) .
a closely related matter is what differences , if any , exist between the behavior of a truly weak field and a stronger field whose projection onto the flow vectors is weak .
alternatively stated , are there differences between the nonlinear `` 2d '' mhd kh instability and the `` @xmath0d '' mhd kh instability ? answers to those basic questions are the objective of this paper .
we find : 1 ) for the cases we have considered with an initially uniform field that the magnetic field transverse to the plane is unimportant and 2 ) there is a transition from the role of the magnetic field as a catalyst to flow self - organization to a role as an added source of energy dissipation that should vanish directly as the initial magnetic field strength projected onto the plane vanishes .
ultimately we must understand the full `` 3d '' problem , in which the perturbation wave vector also lies outside the flow direction . on the other hand it has been difficult to carry out 3d mhd simulations with sufficient numerical resolution to be confident of the results in complex flows such as these . in addition
, it will be useful to compare fully 3d behaviors with 2d flows .
we hope that the current work is a significant , constructive step towards a full understanding of this problem .
the paper plan is as follows . in 2 we will summarize the problem set - up and relevant results from paper i. 3 contains a discussion of new results , while 4 provides a brief summary and conclusion .
we also include an appendix presenting an analytical model for diffusive flux expulsion from a steady vortex , in order to contrast that physics with what we observe in the eddies that form in our simulations .
in order to focus on specific , important physical issues we have chosen to explore an idealization of the mhd kh problem , reserving for the future the more general problem .
we present in this section only a bare outline of our method and some key results from previous work .
a full discussion of the computational setup along with several tests of such issues as adequate numerical resolution and geometry of the computational box can be found in paper i. the geometry of the computations is shown in figure 1 . the only difference from paper i is that in those earlier computations we assumed an aligned field , @xmath9 , whereas we now relax that constraint to include magnetic fields oblique or orthogonal to the flow plane .
we assume the flow to be periodic in @xmath10 and that the @xmath11 boundaries are reflecting ( neither flow nor field lines cross the @xmath11 boundaries ) .
this was a configuration used initially by @xcite , and we followed it in paper i to enable a direct comparison with his results .
the influences of those boundary choices are discussed fully in paper i. in brief , the periodic boundary limits coalescence of structures to scales equal to the box dimension , @xmath12 .
that is more significant than the existence of the reflecting boundaries , which seem to have only minor influence on dynamics in the ( narrow ) central regions where flow organization and dissipation is largely determined .
the initial background flow has uniform density , @xmath13 , gas pressure , @xmath14 , and an adiabatic index , @xmath15 , so that the sound speed , @xmath16 .
the magnetic field , @xmath17 , is also uniform . in paper
i we considered cases with @xmath18 ( and @xmath19 ) , so that @xmath20 , since those were studied by @xcite .
we now add to these a number of new cases as outlined in table 1 .
to facilitate comparisons we identify the simulations from paper i as cases 1 and 2 in table 1 , with the new simulations following .
the velocity in the background state is antisymmetric about @xmath21 according to the relation @xmath22 with @xmath23 .
this describes a smoothly varying flow within a shear layer of full width @xmath24 . for all our simulations
presented here @xmath25 , chosen to make the interactions with the reflecting boundaries small .
the square computational box has @xmath26 .
flow is to the left in the top half - plane and to the right below that . to this state
we add a perturbation , @xmath27 , defined to be a normal mode found from the linearized mhd equations appropriate for the chosen background , periodic in @xmath10 and evanescent in @xmath11 , with period equal to the length of the computational box , @xmath12 .
this was done exactly as in paper i. we note that when @xmath28 there are perturbations in all three vector components of the magnetic and velocity fields . under the conditions we used the computational frame
is comoving with the kh waves .
all flow and field quantities are either symmetric or antisymmetric around two points , which happen with our choice of phases in @xmath27 to be at @xmath21 , @xmath29 . because the velocity field is antisymmetric around these points they are the places where strong vortices tend to form .
the equations we solve numerically are those of ideal compressible mhd ; namely , @xmath30
@xmath31 @xmath32 @xmath33 along with the constraint @xmath34 imposed to account for the absence of magnetic monopoles ( e.g. , @xcite ) .
the isentropic gas equation of state is @xmath35 .
standard symbols are used for common quantities . here , we have chosen rationalized units for the magnetic field so that the magnetic pressure @xmath36 and the alfvn speed is simply @xmath37 .
these equations were solved using a multidimensional mhd code based on the explicit , finite difference `` total variation diminishing '' or `` tvd '' scheme .
the method is an mhd extension of the second - order finite - difference , upwinded , conservative gasdynamics scheme of @xcite , as described by @xcite . the multidimensional version of the code , along with a description of various one and two - dimensional flow tests is contained in @xcite .
the code contains an _ fft_-based routine that maintains the @xmath34 condition at each time step within machine accuracy .
that step does _ not _ compromise the other conservation relations . numerical solution of equations ( [ masscon ] ) - ( [ induct ] ) on a discrete grid leads , through truncation errors , to diffusion of energy , and momentum , as well as to entropy generation .
of course , such effects are also present in nature and are important to defining the character of the flows .
the existence of effective numerical resistivity is necessary , for example , to allow magnetic reconnection to occur in the calculations .
our methods exactly conserve total energy , as well as mass , momentum and magnetic flux , so the exchange between kinetic , thermal and magnetic energies along with entropy production is internally consistent .
there is fairly good evidence that conservative monotonic schemes , as this one is , do a good job of approximately representing physical viscous and resistive dissipative processes that are expected to take place on scales smaller than the grid ( _ e.g. , _
@xcite ) . for the astrophysical environments being
simulated the expected dissipative scales are likely very much smaller than those that can be modeled directly .
recent numerical studies of reconnection suggest in the mhd limit with large kinetic and magnetic reynolds numbers that the local energy dissipation rate through reconnection becomes independent of the value of the resistivity in complex flows ( @xcite ; @xcite ) .
however , when we depend on numerical dissipation , we must be cautious about the possible role of `` uncaptured '' dynamical structures that could be expected on scales smaller than the grid scale , or about magnetic field structures on such scales that could enhance the number of reconnection sites .
in fact , we shall see that for our kh instability - induced flows involving weak fields ( where reconnection topologies are formed on many scales ) _ total energy dissipation is slightly `` enhanced '' when the grid is finer_. this is opposite to what we expect from the effects of reduced numerical diffusion alone , but consistent with an increase in the number of reconnection sites , or `` x '' points , allowed when smaller - scale field structures can be resolved cleanly .. further , we find that for a fixed numerical resolution , addition of a very weak magnetic field substantially enhances the rate of energy dissipation , again as one expects in response to reconnection ( @xcite ) .
thus , we see strong evidence for unsteady , local reconnection as in high reynolds number , mhd turbulence , but also that our numerical solutions are not quite converged in terms of total dissipation
. below we will express all of our results in time units defined by the growth time of the linear instability , @xmath38 , as estimated from graphs presented by @xcite .
that is , we express time as @xmath39 .
we find that especially the initial saturation of the instability , but also the relaxation processes are fairly uniformly expressed in these units .
the time units , @xmath40 , are listed for each case in table 1 , along with the duration of the simulation in units of @xmath41 .
the flows examined in the present paper all have @xmath42 . for comparison
the sound crossing time in the box is @xmath43 ( since @xmath44 ) and the alfvn wave crossing time is @xmath45 ( since @xmath23 ) .
the normalized turnover time for a large eddy is roughly @xmath46 . since it enters prominently into our later discussions , it is helpful here to remind readers in a simple way of what we can expect for the local evolution of the magnetic field in these simulations .
a full analysis of all the subtleties is beyond our scope here , so readers are referred to detailed discussions such as those by @xcite , @xcite , and @xcite .
equation ( [ induct ] ) follows fields when the resistivity is exactly zero .
as already mentioned , our finite difference code will introduce effects that mimic a finite resistivity , although it is not possible to define an exact value for the resistivity , @xmath47 .
the effective resistivity will also depend on grid resolution , decreasing roughly as @xmath48 within smooth flows . despite these limitations we can make good heuristic use of the resistive mhd extension of the induction equation ( [ induct ] ) . for our purposes
it is interesting to cast that equation in the following form : @xmath49 - \eta j^2 + \eta { \bmsy\nabla}\cdot({\bmit j}\times { \bmit b})}{b^2 } , \label{dbdt}\ ] ] where @xmath50 is the lagrangian time derivative , the first term on the right represents field amplification by stretching , and the last two terms containing the resistivity account for magnetic `` annihilation '' and `` diffusion '' .
we have also used equation [ masscon ] .
the current density , @xmath51 .
the term @xmath52 is a dissipative term that balances the joule heating in the analogous energy equation for the gas ( paper i ) .
the last term is written as an expression involving the lorentz force , @xmath53 , to show that it represents the transport of momentum flux in response to resistivity ; the slippage of field lines .
a frozen - in field results when @xmath54 .
that leaves only the time derivatives and one term on the right of equation [ dbdt ] . if there is flow compression or expansion only perpendicular to @xmath55 then the right side of equation [ dbdt ] vanishes and leads to @xmath56 = constant or @xmath57
those are the most common statements of field compression .
however , more generally one needs to include the other ideal mhd term on the right that accounts for `` stretching '' .
in fact , field enhancements due to compression are much more limited than those due to stretching , especially in mildly compressive flows , such as those we are studying .
we also can see from equation ( [ dbdt ] ) that resistive influences on magnetic energy are associated with both joule heating and momentum transport .
in fact reconnection leads to both irreversible heating of the local plasma and to its acceleration .
the physics of reconnection is complex and beyond the scope of this paper .
however , it may be helpful to estimate the dissipation rate using equation [ dbdt ] and the sweet - parker description of reconnection ( @xcite ; @xcite ; @xcite ) . assuming an incompressible flow steadily carrying oppositely directed fields into a current sheet of thickness
, @xmath58 , equation [ dbdt ] gives us a magnetic energy annihilation rate and associated dissipation rate per unit volume , @xmath59 , where @xmath60 is the inflow speed . in this picture plasma flows out from the reconnection region at the alfvn speed , so mass conservation leads to the relation between the current sheet thickness , @xmath58 , and its width , @xmath61 .
then , @xmath62 .
so , the integrated dissipation rate through reconnection depends on the field energy advected into reconnection sites and the summed volumes of all the reconnection sites .
the current density within the current sheet can be estimated as @xmath63 , so that the aspect ratio of the reconnection region is @xmath64 , where @xmath65 is known as the lundquist number of the plasma and is obviously related to the magnetic reynolds number .
consequently the dissipative volume scales as @xmath66 . in 2d ,
dissipative reconnection regions form out of tearing mode instabilities within a current sheet when the aspect ratio , @xmath67 , is small ( @xcite ) , and , hence when the magnetic reynolds number is large . thus
, reconnection is not really steady , and the number of sites and their individual volumes will depend on the magnetic reynolds numbers in the flow .
those are , indeed , the behaviors we see in our simulations .
we already alluded in the introduction to the basic character of nonlinear mhd kh instability properties found from previous work .
our intent here is to explore more fully the behaviors of weak magnetic fields in this situation . in preparation for
that we note from past work several key features for 2d symmetry : \1 ) when there is no magnetic field or if the field is orthogonal to the flow direction a shear layer will `` role up '' .
for the periodic flows considered here the result is a stable `` cat s eye '' vortex whose length equals the imposed periodicity on the space and whose height @xmath68 the length .
as long as the flow is subsonic or `` submagnetosonic '' no shocks are involved and the vortex decays only through viscous diffusion .
we will not deal with the supersonic or supermagnetosonic cases here ( for some work on those see , @xcite ; @xcite and references therein ) . in many astrophysical applications
the kinetic reynolds number of the flow is very large , so we wish to consider similar cases , so that dissipative decay times are long . for our simulations ,
the empirical viscous decay time of the flow is at least four orders of magnitude longer than the duration of our computations ( see case 5 energy evolution curves in figure 3 ) .
thus , the cat s eye represents a `` quasi - steady relaxed state '' , in which the shear layer has spread vertically by horizontal localization of vorticity and become stable .
of course , for non - periodic systems there will be continued spreading due to additional vortex mergers , while in 3d the flows will be unstable to perturbations directed along the third direction . those influences are beyond the scope of our present investigation , however .
\2 ) in the other extreme , if the magnetic tension force produced by a perturbation in the shear layer exceeds the `` lift '' force produced by the perturbation , the perturbed flow is stabilized . for wavevectors aligned with the flow that condition exists in a linear perturbation whenever the field is strong enough that
. then only viscous diffusion contributes to spreading of the original shear layer .
as before , we can neglect that influence over finite times and describe the shear layer as remaining in a quasi - steady relaxed state from the start . for the present computations the critical magnetic field to stabilize a linear perturbation is @xmath70 . for fields slightly weaker than critical , a small , but
finite amplitude perturbation may lead to the same stabilization condition , perhaps after a small amount of quasi - linear growth to the instability .
such was the result obtained for the so - called `` strong - field '' case of paper i , where @xmath71 .
that calculation is listed as case 1 in table 1 , here . under these conditions the flow is never far from laminar and
there is a modest amount of spreading in the shear layer before it also reaches a quasi - steady relaxed state consisting of a broadened , laminar shear layer ( a point made both in paper i and by mbr ) . in either case 1 or case 2 , the total magnetic energy changed very little during the flow evolution . in paper
i we pointed out that for the symmetry imposed the mean vector magnetic field is time invariant ; @xmath72 constant , so any relaxed state with a relatively uniform field will automatically contain magnetic energy close to that of the initial conditions .
this condition just reflects the conservation of magnetic flux on the grid .
the `` relaxed '' magnetic energy is very slightly enhanced , because the field is not quite uniform at the end . a small amount of kinetic energy dissipation takes place in case 1 before the flow becomes relaxed .
that amount is mandated by total energy conservation , the approximate magnetic energy conservation in this case , and the fact that on a fixed space the kinetic energy in a broad , symmetric shear layer is less than in a thin one .
so , any evolution leading to a broadened , laminar shear layer requires an increased thermal energy determined by the final width of the layer , independent of how it got there . in our idealization
the mass in the box is also exactly conserved , so the mean density is constant . consequently , to first order there is no change in thermal energy by way of reversible , adiabatic processes .
most of the increased thermal energy must result from entropy - producing , dissipation of some kind .
putting it simply , for these flows to relax entropy must be generated .
these energy considerations apply to all of the calculations in our study if the relaxed state is a laminar flow .
( it turns out that the kinetic energy of the cat s eye vortex is less than that for the initial flow , as well .
so , its formation must also generate entropy . )
\3 ) for initial fields too weak to prevent formation of the cat s eye vortex through magnetic tension , the shear layer will role up as for unmagnetized fluid flow . in paper
i , however , we saw that when the initial field is only a few times weaker than the critical value for linear stabilization , there follows a dramatic transformation in the flow as the cat s eye develops .
paper i considered a flow with @xmath73 .
we list it in the current table 1 as case 2 .
as the vortex roles up in such cases , magnetic field lines are stretched around it , thus increasing magnetic energy at the expense of kinetic energy of the flow .
the greatest magnetic pressures are produced in thin flux tubes formed between the vortex and its twins in periodic extensions of the space ( see figures 4a & 6 ) . especially within the vortex and around its perimeter , this evolution leads to magnetic reversals , beginning after about one turn - over time for the vortex .
the reversed fields are unstable to tearing mode reconnection .
so , once that happens the magnetic field quickly reorganizes itself , releasing the magnetic stresses and stored magnetic energy . for case 2
this leads , as well , to disruption of the cat s eye and eventually to an almost steady , laminar flow , after formation and disruption of several weaker vortices .
that , final , quasi - steady relaxed state , was similar to the initial conditions , except that the shear layer was broad enough to be stable against perturbations on scales that fit within the periodic box .
the relaxed shear layer had a linear velocity profile ( see mbr for similar points ) .
there are additional interesting characteristics of the relaxed state .
remaining fluctuations in the magnetic and velocity fields were almost exactly correlated , so that they could be described as linearly polarized alfvn waves .
the magnetic energy returned to a level slightly above the initial conditions , with the final excess representing a pair of apparent `` magnetic flux tubes '' bounding a hot central core of the shear layer containing most of the entropy generated during the relaxation process .
this final condition was reached by about @xmath74 .
the relaxed shear layer was broader in this case than case 1 ( see also mbr ) , so that the kinetic energy was also smaller .
thus , even accounting for the slight increase in magnetic energy , we could correctly predict that _ case 2 with a weaker field was necessarily more dissipative than case 1_. that extra dissipation could come about only through the effects of reconnection .
other aspects of this case will be visited presently , since we will encounter them again in some new cases examined here .
thus , from the calculations reported in paper i and in mbr it is obvious that weak magnetic fields can play a major role in the evolution of mhd kh unstable flows .
it is also apparent that these roles involve the exchange of energy and momentum from the gas to the magnetic field and then back to the gas through maxwell stresses and reconnection .
but , it is not yet clear what are the crucial steps in that exchange , nor how it depends on the initial strength of the field , so that the ordinary kh behavior results , if it does , in the limit that the magnetic field becomes vanishingly small .
in addition , although the linear mhd kh instability is not affected by the presence of a ( possibly strong ) field transverse to the flow ( but still aligned to the plane of the shear ) , a field oblique to the flow can carry `` circularly '' polarized alfvn waves .
since we found that the 2d version of the problem generated linearly polarized alfvn waves , it may be important to see if the nonlinear problem depends at all on the orientation of the field , or just on the strength of the field projected onto the plane as in the linear problem .
to meet the objectives at the end of the previous section we have carried out a set of seven new simulations .
they are outlined in table 1 as cases 3 - 9 .
( once again , cases 1 and 2 were discussed in paper i. ) the new simulations were designed to cover a wide range of strengths for the magnetic field in the computational plane .
they include flows that have exactly the same total field strength as case 2 , but in which the initial field is oblique to the computational plane ; @xmath75 ( cases 3 , 5 and 6 ) , as well as flows in which the field is entirely in the computational plane , but is weaker than that in case 2 ( namely cases 4 , 7 , 8 and 9 ) . the plane - projected field strength for cases 6 and 7
is an order of magnitude weaker than for case 2 .
note that cases 3 and 4 are paired to have the same initial planar field strengths , as are cases 6 and 7 .
this enables us to compare efficiently any distinct roles of field strength and orientation with respect to the plane . in case 5
the field is orthogonal to the plane , so we expect ( and see ) no important role for the magnetic field .
as others have noted before , compressible influences are controlled in that case by magnetosonic waves rather than pure sound waves , so there is a very slight modification in response to that ( @xmath76 in case 5 ) .
we will not concern ourselves at all with flows in which the field is strong , by which we mean situations where magnetic tension precludes the nonlinear development of the mhd kh instability . figures
2 and 3 provide a broad overview of the evolution of the new models we computed .
they illustrate the time variation of energy components ( thermal , kinetic and magnetic ) , as well as the pressure minimum ratio , @xmath77 , on the grid at each time .
figure 2 displays results for the runs that were computed with @xmath78 while figure 3 shows the cases computed with @xmath79 . in addition , to provide a sense of the influence of numerical resolution for very weak field cases both case 6h ( @xmath78 ) and case 6 m ( @xmath79 ) are shown together in figure 2 .
resolution issues were discussed in detail for cases 1 and 2 in paper i , where similar energy plots were also given .
we shall add a few additional relevant comments below .
there is one distinctive detail about figures 2 and 3 that is important to their interpretation .
it was apparent to us by comparing animations of the important dynamical quantities that cases 3 and 4 were virtually indistinguishable from one another , and likewise for cases 6 m and 7 .
thus the important issues are somewhat easier to see if we eliminate @xmath80 to first order in figures 2 and 3 .
we plot a reduced energy for the cases with @xmath28 .
the magnetic and kinetic energy plotted include only the planar components ; @xmath81 , where @xmath82 , and @xmath83 . to compensate ,
the total energy is also reduced as @xmath84 , where @xmath85 is the thermal energy .
thus , we ignore the almost constant energy contributions from @xmath80 and @xmath86 .
note , however , it is the total energy , not the reduced total energy , that is conserved . in the definition of @xmath87 , total magnetic pressure from all three components
was used , since it is the entire pressure that exerts a force on the plasma .
we can see that the reduced energy evolution in cases 3 and 4 are almost identical in figure 2 , and the same is true of cases 6 m and 7 in figure 3 .
thus , we observe in @xmath0d , at least , that the transverse magnetic field component ( @xmath80 ) plays no significant role in nonlinear evolution of the instability .
the rationale for some role comes from the observation that finite @xmath80 and @xmath86 enable circularly polarized alfvn waves in @xmath0d , while only linearly polarized alfvn waves are allowed in 2d . in other words , there are twice as many degrees of freedom in @xmath0d for alfvn waves to help disperse perturbations . however , there are two arguments to support our observation of a minimal role . first , the group velocity vector for alfvn waves along which physical information propagates is @xmath88 ; aligned with the magnetic field ( see , @xcite ) . with invariance of quantities along the @xmath89-direction ,
only the group velocity components projected onto the @xmath90 plane are relevant , and they are independent of @xmath80 .
the second point is that even though @xmath80 and @xmath86 are finite in @xmath0d , the above symmetry restricts their contributions to those of magnetic pressure ( see equations [ [ masscon ] ] - [ [ induct ] ] ) ; to the total pressure gradient and the fast and slow wave speeds . in the present case with weak fields , even that has little importance .
for example , the evolution of @xmath87 is almost the same for cases 3 and 4 , where stretching of the field in the plane is dominant in case 3 . for cases 6 m and 7 the values of @xmath87 are always large , because neither the planar nor the transverse fields are very strong .
but at @xmath91 when the stretching of the planar field is greatest , the values of @xmath87 are still similar .
so , our results indicate that for weak fields the extra degrees of freedom allowed in @xmath0d have no significant effect on the kh instability .
these flows are essentially 2d . in one of the most striking findings from these simulations , we find two distinctive evolutionary behaviors for weak field flows , depending on the field strength . when the planar magnetic fields are very weak , it turns out that the evolution of the 2d mhd kh instability is qualitatively similar to the gasdynamic version of the instability .
that is , the cat s eye vortex continues to exist as long as we extend the simulation .
the magnetic field does , through unsteady reconnection , substantially enhance energy dissipation of the vortex over that from the gasdynamical case , however .
this can be seen by comparing case 5 ( @xmath92 ) in figure 3 with case 6 m or case 7 ( @xmath93 ) .
we will characterize such flows as having `` very weak fields '' , and the role of the field as `` dissipative '' .
on the other hand , if the initial fields approach the critical field for linear stabilization within a factor of a few , the cat s eye is disrupted and the field causes the flow to reorganize into a laminar form ( as in case 2 ) .
this we call simply the `` weak - field '' regime , and the field as `` disruptive '' . in either situation
the magnetic energy peaks about the time of the initial reconnection instability . for the disruptive cases
there is a fairly prompt return of the magnetic energy near to its initial value ( see figures 2 and 3 ) . in effect , the magnetic field plays the role of a catalyst , storing energy temporarily and through it modifying the plasma flow .
that role is fairly dramatic , since reconnection leads to dynamical alignment and self - organization in the flow . for dissipative cases ,
the magnetic energy declines much more slowly , and in the high resolution case 6h , seems mostly to oscillate .
that behavior results from the fact that the cat s eye vortex continues to capture magnetic flux , amplifying it and thence dissipating it .
the existence of two qualitatively distinct evolutionary patterns is also quite apparent in the histories of the thermal and kinetic energies displayed in figures 2 & 3 .
the kinetic energy decay in case 6 is almost exponential , with a time constant that can be roughly estimated as @xmath94 . by contrast cases 3 and 4
show before @xmath95 a sharp drop in @xmath96 along with an accompanying increase in @xmath97 , followed by a slow , possibly exponential evolution that is similar to case 6 .
for all cases in table 1 the magnetic energy is always small ( although it can briefly increase by factors between 6 and 20 it never contributes more than a few percents to the total energy ) , so on the face of it the magnetic energy would not seem to be important .
in fact , it can be crucial , as we have already outlined and will discuss more fully , below .
figure 3 , which shows a wider range of projected field strengths , shows similar patterns with some variation in the abruptness of the early energy transition .
we also conclude that as @xmath98 the evolution of the 2d mhd kh instability will smoothly approach that for ordinary gasdynamics .
let us now describe in more detail the characteristics of the two weak - field regimes and also establish the physical boundary between them .
case 6 and case 7 had the weakest finite planar field , @xmath99 , that we considered , and they demonstrate very - weak - field patterns . figure 4 summarizes flow properties for case 6h at times @xmath100 and 30 .
figure 4a shows the flow vorticity component out of the plane , @xmath101 , while figure 4b shows the magnetic field lines as they project onto the plane .
except for minor fine structure that aligns with the planar field , the vorticity at @xmath100 is the same as for a flow with @xmath92 of case 5 .
it also is fairly close to that for all our simulations , except case 1 at this early time .
primarily this image illustrates the formation of the cat s eye and how that concentrates vorticity .
note that the `` corners '' of adjacent cat s eyes overlap , with a shear layer between them .
it is in that shear layer where magnetic fields are most strongly enhanced , as seen for this time in figure 4b .
extensions of that feature around the perimeter also contain concentrations of magnetic flux .
these regions represent flows where gas is accelerated out of a stagnation point midway between vortex centers .
frozen - in fields are thus pulled out or `` stretched '' and amplified . at @xmath102 , the cat s eye vortex is still largely the same , except for some complex , low level vorticity structures outside the main vortex .
their origins are made clearer by examination of the magnetic field structure at this time .
those same regions outside the cat s eye also contain isolated magnetic flux islands and field reversal regions . such magnetic features reveal an environment where magnetic reconnection is active .
the relationship is that reconnective processes not only reorganize the magnetic field topology and release magnetic energy , but they also accelerate the local plasma , and that contributes to the local vorticity .
by contrast to case 2 ( or as we shall see cases 3 and 4 , as well ) , however , magnetic stresses in cases 6 or 7 produce only small modifications to peripheral flows , and are far too weak to disrupt the cat s eye .
the vortex interior shows initial signs in figure 4b of magnetic flux expulsion by @xmath100 , a well - known phenomenon ( @xcite ) . at this early stage ,
just as the vortex is fully formed , there is still some magnetic flux that threads through the `` eye '' .
that does not seem to be the case at the later time . in a close examination of field structures within the vortex
we can find no evidence after about @xmath103 that any magnetic flux threads the vortex .
instead , the field breaks into flux islands within the vortex , and those are annihilated through mergers .
this is , of course , just what high reynolds number ( both kinetic and magnetic ) , nonlinear resistive mhd flows are expected to do in response to reconnection ( @xcite ; @xcite ) .
the field structure in the bottom panel of figure 4b is qualitatively very similar to that found by weiss ( @xcite ) from a classic passive field simulation in a steady vortex with magnetic reynolds number @xmath104 , but very different from that for small magnetic reynolds number ( @xmath105 ) .
our own estimates of effective magnetic reynolds numbers in the case 6h simulation give numbers in excess of @xmath106 ( @xcite ; @xcite ) , so the comparison is very reasonable .
we emphasize , however , that _ the field behaviors seen here , and presumably by weiss , result from localized , inherently time dependent reconnection , not simple flux diffusion_. to demonstrate that we compute in the appendix the equilibrium passive magnetic fields for a simple vortex in a resistive fluid .
figure 5 displays two examples of magnetic field structures predicted by this steady - state resistive mhd theory .
the velocity structure for this model vortex is similar to that observed in the simulated flows for case 6 .
the core , @xmath107 , has a constant vorticity , @xmath108 , ( see figure 4a ) while there is an outer flange in which the velocity decreases to zero at @xmath109 .
solutions depend only on the effective magnetic reynolds number , @xmath110 , within the vortex .
this parameter roughly measures the ratio of the timescale for magnetic field diffusion to the rotation period of the vortex .
when @xmath111 is a few ( @xmath112 in the top panel of the figure ) , flux diffuses quickly enough to almost flatten out field lines to their external pattern except in the vortex core . at the center field lines are actually concentrated by rotation into a quasi - dipole pattern
that dipole results because reconnection into a simpler topology is not permitted .
the steady - state pattern is very different when @xmath111 is large , since field is almost frozen into the vortex .
above @xmath113 the field forms a spiral pattern in the vortex core whose pitch depends on @xmath111 .
shear is greatest in the flange region of this vortex and flux becomes concentrated increasingly into the vortex perimeter .
that particular behavior does resemble case 6 in some ways .
there are multiple field reversals in the perimeter , with that number determined by the balance between field diffusion and addition of new turns to the field through rotation .
the bottom panel in figure 5 shows the steady - state solution for @xmath114 . despite some superficial similarities
, there is an obvious and essential difference between the truly steady - state solution in figure 5 and the `` quasi - steady '' field in the two time dependent simulations ( ours and weiss ) . for the steady solutions the field line that was originally through the vortex center remains and is wrapped into a spiral pattern within the core giving an almost constant field strength there . in
the time dependent cases , reconnection changes the field structure inside the vortex .
the same statement applies to weiss simulation .
if the effective magnetic reynolds number within the flow is large enough for topologies to develop that are tearing - mode unstable , field within the vortex `` breaks '' off from the external field and is destroyed .
thus , we find good agreement with weiss in the nature of eddy flux expulsion and see that it can not be viewed as a steady magnetic diffusion .
the value of @xmath115 never drops to zero inside the vortex of any of our very weak field cases , although it is as much as two orders of magnitude smaller than along the vortex perimeter .
action around the perimeter episodically injects new magnetic flux into the vortex interior , so this process continues as long as we have followed the flow .
examination of the magnetic energy , @xmath116 , for case 6 in figure 2 shows related , episodic peaks on rough intervals , @xmath117 .
that corresponds to about half a turn over - time for the vortex , and represents the interval on which fields along the vortex perimeter are stretched until they become `` folded '' , so as to make them subject to tearing mode reconnection .
so , whatever flux is caught in this flow becomes amplified , increasing the magnetic energy , before reconnection rearranges the field lines .
those lines outside the vortex are relaxed towards the initial field configuration , often through multiple reconnection events , while those inside are isolated into closed islands .
note that the total magnetic flux threading our box is exactly conserved and exactly zero in these computations .
thus any field line entering on the left boundary must at all times extend continuously to the right boundary or exit again on the left .
closed flux islands can exist , but nothing prevents them from being destroyed , since they contribute no net flux .
reconnection , and most obviously flux - island destruction , is irreversible , so that it must be accompanied by entropy generation .
that outcome is very apparent in figures 2 and 3 . recall that to first order total thermal energy changes in these , closed - system simulations reflect nonadiabatic processes .
compare first the evolution of thermal energy in cases 6 m , 7 and 5 .
we see that after some viscous dissipation necessary to form the cat s eye , the case 5 with @xmath118 has almost constant total thermal energy to the end of the simulation .
by contrast cases 6 m and 7 show a steady rise in @xmath85 , with small amplitude oscillations associated with major reconnection events as described above .
thus , the main dynamical impact of the field is enhanced dissipation . the total magnetic energy remains small in all cases , but for the dissipative cases that reflects a near balance between the rate at which kinetic energy is being transferred to magnetic energy and the rate at which magnetic energy is being dissipated . for example , from initial field energy evolution in figure 3 we can crudely estimate for case 6 m that magnetic energy is generated at a rate @xmath119 , which is very close to the mean slope of the thermal energy curve , @xmath85 .
looking next at the case 6m , h @xmath85 plots in figure 2 , we see that energy dissipation is greater in the higher resolution case 6h .
that results despite a smaller numerical resistivity for the higher resolution simulation .
this is because in case 6h a larger amount of flux is caught in the vortex , so more reconnection sites develop .
we estimate that @xmath120 of the total magnetic flux is attached to the vortex at the end of the calculation @xmath121 in case 6h , while only @xmath122 is attached at the end of the calculation @xmath123 in case 6 m .
as the effective magnetic reynolds number increases there is a tendency for decreased resistivity to be countered by an increase in the number of reconnection sites , as our earlier discussion of reconnection theory would suggest .
if we compare simple estimates for rates of magnetic energy generation for cases 6 m and 6h to the rates of thermal energy increase , we see that they are consistent , as we found earlier for case 6 m , alone .
the increased dissipation for the higher resolution simulation is relatively modest , however .
some studies of resistive mhd turbulence suggest , in fact , that at very large reynolds numbers total reconnection rate in a complex flow will be insensitive to the value of the resistivity ( @xcite ; @xcite ) .
if confirmed more generally , that could provide a practical measure of convergence in studies of the present kind .
looking again at thermal energy evolution in figures 2 and 3 we can see a clear behavioral transformation in the sequence : case 5 @xmath124 6(7)@xmath124 8 @xmath124 9 @xmath124 3(4 ) .
this grouping is arranged in order of increased @xmath125 .
there is a sharp rise through the sequence in the amount of dissipation associated with the initial formation of the cat s eye vortex , as well as a more modest increased slope to the subsequent long - term dissipation rate .
the physical character of this transformation is apparent if one compares figure 4b with figure 6 .
the latter shows for case 3 the magnetic pressure distribution at three times , @xmath126 and @xmath127 .
behaviors for this simulation are qualitatively similar to case 2 , as discussed in detail in paper i and outlined in 2.3 . here
we note that , while the field appears wrapped around the vortex at @xmath100 , it has a laminar appearance at @xmath102 .
the velocity field undergoes a similar transition ; the vortex is completely disrupted .
the magnetic field at the intermediate time , @xmath103 , shows aspects of both the other times .
curiously , however , the flow in the dominant vortex pattern there has the opposite vorticity to the original flow .
that feature is short - lived .
we can understand this flow transition from rotational to laminar by examining figures 7 , 8 and 9 , also relating to case 3 . in figure 7
we show the log of @xmath128 with magnetic field lines overlaid at @xmath100 .
the minimum @xmath129 in the strong flux tube connecting vortices , but the magnetic field is dynamically significant most of the way around the perimeter of the vortex .
the maximum @xmath130 in small regions where magnetic reconnection has begun and the field strength has decreased to very small values . to understand how the magnetic field disrupts
the vortex consider the forces involved .
the centripetal force associated with motion around the cat s eye is @xmath131 , since we observe that @xmath132 and @xmath133 for the vortex .
indeed we also confirm that when @xmath118 , the pressure gradient force within the vortex , @xmath134 and is directed towards the vortex center , so that it supplies the necessary force . on the other hand in figure 8 we display the magnetic field lines along with vectors representing the magnetic tension force ; @xmath135 at @xmath136 for case 3 .
it is apparent that magnetic tension forces are concentrated where the field has been pulled into loops by the flow , and that they are directed towards the center of the vortex .
the peak value of @xmath137 . at @xmath136 in case 3 , however , the total pressure gradient is small and actually of the wrong sign to effect significantly the motion of the plasma .
the total force vector field is very similar to the tension force field shown in figure 8 . at the same time
`` x - points '' in the field topology show that reconnection is underway that will isolate the associated field lines .
subsequently , flux islands are formed along the `` axis '' of the cat s eye , and the magnetic tension pulls the plasma frozen into those loops towards the original center of the vortex , disrupting its rotation .
field line segments reconnecting outside the vortex core will tend to relax towards the original field topology .
these observations allow us to estimate simply what minimum initial field , @xmath125 should lead to vortex disruption .
since the early evolution of a weak field is self - similar , we can use the behavior from case 3 to estimate @xmath138 at the time of the first major reconnection in all weak - field cases .
our earlier discussion requires @xmath139 for disruption , leading to the constraint @xmath140 , or alternately , @xmath141 , where @xmath142 is the critical field for stabilization of original instability .
indeed , as figure 3 demonstrates , the transformation between dissipative and disruptive evolution occurs for conditions between those of cases 8 and 9 ; for initial field values between @xmath143 and @xmath144 . figure 9 illustrates why the vortex disruption process is also highly dissipative .
the top panel shows at @xmath136 in case 3 the gas entropy distribution , while the lower panel displays the electric current density , @xmath145 with the field lines overlaid . from this
we can see that excess entropy is concentrated into regions where reconnection is currently active ( highlighted by @xmath145 ) or recently active .
ohmic heating ( @xmath146 ) is partly responsible for the irreversible energy exchange .
the remainder should be viscous dissipation of small - scale , disordered motions .
as we discussed in paper i , the final laminar flow that results in this class of flow includes a central sheet of hot gas containing most of the excess entropy produced through the self - organization of the flow .
we have carried out a series of high resolution mhd simulations of kelvin - helmholtz unstable flows in @xmath0 dimensions .
all of these simulations involve magnetic fields initially too weak to stabilize the flows in the linear regime ; @xmath147 .
thus , since simulations are performed on a periodic space , flows all begin formation of a single `` cat s eye '' vortex .
if the field lying in the computational plane is absent or `` very weak '' the cat s eye structure becomes a persistent , stable feature that represents a `` quasi - steady '' equilibrium .
when there are `` very weak '' magnetic fields in the plane they become wrapped into the vortex and amplified by stretching .
however , within a single turn of the vortex they are subject to tearing mode instabilities leading to magnetic reconnection . that reconnection isolates some magnetic flux within the vortex , which is eventually annihilated .
this is the process through which flux is effectively expelled from a vortex .
as long as the vortex persists this process will repeat .
since reconnection is irreversible , this process is also dissipative and leads to an increase over viscous effects in conversion from kinetic to thermal energy .
we find in this regime that as the initial magnetic field within the computational plane is increased the dissipation rate increases in a similar manner . likewise ,
as we use a finer numerical grid , thus _ reducing the effective numerical resistivity and viscosity , the dissipation rate increases _ , reflecting the increased ability of our code to capture small - scale reconnection events .
this trend is backwards from what one would expect if simple magnetic diffusion were primarily responsible for the reconnection .
it suggests , perhaps that if we had been able to extend these calculations to even higher resolution the energy dissipation rate might have converged to a value independent of the effective resistivity , just as some studies of resistive mhd turbulence find .
the reconnection and expulsion of flux within vortices in our simulations are similar to those in a classic study by weiss of vortex flux expulsion in large - magnetic - reynolds - number flows .
if the initial magnetic field is strong enough that within a single turn of the vortex it is amplified around the vortex perimeter to `` dynamical '' strength ( @xmath148 ) , then the reconnection described in the previous paragraph releases stresses that are capable of disrupting the vortex entirely .
this can happen in a single event or , if the field is only marginally strong enough ( @xmath149 ) , through a succession of dynamical realignment events . in either case
the net result is a laminar , but marginally stable flow , in which the original shear layer is greatly broadened .
thus , as we discussed fully in paper i , such fields can have a remarkable stabilizing influence .
this is despite the fact that their total energy content is a minor fraction of the total , so that they are nominally too weak to be important , according to the usual criteria .
we considered cases in which the magnetic field was entirely within the flow plane and others in which the field was oblique to that plane , in order to examine the role in nonlinear flows of the component out of the plane . for
the @xmath0 d flows we have studied , only the field components in the flow plane have any dynamical significance . in fully 3 d flows , however , we expect further evolution of the `` quasi - steady relaxed states '' of both very weak field ( or dissipative ) cases and weak field ( disruptive ) cases . the cat s eye vortex of very weak field cases is subject to a 3 d instability known as the elliptical instability ( @xcite ; @xcite ) unless the flow lines around the vortex follow perfect circles .
the planar shear flow of weak field cases is stable against linear perturbations but unstable to 3d finite - amplitude perturbations ( @xcite ) .
thus , it will be important to extend the present study to the fully 3 d regime , and we are preparing to do that .
this work by twj , jbg , and af was supported in part by the nsf through grants ast-9318959 and int-9511654 , through nasa grants nagw-2548 and nag5 - 5055 and by the university of minnesota supercomputer institute .
the work by dr was supported in part by seoam scholarship foundation .
we are grateful to b.i .
jun for stimulating and helpful discussions about these results .
in order to show that the flux expulsion from vortex shown in figure 4b requires localized , unsteady reconnection , not just diffusion , here we study the steady state passive field solutions of resistive mhd . under 2d and @xmath0d symmetries , the planar magnetic field can be written as @xmath150 where @xmath151 is the magnetic flux function .
then , the induction equation in ( [ induct ] ) with constant resistivity becomes @xmath152 we consider the evolution of an initially uniform magnetic field , @xmath153 or @xmath154 , within a vortex .
we approximate that the vortex has azimuthal velocity @xmath155 and zero radial velocity .
@xmath108 equals the ( constant ) vorticity within the vortex core .
this is roughly the velocity field of the vortex in figure 4 .
then , the evolution of the magnetic field is described by equation ( [ magfluxeq ] ) with the boundary conditions @xmath156 and @xmath157 at @xmath158 . in a steady state , we can set @xmath159 since the coefficients of the equation do not depend explicitly on @xmath160 .
note that with the given initial magnetic field and boundary conditions , only the solution with @xmath161 is allowed .
then , the equation for @xmath162 is given as @xmath163 where @xmath164 with @xmath165 .
here , @xmath111 is the magnetic reynolds number within a factor of two .
we solve equation ( [ magffuneq ] ) for @xmath162 numerically . in figure 5
, we plot the resulting magnetic field lines for the case with @xmath166 ( highly diffusive case ) and for the case with @xmath167 ( quasi - adiabatic case ) . from dissipation tests we estimate the magnetic reynolds number of the vortex in our time - dependent simulations to be larger than @xmath106 ( see @xcite ; paper i ) .
indeed the fields in figure 4b resemble those in a vortex simulation with magnetic reynolds number @xmath106 reported in @xcite .
the steady - state solution we just described allows only for diffusion of magnetic flux , since no change in the field topology is permitted .
it allows us to see that such a steady - state , diffusive description does not account for flux expulsion from the vortex .
if that were the case the magnetic field lines in figure 4b should have a structure like those in the bottom of figure 5 .
actually , their topologies are fundamentally different in the sense that field in figure 5 threads completely through the vortex , while it does not in figure 4b ( or any similar figure showing additional field lines for this simulation ) .
the reason is that the field in the time dependent simulation is subjected to reconnective instabilities that isolate and then destroy magnetic flux in the interior of the vortex .
ccccccccc 1 & 0.4 & 0 & 0.4 & 2.5&7.5&3.79 & 20@xmath41 & 512 2 & 0.2 & 0 & 0.2&5.0 & 30&1.86 & 20@xmath41 & 512 3 & 0.2 & 45 & 0.14&5.0 & 30&1.71 & 30@xmath41 & 5124 & 0.14 & 0 & 0.14&7.07 & 60&1.71 & 30@xmath41 & 5125 & 0.2 & 90 & 0.0&5.0 & 30&1.59 & 20@xmath41 & 2566h & 0.2 & 85&0.02&5.0 & 30&1.59 & 50@xmath41 & 512 6 m & 0.2 & 85&0.02&5.0 & 30&1.59 & 30@xmath41 & 256 7 & 0.02 & 0 & 0.02&50 & 3000&1.59 & 30@xmath41 & 2568 & 0.04 & 0 & 0.04&25 & 750&1.67 & 30@xmath41 & 2569 & 0.07 & 0 & 0.07&14.3 & 245&1.67 & 20@xmath41 & 256 fig . 1 : : cartoon illustrating the computational setup for these simulations .
there is a central shear layer separating uniform flows . except for the initial perturbation the flow conditions are otherwise uniform . the magnetic field projects onto the computational plane at an angle , @xmath168 . fig . 2 : : time evolution of the high resolution simulations ( cases 3 , 4 , 6h ) , plus the medium resolution simulation case 6 m
shown are the normalized total thermal , kinetic and magnetic energies , as well as the minimum value of the plasma @xmath169 parameter at each time . to emphasize the minimal importance of the transverse magnetic field component , @xmath80 , `` reduced ''
energies are shown that exclude @xmath80 .
fig . 3 : : same as figure 2 , except for medium resolution simulations ( cases 5 , 6 m 7 , 8 , 9 ) .
4a : : greyscale images of the case 6h vorticity component normal to the computational plane at @xmath100 ( top ) and @xmath102 ( bottom ) . to facilitate visualization of structures
, the periodic space has been repeated once .
the vorticity is everywhere positive .
high values take high tones .
4b : : case 6h magnetic field lines projected onto the computational plane plane at @xmath100 ( top ) and @xmath102 ( bottom ) . fig . 5 : : magnetic field lines from an analytical model to study diffusive flux expulsion from a steady eddy with @xmath166 ( top ) and @xmath170 ( bottom ) .
see the text for the definition of @xmath111 .
fig . 6 : : grayscale snapshots of the magnetic pressure distributions for case 3 at @xmath171 and 30 , showing the stages in disruption of the cat s eye vortex by the magnetic field .
high values take high tones .
fig . 7 : : inverted grayscale image of @xmath172 for case 3 at @xmath100 .
projected magnetic field lines are overlaid .
the minimum @xmath173 in the strong flux tube between vortices , while the maximum @xmath130 near the center of the cat s eye .
fig . 8 : : the magnetic field structure for case 3 at @xmath136 , just as substantial magnetic reconnection is underway .
projected field lines are shown , along with arrows that represent magnetic tension forces .
the maximum magnetic tension force is 1.74 fig . 9 : : inverted grayscale image of the gas entropy ( top ) and electrical current ( bottom ) distributions for the case 3 at @xmath136 .
magnetic field lines are also laid on top of the current distribution to emphasize the relationships .
high values take low tones . | we have carried out high resolution mhd simulations of the nonlinear evolution of kelvin - helmholtz unstable flows in @xmath0 dimensions .
the modeled flows and fields were initially uniform except for a thin shear layer with a hyperbolic tangent velocity profile and a small , normal mode perturbation .
these simulations extend work by @xcite and @xcite .
they consider periodic sections of flows containing magnetic fields parallel to the shear layer , but projecting over a full range of angles with respect to the flow vectors .
they are intended as preparation for fully @xmath1 calculations and to address two specific questions raised in earlier work : 1 ) what role , if any , does the orientation of the field play in nonlinear evolution of the mhd kelvin - helmholtz instability in @xmath0d .
2 ) given that the field is too weak to stabilize against a linear perturbation of the flow , how does the nonlinear evolution of the instability depend on strength of the field .
the magnetic field component in the third direction contributes only through minor pressure contributions , so the flows are essentially @xmath2d . in @xcite
we found that fields too weak to stabilize a linear perturbation may still be able to alter fundamentally the flow so that it evolves from the classical `` cat s eye '' vortex expected in gasdynamics into a marginally stable , broad laminar shear layer . in that process
the magnetic field plays the role of a catalyst , briefly storing energy and then returning it to the plasma during reconnection events that lead to dynamical alignment between magnetic field and flow vectors . in our new work
we identify another transformation in the flow evolution for fields below a critical strength . that we found to be @xmath3% of the critical field needed for linear stabilization in the cases we studied . in this `` very weak field ''
regime , the role of the magnetic field is to enhance the rate of energy dissipation within and around the cat s eye vortex , not to disrupt it .
the presence of even a very weak field can add substantially to the rate at which flow kinetic energy is dissipated . in all of the cases we studied magnetic field amplification by stretching in the vortex
is limited by tearing mode , `` fast '' reconnection events that isolate and then destroy magnetic flux islands within the vortex and relax the fields outside the vortex .
if the magnetic tension developed prior to reconnection is comparable to reynolds stresses in the flow , that flow is reorganized during reconnection . otherwise , the primary influence on the plasma is generation of entropy .
the effective expulsion of flux from the vortex is very similar to that shown by @xcite for passive fields in idealized vortices with large magnetic reynolds numbers .
we demonstrated that this expulsion can not be interpreted as a direct consequence of steady , resistive diffusion , but must be seen as a consequence of unsteady fast reconnection .
= cmmi10 scaled1 = cmmi10 = cmmi7 = cmmib10 scaled1 = cmmib10 = cmmib10 scaled 833 = cmbsy10 scaled1 = cmbsy10 = cmbsy10 scaled 833 1 = 1 = 1 = 9 = 9 = 9 = 10 = 10 = 10 = |
alfvn waves have been proposed as an energy delivery mechanism for solar atmospheric processes ranging from heating of the chromosphere @xcite or corona @xcite to the acceleration of the solar wind @xcite . both observations @xcite and numerical models ( e.g. * ? ? ?
* ; * ? ? ?
* ) show that there is enough alfvn wave energy generated in the convection zone to heat the chromosphere , but the mechanisms by which this wave energy is thermalized in the chromosphere , if indeed it is , remain unclear . if alfvn waves are a dominant source of quiet chromospheric heating the observed input flux of @xmath0 must be converted into a @xmath1 local heating rate to balance the energy loss in the chromosphere .
estimates place the maximum heating rate required as @xmath1 dropping to @xmath2 in the upper chromosphere @xcite .
the possibility that alfvn waves may heat the solar atmosphere and drive the solar wind has been studied numerically with many variations . @xcite
extensively studied the 1.5d problem for circularly polarized alfvn waves in expanding flux tubes and showed the importance of coupling to slow modes , fast modes and shocks .
these papers demonstrated that it was certainly possible for alfvn waves to heat the solar atmosphere and drive the solar wind .
later work has extended this to white noise spectrum photospheric drivers in both 1.5d @xcite and 2.5d @xcite .
the main focus of these papers has been on coronal heating and driving the solar wind .
thus , while they contained model atmospheres with chromospheres and transition regions , these were not modelled with very high resolution , nor were detailed heating profiles produced specifically for the chromosphere . despite this
, all papers listed above did show that a heating rate broadly consistent with the requirements of both the chromosphere and corona was possible for photospheric motions with speeds of order @xmath3 km s@xmath4 driving alfvnic perturbations .
@xcite extended their original analysis @xcite and improved the diagnostics for shock heating and concluded that shock heating was the dominant heating mechanism in the chromosphere .
these simulations did not explicitly include resistivity or assess the sensitivity of these results to the chosen driver spectrum . nonetheless they present a compelling case for shock heating of the chromosphere originating from a broad spectrum photospheric , alfvnic driver .
the results presented below represent an extension of this line of work , albeit only in one spatial dimension , to assess the importance of explicit resistive dissipation through electron - ion , electron - neutral and ion - neutral collisions .
also studied is the sensitivity of the results to the chosen driver spectrum and a high - resolution focus on just the chromosphere .
turbulence has long been proposed as a possible mechanism by which the input driver wave energy could be dissipated @xcite .
direct dissipation of driver alfvn waves through resistive effects is too slow to explain the required heating rate .
however , if a turbulent cascade exists which can transfer the driver energy to higher wavenumbers where dissipation does occur there is sufficient wave energy in the chromosphere to balance radiative losses .
this theory does however leave unanswered the question of whether a turbulent cascade to dissipation scales does actually exist in the chromosphere . reduced mhd models have recently concluded that alfvn waves may drive a turbulent cascade , through reflection off the transition region , which could heat the chromosphere @xcite .
the implementation of reduced mhd in @xcite ignores compressive effects so does not permit coupling of alfvn waves to slow modes or the formation of shocks .
studies of the resistive dissipation of alfvn waves by collisions between ions and neutral atoms , begun with @xcite , have recently begun to be re - investigated and extended @xcite .
the temperature variation through the chromosphere means that it contains both regions where the hydrogen is near fully ionized and regions where only heavy ions with a low first ionization potential are ionized @xcite . under low - ionization conditions the ion and neutral gas fluids partly decouple , and energy dissipation due to ion - neutral collisions is orders of magnitude more important than the resistivity resulting from electron - ion coulomb collisions @xcite .
previous work in this area has shown efficient heating of both the upper chromosphere @xcite and the lower chromosphere @xcite due to ion - neutral collisions .
numerical simulations in this area are complicated by the simultaneous requirements to resolve short driver wavelengths and a high alfvn speed in the upper chromosphere .
@xcite also include the hall term in chromospheric heating models further reducing the timestep in the simulations . the computational effort that is required to simultaneously resolve
all of these length and time scales means that 1.5d modeling is often used to model alfvn wave propagation and heating in the chromosphere @xcite .
these have shown that alfvn waves with periods much less than 10 seconds are effectively damped by pedersen resistivity @xcite in the chromosphere , and for incompressible simulations with a kolmogorov driving spectrum the heating is primarily in the lower chromosphere and sufficient to balance radiative losses there @xcite .
there is therefore compelling observational and theoretical evidence that a significant fraction of the heating requirements of the upper chromosphere , and possibly the quiet corona , could be accounted for through the dissipation of alfvn wave energy , or mhd waves in general in geometries which make pure alfvn waves unlikely .
this paper extends the earlier works by concentrating specifically on the chromosphere and varying the spectrum of driven alfvn waves , solving the full compressible mhd equations and including partially ionized plasmas .
section [ sec : numtech ] covers the assumptions made , the numerical method and driver spectra used in the simulations .
since much previous work on turbulent heating @xcite assumed an incompressible plasma incompressible results are presented in section [ sec : incompressible ] .
full compressible simulations results are presented in section [ sec : compressible ] and conclusions which can be drawn from these 1.5d simulations are in section [ sec : conclusions ] .
this paper presents the results of numerical simulations performed using the hall - mhd code lare @xcite .
this code solves the ideal mhd equations explicitly using the lagrangian remap approach and includes the resistive and hall terms using explicit subcycling .
the complete set of mhd equations relevant in this case are @xmath5 where @xmath6 is the mass density , @xmath7 is the ideal gas pressure , @xmath8 is the magnetic field , @xmath9 is the fluid velocity , @xmath10 is the electric field , @xmath11 is the specific internal energy density , @xmath12 is the electron number density , @xmath13 is the proton charge and @xmath14 are the current densities parallel / perpendicular to the local magnetic field .
these equations are in written in lagrangian form where @xmath15 is the usual advective derivative .
shock viscosity is added to the momentum equation through @xmath16 which has a functional form which vanishes for smooth flows in the limit of increasing resolution but remains finite at discontinuities .
thus the shock jump conditions are satisfied through an appropriate shock viscosity as in @xcite . in this formulation shock heating
can therefore be monitored though viscous heating . hence viscous
heating and shock heating are used as synonyms throughout the rest of this paper .
missing from equation ( [ eq : energy ] ) are thermal conduction , radiative losses and all non - local radiation transport . to prevent the atmosphere heating up without bounds due to the viscous dissipation of shocks ( @xmath17 ) and ohmic heating ( @xmath18 ) , both @xmath17 and @xmath18
are calculated in simulations as a diagnostic but only @xmath17 is added to the energy update . if @xmath19 , as in incompressible simulations , ignoring @xmath18 assumes that the background atmosphere is for a slowly varying chromosphere where heating terms are approximately balanced by losses .
shock heating must be included in compressible simulations otherwise the shocks themselves are not treated correctly .
this leads to a heating of the chromosphere . despite this
the ionization state is held fixed at its initial equilibrium value .
since shock heating , which is only weakly sensitive to the ionization state , dominates in these simulations this will not change the heating rates by more than a factor @xmath20 .
the resistive mhd ohm s law for a fully ionized plasma is @xmath21 @xcite where @xmath22 is the classical electron - ion resistivity resulting from coulomb scattering and @xmath23 is the electron - ion collision frequency for ion number density @xmath24 and coulomb logarithm @xmath25 . @xmath26 and the resistivity is anisotropic . normally within solar and space physics , where the lundquist number is high
, this is simplified to @xmath27 with an isotropic resistivity .
this approximations is adopted in this paper so ohmic dissipation resulting from parallel currents is likely to be over estimated by a factor @xmath28 . in the 1.5d simulations in this paper
there can be no current along the background magnetic field so the leading order ohmic dissipation from electron - ion collisions is for perpendicular currents . in equation ( [ eq : ohm ] ) ,
@xmath29 is the classical resistivity resulting from electron collisions with ions and neutrals and @xmath30 the pedersen resistivity @xcite .
explicitly the resistivities used in this paper are : @xmath31 where @xmath32 is the electron collision frequency including collisions with ion and neutrals .
the pedersen resistivity is then given by @xmath33 where @xmath34 is the neutral fraction with @xmath35 the number density of neutrals .
the ion - neutral collision frequency is given by @xmath36 here @xmath37 is the cross - section for ion - neutral collisions for hydrogen from @xcite . for an effective ion mass of 1.2 proton masses @xmath38 .
the pressure gradient term @xmath39 is ignored in equation [ eq : ohm ] as this can only generate a parallel electric field in 1.5d simulations which have no effect on the dynamic equations . in the above descriptions of resistivities
we have adopted the nomenclature used in @xcite .
following @xcite we also use cowling resistivity ( @xmath40 ) , defined as @xmath41 , to refer to that component of perpendicular resistivity which arises purely due to ion - neutral collisions .
it is important to note that the pedersen resistivity is only the inverse of the pedersen conductivity , more commonly used in ionospheric physics , if the hall term is zero .
the inclusion of neutrals in a single - fluid model , as described above , is valid in the same limit as resistive mhd , i.e. timescales larger than the gyro - period , collision times and the plasma period and length - scales larger than the ion gyroradius and collisional mean - free - path .
the only additional constraint , compared to single fluid mhd , is therefore that equations ( 1 - 6 ) are only valid for times larger than the ion - neutral collision time . in the chromosphere
this can be as low as a millisecond so the model is valid up to frequencies of @xmath42 hz . in all simulations
the @xmath43 axis is ignorable , the @xmath44 axis extends vertically up from the solar surface starting at the @xmath45 optically thick surface , and the @xmath46 axis is also ignorable , while velocities in the @xmath46 direction are allowed and associated with alfvnic perturbations .
the upper boundary condition is kept open by the use of both riemann characteristic open boundaries and a damping region where the simulation velocity is smoothly reduced as it approaches the boundary .
the measured reflection of the upper boundary is less than 0.01% . in some sections
lare is converted into a 1.5d incompressible code by setting the forces parallel to the @xmath44 axis to zero .
this sets @xmath47 , which in 1.5d is equivalent to the incompressible condition @xmath48 .
for the simulations presented in this paper , a model atmosphere based on the semi - empirical c7 model of avrett & loeser @xcite is used for the initial temperature profile .
this is then integrated iteratively to provide a density profile which is in hydrostatic equilibrium . on each iteration
the ionization state of the plasma is calculated under the assumption of local thermal equilibrium using a two level athay potential model
@xcite until a converged density is obtained .
the background magnetic field @xmath49 is assumed to be a uniform vertical 50 g .
this is chosen as a representative value averaged along an inter - granular field line .
figure [ fig : eta ] plots neutral density and pedersen diffusivity for the avrett and loeser c7 atmosphere model in this paper .
for comparison , figure [ fig : eta ] also shows the same plots for the two other atmospheric models : the val iii c model of @xcite and the fal a model of @xcite .
it is found that the density of neutral hydrogen predicted by the athay potential model and the associated pedersen resistivity are similar in all cases although the fal a model does predict 5 times greater pedersen resistivity in the upper chromosphere .
figure [ fig : c_a ] shows the temperature and alfvn speed profiles for the avrett and loeser c7 atmosphere model with an imposed background 50 g field . ) as a function of height above the base of the photosphere for different atmospheric models .
black line corresponds to fal a @xcite , green long dashed line - to avrett and loeser c7 @xcite , and blue short dashed line - to val c @xcite model . ]
alfvn waves are introduced into the domain by driving the bottom boundary of the simulation .
the velocity field is prescribed by equation ( [ eq : driver ] ) to give a kinetic energy spectrum consisting of two regions - the low frequency region , where the power in a mode increases with frequency , and a kolmogorov region at higher frequency , where power drops as @xmath50 : @xmath51 @xmath52 is the frequency of mode @xmath53 in the low frequency region and in this paper runs from 0.001hz to 0.01hz , @xmath54 is the frequency of mode @xmath55 in the high frequency region and in this paper runs from 0.01hz to an arbitrary upper cutoff frequency which ranges between 0.1hz and 10hz . @xmath52 and
@xmath54 are both linearly spaced in frequency .
@xmath56 is the velocity component out of the plane of the simulation which is associated with an alfvn wave .
@xmath57 is the number of harmonics used to produce the driver spectrum and is a large enough number to ensure that the spectrum is reproduced smoothly and that further increase in @xmath57 does not lead to changes in the heating rate greater than 1% .
@xmath58 is set to 10,000 for simulations with a 10hz upper cutoff frequency , 1000 for simulations with a 1hz upper cutoff frequency and 100 for simulations with a 0.1 hz upper cutoff frequency .
convergence testing using 100,000 driver elements for a simulation with a 10hz upper cutoff shows that the result is not sensitive to this parameter so long as the total poynting flux through the photospheric lower boundary is kept constant .
@xmath59 is the phase for spectral component @xmath53 or @xmath55 and is selected randomly for each mode .
the amplitude @xmath60 is selected to give a total energy flux at the bottom boundary of @xmath61 .
this corresponds to a velocity field with an rms value of @xmath62 at the @xmath45 optically thick photospheric surface , comparable with observations of photospheric transverse velocities ( see for example @xcite or @xcite ) .
the spectra of the used drivers , for differing cut - off frequencies , are show in figure [ fig : driver ] .
this form of the driver is based on that used by @xcite , although they specified a net average poynting flux of @xmath63 .
there is no observational evidence for such a driver spectrum .
instead this has been chosen so that the r.m.s .
velocity matches observations along with the total poynting flux , which constrains the magnetic field once photospheric densities and velocities are defined .
equation [ eq : driver ] therefore has observationally constrained average velocity and integrated flux .
there is no direct evidence for high frequency alfvn waves thus the choice of a kolmogorov spectrum is motivated purely by the observation that the photospheric motion is turbulent and hence some scale - free power - law dependence is expected .
the results are insensitive to the choice of spectrum in the rising part of the spectrum , which could have been chosen to be flat , other than its effect on the amplitude limiting to specify the net poynting flux . ) . ]
the net poynting flux from the specified velocity driving on the photosphere needs to be calculated from the simulations . in the single fluid mhd model
this can be rewritten as @xmath64 where @xmath9 is the fluid centre of mass velocity , and @xmath8 is the magnetic field . the commonly used approximation @xmath65 where @xmath56 is the velocity amplitude perpendicular to the magnetic field and @xmath66 is the local alfvn speed , fails for the photospheric driver as this is a linear approximation for waves propagating only in one direction . at the photosphere , where there are reflected waves from the transition region ,
this approximation is not valid .
this is demonstrated in figure [ fig : poynting_variants ] , where the time - integrated poynting vector calculated using both equations ( [ eq : approxpoynt ] ) and ( [ eq : fullpoynt ] ) through the lower boundary of a simulation of the solar chromosphere with no dissipation mechanisms present are shown .
these agree up until the first driven alfvn waves from the driver have returned after partial reflection from the transition regions , around 160 seconds , at which point they diverge .
equation ( [ eq : approxpoynt ] ) overestimates the total poynting flux through the boundary by a factor of up to 2 .
this means that estimates of poynting flux obtained using equation ( [ eq : approxpoynt ] ) , whether from simulation or observation , are almost certain to be overestimated . in this paper
the poynting flux is always calculated using the form in equation ( [ eq : fullpoynt ] ) . + ) , orange dashed line using equation ( [ eq : approxpoynt ] ) . ]
using the 1.5d setup as previously described and assuming that the plasma is incompressible ( @xmath67 ) , the mhd equations ( [ eq : momentum])-([eq : ohm ] ) simplify to @xmath68 normalizing to the constant vertical field and the local alfvn speed such that @xmath69 and @xmath70 it is possible to rewrite these equations in terms of the elssser variables @xmath71 , where @xmath72 and @xmath73 correspond to upwards and downwards propagating alfvn waves respectively . after dropping the primes for normalized variable equations ( 11 - 12 )
are , @xmath74 there is no non - linear coupling between upward propagating and downward propagating waves which would lead to a cascade , i.e. no non - linear interaction between @xmath73 and @xmath72 .
alfvn turbulence is suppressed in 1.5d by the assumption of incompressibility .
thus energy introduced by the driver in a given frequency will stay in that frequency until it is either dissipated or leaves the computational domain .
lare is converted to a 1.5d incompressible code as described in section [ sec : numtech ] and used to simulate an alfvn wave spectrum given by equation ( [ eq : driver ] ) propagating into the avrett & loeser c7 model atmosphere .
4096 grid points are used in the @xmath44 direction , and a length of 6000 km above the visible solar surface is simulated for all results , even those for which results are only plotted for the chromosphere . at heights above 4500 km the velocity damping region is applied .
convergence is tested with a simulation of 8192 grid points , resulting in a change in the solution of @xmath75 as measured by the heating rate . ) .
the green dot - dot - dash line is the estimated heating requirement for the quite chromosphere taken from @xcite . ]
simulating an alfvn wave spectrum given by equation ( [ eq : driver ] ) with a high frequency cutoff of 0.1hz , 1hz or 10hz propagating up into the c7 atmosphere ( figure [ fig : heating_height ] ) , it is found that heating is highest in the photosphere with a local maximum at the temperature minimum .
this is comparable to the results in @xcite and @xcite who used a similar model to that presented here .
since resistive dissipation is more effective for shorter wavelengths and there is no turbulent cascade , the choice of upper cutoff frequency for the driver would be expected to affect the efficiency of the dissipation .
this is observed in figure [ fig : heating_height ] where greater heating at all heights is observed if a higher upper cutoff frequency in the driver spectrum is chosen .
the simulations produce heating rates comparable to those in @xcite , with or without the hall term included in ohm s law .
the hall term is found to have negligible effect ( @xmath76 ) on the heating rate in all simulations .
note that the three lines in figure [ fig : heating_height ] are from driver spectra where the amplitudes have been changed so that the net poynting flux is the same in all three cases .
for the 10 hz cut - off the total heating rate in the chromosphere is only of the order of @xmath77 .
the majority of the input energy in these simulations goes into increasing the energy stored in long wavelength alfvn waves in the chromosphere and not into heating .
a fraction of the total input power does leak into the corona , see later , but the continual buildup of alfvn wave energy at long wavelength throughout the simulations , while correct for the incompressible simulations , is unphysical and can only be corrected by the inclusion of compressibility . , equation [ pedersen ] , ( black solid line ) or cowling resistivity only , @xmath78 ( orange dashed line ) with a 0.1hz cut - off in the driver spectrum . ] as with @xcite it is found that heating in the photosphere and low chromosphere is dominated by dissipation due to classical resistivity ( figure [ fig : resist_height ] ) while heating in the upper chromosphere is mainly due to cowling resistivity .
the largest heating rates occur in the parts of the simulation that are dominated by electron - ion and electron - neutral resistivity .
the qualitative features of figure [ fig : heating_height ] can be reproduced with a simple analytic approach .
this assumes that the heating is from a single pass of alfvn waves and all wavelengths are less than the density scale height .
if only alfvn waves are present then the alfvn wave energy @xmath79 is determined by @xmath80 with @xmath81 given by equation ( [ eq : approxpoynt ] ) and @xmath29 the total resistivity . since @xmath82 in steady state ,
this can be solved to give @xmath83.\ ] ] in this linearized wkb approach with only upward propagating alfvn waves the local heating rate @xmath84 is given by @xmath85 where @xmath86 integrating over @xmath87 then gives the local heating rate for an incompressible chromosphere under the assumption that all the heating is from a single upward propagating pass through the chromosphere and transition region , i.e. ignoring reflection .
this analytic heating rate is plotted in figure [ fig : heating_height ] .
the discrepancy between the simplified analytic model and full simulations are due to the wkb approximation and the absence of reflected waves in the analysis leading to equation ( [ eq : heating_analytic ] ) . also plotted on figure [ fig : heating_height ]
is an estimate of the required volumetric heating rate as a function of height taken from @xcite .
this heating rate is not well defined below the temperature minimum . for the incompressible simulations there is no coupling between modes in 1.5d . as a result
it is easy to estimate the effective transmission coefficient of the transition region as a function of frequency without the complication which would arise from mode coupling in a non - linear simulation .
the transmission coefficient is calculated as the ratio of the poynting flux , as a function of frequency , just above the transition region to the poynting flux at the base of the simulation domain .
this is then smoothed with a boxcar moving average of 0.005 hz width and plotted in figure [ trans_coeff ] .
transmission ranges between 40% at 0.5hz to 0.32% at 0.001hz .
only higher frequencies have appreciable transmission coefficients as the shorter wavelengths see a density ramp rather than a discontinuity at the transition region .
for frequencies above 0.5 hz the measured transmission drops due to these frequencies been damped by pedersen resistivity before they reach the transition region .
the overall energy budget from incompressible simulations is therefore that high frequency , i.e. greater than 0.1 hz , alfvn waves can both heat the upper chromosphere and leak into the corona .
the majority of the chromospheric heating in incompressible simulations come from these high - frequency waves .
for the chosen power law spectrum with r.m.s .
amplitude @xmath88 at the photosphere the typical heating rate above 1000 km is of the order of @xmath89 for a 50 g vertical field .
the peak chromospheric heating required to balance radiative and conductive losses are @xmath90 with rates of @xmath91 needed in the upper chromosphere @xcite .
higher heating rates could be achieved with the kolmogorov driver used here by choosing a stronger background field strength or alternatively a higher r.m.s .
velocity driver could be used .
either of these would however increase the poynting flux in the observed low - frequency part of the spectrum and thus contradict observations .
another possibility is that distribution of alfvn waves deviates significantly from a @xmath92 power law and has more energy at higher frequencies .
this would require orders of magnitudes more energy in the high - frequencies contradicting our assumption of a turbulent photospheric driver with a power - law spectra and is not studied further here .
the incompressible model in 1.5d can not be turbulent and there is no cascade of energy to smaller scales .
the heating rate from pedersen resistivity can be made to match the chromospheric heating requirements by ensuring sufficient energy is input into the base of the simulation in frequencies which are damped by ion - neutral collisions .
the heating rates as a function of height are therefore entirely dependent on the user - specified input spectrum which for these frequencies is poorly constrained by observations .
work by hollweg @xcite and others have also shown the importance of ponderomotive coupling of alfvn waves to slow and fast modes in a compressive medium . in this section
we repeat the simulations of section [ sec : incompressible ] for the full set of compressible mhd equations . as before the heating due to pedersen resistivity
is calculated but not added into the simulation thermal energy .
this approach is not possible for the shock heating which is allowed in compressible simulation , as this is required to ensure the correct shock jump conditions .
an inevitable consequence of this is that the model atmosphere heats up due to shock dissipation .
the pedersen resistivity is however kept at the same profile as derived from the initial conditions . in the absence of conductive and radiative losses
the shock heating would ionize the chromosphere and turn - off cowling resistivity , i.e. that part of the pedersen resistivity due to ion - neutral collisions .
this lack of self - consistency is shown below to not be significant as the dominant heating is through shock dissipation not resistivity .
a limitation of the 1.5d model used here is that the net ponderomotive force from a persistent photospheric driver lifts the transition region and changes the atmospheric density structure .
this is shown in figure [ fig : density ] . in a compressible simulation energy cascades to shorter scale - lengths
are now permitted .
this can clearly be seen in figure [ comp_spect_en ] , where the spectral energy in alfvn waves from a driver introducing wave energy between 0.001 hz and 0.1 hz ( black line ) is compared to the energy present in the upper chromosphere ( 2000 km above the photospheric surface ) .
power at frequencies above the driver cutoff are clearly present in the upper chromosphere .
this is not a purely alfvnic cascade as it shows the characteristic -2 spectral power of a shock dominated system , but it does provide a mechanism by which shorter length scales can be created self - consistently within the simulation . as a check of the simulations for consistence against observations we note that the velocity field 2000 km above the photosphere in these simulations is @xmath93 km / s .
transverse velocities observed for spicules by de pontieu et al .
@xcite were @xmath94 km / s . .
both lines are smoothed with a boxcar moving average of width 0.05 hz . [ comp_spect_en ] figure [ shock_plot ] shows that slow mode shocks are produced in the simulations . since the entropy jump across the shock can heat the plasma , care must be taken to correctly reproduce this shock heating . in order to ensure that shocks in the simulations satisfy the entropy condition in the shock jump relations , lare uses the artificial viscosity formulation of @xcite which both viscously changes the velocity and deposits the associated viscous heating . in the limit of infinite resolution
this viscosity goes to zero in smooth regions of the solution and satisfies the shock jump conditions at discontinuities .
however , at finite resolution viscosity is erroneously applied to parts of the solution which are steep on the scale of the grid , but are not true discontinuities .
this means that careful convergence testing of the results of the simulations is required to confirm that the heating is due to shocks and not due to numerical error .
this is shown in figure [ heating_compress ] where it is shown that shock heating is more important in the upper chromosphere than resistive heating , and also that the shock heating rate is converged in these simulations to within 10% in the chromosphere .
while shock heating rates in the photosphere are not converged , they are clearly lower than the well - converged resistive heating rate at those heights
. the light red dash dot line on figure [ heating_compress ] shows the heating rate from the incompressible simulations .
it can be seen that the overall resistive heating rate is similar in both compressible and incompressible simulations , but that additional resistive heating in the upper chromosphere is observed in the compressible simulations . despite this , it is still 4 orders of magnitude lower than the shock heating rate .
figure [ visc_freq ] shows that shock heating in the compressible simulations is the same for both cut - off frequencies above 0.1 hz .
no results are shown for a cut - off of 10 hz as these are the same as those for 1 hz .
the shock heating rate is higher than the heating rate needed to balance conduction and radiation in the chromosphere for a realistic driver amplitude .
this may be explained by artificial effects of these 1.5d simulations .
in particular , the missing vertical structure of the magnetic field and the artificial enhancement of shock strength due to shock coalescence - only possible in 1.5d . it is therefore possible to limit the heating due to pedersen resistivity by imposing a cut - off on photospheric driver frequencies of 0.1 hz while maintaining the same level of shock heating .
the extent to which this may be true in a realistic chromosphere can not be addressed in 1.5d models and must be tested in 2d and 3d . for a driver cut - off frequency of 0.01 hz the shock heating
is reduced by three orders of magnitude
the simulations presented in this paper show that it may be possible to heat the solar chromosphere by dissipating alfvn wave energy . if energy is present in high enough frequency modes then resistive heating from pedersen resistivity is sufficient to heat the solar chromosphere directly .
the presence of such high frequency modes is not currently observable .
however , if compressive effects are included then shock heating is capable of heating the chromosphere for a broad range of driver spectrum in a 1.5d model . in compressible simulations shock heating
dominates over resistive dissipation in the upper chromosphere even for driver spectra with high frequency components .
a problem with all 1.5d models is that shocks can coalesce and as a result the shock heating reported here is certainly an over - estimate .
also absent from the current model is the flux expansion expected in chromospheric flux tubes .
in addition , the possibility that the alfvn turbulent cascade , which requires at least two spatial dimensions and would be fastest in 3d , may increase the relative importance of resistive dissipation by increasing the rate at which long wavelength alfvnic energy cascades to the resistive dissipation scale . despite these omissions from the current 1.5d modelling
there are still a number of important conclusions which can be drawn and questions asked to guide future 2d and 3d work .
* evidence from over the last 30 years , e.g. references from @xcite onwards , have all shown that a low - frequency photospheric driver can excite alfvn waves which are of sufficient energy to heat the chromosphere through shock heating .
the slow modes are generated in the chromosphere via ponderomotive driving .
the present study has focused on the chromosphere in more detail , and higher resolution , than previous studies , included pedersen resistivity and confirmed this potential chromospheric heating mechanism .
* while the incompressible results from @xcite are confirmed the observed heating as a function of height is entirely a function of the user specified driver since no energy cascade is present .
furthermore this is not changed by the inclusion of the hall term .
most importantly , the results presented here show that compressible effects dominate the heating by orders of magnitude over pedersen dissipation in the upper chromosphere . *
a turbulent cascade leading to heating @xcite is a possible heating mechanism where alfvn wave energy can be thermalized .
however , compressible mhd effects must be included to check if in multi - dimensions the dominant energy loss from a low - frequency driver is through an alfvn cascade or through coupling to slow modes .
the present 1.5d study suggests that coupling to slow modes , which subsequently shock and heat the chromosphere , is the dominant heating mechanism . in 2d and
3d the alfvnic turbulent cascade to short scales , and hence dissipation , may be terminated at the dissipation scale of pedersen resistivity .
alternatively , alfvnic wave energy may not reach these scales due to energy loss to slow modes via ponderomotive coupling , through mode coupling at the @xmath95 surface or geometric coupling through flux tube expansion .
while only 1.5d , the current study strongly suggests ponderomotive coupling as the dominant mechanism by which alfvn waves loose energy in the chromosphere .
a final possibility is that energy leaks through the transition region ( figure [ comp_spect_en ] ) faster than it cascades to the dissipation scale . only full simulations in 2d and 3d will be able to determine which of these effects is most important .
this work was supported by stfc grant st / l000733/1 and results were obtained using the dirac hpc facility ( www.dirac.ac.uk ) .
this equipment was funded by bis national e - infrastructure capital grant st / k00042x/1 , stfc capital grants st / h008519/1 and st / k00087x/1 .
dr shelyag is the recipient of an australian research council s future fellowship ( project number ft120100057 ) .
the authors are grateful for the financial support provided by monash - warwick alliance seed fund .
avrett , e. h. in `` solar phenomena in stars and stellar systems '' edited by r.m .
donate and dupree a. k. 1981 ( proceedings of the advanced study institute , bonas , france , august 25-september 5 , 1980 . )
dordrecht , d. reidel publishing co. , 1981 , p. 173 - 198 | physical processes which may lead to solar chromospheric heating are analyzed using high - resolution 1.5d non - ideal mhd modelling .
we demonstrate that it is possible to heat the chromospheric plasma by direct resistive dissipation of high - frequency alfvn waves through pedersen resistivity .
however this is unlikely to be sufficient to balance radiative and conductive losses unless unrealistic field strengths or photospheric velocities are used .
the precise heating profile is determined by the input driving spectrum since in 1.5d there is no possibility of alfvn wave turbulence .
the inclusion of the hall term does not affect the heating rates .
if plasma compressibility is taken into account , shocks are produced through the ponderomotive coupling of alfvn waves to slow modes and shock heating dominates the resistive dissipation . in 1.5d
shock coalescence amplifies the effects of shocks and for compressible simulations with realistic driver spectra the heating rate exceeds that required to match radiative and conductive losses .
thus while the heating rates for these 1.5d simulations are an overestimate they do show that ponderomotive coupling of alfvn waves to sound waves is more important in chromospheric heating than pedersen dissipation through ion - neutral collisions . |
the investigation of small atomic and molecular clusters remains an active area of research , both theoretically@xcite and experimentally.@xcite computer simulations of clusters have revealed a wide diversity of fundamental behavior that is highly sensitive to many factors , such as the cluster aggregate size @xmath4 , the potential used to model the intermolecular interactions , and even the simulation method itself .
the lennard - jones ( lj ) potential is the most commonly used model potential for computer simulations of cluster properties , and it is especially useful for modeling rare gas clusters .
a key aspect of this model when applied to small and medium sized atomic clusters is that the resulting potential surfaces have global minima that are dominated by icosahedral - like structures , with fivefold symmetries that are markedly different than the closest - packed structure found in corresponding bulk systems .
a major motivation for studying clusters is the insight that can be obtained in understanding the transition from finite to bulk behavior . because of their high proportion of surface atoms ,
many clusters have physical properties that typically exhibit a very irregular dependence on their aggregate size , with certain sizes standing out in particular .
for example , the mass spectra of clusters typically exhibit especially abundant sizes that often reflect particularly stable structures , specially reactive clusters , or closed electronic shells,@xcite these `` magic number '' sizes have been of great theoretical interest,@xcite and much work has been done in relating magic number sequences to cluster structure in terms of `` soft '' sphere packings , since many magic number sizes in lj clusters such as @xmath8 , 19 , 55 , and 147 correspond to compact structures that are especially stable.@xcite many cluster properties obtained from simulation studies such as binding energies , root mean square ( rms ) bond length fluctuations and heat capacities also show very irregular dependencies on the cluster size that mostly correspond to their usual magic number sequences . for example , the heat capacity peak values per atom for lj@xmath9@xcite and lj@xmath10@xcite are especially large , which can be attributed to the exceptional thermodynamic stability of their lowest - energy isomers relative to their next higher energy configurations . to more easily compare various cluster property size dependencies in this report
, i have generalized the concept of magic numbers for arbitrary cluster properties to refer to those cluster sizes having particularly enhanced values relative to their immediate neighbors . in a previous study , i examined these generalized magic number effects in lj clusters ranging in size from 4 to 24 atoms.@xcite in this study , i extend the sequence to 60 atoms , which completes the filling of the second mackay icosahedral shell ( to @xmath11 ) and just begins the filling of the third shell ( to @xmath12 ) .
the issue of cluster phase transitions is fundamentally important , and so has been of great interest .
because of their small size , clusters do not have the sharp first order transition characteristic of bulk melting , but instead show changes occurring over a range of temperatures , typically a few degrees k.@xcite although the phase - change range generally decreases as the cluster size increases , coalescing to the bulk transition temperature as @xmath4 approaches infinity , both the range and the location show much variation as a function of @xmath4 .
cluster solid - liquid phase - change regions can be determined from computer simulations typically by identifying characteristic changes in certain cluster properties .
for example , cluster potential energies often change markedly as functions of temperature in the phase - change regions , resulting in distinct peaks in the corresponding heat capacity curves ; rms bond length fluctuations often rise abruptly with increasing temperature at the beginning of the phase - change regions , and the fractions of quenched cluster configurations corresponding to the lowest - energy isomers often decrease sharply with increasing temperature in these regions . defining the phase - change region on the basis of such behavior
is somewhat ambiguous , though , since the temperature ranges associated with these different properties do not always coincide , and occasionally do not even overlap . for a given property , the phase - change region temperature range can also depend on the simulation method ; for example , for ar@xmath9 , the rise in the rms bond length fluctuations as calculated from canonical monte carlo simulations occurred about 7 k lower than that calculated from molecular dynamics simulations.@xcite because of these difficulties , it is convenient to loosely define the solid - liquid phase - change region as the temperature region between the lower - temperature , solid - like region ( where the clusters are mostly confined to their lower energy configurations , undergoing isomerizations only rarely at the higher - temperature end of the region ) and the higher - temperature , liquid - like region ( where the clusters exhibit easy isomerization and diffusion ) .
cluster behavior in the solid - liquid phase - change regions is particularly complicated , with some cluster sizes showing coexistence behavior , where `` solid - like '' and `` liquid - like '' isomers dynamically coexist,@xcite while other cluster sizes show a relatively smooth progression of isomerizations between locally similar but globally distinct configurations occurring throughout this region.@xcite although clusters having magic number sizes such as 13 and 55 have pronounced heat capacity peaks that correspond closely to the solid - liquid phase - change region determined from other cluster properties , other cluster sizes such as @xmath13 , 16 and 17 do not have heat capacity peaks in their phase - change regions.@xcite since heat capcity peaks have often been used to identify phase - change regions,@xcite characterizing their temperature dependencies for sequences of lj clusters can help elucidate the underlying reasons for such varied behavior .
the accurate determination of cluster properties such as the heat capacity is hampered by the poor convergence of standard simulation methods in the phase - change regions .
this poor convergence is a consequence of quasiergodicity , or the incomplete sampling of configuration space.@xcite various methods have been developed in recent years to reduce the systematic errors resulting from quasiergodicity , including histogram methods,@xcite jump - walking methods ( j - walking),@xcite smart walking methods ( s - walking),@xcite and parallel tempering methods.@xcite many of these methods are based on the coupling of configurations obtained from ergodic higher - temperature simulations to the quasiergodic lower - temperature simulations . in my previous study of magic number effects in cluster heat capacities , i used j - walking to successfully overcome quasiergodicity .
monte carlo j - walking methods couple the usual small - scale metropolis moves made by a lower - temperature random walker , with occasional large - scale jumps that move the random walker out of the confined regions of configuration space that it could not otherwise easily escape within the duration of the simulation .
there are two complimentary implementations of j - walking . in tandem
j - walking , higher - temperature walkers and lower - temperature walkers are run in tandem , with the higher - temperature walkers occasionally feeding configurations to the lower - temperature walkers .
because j - walking obeys detailed balance only approximately , the walkers become correlated whenever a lower - temperature walker accepts a jump , and so it is necessary to run several extra metropolis passes after a jump is accepted to break these correlations , thereby decreasing the overall efficiency .
the second method breaks the correlations by running the higher - temperature walkers beforehand , and storing representative samples of the configurations in external files , which the lower - temperature walkers sample periodically in a random fashion .
thus , sampling from an external distribution is more efficient , provided that the required distributions are not too large , and this was the implementation i used in the previous study .
it was also successful for the first few clusters considered in this study , but for clusters ranging in size from @xmath14 to 37 , which had two heat capacity peaks , the heat capacity curves obtained with different j - walker distributions showed poor reproducibility for the lower - temperature peaks .
since the distribution file sizes already totalled more than 2 gb , it seemed prudent to abandon that implementation and switch to the tandem j - walker implementation .
however , a recent study by neirotti , _
et al._@xcite of lj@xmath15 , which is notoriously difficult to simulate accurately , showed that the parallel tempering method was remarkably successful in overcoming quasiergodicity and calculating an accurate heat capacity curve .
parallel tempering is nearly identical to a parallel version of tandem j - walking , except that instead of copying the higher - temperature j - walker s configuration to the lower - temperature walker s configuration whenever a jump is accepted , the two configurations are exchanged .
this has the advantage of maintaining detailed balance , thereby eliminating the need to run extra metropolis passes after accepting an exchange , which makes the method more efficient than tandem j - walking .
thus , i decided to use parallel tempering for all of the cluster sizes examined in this study .
i begin in section [ sec : method ] with a brief description of the computational methods used in this work and a summary of the calculations undertaken .
section [ sec : results ] describes the results obtained for the cluster potential energies , heat capacities , rms bond length fluctuations and quenched configuration distributions , as functions of temperature and cluster size . in section
[ sec : discussion ] , i summarize the important conclusions concerning magic number behavior in lj clusters for @xmath16 .
the clusters studied were modeled by the pairwise additive lennard - jones potential , @xmath17.\end{aligned}\ ] ] because small clusters can dissociate at higher temperatures and make the cluster definition ambiguous , it is customary to confine the clusters by a constraining potential . for classical monte carlo simulations , a perfectly reflecting spherical constraining potential of radius @xmath18 , centered on the cluster s center of mass , is most convenient , and was used in this work .
the classical internal energy and heat capacity were calculated by the usual expressions for an @xmath4-atom cluster ( in reduced units ) , @xmath19 where @xmath20 , @xmath21 , @xmath22 , and @xmath23 .
since so many rare gas cluster simulations have been done for ar , i found it convenient to generate the temperature - dependent curves using ar parameters , with @xmath24 and @xmath25 k. relative rms bond length fluctuations have been calculated in many previous cluster simulations.@xcite these were calculated as @xmath26 where the summations are taken over the entire walk . because of the summations being taken over the entire walk , rms bond length fluctuations can not be determined for the j - walking and parallel tempering simulations , whose configuration jumps and exchanges effectively reduce the long metropolis walks to a collection of very many short excursions about the last accepted jump or exchange configuration .
thus , quasi - ergodicity remains a problem for rms bond length calculations .
caution must also be exercised when interpreting @xmath27 curves , since they are dependent on simulation parameters such as the walk length , as well as the ensemble that is used.@xcite j - walking runs were done for cluster sizes @xmath28 , using sampling from externally stored distributions as described in the preceding study.@xcite distributions for each temperature consisted of several files , totaling @xmath29 configurations , sampled every 100 , 50 , 25 , or 10 passes , depending on the temperature ( 100 passes for higher temperatures having broad distributions , 10 passes for very low temperatures having very narrow distributions ) .
the distributions were generated in stages from higher temperatures to lower , beginning with temperatures well within the cluster liquid region ( 50 to 60 k ) .
data accumulation was done with walks consisting of @xmath30 total passes , with jumps attempted randomly with a frequency of 10% . since any systematic errors in the higher temperature j - walker distributions can feed into the lower temperature distributions and corrupt them , heat capacity peaks occurring at very low temperatures can be very difficult to calculate accurately . in previous studies of 13-atom binary clusters,@xcite
very low temperature heat capacity peaks corresponding to mixing anomalies could be obtained accurately by repeating the full j - walking temperature scan several times , each time with newly generated distributions , and averaging the independent curves .
this procedure was followed in this work as well , and at least five j - walking runs were done for cluster sizes 25 to 30 , with the heat capacity curves showing very good reproducibility in each case . for cluster sizes
31 to 38 though , the curves showed very poor reproducibility for the lower - temperature peaks , and so the method was abandoned . despite the j - walking method s inability to reproduce the lower - temperature heat capacity peaks accurately , the method was still useful for obtaining a listing of the important lowest - energy configurations for each cluster size .
in each case , representative subsets of the distribution files were quenched using steepest descent methods@xcite and all of the unique isomers found were extracted . while this crude method can not identify all of the configurations for clusters of these sizes , it is well suited for identifying most of the lower energy isomers that dominate the low temperature behavior of clusters .
for example , the quenched j - walker distributions for lj@xmath15 contained all of the lowest - energy configurations found by doye _
et al._,@xcite including the global minimum face - centered - cubic ( fcc ) truncated octahedron , which is notoriously difficult to obtain from randomly initialized simulations .
these findings imply that j - walking was successful in accessing all of the important regions of configuration space for these clusters , even though it was unable to do so in the statistically representative manner needed to accurately reproduce the heat capacities at very low temperatures .
thus , higher temperature j - walker distributions were also generated for cluster sizes 39 to 60 to obtain sets of the important low - energy quenched configurations .
metropolis simulations were also run for each cluster .
these provided a check of the j - walking and parallel tempering results for those temperature regions where quasiergodicity in the metropolis runs was not a problem , as well as revealing trends in the systematic errors arising from quasiergodicity in the metropolis runs themselves .
temperature scans were also generated using ar parameters , with the temperature mesh size set to @xmath31 k , and sometimes to 0.5 k in the phase - change regions .
for each temperature , simulations consisted of @xmath32 warmup passes , followed by @xmath33 passes with data accumulation .
the scans in each case were started at the lowest temperature from the global minimum configuration obtained from the corresponding j - walker simulations , and were continued well past the cluster solid - liquid phase - change region , with the final configuration for each temperature used as the initial configuration for the subsequent temperature . to gain a better appreciation of the extent of quasiergodicity in the solid - liquid phase - change region , additional metropolis simulations consisting of @xmath34 total passes per temperature point were run for each cluster over a narrower temperature range encompassing the phase - change region .
parallel tempering@xcite is very closely related to the tandem and parallel walker implementations of j - walking,@xcite with the essential difference being that instead of the higher - temperature configuration being copied over the lower - temperature configuration whenever a jump is accepted , the two configurations are exchanged .
this close similarity between parallel tempering and tandem j - walking makes it very easy to convert a tandem j - walking simulation program to a parallel tempering one . the parallel tempering method s configuration exchange ensures that it obeys detailed balance ( provided that the simulation has been warmed up enough to be in the asymptotic region ) , so there is no need to run the extra metropolis passes to break correlations between the higher - temperature and lower - temperature walkers whenever an exchange move is accepted , which is required for tandem j - walking .
in addition to offering greater efficiency , the obeying of detailed balance allows parallel tempering simulations to vary the exchange - attempt frequency in temperature regions where the exchange - acceptance rate is too low . in j - walking simulations ,
the jump - attempt frequency was typically kept at 10% , since larger values required even more extra metropolis passes to be run whenever a jump was accepted to break the correlations , thereby offsetting the efficiency gains made by increasing the jump - attempt frequency in the first place . as with j - walking ,
the number of temperatures used in a parallel tempering simulation and their spacing can not be chosen arbitrarily , but must be selected to ensure that the exchanges are accepted sufficiently often .
generally , more temperature walkers and smaller spacing are required for the solid - liquid phase - change region , and for very low temperatures . since the heat capacity curves as a functions of temperature typically had peaks in the solid - liquid phase - change regions whose location and height i wanted to determine accurately , the number of temperatures and their spacing was also affected by the need to properly represent the heat capacity curves with a sufficient density of points to allow for good interpolation .
this latter requirement resulted in my choosing temperature spacings in the phase - change region that were small enough that the exchange acceptance ratios were typically quite high .
for very low temperatures , the classical heat capacity curve is nearly linear , thus requiring fewer points to represent .
however , the exchange acceptance ratio decreases rapidly at lower temperatures , requiring more temperature walkers .
since the overall computation time is nearly proportional to the number of walkers , i was able to reduce the number of lower - temperature walkers needed by increasing the exchange - attempt frequency @xmath35 at lower temperatures , according to the relation @xmath36 where @xmath37 .
this expression increases the lowest - temperature exchange - attempt frequency @xmath38 by a quarter of the range above the baseline highest - temperature exchange - attempt frequency , @xmath39 , which was set to 10% .
the parallel tempering simulations done in this study were similar to those done by neirotti _
et al . _ for lj@xmath15.@xcite for each cluster simulation
, a number of random walkers at various temperatures were run in parallel , with each temperature walker occasionally attempting an exchange move with the next higher temperature walker , except for the highest temperature walker , which attempted no exchanges directly ( but did exchange configurations according to attempts made by the next - highest temperature walker ) . since i had the lowest - energy isomers for each cluster from the j - walking distribution quenches , for most of the clusters i initialized the temperature walkers by using the lowest - energy isomer for the lowest - temperature initial configuration , the next lowest - energy isomer for the next lowest - temperature initial configuration , and so on .
for the other clusters , i just used the lowest - energy isomer to initialize all of the initial configurations .
it did not seem to matter which initialization scheme was used , and of course , random configurations could have been used , but at the cost of longer warmup periods . because i initialized the simulations from the lowest - energy isomers ,
i could quickly run preliminary simulations that gave qualitatively good results , which i could use to fine - tune the number of parallel tempering temperature walkers and their spacing .
these preliminary parallel tempering runs consisted of @xmath32 metropolis warmup passes for each temperature walker , followed by @xmath32 parallel tempering warmup passes and then by 10 parallel tempering walks with preliminary data accumulation , with each walk being @xmath32 passes in length .
the configurations for each temperature walker were saved at the end of the warmup period , and at the end of each walk .
the heat capacity curves for each walk were examined for consistency , and if significant variations were found , the parallel tempering run was repeated ( without the metropolis warmup passes ) from the last saved configurations , so that in effect , all of the preceding parallel tempering runs served as warmup runs for the final run . for some of the clusters ,
only parallel tempering warmup passes were run , indicating that the preceding metropolis warmup passes were likely not necessary ( although their computational expense was minor ) .
after the preliminary runs appeared to be sufficiently warmed up , a longer parallel tempering run of @xmath32 warmup passes followed by @xmath33 total passes of data accumulation per temperature point was run ( 10 walks of @xmath40 passes each ) .
again , the heat capacity curves for each walk were checked for consistency ; for a few clusters , the variations in the curves were still too large , and so the run was repeated from the last saved configurations . for lj@xmath15 ,
an especially long warmup was required , consistent with the findings of neirotti _
et al_.@xcite one important difference between my implementation of parallel tempering and that of neirotti _ et al _ concerns the choice of the constraining radius @xmath18 .
ideally , the value of the constraining radius should be large enough that the cluster properties being simulated are not adversely affected by the constraining potential , but not so large that cluster dissociations lead to overly long convergence times . in my previous study of clusters of size 4 to 24 ,
i used @xmath41 , a value large enough not to affect the heat capacity curves , except at higher temperatures , well past the solid - liquid phase - change regions .
thus , for this study , i also set @xmath42 for the metropolis and j - walker simulations .
et al _ used a value of @xmath43 for their simulation of lj@xmath15 , since they had difficulties attaining ergodicity with larger constraining radii . for lj@xmath15 ,
such a small constraining radius was not a problem since the heat capacity peak occurred at a relatively low temperature .
however , for most of the cluster sizes included in this study , such a small value would affect the heat capacity peak height and location . after some preliminary investigations , i chose to set @xmath44 for parallel tempering simulations of cluster sizes 25 to 48 , and to @xmath41 for the larger cluster sizes . figure [ fig : ar30_rc ] illustrates the effects of the constraining radius on clusters ar@xmath45 and ar@xmath46 .
the heat capacity peak for @xmath47 is slightly truncated , but the peak for @xmath44 is in good agreement with the @xmath42 peak , even though the high - temperature side of the peak is substantially affected .
these two examples also show how well parallel tempering was able to overcome the quasiergodicity evident in the metropolis curves .
the two parallel tempering curves for ar@xmath46 show very good agreement for the smaller peak , which is completely absent in the metropolis simulations .
the converged heat capacity curve i obtained for ar@xmath15 was mostly similar to the curve obtained by neirotti _
_ in their parallel tempering study of lj@xmath15,@xcite with a small shoulder in the low - temperature side of the peak .
the reduced peak temperature 0.1653 was also in good agreement with that found by neirotti _
et al . _ ( 0.166 ) , although my peak value was slightly higher ( @xmath48 compared to 5.62 ) and the higher - temperature part of my curve rose substantially higher ; these discrepancies are mostly due to neirotti _
_ having used the smaller constraining radius . as pointed out by neirotti _
_ , the shoulder in the heat capacity curve is in contrast to the small lower - temperature peak found by doyle , wales and miller in their study of lj@xmath15@xcite ( their larger peak location and height were similar to the parallel tempering results , though , at 0.17 and 5.68 , respectively ) .
finally , follow - up parallel tempering simulations of @xmath40 passes per temperature were run for each cluster with periodic quenching of the configurations to determine the population distributions of the lower - energy isomers as functions of temperature .
these were initialized using the final configurations from their corresponding preceding parallel tempering runs . since i was primarily interested in using the quench results to provide qualitative insight into the heat capacity results , rather than determining the quench curves themselves with high accuracy , these shorter walk lengths were sufficient .
much of the magic number behavior of cluster heat capacities can be understood in terms of the magic number behavior of the underlying potential surfaces .
for example , for the small clusters i studied earlier,@xcite the largest heat capacity peaks were observed for sizes @xmath8 , 19 and 23 , corresponding to single , double and triple icosahedra , respectively .
the same magic numbers were found in some corresponding structural properties as well , such as the binding energy differences , @xmath49 .
\label{vmin}\ ] ] the lowest - energy configurations for each of the clusters i examined in this study were obtained from quenches of j - walker distributions .
these were consistent with the lowest - energy configurations found in other studies.@xcite from @xmath1 to 30 , the global minima all consisted of a @xmath50 double icosahedral core with atoms added to tetrahedrally bonded face sites ( anti - mackay overlayers ) , while for @xmath51 to 54 ( except for @xmath52 ) , the added atoms occupied sites corresponding to the outer shell of the @xmath11 mackay icosahedron ( mackay overlayers);@xcite @xmath52 did not fit the pattern , it being a truncated octahedron instead.@xcite clusters for @xmath53 to 60 had added atoms that contributed to the anti - mackay overlayer of the 55-atom core , beginning the building up of the next mackay icosahedron ( @xmath54 ) . although the lowest - energy isomers play a significant role in cluster thermodynamics , other low - energy isomers also play important roles , and the size of the gap between the lowest - energy isomer and the next lowest - energy isomer is an especially important predictor of magic number behavior.@xcite figure [ fig : energylevels ] shows the energy levels of the lower - energy configurations for each cluster size , relative to their global minimum .
sizes @xmath55 and 58 have relatively large gaps between their global minimum and next higher minimum , compared to their immediate neighbors , and so can be considered a magic number sequence ; the gaps between the lowest - energy and next lowest - energy isomers are also plotted in figure [ fig : energyn ] as functions of @xmath4 .
the mackay icosahedron @xmath11 has an especially large gap .
there is also much variation in the densities of the low - lying isomers evident in figure [ fig : energylevels ] , with densities being substantially greater for sizes @xmath56 to 37 than for the other sizes , while sizes @xmath57 to 56 have particularly low densities .
the binding energy differences for the clusters are also displayed in figure [ fig : energyn ] . except for @xmath52 ,
the maxima in the binding energy differences are the same as the maxima in the energy differences between the two lowest - energy isomers . also shown in figure [ fig : energyn ]
is a plot of another indicator of structural magic number behavior , the second finite difference of the energy , @xmath58 these magic numbers are also in agreement with the magic numbers obtained for the binding energy differences , again , except for @xmath52 .
the discrepancy in the magic number sequence for @xmath52 is due to that cluster being unique in this range by having an fcc structure as its global minimum , so that those measures that are relative to the cluster s immediate neighbors are considerably different than those of the other clusters , whose immediate neighbors are structurally similar .
figure [ fig : del2e_nt ] shows a generalization of the zero - temperature @xmath59 curve shown in figure [ fig : energyn ] , with the curves extended to non - zero temperatures to form a @xmath60 surface .
the curves were obtained from parallel tempering simulations ; the temperature scale is for ar . the surface can be seen to encompass two general regions : a flat region at higher temperatures , corresponding to the clusters liquid - like regions , and a corrugated region at lower temperatures , corresponding to the clusters solid - like regions .
the plot shows that many of the clusters are locked in their lowest - energy isomers for extended temperature ranges , especially the magic number sizes @xmath61 , 29 , 39 , 43 , 46 , 49 and 55 the @xmath62 peak values differ very little from their zero - temperature values well into the phase - change region , until their abatement over a relatively narrow temperature range .
the plot also shows that the phase - change temperature regions have a strong dependence on the cluster size , with the melting temperature generally decreasing from about 30 k for @xmath61 to about 20 k for @xmath63 , and then generally increasing to a maximum of about 40 k for @xmath64 .
figure [ fig : potential ] plots the reduced potential energy per particle for each cluster size over a temperature range spanning the solid - like and liquid - like regions .
magic number behavior can be seen for sizes @xmath61 , 29 , 32 , 36 , 39 , 43 , 46 , 49 and 55 by the relatively large differences at lower temperatures between their potential energies and those of their preceding cluster size .
the mackay icosahedron @xmath11 is so stable relative to its neighbors that its curve actually crosses several of the other curves .
the phase regions are apparent in figure [ fig : potential ] by the transition from the irregular spacing between the curves seen for the lower temperature , solid - like regions , to the regular spacing seen for the higher temperature , liquid - like regions .
as was seen in the @xmath60 surface in figure [ fig : del2e_nt ] , the melting temperature generally decreases as the cluster size increases from @xmath1 , reaching a minimum of about 20 k for @xmath4 about 36 , and then increases to a maximum for @xmath11 .
curves of the reduced heat capacity per particle for each cluster size are shown in figures [ fig : cv1 ] , [ fig : cv2 ] and [ fig : cv3 ] , while figure [ fig : magic ] shows the corresponding peak height and temperature as functions of the cluster size ; the peak parameters are also listed in table [ tbl : heat capacity peak ] .
the peak parameters were obtained by smoothing and interpolating the combined parallel tempering data ( 10 walks of @xmath40 passes per temperature point each ) .
the associated parameter uncertainties were estimated by determining the peak height and location for each of the 10 walks separately , and calculating the standard error of the set of these values . as a check
, additional runs of @xmath33 total passes per temperature were done for a few of the clusters , and the standard deviations of the resulting set of peak parameters were compared to the corresponding uncertainty estimates .
these were found to be in good agreement .
the extra check runs also showed the parallel tempering heat capacity results to be very reproducible , unlike the corresponding j - walking results .
each of the clusters had at least one peak in its heat capacity curve , and overall , while their was much variation in the curves , systematic trends are evident . cluster sizes @xmath65 and 55 had peak heights that were substantially greater than those of their immediate neighbors , and thus represent the magic numbers for the heat capacity sequence .
these cluster sizes were also magic numbers for the @xmath59 sequence and the lowest - energy isomer gaps shown in figure [ fig : energyn ] , but the other structural magic numbers ( @xmath66 and 32 ) did not show magic number behavior in their heat capacity curves .
instead , for cluster sizes 25 to 36 , both the heat capacity peak height and the peak temperature showed remarkably smooth variation , with the peak temperature slowly increasing , while the peak height slowly decreased in a nearly linear fashion , until the peak disappeared at @xmath2 .
as can be seen in figure [ fig : cv1 ] , this behavior was somewhat affected by the choice of the constraining radius .
for the metropolis simulations run at @xmath41 , the heat capacity curves had greater slopes at higher temperatures , resulting in the peak disappearing by @xmath67 .
more interesting was the emergence and evolution of a second heat capacity peak , as the first peak gradually disappeared .
beginning with @xmath3 , a small , lower - temperature peak can be seen in figure [ fig : cv1 ] to develop and gradually increase in both height and temperature , until it becomes the dominant peak at @xmath63 .
this trend can be seen to continue in figures [ fig : cv2 ] and [ fig : cv3 ] , until the peak reaches its greatest height and highest temperature for the magic number size @xmath11 .
this behavior suggests partitioning the heat capacity peaks into two sequences , with the first sequence containing the small-@xmath4 peaks that evolve into the large , broad high - temperature peaks as the second mackay shell initially builds up , while the second sequence originates from the smaller , narrower , lower - temperature peaks that evolve into the dominant peaks as the second mackay shell is completed .
the possible beginning of a third sequence corresponding to the initial building up of the third mackay shell can be inferred from the curves in figure [ fig : cv3 ] , where ar@xmath68 has a small , lower - temperature peak , and ar@xmath69 and ar@xmath70 have discernable shoulders in the low - temperature sides of their peaks .
the emergence of the smaller peak nearly coincides with the transition at @xmath51 from the anti - mackay overlayer in the lowest - energy configuration to the mackay overlayer , and the cluster sizes where two prominent heat capacity peaks are evident correspond to those sizes having the greatest density of low - lying isomers ( figure [ fig : energylevels ] ) .
_ found that for clusters in this size range , the high density of isomers was due to the small differences in energy between the various configuration types , anti - mackay , mackay , decahedral , fcc , etc;@xcite similar findings were obtained for clusters bound by a morse potential whose range parameter is similar to the lennard - jones potential.@xcite the two heat capacity sequences make the determination of the cluster melting temperature from heat capacity peak temperatures ambiguous for sizes @xmath51 to 36 .
peaks in heat capacity curves have typically been associated with cluster phase changes , particularly solid - liquid changes.@xcite however , peaks can arise from other circumstances as well , such as mixing anomalies@xcite and premelting effects , which can make the assignment of the cluster melting temperatures difficult .
calvo and spiegelmann defined the melting temperatures of the sodium clusters they studied whose heat capacity curves had multiple peaks , as the temperature associated with the highest peak.@xcite applying such a definition to the clusters shown in figure [ fig : cv1 ] that have two heat capacity peaks would imply that the reduced melting temperature would jump from @xmath71 for @xmath72 to 0.1427 for @xmath63 .
given the structural similarities of these two clusters , such a large difference in the melting temperature would be surprising , and as will be seen later when other cluster properties are examined , the two clusters do in fact have similar melting temperatures .
also , the potential energy curves shown in figure [ fig : potential ] and the @xmath60 surface shown in figure [ fig : del2e_nt ] both suggest that the melting temperature for these clusters is about @xmath73 , and thus intermediate to the two heat capacity peak temperatures . there is much evidence of systematic error due to quasiergodicity in the metropolis results , especially for sizes @xmath51 to 38 , where the metropolis heat capacity curves had lower - temperature peaks that were either missing ( @xmath74 to 34 ) or poorly formed ( @xmath72 to 38 ) . for the other cluster sizes , the metropolis heat capacity curves showed very good agreement with the parallel tempering results , except for temperatures near the heat capacity peak , where the metropolis results for @xmath33 total passes per temperature point were substantially lower than the parallel tempering results , and even the metropolis results for @xmath34 total passes were typically slightly lower .
the fact that parallel tempering simulations of @xmath33 passes could provide significantly better results than similar metropolis simulations of @xmath34 passes , for not much more computational overhead than metropolis simulations of @xmath33 passes , clearly shows the superiority of parallel tempering over the metropolis method .
there have been several previous studies of 55-atom cluster heat capacities done.@xcite the study most directly comparable to the parallel tempering results i obtained for @xmath11 is a j - walking study of ar@xmath10 done by lpez.@xcite he obtained peak parameters of @xmath75 and @xmath76 , compared to my parallel tempering results of 17.45 and 0.294 , respectively .
lpez used the same constraining volume radius as i did for @xmath11 , @xmath42 , but his j - walker distributions consisted of only @xmath77 configurations , which are considerably fewer than the @xmath29 configurations i used for my j - walking studies of cluster sizes @xmath1 to 38 .
given the poor results i obtained with j - walking using larger distributions for smaller cluster sizes , i suspect that the discrepancy in the heat capacity peak parameters is due to lpez having used insufficiently representative j - walker distributions in his study . the heat capacity curve obtained from a metropolis simulation of @xmath34 total passes per temperature that is shown in figure [ fig : cv3 ] had peak parameters of 17.00 and 0.294 , respectively , which are close to the parallel tempering results .
the small discrepancy between these metropolis results and the parallel tempering results is similar to the discrepancies obtained for the other clusters of similar size .
labastie and whetten obtained peak parameters of 15.4 and 0.3 for ar@xmath10 using histogram methods , but their smaller peak height was most likely due to their constraining radius being considerably smaller at @xmath78 .
doye and wales obtained peak parameters of 17.4 and 0.298 for ar@xmath10 using the superposition method,@xcite which is in very good agreement with my results .
the rms bond length fluctuation @xmath79 is another measure that has been used often in cluster simulations to estimate the melting temperature.@xcite according to the lindemann criterion , melting occurs in bulk matter when @xmath79 is about 10% or greater .
however , much care must be exercised in interpreting rms bond length fluctuations obtained from cluster simulations .
first , as calvo and spiegelmann have pointed out , the lindemann criterion was originally proposed for bulk systems , implying that it does not necessarily extend to finite clusters , and the choice of 10% for the critical value is somewhat arbitrary.@xcite second , @xmath27 is dependent on the simulation walk length , with the sharp rise in the curve generally occurring at lower temperatures as the walk length is increased.@xcite for cluster simulations of very long length , a value of @xmath80 can be obtained at temperatures where the cluster spends most of its time in some solid - like form , with only occasional isomerizations to other solid - like forms , rather than undergoing the ready diffusion typical of liquid - like behavior . because @xmath79 is obtained from averages over entire metropolis walks or molecular dynamics trajectories , it can not be estimated from simulations where configuration scrambling occurs , such as j - walking and parallel tempering , and so quasiergodicity issues further complicate matters .
using very long walk - lengths to reduce systematic errors arising from quasiergodicity can result in @xmath27 curves whose transitions occur well before the melting phase - change . despite these caveats ,
rms bond length fluctuation curves have qualitative validity and can provide useful insights , especially in conjunction with other cluster properties .
curves of @xmath27 for each cluster size are shown in figures [ fig : rms1 ] , [ fig : rms2 ] and [ fig : rms3 ] for averaged walk - lengths of @xmath40 and @xmath33 passes per temperature , while figure [ fig : rmsn ] shows the @xmath4-dependence of the temperatures corresponding to the lindemann - like threshold @xmath81 .
the threshold temperatures were estimated by smoothing and interpolating the @xmath27 curves , and then searching for the interpolated temperature corresponding to @xmath81 .
i chose to set the threshold slightly higher than the usual lindemann criterion of @xmath82 , since some of the @xmath27 curves had too much noise near @xmath82 to obtain a reliable estimate of the temperature corresponding to that threshold , and because the threshold is somewhat arbitrary anyway ; the value @xmath81 was also typically near the inflection point of the smoothed curves transition regions .
the @xmath27 curves shown in figures [ fig : rms1 ] to [ fig : rms3 ] are mostly similar , but there are some interesting differences . generally , the steepness of the rise in @xmath79 in the transition region increases with cluster size , with the magic number sizes @xmath83 and 55 having especially sharp rises .
evidence of quasiergodicity can be clearly seen in some curves : the @xmath27 curves for @xmath51 each have a noisy shoulder just before the transition region , while the curves for @xmath84 each have a small peak ; the curves for @xmath52 show the cluster to be locked in its global minimum for all of the lower - temperature metropolis simulations of length @xmath33 passes or fewer , until finally escaping at @xmath85 k. evidence of pre - melting in the anti - mackay overlayer of the third mackay shell can be seen in sizes @xmath53 to 60 by the emergence and evolution of a shoulder in the lower - temperature part of the @xmath27 curves . also indicated on the temperature axes in figures [ fig : rms1 ] to [ fig : rms3 ] are the corresponding heat capacity peak temperatures obtained from the parallel tempering simulations ( likewise , figures [ fig : cv1 ] to [ fig : cv3 ] have the corresponding temperatures associated with the lindemann - like threshold of @xmath81 indicated on their axes ) .
these clearly show that the heat capacity peaks do not necessarily correspond to a cluster phase change . for cluster sizes @xmath1 to 29 ,
the heat capacity peak temperatures are well within the liquid - like region , according to the @xmath27 curves , while for cluster sizes @xmath3 to 36 , where two heat capacity peaks are evident , the higher - temperature heat capacity peak is also clearly in the liquid - like region .
moreover , these temperatures are high enough that quasiergodicity in these regions of the @xmath27 curves is not an issue .
likewise , most of the lower - temperature heat capacity peaks are found to be in the solid - like region of the @xmath27 curves , well before the transition region ( although ergodicity issues regarding the @xmath27 curves in these regions could be involved ) . only for sizes @xmath72 and @xmath5 , are the lower - temperature heat capacity peak temperatures similar to the @xmath27 transition regions .
for the remaining cluster sizes shown in figures [ fig : rms2 ] and [ fig : rms3 ] , however , the heat capacity peak temperatures correspond quite closely to the end of the @xmath27 transition regions , where the curves quickly change slope to mark the start of the liquid - like region .
magic number behavior in the rms bond length fluctuations is evident in figure [ fig : rmsn ] , especially for the threshold temperature values obtained from the longer metropolis simulations of @xmath33 passes .
the threshold temperatures are significantly higher for the magic number sizes @xmath86 and 55 , relative to their immediate neighbors .
these are the same magic numbers seen in figure [ fig : energyn ] for @xmath59 and for the energy gaps between the lowest - energy and next lowest - energy isomers , except for @xmath87
. the @xmath79 data also account for the absence of magic number behavior in the heat capacities for sizes @xmath66 and 32 .
since the large , broad , higher - temperature heat capacity peaks seen slowly diminishing in figures [ fig : energyn ] and [ fig : cv1 ] are located well past the cluster solid - liquid phase - change regions , whatever magic number effects might exist for @xmath66 and 32 are masked by the systematic evolution of the heat capacity peak itself with increasing cluster size .
classical cluster heat capacities are inherently a measure of the width of the potential energy distributions .
typically , for temperatures near the cluster melting temperature , the distributions are much wider than those at lower temperatures ( which span mostly lower - energy , solid - like forms ) and those at higher temperatures ( which span mostly higher - energy , liquid - like forms ) , since these distributions span both types of forms .
however , the distribution width in the liquid region also typically increases with temperature , since the number of accessible configurations also increases .
thus , depending on the energy distribution of the higher - energy isomer potential minima , the inverse dependence of the heat capacity on temperature in eq .
( [ eq : cv ] ) could result in a peak occurring in the liquid - like region of the curve that would have little relation to the cluster s solid - liquid phase change .
the threshold temperatures corresponding to @xmath81 in figure [ fig : rmsn ] also imply that the cluster melting temperature generally decreases as the cluster size increases from @xmath1 , reaching a minimum near @xmath2 , then generally increases with cluster size , reaching a maximum for @xmath11 , and decreases thereafter ; magic number effects superimposed on these general trends account for the local variations .
these general trends are consistent with those inferred from the @xmath60 surface in figure [ fig : del2e_nt ] and the potential energy curves in figure [ fig : potential ] .
casero and soler calculated melting temperatures for lj clusters of size @xmath88 to 34 in the microcannonical ensemble , defining the melting temperature to be the temperature where the free energies of the solid and liquid clusters become equal.@xcite their results indicate a generally increasing trend with cluster size , with reduced melting temperatures for sizes @xmath1 to 34 all above @xmath89 .
these results are very similar to the temperatures seen for higher - temperature heat capacity peak temperatures in figure [ fig : magic ] , and thus are contrary to the melting temperatures implied by the rms bond length fluctuation data shown in figure [ fig : rmsn ] . since the melting temperatures deduced from the @xmath79 data are in good agreement with the melting temperatures obtained from the heat capacity peak temperatures for cluster sizes greater than about @xmath72 , and these temperatures are well below @xmath89 for @xmath90 , it is likely that the reduced melting temperatures for sizes @xmath1 to 34 are also well below @xmath89 .
the temperature dependence on the distribution of quenched isomers is shown in figure [ fig : quench_magic ] for those ar clusters whose heat capacities showed magic number behavior , and in figures [ fig : quench1 ] to [ fig : quench3 ] for some selected non - magic number ar clusters .
each plot shows the percentage of quenches to each of the three lowest - energy isomers , as well as the total percentage of quenches to the remaining higher - energy isomers ; the data for all of these plots were all obtained from parallel tempering simulations .
much of the behavior exhibited by these curves can be rationalized in terms of the energy spacings and densities of the lower - energy isomers shown in figure [ fig : energylevels ] .
the plots for the magic number sizes are all very similar in form , with quenches in the lower - temperature , solid - like region represented solely by the lowest - energy isomer , and quenches to the other isomers only appearing at temperatures near the phase - change region , and higher .
this is a consequence of the large gap between the lowest - energy and next lowest - energy isomers that is typical of magic number clusters .
the only significant differences between the different cluster sizes is the temperature where the fraction of quenches to the lowest - energy isomer drops off substantially , which increases with cluster size , and can be seen to be very similar to the heat capacity peak temperature .
the plots for some non - magic cluster sizes shown in figure [ fig : quench1 ] for @xmath91 are considerably different .
the curves for @xmath1 have a similar form to the magic number curves in figure [ fig : quench_magic ] , but both the decrease in the fraction of quenches to the lowest - energy isomer and the increase in the fraction of quenches to the higher - energy isomers are much more gradual .
this gradual transition is consistent with the more gradual transition for the @xmath27 curve , seen in figure [ fig : rms1 ] .
also , the transition region occurs well below the heat capacity peak temperature ( but near the transition region indicated by @xmath27 ) .
ar@xmath92 has two very closely spaced lowest - energy isomers that are well below its next - higher energy isomers , and most of the lower - temperature quenches are to the second lowest - energy isomer .
the temperature where the fraction of quenches to the second lowest - energy isomer exceeds the fraction to the lowest - energy isomer is similar to the peak temperature of the very small lower - temperature heat capacity peak .
again , the temperature region where the fraction of quenches to the higher - energy isomers increases substantially corresponds well to the transition region in the @xmath27 curves , and is well below the higher - temperature heat capacity peak temperature .
the quench curves for @xmath51 and 32 are similar to the @xmath93 curve , with the fraction of quenches to the lowest - energy isomer dropping rapidly at very low temperatures corresponding closely to the small heat capacity peak temperatures , and the fraction of quenches to higher - energy isomers gradually increasing at temperatures well below the peak temperature of the larger , higher - temperature heat capacity peaks . for ar@xmath94 ,
the density of lower - energy isomers is so great that quenches to higher - energy isomers dominate at temperatures even below the transition region for the @xmath27 curves shown in figure [ fig : rms1 ] .
the lower - temperature shoulders appearing in the @xmath51 @xmath27 curves and the behavior of the corresponding quench curves are indicative of pre - melting and an extended phase - change region .
ar@xmath46 is a magic number according to such structural properties as @xmath95 and @xmath59 , but not according to its heat capacity peak parameters .
its quench curves are also different than the typical magic - number quench curves seen in figure [ fig : quench_magic ] .
the @xmath84 @xmath27 curves show unusual peaks in the solid region , just before their sharp rise .
this region just follows the lower - temperature heat capacity peak , and according to the quench results , almost all of the quenches are to the second lowest - energy isomer .
this indicates that the peaks in the @xmath27 curves are a result of quasiergodicity ( the metropolis heat capacity curves in figure [ fig : cv1 ] also show small , spurious peaks in this region ) .
the quench curves for @xmath96 and 34 indicate easy access to the other low - lying isomers at low temperatures near the small heat capacity peak temperatures .
figure [ fig : quench2 ] shows quench results for some non - magic number clusters for @xmath2 to 48 .
ar@xmath97 , like ar@xmath94 and ar@xmath98 , has an especially high density of low - lying isomers , and its quench curves also show a substantial fraction of quenches to these isomers at temperatures corresponding to the heat capacity peak temperature . for @xmath52 , the temperature region where the fraction of quenches to the lowest - energy fcc isomer drops off and the fraction of quenches to the lowest - energy icosahedral - like isomer ( the second lowest - energy isomer ) increases corresponds to the region just before the heat - capacity peak where a shoulder can be seen in figure [ fig : cv2 ] , which is consistent with other studies of this cluster.@xcite
the quench curves for @xmath99 and 48 are similar to those for @xmath2 , since these clusters all have accessible lower - energy isomers that are only slightly higher than the lowest - energy isomer , but the curves for @xmath100 are similar to the curves for the magic number clusters in figure [ fig : quench_magic ] , indicating that its low - lying isomers are not very accessible .
the isomer energy levels for @xmath101 are similar to those for @xmath102 and 48 , and the corresponding quench curves shown in figure [ fig : quench3 ] are likewise similar to those for @xmath102 and 48 .
ar@xmath103 and ar@xmath104 have some isomers having closely spaced potential energies only slightly above the lowest - energy isomer , and then very few isomers until @xmath105 .
this type of isomer energy distribution is similar to the distributions seen for binary ar - kr clusters , where many mixed isomer configurations had energies similar to the lowest - energy isomer,@xcite and the quench results are similar too , with the low - lying isomer quench curves showing plateaus throughout much of the solid - like temperature region .
however , neither ar@xmath103 nor ar@xmath104 exhibited the small , very low temperature heat capacity peaks that were evident in the binary cluster curves .
ar@xmath106 has two low - energy isomers corresponding to the two unique ways a lone atom can be positioned on a 55-atom mackay icosahedral core , separated by a large energy gap from the other isomers corresponding to rearrangements of the 55-atom core .
the @xmath107 quench curves show substantial fractions of quenches to both isomers throughout much of the solid - like region until the phase - change region , when the higher - energy isomers become accessible . for @xmath108 , there are several lower - energy isomers below @xmath109 that were represented in the quench distributions , but as with @xmath110 and 53 , there was no corresponding lower - temperature heat capacity peak .
both the @xmath108 and 59 quench curves show evidence of premelting , consistent with their corresponding @xmath27 curves in figure [ fig : rms3 ] .
compared to the irregularity with respect to cluster size @xmath4 that characterized the properties of the small clusters of size @xmath111 that i studied previously,@xcite the properties for the clusters sized @xmath112 showed more regularity and consistency .
the lowest - energy isomers for these sizes were all dominated by icosahedral based geometries , except for @xmath52 , whose lowest - energy isomer had fcc geometry ( the lowest - energy icosahedral - like isomer for @xmath52 was only slightly less stable ) . for the icosahedral - like clusters ,
configurations having anti - mackay overlayers were the lowest - energy isomers for @xmath113 , while configurations having mackay overlayers were the lowest - energy isomers for the remaining sizes up to @xmath11 , after which the filling of the third shell began again with anti - mackay overlayers .
correspondingly , the heat capacity peak parameters formed two overlapping sequences as functions of cluster size , with those clusters having anti - mackay overlayer lowest - energy isomers showing a gradual evolution in their heat capacity peaks to higher temperature and smaller size , while those clusters having mackay overlayer lowest - energy isomers had small , lower - temperature peaks that generally shifted to higher temperature and grew in size as @xmath4 increased . for cluster sizes
@xmath114 , the heat capacity peak temperatures did not correspond well to the cluster melting temperature ranges inferred from the temperature dependence of other cluster properties , such as the rms bond length fluctuations and the distributions of quenched isomers .
for each of the cluster sizes from @xmath63 to 60 , though , the melting temperature ranges obtained from the three measures were in qualitatively good agreement . the general trend in the cluster melting temperatures
as a function of the cluster size over the range studied was an initial decrease , reaching a minimum near @xmath63 , then an increase until a maximum at @xmath11 , where the second mackay shell was completed , followed by another decrease as atoms were added to the third mackay shell .
superimposed on this general trend were local maxima corresponding to the magic numbers @xmath61 , 29 , 32 , 36 , 39 , 43 , 46 , 49 and 55 .
the heat capacity peak heights had a magic number series that was a subset of this sequence , with local maxima occuring at @xmath63 , 39 , 43 , 46 , 49 and 55 ; for the other magic number sizes where the heat capacity peaks did not show magic number behavior ( @xmath61 , 29 , and 32 ) , the heat capacity peaks were located well away from the solid - liquid phase - change regions . the magic number sequence for the cluster melting temperatures coincided very well with the magic number sequences obtained for structural properties such as the binding energy difference @xmath95 , the second finite difference of the energy @xmath59 , and the gap between the lowest - energy and next lowest - energy isomers .
the only differences were @xmath115 being a magic number for the @xmath95 sequence instead of @xmath116 , and @xmath87 , which was a magic number for all three structural properties , but was not for either the heat capacity or the rms bond length fluctuation sequences .
this similarity in the magic number sequences is consistent with the widely held view that the thermodynamic behavior of clusters in the phase - change region is dominated by their potential energy surfaces .
the parallel tempering method was found to be especially useful for overcoming quasiergodicity and accurately calculating the heat capacities of these clusters , particularly for the range @xmath3 to 38 , where the metropolis method was hard pressed to provide even qualitatively correct results .
even the j - walking method , which had worked very well for smaller cluster sizes , was unable to determine heat capacities reproducibly for clusters in this range .
given also the relatively small computational overhead for the method compared to metropolis monte carlo , parallel tempering currently stands as the method of choice for cluster monte carlo simulations .
the support of the natural sciences and engineering research council of canada ( nserc ) for part of this research is gratefully acknowledged .
i thank northern digital , inc . for generously providing me the use of their workstations for some of the calculations
reported , and the university of waterloo for the use of their facilities .
i also thank david l. freeman for helpful discussions .
m. r. hoare and p. pal , adv .
* 20 , * 161 ( 1971 ) .
r. s. berry , t. l. beck , h. l. davis and j. jellinek , in _ advances in chemical physics , _ edited by i. prigogine and s. a. rice ( wiley , new york , 1988 ) , vol . 70b , p. 75 .
h. l. davis , j. jellinek , and r. s. berry , 6456 ( 1987 ) .
t. l. beck , j. jellinek , and r. s. berry , 545 ( 1987 ) .
j. e. adams and r. m. stratt , 1332 ( 1990 ) .
j. d. honeycutt and h. c. andersen , j. phys .
* 91 , * 4950 , ( 1987 ) .
j. a. northby , 6166 ( 1987 ) .
d. g. vlachos , l. d. schmidt , and r. aris , 6880 ( 1992 ) ; * 96 , * 6891 ( 1992 ) .
k. doye , d. j. wales , and r. s. berry , , 4234 ( 1995 )
. j. p. k. doye , m. a. miller , and d. j. wales , , 8417 ( 1999 ) .
d. d. frantz , 3747 ( 1995 ) .
r. casero and j. m. soler , 2927 ( 1991 ) .
t. ikeshoji , b. hafskjold , y. hashi , and y. kawazoe , 5126 ( 1996 ) . o. echt , k. sattler , and r. recknagel , 1121 ( 1981 ) .
e. k. parks , b. j. winter , t. d. klots , and s. j. riley , 1882 ( 1991 ) ; e. k. parks , k. p. kearns , and s. .j riley , 10207 ( 1998 ) .
d. prekas , c. lder , and m. velegrakis , 4450 ( 1998 ) .
m. sakurai , k. watanabe , k. sumiyama , and k. suzuki , 235 ( 1999 ) .
d. d. frantz , d. l. freeman , and j. d. doll , 2769 ( 1990 ) .
d. d. frantz , d. l. freeman , and j. d. doll , 5713 ( 1992 ) .
g. e. lopez and d. l. freeman , 1428 ( 1993 ) .
d. d. frantz , 10030 ( 1996 ) ; 1992 ( 1997 ) .
g. e. lpez , 6650 ( 1996 ) .
a. matro and d. l. freeman , and r. q. topper , 8690 ( 1996 ) .
e. curotto , d. l. freeman , and j. d. doll , 1643 ( 1998 ) .
h. xu and b. j. berne , 10299 ( 1999 ) .
c. j. geyer and e. a. thompson , j. am .
* 90 , * 909 ( 1995 ) .
m. r. ghayal and e. curotto , , 4298 ( 2000 ) .
neirotti , f. calvo , d. l. freeman and j. d. doll , 10340 ( 2000 ) ; f. calvo , j. p. neirotti , d. l. freeman and j. d. doll , 10350 ( 2000 ) .
j. p. k. doye , d. j. wales and m. a. miller , , 8143 ( 1998 ) .
j .p . k. doye , m. a. miller and d. j. wales , , 6896 ( 1999 ) .
p.125in*2p0.75inp0.75inp.125in*2p0.75inp0.75 in & & & & & + & & & & & & & & & + 25 & & & [email protected] & [email protected] & 43 & & & [email protected] & [email protected] + 26 & & & [email protected] & [email protected] & 44 & & & [email protected] & [email protected] + 27 & [email protected] & [email protected] & [email protected] & [email protected] & 45 & & & [email protected] & [email protected] + 28 & & & [email protected] & [email protected] & 46 & & & [email protected] & [email protected] + 29 & & & [email protected] & [email protected] & 47 & & & [email protected] & [email protected] + 30 & [email protected] & [email protected] & [email protected] & [email protected] & 48 & & & [email protected] & [email protected] + 31 & [email protected] & [email protected] & [email protected] & [email protected] & 49 & & & [email protected] & [email protected] + 32 & [email protected] & [email protected] & [email protected] & [email protected] & 50 & & & [email protected] & [email protected] + 33 & [email protected] & [email protected] & [email protected] & [email protected] & 51 & & & [email protected] & [email protected] + 34 & [email protected] & [email protected] & [email protected] & [email protected] & 52 & & & [email protected] & [email protected] + 35 & [email protected] & [email protected] & [email protected] & [email protected] & 53 & & & [email protected] & [email protected] + 36 & [email protected] & [email protected] & [email protected] & [email protected] & 54 & & & [email protected] & [email protected] + 37 & & & [email protected] & [email protected] & 55 & & & [email protected] & [email protected] + 38 & & & [email protected] & [email protected] & 56 & & & [email protected] & [email protected] + 39 & & & [email protected] & [email protected] & 57 & & & [email protected] & [email protected] + 40 & & & [email protected] & [email protected] & 58 & [email protected] & [email protected] & [email protected] & [email protected] + 41 & & & [email protected] & [email protected] & 59 & & & [email protected] & [email protected] + 42 & & & [email protected] & [email protected] & 60 & & & [email protected] & [email protected] + | monte carlo methods were used to calculate heat capacities as functions of temperature for classical atomic clusters of aggregate sizes @xmath0 that were bound by pairwise lennard - jones potentials .
the parallel tempering method was used to overcome convergence difficulties due to quasiergodicity in the solid - liquid phase - change regions .
all of the clusters studied had pronounced peaks in their heat capacity curves , most of which corresponded to their solid - liquid phase - change regions .
the heat capacity peak height and location exhibited two general trends as functions of cluster size : for @xmath1 to 36 , the peak temperature slowly increased , while the peak height slowly decreased , disappearing by @xmath2 ; for @xmath3 , a very small secondary peak at very low temperature emerged and quickly increased in size and temperature as @xmath4 increased , becoming the dominant peak by @xmath5 .
superimposed on these general trends were smaller fluctuations in the peak heights that corresponded to `` magic number '' behavior , with local maxima found at @xmath6 and 49 , and the largest peak found at @xmath7 .
these magic numbers were a subset of the magic numbers found for other cluster properties , and can be largely understood in terms of the clusters underlying geometries .
further insights into the melting behavior of these clusters were obtained from quench studies and by examining rms bond length fluctuations . |
within condensed matter physics , spintronics has in itself become a strong field for considerable research , owing to not only its potential applications in electronic technologies but also the many fundamental questions that are raised on the physics of electron spin .
@xcite particularly , the spin - orbit interaction recently has strongly attracted the attentions of theoreticians and experimenters since it opens up the possibility of manipulating electron ( or hole ) spin in non - magnetic materials by electrical means .
@xcite since the theoretical prediction of the spin - hall effect , the application of spintronics has seen considerable advancement .
it was shown that the mott - type skew scattering by impurities would result in separation of opposite spin states via the spin - orbit interaction to the impurity atom .
@xcite this is the extrinsic spin - hall effect .
nevertheless , it has been found that in p - type @xcite ( luttinger model ) and n - type @xcite ( rashba model ) semiconductors , the spin - polarized current ( electron or hole ) can be generated by the intrinsic spin - orbit interaction in non - magnetic structure in the absence of impurity scattering , which is called the intrinsic spin - hall effect ( ishc ) .
the calculation of spin - hall conductivity ( shc ) plays a crucial role in studying the spin - hall effect as it can be in comparison with experimental result .
the extrinsic spin - hall effect was experimentally discovered in the three - dimensional ( 3d ) n - type gaas by optical means via spin accumulation at the edges of a sample @xcite and in two - dimensional ( 2d ) n - type algaas / gaas .
@xcite the magnitude of the experimental value of shc [ @xmath10 in ref .
@xcite agrees with its theoretical value [ @xmath11 .
@xcite however , the sign of the theoretical shc is opposite to the experimental value , and it needs to be further clarified .
@xcite in 2d p - type algaas / gaas @xcite , the experimental value of shc [ @xmath12 also agrees with the theoretical value [ @xmath13 in order of magnitude .
particularly , in ref .
@xcite , the clean limit is considered in the calculation . in 2d
n - type ingan / gan , the strain - dependent intrinsic spin - hall effect detected by optical means is explained in terms of shc in which the strain effect is included .
@xcite in 3d metal pt wire , the large ishc measured electrically throughout the inverse spin - hall effect at room temperature is @xmath14 . @xcite it was theoretically explained in ref .
@xcite on the basis of the huge berry curvature @xcite near the fermi level at the l and x symmetry point in the pt brillouin zone ; the obtained theoretical value of ishc is @xmath15 in the absence of impurity scattering .
most recently , a large spin - hall signal is observed at room temperature in fept / au multi - terminal devices .
@xcite importantly , both the direction of the applied electric field and the strength of the spin - orbit interactions alters the values of the ishc . for the former case , a typical example is the rashba - dresselhaus system .
@xcite when an electric field is applied along the @xmath16 ( @xmath17 $ ] ) ( or @xmath18 @xmath19 $ ] ) direction , we obtain @xmath20 , and these values are equal to the universal constant @xmath8
. however , if an electric field is applied along @xmath21 ( @xmath22 $ ] ) ( or @xmath23 , i.e. , @xmath24 $ ] ) , we obtain @xmath25 , and one of these values has a value higher than @xmath8 .
the later case requires a systematical investigation because the spin - orbit interaction could be very complicated .
for example , it has been proposed that a strained semiconductor results in various k - linear band - splitting .
@xcite nevertheless , we find that strain - induced spin splitting together with the spin - orbit coupling of the host semiconductor can be simplified and expressed in terms of the coefficients of the spin - orbit matrix [ see eq .
( [ hamiltonian - beta ] ) ] . in this study , we focus on generic 2d k - linear spin - orbit coupled systems without impurity scattering , and we systematically investigate the effects of spin - orbit interactions and the direction of the applied electric field on spin - hall current .
we find that the ishc can be calculated analytically and that its unsymmetrical properties can be described using a unified approach .
we show that @xmath3 [ see eq .
( [ detbeta ] ) ] is expressed as the effective coupling of the z - component of spin @xmath4 , and orbital angular momentum @xmath5 .
the decoupling of spin and orbital motion associated with the band - overlapping phenomenon results in the vanishing and sign change of the ishc .
furthermore , by analytically calculating the ishc , we find that the unsymmetrical result of the ishc ( @xmath26 ) is governed by the asymmetric response function @xmath7 [ see eq . ( [ deltabeta ] ) ] , which is the difference in band - splitting in two directions : those of the applied electric field and the spin - hall current .
we find that the direction of the applied electric field alters the magnitude of the asymmetric response function .
consequently , we show that there exists a specific direction of applied electric field such that the asymmetric response function reaches a maximum value . in this case , we show that the ishc also reaches a maximum value in the range @xmath8 to @xmath9 , where @xmath9 is the upper bound value of the ishc . the unsymmetrical result and the maximum asymmetric response function of the ishc can also be deduced from the behavior of the berry curvature at the nearly degenerate area .
the nearly degenerate area refers to the area where the inner and outer bands are very close to each other on the fermi surface .
our present paper is organized as follows . in sec .
[ sec : ishc ] , we define the spin - orbit matrix obtained from the coefficients of the product of momentum and spin . the intrinsic spin - hall conductivity is shown to be proportional to the determinant of spin - orbit matrix @xmath3 .
we use foldy - wouthuysen transformation to show that the effective coupling of the spin @xmath27-component @xmath4 and orbital angular momentum @xmath5 is @xmath28 . in sec .
[ sec : mvshc ] , we analytically calculate the ishc of the generic 2d k - linear spin - orbit coupled system .
the asymmetric response function and the upper bound value of ishc will be discussed . in sec .
[ sec : a - ishc ] , in order to reveal the maximum value of ishc , the direction of applied electric field and its influence on the asymmetric response function is studied . in sec . [
sec : berryc ] , we will show that the unsymmetrical properties of ishc can be deduced from the variation of berry curvature . in sec .
[ sec : esc ] , we discuss the relationship between equilibrium spin current and spin - orbit matrix .
we show that @xmath3 plays the role of color magnetic field strength .
our conclusions are presented in sec .
[ sec : con ] .
the 2d k - linear spin - orbit coupled system hamiltonian in the presence of an applied electric field can be written as @xmath29 where @xmath30 the kinetic energy is @xmath31 and @xmath32 are the pauli spin matrices .
the external potential @xmath33 is @xmath34 .
the generic k - linear spin - orbit coupled 2d systems are related to the spin - orbit matrix @xmath2 , @xmath35 where the coefficients @xmath36 represent the spin - orbit interactions in 2d systems . as an example , consider the rashba system [ @xmath37 @xcite , the pure dresselhaus system [ @xmath38 @xcite and the dirac - type system [ @xmath39 @xcite ; the corresponding spin - orbit matrices for these systems are @xmath40 respectively .
another example is the spin splitting in a bulk strained semiconductor .
@xcite the spin - orbit matrices @xmath41 and @xmath42 denote , respectively , the system with structure inversion asymmetry ( sia ) strain - induced splitting and the system with bulk inversion asymmetry ( bia ) strain - induced splitting .
they are given by @xmath43 where the structure constant @xmath44 and @xmath45 .
@xcite thus , in addition to sia and bulk - inversion - symmetry breaking induced spin - orbit interaction , the strain - induced spin splitting is included in the spin - orbit matrix elements .
accordingly , we do not pose any restrictions on the spin - orbit matrix elements in the following calculations . for calculating ishc , we further rewrite eq .
( [ hamiltonian ] ) in the following form .
@xmath46 where @xmath47 the eigenenergy is @xmath48 for two branches @xmath49 ( @xmath50 for outer band and @xmath51 for inner band ) , where the dispersion term @xmath52 can be written as @xmath53 where @xmath54 the energy dispersion eq .
( [ dis2 ] ) satisfies @xmath55 because the time - reversal symmetry is preserved . for a positive chemical potential ( @xmath56 ) , the fermi momenta for two branches
satisfy the following condition @xmath57 which is the band - splitting at @xmath58 direction on the fermi surface .
the ishc can be evaluated by using the kubo formula @xcite @xmath59 @xmath60 is the spin current - charge current correlation function .
the index @xmath61 represents the direction of applied electric field and @xmath62 the direction of response current .
the conventional definition of spin current is @xmath63 @xcite , and charge current is defined as @xmath64 . from the standard approach
, it can be shown that @xcite @xmath65 where @xmath66 represents the fermi function for two energy branches .
note that the correlation function contains the kinetic term .
next , we focus on spin - hall conductivity in the static case ( @xmath67 ) .
when an electric field is applied in @xmath68 direction , and the spin - hall response in @xmath69 direction is given by @xmath70 substituting eqs .
( [ dxdy ] ) , ( [ dis1 ] ) , ( [ dis2 ] ) , and ( [ diff - fm ] ) into eq.([shc ] ) and using the replacement @xmath71 , after a straightforward calculation , we obtain @xmath72 where @xmath3 stands for the determinant of the spin - orbit matrix @xmath2 @xmath73 equation .
( [ shc-1 ] ) indicates that the spin - hall conductivity vanishes when @xmath74 . to understand the vanishing ishc
, we have to refer to the effective coupling of spin @xmath4 and orbital motion @xmath5 in the presence of an applied electric field . in the following subsection
, we will show that the effective coupling of orbital motion and spin is related to eq .
( [ detbeta ] ) .
we can apply the foldy - wouthuysen transformation @xcite to the hamiltonian eq .
( [ hamiltonian ] ) , and diagonalize the hamiltonian up to some order of @xmath36 . because @xmath3 is order of @xmath75 , the unitary transformation that can diagonalize eq .
( [ hamiltonian ] ) up to second order is given by ( see appendix [ appendix : fw ] ) @xmath76 where @xmath77 and @xmath78 are obtained by using the replacements @xmath79 and @xmath80 in @xmath81 and @xmath82 .
it can be shown that @xmath83=\frac{i\hbar^2}{m}h_{so}.\ ] ] using the unitary transformation eq .
( [ unitary ] ) and eq .
( [ dek ] ) , then eq . ( [ hamiltonian ] ) becomes ( up to the second order of @xmath36 ) @xmath84+v(x , y)+o(\beta^3_{ij } ) .
\end{split}\ ] ] it can be shown that @xmath85=i\sum_{ij}\beta^2_{ij}+\frac{2i}{\hbar}\left(\detbeta\right)\sigma_zl_z,\\\ ] ] where @xmath86 is the orbital angular momentum .
substitute eq .
( [ dhso ] ) into eq .
( [ uhu-1 ] ) , we obtain @xmath87 where @xmath88 equation ( [ uhu - d ] ) shows that the coupling between orbital motion @xmath5 and spin @xmath27 component @xmath89 is proportional to @xmath3 .
therefore , eq . ( [ detbeta ] )
together with eqs .
( [ shc-1 ] ) and ( [ uhu - d ] ) exhibits a discriminant for a non - vanishing spin - hall conductivity : @xmath90 in the rashba - dresselhaus system ( @xmath91 ) , we have @xmath92 , @xmath93 , @xmath94 , @xmath95 , and @xmath96 .
it has been shown that the the vanishing spin - hall conductivity in the rashba - dresselhaus system results from the fact that the orbital motion is decoupled from the spin @xmath27-component when @xmath97 .
@xcite on the other hand , band degeneracy occurs when @xmath98 , namely , the inner band and outer band overlap for some vale @xmath99 .
the solution @xmath99 is given by @xmath100 if @xmath101 , the term @xmath102 is a complex number and the angle @xmath99 does not exist .
the angle @xmath99 exists only when @xmath74
. the degeneracy could be open upon tuning the spin - orbit interactions such that @xmath101 .
therefore , the decoupling of the spin @xmath4 and orbital motion @xmath5 always accompanies the band - overlapping phenomenon .
the decoupling of spin and orbital motion results in a sign change of the ishc and the band - overlapping phenomenon .
in order to evaluate the integral in eq .
( [ shc-1 ] ) , we transform the integral to the contour integral in a complex plane .
if @xmath27 is defined as @xmath103 , the integral becomes a line integral along a closed loop with unit radius .
the function @xmath104 can be rewritten as @xmath105 where @xmath106 symbolizes the complex conjugate and @xmath107 the integral in eq .
( [ shc-1 ] ) can be evaluated by calculating the residue inside the unit circle @xmath108 .
the conditions for the poles appearing in the unit circle indicate the boundary of change of ishc in sign . by using the standard residue theorem @xcite ,
the result is derived as @xmath109,\ ] ] where @xmath110 represents the real part of a complex number .
@xmath111 ( @xmath112 ) is taken from the relative maximum ( minimum ) value of ( @xmath113 , @xmath114 ) .
that is , if @xmath115 then @xmath116 and @xmath117 , and vice versa . equation ( [ shc-2 ] ) can be further simplified . using eq .
( [ lambdas ] ) , we find that @xmath118 ) , it cancels @xmath3 appearing in eq .
( [ shc-1 ] ) . combining eq .
( [ shc-1 ] ) together with ( [ shc-2 ] ) and ( [ shc-3 ] ) , we have @xmath119&;\detbeta\neq0\\ \\ 0&;\detbeta=0\\ \end{array}\right.,\ ] ] where @xmath120 is the sign function .
we have @xmath121 and @xmath122 . the real part of @xmath123 in eq .
( [ shc - result ] ) can be written in terms of coefficients of spin - orbit matrix , @xmath124 note that in eq .
( [ relambda ] ) , there is an absolute value of @xmath3 .
the ishc generally depends on the spin - orbit interaction [ eq .
( [ relambda ] ) ] , and it is not necessarily a universal constant .
we note that the denominator of eq .
( [ relambda ] ) is always positive . nevertheless , the numerator of eq .
( [ relambda ] ) can be either positive or negative .
for convenience in the following discussion , we define the asymmetric response function @xmath7 as @xmath125 we find that the asymmetric response function involves two quantities : @xmath126 is the band - splitting along the direction of the spin - hall response and @xmath127 is the band - splitting along the direction of the applied electric field [ see eq . ( [ diff - fm ] ) ] .
the asymmetric response function is the difference of two specific band - splittings . for @xmath128 ,
the ishc is less than @xmath8 .
therefore , the spin - hall conductivity has an upper bound in magnitude , @xmath129 ) is valid only when @xmath130 in coordinate system @xmath131 .
if @xmath69 axis is along [ 100 ] direction and @xmath68 axis is along [ 010 ] direction , then some spin - orbit coupled systems would satisfy this condition , for example , the pure rashba , the pure dresselhaus , and the rashba - dresselhaus systems .
this result is in agreement with the previous theoretical results @xcite .
interestingly , we find that if @xmath132 , then @xmath133 is negative and the spin - hall conductivity satisfies @xmath134 the ishc still has an upper bounded value @xmath9 ; however , it can exceed the value @xmath8 .
the three conditions are summarized in fig .
[ fig1 ] , where we define @xmath135 and it has been shown that @xmath136 .
the spin - hall conductivity @xmath137 for @xmath138 , @xmath139 for @xmath140 and @xmath141 for @xmath132 .
with @xmath142 for three conditions @xmath138 , @xmath140 , and @xmath132 , where @xmath143 is defined as @xmath144 and @xmath136 . , width=226,height=170 ] when an electric field is applied in the @xmath69 direction , the spin - hall response in the @xmath68 direction is given by @xmath145 the integration in eq .
( [ shcyx ] ) can also be calculated analytically as follows : @xmath146&;\detbeta\neq0\\ \\
0&;\detbeta=0\\ \end{array}\right.\ ] ] unlike the @xmath0 , we have @xmath147 we find that when @xmath148 is larger than @xmath8 , @xmath149 is less than @xmath8 and vice versa . in comparison with @xmath0 [ eq .
( [ shc - result ] ) ] , we find that @xmath0 is in general not equal to @xmath150 in the k - linear system . the symmetrical result ( @xmath20 ) is obtained because the electric field is applied in a direction such that the asymmetric response function vanishes .
both the pure rashba and the pure dresselhaus systems exhibit circular energy dispersion , and the asymmetric response function always vanishes regardless of the direction of the applied electric field . in the rashba - dresselhaus system ,
if the electric field is applied along the direction of @xmath17 $ ] and the spin - hall response occurs along @xmath19 $ ] , the band splitting along the direction of the applied electric field is the same as that along the spin - hall response direction , and thus , the asymmetric response function vanishes .
however , a small change in the direction of the applied electric field would result in a non - vanishing asymmetric response function in the rashba - dresselhaus system .
the influence of the applied electric field on the asymmetric response function is discussed in the next section .
the direction of the applied electric field plays an important role in determining whether the system has an non - vanishing asymmetric response function .
we study the asymmetric response function by rotating the coordinate system from @xmath131 to @xmath151 .
consider counterclockwise rotation of the system along the along @xmath27 axis by an angle @xmath152 ; the relationship between @xmath131 and @xmath151 is given by @xmath153 and @xmath154 .
the term @xmath155 represents the matrix element of the spin - orbit matrix in the new coordinate , and they are given by @xmath156 , @xmath157 , @xmath158 , and @xmath159 .
it can be shown that the value of @xmath3 is independent of the choice of coordinates , i.e. , @xmath160 .
interestingly , it can also be shown that @xmath161 , namely , @xmath162 is also independent of the choice of coordinates . in the coordinate system @xmath151 ,
the term @xmath163 indicates that the electric field is applied along the @xmath164 direction ( @xmath165 ) and the corresponding spin - hall response @xmath166 is along the @xmath167 direction ( @xmath168 ) , where the angle @xmath169 is measured from the positive axis of @xmath167 . on the other hand
, the term @xmath170 indicates that the electric field is applied along the @xmath167 direction ( @xmath168 ) and the corresponding spin - hall response @xmath171 is obtained along the @xmath164 direction ( @xmath165 ) .
therefore , in the new coordinate system , eqs .
( [ shc - result ] ) and ( [ shcyx_2 ] ) are still valid and can be written as @xmath172,\\ & \sigma^z_{y'x'}=-\mathrm{sgn}(\detbeta)\frac{e}{8\pi}\left[1+\mathrm{re}\left(\frac{\lambda'_<}{\lambda'_>}\right)\right ] , \end{split}\ ] ] where @xmath173 the energy dispersion in the new coordinate system is @xmath174 .
equation ( [ shc - newc2 ] ) indicates that the variation in the ishc is altered only by the difference in two band - splittings : the band - splitting along the applied electric field direction and the band - splitting along the spin - hall response direction .
it can be shown that in the new coordinate system , @xmath175 and @xmath176 can be written as @xmath177 where @xmath178 and @xmath179 .
therefore , in general , when @xmath180 , @xmath175 is not equal to @xmath176 , even if @xmath181 .
a small rotation of the direction of the applied electric field would lead to a non - vanishing asymmetric response function . according to eqs .
( [ shc - newc ] ) and ( [ shc - newc2 ] ) , in order to enhance @xmath163 , i.e. , @xmath182 , the band - splitting must satisfy the condition @xmath183 .
this means that the electric field must be applied along the direction with the larger band - splitting in comparison with that in the direction of the spin - hall response . on the other hand ,
the corresponding vaalue of @xmath170 is less than @xmath8 .
conversely , if we want to enhance @xmath170 , i.e. @xmath184 , then we must have @xmath185 .
this means that the electric field must be applied along the direction with larger band - splitting in comparison with that in the direction of the spin - hall response .
therefore , we conclude that in order to obtain the ishc @xmath186 with @xmath187 ( @xmath186 can be @xmath188 or @xmath189 ) , the band splitting along the direction of the applied electric field must be larger than that along the direction of the spin - hall response . as indicated in sec .
[ sec : a - ishc ] , when @xmath187 , @xmath186 still has an upper bound value of @xmath9 .
this means that we can further enhance @xmath186 by finding the maximum value of the asymmetric response function .
in fact , it can be shown that ( see appendix [ appendix : mnbs ] ) when all the strengths of the spin - orbit interactions are fixed , the maximum value of @xmath190 exists for a specific direction @xmath191 [ see eq .
( [ thetam ] ) ] .
this also implies that the direction of the largest band - splitting is always perpendicular to that of the smallest band - splitting on the fermi surface .
furthermore , the existence of the maximum value @xmath190 on the fermi surface provides us a method to obtain the maximum ishc with respect to these fixed values of @xmath36 . in the next section , we explain how the enhanced spin - hall response is the manifestation of berry curvature at the nearly degenerate area .
in this section , we analyze the berry curvature in a system with fixed spin - orbit interactions , and for convenience , the direction of the electric field is fixed while we rotate the system ( see fig . [ figrotation ] ) . the spin - hall conductivity @xmath0 [ eq .
( [ shc ] ) ] can be written in terms of the berry curvature as @xmath192 and it can be shown that @xmath193 the energy dispersion @xmath194 is given by @xmath195 , where @xmath196 , @xmath197 , and @xmath198 . .
the band splitting @xmath199 is defined as @xmath200 .
the circular area ( gray ) represents the nearly degenerate area .
, width=302,height=132 ] we note that the berry curvature of outer band ( @xmath201 ) is opposite to that of the inner band ( @xmath202 ) in sign .
therefore , when both bands are occupied ( @xmath56 ) , the only contribution to the spin - hall conductivity is the berry curvature of the outer band .
namely , at a fixed value of @xmath58 , the berry curvature of the inner band cancels that of the outer band at every @xmath203 point with @xmath204 .
equation ( [ shc_berry ] ) becomes @xmath205 on the other hand , for @xmath206 , we have @xmath207 where @xmath208
we plot the variation in the berry curvatures @xmath209 and @xmath210 along the path @xmath211 .
the direction of the electric field is fixed along @xmath68 for obtaining @xmath212 or along @xmath69 for obtaining @xmath213 .
the peak value of @xmath212 ( @xmath213 ) is denoted as @xmath214 ( @xmath215 ) .
the peak value refers to the value of the berry curvature at the nearly degenerate area ( see fig .
[ figbc ] ) .
the ishcs are @xmath216 and @xmath217 , where @xmath218/(\sum_{ij}\beta_{ij}^2 + 2|\detbeta|)$ ] .
we select a system with a non - spherical energy dispersion ( the nearly degenerate area exists ) and @xmath219 .
we use the following coefficients of the spin - orbit matrix : @xmath220 ev nm , @xmath221 ev nm , @xmath222 ev nm , @xmath223 ev nm . the fermi energy is 2.67 mev .
the particle mass is 0.08 in units of the bare electron mass .
the energy dispersion is non - spherical , as shown in the left - hand side of fig .
[ figbc](a ) .
the thick line represents the direction with the largest band - splitting on the fermi surface . by using eq .
( [ thetam ] ) , we have @xmath224 and @xmath225 .
the asymmetric response function vanishes at @xmath226 , and it can be obtained by using @xmath227 .
the result is @xmath228 and @xmath229 .
( ev cm)@xmath230 vs @xmath152 .
( b ) variation of the peak values of the berry curvatures ( @xmath214 and @xmath215 ) with @xmath152 .
the berry curvatures have units of @xmath231 . the peak values @xmath214
( @xmath232 ) and @xmath215 ( @xmath233 ) correspond to fig .
[ figbc](a ) to ( h ) .
, width=302,height=415 ] the variations of asymmetric response function and the peak values of the berry curvatures with @xmath152 are shown in fig .
[ figpeakbc ] . when the system is not rotated ( @xmath234 ) , the band splitting at @xmath235 is less than that at @xmath236 [ fig .
[ figpeakbc](a ) ] , and we have @xmath237 .
we find that @xmath214 is larger than @xmath215 as shown in fig .
[ figpeakbc](b ) .
we now rotate the system with an angle @xmath238 such that the band splitting at @xmath235 is equal to that at @xmath236 , resulting a vanishing asymmetric response function , i.e. , @xmath239 [ fig .
[ figpeakbc](a ) ] .
this results in @xmath240 . in fig .
[ figpeakbc](b ) , @xmath214 is equal to @xmath215 when @xmath238 .
if we further rotate the system such that the band splitting at @xmath235 is larger than that at @xmath236 ( @xmath241 ) , we obtain @xmath242 .
the corresponding @xmath215 is now larger than @xmath214 as can be seen in fig .
[ figpeakbc](b ) . if we rotate the system by @xmath243 such that the largest band splitting is now located along the @xmath69 direction , the magnitude of the asymmetric response function in this case reaches a maximum value as shown in fig .
[ figpeakbc](a ) .
we still have @xmath242 , but @xmath149 would be very close to @xmath9 , and @xmath244 is less than @xmath8 .
we find that @xmath215 is not only considerably larger than that of @xmath214 , but is also the maximum value in comparison with the other peak values of the berry curvatures .
when @xmath245 , the asymmetric response function vanishes . in this case , we obtain @xmath246 and @xmath247 .
when the maximum peak value of the berry curvature @xmath214 ( @xmath215 ) is obtained at @xmath248 ( @xmath243 ) , the magnitude of the asymmetric response function reaches a maximum value .
we have @xmath249 , but @xmath250 would be very close to @xmath9 , and @xmath251 is less than @xmath8 .
we select some specific angles [ @xmath252 shown in fig .
[ figpeakbc](a ) and ( b ) ] and plot the corresponding berry curvatures along the path @xmath253 and the orientations of the energy dispersions ( see fig . [ figbc ] ) . as shown in fig .
[ figbc ] , both the berry curvatures ( @xmath212 and @xmath213 ) have significant values ( peak values ) in the nearly degenerate area . when the system is rotated , both the positions of nearly degenerate area and @xmath214 and @xmath215 change together .
furthermore , @xmath215 increases and @xmath214 decreases as seen from fig .
[ figbc](a ) to ( d ) . on the other hand , in fig .
[ figbc](e ) to ( h ) , @xmath215 decreases and @xmath214 increases [ see also fig .
[ figpeakbc](b ) ] . and
the corresponding variation in the berry curvature @xmath212 ( thick line ) and @xmath213 ( broken line ) along the path @xmath253 , @xmath254 . both the berry curvatures ( @xmath212 and @xmath213 ) have units of @xmath231 .
( a)@xmath255 , ( b)@xmath256 , ( c)@xmath257 , ( d)@xmath258 , ( e)@xmath259 , ( f)@xmath260 , ( g)@xmath261 , ( h)@xmath262.,width=604,height=529 ] the variation in the ishcs ( @xmath148 and @xmath149 ) corresponding to the berry curvature variations in fig . [ figbc](a ) to ( h ) is shown in fig .
[ figshcvsn ] , where @xmath216 , @xmath217 and @xmath218/(\sum_{ij}\beta_{ij}^2 + 2|\detbeta|)$ ] .
the ishc @xmath148 decreases in fig .
[ figshcvsn](a ) to ( d ) and then increases in fig .
[ figshcvsn](e ) to ( h ) .
the ishc @xmath149 increases in fig .
[ figshcvsn](a ) to ( d ) and subsequently decreases in fig .
[ figshcvsn](e ) to ( h ) . in the cases shown in fig .
[ figshcvsn](b ) and ( f ) , @xmath263 and both the berry curvatures have the same peak values .
the behavior of the berry curvature at the nearly degenerate area is in agreement with our conclusions in sec .
[ sec : a - ishc ] .
( @xmath232 ) and @xmath206 ( @xmath233 ) corresponding to different rotation angles @xmath152 corresponding to fig .
[ figbc](a ) to ( h ) and @xmath252 in fig .
[ figpeakbc].,width=264,height=377 ] in the case of spherical energy dispersion ( there is no nearly degenerate area ) , it can be shown that the berry curvature @xmath212 equivalent to @xmath213 shifted by @xmath264 .
as in the non - spherical case , the berry curvature still exhibits two significant responses along the directions of the spin - hall current , but the shape and peak value of the berry curvature do not change when we rotate the system .
this means that @xmath265 regardless of the orientation of the system .
the magnitude of the asymmetric response function always vanishes in this case , and the ishc is a universal constant @xmath8 . it must be emphasized that the angle @xmath191 enables us to find the maximum value of the asymmetric response function for some fixed @xmath36 .
if we have another set of values @xmath266 , the corresponding maximum value of the asymmetric response function is in general different from that with @xmath36 .
the magnitude of the ishc may further be enhanced by tuning the spin - orbit interactions to change the maximum value of the magnitude of the asymmetric response function , but it still has an upper bound of @xmath9 .
the measurable responses caused by the spin - hall effect are very different from those in the present idealized system , which is infinite in size and does not include impurity scattering .
measurable quantities such as spin accumulation , however , depend on boundary conditions .
the conserved spin - current considered in the present paper may correspond to smooth boundaries .
@xcite however , the presence of impurities can drastically affect clean limit results @xcite . in ref .
@xcite , it was shown that impurity scattering does not suppress the spin - hall conductivity in the spatially random rahsba spin - orbit coupled system .
in particular , the su(2 ) formulation on extrinsic mechanism of spin hall conductivity was recently investigated in ref .
@xcite . however , the effects of a finite size and impurity scattering are beyond the scope of the present paper
. hopefully , our interesting predictions of higher intrinsic ishc would stimulate measurements in 2d semiconductor systems in the near future .
we now turn to the discussion on equilibrium spin current in this generic k - linear spin - orbit coupled system . in ref .
@xcite , it was shown that even in thermodynamic equilibrium , spin current for the rashba - dresselhaus system does not vanish in the absence of external fields .
this phenomena has arisen many discussions on the definition of spin current .
@xcite the possibilities to detect the equilibrium spin currents have been studied in refs .
@xcite and @xcite .
we calculate the equilibrium spin current by using conventional definition of spin current . in the case of the positive chemical potential ( @xmath56 ) ,
two branches are populated . in the absence of external fields ,
the equilibrium spin - current is the sum of the in - plane spin currents of the two branches @xmath267 where @xmath268 and @xmath269 is the eigenstate of hamiltonian eq .
( [ hamiltonian ] ) . from eq .
( [ diff - fm ] ) and @xmath270 , a straightforward calculation yields @xmath271 where @xmath272 . for specific systems ,
the result is in agreement with the previous results . in the pure rashba system , where @xmath273 and @xmath274 , we have @xmath275 and @xmath276 . in the pure dresselhaus system , where @xmath277 and @xmath278 , we have @xmath279 and @xmath280 .
it is interesting to note that the equilibrium spin current @xmath281 is related to the inverse of spin - orbit matrix @xmath282 via @xmath283 .
we find that @xmath3 also appears in the expression of equilibrium spin current eq .
( [ escr ] ) .
however , @xmath284 occurs in the third order of @xmath36 . in this sense ,
( [ uhu ] ) fails to explain the physical meaning of @xmath3 in this case ( see appendix [ appendix : fw ] ) .
recently , the equilibrium spin current in k - linear spin - orbit coupled systems is found to be link to the non - abelian su(2 ) gauge theory , where the pauli spin matrix serves as a color index in the gauge field .
@xcite the resulting color current satisfies covariant conservation .
the equilibrium spin current obtained from the covariant conserved color current in the rashba - dresselhaus systems is in agreement with ref .
@xcite . in the following
, we apply this formalism to the generic k - linear systems .
the hamiltonian eq . ( [ hamiltonian ] ) can be written in terms of su(2 ) gauge field @xmath285 , where @xmath286 and @xmath287 .
we have @xmath288 the equilibrium spin current denoted as @xmath289 @xcite is proportional to @xmath290 , where @xmath291 $ ] is the field strength .
the physical meaning of @xmath3 in equilibrium spin current is now clear . the field strength in the su(2 ) non - abelian gauge field
is given by @xcite @xmath292\\ & = -i[a_x^x\sigma_x+a_x^y\sigma_y , a_y^x\sigma_x+a_y^y\sigma_y]\\ & = -i\left(a_x^xa_y^y[\sigma_x,\sigma_y]+a_x^ya_y^x[\sigma_y,\sigma_x]\right)\\ & = 2\sigma_z\frac{m^2}{\hbar^4}\left(\detbeta\right ) , \end{split}\ ] ] where eq .
( [ gf ] ) is used .
we have @xmath293 and @xmath294 .
that is , @xmath3 plays the role of color magnetic field strength @xmath295 .
in conclusion , we have shown that in 2d and k - linear spin - orbit coupled systems , the properties of the intrinsic spin - hall conductivity are governed by two quantities : the effective coupling of spin and orbital motion ( reflected by @xmath3 ) and the asymmetric response function ( @xmath7 ) .
the effective coupling of spin and orbital motion is a discriminant for determining whether or not the spin - hall conductivity vanishes .
the decoupling of spin and orbital motion associated with band - overlapping phenomenon explains the physical origin of the sign change of the intrinsic spin - hall conductivity .
furthermore , the dependence of spin - orbit interaction on the spin - hall effect and the resulting unsymmetrical properties are related to the asymmetric response function , which is determined by the difference in band - splitting along two directions : those of the applied electric field the and spin - hall current .
we varied the orientation of the system and studied the variation in the berry curvature and the corresponding spin - hall response .
we found that maximum intrinsic spin - hall conductivity occurs along the direction of the nearly degenerate area , which also leads to the maximization of the berry curvature and the magnitude of the asymmetric response function .
the position of the nearly degenerate area can be determined analytically .
we also showed that the intrinsic spin - hall conductivity has an upper bound value of @xmath9 .
in addition , we showed that the equilibrium spin current is proportional to @xmath296 , and @xmath3 determines the field strength of the su(2 ) non - abelian gauge field in equilibrium spin current .
we thank the national science council of taiwan for the support under contract no .
nsc 101 - 2112-m-110 - 013-my3 .
a unitary transformation can be generally written as @xmath297 , where the hermitian matrix @xmath298 can be expend in order of @xmath36 , i.e. @xmath299 .
that is , @xmath300 represents the term proportional to the order of @xmath36 , @xmath301 the order of @xmath75 and so on .
follow the approach of foldy - wouthuysen transformation , the hamiltonian @xmath302 under the unitary transformation @xmath297 is given by @xmath303 where @xmath304\\ & h^{(2)}=[is^{(2)},\varepsilon_{\mk}]+[is^{(1)},h_{so}]+\frac{1}{2!}[is^{(1)},[is^{(1)},\epsilon_{\mk}]]\\ \end{split}\ ] ] and @xmath305+[is^{(2)},h_{so}]+\frac{1}{2!}[is^{(1)},[is^{(2)},\varepsilon_{\mk}]]\\ & + \frac{1}{2!}[is^{(2)},[is^{(1)},\varepsilon_{\mk}]]+\frac{1}{2!}[is^{(1)},[is^{(1)},h_{so}]]\\ & + \frac{1}{3!}[is^{(1)},[is^{(1)},[is^{(1)},\varepsilon_{\mk } ] ] ] .
\end{split}\ ] ] because @xmath306 is an odd matrix , we have to find a matrix @xmath300 to cancel this term .
namely , we require @xmath307 and @xmath308=0.\ ] ] on the other hand , we note that @xmath306 is made up of the linear momentum @xmath309 , i.e. , @xmath310 and @xmath311 is proportional to @xmath312 , @xmath300 is obtained by the replacements @xmath79 and @xmath80 in @xmath306 .
take into account the constant @xmath313 , we have @xmath314 substitute eq .
( [ s1 ] ) into @xmath315 , after a straightforward calculation , we find that the last two terms @xmath316+\frac{1}{2!}[is^{(1)},[is^{(1)},\epsilon_{\mk}]]$ ] gives a diagonalized form @xmath317 .
this means that @xmath315 is already diagonalized , and thus , we can choose @xmath318 substitute eqs .
( [ s1 ] ) and ( [ s2 ] ) into @xmath319 [ eq .
( [ h3 ] ) ] , we obtain @xmath320\\ & -\frac{2m^2}{3\hbar^4}\left(\detbeta\right)[\{x , l_z\}(\sigma_y\beta_{xx}-\sigma_x\beta_{yx})\\ & + \{y , l_z\}(\sigma_y\beta_{xy}-\sigma_x\beta_{yy } ) ] .
\end{split}\ ] ] we find that the term in @xmath321 $ ] is composed of odd matrices .
therefore , we must require @xmath322 . in this sense ,
the next diagonalized part is order of @xmath323 .
as shown in sec . [ sec : mvshc ] and sec .
[ sec : a - ishc ] , the value @xmath324 determines whether the ishc is larger than @xmath8 or equal to @xmath8 .
furthermore , if the magnitude @xmath190 increases upon varying @xmath152 , the ishc would approach a maximum value with respect to the fixed value of @xmath36 .
we will show that the maximum value of @xmath190 exists for some angle @xmath191 on the fermi surface for fixed spin - orbit interactions .
first , we show that when @xmath175 reaches the maximum value for some @xmath191 , @xmath176 must reach the minimum at @xmath191 and vice versa . from eq .
( [ gammaarb ] ) , the condition @xmath325 at some @xmath191 gives @xmath326 where @xmath327 and @xmath328 .
it can also be shown that @xmath329 .
on the other hand , we redefine the parameters @xmath330 , @xmath331 and @xmath332 as @xmath333 , @xmath334 , and @xmath335 .
the second derivative gives @xmath336 and @xmath337 . because the second derivative are opposite in sign for @xmath168 and @xmath165
, this implies that when @xmath338 has the maximum ( minimum ) value , @xmath339 has the the minimum ( maximum ) value . in conclusion , @xmath190 must be the maximum value when @xmath340 .
the energy dispersion of the @xmath341 system is of the form @xmath342 .
interestingly , it can be shown that for @xmath340 , the energy dispersion @xmath194 in general has the same form as that of the @xmath341 system .
because we require @xmath343 to have the form @xmath344 , the coefficient @xmath345 must be zero at some @xmath152 , and it is obtained from the equation : @xmath346 .
this indeed gives @xmath340 .
consider the rashba - dresselhaus system ( @xmath91 ) .
if @xmath69 and @xmath68 lie respectively along the @xmath19 $ ] and @xmath17 $ ] directions , then @xmath181 and @xmath347 .
equation ( [ thetam ] ) implies that @xmath348 .
( see fig .
[ figrd ] ) for @xmath349 , the resulting dispersion in the new coordinate system is given by @xmath350 and we have @xmath351 . as a result , in order to obtain a large spin - hall current ( @xmath187 ) , an electric field must be applied along the @xmath164 direction because the nearly degenerate area is located at @xmath352 .
for @xmath353 , we have @xmath354 , and in this case , @xmath355 .
the electric field must be applied along the @xmath356 direction for obtaining a large spin - hall current . as shown in fig .
[ figrd ] , @xmath356 is obtained as the rotation of @xmath167 by @xmath264 , and thus , it is parallel to @xmath164 . in the rashba - dresselhaus system ,
@xmath22 $ ] and @xmath359 $ ] are nonequivalent axes .
the corresponding band - splitting values are @xmath360 and @xmath361 .
we change the coordinate ( @xmath69 , @xmath68 ) to ( @xmath167 , @xmath164 ) such that @xmath167 and @xmath164 are parallel to @xmath22 $ ] and @xmath359 $ ] , respectively . in this case
, we have @xmath349 .
the resulting effective spin - orbit matrix is @xmath362 we have @xmath363 and @xmath364
. the asymmetric response function @xmath365 corresponding to the spin - orbit matrix in eq .
( [ rdasym ] ) is @xmath366 . using eq .
( [ shc - newc2 ] ) , we can show that @xmath367.\ ] ] for @xmath368 , the ishcs are @xmath369 and @xmath370 .
for @xmath371 , the ishcs are @xmath372 and @xmath373 .
@xcite let us suppose that @xmath374 and @xmath375 , and in this case , we have @xmath351 .
the rashba - dresselhaus system has a smaller band - splitting of @xmath360 along the @xmath22 $ ] direction on the fermi surface .
on the other hand , the system has a larger band - splitting [ @xmath361 ] along the @xmath359 $ ] direction .
when the electric field is applied along the @xmath164 direction ( @xmath359 $ ] ) , the spin - hall response along the @xmath167 direction indeed has a value larger than @xmath8 , i.e. , @xmath182 , as shown above .
interestingly , when @xmath377 is very close to @xmath378 in magnitude , eq .
( [ app : lambda ] ) is very close to unity .
the ishc in the rashba - dresselhaus system would transit from @xmath379 to @xmath380 upon tuning the rashba coupling via gate voltage.@xcite rashba coupling and dresselhaus coupling are usually of the same order of magnitude in the gaas quantum well .
@xcite in the ii - vi semiconductor , rashba coupling is larger than dresselhaus coupling , while in the iii - v semiconductor , dresselhaus coupling would be larger than rashba coupling .
@xcite in the narrow - gap compounds , rashba coupling dominates .
@xcite g. a. prinz , science * 282 * , 1660 ( 1998 ) ; s. a. wolf , d. d. awschalom , r. a. buhrman , j. m. daughton , s. von molnar , m. l. roukes , a.y .
chtchelkanova , and d. m. treger , science * 294 * , 1488 ( 2001 ) .
i. zutic , j. fabian , and s. d. sarma , rev .
mod . phys . * 76 * , 323 ( 2004 ) .
j. e. hirsch , phys .
* 83 * , 1834 ( 1999 ) .
s. murakami , n. nagaosa , and s .- c .
zhang , science * 301 * , 1348 ( 2003 ) ; s. murakami , n. nagaosa , s .- c .
zhang , phys .
b * 69 * , 235206 ( 2004 ) .
j. sinova , d. culcer , q. niu , n. a. sinitsyn , t. jungwirth , and a. h. macdonald , phys . rev
. lett . * 92 * , 126603 ( 2004 ) .
y. k. kato , r. c. myers , a. c. gossard , and d. d. awschalom , science * 306 * , 1910 ( 2004 ) .
n. p. stern , s. ghosh , g. xiang , m. zhu , n. samarth and d. d. awschalom , phys .
* 97 * , 126603 ( 2006 ) .
v. sih , r. c. myers , y. k. kato , w. h. lau , a. c. gossard , and d. d. awschalom , nature physics * 1 * , 31 ( 2005 ) . h. engel , e. i. rashba and b. i. halperin , phys .
lett . * 95 * , 166605 ( 2005 ) .
j. wunderlich , et al , phys .
lett . * 94 * , 047207 ( 2005 ) .
b. a. bernevig and s .- c .
zhang , phys .
lett . * 95 * , 016801 ( 2005 ) . h.
j. chang , t .- w .
chen , j. w. chen , w. c. hong , w. c. tsai , y. f. chen and g. y. guo , phys .
lett . * 98 * , 136403 ( 2006 ) .
e. saitoh , m. ueda , h. miyajima and g. tatara , appl .
. lett . * 88 * , 182509 ( 2006 ) ; s. o. valenzuela and m. tinkham , nature ( london ) * 442 * , 176 ( 2006 ) ; k. ando , y. kajiwara , s. takahashi , s. maekawa , k. takemoto , m. takatsu and e. saitoh , phys . rev .
b * 78 * , 014413 ( 2008 )
. t. kimura , y. otani , t. sato , s. takahashi , and s. maekawa , phys .
lett . * 98 * , 156601 ( 2007 ) .
g. y. guo , s. murakami , t .- w .
chen and nagaosa , phys .
* 100 * , 096401 ( 2008 ) .
a. bohm , a. mostafazadeh , h. koizumi , q. niu and j. zwanziger , _ the geometric phase in quantum systems _
( springer , berlin , 2003 ) .
t. seki , y. hasegawa , s. mitani , s. takahashi , h. imamura , s. maekawa , j. nitta and k. takahashi , nature mater .
* 7 * , 125 ( 2008 ) .
shen , phys .
b * 70 * , 081311(r ) ( 2004 ) ; n. a. sinitsyn , e. m. hankiewicz , w. teizer , and j. sinova , phys . rev .
b * 70 * , 081312(r ) ( 2004 ) ; m .- c . chang , phys .
b * 71 * , 085315 ( 2005 ) ; t .- w . chen , c .- m .
huang and g. y. guo , phys .
b * 73 * , 235309 ( 2006 ) . b. a. bernevig and s .- c . zhang , phys .
b * 72 * , 115204 ( 2005 ) ; y. kato , r. c. myers , a. c. gossard , d. d. awschalom , nature ( london ) * 427 * , 50 ( 2004 ) ; y. kato , r. c. myers , a. c. gossard , d. d. awschalom , phys . rev
. lett . * 93 * , 176601 ( 2004 ) .
e. i. rashba , sov .
solid state * 2 * , 1224 ( 1960 ) ; y. a. bychkov and e. i. rashba , j. phys .
c * 17 * , 6039 ( 1984 ) .
g. dresselhaus , phys . rev . * 100 * , 580 ( 1955 ) .
j. w. mcclure , phys .
rev . * 104 * , 666 ( 1956 ) ; d. p. divincenzo and e. j. mele , phys .
b * 29 * , 1685 ( 1984 ) ; v. p.
gusynin and s. g. sharapov , * 95 * , 146801 ( 2005 ) ; 19.k . s. novoselov , a. k. geim , s. v. morozov , d. jiang , y. zhang , s. v. dubonos , i. v. grigorieva , and a. a. firsov , science * 306 * , 666 ( 2004 ) ; k. s. novoselov , a. k. geim , s. v. morozov , d. jiang , m. i. katsnelson , i. v. grigorieva , s. v. dubonos , and a. a. firsov , nature ( london ) * 438 * , 197 ( 2005 ) ; y. zhang , j. p.
small , m. e. s. amori , and p. kim , phys .
lett . * 94 * , 176803 ( 2005 ) ; y .- w .
tan , h. l. stormer , and p. kim , nature ( london ) * 438 * , 201 ( 2005 ) ; y. zhang , y .- w .
tan , h. l. stormer , and p. kim , * 438 * , 201 ( 2005 ) .
n. a. sinitsyn , a. h. macdonald , t. jungwirth , v. k. dugaev and j. sinova , phys .
b * 75 * , 045315 ( 2007 ) . g. e. pikus and a. n. titkov , _ optical orientation_(north - holland , amsterdam . 1984 ) ,
in constructing the spin - orbit matrix @xmath42 , we have neglected the term @xmath381 which exists when @xmath382 . however , since @xmath383 is not influenced by the in - plane electric field , @xmath384 , we can approximately neglect this term .
( see also @xcite ) g. d. mahan , _ many - particle physics _
( kluwer academic / plenum publisher , new york , 2000 ) .
d. culcer , j. sinova , n. a. sinitsyn , t. jungwirth and a. h. macdonald and q. niu , phys .
lett . * 93 * , 046602 ( 2004 ) ; j. shi , p. zhang , d. xiao and q. niu , phys .
lett . * 96 * , 076604 ( 2006 ) .
b. a. bernevig , phys .
b * 71 * , 073201 ( 2005 ) .
j. sinova , s. murakami , s .- q . shen and m
choi , solid state commun . * 138 * , 214 ( 2006 ) ; j. i. inoue , g. e. w. bauer and l. w. molenkamp , phys . rev .
b * 70 * , 041303(r ) ( 2004 ) ; e. g. mishchenko , a. v. shytov and b. i. halperin , phys
lett . * 93 * , 226602 ( 2004 ) ; o. chalaev and d. loss , phys .
b * 71 * , 245318 ( 2005 ) ; o. v. dimitrova , phys . rev .
b * 71 * , 245327 ( 2005 ) .
j. schliemann and d. loss , phys .
b * 69 * , 165315 ( 2004 ) ; a. g. malshukov and k. a. chao , phys . rev . b*71 * , 121308 ( r ) ( 2005 ) ; a. khaetskii , phys . rev .
b * 73 * , 115323 ( 2006 ) ; n. sugimoto , s. onoda . s. murakami and n. nagaosa , phys .
b * 73 * , 113305 ( 2006 ) .
a. a. burkov , alvaro s. nunez and a. h. macdonald , phys .
b * 70 * , 155308 ( 2004 ) .
k. nomura , j. sinova , t. jungwirth , q. niu , and a. h. macdonald , phys .
b * 71 * , 041304 ( 2005 ) ; b. k. nikolic , l. p. zarbo , and s. souma , phys . rev .
b * 72 * , 075361 ( 2005 ) ; l. sheng , d. n. sheng , c. s. ting , phys . rev
. lett . * 94 * , 016602 ( 2005 ) ; e. m. hankiewicz , l. w. molenkamp , t. jungwirth , and j. sinova , phys .
b * 70 * , 241301(r ) ( 2004 ) .
v. k. duagev , m. inglot , e. ya .
sherman , and j. barna@xmath385 , phys .
b * 82 * , 121310(r ) ( 2010 ) .
r. raimondi , p. schwab , c. gorini , and g. vignale , ann .
( berlin ) 524 , 153 ( 2012 ) [ arxiv:1110.5279 ] .
chen , h .- c . hsu and g. y. guo , phys .
b * 80 * , 165302 ( 2009 ) .
e. i. rashba , phys .
b * 68 * , 241315(r ) ( 2003 ) .
e. i. rashba , phys .
b * 70 * , 161201(r ) ( 2004 ) ; a. a. burkov , a. s. n@xmath386ez and a. h. macdonald , phys . rev .
b * 70 * , 155308 ( 2004 ) ; e. b. sonin , phys . rev .
b * 76 * , 033306 ( 2007 ) . q .-
f . sun , x. c. xie and j. wang , phys .
* 98 * , 196801 ( 2007 ) ; phys .
b * 77 * , 035327 ( 2008 ) . v. a. sablikov , a. a. sukhanov and y. ya .
tkach , phys .
b * 78 * , 153302 ( 2008 ) .
drouhin , g. fishman and j .- e .
wegrowe , phys .
b * 83 * , 113307 ( 2011 ) .
e. b. sonin , phys .
lett . * 99 * , 266602 ( 2007 ) .
i. v. tokatly , phys .
* 101 * , 106601 ( 2008 ) .
j. nitta , t. akazaki , h. takayanagi and t. enoki , phys .
rev . lett . * 78 * , 1335 ( 1997 ) .
e. i. rashba , physica e : low - dimensional systems and nanostructures * 34 * , 31 ( 2006 ) ; hyun c. lee and s .- r .
eric yang , phys .
b * 72 * , 245338 ( 2005 ) .
b. jusserand , d. richards , g. allan , c. priester and b. etienne , phys .
b * 51 * , 4707 ( 1995 ) ; w. knap , c. skierbiszewski , a. zduniak , e. litwin - staszewska , d. bertho , f. kobbi , j. l. robert , g. e. pikus , f. g. pikus , s. v. iordanskii , v. mosser , k. zekentes , and yu .
b. lyanda - geller , phys .
b * 53 * , 3912 ( 1996 ) ; j. b. miller , d. m. zumbuhl , c. m. marcus , y. b. lyanda - geller , d. goldhaber - gordon , k. campman , and a. c. gossard , phys . rev
. lett . * 90 * , 076807 ( 2003 ) . | we analytically calculate the intrinsic spin - hall conductivity ( ishc ) ( @xmath0 and @xmath1 ) in a clean , two - dimensional system with generic k - linear spin - orbit interaction .
the coefficients of the product of the momentum and spin components form a spin - orbit matrix @xmath2 .
we find that the determinant of the spin - orbit matrix @xmath3 describes the effective coupling of the spin @xmath4 and orbital motion @xmath5 .
the decoupling of spin and orbital motion results in a sign change of the ishc and the band - overlapping phenomenon .
furthermore , we show that the ishc is in general unsymmetrical ( @xmath6 ) , and it is governed by the asymmetric response function @xmath7 , which is the difference in band - splitting along two directions : those of the applied electric field and the spin - hall current .
the obtained non - vanishing asymmetric response function also implies that the ishc can be larger than @xmath8 , but has an upper bound value of @xmath9 .
we will that the unsymmetrical properties of the ishc can also be deduced from the manifestation of the berry curvature at the nearly degenerate area . on the other hand , by investigating the equilibrium spin current , we find that @xmath3 determines the field strength of the su(2 ) non - abelian gauge field . |
the ability to monitor the gamma - ray sky continuously is extremely important .
the majority of gamma - ray sources are variable , exhibiting flares and transient outbursts on time scales from seconds to years . while there are currently several all - sky monitors in the hard x - ray energy range providing daily light curves , e.g. the all - sky monitor ( asm ) on the _ rossi x - ray timing explorer _ ( _ rxte _ ) from 210 kev @xcite , the gas slit camera ( gsc ) on the _ monitor of all - sky x - ray image _ ( _ maxi _ ) from 1.520 kev @xcite , and the burst alert telescope ( bat ) on _ swift
_ from 1550 kev @xcite , there has not been an all - sky monitor in the low energy gamma - ray region since the burst and transient source experiment ( batse ) instrument on the _ compton gamma - ray observatory _ ( _ cgro _ ) , which was sensitive from 201800 kev @xcite .
the gamma - ray satellite _ fermi _ was launched on 2008 june 11 and commenced science operations on 2008 august 12 . _ fermi _ contains two instruments : the large area telescope ( lat ) , sensitive to gamma rays from @xmath4 mev to @xmath5 gev @xcite ; and the gamma - ray burst monitor ( gbm ) , which is sensitive to x - rays and gamma rays from 8 kev to 40 mev @xcite . with its wide field of view , gbm can be used to provide nearly continuous full - sky coverage in the hard x - ray / soft gamma - ray energy range .
it is the only instrument currently in orbit that can perform all - sky monitoring above 100 kev ( but below the 30 mev threshold of the _ fermi _ lat ) with useable sensitivity .
the _ swift _ /bat sensitivity drops off rapidly above 100 kev , and its energy range effectively ends at 195 kev . _ integral _ , which has a relatively narrow field of view , can not make continuous observations of a large number of individual sources . also , gbm is not limited by solar pointing constraints , as are most other instruments , which allows the monitoring of sources at times during which other instruments can not . the earth occultation technique , used very successfully with batse , has been adapted to gbm to obtain fluxes for an input catalog of known or potential sources . a catalog of 82 sources is currently being monitored and regularly updated .
this catalog contains predominantly galactic x - ray binaries , but also includes the crab , the sun , two magnetars , five active galactic nulcei ( agn ) , and two cataclysmic variables . at energies above 100 kev , six persistent sources ( the crab , cyg x-1 , swift j1753.5 - 0127 , 1e 1740 - 29 , cen a
, grs 1915 + 105 ) and two transient sources ( gx 339 - 4 and xte j1752 - 223 ) have been detected in the first two years of observations . in section [ sec : obs ] , we briefly describe the gbm instrument and outline the earth occultation technique as applied to gbm , in section [ sec : res ] we present the light curves for the eight sources , and in section [ sec : dis ] we discuss the results , the gbm capabilities and future work .
gbm consists of 14 detectors : 12 nai detectors , each 12.7 cm in diameter and 1.27 cm thick ; and two bgo detectors , 12.7 cm in diameter and 12.7 cm thick .
the nai detectors are located on the corners of the spacecraft , with six detectors oriented such that the normals to their faces are perpendicular to the z - axis of the spacecraft ( the lat is pointed in the @xmath6-direction ) , four detectors pointed at @xmath7 from the z - axis , and 2 detectors pointed @xmath8 off the z - axis .
together , these 12 detectors provide nearly uniform coverage of the unocculted sky in the energy range from 8 kev to 1 mev .
the two bgo detectors are located on opposite sides of the spacecraft and view a large part of the sky in the energy range @xmath9 kev to @xmath10 mev .
it should be noted that none of the gbm detectors have direct imaging capability .
single crab occultation step seen in the ctime raw counts data of a single gbm nai detector ( nai 2 ) in the 1225 kev band with 2.048-second time bins .
the crab was @xmath11 from the normal to the detector .
the time window is centered on the calculated occultation time for 100 kev.,width=291 ] known sources of gamma - ray emission can be monitored with non - imaging detectors using the earth occultation technique , as was successfully demonstrated with batse @xcite .
when a source of gamma rays is occulted by the earth , the count rate measured by a detector will drop , producing a step - like feature .
when the source reappears from behind the earth s limb , the count rate will increase , producing another step .
the occultation has a finite transition time due to the effect of absorption in the earth s atmosphere .
the shapes of the individual occultation steps depend on energy and occultation angle .
the occultation angle , @xmath12 , is defined as the elevation angle of the source being occulted with respect to the plane of the _ fermi _ orbit .
the transmission through the atmosphere as a function of time is modeled as @xmath13 $ ] , where @xmath14 is the mass attenuation coefficient of gamma rays at energy @xmath15 in air and @xmath16 is the air mass along the line of sight at a given altitude @xmath17 based on the @xcite .
this requires instantaneous knowledge of the spacecraft position , the direction to the source of interest as seen from the spacecraft , and a model of the earth that includes its oblateness . _
fermi _ was launched into a @xmath18 inclination orbit at an altitude of 555 km .
the orbital period is 96 minutes , and individual occultation steps last for @xmath19 seconds .
figure [ crabstep ] shows a single step due to a crab occultation in the count rate in the 1225 kev band of a single gbm nai detector observing the occultation nearly face - on .
the diameter of the earth as seen from _ fermi _ is @xmath20 , so roughly 30% of the sky is occulted by the earth at any one time .
one complete orbit of the spacecraft allows over 85% of the sky to be observed .
the precession of the orbital plane allows the entire sky to be occulted every @xmath0 days ( half the precession period for the _ fermi _ orbit ) , though the exposure is not uniform .
the earth occultation technique was developed for batse using two separate approaches , one at the nasa marshall space flight center @xcite and the other at the nasa jet propulsion laboratory @xcite . for gbm , we follow the @xcite approach .
the primary difference in the implementation of the occultation technique between gbm and batse arises from the different pointing schemes of the respective missions .
_ cgro _ was three - axis stabilized for each viewing period , which typically lasted for two weeks .
this meant that a source remained at a fixed orientation with respect to the detectors through an entire viewing period .
in contrast , _ fermi _ scans the sky by pointing in a direction @xmath21 ( august 2008september 2009 ) or @xmath22 ( october 2009present ) north of the zenith for one orbit ; it then rocks to @xmath21 or @xmath22 south for the next orbit , continuing to alternate every orbit unless the spacecraft goes into a pointed mode ( which occurs rarely ) .
in addition , the spacecraft performs a roll about the z - axis as it orbits .
because the orientation of a source with respect to the gbm detectors varies as a function of time , the detector response as a function of angle must be accounted for . a detailed
instrument modeling and measurement program has been used to develop the gbm instrument response as a function of direction @xcite , which is incorporated into the occultation analysis .
it should be noted that the gbm occultation sensitivity exceeds that of batse at energies below @xmath1 kev and above @xmath2 mev , since gbm has only a be window on the nai detectors instead of the plastic scintillator , aluminum honeycomb , and aluminum window that covered the front of the batse scintillators @xcite . due to its larger area ,
batse was more sensitive than gbm between 25 kev and 1.5 mev .
gbm has two continuous data types : ctime data with nominal 0.256-second time resolution and 8-channel spectral resolution and cspec data with nominal 4.096-second time resolution and 128-channel spectral resolution .
the results presented in this paper use the low - spectral resolution ctime data .
detailed spectral analyses using the higher resolution cspec data are reserved for future work .
the data are selected to remove gamma - ray bursts , solar flares , electron precipitation events , cosmic ray events , terrestrial gamma - ray flashes , etc .
the occultation technique requires an input catalog of predetermined source locations , and currently we are monitoring 82 sources . for each day
, the occultation times for each source are calculated using the known spacecraft positions .
the time of the occultation step is taken to be the time for which the transmission of a 100 kev gamma ray through the atmospheric column is 50% .
the time at which the atmospheric transmission reaches 50% is energy dependent , e.g. for energies less than 100 kev , a setting step will occur earlier ( see fig . [ crabstep ] ) .
this energy dependence is accounted for in the calculation of the atmospheric transmission function , @xmath23 .
for each occultation step , a 4-minute window is defined that is centered on the 100 kev occultation time . for each energy band
, the count rates in the window are fit separately for each detector viewing the source of interest . in each of these detectors ,
the count rates are fitted with a quadratic background plus source models for the source of interest and each interfering source that appears in the window .
the source models consist of @xmath23 and a time - dependent model count rate , derived from the time - dependent detector response convolved with an assumed source spectrum .
each source model is multiplied by a scaling factor , and the source flux is then computed by a joint fit to the scaling factors across all detectors in the fit . the best - fit scaling factor
is then multiplied by the assumed source flux model integrated over the energy band to obtain the photon flux .
up to 31 occultation steps are possible for a given source in a day , and these steps are summed to get a single daily average flux .
this technique can be used with either the nai or bgo detectors , though the analysis presented here uses only the nai detectors .
a more complete description of the gbm implementation of the occultation technique will be given in @xcite .
lccccccccc cyg x-1 & 1151.0 & 3.7 & 312 & 1130.7 & 6.9 & 163 & 529.0 & 49.5 & 10.7 + crab & 1000.0 & 3.3 & 307 & 1000.0 & 6.3 & 158 & 1000.0 & 48.0 & 20.9 + xte j1752 - 223 & 730.8 & 14.2 & 51 & 563.1 & 26.7 & 21 & 226.1 & 204.4 & 1.1 + cen a & 72.4 & 3.6 & 20 & 104.2 & 6.7 & 16 & @xmath24 & & + swift 1753.5 - 0127 & 121.0 & 4.4 & 28 & 126.8 & 8.2 & 15 & 104.5 & 62.0 & 1.7 + 1e 1740 - 29 & 116.3 & 4.7 & 25 & 92.3 & 8.8 & 11 & 126.8 & 65.0 & 2.0 + grs 1915 + 105 & 128.1 & 3.6 & 35 & 54.9 & 6.8 & 8.0 & @xmath25 & & + gx 339 - 4 & 399.4 & 18.3 & 22 & 249.5 & 33.8 & 7.4 & @xmath26 & &
+ in @xcite , the measured gbm 1250 kev light curves are compared to the _ swift _ bat 1550 kev light curves for several sources over the same time intervals , and it is seen that the fluxes measured by the two instruments compare well . at energies above the @xmath27 kev upper energy limit of the _ swift _ 22-month catalog @xcite , however ,
the gbm observations provide the only wide - field monitor available for the low energy gamma - ray sky . of the catalog sources being monitored with gbm , six persistent sources have been detected above 100 kev with a statistical significance of at least @xmath3 after two years of observations , as well as two transient sources .
table [ flux_table ] gives the fluxes averaged over all 730 days from 2008 august 12 ( mjd 54690 , the beginning of science operations ) to 2010 august 11 ( mjd 55419 ) for the persistent sources , and over all of the days of the flares for the transient sources .
also given are the significances for each energy band .
the errors are statistical only .
the sources are sorted by their detection significance in the 100300 kev band .
the six persistent sources crab , cyg x-1 , cen a , grs 1915 + 105 , 1e 1740 - 29 , and swift j1753.5 - 0127 are detected by gbm at energies above 100 kev . in figures [ crab][swift ] we show light curves for these sources generated from the gbm data in several broad energy bands with five - day resolution .
these persistent sources demonstrate the capabilities of the gbm earth occultation monitoring . the crab emission in the hard x - ray / low energy gamma - ray regime contains a combination of pulsar and pulsar wind nebula contributions .
figure [ crab ] shows the light curves measured by gbm in four broad energy bands from 12 kev up to 500 kev .
the spectrum in this regime has been shown by analysis of batse occultation data @xcite and data from spi on board _ integral _ @xcite to agree with the spectrum measured with other instruments at lower x - ray energies , and then to steepen near 100 kev .
results of the batse analysis can be described by a broken power law , while results of the spi analysis suggest a smoothly steepening spectrum .
the batse analysis further noted a distinct hardening of the spectrum near 650 kev , although this has not been confirmed by _
integral _ or the comptel instrument on _ cgro _ @xcite .
gbm light curve for the crab .
the horizontal scale is in modified julian days over the 730 day gbm exposure period , and has been binned 5 days per data point .
the dashed horizontal lines show the average flux in each of four energy bands increasing from top to bottom . in the bottom plot
, the solid line marks the zero flux level .
note that the apparent `` flare '' near mjd 55180 is due to a giant outburst in the nearby accreting pulsar a0535 + 262.,width=294 ] the _ integral _ spectral measurements are consistent with either a smoothly steepening spectrum or a spectrum of the form @xmath28 photons @xmath29 s@xmath30 kev@xmath30 , where @xmath31 for @xmath32 kev and @xmath33 for @xmath34 kev @xcite .
this corresponds to a ( 50100 kev)/(1250 kev ) flux ratio of @xmath35 for _ integral _ compared to @xmath36 for gbm . the ( 100300 kev)/(1250 kev ) flux ratio @xmath37 corresponding to the _ integral _ spectrum compares to the gbm value of @xmath38 , and the ( 300500 kev)/(1250 kev ) flux ratio @xmath39 for _ integral _ corresponds to the gbm value of @xmath40 .
the gbm measurements suggest a somewhat flatter spectrum than that derived from _ integral _ , particularly above 100 kev , and are best described by a spectrum with @xmath41 .
these results demonstrate that gbm is able to see significant emission above 300 kev at a level consistent with the canonical hard spectrum of the crab .
future analysis with the finer spectral resolution of the gbm cspec data will allow a better determination of the the break energy and spectral index above the break . the gbm light curve for cyg x-1 over 730 days .
the light curve has been binned 5 days per data point .
the fluxes are in crab units , and the dashed and solid lines mark the average flux and zero flux levels , respectively.,width=294 ] cygnus x-1 is a high - mass x - ray binary and was one of the first systems determined to contain a black hole @xcite .
the x - ray emission is bimodal , with the @xmath42 kev emission anticorrelated with the @xmath43 kev emission @xcite .
it has been observed to emit significant emission above 100 kev including a power law tail extending out to greater than 1 mev @xcite .
based on batse occultation analysis , @xcite have shown that in the high gamma - ray intensity ( hard ) state , the spectrum consists of a comptonized shape below 200300 kev with a soft ( @xmath44 ) power - law tail extending to at least 1 mev . in the low - intensity ( soft ) state , however , the spectrum takes on a different shape : in this case , the entire spectrum from 30 kev to 1 mev is characterized by a single power law with a harder index @xmath45 . similar behavior has been reported for two low - mass x - ray binaries ( lmxbs ) that also contain black holes , the transient gamma - ray sources gro j0422 + 32 @xcite and gro j1719 - 24 @xcite .
figure [ cygx1 ] shows the gbm light curves .
the light curves show significant variability with emission above 300 kev up until about mjd 55355 .
starting at about mjd 55355 , the 100300 kev band emission began to decrease @xcite , dropping from an average level of about 1200 mcrab down to nearly undetectable levels on mjd 5540555406 .
on mjd 55374 maxi detected a rapid rise in the soft 24 kev band @xcite , which combined with the decrease in the low energy gamma - ray flux indicated a transition to a thermally - dominated soft state .
as of mjd 55419 the one - day average gbm light curves show that the 1250 kev flux has begun to rise , while the 100300 kev flux remains at a low level of @xmath46 mcrab , consistent with the @xmath47 state of @xcite .
we will continue to monitor cyg x-1 during the thermally - dominated state and follow its transition back to the low / hard state .
the gbm light curves ( fig .
[ cygx1 ] ) reveal significant emission above 300 kev , consistent with the power law tail observed when cyg x-1 is in its low / hard state .
the 50100 kev flux level observed by gbm over the full observation period ( table [ flux_table ] ) is 1.15 crab , consistent with the batse high gamma - ray state ( @xmath48 of @xcite ) .
the observed gbm flux ratios are @xmath49 , @xmath50 , and @xmath51 , inconsistent with a single power law .
a single power law with @xmath52 would yield flux ratios 0.158 , 0.105 , and 0.021 . instead , the gbm ratios suggest a spectrum that appears significantly flatter at low energies and steeper at high energies , consistent with the behavior reported earlier from the batse analysis . the gbm light curve for cen a. the light curve has been binned 5 days per data point .
the fluxes are in crab units , and the dashed and solid lines mark the average flux and zero flux levels , respectively.,width=294 ] the relatively nearby radio galaxy cen a is a seyfert 2 galaxy that is the brightest agn in hard x - rays / low energy gamma rays .
it has powerful jets aligned at approximately @xmath53 from the line of sight and is seen to vary on time scales of tens of days to years .
it has been observed at hard x - ray energies by _ osse _
@xcite , _ integral _ and _ rxte _ @xcite , and at energies @xmath54 mev by comptel @xcite .
the observations below 150 kev are consistent with a hard spectrum with a power law index @xmath55 .
the combined osse and comptel data are consistent with a steepening of the spectrum at 150 kev to @xmath56 , with the spectrum then extending unbroken to beyond 10 mev . the gbm light curve for cen a is shown in fig . [ cena ] .
because cen a is relatively far below the equatorial plane , with a declination @xmath57 , its beta angle ( which ranges between @xmath58 ) can be larger than the half - angle size of the earth as seen from _ fermi _ ( @xmath59 ) .
when this happens , cen a is not occulted .
this causes periodic gaps in the light curve , with the period of the gaps equal to the precession period of the orbit .
the fluxes as measured by gbm and given in table [ flux_table ] are consistent with the hard spectrum measured by previous instruments .
the flux ratios measured by gbm are @xmath60 and @xmath61 respectively .
an unbroken @xmath62 power law up to 300 kev would result in flux ratios of 0.178 and 0.129 , respectively .
a @xmath62 power law extending up to 500 kev would result in a ( 300500 kev)/(1250 kev ) flux ratio of 0.028 , while gbm measures essentially no flux above 300 kev , @xmath63 , consistent with a steepening or cut off somewhere near 300 kev .
the gbm results appear to be consistent with the steepening seen in the osse - comptel spectra .
grs 1915 + 105 light curve .
the light curve has been binned 5 days per data point .
the fluxes are in crab units , the fluxes are in crab units , and the dashed and solid lines mark the average flux and zero flux levels , respectively.,width=294 ] the galactic microquasar grs 1915 + 105 is a lmxb with the compact object being a massive black hole @xcite .
it was highly variable over the 9-year observation period of the _ cgro _ mission @xcite with significant emission observed out to @xmath64 mev @xcite . combining the batse data from multiple outbursts yields a spectrum best fit by a broken power law , with spectral index @xmath65 below 300 kev flattening to @xmath66 above 300 kev .
the spectrum derived from batse data shows no evidence of the thermal spectrum seen in cyg x-1 . by contrast , observations with _ integral_/spi in the 20500 kev energy range @xcite showed evidence for a time - variable thermal comptonization component below @xmath67 kev along with a relatively steady , hard power law at higher energies , indicating that different emission regions are likely responsible for the soft and hard emission .
the gbm daily fluxes integrated over 730 days ( table [ flux_table ] ) show significant emission above 100 kev , consistent with the relatively hard power law spectrum seen in batse and spi data .
the gbm light curve ( fig . [ grs1915 ] ) shows distinct variability below 100 kev , with statistics above 100 kev insufficient to determine the level of variability of the emission .
the flux ratios observed by gbm ( @xmath68 and @xmath69 ) are close to the flux ratios expected from a power law spectrum with @xmath70 ( @xmath71 and @xmath72 ) .
the gbm light curve for 1e 1740 - 29 .
the light curve has been binned 5 days per data point .
the fluxes are in crab units , and the dashed and solid lines mark the average flux and zero flux levels , respectively.,width=294 ] the black hole candidate 1e 1740 - 29 ( also known as 1e 1740.7 - 2942 ) is a lmxb very near the galactic center . with a large double - ended radio jet , it was the first source identified as a microquasar , and spends most of its time in the low / hard state @xcite .
_ integral _ observations indicate the presence of significant emission up to at least 500 kev with a steepening of the spectrum near 140 kev @xcite . the spectrum can be modeled either with a thermalized compton spectrum and a high energy power law tail , or with two superimposed thermal compton components .
evidence for a broad 511 kev line observed by sigma @xcite suggests that 1e 1740 - 29 may be a source of positrons .
the gbm results ( fig .
[ 1e1740 ] ) are consistent with the high energy component observed when 1e 1740 - 29 is in the low / hard state .
below 100 kev and above 300 kev , gbm sees approximately @xmath73 higher flux than _ integral _ , while in the 100300 kev band , gbm observes approximately 90% of the level reported by _
integral_. the gbm light curve for swift j1753.5 - 0127 .
the light curve has been binned 5 days per data point .
the fluxes are in crab units , with the average flux ( dashed lines ) and zero flux ( solid lines ) levels shown.,width=294 ] the x - ray nova swift j1753.5 - 0127 ( fig .
[ swift ] ) is a lmxb with the compact object likely being a black hole @xcite .
_ swift _ discovered this source when it observed a large flare in 2005 july @xcite .
the source did not return to quiescence but settled into a low intensity hard state @xcite . _ integral _ observations @xcite showing emission up to @xmath74 kev were compatible with thermal comptonization modified by reflection , with evidence for separate contributions from a jet , disk , and corona .
batse occultation measurements from 19912000 showed no significant emission from this source above 25 kev @xcite .
the gbm results are consistent with this source still remaining in a hard state , with significant emission in excess of 100 mcrab above 100 kev .
the light curves show that the emission from the higher energy bands declined beginning about mjd 55200 , and that it increased again beginning about mjd 55325 and is currently at or just below its two - year average .
the spectrum is inconsistent with a single power law , and future work using the gbm cspec data will allow a more detailed analysis of the spectrum
. we will continue to monitor this source while it is in the low / hard state . the gbm light curve for xte j1752 - 223 .
the light curve has been binned 5 days per data point .
the vertical dashed lines at mjd 55129 and mjd 55218 mark the flaring region used to derive the average fluxes in table [ flux_table ] .
the fluxes are in crab units , and the horizonial solid lines mark the zero flux levels.,width=294 ] the new transient black hole candidate xte j1752 - 223 , discovered by _
@xcite , was observed by gbm to rise from undetectable on 2009 october 24 ( mjd 55128 ) to @xmath75 mcrab ( 1225 kev ) , @xmath76 mcrab ( 2550 kev ) , @xmath77 mcrab ( 50100 kev ) , and @xmath78 mcrab ( 100300 kev ) on 2009 november 2 @xcite .
the light curve is variable , especially in the 1225 kev band , where the flux initially rose to about 240 mcrab ( 25 - 28 oct ) , suddenly dropped to non - detectable on october 29 - 30 , then rose again during the period october 31 to november 2 ( mjd 5513555137 ) .
the flux remained relatively constant until november 25 ( mjd 55160 ) when it began to rise again , peaking in the high energies on 2009 december 20 ( mjd 55185 ) .
after an initial slow decline , the high energy flux rapidly declined back to the pre - flare levels .
the light curve for the entire mission to date , with 5-day resolution , is shown in fig .
[ xtej1752 ] .
the fluxes for xte j1752 - 223 in table [ flux_table ] are integrated over the days when xte j1752 - 223 was observed to be in a high gamma - ray intensity state , mjd 5512955218 .
rxte measurements indicate a black hole low / hard spectrum with a power law component ( @xmath79 ) superimposed on a weak black body ( kt @xmath80 kev ) .
a 6.4 kev iron line is also seen , with combined spectral and timing properties similar to those observed in cyg x-1 and gx 339 - 4 @xcite .
results from _
rxte_/hexte analysis have shown evidence for emission up to 200 kev @xcite , best fit with a broken power law with a break energy near 130 kev , again markedly similar to cyg x-1 .
the flux ratios measured with gbm ( @xmath81 ) are similar to those observed for cyg x-1 and are consistent with a @xmath82 spectrum steepening to at least @xmath83 above 100 kev . the gbm light curve for gx 339 - 4 .
the light curve has been binned 5 days per data point .
the vertical dashed lines at mjd 55244 and mjd 55289 mark the flaring region used to derive the average fluxes in table [ flux_table ] .
the fluxes are in crab units , and the horizontial solid lines mark the zero flux levels.,width=294 ] the highly variable lmxb and black hole candidate gx 339 - 4 @xcite is characterized by rapid time variability and low / hard x - ray states similar to those of cyg x-1 @xcite .
the results of analysis of both batse @xcite and _ integral _
@xcite data have indicated the presence of high energy emission above 200 kev during previous outbursts , with the _ integral _ spectrum fitted by a thermal comptonization component together with synchrotron or self - synchrotron emission possibly originating at the base of a jet .
gx 339 - 4 was observed by maxi to begin a large flare event starting on 2010 january 3 @xcite .
the flux observed by gbm began to increase starting in early 2010 january and continued to increase up to a level of @xmath84 mcrab ( 1225 kev ) , @xmath85 mcrab ( 2550 kev ) , @xmath86 mcrab ( 50100 kev ) , and @xmath87 mcrab ( 100300 kev ) by early - april 2010 , after which it began to rapidly decrease .
it returned to quiescence in the higher energy bands by mid - april and in the 1250 kev band by the end of april .
the fluxes for gx 339 - 4 in table [ flux_table ] are integrated over the days when gx 339 - 4 was observed to be in a high gamma - ray state , mjd 5524455289 .
similar to cyg x-1 and xte j1752 - 223 , the flux ratios measured by gbm ( @xmath88 and @xmath89 , with no measurable intensity above 300 kev ) appear consistent with a @xmath82 power law steepening above 100 kev to at least @xmath83 .
note that there was a weaker double - peaked flare starting around mjd 54888 @xcite and lasting until around mjd 55000 ( see fig .
[ gx339 ] ) . while the light curve in the 100300 kev band is suggestive of positive emission in this band , it is only marginally significant ( @xmath90 ) with an average flux of @xmath91 mcrab .
using the earth occultation technique , the gbm instrument on _ fermi _ has been monitoring the gamma - ray sky in the @xmath92 kev energy range , providing daily measurements for a catalog of 82 sources , including 72 x - ray binaries , five agn , two magnetars , two cataclysmic variables , the crab , and the sun . after the first two years of the _ fermi _ mission , the earth occultation technique applied to the gbm ctime data has been used to detect six persistent sources and two transients at energies above 100 kev with a significance greater than @xmath3 , demonstrating the capability of gbm to observe and monitor such sources .
two of the sources , the crab and cyg x-1 , were detected with high significance above 300 kev .
light curves of all eight sources were presented in four broad energy bands from 12500 kev .
the outbursts from the transient sources xte j1752 - 223 and gx 339 - 4 were clearly visible in the 1250 kev , 50100 kev , and 100300 kev broad bands .
xte j1752 - 223 was a previously unknown source , and the gbm light curves in the hard x - ray / low energy gamma - ray energy bands are consistent with the initial classification of this object as a black hole candidate in a bright low / hard state .
the steep decline of the hard x - ray emission starting around mjd 55215 corresponded to an increase in the soft x - ray flux @xcite , indicating the transition from the low / hard state to a soft state .
when xte j1752 - 223 returned to the low / hard state around mjd 55282 @xcite , the hard x - ray emission was below the sensitivity limit of gbm .
monitoring of cyg x-1 at the onset of a recent state transition showed a steady decrease in the 100300 kev flux that began about 19 days before the soft x - ray flux began to rise .
as of mjd 55419 , cyg x-1 remains in a soft state , and we will continued to monitor cyg x-1 in anticipation of the transition back to the canonical hard state . while the gbm ctime data used here does not have enough spectral resolution to produce detailed spectra , the flux ratios between the 1250 kev broad band and the 50100 , 100300 , and 300500 kev broad bands for the crab are generally consistent with those inferred from the measured _ integral _ spectrum , with the gbm results suggesting a slightly harder spectrum .
the flux ratios observed for the transient sources xte j1752 - 223 and gx 339 - 4 are similar to cyg x-1 when it is in its canonical low / hard state , again consistent with these transients being observed in bright low / hard states .
future work will use the gbm cspec data , with its finer energy binning , to examine the detailed spectra for all of these sources , with particular emphasis on the low energy gamma - ray energy range .
also , the bgo detectors , with their greater sensitivity at higher energies , will be used to obtain additional measurements at energies above 150 kev .
several of the detected sources have spectral breaks or cutoffs in the 100300 kev range , and we will look for these features and monitor their evolution over time . we will continue to add to the list of sources being monitored as appropriate .
we have detected cen a with gbm in all energy bands up to 300 kev .
batse detected several bl lac objects known to exhibit flaring activity on the time scale of days @xcite .
the flaring behavior of blazars is sometimes accompanied by a shift upwards in the peak of the synchroton spectrum ( e.g. above 100 kev in the 1997 mrk 501 flare @xcite ) , making detection of these sources possible with gbm .
while none have been detected so far , we will continue to expand our monitoring program to include the blazars with flares detected in other wavelengths .
we plan to expand our monitoring program to include other agn as well .
we continue to fine tune the algorithms and work to reduce the systematic errors in the flux determination . in the case of the batse earth occultation analysis
, there was clear evidence for the presence of sources in the data which were not in the occultation input catalog and which caused uncertainty in the assignment of fluxes to the sources that were in the input catalog .
we see similar evidence for these uncataloged sources with gbm , especially below about 50 kev . to address this ,
our approach is two - fold : ( 1 ) we compare our gbm measurements with overlapping energy bands from other operating missions and regularly update our catalog as new transient outbursts are observed with gbm or other instruments .
light curves are regenerated if needed after catalog updates .
( 2 ) we are developing an imaging technique for gbm to produce an all - sky map of hard x - ray / soft gamma - ray sources .
this map will then be used to identify sources not currently in the gbm occultation catalog , to expand the catalog , and to reduce the uncertainties in the measured fluxes .
this work is supported by the nasa fermi guest investigator program . at lsu ,
additional support is provided by nasa / louisiana board of regents cooperative agreement nnx07at62a .
a.c.a . wishes to thank the spanish ministerio de ciencia e innovacin for support through the 2008 postdoctoral program micinn / fulbright under grant 2008 - 0116 .
atwood , w. b. et al .
2009 , , 697 , 1071 bissaldi , e. et al .
2009 , exp .
, 24 , 47 bolton , c. t. 1972 , , 235 , 271 bouchet , l. , del santo , m. , jourdain , e. , roques , j. p. , bazzano , a. , & decesare , g. 2009 , , 693 , 1871 bouchet , l. et al .
1991 , , 383 , l45 cadolle bel , m. et al .
2007 , , 659 , 549 caballero - garca , m. d. et al .
2009 , , 692 , 1339 case , g. l. , cherry , m. l. , fannin , c. , rodi , j. , ling , j. c. , & wheaton , w. a. 2005 , suppl . , 5 , 341 case , g. l. , cherry , m. l. , ling , j. c. , & wheaton , w. a. 2008 , in 2nd kolkata conference on observational evidence for black holes in the universe ( aip conf . proc .
1053 ) , ed . s. k. chakrabarti & a. s. majumdar ( melville , ny : aip ) , 145 case , g. l. , wilson - hodge , c. a. , cherry , m. l. , kippen , r. m. , ling , j. c. , radocinski , r. g. , & wheaton , w. a. 2007 , in the first glast symp .
( aip conf .
921 ) , ed .
s. ritz , p. michelson , & c. a. meegan ( melville , ny : aip ) , 538 case , g. l. et al .
( in preparation , 2010 ) connaughton , v. , robinson , c. r. , mccollough , m. l. , & laurent - muehleisen , s. 1999 , in bl lac phenomenon ( asp conf . series 159 ) , ed . l. o. takalo & a. sillanp ( san francisco , ca : asp ) , 209 cowley , a. p. , schmidtke , p. c. , hutchings , j. b. , & crampton , d. 2002 , , 123 , 1741 dolan , g. f. et al . 1977 , , 267 , 813 doxsey , r. et al . 1979 , , 228 , l67 droulans , r. & jourdain , e. 2009 , 494 , 229 fishman , g. j. et al .
1989 , in proc .
gamma - ray observatory science workshop , ed .
w. johnson ( greenbelt , md : gsfc ) , 2 gehrels , n. et al .
2004 , , 611 , 1005 greiner , j. , cuby , j. g. , & mccraughrean , m. j. 2001 , , 414 , 522 harmon , b. a. et al .
1994 , , 425 , l17 harmon , b. a. et al .
2002 , , 138 , 149 harmon , b. a. et al .
2004 , , 154 , 585 homan , j. 2010 , atel , 2387 hoover , a. s. et al .
2008 , in gamma - ray bursts 2007 ( aip conf . proc .
1000 ) , ed .
m. galassi , d. palmer , & e. fenimore ( melville , ny : aip ) , 565 jourdain , e. & roques , j. p. 2009
, , 704 , 17 kinzer , r. l. et al .
1995 , , 449 , 105 kuiper , l. et al .
2001 , , 378 , 918 levine , et al .
1996 , , 469 , l33 ling , j. c. , mahoney , w. a. , wheaton , w. a. , & jacobson , a. s. 1987 , , 321 , l117 ling , j. & wheaton , w. 2003a , , 584 , 399 ling , j. & wheaton , w. 2003b , , 598 , 334 ling , j. & wheaton , w. 2005a , , 622 , 492 ling , j. & wheaton , w. 2005b , suppl . , 5 , 80 ling , j. c. et al .
1997 , , 484 , 375 ling , j. c. et al .
2000 , , 127 , 79 markwardt , c. b. et al .
2009a , atel , 1945 markwardt , c. b. et al .
2009b , atel , 2258 matsuoka , m. et al .
2009 , , 61 , 999 mcconnell , m. et al .
2000 , , 523 , 928 meegan , c. et al .
2009 , , 702 , 791 miller , j. m. , homan , j. , & miniutti , g. 2006 , , 652 , l113 mirabel , i. f. , rodrguez , l. f. , cordier , b. , paul , j. , & lebrun , f. 1993 , , 358 , 215 much , r. et al .
1996 , , 120 , 703 muoz - darias , t. , motta , s. , belloni , t. & homan , j. 2010a , atel , 2518 muoz - darias , t. , motta , s. , pawar , d. , belloni , t. m. , campana , s. , & bhattacharya , d. 2010b , mnras , 404 , l94 negoro , h. et al . 2010a , atel , 2396 negoro , h. et al . 2010b , atel , 2711 paciesas , w. s. et al .
1993 , , 97 , 253 paciesas , w. s. , deal , k. j. , harmon , b. a. , wilson , c. a. , zhang , s. n. , & fishman , g. j. 1995 , annals of the new york acad . of sci . , 759 , 308 paczynski , s. 1974 , , 34 , 161 palmer , d. et al .
2005 , atel , 546 petry , d. et al .
2000 , , 536 , 742 rothschild , r. e. et al . 2006 , , 641 , 801 samimi , j. et al .
1979 , , 278 , 434 shaposhnikov , n. , markwardt , c. b. & swank , j. h. 2009 , atel , 2269 shaw , s. e. et al .
2004 , , 418 , 1187 steinle , h. et al .
1998 , , 330 , 97 sunyaev , r. , et al .
1991 , , 383 , l49 tueller , j. et al .
2010 , , 186 , 378 us committee on extension to the standard atmosphere 1976 , us standard atmosphere , noaa - s / t 76 - 1562 , ( washington , dc : u.s .
govt printing office ) wilson - hodge , c. a. & case , g. 2010 , atel , 2721 wilson - hodge , c. a. , camero - arranz , a. , case , g. , chaplin , v. , & connaughton , v. 2009b , atel , 2280 wilson - hodge , c. a. et al .
2009a , in proceedings of the 2009 _ fermi _ symposium , econf c091122 wilson - hodge , c. a. et al .
( in preparation , 2010 ) yamaoka , k. et al .
2010 , atel , 2380 zdziarski , a. a. , grove , j. e. , poutanen , j , rao , a. r. , & vadawale , s. v. 2001 , , 554 , l45 | the nai and bgo detectors on the gamma - ray burst monitor ( gbm ) on _ fermi _ are now being used for long - term monitoring of the hard x - ray / low energy gamma - ray sky . using the earth occultation technique as demonstrated previously by the batse instrument on the _ compton gamma - ray observatory _ , gbm can be used to produce multiband light curves and spectra for known sources and transient outbursts in the 8 kev to 1 mev energy range with its nai detectors and up to 40 mev with its bgo detectors . over 85% of the sky
is viewed every orbit , and the precession of the _ fermi _ orbit allows the entire sky to be viewed every @xmath0 days with sensitivity exceeding that of batse at energies below @xmath1 kev and above @xmath2 mev .
we briefly describe the technique and present preliminary results using the nai detectors after the first two years of observations at energies above 100 kev .
eight sources are detected with a significance greater than @xmath3 : the crab , cyg x-1 , swift j1753.5 - 0127 , 1e 1740 - 29 , cen a , grs 1915 + 105 , and the transient sources xte j1752 - 223 and gx 339 - 4 .
two of the sources , the crab and cyg x-1 , have also been detected above 300 kev . |
the watershed is a versatile and powerful segmentation tool .
its use for segmentation is due to ch.lantujoul and s.beucher @xcite .
it may be applied on an image considered as a topographic surface @xcite , @xcite , @xcite .
an image may be considered as a node weighted graph ; the nodes are the pixels of the image , weighted by their grey tone ; the edges connect neighboring nodes and are not weighted . . or the watershed may be applied on an edge weighted graph , such as a region adjacency graph @xcite , where the nodes are unweighted and represent the catchment basins , and the edges connecting neighboring basins are weighted by the altitude of the pass point separating two basins . in the first case
, one has to find the watershed on a node weighted graphs , in the second on an edge weighted graph @xcite .
definitions and algorithms are not the same in both worlds , although the have the same physical inspirations . the rain model , where the destiny of a drop of water falling on the surface defines the catchment basins ; a catchment basin is the attraction zone of a minimum , i.e. the set of nodes from where a drop of water may reach this minimum .
catchment basins generally overlap.the flooding model where the relief is flooded from sources placed at the regional minima and meet for forming a partition .
the later method being often implemented as shortest distance algorithms @xcite , @xcite .
there is no opposition between these models as the same trajectory may be followed from bottom to top , and we have a flooding model or from top to bottom and we have a rain model . a good review on the watershed may be found in @xcite , and in the recent book @xcite .
this paper aims at showing the equivalence between edge or node weighted graphs for the construction of the watershed .
a _ non oriented graph _
@xmath0 $ ] is a collection @xmath1 of vertices or nodes and of edges @xmath2 an edge @xmath3 being a pair of vertices ( see @xcite,@xcite ) . _
a chain _ of length @xmath4 is a sequence of @xmath4 edges @xmath5 , such that each edge @xmath6 of the sequence @xmath7 shares one extremity with the edge @xmath8 @xmath9 , and the other extremity with @xmath10 @xmath11 . a _ path _ between two nodes @xmath12 and
@xmath13 is a sequence of nodes @xmath14 such that two successive nodes @xmath15 and @xmath16 are linked by an edge . _ a cycle
_ is a chain or a path whose extremities coincide . _
a cocycle _ is the set of all edges with one extremity in a subset @xmath17 and the other in the complementary set @xmath18 the subgraph spanning a set @xmath19 is the graph @xmath20 $ ] , where @xmath21 are the edges linking two nodes of @xmath22 the partial graph associated to the edges @xmath23 is @xmath24.$ ] _ a connected graph _ is a graph where each pair of nodes is connected by a path . in a graph @xmath0
, $ ] edges and nodes may be weighted : @xmath25 is the weight of the edge @xmath26 and @xmath15 the weight of the node @xmath27 the weights take their value in the completely ordered lattice @xmath28 . [
[ regional - minima ] ] regional minima + + + + + + + + + + + + + + + a subgraph @xmath29 of an edge weighted graph @xmath30 is a flat zone , if any two nodes of @xmath29 are connected by a chain of uniform altitude . a subgraph @xmath29 of a graph @xmath30 is a regional minimum
if @xmath29 is a flat zone and all edges in its cocycle have a higher altitude . [ [ catchment - basins ] ] catchment basins + + + + + + + + + + + + + + + + a chain @xmath31 is a _ flooding chain _
, if each edge @xmath32 is one of the lowest edges of its extremity @xmath33 and if along the chain the weigths of the edges is never increasing .
the catchment basin of a minimum @xmath34 is the set of nodes linked by a flooding chain with a node within @xmath35 [ [ regional - minima-1 ] ] regional minima + + + + + + + + + + + + + + + a subgraph @xmath29 of a node weighted graph @xmath30 is a flat zone , if any two nodes of @xmath29 are connected by a path along which all nodes have the same altitude . a subgraph @xmath29 of a graph @xmath30 is a regional minimum if @xmath29 is a flat zone and all neighboring nodes have a higher altitude . [
[ catchment - basins-1 ] ] catchment basins + + + + + + + + + + + + + + + + a _ flooding path _ between two nodes is a path along which the weigths of the nodes is never increasing .
the catchment basin of a minimum @xmath34 is the set of nodes linked by a non ascending path with a node within @xmath35 the definition of the catchment basins is the loosest possible , compatible with the physical inspiration of the rain model : a drop of water falling on a surface can not go upwards . with this definition
, the same node may belong to various catchment basins .
in other words there are large overlapping zones of the catchment basins . as most algorithm aim at producing a partition
, they propose various methods for suppressing these overlapping zones .
we want to show the equivalence of the watershed on node and edge weighted graphs .
we first present the outline of the method on the two simple graphs in fig.[wshdex1 ] , the left one being edge weighted and the right being node weighted .
[ ptb ] wshdex1-eps-converted-to.pdf consider first the edge weighted graph .
it has 4 nodes a , b , c and d separated by weighted edges . in a flooding chain ,
each node is linked with the next node by one of its lowest edges .
for this reason , if an edge is not the lowest edge of one of its extremities , it will never be crossed by a flooding chain .
it is the case for the edge bc : the lowest adjacent edge of b is ab and the lowest adjacent edge of c is cd .
for this reason , this edge bc can be suppressed from the graph ( as presented in the second line ) , without modifying the flooding chains of the original graph . in a last step
we assign weights to the nodes : each node gets the weight of its lowest adjacent edge , as represented in the third line .he resulting graph is called flooding graph . in our case , it has two regional minima , which are identical if one considers them from the point of view of the edge weights or the node weights . for this simple graph , they constitute the catchment basins .
consider now the node weigthed graph on the right of fig.[wshdex1 ] .
it has two isolated regional minima .
one adds a loop edge linking each isolated node with itself as shown in fig.[wshdex1 ] .
this modification does not change the flooding paths of the initial graph .
the last step consists in assigning , as weight to each edge , the maximal weight of its extremities .
the added loops algo get an edge weight . as a result we get again a flooding graph with the following features : * the edges spanning the regional minima of the node weighted graph are the regional minima of the edge weighted graph .
* the lowest adjacent edge of a node has the same weight as this node .
for this reason each flooding path of the node weighted graph is simultaneously a flooding chain of the edge weighted graph .
having an identity between the minima and between flooding paths and chains , the catchment basins of both graphs are the same . in our case
we have two basins with an overlapping zone containing the node f@xmath36
we define two operators between edges and nodes : - an erosion @xmath37 _ { ij}=n_{i}\wedge n_{j}$ ] and its adjunct dilation @xmath38 _ { i}=\underset{(k\text{\thinspace neighbors\thinspace of\thinspace}\,i)}{\bigvee e_{ik}}$]- a dilation @xmath39 _ { ij}=n_{i}\vee n_{j}$ ] and its adjunct erosion @xmath40 _ { i}=\underset{(k\text{\thinspace neighbors\thinspace of\thinspace}\,i)}{\bigwedge e_{ik}}$ ] the pairs @xmath41 and @xmath42 are adjunct operators .
the pairs @xmath43 and @xmath44 dual operators .
we call the first pair * flooding adjunction * as we may give it a physical explanation .
let us consider a region adjacency graph of a topographical surface , where @xmath15 and @xmath45 represent the flood level in the basins @xmath46 and @xmath47 and @xmath25 represents the altitude of the pass pont between both basins . then : the altitudes of the nodes @xmath46 and @xmath47 the lowest flood covering @xmath46 and @xmath48 has the altitude @xmath49 _ { ij}=n_{i}\vee n_{j}$ ] if @xmath46 represents a catchment basin , @xmath50 the altitude of the pass points with the neighboring basin @xmath51 then the highest level of flooding without overflow through an adjacent edge is @xmath40 _ { i}=\underset{(k\text{\thinspace neighbors\thinspace of\thinspace}\,i)}{\bigwedge e_{ik}}.$ ] as @xmath52 and @xmath53 are adjunct operators , the operator @xmath54 is a closing on @xmath4 and @xmath55 is an opening on @xmath56 .
we consider first an edge weighted graph @xmath58 and study the effect of the opening @xmath57 on its edge weights .
fig.[wfal22 ] presents from left to right : 1 ) an edge weighted graph , , 2 ) the result of the erosion @xmath52 , 3 ) the subsequent dilation producing an opening . the edges in red are those whose weight has been reduced by the opening .
the others , invariant by the opening * * * * @xmath57 are the edges which are , as we establish below , the lowest edge of one of their extremities .
two possibilities exist for an edge @xmath26 with a weight @xmath59 the edge @xmath26 has lower neighboring edges at each extremity . hence @xmath60 and @xmath61 ; hence @xmath62 the edge @xmath26 is not invariant by the opening @xmath57 the edge @xmath26 is the lowest edge of the extremity @xmath27 then @xmath63 and @xmath64 ; hence @xmath65 the edge @xmath26 is invariant by the opening @xmath57 hence the edges which are invariant by @xmath57 are all edges which are the lowest edges for one of their extremities .
the operator keeping for each node only its lowest adjacent edges is written @xmath66as each node has at least one lowest neighboring edge , the resulting graph @xmath67 spans all the nodes .
the resulting graph also contains all flooding chains of the initial graph ( since in a flooding chain , each edge is the lowest edge of one of its extremities ) .
consider an edge weighted graph @xmath58 invariant by the opening @xmath68 we assign to the nodes the weights @xmath69we call @xmath70 the graph on which one only considers the node weights .
if an edge weighted graph @xmath58 is invariant by the opening @xmath71 then its regional minima edges span the regional minima nodes of @xmath72 * proof : * a regional minimum @xmath34 of the graph @xmath58 is a plateau of edges with altitude @xmath73 , with all adjacent edges in the cocycle having a weight higher than @xmath74if a node @xmath46 belongs to this regional minimum , its adjacent edges have a weight @xmath75 but it has at least one neighboring edge with weight @xmath73 : hence the weight of @xmath46 is @xmath76 consider now an edge @xmath77 with @xmath78 inside the regional minimum and @xmath79 outside
. then @xmath80 . as @xmath81is invariant by @xmath71
the edge @xmath82 is one of the lowest edges of the nodes @xmath79 : thus the weight of @xmath79 is @xmath83this shows that the nodes spanned by the regional minimum @xmath34 form a regional minimum of the graph @xmath72 consider now a node weighted graph @xmath85the closing @xmath84 is obtained by a dilation @xmath53 of the node weights followed by an erosion @xmath69 the closing @xmath84 replaces each isolated node constituting a regional minimum by its lowest neighboring node and leaves all other nodes unchanged .
consider a node with a weight @xmath73 belonging to a regional minimum : consider the case where the node @xmath46 is an isolated regional minimum .
then @xmath53 assigns to all edges adjacent to @xmath46 a weight bigger than @xmath74the subsequent erosion @xmath52 assigns to @xmath46 the smallest of these weights , which is the weight of the smallest neighbor . the node @xmath46 is not invariant by the closing @xmath86 suppose that @xmath46 belongs to a regional minimum which is not isolated .
the dilation @xmath53 assigns to each edge adjacent to @xmath46 a weight @xmath87if @xmath46 has a a neighbor @xmath48 with a weight @xmath88then @xmath53 assigns to the edge @xmath26 the weight @xmath74the subsequent erosion @xmath52 assigns to @xmath46 the smallest of these weights , that is @xmath89 the node @xmath46 is invariant by the closing @xmath86 hence if @xmath30 has isolated regional minima , it is not invariant by @xmath90if we add a loop edge linking each isolated regional minimum with itself we obtain a graph invariant by @xmath91indeed , if @xmath46 is an isolated regional minimum with weight @xmath92 we add a loop edge @xmath93 ; the dilation @xmath53 assigns to the loop the weight @xmath73 and to all other edges adjacent to @xmath94 a weight @xmath95the subsequent erosion @xmath52 assigns to @xmath46 the smallest of these weights , which is the weight of the loop , i.e. @xmath89 we write @xmath96 the operator which adds to a node weighted graph a loop between each isolated regional minimum and itself .
consider a node weighted graph @xmath70 invariant by the closing @xmath86 we assign to the edges the weights @xmath97we call @xmath58 the graph on which one only consider the edge weights .
if @xmath30 is invariant by the closing @xmath98 then the edges spanning the regional minima nodes of @xmath70 form the regional minima edges of @xmath99 * proof : * a regional minimum @xmath34 of @xmath70 is a plateau of pixels with altitude @xmath73 , containing at least two nodes ( there are no isolated regional minima as @xmath70 is invariant by @xmath84 ) .
all internal edges of the plateau get the valuation @xmath73 by @xmath100 if an edge @xmath26 has the extremity @xmath46 in the minimum and the extremity @xmath48 outside , then @xmath101hence , for the graph @xmath102 the edges spanning the nodes of @xmath34 form a regional minimum .
we consider now a graph @xmath103 on which both nodes and edges are weighted . if we consider only the edge weights we write @xmath58 and @xmath70 if we consider only the node weights .
* definition : * an edge and node weighted graph @xmath104 $ ] is a flooding graph iff its weight distribution @xmath105 verifies both @xmath106 and @xmath107 in a flooding graph , the weight distribution @xmath105 verifies @xmath108 and @xmath109 showing that @xmath110 and @xmath111 as @xmath30 is invariant by @xmath71 all its edges are the lowest edge of one of their extremities . and as @xmath30 is invariant by @xmath98 it has no isolated regional minimum . we have established earlier that : * if a graph @xmath58 is invariant by @xmath71 then its regional minima edges span the regional minima nodes of @xmath112 * if a graph @xmath70 is invariant by @xmath98 then the edges spanning the regional minima nodes of @xmath70 form the regional minima edges of @xmath99 as a flooding graph is both invariant by @xmath57 and by @xmath98 we get the following theorem . * * if @xmath30 is a flooding graph then the node weighted graph @xmath70 and the edge weighted graph @xmath58 have the same regional minima subgraph .
more precisely , the regional minima nodes of @xmath70 are spanned by the regional minima edges of @xmath99 fig.[wfal28 ] presents the same flooding graph , on the left with its edge weights and on the right with its node weights : they have exactly the same regional minima . consider a flooding path on @xmath113 i.e. a path @xmath114 of neighboring nodes with a non increasing weight , starting at node @xmath115 and ending at node @xmath116 belonging to a regional minimum of @xmath70 .
as @xmath117 , the weight of the edge @xmath118 obtained by @xmath53 is equal to the weight of @xmath15 and is one of teh lowest edges of @xmath119 this shows , that the series of edges @xmath120 form a flooding chain of @xmath58 , ending in a regional minimum of @xmath121 inversely , consider a flooding chain@xmath122 $ ] of @xmath99 the weights of the edges are not increasing .
furthermore , the edge @xmath118 is the lowest edge of the node @xmath15 . as in a flooding graph the lowest edge of a node has the same weight than this node , the edge @xmath123 has the same weight than the node @xmath124thus the path @xmath114 also is a flooding path of @xmath70 ending a regional minimum of @xmath72 there is a one to one correspondance between the regional minima of @xmath58 and @xmath70 ; there is also a one to one correspondance between the flooding paths and the flooding tracks , with the same weight distribution .
this shows that both graphs , @xmath58 on which we only consider the edge weights and @xmath70 on which we only consider the node weights , have the same catchment basins .
consider an arbitrary edge weighted graph @xmath58 with edge weights @xmath125 and without node weights .
the operator @xmath67 suppresses all edges which are not invariant by @xmath126the remaining edges are invariant by @xmath57 and verify : @xmath127 if we assign to the nodes of @xmath58 weights equal to @xmath128 then @xmath129 and the resulting graph is a flooding graph .
[ [ illustration ] ] illustration + + + + + + + + + + + + fig.[wshdex4]a presents an edge weighted graph @xmath130 on which the node weights are produced by the erosion @xmath131 in fig.[wshdex4]b a dilation @xmath53 applied after the erosion produces an opening of the initial edge weights .
the weights with a red color are those which are lowered by the opening .
these edges are suppressed producing the graph @xmath67 in fig.[wshdex4]c .
the erosion @xmath132 applied on @xmath30 or on @xmath67 produces the same node weights .
the resulting graph is a flooding graph .
[ wshdex4]d shows the regional minima of the complete graph if one considers only the node weights .
in contrast , fig .
[ wshdex4]e shows the regional minima of the flooding graph , if one only considers the node weights .
they are not identical . if we consider only the edge weights of the flooding graph , we write @xmath58 and @xmath70 if we consider only the node weights .
one verifies that the edges spanned by the regional minima nodes of @xmath70 in fig .
[ wshdex4]e are spanned by the regional minima edges of @xmath58 in fig.[wshdex4]f .
one also verifies that flooding paths and flooding chains are identical , each node being followed by an edge with the same weight .
[ ptb ] wshdex4-eps-converted-to.pdf consider an arbitrary node weighted graph @xmath70with node weights @xmath133 and without edge weights .
the operator @xmath134 adds a loop edge between each isolated regional minimum and itself , producing a graph invariant by @xmath91the nodes verify @xmath135 if we assign to the edges of @xmath70 weights equal to @xmath136 then @xmath137 and the resulting graph is a flooding graph .
[ [ illustration-1 ] ] illustration + + + + + + + + + + + + fig.[wshdex7]a presents a node weighted graph .
it is not invariant by @xmath84 as it has isolated regional minima .
one adds a loop edge linking each isolated regional minimum with itself , producing the graph in fig.[wshdex7]b .
the dilation @xmath53 produces the edge weights .
the resulting edge and node weighted graph in [ wshdex7]c is a flooding graph .
the regional minima have been highlighted by distinct colors .
again , the identity between flooding paths and flooding chains is clearly visible .
[ ptb ] wshdex7-eps-converted-to.pdf we have presented how to derive from any node or edge weighted graph a flooding graph with the same regional minima .
to each flooding path corresponds a flooding chain with the same weight distribution , as each node is followed by an edge with the same weight .
the reverse also is true : to each flooding chain corresonds a flooding path with the same weight distribution . the catchment basins , as defined earlier , are exactly the same .
as all non ascending paths and chains are accepted for defining the catchment basins , we also have the largest overlaping zones between them .
two classes of watershed algorithms have been developed , the ones for node weighted graphs , the others for edge weighted graphs .
since we have a one to one correspondance between flooding paths and flooding chains , each algorithm developed for a node ( resp .
edge ) weighted graph may now also be applied for an edge ( .resp node ) weighted graph , by applying it on the associated flooding graph .
fig.[wshdex8]a1 presents a flooding graph with three regional minima .
the three associated catchment basins overlap at a node p with weight 4 . in order to break the tie , only one edge adjacent to the node @xmath138 should be kept , and the others suppressed .
three solutions are possible , illustrated by the figures [ wshdex8]a2 , b2 and c2 .
there is no means to decide between one or the other solution if one considers only the weight on one edge .
if one considers the first two edges of each flooding path , we obtain a lexicographic measure , by concatenating the weight of the first and the second , illustrated in fig .
[ wshdex8]b1 .
there remains now two choices between the edges with weights @xmath139 represented in figures figures [ wshdex8]b2 and c2 .
considering a third edge along the flooding paths leaves only one choice , as the edge with the weight @xmath140 is the lowest . the corresponding waterhed segmentation is presented in fig .
[ wshdex8]c2 .
this example shows that we are able to reduce the number of partitions associated to a flooding graph , if one considers not only the first neighboring nodes or edges in the flooding paths or chains , but a number of nodes , ordered in a lexicographic order .
[ ptb ] wshdex8-eps-converted-to.pdf by defining a lexicographic order among the flooding chains of depth @xmath141 has permitted to break the ties , leaving only one solution .
given a node or weighted graph we first derive its flooding graph @xmath142.$ ] we associate to @xmath30 an oriented graph @xmath143 by replacing the edge @xmath144 by an arrow @xmath145 if @xmath146and@xmath147by two arrows @xmath145 and @xmath148 if @xmath149 . the loop edge linking an isolated regional minimum node @xmath34 with itself also is replaced by an arrow @xmath150 the graph @xmath143 verifies the property @xmath151 : @xmath152 there exists at least an oriented path @xmath153 of @xmath143 ( with a positive or null length ) linking @xmath138 with a regional minimum .
we define the catchment basin of a regional minimum as the set of nodes linked by an oriented path with this minimum .
obviously , each node belongs to at least one catchment basins .
catchment basins may overlap and form a watershed zone when two paths having the same node as origin reach two distinct regional minima .
we aim at pruning the graph @xmath154 without any arbitrary choices , and get a partial graph @xmath155 for which the property @xmath151 still holds but the watershed zones are smaller .
as soon the path @xmath153 reaches a regional minimum , it may be prolonged into a path of infinite length , by infinitely cycling between @xmath156 nodes within the regional minimum or along the loop joining each isolated regional minimum with itself .
all oriented paths or chains are thus of infinite length .
and we may consider them , either in their full infinite length or consider only the first @xmath157 edges .
we now define a family of preorder relations ( order relation without antisymmetry ) between the paths of @xmath158 the lexicographic preorder relation of length @xmath157 compares the infinite paths @xmath159 and @xmath160 by considering the @xmath157 first nodes and edges : @xmath161 if @xmath162 or there exists @xmath163 such that [ c]l@xmath164 + @xmath165 @xmath166 if @xmath161 or if @xmath167 this preorder relation is total , as it permits to compare all paths ; for this reason , among all paths linking a node @xmath138 with a regional minimum , there exists always at least one which is the smallest for @xmath168 we say that this path is the steepest for the lexicographic order of depth @xmath169 for @xmath170 we consider the infinite paths and we simply write @xmath171 . if @xmath172 and@xmath173 are two paths of infinite length verifying @xmath174 , then the paths @xmath175 and@xmath176 obtained by skipping the @xmath177 first nodes also verify @xmath178if @xmath172 is the smallest path linking its origin with a regional minimum , then @xmath175 is the smallest path leading from @xmath179 to the same regional minimum .
consider two lexicographic order relations @xmath180 and @xmath181 with @xmath182 then for @xmath183 and @xmath184 @xmath185 or equivalently @xmath186 the steepest path for the lexicographic order @xmath177 also is steepest for the lexicographic order @xmath157 . as a consequence , a catchment basin for @xmath187
is included in the catchment basin for @xmath188 for increasing values of @xmath51 the catchment basins become larger , are nested , and the watershed zones are reduced or vanish . for @xmath170
a node is linked by two minimal paths with two distinct minima , only if these two paths have exactly the same weights , which seldom happens in natural images .
in particular , if the regional minima have distinct weights , the catchment basins form a partition .
we associate to each order relation @xmath189 of length @xmath157 a pruning operator @xmath190the pruning @xmath191 suppresses each edge which is not the first edge of a steepest path for @xmath189 among all paths with the same origin @xmath36after pruning , each node outside the regional minima is the origin of one or several @xmath192steepest flooding tracks or @xmath192steepest flooding paths we say that the graph @xmath193 has a k - steepness or is k - steep . as for @xmath182 @xmath194 , we have @xmath195 . furthermore @xmath196 each pruning @xmath193 suppresses a number of edges still present in @xmath197 ; it suppresses them all , without doing any arbitrary choices between them .
[ [ particular - k - steep - graphs ] ] particular k - steep graphs + + + + + + + + + + + + + + + + + + + + + + + + + applied to an arbitrary graph , the pruning @xmath198 suppresses the edges which are not the lowest edge of one of their extremities . in a flooding graph , each edge it the lowest edge of one of its extremities and @xmath199 is inoperant .
the pruning @xmath200 keeps for each node @xmath46 the adjacent edges linking @xmath46 with one of its lowest neighboring nodes .
the pruning @xmath201 only keeps the first edge of the steepest paths .
* lemma : * any oriented path in @xmath193 of length @xmath157 is of maximal steepness for @xmath168 for this reason , a node @xmath138 belongs to a @xmath192catchment basin associated to a node @xmath34 in a regional minimum , if there exists an oriented path in @xmath193 from @xmath138 to @xmath202for increasing values of @xmath51 the catchment basins are decreasing , and so are the overlapping zones between them .
the operator @xmath191 defined above has nice properties but is not a local operator .
it is however possible to implement it using only local operators as we present now .
the adjunctions @xmath203 and @xmath204 were defined for non oriented graphs .
we now define the equivalent operators for oriented graphs .
the erosion from arrows to nodes assigns to each node @xmath138 the minimal weight of all arrows having @xmath138 as origin : @xmath205 the dilation is obtained by adjunction . consider a weight distribution @xmath78 on the nodes of the oriented graph and a weight distribution @xmath56 on the arrows .
@xmath206for @xmath207 the dilation @xmath208 assigns to the arrow @xmath145 the weight of its origin @xmath209 we also will need the dual erosion @xmath210 assigning to the arrow @xmath211 the weight of its extremity @xmath212 .
consider an arrow @xmath213the erosion @xmath214 assigns to the node @xmath138 the minimal weight of all arrows having @xmath138 as origin .
the subsequent dilation assigns to the arrow @xmath145 the weight of @xmath215 i.e. the the minimal weight of all arrows having @xmath138 as origin .
thus if @xmath216 leaves the arrow @xmath145 unchanged , it means that this arrow is one of the lowest arrows having @xmath138 as origin .
we define a pruning operator @xmath217 which cuts all arrows which are not invariant by the opening @xmath218 i.e. which are not one of the lowest edges of their origin .
we say that an edge and node weighted graph is an oriented flooding graph if the weights of the nodes and of the arrows verify : for the node weights @xmath219 and for the edge weights : @xmath220 such a graph is invariant by the opening @xmath221 and by the closing @xmath222 we start with a node and edge weighted flooding graph @xmath85as explained above , we associate to @xmath30 an oriented graph @xmath143 by replacing the edge @xmath144 by an arrow @xmath145 if @xmath223and@xmath147by two arrows @xmath145 and @xmath148 if @xmath149 . the loop edge linking an isolated regional minimum node @xmath34 with itself also is replaced by an arrow @xmath150 it is easy to show that if @xmath30 is a flooding graph , then @xmath143 is an oriented flooding graph . in order to identify edges which belong to flooding chains of maximal lexicographic @xmath192steepness , we have to shift the weight distribution of nodes and weights along the oriented flooding track upwards .
this is obtained thanks to the erosion @xmath224 which assigns to the arrow @xmath145 the weight of its extremity @xmath225 after this erosion , we get an edge weight distribution which is not invariant anymore by the opening @xmath226applying the pruning operator @xmath217 leaves a graph which is invariant by @xmath226a final erosion @xmath227 assigns to the nodes their new weights . thus we have applied to @xmath143 the following operators : @xmath228 from nodes to arrows , the pruning operator @xmath229 on the edges followed by a last erosion @xmath230 from the edges to the nodes , producing node weights @xmath231the resulting graph is an oriented flooding graph .
indeed the edges verify @xmath232we call @xmath233 the succession of these three operators . in every day words ,
the operator @xmath233 does the following : each node is assigned the minimal weight of all arrows for which it is origin ; each arrow is assigned a weight equal to its extremity ; each arrow with a higher weight than its origin is suppressed . every time that we apply @xmath234 new edges are pruned and the weight distribution along the flooding paths and chains moves upwards .
for @xmath235 we obtain a graph @xmath236 which contains only flooding paths and chains of steepness @xmath237 [ [ case - of - an - initially - edge - weighted - graph ] ] case of an initially edge weighted graph + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the fig .
[ wshdex9]a represents an edge weighted graph on which the regional minima edges are indicated in red .
[ wshdex9]b presents the associated flooding graph @xmath30 .
we associate to @xmath30 an oriented graph @xmath143 by replacing the edge @xmath144 by an arrow @xmath145 if @xmath146and@xmath147by two arrows @xmath145 and @xmath148 if @xmath149 and get fig .
[ wshdex9]c .
the operator @xmath233 produces a new weight distribution for both nodes and edges in fig .
[ wshdex9]d .
furthermore two arrows are cut as they are not invariant by the opening @xmath238 as each connected component contains only one regional minimum , we can stop the pruning and we obtain the final watershed partition .
this partition is indicated in false color on top of the flooding graph @xmath30 in fig.[wshdex9]e and also in false color on top of the complete initial graph in fig.[wshdex9]f .
[ ptb ] wshdex9-eps-converted-to.pdf [ [ case - of - an - initially - node - weighted - graph ] ] case of an initially node weighted graph + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + fig.[wshdex10]a presents a node weighted graph .
it contains a number of some difficulties , like the presence of 2 plateaux and of a buttonhole at the node with weight @xmath239 in a first step we add loop edges linking each isolated regional minimum with itself as shown in fig.[wshdex10]b .
the next step produces edge weights by the dilation @xmath53 of the node weights , yielding the initial flooding graph @xmath30 in fig.[wshdex10]c the oriented graph @xmath143 is produced in fig.[wshdex10]d . with each new applications of the operator @xmath233
the graph @xmath143 is further pruned , yielding the graphs of fig.[wshdex10 ] e , f and g. fig.[wshdex10]h transports the arrows of the graph fig.[wshdex10]g onto the graph @xmath240the remaining flooding tracks are clearly visible , on a graph which has kept the initial node weights .
[ ptb ] wshdex10-eps-converted-to.pdf
we now summarize the results of this study . starting from a node or edge weighted graph , we want to construct a watershed partition , without arbitrary choices , or with a minimum number of arbitrary choices . in a first step
we derive the flooding graph .
an edge weighted graph looses the edges which are not the lowest neighboring edges of one of their extremities , but keeps the edge weights .
the node weights are obtained by the erosion @xmath241to a node weighted graph are added some loops linking each isolated regional minimum with itself .
the node weights keep their initial weights and the edge weights are obtained by the dilation @xmath242 in both cases we obtain a graph with a perfect coupling between edge and node weights ; furthermore , the edge regional minima span the node regional minima .
and the flooding chains span the flooding paths , each node being followed by an edge with the same weight .
* method 1 : * the ties are broken by a watershed algorithm applied on the flooding graph . any algorithm of the literature developed for node ( resp .
edge ) weighted graphs may now be applied to the flooding graph , even if the initial graph is an edge ( resp .
node ) weighted graph .
the algorithm of b.marcotegui et al @xcite for constructing the watershed on an edge weighted graph is derived from prim s algorithm for constructing a minimum spanning tree or forest .
this algorithm is myopic and considers only flooding chains with a lexicographic depth equal to 1 .
the algorithms proposed in @xcite are or the same myopic type . using an algorithm @xcite based on the topographic distance @xcite,@xcite , and
applied to node weighted graphs , choses flooding chains with a lexicographic depth equal to 2 , and for this reason are more selective .
such an algorithm , designed for node weighted graphs may now be applied to a graph which initially was edge weighted .
the arbitrary choices for producing a watershed partition is taken in charge by the watershed algorithms applied to the flooding graph . in the case of @xcite or @xcite , by the scheduling of the shortest distance algorithm ( ultrametric flooding distance for node weighted graphs or topographic distance for node weighted graphs ) .
these are only a few examples of algorithms which may be used , among a large number of others .
we have seen how to reduce the number of flooding tracks and paths . the flooding graph @xmath30 is transformed into an oriented graph @xmath243the operator @xmath233 prunes further this graph , again without arbitrary choices .
after @xmath157 iterations , only the flooding paths or flooding tracks of the graph @xmath30 with a lexicographic depth @xmath244 remain . we then `` transport '' the arrows of the graph @xmath236 onto the flooding graph @xmath30 and get a graph @xmath245 we keep an edge @xmath144 of @xmath30 if and only if there exists an arrow @xmath145 or @xmath148 in the graph @xmath246like that we obtain a flooding graph @xmath30 in which the only remaining flooding tracks and paths have a steepness @xmath237 a node @xmath138 belongs to two catchment basins , if there exists two flooding paths towards the corresponding minima , with identical @xmath244 first edges forming a track of @xmath244 maximal steepness
. fig.[wshdex11 ] presents 4 pruning states of the node weighted graph in fig.[wshdex10]a .
fig.[wshdex11]a presents the associated flooding graph ( without the loops on the isolated regional minima ) , where the edge weights have been added .
two types of methods may be applied to this graph .
* method 2 : * after @xmath157 steps of pruning , the graph @xmath247 has only flooding paths and chains of steepness @xmath248again any algorithm developed for node or edge weighted graph may be applied to this graph .
but as the remaining flooding paths are extremely scarce , even the loosest and most myopic algorithms will do a good job .
in particular the algorithm by b.marcotegui which normally selects flooding paths of steepness @xmath249 now selects paths of steepness @xmath250 as they are the only available .
the same is true for the algorithm based on the topographic distance , selecting flooding paths of steepness @xmath251 the remaining arbitrary choice , if any , is taken in charge by the scheduling of the shortest distance algorithm ( ultrametric flooding distance for node weighted graphs or topographic distance for node weighted graphs ) .
[ ptb ] wshdex11-eps-converted-to.pdf * method 3 : * we continue the pruning @xmath236 until there remain no arrows which are head to tail , except in the regional minima .
this is the case in fig.[wshdex10]f .
there exists a node with weight @xmath252 which is the origin of two arrows with weight @xmath253if we arbitrarily suppress one of them , we are also done and have a partition .
there are two solutions possible at this stage of pruning .
thus if a node is the origin of several arrows , one leaves only one .
( this method has often been used in hardware implementations , but without preliminary pruning @xcite,@xcite ) . for @xmath170 we obtain a maximal pruning of the flooding graph .
in fact we may stop as soon the graph is cut into a number of components containing each one regional minimum .
the node and edge weights remain identical as in the graph @xmath130 but there remains only a minimal number of flooding paths and chains .
a node will be linked to two distinct minima by two flooding chains only if there exists two paths with exactly the same weight distributions towards these two minima ; this will rarely happen in natural images .
if the minima have distinct weights , each node is linked with one and only one minimum by a flooding path of maximal steepness .
* method 4 : * we continue the pruning @xmath236 until each connected component contains only one regional minimum and we are done .
if this can not be achieved , then we resort to method 3 . after an infinite number of prunings
, @xmath254 does not contain any head and tail arrows and method 3 can be applied .
* method 5 : * we slightly change the weights of the minima so that they are all distinct .
we continue the pruning @xmath236 until each connected component contains only one regional minimum and we are done . we know that this will happen for @xmath255
we have established that node and weighted graphs represent the same topography , with the same minima and the same catchment basins .
we have presented a method to reduce the overlapping zones of the catchment basins without arbitrary choices .
we finally have presented how to introduce such arbitrary choices if needed . | watersheds have been defined both for node and edge weighted graphs .
we show that they are identical : for each edge ( resp .
node ) weighted graph exists a node ( resp .
edge ) weighted graph with the same minima and catchment basin . |
consider a typical design problem with more than one objectives ( design criteria ) .
for example , we may want to design a network that provides the maximum capacity with the minimum cost , or we may want to design a radiation therapy for a patient that maximizes the dose to the tumor and minimizes the dose to the healthy organs .
in such _ multiobjective _ ( or _ multicriteria _ ) problems there is typically no solution that optimizes simultaneously all the objectives , but rather a set of so - called _ pareto optimal _ solutions , i.e. , solutions that are not dominated by any other solution in all the objectives .
the trade - off between the different objectives is captured by the _ trade - off _ or _
pareto curve _
( surface for three or more objectives ) , the set of values of the objective functions for all the pareto optimal solutions .
multiobjective problems are prevalent in many fields , e.g. , engineering , economics , management , healthcare , etc .
there is extensive research in this area published in different fields ; see @xcite for some books and surveys . in a multiobjective problem
, we would ideally like to compute the pareto curve and present it to the decision maker to select the solution that strikes the ` right ' balance between the objectives according to his / her preferences ( and different users may prefer different points on the curve ) .
the problem is that the pareto curve has typically an enormous number of points , or even an infinite number for continuous problems ( with no closed form description ) , and thus we can not compute all of them .
we can only compute a limited number of solutions ( points ) , and of course we want the computed points to provide a good approximation to the pareto curve so that the decision maker can get a good sense of the range of possibilities in the design space .
we measure the quality of the approximation provided by a computed solution set using a ( multiplicative ) approximation ratio , as in the case of approximation algorithms for single - objective problems .
assume ( as usual in approximation algorithms ) that the objective functions take positive values .
a set of solutions @xmath0 is an _
@xmath1-pareto set _ if the members of @xmath0 approximately dominate within @xmath2 every other solution , i.e. , for every solution @xmath3 there is a solution @xmath4 such that @xmath5 is within a factor @xmath2 or better of @xmath3 in all the objectives @xcite . often , after computing a finite set @xmath0 of solutions and their corresponding points in objective space ( i.e. , their vectors of objective values ) , we ` connect the dots ' , taking in effect also the convex combinations of the solution points . in problems where the solution points
form a convex set ( examples include multiobjective flows , linear programming , convex programming ) , this convexification is justified and provides a much better approximation of the pareto curve than the original set @xmath0 of individual points .
a set @xmath0 of solutions is called an _
@xmath1-convex pareto set _ if the convex hull of the solution points corresponding to @xmath0 approximately dominates within @xmath2 all the solution points @xcite . even for applications with nonconvex solution sets ,
sometimes solutions that are dominated by convex combinations of other solutions are considered inferior , and one is interested only in solutions that are not thus dominated , i.e. , in solutions whose objective values are on the ( undominated ) boundary of the convex hull of all solution points , the so - called _ convex pareto set_. note that every instance of a multiobjective problem has a unique pareto set and a unique convex pareto set , but in general it has many different @xmath1-pareto sets and @xmath1-convex pareto sets , and furthermore these can have drastically different sizes .
it is known that for every multiobjective problem with a fixed number @xmath6 of polynomially computable objective functions , there exist an @xmath1-pareto set ( and also @xmath1-convex pareto set ) of polynomial size , in particular of size @xmath7 , where @xmath8 is the bit complexity of the objective functions ( i.e. , the functions take values in the range @xmath9 $ ] ) @xcite . whether such approximate sets can be constructed in polynomial time is another matter : necessary and sufficient conditions for polynomial time constructibility of @xmath1-pareto and @xmath1-convex pareto sets are given respectively in @xcite . the most common approach to the generation of pareto points ( called weighted - sum method ) is to give nonnegative weights @xmath10 to the different objective functions @xmath11 ( assume for simplicity that they are all minimization objectives ) and then optimize the linear combining function @xmath12 ; this approach assumes availability of a subroutine @xmath13 that optimizes such linear combinations of the objectives . for any set of nonnegative weights ,
the optimal solution is clearly in the pareto set , actually in the convex pareto set .
in fact , the convex pareto set is precisely the set of optimal solutions for _ all _ possible such weighted linear combinations of the objectives .
of course , we can not try all possible weights ; we must select carefully a finite set of weights , so that the resulting set of solutions provides a good approximation , i.e. , is an @xmath1-convex pareto set for a desired small @xmath1 .
it is shown in @xcite that a necessary and sufficient condition for the polynomial time constructibility of an @xmath1-convex pareto set is the availability of a polynomial time @xmath13 routine for the approximate optimization of nonnegative linear combinations . in a typical multiobjective problem ,
the @xmath13 routine is a nontrivial piece of software , each call takes a substantial amount of time , thus we want to make the best use of the calls to achieve as good a representation of the solution space as possible .
ideally , we would like to achieve the smallest possible approximation error @xmath1 with the fewest number of calls to the @xmath13 routine .
that is , given @xmath14 , compute an @xmath1-convex pareto set for the instance at hand using _ as few comb calls as possible_. ( in @xcite we study a different cost metric ; we explain the difference at the end of this section , and we note that both metrics are relevant for different aspects of decision making in multiobjective problems . ) we measure the performance of an algorithm by the ratio of its cost , i.e. , the number of calls that it makes , to the minimum possible number required for the instance at hand ( as is usual in approximation algorithms ) .
let @xmath15 be the number of points in the smallest @xmath1-convex pareto set for an instance @xmath16 . clearly , every algorithm that computes an @xmath1-convex pareto set needs to make at least @xmath15 calls .
the _ performance ( competitive ) ratio _ of an algorithm @xmath17 that computes an @xmath1-convex pareto set using @xmath18 calls for each instance @xmath16 is @xmath19 .
an important point that should be stressed here is that , as in the case of online algorithms , the algorithm does not have complete information about the input , i.e. , the ( convex ) pareto curve is _ not _ given explicitly , but can only be accessed indirectly through calls to the @xmath13 routine ; in fact , the whole purpose of the algorithm is to obtain an approximate knowledge of the curve . in this paper
we investigate the performance of the _ chord _ algorithm , a simple , natural greedy algorithm for the approximate construction of the pareto set .
the algorithm and variants of it have been used often , under various names , for multiobjective problems @xcite as well as several other types of applications involving the approximation of curves , which we will describe later on .
we focus on the bi - objective case ; although the algorithm can be defined ( and has been used ) for more objectives , most of the literature concerns the bi - objective case , which is already rich enough , and covers also most of the common uses of the algorithm .
we now briefly describe the algorithm .
let @xmath20 , @xmath21 be the two objectives ( say minimization objectives for concreteness ) , and let @xmath22 be the ( unknown ) convex pareto curve .
first optimize @xmath20 , and @xmath21 separately ( i.e. , call @xmath13 for the weight tuples @xmath23 and @xmath24 ) to compute the leftmost and rightmost points @xmath25 of the curve @xmath22 .
the segment @xmath26 is a first approximation to @xmath22 ; its quality is determined by a point @xmath27 that is least well covered by the segment .
it is easy to see that this worst point @xmath28 is a point of the pareto curve @xmath22 that minimizes the linear combination @xmath29 , where @xmath30 is the absolute value of the slope of @xmath26 , i.e. , it is a point of @xmath22 with a supporting line parallel to the ` chord ' @xmath26 .
compute such a worst point @xmath28 ; if the error is at most @xmath1 , then terminate , otherwise add @xmath28 to the set @xmath0 to form an approximate set @xmath31 and recurse on the two intervals @xmath32 and @xmath33 . in section [ sec : prelim ]
we give a more detailed formal description ( for example , in some cases one can determine from previous information that the maximum possible error in an interval is upper bounded by @xmath1 and do not need to call @xmath13 ) .
the algorithm is quite natural , it has been often reinvented and is commonly used for a number of other purposes .
as pointed out in @xcite , an early application was by archimedes who used it to approximate a parabola for area estimation @xcite .
in the area of _ parametric optimization _ , the algorithm is known as the eisner - severance " method after @xcite .
note that parametric optimization is closely related to bi - objective optimization .
for example , in the parametric shortest path problem , each edge @xmath34 has cost @xmath35 that depends on a parameter @xmath36 .
the length of the shortest path is a piecewise linear function of @xmath36 whose pieces correspond to the vertices of the convex pareto curve for the bi - objective shortest path problem with cost vectors @xmath37 on the edges .
a call to the @xmath13 routine for the bi - objective problem corresponds to solving the parametric problem for a particular value of the parameter .
the chord algorithm is also useful for the approximation of convex functions , and for the approximation and smoothening of convex and even non - convex curves . in the case of functions ,
an appropriate measure of distance between the function @xmath38 and the approximation is the vertical distance , while for curves a natural measure is the hausdorff distance ; note that for a given curve and approximating segment @xmath39 , the same point @xmath28 of the curve with supporting line parallel to @xmath39 maximizes also the above distances . in the context of smoothening and compressing curves and polygonal lines for graphics and related applications , the chord algorithm is known as the ramer - douglas - peucker algorithm , after @xcite who independently proposed it .
previous work has analyzed the chord algorithm ( and variants ) for achieving an @xmath1-approximation of a function or curve with respect to vertical and hausdorff distance , and proved bounds on the cost of the algorithm as a function of @xmath1 : for all convex curves of length @xmath40 , ( under some technical conditions on the derivatives ) the algorithm uses at most @xmath41 calls to construct an @xmath1-approximation , and there are curves ( for example , a portion of a circle ) that require @xmath42 calls @xcite .
note however that these results do not tell us what the performance ratio is , because for many instances , the optimal cost @xmath43 may be much smaller than @xmath44 , perhaps even a constant . for example
, if @xmath22 is a convex polygonal line with few vertices , then the chord algorithm will perform very well for @xmath45 ; in fact , as shown by @xcite in the context of parametric optimization , if there are @xmath46 breakpoints , then the algorithm will compute the exact curve after @xmath47 calls .
( the problem is of course that in most bi - objective and parametric problems , the number @xmath46 of vertices is huge , or even infinite for continuous problems , and thus we have to approximate . ) in this paper we provide sharp upper and lower bounds on the performance ( competitive ) ratio of the chord algorithm , both in the worst case and in the average case setting . consider a bi - objective problem where the objective functions take values in @xmath9 $ ] .
we prove that the worst - case performance ratio of the chord algorithm for computing an @xmath1-convex pareto set is @xmath48 .
the upper bound implies in particular that for problems with polynomially computable objective functions and a polynomial - time ( exact or approximate ) comb routine , the chord algorithm runs in polynomial time in the input size and @xmath49 .
we show furthermore that there is no algorithm with constant performance ratio . in particular ,
every algorithm ( even randomized ) has performance ratio at least @xmath50 .
similar results hold for the approximation of convex curves with respect to the hausdorff distance .
that is , the performance ratio of the chord algorithm for approximating a convex curve of length @xmath40 within hausdorff distance @xmath1 , is @xmath51 .
furthermore , every algorithm has worst - case performance ratio at least @xmath52 .
we also analyze the expected performance of the chord algorithm for some natural probability distributions .
given that the algorithm is used in practice in various contexts with good performance , and since worst - case instances are often pathological and extreme , it is interesting to analyze the average case performance of the algorithm .
indeed , we show that the performance on the average is exponentially better . note that chord is a simple natural greedy algorithm , and is not tuned to any particular distribution .
we consider instances generated by a class of product distributions that are `` approximately '' uniform and prove that the expected performance ratio of the chord algorithm is @xmath53 ( upper and lower bound ) . again
similar results hold for the hausdorff distance
. * related work .
* there is extensive work on multiobjective optimization , as well as on approximation of curves in various contexts .
we have discussed already the main related references .
the problem addressed by the chord algorithm fits within the general framework of determining the shape by probing @xcite .
most of the work in this area concerns the exact reconstruction , and the analytical works on approximation ( e.g. , @xcite ) compute only the worst - case cost of the algorithm in terms of @xmath1 ( showing bounds of the form @xmath54 ) .
there does not seem to be any prior work comparing the cost of the algorithm to the optimal cost for the instance at hand , i.e. , the approximation ratio , which is the usual way of measuring the performance of approximation algorithms .
the closest work in multiobjective optimization is our prior work @xcite on the approximation of convex pareto curves using a different cost metric .
both of the metrics are important and reflect different aspects of the use of the approximation in the decision making process .
consider a problem , say with two objectives , suppose we make several calls , say @xmath46 , to the @xmath13 routine , compute a number of solution points , connect them and present the resulting curve to the decision maker to visualize the range of possibilities , i.e. , get an idea of the true convex pareto curve .
( the process may not end there , e.g. , the decision maker may narrow the range of interest , followed by computation of a better approximation for the narrower range , and so forth ) . in this scenario ,
we want to achieve as small an error @xmath1 as possible , using as small a number @xmath46 of calls as we can , ideally , as close as possible to the minimum number @xmath15 that is absolutely needed for the instance . in this
setting , the cost of the algorithm is measured by the number of calls ( i.e. , the computational effort ) ; this is the cost metric that we study in this paper , and the performance ratio is as usual the ratio of the cost to the optimum cost .
consider now a scenario where the decision maker does not just inspect visually the curve , but will look more closely at a set of solutions to select one ; for instance a physician in the radiotherapy example will consider carefully a small number of possible treatments in detail to decide which one to follow . since human time is much more limited than computational time ( and more valuable , even small constant factors matter a lot ) , the primary metric in this scenario is the number @xmath55 of selected solutions that is presented to the decision maker for closer investigation ( we want @xmath55 to be as close as possible to @xmath15 ) , while the computational time , i.e. , the number @xmath46 of calls , is less important and can be much larger ( as long as it is feasible of course ) .
this second cost metric ( the size @xmath55 of the selected set ) is studied in @xcite for the convex pareto curve ( and in @xcite in the nonconvex case ) . among other results ,
it is shown there that for all bi - objective problems with an exact @xmath13 routine and a continuous convex space , an optimal @xmath1-convex pareto set ( i.e. , one with @xmath15 solutions ) can be computed in polynomial time using @xmath56 calls to @xmath13 in general , ( though more efficient algorithms are obtained for specific important problems such as bi - objective lp ) .
for discrete problems , a @xmath57-approximation to the minimum size can be obtained in polynomial time , and the factor @xmath57 is inherent . as remarked above , both cost metrics are important for different stages of the decision making .
recall also that , as noted earlier , the chord algorithm runs in polynomial time , and furthermore , one can show that its solution set can be post - processed to select a subset that is a @xmath1-convex pareto set of size at most @xmath58 .
* structure of the paper . * the rest of the paper is organized as follows : section [ sec : prelim ] describes the model and states our main results , section [ sec : worst ] concerns the worst - case analysis , and section [ sec : average ] the average - case analysis .
section [ sec : concl ] concludes the paper and suggests the most relevant future research directions .
this section is structured as follows : after giving basic notation , in section [ ssec : defs ] we describe the relevant definitions and framework from multiobjective optimization . in section [ ssec : chord - description ] we provide a formal description of the chord algorithm in tandem with an intuitive explanation .
finally , in section [ ssec : results ] we state our results on the performance of the algorithm as well as our general lower bounds .
* notation .
* we start by introducing some notation used throughout the paper . for @xmath59
, we will denote @xmath60 : = \{1,2,\ldots , n\}$ ] and @xmath61 : = \ { i , i+1 , \ldots , j \}$ ] . for @xmath62 , we denote by @xmath63 the line segment with endpoints @xmath64 and @xmath28 , @xmath65 denotes its length , @xmath66 is the triangle defined by @xmath67 and @xmath68 is the internal angle of @xmath66 formed by @xmath63 and @xmath69
. we will use @xmath70 and @xmath71 as the two coordinates of the plane . if @xmath64 is a point on the plane , we use @xmath72 and @xmath73 to denote its coordinates ; that is , @xmath74 .
we will typically use the symbol @xmath36 to denote the ( absolute value of the ) slope of a line , unless otherwise specified .
sometimes we will add an appropriate subscript , i.e. , we will use @xmath75 to denote the slope of the line defined by @xmath64 and @xmath28 . for a lebesgue measurable set @xmath76
we will denote its area by @xmath77 .
we describe the general framework of a bi - objective problem @xmath78 to which our results are applicable .
a bi - objective optimization problem has a set of valid instances , and every instance has an associated set of feasible solutions , usually called the solution or decision space .
there are two objective functions , each of which maps every instance
solution pair to a real number .
the problem specifies for each objective whether it is to be maximized or minimized .
consider the plane whose coordinates correspond to the two objectives .
every solution is mapped to a point on this plane .
we denote the objective functions by @xmath70 and @xmath71 and we use @xmath79 to denote the objective space ( i.e. , the set of @xmath57-vectors of objective values of the feasible solutions for the given instance ) . as usual in approximation , we assume that the objective functions are polynomial time computable and take values in @xmath80 $ ] , i.e. , @xmath81 ^ 2 $ ] , where @xmath8 is polynomially bounded in the size of the input . note that this framework covers all discrete combinatorial optimization problems of interest ( e.g. , shortest paths , spanning tree , matching , etc ) , but also contains many continuous problems ( e.g. , linear and convex programs , etc ) . throughout this paper we assume , for the sake of concreteness , that both objective functions @xmath70 and @xmath71 are to be minimized .
all our results hold also for the case of maximization or mixed objectives .
let @xmath82 .
we say that @xmath64 _ dominates _
@xmath28 if @xmath83 ( coordinate - wise ) .
we say that @xmath64 _
@xmath28 ( @xmath85 ) if @xmath86 .
let @xmath87 .
the _ pareto set _ of @xmath88 , denoted by @xmath89 , is the subset of undominated points in @xmath88 ( i.e. , @xmath90 iff @xmath91 and no other point in @xmath88 dominates @xmath64 ) .
the convex pareto set of a @xmath88 , denoted by @xmath92 , is the minimum subset of @xmath88 whose convex combinations dominate ( every point in ) @xmath88 .
we also use the term _ lower envelope _ of @xmath88 to denote the pareto set of its convex hull , i.e. , @xmath93 .
in particular , if @xmath88 is convex its lower envelope is identified with its pareto set .
if @xmath88 is finite , its lower envelope is a convex polygonal chain with vertices the points of @xmath92 .
note that , for any set @xmath88 , the lower envelope @xmath94 is a convex and monotone decreasing planar curve .
for @xmath95 we will denote by @xmath96 the subset of @xmath94 with endpoints @xmath64 , @xmath28
. an _ @xmath84-convex pareto set _ of @xmath88 ( henceforth @xmath84-cp ) is a subset @xmath97 of @xmath88 whose convex combinations @xmath84-cover ( every point in ) @xmath88 .
note that such a set need not contain points dominated by convex combinations of other points , as they are redundant .
if a set contains no redundant points , we call it non - redundant .
let @xmath98 , @xmath99 and @xmath100 , be a non - redundant set . by definition
, @xmath0 is an @xmath84-cp for @xmath88 if and only if the polygonal chain @xmath101 @xmath84-covers @xmath92 .
we define the _ ratio distance _ from a point @xmath64 to a point @xmath28 as @xmath102 .
( note the asymmetry in the definition . ) intuitively , it is the minimum value of @xmath103 such that @xmath28 @xmath84-covers @xmath64 .
we also define the ratio distance between sets of points . if @xmath104 , then @xmath105 . as a corollary of this definition ,
the set @xmath106 is an @xmath84-cp for @xmath88 if and only if @xmath107 .
the above definitions apply to any set @xmath87 .
let @xmath78 be a bi - objective optimization problem in the aforementioned framework .
for an instance of @xmath78 , the set @xmath88 corresponds to the objective space @xmath79 ( for the given instance ) .
we do not assume that the objective space @xmath79 is convex ; it may well be discrete or a continuous non - convex set .
it should be stressed that the objective space is not given explicitly , but rather implicitly through the instance . in particular , we access the objective space @xmath79 of @xmath78 via an oracle @xmath108 that ( exactly or approximately ) minimizes non - negative linear combinations @xmath109 of the objectives . that is , the oracle takes as input a parameter @xmath110 and outputs a point @xmath111 ( i.e. , a feasible point ) that ( exactly or approximately ) minimizes the combined objective function @xmath112 .
we use the convention that , for @xmath113 , the oracle minimizes the @xmath70 objective . formally , for @xmath114 , we denote by @xmath115 the problem of optimizing the combined objective @xmath116 over @xmath79 .
let @xmath117 be an accuracy parameter .
then , for @xmath114 , we will denote by @xmath118 a routine that returns a point @xmath111 that optimizes @xmath119 up to a factor of @xmath120 , i.e. , @xmath121 .
in other words , the @xmath122 routine is an `` approximate optimization oracle '' for the objective space @xmath79 .
we say that the @xmath108 problem has a polynomial time approximation scheme ( ptas ) , if for any instance of @xmath78 and any @xmath123 there exists a routine @xmath118 ( as specified above ) that runs in time polynomial in the size of the instance .
as shown in @xcite , for any bi - objective problem in the aforementioned framework , there is a ptas for constructing an @xmath84-convex pareto set if and only if there is a ptas for the @xmath108 problem .
we now provide a geometric characterization of @xmath118 that will be crucial throughout this paper .
consider the point @xmath111 returned by @xmath118 and the corresponding line @xmath124 through @xmath28 with slope @xmath125 , i.e. , @xmath126 .
then there exists no solution point ( i.e. , no point in @xmath79 ) below the line @xmath127 .
geometrically , this means that we sweep a line of absolute slope @xmath36 , until it touches ( exactly or approximately ) the undominated boundary ( lower envelope ) of the objective space @xmath79 . for @xmath128 ,
the routine returns a point on the lower envelope @xmath129 , while for @xmath123 it returns a ( potentially ) dominated point of @xmath79 `` close '' to the boundary ( where the notion of `` closeness '' is quantitatively defined by the aforementioned condition ) .
see figure [ fig : comb ] for an illustration .
if @xmath28 is the ( feasible ) point in @xmath79 returned by @xmath118 , then we write @xmath130 .
we assume that either @xmath128 ( i.e. , we have an exact routine ) , or we have a ptas . for @xmath128 , i.e. , when the optimization is exact , we omit the subscript and denote the comb routine simply by @xmath115
. we will denote by @xmath131 the size of an optimum @xmath84-convex pareto set for the given instance , i.e. , an @xmath84-convex pareto set with the minimum number of points .
note that , obviously every algorithm that constructs an @xmath84-cp , must certainly make at the very least @xmath131 calls to @xmath108 , just to get @xmath131 points which are needed at a minimum to form an @xmath84-cp ; this holds even if the algorithm somehow manages to always be lucky and call @xmath108 with the right values of @xmath36 that identify the points of an optimal @xmath84-cp .
having obtained the points of an optimal @xmath84-cp , another @xmath131 many calls to comb with the slopes of the edges of the polygonal line defined by the points , suffice to verify that the points form an @xmath84-cp .
hence , the `` offline '' optimum number of calls is at most @xmath132 .
let @xmath133 be the number of @xmath108 calls required by the chord algorithm on instance @xmath79 .
the _ worst - case performance ratio _ of the algorithm is defined to be @xmath134 .
if the inputs are drawn from some probability distribution @xmath135 , then we will use the _ expected performance ratio _
@xmath136 $ ] as a measure ( note that we shall omit the subscript `` @xmath137 '' when the underlying distribution over instances will be clear from the context ) .
while the main focus of this paper is on approximation of multiobjective optimization problems , our analysis also applies ( with minor modifications ) to related settings ( in which the the chord algorithm has been used ) .
consider for example the following classical _ curve simplification _ problem : given a convex curve @xmath138 of length ( at most @xmath139 on the plane , find the minimum number of points on @xmath138 so that the corresponding convex polygonal chain @xmath140 approximates the curve @xmath138 within distance error @xmath84 .
a popular distance measure in this setting is the
_ hausdorff distance _ of @xmath140 from @xmath138 , i.e. , the maximum euclidean distance of a point in the actual curve from the approximating curve .
note that the hausdorff distance is invariant under translation , while our ratio distance is invariant under scaling .
we remark that our upper and lower bounds for the performance of the chord algorithm wrt the ratio distance apply with minor modifications for the hausdorff distance .
this can be seen as follows : a curve of length @xmath40 located anywhere on the plane can be scaled down by @xmath40 and translated to the unit square @xmath141 $ ] . by definition , approximating the original curve within hausdorff distance @xmath84 is equivalent to approximating the new curve within hausdorff distance error @xmath142 .
a simple calculation shows that for a convex curve in the unit square @xmath143 $ ] the two metrics ( hausdorff and ratio distance ) are within a constant factor of each other .
hence , the upper and lower bounds on the approximation wrt ratio distance give immediately corresponding bounds wrt hausdorff .
we also define the horizontal distance .
we use this distance as an intermediate tool for our lower bound construction in section [ ssec : worst - lower ] .
the _ horizontal distance _ from a point @xmath64 to a point @xmath28 is defined @xmath144 .
the horizontal distance from @xmath64 to a line @xmath145 ( that is not horizontal ) is @xmath146 , where @xmath147 is the @xmath71-projection of @xmath64 on @xmath145 ( i.e. , the point that a horizontal line from @xmath64 intersects @xmath145 ) . _
all the upper bounds of this paper on the performance of the chord algorithm hold under the assumption that we have a ptas for the comb problem . on the other hand ,
our lower bounds apply even for the special case that an exact routine is available . for the clarity of the exposition
, we describe the chord algorithm and prove our upper bounds for the case of an exact comb routine .
we then describe the simple modifications in the algorithm and analysis for the case of an approximate routine . _
we have set the stage to formally describe the algorithm .
let @xmath78 be a bi - objective problem with an efficient exact @xmath108 routine . given @xmath148 and an instance @xmath79 of @xmath78 ( implicitly via @xmath108 ) , we would like to construct an @xmath84-cp for @xmath79 using as few calls to @xmath108 as possible . as mentioned in the introduction , a popular algorithm for this purpose
is the chord algorithm that is the main object of study in this paper . in table [
table : chord ] below we describe the algorithm in detailed pseudo - code .
the pseudo - code corresponds exactly to the description of the algorithm in the introduction .
the ( recursively defined ) routine chord is called from the main algorithm and returns a set of points @xmath149 that is an @xmath84-cp for @xmath79 .
the recursive description of the algorithm is quite natural and will be useful in the analysis .
p3inp3.8 in * chord algorithm * ( _ input : _
@xmath79 , @xmath84 ) + @xmath150 ; + @xmath151 ; + @xmath152 ; + * return * @xmath153 . & @xmath154 + * if * @xmath155 * return * @xmath156 ; + @xmath157 ; @xmath158 ; + * if * @xmath159 * return * @xmath156 ; + @xmath160 line parallel to @xmath161 through @xmath28 ; + @xmath162 ; @xmath163 ; + @xmath164 ; @xmath165 ; + * return * @xmath166 .
let us proceed to explain the notation used in the pseudo - code in tandem with some intuitive understanding of the algorithm .
first , the feasible points @xmath167 minimize the @xmath70-objective and @xmath71-objective respectively .
( note that these points may be dominated , i.e. , are not necessarily the extreme points of @xmath168 ; however , this does not affect our analysis . ) by monotonicity and convexity , the lower envelope is contained in the right triangle @xmath169 , i.e. , @xmath170 .
( note that the point @xmath171 is not a feasible point , but is solely defined for the purpose of `` sandwiching '' the lower envelope . )
the chord routine takes as input ( i ) the desired error tolerance @xmath84 , and ( ii ) the ordered @xmath172-set of points @xmath173 . in every recursive call of the chord routine ,
the points @xmath174 ( left ) and @xmath175 ( right ) are ( feasible ) points of the lower envelope , i.e. , @xmath176 .
moreover , the point @xmath3 is a ( not necessarily feasible ) point and the following conditions are satisfied : * the point @xmath3 lies to the right of @xmath174 , to the left of @xmath175 and below the line segment @xmath161 .
in particular , this implies that the triangle @xmath177 is either right or obtuse , i.e. , @xmath178 .
* the subset of the lower envelope ( convex curve ) with endpoints @xmath174 and @xmath175 is contained in @xmath177 , i.e. , @xmath179 .
see figure [ fig : chord ] for an illustration .
the red curve represents the lower envelope between the points @xmath174 and @xmath175 , i.e. , the unknown curve we want to approximate .
( note that the feasible points @xmath180 are the results of previous recursive calls . )
the point @xmath158 is the feasible point in @xmath129 computed in the current iteration ( recursive call ) .
we remark that this point is at maximum ratio distance from the `` chord '' @xmath161 among all points of @xmath181 .
the chord routine will recurse on the triangles @xmath182 and @xmath183 . note that , by construction , the line @xmath184 is parallel to @xmath161 . during the execution of the algorithm ,
we `` learn '' the objective space in an `` online '' fashion .
after a number of iterations , we have obtained information that imposes an `` upper '' and a `` lower '' approximation to @xmath129 . in particular , the computed solution points define a polygonal chain that is an upper approximation to @xmath129 and the supporting lines at these points define a lower approximation . as the number of iterations increases , these bounds become more and more refined , hence we obtain a better approximation to the curve .
consider for example figure [ fig : chord ] . before the current iteration of the chord routine ,
the only information available to the algorithm is that the lower envelope ( between points @xmath174 and @xmath175 ) lies in @xmath177 , i.e. , @xmath161 is an upper approximation and the polygonal chain @xmath185 is a lower approximation . given the information the algorithm has at this stage , the potential error of this initial approximation is the ratio distance @xmath186 .
( if this distance is at most @xmath84 , then the segment @xmath161 @xmath84-covers the subset of the lower envelope between @xmath187 and the routine terminates .
otherwise , the point @xmath28 is computed and the approximation is refined , if necessary , as explained above . ) after the current iteration , the upper approximation is refined to be @xmath188 and the lower approximation is @xmath189 . the potential error
given the available information now is @xmath190 . by applying these arguments recursively
, we get that the chord algorithm always terminates and , upon termination , it outputs an @xmath84-cp for @xmath79 ( see lemma [ lem : chord - finds - eps - convex ] for a rigorous proof ) .
consider the recursion tree built by the chord algorithm .
every node of the tree corresponds to a triangle @xmath177 ( input to the chord routine at the corresponding recursive call ) . in the analysis
, we shall use the following convention : there is no node in the recursion tree if , at the corresponding step , the routine terminates without calling @xmath108 ( i.e. , if @xmath155 in the aforementioned description ) .
the pseudo - code of table [ table : chord ] is specialized for the ratio distance , but one may use other metrics based on the application . in the context of convex curve simplification ,
our upper and lower bounds for the chord algorithm also apply for the hausdorff distance ( i.e. the maximum euclidean distance of a point in the actual curve from the approximate curve ) .
we are now ready to state our main results .
our first main result is an analysis of the chord algorithm on worst - case instances that is tight up to constant factors . in particular , for the ratio distance we prove [ thm : worst ] the worst - case performance ratio of the chord algorithm ( wrt the ratio distance ) is @xmath191 the lower bound on the performance of the chord algorithm is proved in section [ ssec : worst - lower ] . in section [ ssec : worst - lower - gen ] , we also prove a general lower bound of @xmath192 on the performance ratio of _ any _ algorithm in the @xmath108 based model , as well as lower bounds for approximating with respect to the horizontal ( resp .
vertical ) distance . in section [ ssec : worst - upper ]
we give a proof of the upper bound . in section [ sssec : worst - upper - simple ]
we start by presenting the slightly weaker upper bound of @xmath193 ; this result has the advantage that its proof is simple and intuitive .
the proof of the asymptotically tight upper bound requires a more careful analysis and is presented in section [ sssec : worst - upper - final ] .
_ it turns out that the hausdorff distance behaves very similarly to the ratio distance .
in particular , by essentially identical proofs , it follows that the performance ratio of the chord algorithm for approximating a convex curve of length @xmath40 within hausdorff distance @xmath1 , is @xmath51 .
furthermore , every algorithm has worst - case performance ratio at least @xmath52 . _ in the process , we also analyze the chord algorithm with respect to the horizontal distance metric ( or by symmetry the vertical distance ) . we show that in this setting the performance ratio of the algorithm is unbounded . in fact , we can get a strong lower bound in this case : _ any _ algorithm with oracle access to @xmath108 has an unbounded performance ratio ( theorem [ thm : hd - lb - gen - ws ] ) , even on instances that lie in the unit square and the desired accuracy is a constant .
our second main result is an asymptotically tight analysis of the chord algorithm in an average case setting ( wrt the ratio distance ) .
our random instances are drawn from standard distributions that have been widely used for the average case analysis of geometric algorithms in a variety of settings .
in particular , we consider ( i ) a poisson point process on the plane and ( ii ) @xmath55 points drawn from an `` un - concentrated '' product distribution .
we now formally define these distributions .
[ def : ppp ] a ( spatial , homogeneous ) _ poisson point process ( ppp ) _ of intensity @xmath194 on a bounded subset @xmath195 is a collection of random variables @xmath196 ( representing the number of points occurring in every subset of @xmath197 ) , such that : ( i ) for any lebesgue measurable @xmath88 , @xmath198 is a poisson random variable with parameter @xmath199 ; ( ii ) for any collection of _ disjoint _ subsets @xmath200 the random variables @xmath201 \}$ ] are mutually independent .
[ def : delta - balanced distn ] let @xmath88 be a bounded lebesgue - measurable subset of @xmath202 , and let @xmath203 be a distribution over @xmath88 .
the distribution @xmath203 is called @xmath204-_balanced _ , @xmath205 , if for all lebesgue measurable subsets @xmath206 , @xmath207 $ ] , where @xmath208 is the uniform distribution over @xmath88 .
we assume that @xmath204 is an absolute constant and we omit the dependence on @xmath204 in the performance ratio below .
we prove : [ thm : average ] for the aforementioned classes of random instances , the expected performance ratio of the chord algorithm ( wrt to the ratio distance ) is @xmath209 . the upper bound proof is given in section [ ssec : average - upper ] and the lower bound one in section [ ssec : average - lower ] .
we first present detailed proofs for the case of ppp and then present the ( more involved ) case of product distributions .
( we note that similar results apply also for approximation under the hausdorff distance . )
in section [ ssec : worst - lower - chord ] we prove a tight lower bound for the chord algorithm for the ratio distance metric . in section [ ssec : worst - lower - gen ]
we show our general lower bounds , both for the ratio distance and the horizontal distance .
our main result in this section is a proof of the lower bound statement in theorem [ thm : worst ] .
in fact , we show a stronger statement that also rules out the possibility of constant factor bi - criteria approximations , i.e. , it applies even if the chord algorithm is allowed ( ratio distance ) error @xmath210 and we compare it against the optimal @xmath84-cp set .
[ thm : rd - lb - ws ] let @xmath211 be an absolute constant .
let @xmath148 be smaller than a sufficiently small constant and @xmath212 be large enough .
there exists an instance @xmath213 such that @xmath214 and @xmath215 .
the lower bound applies even if an exact @xmath108 routine is available , hence we restrict ourselves to this case .
before we proceed with the formal proof , we give an explanation of our construction for the case @xmath216 and @xmath217 .
the rough intuition is that the algorithm can perform poorly when the input instance is `` skewed '' , i.e. , we have a triangle @xmath218 where @xmath219 . for such instances one can force the algorithm to select many `` redundant '' points ( hence , perform many calls to @xmath108 ) to obtain a certificate it has found an @xmath84-cp set , even when few points ( calls ) suffice .
for the special case under consideration , the `` hard '' instance has endpoints @xmath220 and @xmath221 , where @xmath84 is sufficiently small ( to be specified later ) .
initially , the only available information to the algorithm is that the convex pareto set for the given instance lies in the right triangle @xmath222 , where @xmath223 . observe that , for an instance with these endpoints , the initial error @xmath224 is roughly equal to @xmath225 and one intermediate point @xmath226 together with the rightmost point of the curve always suffice to form an @xmath84-cp set , i.e. , the optimal size is no more than @xmath227 our construction will define a sequence of points @xmath228 ( ordered in increasing @xmath70-coordinate ) which will form the convex pareto set for the corresponding instance and that force the chord algorithm to select _ all _ the @xmath229 s ( in order of increasing @xmath230 ) , until it finds @xmath231 .
that is , the chord algorithm will monotonically converge to the optimal point by visiting all the vertices of the instance in order ( in increasing @xmath70-coordinate ) .
let @xmath232 be the slope of @xmath39 .
the algorithm starts by calling @xmath233 to find a solution point at maximum ratio distance from the chord @xmath39 .
our construction guarantees that @xmath234 .
that is , if @xmath235 is the line parallel to @xmath39 through @xmath236 , @xmath235 supports the objective space .
the point @xmath236 is selected on the line segment @xmath237 so that @xmath238 , i.e. , it is obtained by subdividing ( the length of ) @xmath237 geometrically with ratio @xmath239 for an appropriate @xmath239 ( to be specified next ) . consider the point @xmath240 .
the error of the approximation @xmath241 equals @xmath242 ( note that the error to the left of @xmath236 is @xmath243 ) . if @xmath244 , we are already done , since @xmath245 , which implies @xmath246 . on the other hand , if @xmath247 , we are also done since @xmath248 , hence @xmath246 . if @xmath249 , it is not hard to show that @xmath250 hence , after the first call to @xmath108 , the error has decreased by an additive of @xmath251 and the algorithm will recurse on the triangle @xmath252 .
note that @xmath253 .
the algorithm proceeds by calling @xmath254 and this call will return the point @xmath255 .
let @xmath256 be the ( supporting ) line parallel to @xmath257 through @xmath255 .
similarly , @xmath256 supports the objective space .
the point @xmath255 is selected by repeating our `` geometric subdivision trick . '' recall that there are no feasible points below the line @xmath258 .
let @xmath259 be the projection of @xmath255 on @xmath237 .
we select @xmath255 on the segment @xmath258 so that @xmath260 .
the error of the approximation @xmath261 equals @xmath262 , where @xmath263 .
if @xmath264 , we have that @xmath265 i.e. , after the second step of the algorithm , the error has decreased by another additive @xmath266 and the algorithm will recurse on the triangle @xmath267 .
we can repeat this process iteratively , where ( roughly ) in step @xmath230 we select @xmath229 on the line @xmath268 , so that the length of the projection satisfies @xmath269 .
this iterative process can continue as long as @xmath270 . also note that the number @xmath271 of iterations can not be more than @xmath272 because @xmath273 and @xmath274 . since , @xmath275 it turns out that the optimal choice of parameters is @xmath276 .
we stress that the actual construction is more elaborate than the one presented in the intuitive explanation above . also , to show a bi - criterion lower bound , we need to add one more point @xmath277 so that the chord algorithm selects @xmath228 until it forms a @xmath278-cp , while the point @xmath277 ( along with the rightmost point @xmath279 ) suffice to form an @xmath84-cp .
the formal proof comes in two steps .
we first analyze the chord algorithm with respect to the horizontal distance metric .
we show that the performance ratio of the algorithm is unbounded in this setting ( this statement also holds for the vertical distance by symmetry ) .
in particular , for every @xmath280 , there exists an instance @xmath281 ( lying entirely in the unit square ) so that the chord algorithm ( applied for additive error @xmath282 ) has performance ratio @xmath239 .
we then show that , for an appropriate setting of the parameters in @xmath283 , we obtain the instance @xmath284 that yields the desired lower bound with respect to the ratio distance .
_ * step 1 : * _ the instance @xmath285 lies in the triangle @xmath286 , where @xmath287 , @xmath288 and @xmath223 .
the points @xmath289 and @xmath279 are ( the extreme ) vertices of the convex pareto set .
we introduce two additional parameters . the first one , @xmath280 , is the ratio used in the construction to geometrically subdivide the length of the line @xmath237 in every iteration .
the second one , @xmath290 with @xmath291 $ ] , is the number of iterations and equals the number of vertices in the instance .
we define a set of points @xmath292 ordered in increasing @xmath70coordinate and decreasing @xmath71coordinate
. our instance will be the convex polygonal line with vertices the points of @xmath293 .
we set @xmath294 and @xmath295 .
the set of points @xmath296 is defined recursively as follows : 1 .
the point @xmath236 has @xmath297 and @xmath298 .
2 . for @xmath299 $ ]
the point @xmath229 is defined as follows : let @xmath300 denote the line parallel to @xmath301 through @xmath302 .
the point @xmath229 is the point of this line with @xmath303 .
the algorithm is applied on this instance with desired horizontal distance error @xmath304 also denote @xmath305 note that @xmath306 .
see figures [ chord - fig - hd1 ] and [ chord - fig - hd2 ] for a graphic illustration of the worst - case instances for the chord algorithm .
we would like to stress that the figures are not drawn to scale .
in particular , in the figures below we have @xmath307 , while the actual lower bound for the ratio distance applies for @xmath308 ; in particular , for @xmath309 and @xmath310 . we show the following : [ claim : chd - hd ] the chord algorithm applied to the instance @xmath281 and error bound @xmath311 wrt horizontal distance selects the sequence of points @xmath312 , while the set @xmath313 attains error @xmath306 . for
@xmath314 $ ] , let @xmath315 be the intersection of the line @xmath316 the line parallel to @xmath317 through @xmath229with @xmath318 .
the error of @xmath313 is exactly @xmath319 .
observe that @xmath320 . by a simple geometric argument we obtain @xmath321 , which yields the second statement .
for the first statement , we show inductively that the recursion tree built by the algorithm for @xmath281 is a path of length @xmath322 and at depth @xmath323 , for @xmath324 $ ] , the chord subroutine selects point @xmath229 .
the proof amounts to noting that the error of the approximation @xmath325 is @xmath326 , which is @xmath327 for @xmath328 and @xmath329 for @xmath330 . to provide the details we need some notation . for @xmath324 $ ]
, we denote @xmath331 .
let @xmath332 be the @xmath71-projection of @xmath229 on @xmath317 , so that @xmath333 . recall that @xmath315 denotes the intersection of the line @xmath334 the line parallel to @xmath317 through @xmath229with @xmath318 .
if @xmath335 is the @xmath71-projection of @xmath236 on @xmath336 , we have @xmath337 .
( see figure [ chord - fig - hd2 ] for an illustration of these definitions . )
we start with the following claim : [ lem : hd - induction ] for all @xmath324 $ ] , we have @xmath338 . by induction on @xmath230 . for the induction basis ( @xmath339 ) , we observe that the triangles @xmath340 and @xmath341 are similar , hence @xmath342 which yields @xmath343 as desired .
suppose that the claim is true for @xmath314 $ ] .
we will prove it for @xmath344 .
we similarly exploit the similarity of the triangles @xmath345 and @xmath346 , from which we get @xmath347 where the second equality follows from the collinearity of @xmath348 .
observe that the third term in ( [ eq : one ] ) is equal to @xmath349 by construction .
now note that , because of the parallelogram @xmath350 , we have @xmath351 .
hence , ( [ eq : one ] ) and the induction hypothesis imply @xmath352 which completes the proof of the claim . since @xmath353 ( as noted in the proof of claim [ lem : hd - induction ] ) , it follows that for all @xmath314 $ ] we have @xmath354 we will start by showing that the set @xmath355 is an @xmath356-convex pareto under the horizontal distance .
first , note that the error to the right of @xmath277 , i.e. , the distance of the lower envelope from @xmath357 is in fact zero ( since @xmath358 is the rightmost edge of the lower envelope ) .
it suffices to bound from above the error to its left .
since @xmath359 has absolute slope larger than @xmath39 , the unique point ( of the lower envelope ) at maximum distance from @xmath359 is @xmath236 .
thus , we have that @xmath320 . from the similarity of the triangles @xmath360 and @xmath361 we get @xmath362 hence , @xmath363 as desired .
we now proceed to analyze the behavior of the chord algorithm .
we will show that the algorithm selects the points @xmath364 ( in this order ) till it terminates .
formally , we consider the recursion tree built by the algorithm for the instance @xmath281 and prove that it is a path of length @xmath322 .
in particular , for all @xmath324 $ ] , at depth @xmath323 , the chord subroutine selects point @xmath229 .
we prove the aforementioned statement by induction on the depth @xmath6 of the tree .
recall that the chord algorithm initially finds the extreme points @xmath289 and @xmath279 . for @xmath365 ( first recursive call ) ,
the algorithm selects a point of the lower envelope with maximum horizontal distance from @xmath39 . by construction ,
all the points ( of the lower envelope ) in the line segment @xmath366 have the same ( maximum ) distance from @xmath39 ( since @xmath366 is parallel to @xmath39 ) .
hence , any of those points may be potentially selected .
since @xmath108 is a black - box oracle , we may assume that indeed @xmath236 is selected .
the maximum error after @xmath236 is selected equals @xmath367 .
hence , the algorithm will not terminate after it has selected @xmath236 . for the inductive step ,
we assume that the recursion tree is a path up to depth @xmath368 $ ] and the algorithm selected the points @xmath369 up to this depth .
we analyze the algorithm at depth @xmath370 . at depth
@xmath370 the information available to the algorithm is that ( i ) the error to the left of @xmath371 is @xmath243 and ( ii ) the error to its right is @xmath372 ( since @xmath373 ) . hence , the algorithm does not terminate and it calls @xmath108 to find a point between @xmath371 and @xmath279 at maximum distance from @xmath374 . by construction , the points of maximum distance are those belonging to the line @xmath375 ( which is parallel to @xmath374 ) ; similarly , we can assume @xmath376 is selected .
at this point of the execution the algorithm has the information that the error to the left of @xmath376 is @xmath243 and the error to its right is @xmath377 , unless @xmath378 .
this completes the induction and the proof of lemma [ claim : chd - hd ] . *
_ step 2 : _ * the instance @xmath284 is obtained from @xmath379 by appropriately setting the four relevant parameters . in particular , 1 .
fix @xmath380 , @xmath381 , @xmath382 and @xmath383 2
. set @xmath384 3 . also , define @xmath385 and @xmath386 .
observe that , under this choice of parameters , we have @xmath387 and @xmath388 . our main lemma for this step is the following : [ claim : chd - rd ] the chord algorithm applied to @xmath284 and error bound @xmath389 wrt ratio distance selects ( a superset of ) the points @xmath390 , while the set @xmath391 forms an @xmath392-convex pareto set .
the main idea of the proof is that for the particular instance under consideration , the horizontal distance metric is a very good approximation to the ratio distance . as a consequence
, one can show that the behavior of the chord algorithm in both metrics is similar .
the reason we `` lose '' a constant factor in the number of points ( i.e. , the chord algorithm selects @xmath393 points under the ratio distance as opposed to @xmath394 under the horizontal distance ) is due to the error term in the approximation between the metrics .
( the factor of @xmath395 is not important ; any constant factor bigger than @xmath396 would suffice . )
we now proceed with the details .
consider the set @xmath397 defining the instance @xmath284 .
we will need the following lemma that quantifies the closeness of the two metrics in our setting .
[ lem : hd - vs - rd ] let @xmath398 , @xmath288 and @xmath223 .
consider a point @xmath399 in @xmath286 and let @xmath36 be the absolute slope of @xmath400 .
if @xmath401 is the @xmath70-projection of @xmath399 on @xmath318 , then for any point @xmath402 in @xmath403 we have @xmath404 the proof , though elementary , requires some careful calculations .
let @xmath405 be the @xmath70-projection of @xmath399 on the segment @xmath318 , that is @xmath406 .
clearly , @xmath407 if @xmath36 is the absolute slope of @xmath400 , we have that @xmath408 fix a point @xmath409 .
we want to show that @xmath410 , the horizontal distance of @xmath402 from @xmath400 , is a good approximation to the corresponding ratio distance @xmath411 when @xmath36 is large .
( see figure [ chord - fig - hrd ] for an illustration . )
we first calculate the horizontal distance @xmath410 .
observe that @xmath412 where @xmath64 is the @xmath71-projection of @xmath402 on @xmath413 , i.e. , @xmath414 and @xmath415 it is clear that @xmath416 from the similarity of the triangles @xmath417 and @xmath418 it follows @xmath419 since @xmath420 , @xmath421 and @xmath422 we get @xmath423 therefore , @xmath424 the ratio distance @xmath425 by definition is such that @xmath426 . we thus get @xmath427 or equivalently @xmath428 substitution yields that @xmath429 observe that the numerator of the above fraction equals @xmath410 and the denominator is bigger than @xmath396 .
hence , @xmath430 . for the other inequality we can write : @xmath431 as desired
. to obtain ( [ eq : ineq ] ) , we bound each term in ( [ eq : prod ] ) separately .
the first term is clearly at most @xmath40 . as for the second term , first note that the denominator is at least @xmath396 . to bound the numerator from above
, we claim that @xmath432 . to see this we use our assumption that @xmath402 lies in @xmath403 .
in particular , this implies that @xmath402 lies below ( or on ) the line segment @xmath400 , i.e. , @xmath433 which gives @xmath434 as desired .
this completes the proof of lemma [ lem : hd - vs - rd ] .
we may now proceed with the proof of lemma [ claim : chd - rd ] . by lemma [ claim : chd - hd ] ,
the set @xmath391 attains horizontal distance error at most @xmath392 . by the first inequality of ( [ eqn : hd - vs - rd ] ) ,
the ratio distance error of @xmath391 is at most as big , hence the second statement of lemma [ claim : chd - rd ] follows . by lemma [ claim : chd - hd ] , the chord algorithm under the horizontal distance metric selects all the points @xmath435 in order until it guarantees an @xmath436-approximation .
in paricular , after the algorithm has selected the subset @xmath325 , for @xmath437 $ ] , the horizontal approximation error is @xmath438 we remark that the error term @xmath439 in the rhs of ( [ eqn : hd - vs - rd ] ) leads to the `` constant factor loss '' , i.e. , the fact that the chord algorithm under the ratio distance picks @xmath393 ( as opposed to @xmath394 ) points .
( also note that , since the ratio distance is a lower bound for the horizontal distance , the chord algorithm under the former metric will select at most @xmath394 points . )
suppose we invoke the chord algorithm with desired error of @xmath243 , i.e. , we want to reconstruct the lower envelope exactly . then the algorithm will select the points @xmath229 in order of increasing @xmath230 .
it is also clear that the error of the approximation decreases monotonically with the number of calls to @xmath108 .
hence , to complete the proof , it suffices to show that after the algorithm has selected @xmath440 , the ratio distance error will be bigger than @xmath441 . to do this , we appeal to lemma [ lem : hd - vs - rd ] . the ratio distance error of the set @xmath442 is @xmath443 .
an application of ( the second inequality of ) ( [ eqn : hd - vs - rd ] ) for @xmath444 , @xmath445 gives @xmath446 for the first term of the rhs , it follows from ( [ eqn : hd - error2 ] ) by substitution that @xmath447 and we similarly get @xmath448 hence , @xmath449 where @xmath450 consider the error term @xmath451 .
we will show that @xmath452 which concludes the proof .
the first summand @xmath453 is negligible compared to @xmath454 , as long as @xmath455 to bound the second summand from above we need a lower bound on the slope @xmath456 .
recall ( equation ( [ eq : xistar ] ) in lemma [ lem : hd - induction ] ) that @xmath457 .
it is also not hard to verify that @xmath458 also recall that @xmath459 hence , @xmath460 it is straightforward to check that for the chosen values of the parameters , @xmath461 , hence the second summand will also be significantly smaller than @xmath462 . in conclusion ,
@xmath463 and lemma [ claim : chd - rd ] follows .
this also completes the proof of theorem [ thm : rd - lb - ws ] .
we can show an information theoretic lower bound against any algorithm that uses @xmath108 as a black box with respect to the ratio distance metric . even though the bound we obtain is exponentially weaker than that attained by the chord algorithm , it rules out the possibility of a constant factor approximation in this model .
in particular , we show : [ thm : rd - lb - ws - gen ] any algorithm ( even randomized ) with oracle access to a @xmath108 routine , has performance ratio @xmath464 with respect to the ratio distance .
let @xmath17 be a general algorithm with oracle access to @xmath108 .
the algorithm is given the desired error @xmath84 , and it wants to compute an @xmath84-cp set . to do this , it queries the @xmath108 routine on a sequence of ( absolute ) slopes @xmath465 and terminates when it has obtained a `` certificate '' that the set of points @xmath466 forms an @xmath84-cp set .
the queries to @xmath108 can be adaptive , i.e. , the @xmath230-th query @xmath467 of the algorithm @xmath17 can depend on the information ( about the input instance ) the algorithm has obtained from the previous queries @xmath468 . on the other hand , the adversary also specifies the input instance adaptively , based on the queries made by the algorithm . we will define a family of instances @xmath469 and an error @xmath148 with the following properties : 1 .
each instance @xmath470 has @xmath471 2 . in order for an algorithm @xmath17 to have a certificate it found an @xmath84-cp set for an ( adversarially chosen ) instance @xmath470
, it needs to make at least @xmath472 calls to @xmath108 .
our construction uses the lower bound example for the chord algorithm from section [ ssec : worst - lower - chord ] essentially as a black box .
consider the instance @xmath473 ( note that for @xmath216 we have that @xmath474 and @xmath475 ) and let @xmath476 be the corresponding set of points , where @xmath294 and @xmath477 and @xmath478 . our family of input instances @xmath469 consists of all `` prefixes '' of @xmath293 , i.e. , @xmath479 , where @xmath480 , @xmath324 $ ] .
by the properties of @xmath293 , each @xmath481 defines a convex monotone decreasing polygonal curve , i.e. , all its points are vertices of the corresponding convex pareto .
moreover , the set @xmath482 is an @xmath483-cp set for @xmath481 , where @xmath484 .
note that @xmath485 , @xmath324 $ ] , and that @xmath486 where the first inequality holds from ( [ eqn : hd - vs - rd ] ) , and the rest follow by construction .
this proves property @xmath487 above .
we now proceed to prove @xmath488 .
we claim that , given an arbitrary ( unknown ) instance @xmath470 , in order for an algorithm @xmath17 to have a certificate it discovered an @xmath84-cp set , @xmath17 must uniquely identify the instance @xmath16 . in turn
, this task requires @xmath489 @xmath108 calls .
indeed , consider an unknown instance @xmath470 and let @xmath229 be the rightmost point of @xmath16 ( excluding @xmath279 ) the algorithm @xmath17 has discovered up to the current points of its execution . at this point
, the information available to the algorithm is that @xmath490 .
hence , the error the algorithm can guarantee for its current approximation is @xmath491 the first inequality above follows from lemma [ lem : hd - vs - rd ] .
also note that the term in the rhs is minimized for @xmath330 and ( by the same analysis as in lemma [ claim : chd - rd ] ) the minimum value is bigger than @xmath436 .
it remains to show that an adversary can force any algorithm to make at least @xmath492 queries to @xmath108 until it has identified an unknown instance of @xmath469 .
clearly , identifying an unknown instance @xmath470 , i.e. , finding the index @xmath145 such that @xmath493 , is equivalent to identifying @xmath494 the rightmost point of @xmath16 to the left of @xmath279 .
first , we can assume the algorithm is given the extreme points @xmath495 beforehand .
a general algorithm @xmath17 is allowed to query the @xmath108 routine for any slope @xmath496 .
suppose that @xmath493 .
then , for @xmath497 , the @xmath108 routine returns : ( i ) @xmath229 if @xmath498 , ( ii ) @xmath494 if @xmath499 , and ( iii ) @xmath279 if @xmath500 .. that is , the information obtained from a query @xmath501 is whether @xmath498 , @xmath499 or @xmath500 .
so , for our class of instances , a general deterministic algorithm @xmath17 is equivalent to a ternary decision tree with the corresponding structure .
the tree has @xmath502 many leaf nodes and there are at most @xmath503 leaves at depth @xmath6 .
every internal node of the tree corresponds to a query to the @xmath108 routine , hence the depth measures the worst - case number of queries made by the algorithm .
it is straightforward that any such tree has depth @xmath492 and the theorem follows .
( if we allow randomization , the algorithm is a randomized decision tree .
the expected depth of any such tree remains @xmath492 which yields the theorem for randomized algorithms as well . )
we next show that , for the horizontal distance metric ( or vertical distance by symmetry ) , any algorithm in our model has an unbounded performance ratio , even for instances that lie in the unit square and for desired error @xmath282 .
[ thm : hd - lb - gen - ws ] any algorithm with oracle access to a @xmath108 routine has an unbounded performance ratio with respect to the horizontal distance , even on instances that lie in the unit square and for approximation error @xmath282 .
let @xmath17 be an algorithm that , given the desired error @xmath84 , computes an @xmath84-approximation with respect to the horizontal distance .
fix @xmath504 .
we show that an adaptive adversary can force the algorithm to make @xmath505 queries to @xmath108 even when @xmath506 queries suffice .
that is , the performance ratio of the algorithm is @xmath505 ; since @xmath239 can be arbitrary the theorem follows .
our family of ( adversarially constructed ) instances all lie in the unit square and are obtained by a modification of the lower bound instance @xmath507 for chord under the horizontal distance ( see step 1 in section [ ssec : worst - lower - chord ] ) . since the horizontal distance is invariant under translation , we can shift our instances so that the points @xmath508 and @xmath509 are the leftmost and rightmost points of the convex pareto set . after this shift ,
the point @xmath171 is identified with the origin .
consider the initial triangle @xmath169 , where @xmath510 , @xmath511 and @xmath512 , and let @xmath513 be the first query made by the algorithm .
given the value of @xmath514 , the adversary adds a vertex @xmath236 to the instance so that @xmath515
. the strategy of the adversary is the following : the point @xmath236 belongs to @xmath237 , i.e. , @xmath516 . if @xmath517 , then the point @xmath236 has @xmath518 otherwise , @xmath519 .
let @xmath235 be the line with slope @xmath514 through @xmath236 and @xmath520 be its intersection with @xmath521 .
the current information available to the algorithm is that the point @xmath236 is a feasible point and that there are no points below the line @xmath235 .
hence , the horizontal distance error it can certify is @xmath522 on the other hand , if the true convex pareto set was @xmath523 , the set @xmath524 would attain error @xmath525 note that @xmath526 after its first query , the algorithm knows that the error to the left of @xmath236 is @xmath243 , hence it needs to focus on the triangle @xmath252 .
therefore , it is no loss of generality to assume that @xmath527 . given @xmath528 , the adversary adds a vertex @xmath255 to the instance so that @xmath529 . its strategy is to place @xmath255 on @xmath530 and to set @xmath531 if @xmath532 and @xmath533 otherwise
let @xmath256 be the line with slope @xmath528 through @xmath255 and @xmath534 be its intersection with @xmath521 .
the information available to the algorithm after its second query is that the point @xmath255 is a feasible point and that there are no points below the line @xmath256 .
hence , the horizontal distance error it can certify is @xmath535 on the other hand , if the true convex pareto set was @xmath536 , the set @xmath537 would attain error @xmath538 similarly , @xmath539 the adversary argument continues by induction on @xmath230 .
the induction hypothesis is the following : after its @xmath230-th query , the algorithm has computed a set of points @xmath325 , @xmath540 , @xmath541 $ ] , where @xmath542 is the @xmath271-th query and @xmath543 .
( the @xmath544 s are vertices of the convex pareto set of the corresponding instance ordered left to right ) .
let @xmath545 be the line with slope @xmath467 through @xmath229 and @xmath546 be its intersection with @xmath521 .
then @xmath547 we now prove the induction step .
we start by noting that the error to the left of @xmath229 is zero , so the algorithm needs to focus on the triangle @xmath548 ; this implies that ( without loss of generality ) @xmath549 .
given @xmath550 , the adversary adds a vertex @xmath551 such that @xmath552 .
similarly , the strategy of the adversary is to place @xmath551 on @xmath553 and to set @xmath554 if @xmath555 and @xmath556 otherwise .
let @xmath557 be the line with slope @xmath550 through @xmath551 and @xmath546 be its intersection with @xmath521 .
the information available to the algorithm after its @xmath558-th query is that the point @xmath551 is a feasible point and that there are no points below the line @xmath557 .
hence , the horizontal distance error it can certify is @xmath559 on the other hand , if the true convex pareto set was @xmath560 , the set @xmath561 would attain error @xmath562 . now note that @xmath563 where the second inequality uses the inductive hypothesis and the definition of @xmath551 .
( see figure [ fig : gen - lb - hd ] for an illustration ; the figure depicts the case that @xmath564 . )
this completes the induction .
the overall adversary argument is obtained from the above construction for @xmath565 .
that is , the adversarially constructed instance is the set @xmath566 , where @xmath567 ( recall that @xmath467 is the @xmath230-th query ) .
the set @xmath568 has error @xmath569 while the algorithm can certify error @xmath570 thus , after @xmath571 steps , the error of the algorithm remains more than @xmath282 , while @xmath172 queries suffice to attain error @xmath572 .
this completes the proof of theorem [ thm : hd - lb - gen - ws ] . in this section
we establish the upper bound statement of theorem [ thm : worst ] , i.e. , we show that the performance ratio of the chord algorithm for the ratio distance is @xmath573 . for the sake of the exposition
, we start by showing the slightly weaker upper bound of @xmath193 .
the proof of the asymptotically tight upper bound is more involved and builds on the understanding obtained from the simpler argument presented first .
before we proceed with the argument , some comments are in order .
perhaps the most natural approach to prove an upper bound would be to argue that the error of the approximation constructed by the chord algorithm decreases substantially ( say by a constant factor ) in every iteration ( subdivision ) or after an appropriately defined `` epoch '' of a few iterations .
this , if true , would yield the desired result since the initial error can not be more than @xmath574 .
unfortunately , such an approach badly fails , as implied by the construction of theorem [ thm : rd - lb - ws ] .
recall that , in the simplest setting of that construction , the initial error is @xmath225 and decreases by an additive @xmath266 in every iteration , where @xmath575 .
hence , the error decreases by a sub - constant factor in every iteration .
in fact , we note that such an argument can not hold for _ any _ algorithm in our setting ( i.e. , given oracle access to @xmath108 ) , as follows from our general lower bound ( theorem [ thm : rd - lb - ws - gen ] ) .
our approach is somewhat indirect .
we prove that the _ area _ between the ( implicitly constructed ) `` upper '' and `` lower '' approximation decreases by a constant factor in every iteration of the algorithm ( lemma [ lem : area - half ] ) .
this statement can be viewed as a `` potential function type argument . '' to obtain an upper bound on the performance ratio , one additionally needs to relate the area to the ratio distance .
indeed , we show that , when the area ( between the upper approximation and the lower approximation ) has become `` small enough '' ( roughly at most @xmath576 ) , the error of the approximation ( ratio distance of the lower approximation from the upper approximation ) is at most @xmath84 ( lemma [ lemma : small - area - means - done ] ) .
we combine the above with a simple charging argument ( lemma [ lem : opt - lb ] ) to get the desired performance guarantee .
formally , we prove : [ thm : ws - ub - area ] let @xmath577 be the triangle at the root of the chord algorithm s recursion tree , let @xmath578 be its area , and denote @xmath579 .
the algorithm finds an @xmath84-cp set after @xmath580 calls to @xmath108 , where @xmath581 .
the claimed upper bound on the performance ratio follows from the previous theorem , since @xmath582 ^ 2 $ ] , which implies @xmath583 and @xmath584 . to prove theorem [ thm : ws - ub - area ]
we will need a few lemmas .
we start by proving correctness , i.e. , we show that , upon termination , the algorithm computes an @xmath84-cp set for @xmath79
. this statement may be quite intuitive , but it requires a proof .
the following claim formally states some basic properties of the algorithm : [ clm : sandwich ] let @xmath585 be the triangle processed by the chord algorithm at some recursive step
. then the following conditions are satisfied : ( i ) @xmath586 ; in particular , @xmath587 and @xmath588 , ( ii ) @xmath589 , @xmath590 and @xmath591 lies below the line @xmath592 , and ( iii ) @xmath593 . that is , whenever the chord routine is called with parameters @xmath173 ( see table [ table : chord ] ) , the points @xmath174 and @xmath175 are feasible solutions of the lower envelope and the segment of the lower envelope between them is entirely contained in the triangle @xmath177 .
consider the node ( corresponding to ) @xmath594 in the recursion tree built by the algorithm .
we will prove the claim by induction on the depth of the node .
the base case corresponds to either an empty tree or a single node ( root ) .
the chord routine is initially called for the triangle @xmath595 . conditions ( i ) and ( ii )
are thus clearly satisfied .
it follows from the definition of the @xmath108 routine that there exist no solution points strictly to the left of @xmath289 and strictly below @xmath279 .
hence , by monotonicity and convexity , we have @xmath596 , i.e. , condition ( iii ) is also satisfied .
this establishes the base case . for the induction step , suppose the claim holds true for every node up to depth @xmath597
. we will prove it for any node of depth @xmath370 .
indeed , let @xmath598 be a depth @xmath370 node and let @xmath585 be @xmath598 s parent node in the recursion tree . by the induction hypothesis , we have that ( i ) @xmath586 ; @xmath587 and @xmath588 ( ii ) @xmath589 , @xmath590 and @xmath591 lies below the line @xmath592 and ( iii ) @xmath593 .
we want to show the analogous properties for @xmath598 .
we can assume without loss of generality that @xmath598 is @xmath594 s left child ( the other case being symmetric ) .
then , it follows by construction ( see table [ table : chord ] ) that @xmath599 where @xmath600 .
we claim that @xmath601 .
indeed , note that @xmath602 ( as follows from property ( ii ) of the induction hypothesis ) . by monotonicity and convexity of the lower envelope , combined with property ( iii ) of the induction hypothesis , the claim follows .
now note that @xmath603 and @xmath604 .
hence , property ( i ) of the inductive step is satisfied .
since @xmath605 has non - positive slope ( as follows from ( ii ) of the inductive hypothesis ) , @xmath606 lies to the left and above @xmath607 ; similarly , since @xmath592 has negative slope , @xmath607 lies to the left and above @xmath229 .
we also have that @xmath608 , where the first inequality follows from the fact that @xmath601 . property ( ii )
follows from the aforementioned . by definition of the chord routine
, we have @xmath604 and there are no solution points below @xmath609 . by convexity , we get that @xmath610 lies below @xmath611 .
hence , property ( iii ) also follows .
this proves the induction and the claim . by exploiting the above claim
, we can prove correctness : [ lem : chord - finds - eps - convex ] the set of points @xmath293 computed by the chord algorithm is an @xmath84-cp set .
let @xmath612 be the set of feasible points in @xmath129 output by the algorithm , where the points of @xmath293 are ordered in increasing order of their @xmath70-coordinate ( decreasing order of their @xmath71-coordinate ) .
note that all the @xmath229 s are in convex position .
we have that @xmath613 .
so , it suffices to show that , for all @xmath230 , @xmath614 .
since the algorithm terminates with the set @xmath293 , it follows that , for all @xmath230 , the chord routine was called by the algorithm for the adjacent feasible points @xmath615 and returned without adding a new feasible point between them .
let @xmath591 be the corresponding third point ( argument to the chord routine ) .
then , by claim [ clm : sandwich ] , we have that @xmath591 is between @xmath229 and @xmath551 in both coordinates and below the segment @xmath616 and moreover @xmath617 .
since the chord routine returns without adding a new point on input @xmath618 , it follows that either @xmath619 or the point @xmath620 satisfies @xmath621 . in the former case , since @xmath622 , we obtain @xmath623 as desired . in the latter case , we have @xmath624 .
that is , we claim that @xmath625 is a point of @xmath626 ( by claim [ clm : sandwich ] ) with maximum ratio distance from @xmath627 .
since the feasible points in @xmath628 lie between @xmath616 and its parallel line through @xmath625 , the claim follows .
this completes the proof of lemma [ lem : chord - finds - eps - convex ] . to bound from above the performance ratio
we need a few more lemmas .
our first lemma quantifies the area shrinkage property .
it is notable that this is a statement independent of @xmath1 .
[ lem : area - half ] let @xmath629 be the triangle processed by the chord algorithm at some recursive step
. denote @xmath630 .
let @xmath631 , @xmath632 be the triangles corresponding to the two new subproblems .
then , we have @xmath633 let @xmath634 be the projections of @xmath229 on @xmath635 and @xmath605 respectively ; see figure [ fig : chord - area ] for an illustration .
let @xmath636\ ] ] where the equality holds because the triangles @xmath594 and @xmath637 are similar ( recalling that @xmath638 ) . hence , we get @xmath639 we have @xmath640 where ( [ uses : y ] ) follows from ( [ eq : y ] ) . by using ( [ eq :
y2 ] ) and expanding we obtain @xmath641 as desired .
our second lemma gives a convenient lower bound on the value of the optimum .
[ lem : opt - lb ] consider the recursion tree @xmath642 built by the algorithm and let @xmath643 be the set of lowest internal nodes , i.e. , the internal nodes whose children are leaves .
then @xmath644 .
recall that , by convention , there is no node in the tree if , for a triangle , the chord routine terminates without calling @xmath108 .
each lowest internal node of the tree corresponds to a triangle @xmath645 with the property that the ratio distance of the convex pareto set from the line segment @xmath592 is strictly greater than @xmath84 ( as otherwise , the node would be a leaf ) .
each such triangle must contain a point @xmath646 of an optimal @xmath84-cp set .
any two nodes of @xmath643 are not ancestor of each other , and therefore the corresponding triangles are disjoint ( neighboring triangles can only intersect at an endpoint ) .
thus , each one of them must contain a distinct point of the optimal @xmath84-cp set , and hence the lemma follows .
finally , we need a lemma that relates the ratio distance within a triangle to its area . we stress that the lemma applies only to triangles @xmath585 considered by the algorithm .
[ lemma : small - area - means - done ] consider a triangle @xmath585 considered in some iteration of the chord algorithm such that @xmath647 .
let @xmath648 . if @xmath649 , then @xmath650 .
the basic idea of the argument is that the worst - case for the area - error tradeoff is ( essentially ) when @xmath651 , and the triangle @xmath594 is right and isosceles ( i.e. , @xmath652 ) .
we now provide a formal proof .
let @xmath585 be a triangle considered by the algorithm .
the points @xmath653 and we have @xmath654 , @xmath655 and the point @xmath591 lies below the line @xmath592 .
note the latter imply that @xmath656 ( see figure [ fig : area - distance ] . )
we will relate the area @xmath657 to the ratio distance @xmath658 .
consider the intersection @xmath659 of the lines @xmath592 and @xmath660 , where @xmath661 denotes the origin .
then we have that @xmath662 . from the definition of the ratio distance
it follows that for any point @xmath663 it holds @xmath664 ( in fact , @xmath659 is the unique minimizer ) .
let @xmath665 be the projection of @xmath591 on @xmath592 .
it follows that @xmath666 for the area we have that @xmath667 . since @xmath594 is either right or obtuse
, the length of its largest base is at least twice the length of the corresponding height , i.e. , @xmath668 .
hence , @xmath669 . by expanding and using ( [ eq : star ] ) , we get @xmath670 and the lemma follows . at this point
we have all the tools we need to complete the proof of theorem [ thm : ws - ub - area ] . lemma [ lem : chord - finds - eps - convex ] gives correctness . to bound the performance ratio we proceed as follows : first , by lemma [ lem : area - half ] ,
when a node of the tree is at depth @xmath671 , the corresponding triangle will have area at most @xmath672 .
hence , by lemma [ lemma : small - area - means - done ] ( noting that @xmath673 ) , it follows that the depth of the recursion tree @xmath642 is @xmath674 .
every internal tree node is an ancestor of a node in @xmath643 .
the chord algorithm makes one query for every node of the tree , hence @xmath675 .
lemma [ lem : opt - lb ] now implies that @xmath676 , which concludes the proof . in this subsection , we prove the asymptotically tight upper bound of @xmath677 on the worst - case performance ratio of the chord algorithm .
the analysis is more subtle in this case and builds on the intuition obtained from the simple analysis of the previous subsection .
the proof bounds in effect the length of paths in the recursion tree that consist of nodes with a single child .
it shows that if we consider any @xmath1-convex pareto set @xmath0 , the segment of the lower envelope between any two consecutive elements of @xmath0 can not contain too many points of the solution produced by the chord algorithm .
[ thm : ws - ub - tight ] the worst - case performance of the chord algorithm ( with respect to the ratio distance ) is @xmath677 .
we begin by analyzing the case @xmath678 ( i.e. , the special case that one intermediate point suffices and is required for an @xmath84-approximation ) and then handle the general case .
it turns out that this special case captures most of the difficulty in the analysis .
let @xmath289 ( leftmost ) and @xmath279 ( rightmost ) be the extreme points of the convex pareto curve as computed by the algorithm .
we consider the case @xmath678 , i.e. , ( i ) the set @xmath679 is _ not _ an @xmath84-cp and ( ii ) there exists a solution point @xmath226 such that @xmath680 is an @xmath84-cp .
fix @xmath280 with @xmath681 .
we will prove that , for an appropriate choice of the constant in the big - theta , the chord algorithm introduces at most @xmath239 points in either of the intervals @xmath682 $ ] , @xmath683 $ ] .
suppose , for the sake of contradiction , that the chord algorithm adds more points than that in the segment @xmath684 ( the proof for @xmath685 being symmetric . ) we say that , in some iteration of the chord algorithm , a triangle is _ active _ , if it contains the optimal point @xmath226 .
in each iteration , the chord algorithm has an active triangle which contains the optimal point @xmath226 . outside that triangle
, the algorithm has constructed an @xmath84-approximation .
we note that the chord algorithm may in principle go back and forth between the two sides of @xmath226 ; i.e. , in some iterations the line parallel to the chord touches the lower envelope to the left of @xmath226 and in other iterations to the right .
let @xmath686 be the initial triangle .
we focus our attention on the ( not necessarily consecutive ) iterations of the chord algorithm that add points to the left of @xmath226 .
we index these iterations in increasing order with @xmath687 $ ] .
consider the `` active '' triangle @xmath688 generated in each such iteration , where @xmath606 is the new point of the curve added in the iteration .
we let @xmath689 be the initial triangle @xmath686 .
it is clear that ( i ) @xmath690 lies to the right of @xmath606 , ( ii ) @xmath691 lies to the left of ( or is equal to ) @xmath692 and ( iii ) @xmath693 .
also , let us denote @xmath694 , that is @xmath695 is the @xmath279-vertex ( i.e. , the vertex that lies to the right of @xmath226 ) of the active triangle in the last iteration @xmath696 .
note that @xmath695 could be equal to the point @xmath279 ( which would happen if the chord algorithm introduces points only to the left of @xmath226 , i.e. , proceeds towards @xmath226 monotonically ) , but in general this will not be the case .
let @xmath697 be the intersection point of the line @xmath605 with the line @xmath698 ; see figure [ tight - upper ] . note that , to the left of @xmath226 the convex pareto curve has no points below the line @xmath698 .
all the points of the convex pareto curve that are left of @xmath226 and lie in the active triangle @xmath688 ( these are the potentially not @xmath84-covered points ) are actually in the triangle @xmath699 .
consider the line that goes through @xmath606 and is parallel to @xmath700 and let @xmath701 be its intersection with @xmath698 .
note that the line @xmath605 is parallel to @xmath702 ( by construction of the algorithm , this is the line that added the point @xmath606 and formed the active triangle @xmath688 ) and @xmath692 lies to the right of ( or is equal to ) @xmath695 , so the line @xmath703 is to the left of ( or is equal to ) @xmath605 , hence @xmath701 lies to the left of ( or is equal to ) @xmath697 .
now , let @xmath11 be the intersection point of @xmath704 with the line @xmath698 .
clearly , @xmath11 lies to the left of @xmath701 .
furthermore , @xmath11 lies to the right of ( or is equal to ) @xmath705 .
the reason is that , below @xmath706 the curve has no points ( strictly ) to the left of the line @xmath707 , so @xmath606 is to the right of the line ( or on the line ) .
if a point @xmath708 is above a line @xmath145 , in the sense that the segment connecting @xmath64 to the origin @xmath661 intersects @xmath145 , say at some point @xmath28 , then we define the _ excess ratio distance _ of @xmath64 from @xmath145 to be @xmath709 , i.e. , the ratio distance of @xmath28 from @xmath64 .
let @xmath710 be the excess ratio distance of @xmath606 from the line @xmath698 for @xmath711 .
let @xmath712 be the excess ratio distance of @xmath695 from the line @xmath703 for @xmath713 . referring to figure [ tight - upper ] , @xmath714 and @xmath715 .
let @xmath716 and let @xmath717 . by definition
, we have @xmath718 for all @xmath230 .
furthermore , @xmath719 is the excess ratio distance of @xmath720 from the line @xmath721 .
since the coordinates of the points are in @xmath722 $ ] , it follows that @xmath723 .
thus , we get @xmath724 [ claim : tu1 ] @xmath725 and @xmath726 .
let @xmath727 be the intersection of the segment @xmath728 with the line @xmath698 and @xmath729 the intersection of @xmath730 with @xmath698 . by definition , @xmath731 , and @xmath732 .
let @xmath733 be the intersection of @xmath730 with the line from @xmath606 parallel to @xmath698 . from the similar triangles
@xmath734 and @xmath735 , we have @xmath736 .
therefore , @xmath737 . from the similar triangles @xmath738 and @xmath739 , the latter ratio is equal to @xmath740 , yielding the first equality for @xmath741 in the claim .
the second equality follows from the similar triangles @xmath742 and @xmath743 , since @xmath703 is parallel to @xmath700 .
the second equality implies then the expression for @xmath744 . from the claim we have , @xmath745 and
since @xmath705 lies left of @xmath11 , we have @xmath746 .
therefore , @xmath747 let @xmath748 be the intersections of @xmath749 with the lines @xmath703 and @xmath707 respectively ; see figure [ tight - upper ] . clearly , @xmath750 , and hence @xmath751 . thus , @xmath752
. therefore , @xmath753 [ claim : tu2 ] @xmath754 and @xmath755 .
since the last iteration of the chord algorithm adds a new point @xmath756 , the segment @xmath757 does not @xmath84-cover all the pareto points left of @xmath226 in the active triangle .
these points are all in the triangle @xmath758 .
the ratio distance of any point in this triangle from @xmath757 is upper bounded by both @xmath759 and by @xmath760 .
it follows that @xmath754 and @xmath755 .
thus , we get @xmath761 and @xmath762 [ claim : tu3 ] @xmath763 . if @xmath764 then the claim follows from inequality ( [ g - bound ] ) .
so suppose that @xmath765 .
the point @xmath766 is at ratio distance at most @xmath84 from the line @xmath684 since @xmath767 is an @xmath1-convex pareto set .
therefore , @xmath226 is at most excess ratio distance @xmath84 from the line @xmath768 , because @xmath768 is parallel to @xmath769 and @xmath695 is to the right of @xmath226 . since @xmath765
, it follows that @xmath770 and hence @xmath771 .
therefore , @xmath772 since @xmath773 ( as @xmath774 is left of @xmath226 ) , we conclude that @xmath775 thus , we have a lower bound on the product of the @xmath741 s from inequality ( [ p - bound ] ) and on the product of the @xmath776 s from claim [ claim : tu3 ] .
it is easy to see ( and is well - known ) that for a fixed product of the @xmath741 s , the product of the @xmath776 s is maximized if all factors are equal .
we include a proof for convenience .
let @xmath777 for @xmath778 .
the maximum of @xmath779 subject to @xmath780 is achieved when all the @xmath781 are equal .
suppose @xmath782 .
then @xmath783 is maximized subject to @xmath784 , when @xmath785 is minimized , which happens when @xmath786 by the arithmetic - geometric mean inequality ( @xmath787 with equality iff @xmath786 . for general @xmath788 ,
if the @xmath781 s maximize @xmath779 subject to @xmath780 then we must have @xmath789 for all pairs @xmath790 , because otherwise replacing @xmath791 by their geometric mean will increase @xmath779 .
thus , for any value of @xmath792 , the product @xmath793 is maximized when @xmath794 for all @xmath795 and @xmath796 . since @xmath797
, we must have @xmath798 because @xmath799 .
therefore , @xmath800 , hence @xmath801 , which implies that @xmath802 .
we now proceed to analyze the general case , essentially by reducing it to the aforementioned special case .
suppose that the optimal solution has an arbitrary number of points , i.e. , has the form @xmath803 .
charge the points computed by the chord algorithm to the edges of the optimal solution as follows : if a point belongs to the portion of the lower envelope between the points @xmath302 and @xmath804 ( where we let @xmath805 and @xmath806 , then we charge the point to the edge @xmath807 ; if the chord algorithm generates a point @xmath229 of the optimal solution then we can charge it to either one of the adjacent edges .
we claim that every edge of @xmath808 is charged with at most @xmath809 points of the chord algorithm , where @xmath810 is the same number as in the above analysis for the @xmath678 case . to see this , consider any edge @xmath807 of @xmath808 .
let @xmath811 be the first point generated by the chord algorithm that is charged to this edge , i.e. , @xmath811 is the first point that lies between @xmath302 and @xmath804 .
we claim that the chord algorithm will generate at most @xmath239 more points in each of the two portions @xmath812 and @xmath813 of the lower envelope .
the argument for the two portions is symmetric .
consider the portion @xmath814 .
the proof that the chord algorithm will introduce at most @xmath239 points in this portion is identical to the proof we gave above for the @xmath678 case , with @xmath811 in place of @xmath289 and @xmath229 in place of @xmath226 .
the only fact about the assumption @xmath678 that was used there was that the edge @xmath684 @xmath1-covers the portion of the lower envelope between @xmath289 and @xmath226 .
it is certainly true here that the segment @xmath815 @xmath1-covers the portion @xmath814 , since the edge @xmath816
@xmath1-covers @xmath817 . hence ,
by the same arguments , the chord algorithm will generate at most @xmath239 points between @xmath811 and @xmath229 .
thus , the algorithm will generate no more than @xmath818 points overall , and hence its performance ratio is @xmath677 .
this completes the proof .
_ we briefly sketch the differences in the algorithm and its analysis for the case of an approximate comb routine .
first , in this case , the description of the chord algorithm ( table [ table : chord ] ) has to be slightly modified ; this is needed to guarantee that the set of computed points is indeed an @xmath84-cp .
in particular , in the chord routine , we need to check whether @xmath819 for an appropriate @xmath820 .
in particular , we choose @xmath821 such that @xmath822 , where @xmath823 is the accuracy of the approximate comb routine , i.e. , the routine @xmath122 .
consider the case that the @xmath122 routine always returns feasible points that belong to a @xmath120 scaled version of the lower envelope .
the same analysis as in the current section establishes that the chord algorithm performs at most @xmath824 calls to @xmath122 in this setting .
if @xmath821 is `` close '' to @xmath84 ( say , @xmath825 ) the first term is clearly @xmath826 .
hence , to prove the desired upper bound , it suffices to show that @xmath827 .
( it is clear that @xmath828 , but in principle it may be the case that @xmath829 is arbitrarily larger . )
this is provided to us by a planar geometric lemma from @xcite ( lemma 5.1 ) which states that if @xmath830 then @xmath831 .
selecting @xmath832 suffices for the above and completes our sketched description . _
in section [ ssec : average - upper ] we present our average case upper bounds and in section [ ssec : average - lower ] we give the corresponding lower bound . in section [ ssec : ppp - upper ]
we start by proving our upper bound for random instances drawn from a poisson point process ( ppp ) .
the analysis for the case of unconcentrated product distributions is somewhat more involved and is given in section [ ssec : prod - upper ] .
* overview of the proofs . *
the analysis for both cases has the same overall structure , however each case has its difficulties .
we start by giving a high - level overview of the arguments . for the sake of simplicity ,
in the following intuitive explanation , let @xmath55 denote : ( i ) the expected number of points in the instance for a ppp and ( ii ) the actual number of points for a product distribution .
similarly to the simple proof of section [ sssec : worst - upper - simple ] for worst - case instances , to analyze our distributional instances we resort to an indirect measure of progress , namely the area of the triangles maintained in the algorithm s recursion tree .
we think that this feature of our analysis is quite interesting and indicates that this measure is quite robust . in a little more detail , we first show ( see lemma [ lem : area recursion ] for the case of ppp ) that every subdivision performed by the algorithm decreases the area between the upper and lower approximations by a significant amount ( roughly at an exponential rate ) with high probability .
it then follows that at depth @xmath833 of the recursion tree , each `` surviving triangle '' contains an expected number of at most @xmath833 points with high probability .
we use this fact , together with a charging argument in the same spirit as in the worst - case , to argue that the expected performance ratio is @xmath833 in this case . to analyze the expected performance ratio in the complementary event , we break it into a `` good '' event , under which the ratio is @xmath834 with high probability , and a `` bad '' event , where it is potentially unbounded ( in the poisson case ) or at most @xmath55 ( for the case of product distributions ) .
the potential unboundedness of the performance ratio in the poisson case creates complications in bounding the expected ratio of the algorithm over the entire space .
we overcome this difficulty by bounding the upper tail of the poisson distribution ( see claim [ claim : ppp - tail ] ) . in the case of product distributions , the worst case bound of @xmath55 on
the competitive ratio is sufficient to conclude the proof , but the technical challenges present themselves in a different form . here ,
the `` contents '' of a triangle being processed by the algorithm depend on the information coming from the previous recursive calls making the analysis more involved .
we overcome this by understanding the nature of the information provided from the conditioning . * on the choice of parameters . *
a simple but crucial observation concerns the interesting range for the parameters of the distributions .
suppose that we run the chord algorithm with desired error @xmath148 on some random instance that lies entirely in the set @xmath835 ^ 2 $ ] . then , it is no loss of generality to assume that the number of random points in the instance ( expected number for the ppp case ) is upper bounded by some fixed polynomial in @xmath836 and @xmath837 . if this is not the case , it is easy to show that the chord algorithm makes at most a constant number of @xmath108 calls in expectation .
for the analysis of the ppp case we will make crucial use of the following technical claim : [ claim : ppp - tail ] let @xmath838 be a @xmath839 random variable , with @xmath840 , and let @xmath841 be some event .
then @xmath842 \pr[{\cal e } ] \le \max\left\{{1 \over \nu } , o(\nu^3 ) \pr[{\cal e } ] \right\}.\ ] ] let @xmath843 be such that @xmath844 < \pr[{\cal e } ] \le \pr[x \ge k^*]$ ] . clearly , @xmath842 \pr[{\cal e } ] \leq { \mathop{\textstyle \sum}}_{i = k^*}^{+ \infty } i \cdot \pr[x = i ] = { \mathop{\textstyle \sum}}_{i = k^*}^{+ \infty } i \cdot { e^{- \nu } \nu^i \over i ! } = \nu { \mathop{\textstyle \sum}}_{i = k^*-1}^{+ \infty } { e^{- \nu } \nu^{i } \over i ! } = \nu \pr[x \ge k^*-1].\ ] ] we distinguish two cases . if @xmath845 , then chebyshev s inequality yields @xmath846 \le { 1 \over \nu^2}$ ] which gives @xmath842 \pr[{\cal e } ] \le { 1\over \nu}.\ ] ] if @xmath847 , then @xmath848 \le { k^*+1 \over \nu } \pr[x = k^*+1 ] \le o(\nu ) \pr[{\cal e}]\ ] ] and @xmath849 \le { ( k^*+1)^2 \over \nu^2 } \pr[x = k^*+1 ] \le o(\nu^2 ) \pr[{\cal e}].\ ] ] hence , @xmath850 \pr[{\cal e } ] & \le & \nu \cdot \pr[x \ge k^*-1]\\ & \le & \nu \cdot \big(\pr[{\cal e}]+ \pr[x = k^*-1 ] + \pr[x = k^*]\big)\\ & \le & o(\nu^3 ) \cdot \pr[{\cal e}].\end{aligned}\ ] ] this concludes the proof of the claim .
we start by pointing out that if the intensity of the ppp is very large , the chord algorithm will terminate after a constant number of calls in expectation : [ fact : dens - ppp ] let @xmath595 be at the root of the chord algorithm s recursion tree and let @xmath194 denote the intensity of the ppp .
let @xmath851 and @xmath852 . if @xmath853 , then @xmath854 = o(1)$ ] .
first note we can clearly assume that @xmath855 and @xmath856 .
let @xmath857 and @xmath858 .
let @xmath859 be the right triangle of maximum area whose hypotenuse is parallel to @xmath39 .
( this is the shaded triangle in figure [ fig : ub - ac ] . )
we claim that @xmath860 indeed , it is clear that @xmath171 is a vertex of @xmath861 and that either @xmath862 or @xmath863 ( or both ) are vertices .
hence , one of the edges of @xmath861 has length at least @xmath864 . since @xmath30 is the slope of the hypotenuse , the other edge has length at least @xmath865 .
if there is a feasible point in @xmath861 , the chord algorithm will find it by calling @xmath233 and terminate ( since such a point forms an @xmath84-cp set ) .
let @xmath866 be the number of random points that land in the triangle @xmath861 .
note that @xmath866 is a @xmath867 random variable , with @xmath868 .
we can write @xmath869 = { \mathbb{e}}[{\mathrm{chord}}_{{\epsilon}}(t_1 ) \mid x^{\ast } = 0 ] \pr[x^{\ast}=0 ] + { \mathbb{e}}[{\mathrm{chord}}_{{\epsilon}}(t_1 ) \mid x^{\ast } \ge 1 ] \pr[x^{\ast } \ge 1].\ ] ] observe that the second term is bounded from above by a constant , hence it suffices to bound the first term . recall that the number of calls performed by the chord algorithm is at most twice the number @xmath870 of feasible points in the root triangle @xmath577 .
therefore , we can write @xmath871 \pr[x^{\ast}=0 ] \le 2 { \mathbb{e}}[y_1 \mid x^{\ast } = 0 ] \pr[x^{\ast}=0].\ ] ] recall that @xmath870 is a @xmath872 random variable , where @xmath873 and note that we can assume wlog that @xmath874 ( since otherwise the expected number of feasible points in @xmath577 is at most @xmath396 and we are done ) .
hence , by claim [ claim : ppp - tail ] the rhs above is bounded by @xmath875 \right\}.\ ] ] thus , to complete the proof it suffices to show that @xmath876 = o(1)$ ] .
first note that this quantity equals @xmath877 which is at most @xmath878 by monotonicity ( which holds for our choice of @xmath879 ) .
the latter expression is at most @xmath880 , where @xmath881 , which is easily seen to be absolutely bounded .
the same proposition also applies for the case that we have an approximate @xmath122 routine ( in this case , we replace @xmath84 by an appropriate @xmath882 so that @xmath883 ) .
we remark that a similar proposition can be shown for the hausdorff distance ; this does not hold for the case of the horizontal / vertical distance , which is why the average case bounds of this section do not apply for the latter metrics .
note that by assumption @xmath884 ^ 2 $ ] which implies that @xmath885 and @xmath584 .
we also have that @xmath886 ( since otherwise the set @xmath887 is an @xmath84-cp ) which gives @xmath888 therefore , @xmath889 the main result of this section is the following theorem , which combined with proposition [ fact : dens - ppp ] , yields the desired upper bound of @xmath890 on the expected performance ratio .
[ th : competitive ratio ppp ] let @xmath577 be the triangle at the root of the chord algorithm s recursion tree , and suppose that points are inserted into @xmath577 according to a poisson point process with intensity @xmath194 .
the expected performance ratio of the chord algorithm on this instance is @xmath891 .
the proof of theorem [ th : competitive ratio ppp ] will require a sequence of lemmas . throughout this section , we will denote @xmath892 . recall that the number of queries performed by the chord algorithm is bounded from above by twice the total number of points in the triangle @xmath577 .
since the expected number of points in @xmath577 is @xmath893 , we have : [ lem : super pessimistic competitive ratio ] the expected performance ratio of the chord algorithm is @xmath894 . hence , we will henceforth assume that @xmath895 is bounded from below by a sufficiently large positive constant .
( if this is not the case , the expected total number of points inside @xmath577 is @xmath506 and the desired bound clearly holds . )
our first main lemma in this section is an average case analogue of our lemma [ lem : area - half ] : the lemma says that the area of the triangles maintained by the algorithm decreases _ geometrically _ ( as opposed to linearly , as is the case for arbitrary inputs ) at every recursive step ( with high probability ) . intuitively , this geometric decrease is what causes the performance ratio to drop by an exponential in expectation .
[ lem : area recursion ] let @xmath585 be the triangle processed by the chord algorithm at some recursive step .
denote @xmath896 .
let @xmath897 and @xmath898 .
for all @xmath899 , with probability at least @xmath900 conditioning on the information available to the algorithm , @xmath901 it follows from the properties of the chord algorithm ( see e.g. , claim [ clm : sandwich ] ) that , before the @xmath108 routine on input @xmath594 is invoked , the following information is known to the algorithm , conditioning on the history ( see figure [ fig : chord - avg ] ) : * there exist solution points at the locations @xmath544 , for all @xmath902 $ ] .
* there is no point below the line @xmath903 , for all @xmath541 $ ] .
* there is no point below the line @xmath904 , for all @xmath541 $ ] . by the definition of the poisson point process , conditioning on the above information
, the number of points in a region of area @xmath0 inside @xmath594 follows a poisson distribution with parameter @xmath905 .
hence , letting @xmath906 being parallel to @xmath592 so that the triangle @xmath907 has area @xmath908 , it follows that , with probability at least @xmath909 , @xmath910 ( i.e. , the triangle contains a feasible point ) .
hence , with probability at least @xmath909 the point @xmath229 is contained in @xmath861 .
we bound from above the area of @xmath911 by the area of @xmath912 ( clearly , @xmath913 ) and similarly the area of @xmath914 by the area of @xmath915 . from the similarity of the triangles @xmath594 and @xmath861 we get @xmath916 hence , @xmath917 which gives @xmath918 .
therefore , @xmath919 finally , @xmath920 this concludes the proof of lemma [ lem : area recursion ] .
let us choose @xmath921 , and let @xmath598 be a triangle maintained by the algorithm at depth @xmath6 of the recursion .
it follows from lemma [ lem : area recursion ] that , with probability at least @xmath922 , @xmath923 where to bound the probability of the above event we have taken a union bound over the events on the path of the recursion tree connecting @xmath598 to the root of the recursion .
now consider the top @xmath924 levels of the recursion tree of the algorithm .
using lemma [ lem : area recursion ] and a union bound it follows that , with overall probability at least @xmath925 , the area of _ every _ triangle at depth ( exactly ) @xmath926 of the recursion tree is bounded from above by @xmath927 where we used our assumption on the range of @xmath171 .
let @xmath928 be the event that all the nodes ( triangles ) maintained by the algorithm at depth @xmath926 of the recursion tree ( if any ) have area at most @xmath929 .
in the aforementioned , we argued that the probability of the event @xmath928 is at least @xmath930 .
we now show the following : [ lem : competitive ratio in the good event ] conditioning on @xmath928 , the expected performance ratio of the chord algorithm is @xmath931 .
let @xmath932 be the recursion tree of the algorithm and @xmath933 be obtained from @xmath932 by pruning it at level @xmath926 .
let @xmath934 be the set of nodes of @xmath933 and let @xmath935 be the subset of nodes in @xmath934 that lie at depth @xmath926 from the root .
clearly @xmath936 is a subset of the leaves of @xmath933 . for a triangle ( node ) @xmath598 maintained by the algorithm at depth @xmath926 of the recursion , that is @xmath937 , we let the random variable @xmath938 denote the number of points inside @xmath598 . also , denote by @xmath939 the set of lowest internal nodes of the tree @xmath933 . by ( a straightforward analogue of ) lemma [ lem : opt - lb ] , we have @xmath940 . also , since @xmath933 is a depth @xmath926 binary tree , it holds @xmath941 .
we condition on the information @xmath942 available to the algorithm in the first @xmath926 levels of its recursion - tree ( without the information obtained from processing
i.e. , calling @xmath108 for any triangle at depth @xmath926 ) . by assumption
, @xmath942 satisfies the event @xmath943 .
conditioning on the information @xmath942 , for all @xmath937 , @xmath938 follows a poisson distribution with parameter @xmath944 .
so , given that the event @xmath943 holds , we have @xmath945 \le \nu \cdot s^{\ast \ast}$ ] .
note that the chord algorithm makes a query to @xmath108 for every node in the tree @xmath933 . also recall that the number of queries performed by the algorithm on a triangle @xmath598 containing a total number of @xmath938 points is at most @xmath946 .
hence , the expected total number of queries made can be bounded as follows : @xmath947 & \le & |{\cal v}_{d^{\ast}}| + 2 \cdot { \mathop{\textstyle \sum}}_{t \in { \cal l}_{d^{\ast } } } \mathbb{e}[x_t~|~{\cal f}]\\ & \le & |{\cal v}_{d^{\ast}}| + 2 \left|{\cal l}_{d^{\ast}}\right| \cdot \nu \cdot s^{\ast \ast}\\ & \le & |{\cal v}_{d^{\ast}}| + 4 |{\cal l}'_{d^{\ast}}| \cdot \nu \cdot s^{\ast \ast}\\ & \le & |{\cal l}'_{d^{\ast}}| \cdot ( 2d^{\ast}+ 4 \nu \cdot s^{\ast \ast}),\end{aligned}\ ] ] where the third inequality uses the fact that @xmath948 .
so , conditioning on the information @xmath942 , the expected performance ratio of the algorithm is @xmath949 & \le & \mathbb{e}\left [ { { \mathrm{chord}}_{{\epsilon } } \over |{\cal l}'_{d^{\ast}}|}~\vline~{\cal f}\right ] \\ & \le & ( 2d^{\ast}+ 4 \nu \cdot s^{\ast \ast } ) \\ & = & o \left ( \log \log ( \nu s_1 ) \right).\end{aligned}\ ] ] integrating over all possible @xmath942 in @xmath943 concludes the proof of lemma [ lem : competitive ratio in the good event ] . from lemma [ lem : competitive ratio in the good event ]
it follows that @xmath950=o\left ( \log \log ( \nu s_1 ) \right),\ ] ] and from the preceding discussion we have that @xmath951 \le { 2 \over ( { \ln ( \nu s_1)})^{c-1}}.$ ] hence , we have established the following .
[ lem : good case competitive ratio ] for @xmath952 , there exists an event @xmath953 , with @xmath954 \ge 1-{2 \over ( { \ln ( \nu s_1)})^{c-1}}$ ] , such that the expected performance ratio of the algorithm conditioning on @xmath928 is @xmath955 .
let @xmath956 be the event that all the triangles at the level @xmath957 of the recursion tree of the algorithm ( if any ) have area at most @xmath958 . with the same technique , but using different parameters in the argument , we can also establish the following : [ lem : pessimistic competitive ratio ] for @xmath959 , there exists an event @xmath960 , with @xmath961 \ge 1-{2 \over ( { \nu s_1})^{c'-1}}$ ] , such that the expected performance ratio of the algorithm conditioning on @xmath960 is @xmath962 .
we want to use lemmas [ lem : good case competitive ratio ] and [ lem : pessimistic competitive ratio ] together with proposition [ lem : super pessimistic competitive ratio ] to deduce that the expected performance ratio of the algorithm is @xmath955 .
this may seem intuitive , but it is in fact not immediate . for technical purposes let us define the event @xmath963 , where @xmath943 is the event defined in the proof of lemma [ lem : good case competitive ratio ] .
it is easy to see that conditioning on @xmath964 the expected performance ratio of the algorithm can still be bounded by @xmath965 , since this expectation is affected only by whatever happens at level @xmath966 of the recursion tree and below . on the other hand , using the fact that @xmath954 \ge 1-{2 \over ( { \ln ( \nu s_1)})^{c-1}}$ ] , it follows that @xmath967 \le { 2 \over ( { \ln ( \nu s_1)})^{c-1}}.\ ] ] we bound the expectation of the performance ratio using the law of total probability as follows : @xmath968 & \le & \mathbb{e}\left [ { { \mathrm{chord}}_{{\epsilon } } \over { \mathrm{opt}}_{{\epsilon}}}~\vline~{\cal a}\right ] \cdot \pr[{\cal a } ] + \mathbb{e}\left [ { { \mathrm{chord}}_{{\epsilon } } \over { \mathrm{opt}}_{{\epsilon}}}~\vline~{\cal c}\right ] \cdot \pr[{\cal c } ] \nonumber \\ & + & \mathbb{e}\left [ { { \mathrm{chord}}_{{\epsilon } } \over { \mathrm{opt}}_{{\epsilon}}}~\vline~\overline{{\cal a}\cup { \cal c}}\right ] \cdot \pr\left[\overline{{\cal a}\cup { \cal c}}\right ] \nonumber \\ & \le & o\left ( \log \log ( \nu s_1 ) \right ) \cdot \left ( 1-{2 \over ( { \ln ( \nu s_1)})^{c-1}}\right ) + o\left ( \log(\nu s_1 ) \right ) \cdot { 2 \over ( { \ln ( \nu s_1)})^{c-1 } } \nonumber \\ & + & \mathbb{e}\left [ { { \mathrm{chord}}_{{\epsilon } } \over { \mathrm{opt}}_{{\epsilon}}}~\vline~\overline{{\cal a}\cup { \cal c}}\right ] \cdot \pr\left[\overline{{\cal a}\cup { \cal c}}\right ] .
\label{eqn : ena } \end{aligned}\ ] ] where ( [ eqn : ena ] ) follows from lemmas [ lem : good case competitive ratio ] and [ lem : pessimistic competitive ratio ] . to conclude
, we need to bound the last summand in the above expression .
note first that @xmath969 .
hence , @xmath970 \le { 2 \over ( { \nu s_1})^{c'-1}}.\ ] ] we again use the fact that the number of queries made by the chord algorithm ( hence , also the performance ratio ) is bounded by twice the total number of points in the triangle at the root of the recursion tree .
this number @xmath838 follows a poisson distribution with parameter @xmath893 .
hence , we have @xmath971 \cdot \pr\left[\overline{{\cal a}\cup { \cal c}}\right ] \le 2 \cdot \mathbb{e}\left [ x~\vline~\overline{{\cal a}\cup { \cal c}}\right ] \cdot \pr\left[\overline{{\cal a}\cup { \cal c}}\right].\ ] ] to bound the right hand side of the above inequality we use claim [ claim : ppp - tail ] and obtain : @xmath972 \cdot \pr \left [ \overline{{\cal a } \cup { \cal c } } \right ] & \le & \max\left\{{1 \over \nu s_1 } , o((\nu s_1)^3 ) \cdot \pr\left[\overline{{\cal a}\cup { \cal c}}\right ] \right\ } \\ &
\le & \max \left\ { { 1\over \nu s_1 } , o((\nu s_1)^3 ) { 2 \over ( { \nu s_1})^{c'-1}}\right\}.\end{aligned}\ ] ] choosing @xmath973 , the above rhs becomes @xmath506 .
plugging this into with @xmath974 gives @xmath975 = o\left ( \log \log ( \nu s_1 ) \right).\end{aligned}\ ] ] this concludes the proof of theorem [ th : competitive ratio ppp ] .
we start by proving the analogue of proposition [ fact : dens - ppp ] .
[ fact : dens - unif ] let @xmath595 be at the root of the chord algorithm s recursion tree and suppose that @xmath55 points are inserted into @xmath577 independently from a @xmath204-balanced distribution , where @xmath976 . let @xmath851 , @xmath852 and @xmath977 . if @xmath978 , then @xmath854 = o(1)$ ] . as in proposition [ fact : dens - ppp ] we have that @xmath979 ( see figure [ fig : ub - ac ] ) .
the probability that a random point falls into @xmath861 is at least @xmath980 , hence the probability than none of the @xmath55 points falls into @xmath861 is at most @xmath981 .
as before , if there is a feasible point in @xmath861 , the chord algorithm will find it and terminate . in all cases ,
the algorithm terminates after at most @xmath982 calls to @xmath108 .
hence , @xmath983 \leq 3 + 2n \cdot ( 1-t^{\ast})^n$ ] . the last summand is bounded by @xmath984 which is at most @xmath985 for all @xmath986 by monotonicity .
the latter quantity equals @xmath987 , where @xmath988 , which is easily seen to be @xmath506 . recalling that @xmath885 and @xmath989 we deduce that @xmath990 the main result of this section is devoted to the proof of the following theorem : [ th : competitive ratio product delta - balanced distribution ]
let @xmath577 be the triangle at the root of the chord algorithm s recursion tree , and suppose that @xmath55 points are inserted into @xmath577 independently from a @xmath204-balanced distribution , where @xmath205 .
the expected performance ratio of the chord algorithm on this instance is @xmath991 .
combined with proposition [ fact : dens - unif ] , the theorem yields the desired upper bound of @xmath992 .
the proof has the same overall structure as the proof of theorem [ th : competitive ratio ppp ] , but the details are more elaborate .
we emphasize below the required modifications to the argument .
since the performance ratio of the chord algorithm is at most @xmath982 on any instance with @xmath55 points , we will assume that @xmath55 is lower bounded by a sufficiently large absolute constant ( @xmath993 suffices for our analysis ) .
we start by giving an area shrinkage lemma , similar to lemma [ lem : area recursion ] .
( see figure [ fig : chord - avg ] for an illustration . )
[ lem : area recursion product distn ] let @xmath585 be a triangle processed by the chord algorithm at recursion depth at most @xmath994
. denote @xmath995 .
let @xmath897 , @xmath898 and @xmath996 . for all @xmath899 , with probability
at least @xmath997 conditioning on the information available to the algorithm ,
@xmath998 we follow the proof of lemma [ lem : area recursion ] with the appropriate modifications .
let @xmath594 be the triangle maintained by the algorithm at some node of the recursion tree , and suppose that @xmath594 is at recursion depth at most @xmath999 .
the information available to the algorithm when it processes @xmath594 ( before it makes the query @xmath1000 ) is : * the location of all points @xmath544 , @xmath902 $ ] is known ; by our assumption on the depth it follows that @xmath1001 .
* there is no point below the line ( defined by ) @xmath903 , or below the line @xmath904 , for all @xmath541 $ ] . given this information , the probability that , among the remaining @xmath1002 points ( whose location is unknown ) , none falls inside a triangle @xmath598 of area @xmath1003 inside @xmath594 , is at most @xmath1004 indeed ,
let @xmath1005 be the subset of the root triangle which is available for the location of @xmath229 ; this is the convex set below the line @xmath1006 , to the right of all lines @xmath1007 , for @xmath541 $ ] , and above all lines @xmath1008 , for @xmath541 $ ] .
the probability that a point whose location is unknown falls inside @xmath1009 is @xmath1010 \over { \cal d}[{t}^{\ast}_i ] } \ge { ( 1-\gamma ) { \cal u}\left [ t \right ] \over { \cal u}[{t}^{\ast}_i ] / ( 1-\gamma ) } \ge { ( 1-\gamma ) { \cal u}\left [ t \right ] \over { \cal u } [ { t_1}]/ ( 1-\gamma ) } = { ( 1-\gamma)^2 } { s(t ) \over s_1}.\ ] ] choosing @xmath1011 , the probability becomes @xmath1012 hence , the probability that @xmath598 is empty is at most @xmath1013 the proof of the lemma is concluded by identical arguments as in the proof of lemma [ lem : area recursion ] . using lemma [ lem : area recursion product distn ] and the union bound , we can show that , with probability at least @xmath1014 the following are satisfied : * all triangles maintained by the algorithm at depth @xmath1015 of its recursion tree have area at most @xmath1016 * for every node ( triangle ) @xmath230 in the first @xmath1017 levels of the recursion tree @xmath1018 where @xmath1019 is defined as in the statement of lemma [ lem : area recursion product distn ] .
the proof of the second assertion above follows immediately from lemma [ lem : area recursion product distn ] and the union bound .
the first assertion is shown similarly to the analogous assertion of theorem [ th : competitive ratio ppp ] .
for the above we assumed that @xmath1020
. now let us call @xmath1021 the event that the above assertions are satisfied .
we can show the following .
[ lem : competitive ratio in the good event - product ] suppose @xmath1022 .
conditioning on the event @xmath1021 , the expected performance ratio of the chord algorithm is @xmath1023 .
the proof is in the same spirit to the proof of lemma [ lem : competitive ratio in the good event ] , but more care is needed .
we need to argue that , under @xmath1021 , the expected number of points falling inside a triangle at depth @xmath1024 of the recursion tree is @xmath1025 . using rationale similar to that used in the proof of lemma [ lem : area recursion product distn ] above , we have the following : let @xmath594 be the triangle maintained by the algorithm at a node @xmath230 at depth @xmath1015 of the recursion tree .
let also @xmath64 be a point whose location is unknown to the algorithm ( conditioning on the information known to the algorithm after processing the first @xmath1026 levels of the recursion tree ) .
the probability that the point @xmath64 falls inside @xmath594 is @xmath1027 \over { \cal d}[{t}^{\ast}_i ] } \le { { \cal u}\left [ t_i \right ] / ( 1-\gamma ) \over ( 1-\gamma ) { \cal u}[{t}^{\ast}_i ] } , \ ] ] where @xmath1028 is the region below the line @xmath1006 , above the lines @xmath903 , for all @xmath271 in the first @xmath1024 levels of the recursion tree , and above the lines @xmath1008 , for all @xmath271 in the first @xmath1015 levels of the recursion tree . to upper bound the probability that @xmath64 falls inside @xmath594
, we need a lower bound on the size of the area @xmath1029 .
such bound can be obtained by noticing that @xmath1030 where the summation ranges over all @xmath271 in the first @xmath994 levels of the recursion tree .
hence , @xmath1031 where we used that @xmath1022 .
hence , the probability that a point falls inside @xmath594 is at most @xmath1032 / ( 1-\gamma ) \over ( 1-\gamma ) { \cal u}[{t}^{\ast}_i ] } \le { 1 \over ( 1-\gamma)^2 } \cdot { s_1 \cdot { e \cdot c \cdot \ln \ln n
\over n } \over s_1/2 } \le { 2 \cdot e \cdot c \cdot \ln \ln n \over ( 1-\gamma)^2 n}.\ ] ] therefore , the expected number of points falling in @xmath594 is at most @xmath1033 the final part of the proof is a charging argument identical to the one in lemma [ lem : competitive ratio in the good event ] .
we have thus established the following : [ lem : good case competitive ratio - product ] for @xmath1034 , there exists an event @xmath1021 , with @xmath1035 \ge 1-{2 \over \left ( \ln n \right)^{{{0.5 c ( 1-\gamma)^2}}-1}},$ ] such that the expected performance ratio of the chord algorithm conditioning on @xmath1021 is @xmath1036 .
we can also show the analogue of lemma [ lem : pessimistic competitive ratio ] .
namely , [ lem : pessimistic competitive ratio - product ] for @xmath1037 , there exists an event @xmath1038 , with @xmath1039 \ge 1-{2 \over n^{{{0.5 c ' ( 1-\gamma)^2}}-1}}$ ] , such that the expected performance ratio of the chord algorithm conditioning on @xmath1038 is @xmath1040 .
the proof is similar to the proof of lemma [ lem : good case competitive ratio - product ] , except that the bound is now a bit trickier . for @xmath1037 ,
let @xmath1038 be the event that * all the triangles maintained by the algorithm at depth @xmath1041 of its recursion tree have area at most @xmath1042 * for every node @xmath230 inside the first @xmath1043 levels of the recursion tree @xmath1044 where @xmath1045 is defined as in the statement of lemma [ lem : area recursion product distn ] . using arguments similar to those in the proof of lemma [ lem : area recursion product distn ] and the union bound , we obtain that @xmath1046 \ge 1-{2 \over n^{{{0.5 c ' ( 1-\gamma)^2}}-1}}.\ ] ] now let @xmath1047 and @xmath933 be the recursion tree of the algorithm pruned at level @xmath926 .
we also define the set @xmath1048 as in the proof of lemma [ lem : competitive ratio in the good event ] , ( but with @xmath1049 replaced by @xmath1050 ) . as in that proof , any @xmath84-cp set needs to use at least @xmath1051 points .
moreover , the total number of nodes inside @xmath933 is at most @xmath1052 whenever the algorithm processes a triangle , a planar region of area at most @xmath1053 is removed from @xmath577 ( the root triangle ) .
therefore , after finishing the processing of the first @xmath1054 levels of the tree , a total area of at most @xmath1055 is removed from @xmath577 .
we distinguish two cases . if @xmath1056 , then the size of the optimum is at least @xmath1057 points . since there is a total of @xmath55 points ( and the algorithm never performs more than @xmath1058 @xmath108 calls ) , it follows that in this case the performance ratio is @xmath1059 . on the other hand ,
if @xmath1060 , then the total area that has been removed from @xmath577 is at most @xmath1061 hence , the remaining area is at least @xmath1062 , assuming @xmath1063 . given this bound it follows that the expected number of points inside a triangle at level @xmath1064 of the recursion tree is at most @xmath1065 using the aforementioned and noting that the performance ratio `` paid '' within the first @xmath1066 levels of the recursion tree is at most @xmath1067 , we can conclude the proof via arguments parallel to those in the proof of lemma [ lem : competitive ratio in the good event ] .
now let us choose @xmath1068 and @xmath1069 . from lemmas [ lem : good case competitive ratio - product ] and
[ lem : pessimistic competitive ratio - product ] we have that @xmath1070 \ge 1- { 2 \over ( \ln n)^3}~~\text{and}~~\pr[{\cal b}_{c ' } ] \ge 1- { 2 \over n}.\ ] ] given this , we can conclude the proof theorem [ th : competitive ratio product delta - balanced distribution ] .
the argument is the same as the end of the proof of theorem [ th : competitive ratio ppp ] , except that we can now trivially bound the performance ratio of the algorithm by @xmath982 in the event @xmath1071 . in this section
we show that our upper bounds on the expected performance of the algorithm are tight up to constant factors .
in particular , for the case of the poisson point process we prove : [ thm : lower bound ppp ] let @xmath577 be the triangle at the root of the chord algorithm s recursion tree , and suppose that points are inserted into @xmath577 according to a poisson point process with intensity @xmath194 .
there exists an infinite family of instances ( parameterized by @xmath84 and @xmath8 ) on which the expected performance ratio of the chord algorithm is @xmath1072 . in particular
, we can select the parameters so that @xmath1073 , which yields a lower bound of @xmath1074 .
the lower bound construction is reminiscent to the worst - case instance of section [ ssec : worst - lower ] . in particular
, the initial triangle @xmath1075 ( at the root of the recursion tree ) will be right and @xmath1076 . to avoid clutter in the expressions
, we present the proof for the case @xmath217 .
the generalization for all values of @xmath8 is straightforward .
we fix @xmath1077 , @xmath1078 and @xmath1079 and select the intensity of the poisson process to be @xmath1080 .
note that for this setting of the parameters we have that @xmath1081 , we thus obtain an @xmath1082 lower bound .
given the endpoints @xmath1083 , it is clear that @xmath1084 with probability @xmath396 .
hence , it suffices to show that the chord algorithm makes @xmath1085 @xmath108 calls in expectation before it terminates . to show this ,
we are in turn going to prove that for @xmath84 being a sufficiently small positive constant , with constant probability , there exists a path @xmath1086 of length @xmath1087 in the recursion tree of the algorithm .
as shown in lemma [ lem : hd - vs - rd ] , for such instances the ratio distance is very well approximated by the horizontal distance .
in particular , consider the triangle @xmath585 ( see figure [ fig : chord - avg - lb ] ) , where @xmath1088 .
for any point @xmath1089 , we have that @xmath1090 hence , as long as @xmath1091 ( the slope of @xmath592 ) is sufficiently large , we can essentially use the horizontal distance metric as a proxy for the ratio distance .
indeed , this will be the case for our lower bound instance below .
we now proceed to formally describe the construction .
the path @xmath1086 of length @xmath1087 will be defined by the triangles with @xmath1092 , i.e. , the ones corresponding to the right subproblem in every subdivision performed by the algorithm . for notational convenience
, we shift the coordinate system to the point @xmath1093 , so that @xmath1094 , @xmath1095 and @xmath1096 .
( note that the horizontal distance is invariant under translation . )
we label the triangles in the path @xmath1097 by @xmath577 , @xmath1098 , @xmath1099 , and we let the vertices of triangle @xmath594 be @xmath1100 , @xmath1101 , and @xmath1092 .
suppose that when the chord algorithm processes the triangle @xmath594 , the @xmath108 routine returns the point @xmath229 on a line @xmath1102 parallel to @xmath592 ( as in figure [ fig : chord - avg - lb ] ) .
note that @xmath1103 and @xmath1104 .
let @xmath1105 .
let @xmath1106 .
the theorem follows easily from the next lemma : [ lem : long path ] let @xmath84 be a sufficiently small positive constant . with probability at least @xmath1107 , for all @xmath437 $ ] , @xmath1108 where @xmath1109 and @xmath1110 .
it is clear that ( [ eqn : rec - bound ] ) is satisfied for @xmath339 .
we are going to show by induction that , if ( [ eqn : rec - bound ] ) holds for some @xmath230 , it also holds for @xmath344 with probability at least @xmath1111 .
the theorem then follows by a union bound over all @xmath437 $ ] . by the similarity of the triangles
@xmath1112 and @xmath1113 we have @xmath1114 from the properties of the poisson point process we have the following : conditioning on the information available to the algorithm when it processes the triangle @xmath594 , if a measurable region inside @xmath594 has area @xmath1115 , then the number of points inside this region follows a poisson distribution with parameter @xmath1116 . hence , with probability at least @xmath1117 , any such region contains at least one point .
hence , with probability at least @xmath1118 we have that @xmath1119 . note that @xmath1120 . using and the induction hypothesis
, this implies that , with probability at least @xmath1118 , @xmath1121 hence @xmath1122 as desired . on the other hand ,
if the area of a region is no more than @xmath1123 the probability that a point is contained in that region is at most @xmath1124 .
similarly , this implies that , with probability at least @xmath1125 , @xmath1126 by the properties of the poisson point process it follows that the point @xmath229 is uniformly distributed on the segment @xmath1102 .
hence , with probability at least @xmath1127 , it holds @xmath1128 a union bound concludes the proof .
we now show how the theorem follows from the above lemma .
first note that , for all @xmath230 it holds @xmath1129 hence , by lemma [ lem : long path ] and the choice of @xmath394 , it is easy to check that with probability at least @xmath1130 , we have @xmath1131 and @xmath1132
the latter condition implies that the horizontal distance is a very good approximation to the ratio distance .
the latter , combined with the first condition , implies that the node ( corresponding to the triangle ) @xmath1133 is not a leaf of the recursion tree .
that is , all the triangles @xmath1134 survive in the recursion tree , since the chord algorithm does not have a certificate that the points already computed are enough to form an @xmath84-cp set .
this concludes the proof of the theorem . an analogous result can be shown similarly for the case of @xmath55 points drawn from a balanced distribution
we studied the chord algorithm , a simple popular greedy algorithm that is used ( under different names ) for the approximation of convex curves in various areas .
we analyzed the performance ratio of the algorithm , i.e. , the ratio of the cost of the algorithm over the minimum possible cost required to achieve a desired accuracy for an instance , with respect to the hausdorff and the ratio distance .
we showed sharp upper and lower bounds , both in a worst case and in an average setting . in the worst case
the chord algorithm is roughly at most a logarithmic factor away from optimal , while in the average case it is at most a doubly logarithmic factor away .
we showed also that no algorithm can achieve a constant ratio in the worst - case , in particular , at least a doubly logarithmic factor is unavoidable .
we leave as an interesting open problem to determine if there is an algorithm with a better performance than the chord algorithm ( both in the worst - case and in average case settings ) , and to determine what is the best ratio that can be achieved .
another interesting direction of further research is to analyze the performance of the chord algorithm in three and higher dimensions , and to characterize what is the best performance ratios that can be achieved by any algorithm . | the chord algorithm is a popular , simple method for the succinct approximation of curves , which is widely used , under different names , in a variety of areas , such as , multiobjective and parametric optimization , computational geometry , and graphics .
we analyze the performance of the chord algorithm , as compared to the optimal approximation that achieves a desired accuracy with the minimum number of points .
we prove sharp upper and lower bounds , both in the worst case and average case setting . |
the first papers about simple population models with complex dynamics are @xcite .
the main method of analysis of these models is the reduction to discrete equations : the logistic and ricker s equation @xcite . in @xcite ,
liu and gopalsamy investigated the following equation with piecewise constant argument @xmath0 ) \big\},\:\:\:t>0,\ ] ] where @xmath1 are positive constants and @xmath2 $ ] denotes the greatest integer function .
the authors showed that for certain parameter values of @xmath3 and @xmath4 , equation ( [ e5 ] ) generates li - yorke chaos @xcite . in the present paper
the approaches of @xcite and @xcite are developed for different types of the logistic equation with piecewise constant argument .
transformations of the dependent and independent variables are used to obtain convenient discrete equations for the dynamic analysis . the li - yorke theorem @xcite
is referred to prove chaos .
a connection between solutions of continuous models and discrete equations is used to make appropriate simulations . in the paper following equations
are considered : @xmath5))\,x([t]),\\ \label{e8}\frac{d\,x(t)}{dt}&=&(a - b\,x([t+1]))\,x(t),\\ \label{e9}\frac{d\,x(t)}{dt}&=&(a - b\,x([t+1]))\,x([t+1]).\end{aligned}\ ] ] it is seen that we suggest to involve not only delayed , but also advanced arguments in the population models .
although the role of delay in the population dynamics has been discussed vitally @xcite , the anticipation phenomena has not been considered yet .
anticipatory assumption in a population model may mean that _ a will , a wish , an anticipation _ is taken into account .
it can be assumed that anticipation is a prediction reached by the decisions of the present time .
we introduce anticipation in our population models via function @xmath6 $ ] @xcite .
we show that for critical values of the parameters @xmath7 the solutions of the differential equations ( [ e7]),([e8]),([e9 ] ) show intermittency which is `` almost periodic '' behavior interrupted by chaotic motions @xcite . in the population models ,
we consider @xmath8 , where @xmath9 denotes the size of the population at time @xmath10 and @xmath11 is a positive integer , let say , the average value of a population .
thus , @xmath12 does not represent the size of the population and it can be negative .
let us start with equation ( [ e7 ] ) .
if @xmath13 , @xmath14 , it takes the following form @xmath15 then , @xmath16 and , hence @xmath17 if one makes the change of variable @xmath18 in ( [ e13 ] ) , then obtains @xmath19 where @xmath20 .
the right - hand side of equation ( [ e14 ] ) is the logistic map @xmath21 when @xmath22 , @xmath23 generates chaos through period - doubling ( see @xcite for more details ) . since this map is obtained from the solution of the equation ( [ e11 ] ) and @xmath20 , it is obvious that equation ( [ e7 ] ) can generate chaos for @xmath24 . in @xcite , one can find that @xmath23 has intermittent behavior at @xmath25 .
therefore , equation ( [ e12 ] ) , and consequently , ( [ e7 ] ) , displays intermittency , too .
let @xmath26 , @xmath27 and @xmath28 .
one can see the intermittency phenomena for equation ( [ e7 ] ) in figure [ fig : intermittency x ] . ) ) \,x([t])$ ] with @xmath28 , @xmath29 , @xmath27.,width=360 ] let us consider another equation ( [ e8 ] ) .
if @xmath30 , then @xmath31 and @xmath32 hence , @xmath33 now the transformation @xmath34 , @xmath35 , in ( [ e17 ] ) yields @xmath36 where @xmath37 is a negative integer . the right - hand side of equation ( [ e18 ] ) is a function of the form @xmath38 in their article , may and oster @xcite discussed the behavior of the following discrete - time equation : @xmath39 they proved that for certain values of @xmath40 equation ( [ mo ] ) has fixed points of period @xmath41 for @xmath42 and it generates chaos .
below we will try to extend their results to equation ( [ e160 ] ) for @xmath43 and @xmath44 .
now let us consider the fixed points of @xmath45 , @xmath46 and @xmath47 for different values of @xmath3 with @xmath44 .
consider the value @xmath48 which is borrowed from @xcite , the mapping @xmath47 is tangent to @xmath49 line , as shown in figure [ periodic2teget ] .
and @xmath49 when @xmath50 and @xmath44.,width=360 ] when @xmath51 , the mapping @xmath47 has extra fixed points which are denoted by black stars in figure [ periodic3 ]
. then , there exist period three points which are not period one and two
. consequently , equation ( [ e8 ] ) admits the chaos through li and yorke theorem @xcite .
and @xmath49 when @xmath51 and @xmath44.,width=360 ] for values of @xmath3 just above @xmath52 , the system displays intermittency . in common ,
simulations of the corresponding discrete equation ( [ e18 ] ) are realized , but we propose to see the complex behavior in its original form .
thus , to compute @xmath53 for @xmath54 let us apply the following program .
first , we fix @xmath55 with a negative integer @xmath37 .
then , we calculate the sequence @xmath56 , by using ( [ e18 ] ) and , then @xmath34 , @xmath35 and @xmath57 .
substituting values of @xmath58 and @xmath59 in ( [ e160 ] ) , we obtain the solution of equation ( [ e110 ] ) . when @xmath60 , @xmath44 and @xmath61 , the result of simulation is seen in the figure [ fig : intermittency x(t ) ] .
for @xmath60 , @xmath44 , @xmath62.,width=360 ] let @xmath63 and @xmath64 . in figure [ chaos_2 ] , one can see that equation ( [ e160 ] ) generates chaos . for @xmath63 and @xmath65.,width=360 ] our last system is as follows @xmath66))\,x([t+1]),\ ] ] where @xmath3 and @xmath4 are constants with @xmath67 the corresponding discrete - time equation is @xmath68 let @xmath34 and @xmath69 , the last equation can be written in the following form @xmath70 where @xmath37 is a negative integer .
the right - hand side of equation ( [ e21 ] ) is a function of the form @xmath71 where @xmath72 .
similarly to the equation ( [ e15 ] ) , the last one generates complex dynamics .
we discuss the complex behavior of different types of logistic equations with piecewise constant argument of delay and advance types .
the idea of anticipation and piecewise constant argument are used together .
transformations of the space and time variables are used to obtain proper discrete - time equations .
the parameter values of the discrete - time equations which cause chaos and intemittency are utilized to get analogues for continuous solutions .
simulations of the continuous dynamics are given .
m. u. akhmet , h. ktem , s. pickl , g. w. weber , an anticipatory extension of malthusian model , seventh international conference on computing anticipatory systems , computing anticipatory system , casys05 , 2006 , pp .
260 - 264 . | we consider the logistic equation with different types of the piecewise constant argument .
it is proved that the equation generates chaos and intermittency .
li - yorke chaos is obtained as well as the chaos through period - doubling route .
basic plots are presented to show the complexity of the behavior .
logistic equation ; piecewise constant argument ; chaos ; intermittency . |
to solve the grpa equations ( 8) and ( 9 ) , the bare susceptibilities @xmath105 and @xmath106 are evaluated using the mean - field green s function in eq .
( 2 ) to give @xmath107 where we introduced the tensor @xmath108 following ref .
@xcite , we introduce a column vector @xmath109(@xmath110 ) and a @xmath111 matrix @xmath112 as @xmath113 the grpa equations are cast into the form @xmath114 the above equations are easily solved to give we derive the analytic expression of the energy dispersion of the higgs mode for small momentum @xmath117 following the approach of ref .
the gap equation ( [ gapeq ] ) can be rewritten as @xmath118 we denote @xmath119 , @xmath120 , @xmath121 , and @xmath69 . subtracting eq .
( [ gapeq2 ] ) from @xmath122=0 $ ] which is equivalent to eq .
( [ polehiggs ] ) , we obtain @xmath123 where @xmath124 here , we have replaced @xmath125 . at @xmath77 , eq . ( [ higgs_q ] ) reduces to @xmath126}=0 , \label{higgs_q0}\end{aligned}\ ] ] where @xmath127 and @xmath128 is the density of states of the fermion energy band . if we set @xmath129 , the denominator in the integrand of eq .
( [ higgs_q0 ] ) is proportional to @xmath130 , while the numerator is proportional to @xmath131 for small @xmath131 because @xmath132 .
the integral in eq .
( [ higgs_q0 ] ) is thus well defined in the limit @xmath133 . in this limit , eq .
( [ higgs_q0 ] ) is satisfied when @xmath134 and consequently the energy of the higgs mode is obtained as @xmath135 .
@xmath138=-t^2\left\{(\sin(\bm p\cdot\bm a_1)+\sin(\bm p\cdot\bm a_3))(\bm q\cdot\bm a_1)+(\sin(\bm p\cdot\bm a_2)-\sin(\bm p\cdot\bm a_3))(\bm q\cdot\bm a_2)\right\},\\ & & w_1'={\rm im}[\gamma^*\delta\gamma_1]=t^2\left\{(1+\cos(\bm p\cdot\bm a_1)+\cos(\bm p\cdot\bm a_3))(\bm q\cdot\bm a_1)+(1+\cos(\bm p\cdot\bm a_2)+\cos(\bm p\cdot\bm a_3))(\bm q\cdot\bm a_2)\right\},\\ & & w_2={\rm re}[\gamma^*\delta\gamma_2]=-\frac{t^2}{2}\left\{(1+\cos(\bm p\cdot\bm a_1)+\cos(\bm p\cdot\bm a_3))(\bm q\cdot\bm a_1)^2+(1+\cos(\bm p\cdot\bm a_2)+\cos(\bm p\cdot\bm a_3))(\bm q\cdot\bm a_2)^2\right\},\\ & & w_2'={\rm im}[\gamma^*\delta\gamma_2]=-\frac{t^2}{2}\left\{(\sin(\bm p\cdot\bm a_1)+\sin(\bm p\cdot\bm a_3))(\bm q\cdot\bm a_1)^2+(\sin(\bm p\cdot\bm a_2)-\sin(\bm p\cdot\bm a_3))(\bm q\cdot\bm a_2)^2\right\},\\ & & |\gamma_{\bm p+\bm q}|^2\simeq |\gamma_{\bm p}|^2+s_1+s_2,\ \
s_1=2w_1,\ s_2=|\delta\gamma_1|^2 + 2w_2,\\ & & |\delta\gamma_1|^2=t^2\left\{(\bm q\cdot\bm a_1)^2 + 2\cos(\bm p\cdot\bm a_3)(\bm q\cdot\bm a_1)(\bm q\cdot\bm a_2)+(\bm q\cdot\bm a_2)^2\right\},\end{aligned}\ ] ] one finds that eq .
( [ higgs_q_sub ] ) becomes @xmath139 since the factor @xmath140 is odd for @xmath141 , the summation for @xmath141 including this factor vanishes .
similarly , @xmath142 can be approximated as @xmath143 in evaluating further eqs .
( [ x ] ) , ( [ ref ] ) , and ( [ imf ] ) , we encounter the factor @xmath144 here , we replace @xmath145 in the integrand with its value at the @xmath146 ( @xmath147 ) point , i.e. , @xmath148 . at half - filling ,
since the fermi level is at the @xmath146 ( @xmath147 ) point , this replacement is justified for small @xmath10 . as a result , @xmath149 and
@xmath150 vanish and we finally obtain @xmath151 where @xmath74 is the fermi velocity . consequently , the dispersion of the higgs mode is obtained as @xmath152 in the limit @xmath153 , @xmath154 is approximated as @xmath155 which is plotted in fig . 1 ( f ) . on the other hand , at the transition point with @xmath85 ,
the dispersion for small @xmath10 coincides with that of the lower edge of the continuum as @xmath156 we derive the energy dispersion of the ab / leggett mode for small momentum . subtracting eq .
( [ gapeq2 ] ) from @xmath157=0 $ ] which is equivalent to eq . ( 11 ) , we obtain @xmath158 where @xmath159 setting @xmath77 , we obtain @xmath160 where @xmath161 thus , eq . ( [ ab_q ] ) reduces to @xmath162 from @xmath163 , we obtain the gapless ab / leggett mode : @xmath164 .
note that the term within parentheses in the numerator of eq .
( [ yq0 ] ) at @xmath165 is found to give @xmath166 for small @xmath136 , expanding @xmath167 , @xmath142 , and @xmath168 to second order in @xmath136 , we obtain @xmath169 in eqs .
( [ yq ] ) and ( [ frq ] ) , we used the same approximation as the one for eq .
( [ x - r ] ) . as a result
, we obtain @xmath170 we have used eq .
( [ d_2 ] ) to derive the final expression .
the pole is thus given by @xmath171 note that @xmath172 from the gap equation ( 2 ) . at the transition point ( @xmath85 ) , @xmath173 and thus the ab / leggett mode becomes degenerate with the higgs mode as well as the edge of the continuum as @xmath174 . | we study the higgs amplitude mode in the @xmath0-wave superfluid state on the honeycomb lattice inspired by recent cold atom experiments .
we consider the attractive hubbard model and focus on the vicinity of a quantum phase transition between semi - metal and superfluid phases . on either side of the transition
, we find collective mode excitations that are stable against decay into quasiparticle - pairs . in the semi - metal phase ,
the collective modes have `` cooperon '' and exciton character .
these modes smoothly evolve across the quantum phase transition , and become the anderson - bogoliubov mode and the higgs mode of the superfluid phase .
the collective modes are accommodated within a window in the quasiparticle - pair continuum , which arises as a consequence of the linear dirac dispersion on the honeycomb lattice , and allows for sharp collective excitations .
bragg scattering can be used to measure these excitations in cold atom experiments , providing a rare example wherein collective modes can be tracked across a quantum phase transition . _
introduction_ spontaneous symmetry breaking of continuous symmetries gives rise to two typical collective excitations - gapless goldstone modes and a gapped amplitude mode , also called the higgs mode @xcite .
while the goldstone mode has been observed in various contexts , the higgs mode has evaded observation with rare exceptions such as nbse@xmath1 , which has coexisting charge density wave and superconducting order @xcite and multiferroic ba@xmath1coge@xmath1o@xmath2 @xcite .
remarkably , two recent experiments have successfully observed this mode by tracking collective excitations across a quantum phase transition .
the first involves pressure studies of tlcucl@xmath3 , a magnetic material which undergoes a transition from dimer order to magnetic order @xcite .
the second is the realization of the bose - hubbard model in ultracold gases , with a visible amplitude mode near the superfluid - mott transition @xcite . in this letter
, we propose a novel scheme to observe the higgs mode in a _ fermi _ superfluid .
hitherto , the higgs mode has never been observed in fermi superfluids as it decays into pairs of quasiparticles .
our proposal circumvents this issue by exploiting a special feature of the honeycomb lattice geometry which allows for a window in the quasiparticle - pair continuum the higgs mode survives as a stable excitation inside this window .
inspired by the recent realization of the honeycomb optical lattices in cold atom experiments @xcite , we study the attractive hubbard model in this geometry : @xmath4 parameter @xmath5 denotes hopping amplitude between nearest neighbor ( @xmath6 ) and next - nearest neighbor sites ( @xmath7 ) .
@xmath8 is an on - site attractive interaction and @xmath9 is the chemical potential . we envisage a setup with a deep optical lattice to trap two hyperfine species of fermions , and a magnetic field on the attractive side of a feshbach resonance @xcite .
this model hosts a superfluid state of dirac fermions , with several interesting implications @xcite . in this proposal , we make use of two key features : ( i ) strictly at half - filling , there is an interaction - tuned quantum phase transition from a semi - metal phase to an @xmath0-wave superfluid .
this has been demonstrated by sophisticated quantum monte carlo simulations on very large system sizes @xcite .
this transition is a consequence of the dirac cone dispersion which leads to vanishing density of states at the fermi level , thereby necessitating a critical interaction strength to induce superfluid order @xcite .
( ii ) in the semi - metal phase , the two - particle continuum has a window structure , again a consequence of the dirac cone dispersion @xcite .
a collective mode excitation propagating inside this window is stable against decay into quasiparticle - pairs .
we show that this window structure persists in the superfluid phase , thus allowing for a stable higgs mode excitation . as obtained from the mean - field theory ( b ) .
evolution of the elementary excitations across the quantum critical point ( qcp ) of the semi - metal ( sm ) to superfluid ( sf ) phase transition ( c)-(f ) . the dash - dot line in each panel and
the dash - dot - dot lines in ( f ) show the asymptotic dispersions of the continuum edge and the superfluid collective modes for small @xmath10 , respectively . ]
fig .
[ fig1](b ) shows the phase diagram of this model at half - filling .
our key findings are summarized in fig .
[ fig1](c)-(f ) : ( i ) there are sharp collective mode excitations on either side of the transition . the two - particle continuum is shown as the shaded region , note the window structure .
( ii ) in the semi - metal phase , there are three degenerate collective modes with `` cooperon '' and exciton character .
( iii ) on the superfluid side , there is a goldstone mode and remarkably , a distinct superfluid amplitude ( higgs ) mode .
these modes can be observed in a cold atoms experiment using bragg scattering .
this is a rare example wherein relevant collective excitations can be tracked across a quantum phase transition .
_ mean field theory_ the eth group @xcite has studied fermions loaded onto a honeycomb optical lattice with tunable anisotropy .
we consider the attractive hubbard model in the _ isotropic _ honeycomb lattice . as discussed in ref .
@xcite , the isotropic limit is expected to have the highest superfluid transition temperature and is the most promising for experimental realization .
we decompose the hubbard interaction in the superfluid channel using the order parameter @xmath11 , taken to be real . for brevity ,
we introduce a vector operator consisting of creation and annihilation operators @xmath12 ( @xmath13 and @xmath14 denote the two sublattices as shown in fig .
[ fig1](a ) ) .
the mean field hamiltonian can be written as @xmath15 , where @xmath16 , @xmath17 and @xmath18-\mu$ ] with @xmath19 and @xmath20 being the two basis vectors shown in fig .
[ fig1 ] ( a ) .
we take the lattice spacing to be unity . @xmath21 and @xmath22 are the pauli matrices in the nambu and sublattice space , respectively .
the single - particle green s function for the mean - field hamiltonian is given by @xmath23^{-1 } \equiv\left ( \begin{array}{cc } \hat g^{aa}(p ) & \hat g^{ab}(p)\\ \hat g^{ba}(p ) & \hat g^{bb}(p ) \end{array } \right).\label{greenf}\end{aligned}\ ] ] here , we denote @xmath24 , where @xmath25 is the fermion matsubara frequency .
the gap and number equations are obtained from the off - diagonal and diagonal elements of the green s function @xmath26 as @xcite ( hereafter , we restrict ourselves to zero temperature ) @xmath27 where @xmath28 is the spectrum of the bogoliubov quasiparticles , @xmath29 , and @xmath30 is the number of lattice sites .
at half - filling , the self - consistent solution of @xmath31 becomes non - zero for @xmath32 indicating a transition from semi - metal to superfluid phases @xcite . for @xmath33 ,
mean - field theory gives @xmath34 .
quantum monte carlo gives the same transition , except with @xmath35 renormalized to @xmath36 @xcite . in the rest of this letter , we use mean - field results with the understanding that fluctuations will renormalize @xmath8 quantitatively .
we note that @xmath35 weakly depends on the value of @xmath37 .
_ generalized random phase approximation ( grpa)_ on either side of the critical point , there are low - lying density and pairing fluctuations .
we use a generalized random phase approximation ( grpa ) scheme to evaluate density and pairing response functions .
we follow the green s function approach of ct and griffin @xcite .
we denote matrix susceptibilities containing the response to weak density and pairing perturbations , respectively , as @xmath38 where @xmath39 ( @xmath40 is a boson matsubara frequency ) .
any susceptibility @xmath41 is defined as @xmath42 where @xmath43 ( @xmath44 denotes the unit cell , @xmath45 the sublattice , and @xmath46 an imaginary time ) , @xmath47 , @xmath48 and @xmath49 .
the density and pair annihilation operators are written as @xmath50 and @xmath51 , respectively .
the grpa equations read @xcite @xmath52 where @xmath53 is either @xmath54 or @xmath55 and @xmath56 .
@xmath57 denotes the bare susceptibility @xcite , @xmath58 includes an infinite sum over ladder diagrams , while @xmath59 is the final result which also includes bubble diagrams .
_ undamped higgs mode_ in the superfluid phase , we solve grpa equations ( [ grpa2 ] ) and ( [ grpa1 ] ) to evaluate the amplitude and phase correlation functions @xmath60 and @xmath61 .
the amplitude and phase fluctuation operators are given by @xmath62 and @xmath63 , respectively .
for the case of @xmath33 , the expressions simplify and we can identify their respective poles , which we denote `` higgs '' and `` ab / leggett '' .
these poles satisfy @xmath64 we have defined @xmath65 here , we have denoted @xmath66 , @xmath67 , @xmath68 , and @xmath69 . on the other hand , solving eqs .
( [ grpa2 ] ) and ( [ grpa1 ] ) for density response , we find that @xmath70 when @xmath33 .
thus , the density response function only retains the ab / leggett pole given in eq .
( [ poleab ] ) . setting @xmath71 in eq .
( [ poleab ] ) , we recover the gap equation ( [ gapeq ] ) .
thus , the superfluid phase has _ gapless _
collective mode(s ) arising from phase fluctuations .
in fact , at half - filling , the ab / leggett pole in eq .
( [ poleab ] ) is a double pole corresponding to two gapless modes : the anderson - bogoliubov ( ab ) mode and the leggett mode @xcite .
the ab mode is the usual goldstone mode associated with u(1 ) symmetry breaking @xcite .
the leggett mode is composed of out - of - phase fluctuations between sublattices @xcite - it acquires a gap away from half - filling @xcite .
the ab and leggett modes become degenerate at half - filling reflecting a special pseudospin su(2 ) symmetry of the hubbard model @xcite . for small @xmath72 ,
eq .
( [ poleab ] ) gives the dispersion relation of the ab / leggett mode to be @xcite @xmath73 where @xmath74 is the fermi velocity at the dirac points .
setting @xmath75 , eq .
( [ polehiggs ] ) also reduces to the gap equation .
thus , there exists a _
gapped _
collective mode with the energy gap @xmath76 at @xmath77 .
this is the ` higgs ' mode or the amplitude mode @xcite arising from amplitude fluctuations of the superfluid order parameter .
it can be understood using the mechanical analog of motion along the radial direction of the famous `` mexican hat '' potential ; the energy gap stems from the finite curvature of the potential along the radial direction .
remarkably , the higgs mode disperses below the quasiparticle pair continuum in figs .
[ fig1](e ) and ( f ) . in particular , close to the @xmath55 point
, it is well separated from the lower edge of the continuum .
this is to be contrasted with the case of typical superfluids : due to the underlying fermi surface , the continuum exhibits a horizontal edge near @xmath78 @xcite .
the higgs mode therefore enters the continuum , becomes heavily damped and is unobservable . in our case , the higgs mode in fig .
[ fig1 ] is undamped over large sections of the brillouin zone . for @xmath72 , solving eq .
( [ polehiggs ] ) , the higgs mode has the dispersion relation @xmath79 @xcite .
the higgs mode disperses below the lower edge of the continuum which is given by @xmath80\simeq 4\delta^2 + 2v_f^2q^2 $ ] ( @xmath72 ) .
this window or arch in the continuum , shown in fig .
[ fig1 ] , is a consequence of the dirac - like dispersion of underlying fermions .
the higgs mode stays undamped as long as it lies within this window .
even if we go slightly away from half - filling , the higgs mode survives undamped .
close to @xmath81 or @xmath82 , the window disappears because of the presence of the fermi surface - as a result , the higgs mode is strongly damped .
the higgs mode , the ab mode , and the lower edge of the continuum become degenerate at the qcp for @xmath72 : @xmath83 .
the ab mode and leggett mode are strongly coupled with density fluctuations ; they appear as poles in the density response function @xmath84 . however , when @xmath33 , the higgs mode has no corresponding pole in the density response function .
thus , the higgs mode is composed of pure amplitude fluctuations and can not be excited by a density perturbation .
this reflects the underlying su(2 ) pseudospin symmetry @xcite in the problem . a small finite @xmath37 breaks this symmetry and forces the higgs mode to acquire a density component :
the density response then shows a peak at the higgs mode , as shown in fig.[fig2 ] . _ collective modes in the semi - metal phase : cooperons and excitons_ in the semi - metal phase , setting @xmath85 , density and pair response functions become decoupled in the grpa equations ( [ grpa2])-([grpa1 ] ) .
the susceptibilities satisfy the usual rpa equations @xmath86 , @xmath87 .
the bare susceptibility @xmath88 describes a single rung diagram with particle - particle ( hole - hole ) excitations , and @xmath89 describes a single bubble diagram with particle - hole excitations .
they are given by @xmath90,\ \
\ \label{barem}\\ & \chi_{n_\sigma n}^{0\nu_1\nu_2}(q)=\frac{1}{2n}\sum_{\bm p}\left[\frac{-\kappa^{\nu_1\nu_2}_{\bm p+\bm q}\eta^{\nu_2\nu_1}_{\bm p}}{\xi_{\bm p+\bm q}^+ -\xi_{\bm p}^--i\omega_n}+\frac{\eta_{\bm p+\bm q}^{\nu_1\nu_2}\kappa_{\bm p}^{\nu_2\nu_1}}{\xi_{\bm p+\bm q}^- -\xi_{\bm p}^+-i\omega_n}\right],\ \
\ \label{barel}\end{aligned}\ ] ] where @xmath91 and @xmath92 .
we denote @xmath93 . from the denominator in eq .
( [ barem ] ) , we see that pairing response arises from particle particle ( or hole hole ) excitations .
an undamped pairing mode , occurring below the particle particle continuum in fig .
[ fig1](c ) , is therefore a two particle bound state with well defined momentum and energy .
it can be understood as a preformed cooper pair we call this a _ cooperon _ excitation @xcite . similarly , from eq . ( [ barel ] ) , we see that density response arises from particle hole excitations .
an undamped density mode is thus a particle - hole bound state
we call this an _ exciton_. the dispersions of cooperons and excitons are determined by the poles of the corresponding response functions , giving @xmath94 and @xmath95 . with @xmath33 ,
these reduce to the identical equation @xmath96 thus , the cooperon and the exciton are degenerate when @xmath33 .
their dispersion is shown in figs .
[ fig1 ] ( c ) and ( d ) the modes are undamped as they lie below the two particle continuum . in particular
, they are well separated from the continuum in the vicinity of the @xmath55 points .
we suggest that experiments should probe this region to observe the collective excitations .
this feature of the @xmath55 points can be understood from the single particle band structure which has saddle points at these wavevectors .
they consequently have a very large density of states which provides large phase space for the hubbard interaction to form two particle bound states . these collective modes in the semi - metal phase
were predicted many years ago using an insightful single - cone approximation ref .
@xcite reported a triplet exciton mode in the repulsive hubbard model .
the authors identified the window structure in the continuum as capable of accommodating stable modes .
mapping their results to the attractive hubbard case @xcite , the triplet excitons translate to cooperon and exciton modes .
we reaffirm their prediction starting from a microscopic picture taking into account the sublattice structure .
our expressions also agree with those of ref .
@xcite which only considers the @xmath97 segment and concludes that there is no undamped mode . however , we find an undamped mode in the @xmath98 and @xmath99 directions .
_ cooperon condensation_ as we approach the critical point from the semi - metal side , the energy of the cooperon and exciton decreases progressively ( see fig . [
fig1](c ) and ( d ) ) . precisely at the transition
, the cooperon `` softens '' at @xmath77 and undergoes condensation .
in fact , setting @xmath71 , the cooperon pole in eq .
( [ pole_cooperon ] ) reduces to the gap equation ( [ gapeq ] ) .
since cooperons and excitons are degenerate for @xmath33 , the exciton can also condense at the critical point .
that gives rise to the sublattice
cdw state - which is degenerate with the superfluid state due to su(2 ) pseudospin symmetry . for @xmath100 , this degeneracy
is lifted in favour of the superfluid and the cooperon condenses preferentially . as we cross @xmath35 and enter the superfluid phase , cooperons and excitons hybridize to become the ab , leggett and higgs modes .
the excitonic component , when present , allows these modes to have peaks in the density response function .
the cooperonic component manifests as peaks in the pairing response .
thus , the collective modes evolve smoothly across the qcp and carry signatures of the underlying spontaneous symmetry breaking
. _ visibility of the higgs mode_ /\pi$ ] for @xmath101 ( a ) .
the cross section for the momentum at the @xmath55 point ( b ) .
the upper peak corresponds to the higgs mode . ]
our higgs mode is stable against decay into pairs of fermions due to the window structure in the two particle continuum .
however , if we go beyond rpa level , the higgs mode may decay by emitting ab / leggett excitations which are lower in energy .
we argue that this merely leads to broadening of the higgs mode .
in our semi - metal to superfluid transition , due to the pseudospin symmetry present when @xmath33 , the order parameter can be thought of as an o(3 ) object .
quantum monte carlo simulations of the o(3 ) ordering transition show that the higgs mode survives although it is broadened by decay into goldstone bosons @xcite .
we expect this to be true in our case as well .
we suggest bragg spectroscopy measurements on a fermi superfluid in a honeycomb optical lattice as a way to measure this undamped higgs mode . in this technique ,
a two - photon process imparts a density-``kick '' to the system .
the response to this perturbation can be quantified by measuring the momentum transferred or the energy absorbed .
the momentum transferred is a measure of the dynamic structure factor related to the density response function @xmath102/\pi$ ] @xcite it can detect collective modes as long as they have a density component . for any small @xmath100 , the higgs mode in our model has a density component which makes it visible to bragg spectroscopy .
a small @xmath37 hopping is expected to be present in an optical lattice setup any way @xcite .
figure [ fig2 ] shows @xmath103 for @xmath104 .
the sharp intensity peak for the higgs mode can be clearly seen below the continuum .
an alternative approach is to measure the energy absorption in response to a weak shaking of the optical lattice @xcite . _
acknowledgements_we acknowledge a. paramekanti for fruitful dicussions .
s.t . thanks m. sigrist , t. esslinger , l. tarruell , y. ohashi , s. okada , s. konabe , k. asano , s. kurihara , and k. kamide for discussions .
s. t. was supported by grant - in - aid for scientific research , no .
24740276 .
c. m. varma , j. low temp . phys .
* 126 * , 902 ( 2002 ) .
r. sooryakumar and m. v. klein , phys .
rev .
lett .
* 45 * , 660 ( 1980 ) .
p. b. littlewood and c. m. varma , phys .
rev .
lett . * 47 * , 811 ( 1981 ) ; phys .
rev .
b * 26 * , 4883 ( 1982 ) .
k. penc _
et al . _ ,
phys .
rev .
lett . * 108 * , 257203 ( 2012 ) .
ch .
regg _
et al . _ ,
phys .
rev .
lett .
* 100 * , 205701 ( 2008 ) .
u. bissbort _
et al . _ ,
phys .
rev .
lett . * 106 * , 205303 ( 2011 ) .
m. endres _
et al . _ , nature ( london ) , * 487 * , 454 ( 2012 ) .
l. tarruell , d. greif , t. uehlinger , g. jotzu , and t. esslinger , nature * 483 * , 302 ( 2012 ) .
i. bloch , j. dalibard , and w. zwerger , rev .
mod .
phys .
* 80 * , 885 ( 2008 ) .
e. zhao and a. paramekanti , phys .
rev .
lett . * 97 * , 230404 ( 2006 ) .
s. tsuchiya , r. ganesh , and a. paramekanti , phys .
rev .
a * 86 * , 033604 ( 2012 ) .
s. sorella , y. otsuka , and s. yunoki , scientific rep .
* 2 * , 992 ( 2012 ) .
ref .
@xcite has used quantum monte carlo simulations to establish the phase diagram of the _ repulsive _ hubbard model with @xmath33 . however
, this model can be mapped onto the _ attractive _ model using a sublattice - dependent particle - hole transformation .
see , e.g. , s .- c .
zhang , phys .
rev .
lett .
* 65 * , 120 ( 1990 ) .
p. nozieres and f. pistolesi , eur .
phys .
j. b * 10 * , 649 ( 1999 ) .
g. baskaran and s. a. jafari , phys .
rev .
lett . * 89 * , 016402 ( 2002 ) .
r. ct and a. griffin , phys .
rev .
b * 48 * , 10404 ( 1993 ) . see supplementary material for explicit expressions for the bare susceptibility @xmath57 . p. w. anderson , phys . rev . * 110 * , 827 ( 1958 ) .
y. nambu , phys . rev . * 117 * , 648 ( 1958 ) .
a. j. leggett , prog .
. phys . * 36 * , 901 ( 1966 ) .
see supplementary material for the detailed derivation of the asymptotic dispersions of the ab / leggett mode and the higgs mode .
t. m. rice , k .- y .
yang , and f. c. zhang , rep .
. phys . *
75 * , 016502 ( 2012 ) . n. m. r. peres , m. a. n. arajo , and a. h. castro neto , phys .
rev .
lett .
* 92 * , 199701 ( 2004 ) .
s. gazit , d. podolsky , and a. auerbach , arxiv:1212.3759 ( 2012 ) .
l. pitaevskii and s. stringari , _ bose - einstein condensation _ , ( oxford university press , oxford , 2003 ) .
j. ibaez - azpiroz _ et al .
_ , phys .
rev .
a * 87 * , 011602(r ) ( 2013 ) . |
in this article we consider the inverse problem of computing the refractive index of a medium from ultrasound time - of - flight ( tof ) measurements . on the one side
this task is a tomographic problem of its own and often called the _
inverse kinematic problem _ which has important applications , e.g. in seismics . on the other side the knowledge of the sound speed , resp .
refractive index , of an object @xmath0 under consideration is essential for inverse problems in inhomogeneous media such as photoacoustic or ultrasound vector tomography .
the idea is very simple : an ultrasound signal is emitted at a transmitter @xmath1 and its travel time is acquired at a detector @xmath2 , see figure [ motivation ] .
of course the tof depends of the sound speed @xmath3 in the medium . the _ refractive index _
@xmath4 , where @xmath5 denotes the constant sound speed of a reference medium like water or air outside the object , causes refractions of the ultrasound beam . assuming a constant @xmath6 ( e.g. by setting @xmath7 ) in applications such as photoacoustic or vector tomography
hence might cause severe artifacts .
we emphasized this in figure [ motivation ] .
there , the blue triangle would be detected at the wrong place if we assume that the ultrasound signal travels along a straight line and that there is no refraction at all . + to a detector @xmath2 .
if @xmath8 , then the signal no longer travels along straight lines and reconstructions which neglect refraction show artifacts .
e.g. the blue triangle would be detected at the wrong ( dotted ) place if we assumed straight lines as signal paths . ]
let us briefly illuminate the situation in vector tomography .
norton @xcite derived as mathematical model to compute a solenoidal vector field @xmath9 from tof measurements the doppler transform along straight lines @xmath10 where @xmath11 denotes the vector of direction of the line @xmath12 which connects the points @xmath1 and @xmath2 .
hence , to solve the inverse problem of computing @xmath9 from data @xmath13 it is necessary to know the sound speed @xmath3 . but _
fermat s pronciple _ says that the propagation paths are geodesic curves of the riemannian metric [ n - metric ] s^2 = ^2 ( x ) |x|^2 leading to an improved model @xmath14 where @xmath15 is a geodesic curve associated with the metric ( [ n - metric ] ) and connecting the points @xmath1 and @xmath2 .
the problem now arises that the integrand determines the integration curve @xmath15 turning the inverse problem into a nonlinear one .
also in photoacoustic tomography a variable sound speed leads to quite other analysis and numerics , see e.g. @xcite .
hence , following fermat s principle , then instead of the euclidean space @xmath16 with the euclidean metric tensor @xmath17 , we have to consider the riemannian manifold @xmath18 with metric tensor @xmath19 .
+ this is motivation to investigate the following inverse problem : given tof measurements @xmath20 for transmitter / detector pairs @xmath21 , compute the index of refraction @xmath22 satisfying [ ip - refractive ] r ( ) = u^meas , where the ray transform @xmath23 is given as [ ray - transform ] r ( ) ( x , ) : = _ ^_x , s = _
^0 ( ^_x,(t ) ) |^_x,(t)| t . here
, @xmath24 denotes a geodesic curve of the metric ( [ n - metric ] ) with @xmath25 and @xmath26 for @xmath27 a tangent vector at @xmath28 . +
this inverse problem and its research has a long lasting history .
we summarize some important references in this context .
herglotz @xcite was among the first researchers who has taken inhomogeneities into account .
he investigated the earth s inner structure by considering travel times of seismic waves .
mukhumetov @xcite proved that the determination of simple metrics in two dimensions from travel times is possible .
the extension to the three - dimensional case was achieved independently by romanov @xcite and mukhumetov @xcite .
general results on the inverse kinematic problem have been proven by stefanov and uhlmann in @xcite , chung et al .
@xcite and sharafutdinov @xcite ; a further uniqueness and stability result can be found in @xcite .
the 2d problem for anisotropic metrics was solved by pestov and uhlmann @xcite ; the approach contained therein is constructive .
the question of a unique solution , the so called _ boundary rigidity problem _ , is not entirely solved by now .
first results in 2d were achieved by michel @xcite , croke @xcite and otal @xcite .
pestov and uhlmann @xcite showed uniqueness for simple , two - dimensional riemannian manifolds .
a microlocal treatment can be found in stefanov and uhlmann @xcite .
local and semi - global results were presented by croke et al .
@xcite , stefanov and uhlmann @xcite , gromov et al .
@xcite , croke @xcite , @xcite and lassas et al .
@xcite , partly for special metrics only . for more references concerning analytical results for the inverse kinematic problem and the boundary rigidity problem
we refer to the book of sharafutdinov @xcite and the references therein .
based on the pestov - uhlmann reconstruction formulas from @xcite monard @xcite derived a numerical solver for the linear geodesic ray transform .
another numerical solution scheme which relies on beylkin s theory @xcite is presented in @xcite .
the influence of refraction to reconstruction results in 2d emission tomography have been studied in @xcite , a numerical solver for the geodesic ray transform based on b - splines is presented in @xcite . a further numerical solver for the inversion of @xmath23 ( [ ray - transform ] )
is found in klibanov and romanov @xcite ; here the linearization is done by replacing the geodesic curves by straight lines .
+ the novelty of our article is twofold : on the one hand we consider the nonlinear problem and our numerical solver linearizes @xmath23 in each iteration step using the old iterate @xmath29 to compute the geodesic curves . on the other hand we use an explicit representation of the geodesic backprojection operator and
show how to implement it . to this
end the construction of a so called _ geodesic projection _ was necessary .
+ _ outline . _ in section 2 we provide essential results from riemannian geometry which are necessary for our further considerations as well as the mathematical model for the inverse problem .
additionally we collect some mathematical properties of the nonlinear forward operator @xmath23 ( [ ray - transform ] ) .
the iterative solver which we develop in this article demands for evaluation of integrals along geodesic curves .
the computation of these curves is done using the method of characteristics which is outlined in section 2.5 .
the regularizing solution scheme is subject of section 3 .
we formulate an appropriate tikhonov functional and linearize @xmath23 for its minimization .
the derivative of the so arising functional contains the backprojection operator of the geodesic ray tranform .
we give an explicit expression of this operator using the concept of _ phase functions _ and _ geodesic projection _ yielding in that sense an analogon to the conventional 2d backprojektion operator in euclidean geometry ( section 3.1 ) . the iterative minimization scheme and its implementation
is described in section 3.2 .
section 4 finally contains numerical evaluations of the method for several refractive indices @xmath6 with exact and noisy data .
section 5 concludes the article .
we collect some fundamental results from differential geometry which are useful for our later considerations . throughout the article we assume @xmath30 to be a compact and convex domain which is seen as a submanifold of @xmath31 .
+ [ def:2_riemmetrik ] on @xmath30 we define a _
riemannian metric _ as a differentiable mapping @xmath32 , such that @xmath33 is a positive definite , symmetric bilinear form on the tangent space @xmath34 in @xmath28 .
we have for @xmath35 * @xmath36 , * @xmath37 , if @xmath38 and * for diffentiable vector fields @xmath39 is @xmath40 a differentiable mapping . here
@xmath41 is the tangent bundle on @xmath42 .
+ a representation of the metric @xmath43 with respect to local coordinates is given by g_x = _ i , j=1^n g_ij(x ) dx^i|_x dx^j|_x .
[ gl:2_defimetriksum ] the third condition is then equivalent to the requirement that the coefficient functions @xmath44 are differentiable independently of the chart .
the local coordinates @xmath45 are called _
metric tensor_. + if it is convenient , then we use the einstein notation .
that means , that we sum up over doubled indices .
the representation then becomes @xmath46 the tuple @xmath18 is called _
riemannian manifold_. + let @xmath47 be the metric tensor of euclidean geometry ( @xmath48 ) .
then @xmath49 is a riemannian manifold and is called _
euclidean space_. + we introduce a specific metric tensor which plays a crucial role when studying ultrasound wave propagation in an inhomogeneous medium .
let @xmath3 be the speed of sound at @xmath50 and @xmath5 be the sound speed of a reference medium ( e.g. air or water ) .
then , @xmath4 denotes the _ index of refraction_. especially we assume @xmath51 in @xmath52 .
+ [ lem:2_messgeometrie ] for @xmath50 let [
metric - n ] g_x^ = g_ij^(x ) dx^i dx^j with metric tensor @xmath53 where the index of refraction @xmath54 is supposed to be positive and differentiable .
then @xmath55 is a riemannian manifold .
the element of length is then given as @xmath56 the submanifold @xmath30 can be canonically embedded into @xmath31 such that we can choose the identity as chart which is differentiable .
the metric tensor @xmath57 satisfies the requirements of definition [ def:2_riemmetrik ] , since it is symmetric ( @xmath58 ) , differentiable ( the component functions @xmath6 are differentiable ) and positive definite ( @xmath59 ) .
+ for simplicity we set @xmath60 for the rest of the article .
our aim is to model the propagation of ultrasound waves in a medium with variable sound speed by geodesic curves associated with the metric tensor ( [ metric - n ] ) .
this is due fermat s principle ( see section 2.3 ) .
we summarize the basics of geodesics . for details
we refer to standard textbooks such as @xcite .
let @xmath18 be a riemannian manifold , @xmath61 $ ] .
the curve @xmath62 is called _ geodesic curve _ or _ geodesic _ , if it satisfies the _ geodesic equation _ ^i + _ jk^i ( ) ^j ^k = 0 .
[ gl:2_geoddgl ] the _ christoffel symbols _
@xmath63 for @xmath64 are given as _
jk^i(x ) : = g^ip(x ) ( ( x ) + ( x ) - ( x ) ) , [ gl:2_christoffel ] where @xmath65 denote the coefficients of the inverse of the metric tensor @xmath66 .
we define the _ length _ @xmath67 of @xmath15 by t ( ) & = & _ ds + & = & _ _ 0^_1 dt . + equipped with initial conditions @xmath68 as starting point and @xmath69 as an initial direction , it follows by the picard - lindelf theorem that has a unique solution which is then denoted by @xmath70 . if we want to point out on which metric the geodesic depends , then we write @xmath71 . for @xmath72
the length @xmath73 coincides with the tof of an ultrasound signal emitted from @xmath28 in direction @xmath74 .
[ d - distance ] let @xmath18 be a riemannian manifold and @xmath75 .
we define the _ distance _ between @xmath28 and @xmath76 by @xmath77 any curve @xmath78 in @xmath42 which attains the infimum is called _ shortest curve _ from @xmath28 to @xmath76 .
+ in the euclidean space @xmath79 all geodesic curves are straight lines and vice versa .
particularly all geodesics are at the same time shortest paths between two points . on the sphere
@xmath80 equipped with the euclidean metric induced from @xmath81 , all great circles are geodesic curves .
but for two separate points on the sphere there are two geodesic curves which connect them and in general only one of them is a shortes curve between these points .
[ kor:2_stetigegeodaeten ] let @xmath82
. then all geodesic curves and their first derivatives depend continuously on the initial values and the refractive index @xmath6 . + setting z_1(t ) & : = & _ 1(t ) + z_2(t ) & : = & _ 2(t ) + z_3(t ) & : = & z_1(t ) = _ 1(t ) + z_4(t ) & : = & z_2(t ) = _ 2(t ) . with initial values @xmath83 und @xmath69 , we transform the geodesic equation ( [ gl:2_geoddgl ] ) in a system of first order @xmath84 where f_1(t , z(t ) ) & = & z_3(t ) , + f_2(t , z(t ) ) & = & z_4(t ) , + f_3(t , z(t ) ) & = & -^-1(z(t ) ) + & & + f_4(t , z(t ) ) & = & - ^-1(z(t ) ) . thereby @xmath85 for @xmath86 , @xmath87 compact . because @xmath6 is continuously differentiable
, @xmath88 is continuously differentiable , too .
furthermore @xmath88 fullfills a lipschitz condition , which we show by using the mean value theorem . for
any @xmath89 let @xmath90 the mean value theorem guarantees the existence of @xmath91 such that @xmath92 for all @xmath93 . since @xmath94 and @xmath95 , it is easy to prove that @xmath96 the assertion now follows from ( * ? ? ?
iii ) .
+ a riemannian metric @xmath97 on a compact manifold @xmath42 is called _ simple _ , if the boundary @xmath98 is strictly convex and every two points @xmath99 are connected by a unique geodesic curve which depends smoothly on @xmath100 .
a geodesic @xmath101\to m$ ] is called _ maximal _ if it can not be extended to a segment @xmath102 $ ] for any @xmath103 .
the metric @xmath97 is called _ dissipative _ , if it is simple and if for every point @xmath50 and vector @xmath104 the maximal geodesic @xmath105 is defined on a finite segment @xmath106 $ ] .
+ if the metric @xmath97 is simple , then obviously every geodesic is also a shortest curve in the sense of definition [ d - distance ] .
our modelling bases on an important physically axiom , _
principle_. it can be summarized as follows : + _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ a wave signal , which moves from one point to another , always follows the locally shortest path , such that the time of flight is at its minimum .
this means that the acceleration in every point disappears in path direction .
+ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ as a consequence of this axiom it follows that the signals move along geodesic curves . according to this axiom we are going to model ultrasound beams in an inhomogeneous medium with refractive index @xmath22 as geodesic curves associated with the metric [ metric - fermat ] g^ ( x ) = ^2 ( x ) ( _ ij ) .
we have now all ingredients together to describe the mathematical model of our measurement process .
[ def:2_laufzeitabbildung ] let @xmath55 be a compact riemannian manifold , where @xmath57 is the metric ( [ metric - fermat ] ) .
we call the mapping [ tof - mapping ] u : t^0 m ^+ , ( x , ) _ _ -(x,)^0 ( _ x,(t ) ) dt _ time - of - flight mapping _
( tof mapping ) , where @xmath107 @xmath108 is parametrised with respect to arc length and @xmath109 : \gamma_{x,\xi}^{\tn}(\tau ) \cap \partial m
\neq \emptyset \right\}\ ] ] is the moment where @xmath110 intersects the boundary @xmath98 for the first time . + the following definition addresses the practical situation that we have measurements at the boundary @xmath98 . in this case
we have to distinguish whether the ultrasound wave enters or leaves the domain @xmath42 .
+ we have @xmath111 ( red arrow ) at @xmath98 and @xmath74 is the tangential vector ( dotted arrow ) in @xmath28 .
there is an `` outflow '' with measured data @xmath112 . in the case of an `` inflow '' , @xmath113
, the geodesic @xmath15 would be outside from @xmath42 and hence @xmath114 . ]
let @xmath115 we call @xmath116 _ outflow _ and @xmath117 _ inflow _ , where @xmath118 is the outer normal at @xmath42 in @xmath28 .
we have @xmath119 and @xmath120 are compact manifolds .
+ the situation is illustrated in figure [ fig:2_messrichtung ] .
the following lemma is proven in @xcite .
+ [ l - tau - smooth ] let @xmath18 be a compact , dissipative riemannian manifold . then the mapping @xmath121 is a smooth function .
+ the data acquired by tof measurements can now be modeled as integrals of @xmath6 along geodesics associated with the metric @xmath57 . to this end
let @xmath122 and @xmath123 the solution of the geodesic equation ( [ gl:2_geoddgl ] ) with respect to the metric @xmath57 .
the mapping @xmath124 with u^meas(x , ) = 0 & ( x , ) _ -m + _ ^_x , ( y ) l(y ) & ( x , ) _
+ m + , [ gl:2_randwerte ] assigns a geodesic curve starting in @xmath125 with tangent @xmath27 its travel time until it leaves the domain .
this is why we call @xmath20 the _ tof mapping_. the inverse problem consists of computing the refractive index @xmath6 from @xmath20 .
the forward operator is given by the ray transform r ( ) ( x , ) : = _ ^_x , ( z ) l(z ) , ( x , ) _
+ m .[gl : operatorgleichung ] for @xmath126 we furthermore define the _ linearized ray transform _ @xmath127 where @xmath128 is the solution of with respect to the metric @xmath129 and initial values @xmath130 and @xmath131 .
if @xmath132 , we obviously have @xmath133 the inverse problem of determining the refractive index @xmath6 from tof measurements finally means to find a solution of [ ip - tof ] r ( ) = u^meas .
note that the crucial point is that the curve along which @xmath23 integrates depends on the integrand @xmath6 turning ( [ ip - tof ] ) into a highly nonlinear , ill - posed problem . from now on
we assume that the refractive index @xmath135 and @xmath55 is a compact , dissipative riemannian manifold ( cdrm ) .
we recall that this implies that for any two points @xmath99 there exists a unique , maximal geodesic @xmath15 connecting @xmath28 and @xmath76 which at the same time is the shortest path between these points . for proving the continuity of @xmath23
it is useful to introduce the phase function following the outlines of guillemin and sternberg @xcite .
compare also ( * ? ? ?
* section 2 ) . + for @xmath136 and @xmath137 , @xmath138 , we denote by @xmath139\to\rr^2 $ ] the unique solution of the geodesic equation ( [ gl:2_geoddgl ] ) with respect to the refractive index @xmath6 and initial values [ phase - setting ] _ , s(0 ) = r + s ^ _ ,
s(0 ) = . here
, @xmath140 and @xmath141 is the unique parameter where @xmath142 intersects the boundary @xmath98 for the second time and @xmath143 is perpendicular to @xmath11 .
the situation is illustrated in figure [ fig:2_phasenfunktion ] .
it is now possible to define @xmath142 as level curves of a phase function @xmath144 .
+ and @xmath145 ( red arrows ) a point @xmath146 is defined by the two parameters @xmath147 ( blue ) such that @xmath148 .
the black curves represent geodesics passing @xmath42 initiating from the straight line @xmath149 with vector of direction @xmath11 . ]
[ lem:2_phasenfunktion ]
let @xmath97 be a metric such that for @xmath150 and @xmath151 with @xmath152 the geodesic curves @xmath142 and @xmath153 do not intersect . then the matrix @xmath154 is regular for fixed @xmath11 and there exists a _ phase function
_ @xmath155 such that the geodesic @xmath156 is implicitly given and uniquely determined by @xmath157 + in that sense geodesic curves can be interpreted as manifolds of constant phase of a wave field
. + let @xmath158 and consider @xmath159 , i.e. the unit disk equipped with the euclidean metric .
then the geodsics are straight lines and @xmath160 is given by @xmath161 and @xmath162 is the usual parametrization of straight lines in parallel geometry . here , @xmath163 and @xmath164 for @xmath165 .
please note that by ( [ phase - setting ] ) the normal vector of a line is @xmath166 instead of @xmath11 as usual .
+ in euclidean geometry ( i.e. @xmath7 ) it is easy to determine the boundary intersection points @xmath167 of the straight line starting at @xmath50 with vector of direction @xmath168 .
for @xmath8 the situation is different .
this is why we construct a so - called _ geodesic projection _ , a mapping which determines the intersection point with the boundary and the corresponding tangential vector of a geodesic curve that passes through @xmath169 .
+ [ def:2_geodprojektion ] adopt the assumptions of lemma [ lem:2_phasenfunktion ] .
the mapping @xmath170 , given by @xmath171 where @xmath172 is uniquely determined by @xmath173 is called _ geodesic projection _ on the boundary @xmath174 .
+ the construction of @xmath175 is illustrated in figure [ fig:2_projektion ] .
the geodesic projection links the 2d parallel geometry in ct to the more general case of an inhomogeneous medium with given metric @xmath57 ; a fact which proves very useful for defining the backprojection operator and the implementation of our numerical solution approach for ( [ ip - tof ] ) .
please note that we need the assumptions of lemma [ lem:2_phasenfunktion ] and the existence of a phase function that @xmath176 is well - defined . + . at first
the green point @xmath177 is mapped by @xmath144 ( blue arrow ) on @xmath178 ( blue point ) and subsequently it is mapped to the boundary @xmath125 ( red point ) . ]
if @xmath179 generates a metric @xmath129 which is dissipative , then the linearized operator @xmath134 is continuous .
+ [ sat:2_stetlinvo ] let @xmath126 such that the corresponding metric @xmath180 turns @xmath181 into a cdrm .
then , the linearized ray transform @xmath182 is continuous .
+ under the assumption that the functions @xmath183 generating dissipative metrics @xmath129 are dense in @xmath184 , then we even can prove the continuity of the nonlinear operator @xmath23 .
+ suppose that the subset @xmath185 is dense in @xmath184 .
then @xmath186 is continuous .
+ let @xmath187 be arbitrary and @xmath188 be a sequence in @xmath189 such that @xmath190 as @xmath191 .
then we can estimate & & r(a ) - r(a_k)_l^2(_+ m ) + & = & r(a ) - r_a_k(a ) + r_a_k(a ) - r(a_k ) _
l^2(_+ m ) + & & r(a ) - r_a_k(a ) _
l^2(_+ m ) + r_a_k(a - a_k ) _ l^2(_+ m ) . since @xmath192
we have that @xmath193 is linear and continuous due to theorem [ sat:2_stetlinvo ] and thusly @xmath194 to tackle the first term we assume @xmath195 , @xmath196 to be respective solutions of the geodesic equation ( [ gl:2_geoddgl ] ) parametrized on @xmath197 $ ] .
we obtain & & r(a ) - r_a_k(a ) _
l^2(_+ m ) + & = & _ ^a_x , a(z ) l(z ) - _ ^a_k_x , a(z ) l(z ) _
l^2(_+ m ) + & = & _ 0 ^ 1 a(^a_x,(t))|^a_x,(t ) | t - _ 0 ^ 1 a(^a_k_x,(t ) ) | ^a_k_x,(t ) | t _ l^2(_+ m ) + & & _ y m |a(y)| _ 0 ^ 1 |^a_x,(t ) - ^a_k_x,(t ) | t _ l^2(m ) .
let @xmath198 be arbitrary . for @xmath199
sufficiently large we deduce from lemma [ kor:2_stetigegeodaeten ] and lemma [ l - tau - smooth ] @xmath200 } \|\dot \gamma^a_{x,\xi}(t ) - \dot \gamma^{a_k}_{x,\xi}(t ) \|
< \frac{\varepsilon}{2\sup_{y\in m } |a(y)| } .\ ] ] a sufficiently large @xmath199 furthermore assures that @xmath201 putting all this together we conclude r(a ) - r(a_k)_l^2(_+ m ) & & r(a ) - r_a_k(a ) _
l^2(_+ m ) + r_a_k(a - a_k ) _ l^2(_+ m ) + & & /2 + / 2 = if only @xmath199 is sufficiently large .
this proves the theorem .
+ unfortunately by now there is no proof that @xmath189 is dense in @xmath184 or not .
+ for completeness we give two existence and uniqueness results which can be found in @xcite .
+ let @xmath126 such that it induces a simple riemannian metric .
* if the measured data @xmath20 is generated by a refractive index @xmath82 , then the operator equation @xmath202 has a unique solution @xmath203 . *
if the measured data @xmath20 is generated by a refractive index @xmath204 , then the nonlinear operator equation @xmath205 has a unique solution @xmath206 .
+ our numerical solution scheme for ( [ ip - tof ] ) is iterative and demands for evaluating the forward operator , i.e. the computation of the tof measurements for a given metric , in each iteration step .
we do this using the fact that the tof function @xmath207 ( [ tof - mapping ] ) satisfies a partial differential equation where the differential operator is the so called _ geodesic vector field_. [ def:3_geodaetischerfluss ] let @xmath18 be a riemannian manifold and @xmath208 the tof mapping ( [ tof - mapping ] ) .
we call @xmath209 defined by @xmath210 _ geodesic vector field_. + there s a fundamental connection between the geodesic vector field and the refractive index regarding our measure geometry .
[ sat:3_transportgleichung ] let @xmath18 be a riemannian manifold with the metric tensor given as @xmath211 .
then for all @xmath212 the transport equation @xmath213 holds true
. a proof can be found in ( * ? ? ?
* section 1.2 ) .
[ bem:3_fluss ] the differential operator @xmath214 has a specific geometrical meaning . let @xmath215 be the solution of @xmath216 then we have @xmath217 for a geodesic @xmath70 .
this is the _ geodesic flow _ associated with the vector field @xmath218 .
+ [ kor:3_transportgleichung ] let @xmath135 be positive and the metric @xmath97 on @xmath42 defined by @xmath219 with metric tensor @xmath220 then hu(x , ) & = & ^i ( x , ) + ^-1(x ) ( ( x ) ^2 - 2 ^i , ( x ) ) ( x , ) + & = & ( x ) [ gl:3_transportgleichung ] holds true for all @xmath212
. + let @xmath212 . for @xmath221
the christoffel symbols are given by ( ^1_ij(x ) ) _ i , j=1 ^ 2 = ^-1(x ) & + & - .
for the partial derivatives we have ^1_jk(x)^j^k & = & ^-1(x ) _ 1 & _ 2 & + & - _ 1 + _ 2 + & = & ^-1(x ) ( _ 1 ^ 2 + 2 _ 1 _ 2 - _ 2 ^ 2 ) + & = & ^-1(x ) ( 2_1 ^ 2 + 2 _ 1 _ 2 - ^2 ) + & = & ^-1(x ) ( 2 _ 1 - ^2 ) + & = & ^-1(x ) ( 2 _ 1 , ( x ) - ^2 ) . in the same way we compute @xmath222 and
the proposition follows immediately .
+ in theorem [ sat:3_transportgleichung ] we transformed the inverse problem ( [ ip - tof ] ) into a parameter identification problem for the transport equation with boundary conditions .
again we recognize that the problem is highly nonlinear , because the refractive index @xmath6 appears as source - term as well as parameter in the differential operator . + according to the transport equation we want to develop a method to compute the geodesic curves for a given refractive index . to this end
we use the method of characteristics .
let @xmath223 be the graph of @xmath207 and @xmath224 a curve in @xmath225 with @xmath86 , @xmath226 a compact interval .
differentiating with respect to @xmath227 yields @xmath228 then @xmath229 is a normal of the curve at @xmath227 since & & _ x u ( x(t),(t ) ) + _ u ( x(t),(t ) ) + -1 ( x ( t ) , ( t ) , x^i(t ) + ^i(t ) ) + & = & _ x u ( x(t),(t ) ) x(t ) + _ u(x(t),(t ) ) ( t ) - x^i(t ) + & & - ^i(t ) + & = & 0 .
from @xmath230 we get that @xmath231 is a tangent vector , because & & _ x u ( x(t),(t ) ) + _ u ( x(t),(t ) ) + -1 ( ( t ) , ( -^i_jk(x(t ) ) ^j(t ) ^k(t))_i , ( x(t ) ) ) + & = & _ x u(x(t),(t ) ) - ^i_jk(x(t))^i(t)^j(t ) - ( x(t ) ) + & = & hu(x(t),(t ) ) - ( x(t ) ) + & = & 0 .
if we choose a boundary point @xmath232 and a direction @xmath233 , we finally obtain the initial value problem x_i(t ) = _ i(t ) & t i , i=1,2 + _ i(t ) = -^i_jk(x(t ) ) ^j(t ) ^k(t ) & t i , i=1,2 + z(t ) = ( x(t ) ) & t i + & + x(0 ) = x_0 & + ( 0 ) = _ 0 & + z(0 ) = 0 & . [ gl:3_awp ]
the solution is a characteristic , i.e. a curve on @xmath225 , which is the union of all characteristics .
but the characteristic curves of @xmath218 are just the geodesics of @xmath55 . solving ( [ gl:3_awp ] )
thus follows a geodesic curve starting at the boundary until it meets @xmath98 again say at @xmath234 .
then @xmath235 is the tof of the ultrasound signal propagating from @xmath236 in direction @xmath237 until it leaves the boundary @xmath98 at @xmath238 .
we intend to solve the nonlinear inverse problem ( [ ip - tof ] ) iteratively by a steepest descent method for a tikhonov functional where we use the linearized operator @xmath134 . in a first subsection
we define this functional , the iteration scheme and some of its properties . the second subsection then is devoted to deal with implementation issues of this method . + again we assume @xmath239 to be compact . for @xmath240 and @xmath89
we consider the tikhonov functional @xmath241 defined by [ def:4_tikhonovfunktional ] j_(f ) = r(f ) - u^meas _
y^2 + f - 1 _ l^p(m)^p .
here we used the notation @xmath242 .
the penalty term @xmath243 thereby ensures that the solution gives a refractive index that does not vary too strongly from @xmath7 , i.e. a homogeneous medium .
this is a supposed a priori information about the exact solution which we want to incorporate .
furthermore this term of course yields stability of the solution process .
the choice @xmath244 furthermore allows also for sparse solutions . because of the nonlinearity we can not expect that there exists a unique minimizer of @xmath245 .
usually a stationary point @xmath246 of @xmath247 is searched by the condition @xmath248 . there
we face a further problem : the computation of a gteaux derivative @xmath249 of @xmath23 .
the computation is still object of current research .
this is why we linearize the problem by using @xmath134 for fixed @xmath250 instead of the nonlinear operator @xmath23 . in this way
a second tikhonov functional @xmath251 comes up which is defined by @xmath252 with @xmath179 fixed . since @xmath134 is linear , it is simple to compute the subdifferential @xmath253 .
the last ingredient we need for doing so is the adjoint @xmath254 of @xmath134 .
+ [ sat:4_adjoperator ] let @xmath255 be fixed
. then the adjoint operator @xmath256 is given by [ ra - adjoint ] r_a^*(y ) = _ s^1 ( ( , ( y , ) ) ) | ( , ( y , ) ) | , ym , where @xmath257 \to \partial_+ \omega m$ ] is the geodesic projection from definition [ def:2_geodprojektion ] , @xmath258 denotes the jacobian of @xmath259 evaluated at @xmath260 , @xmath261 and @xmath262 satisfies @xmath263 for @xmath264 and @xmath265 for @xmath244 .
+ let @xmath266 and @xmath267 .
then we have r_a f(x , ) , ( x , ) _ y & = & _ _ + m r_a f(x , ) ( x , ) x + & & = _ _ + m _ ^a_x , f ( z ) ( x , ) l(z ) x + & & = _
s^1 _ -r^r _ m f ( y ) ( s-(y , ) ) ( ( , s ) ) y | ( , s ) |s d + & & = _
s^1 _ -r^r ( ( , s ) ) ( s-(y , ) ) | ( , s ) |s y + & & = _ m f ( y ) _
s^1 ( ( , ( y , ) ) ) | ( , ( y , ) ) | dy + & & = f , r_a^ * _ l^p l^p^*. + note that @xmath254 is an analogon to the classical backprojection operator , since it integrates along all geodesics passing a given point @xmath169 .
the numerical approximation of the geodesic projection @xmath268 will prove the part of our method which is most time consuming .
+ consider again the 2d parallel geometry in the unit disk @xmath269 , equipped with the euclidean geometry @xmath47 .
then the geodesics are straight lines and the geodesic projection @xmath270 is given by @xmath271 where @xmath163 , @xmath272 . in that sense
we can equivalently formulate @xmath176 as mapping @xmath273 by @xmath274 and obtain @xmath275 since in @xmath159 the phase function is given as @xmath276 , we obtain the usual backprojection operator in @xmath159 @xmath277 note that in the last identity we substituted @xmath166 by @xmath11 to stick on the conventional notation .
the weight @xmath278 comes from the fact that the lines @xmath279 here are parametrized by a boundary point @xmath280 and a unit tangential vector @xmath74 in @xmath28 with @xmath281 instead of the conventional parametrization by a normal vector @xmath11 and offset @xmath136 . + if we would just minimize @xmath282 for fixed @xmath183 we would completely neglect the influence of the integrand @xmath88 to the integration curves .
thus the idea of our algorithm is to use a steepest descent method which uses the actual iterate @xmath283 to compute the new one @xmath29 by minimizing @xmath284 .
this of course means to compute a new set of geodesic curves in each iteration step .
this leads to the following , adaptive iteration scheme .
+ [ alg:4_iteradapmin ] _ input : _ initial value @xmath285 , sequence of regularization parameters @xmath286 . * compute for @xmath285 the set of geodesics @xmath287 and set @xmath288 . * repeat for @xmath289 * * minimize @xmath290 , i.e. compute @xmath291 * * compute the new set of geodesics @xmath292 and evaluate @xmath293 .
_ output : _
@xmath294 for a stopping index @xmath295 , set of geodesics @xmath296 . + the minimization step ( s1a ) is done iteratively using the landweber method , compare e.g. @xcite .
let @xmath29 be the actual iterate in step ( s1 ) , then this iteration reads as _ k^0 & = & _ k + _ k^l+1 & = & _ k^l - _ l ( a_k^ * ( a_k_k^l - u^meas ) + _ k , l^ * ) , l=0,1,2 , + _ k+1 & = & _ k^l^ * with an element @xmath297 , @xmath298 , regularization parameter @xmath299 , step size @xmath300 , @xmath301 and reasonable stopping index @xmath302 . for @xmath303
we have @xmath304 for @xmath305 .
for @xmath244 we use the standard soft threshold algorithm from @xcite .
this yields the following version of step ( s1 ) from algorithm [ alg:4_iteradapmin ] as it is implemented .
+ [ alg:4_abstieglintikh ] _ input : _ current iterate @xmath29 , regularization parameter @xmath299 . *
define the operator @xmath306 and its adjoint @xmath307 by ( [ ra - adjoint ] ) .
* set @xmath308 and iterate for @xmath309 * * compute the residuum @xmath310 and stop if @xmath311 .
* * evaluate the adjoint operator @xmath312 .
* * determine a new search direction by @xmath313 in case @xmath303 or use soft thresholding for @xmath244 .
* * choose a decent step size @xmath314 and update @xmath315 * set @xmath316 .
_ output : _ new iterate @xmath317 .
+ [ bem:4_hilfsproblem ] * the evaluation of the linearized forward operator @xmath318 in step ( s1a ) is computed quickly and depends on the discretization of the geodesics and the evaluation of @xmath29 .
* the computation of the adjoint in step ( s1b ) however is very time consuming .
the difficulty lies in the interpolation of the geodesics .
the challenge is to find the geodesics that pass through a point @xmath177 , emitted from a direction @xmath319 at the boundary .
* the step size @xmath320 in step ( s1d ) can be determined by a line - search algorithm . * for convergence of the landweber method in algorithm [ alg:4_abstieglintikh ] to the minimum norm solution we refer to @xcite .
our aim is to prove that algorithms [ alg:4_iteradapmin ] , [ alg:4_abstieglintikh ] generate a sequence of iterates @xmath321 of decreasing values @xmath322 for the tikhonov functional @xmath247 ( [ def:4_tikhonovfunktional ] ) . to this end
we define the mapping @xmath323 by @xmath324 agorithm [ alg:4_iteradapmin ] reads then as [ gl:4_iterationt ] _ k+1 : = ( _ k ) = _ f l^p ( m ) j^_k_(f ) with given initial value @xmath325 . assuming that @xmath326 is a contraction
, then banach s fixed point theorem says that @xmath321 converges to a unique @xmath327 satisfying @xmath328 [ kor:4_minimaleigenschaftfixpunkt ] let @xmath329 be a contraction and @xmath327 be the unique fixed point of @xmath326 in @xmath330
. then @xmath331 and @xmath332 holds true for all @xmath266 and @xmath333 .
+ since @xmath334 we have @xmath335 for all @xmath266 .
note that in general we do not have [ gl:4_minimaleigenschaftgeodaeten ] j_(^ * ) j_^a(^ * ) ax_p but assuming that ( [ gl:4_minimaleigenschaftgeodaeten ] ) holds at least in a neighborhood of @xmath327 , then algorithm ( [ alg:4_iteradapmin ] ) in fact generates a decreasing sequence @xmath336 .
+ [ sat:4_lokalekonvergenz ] let @xmath337 be the sequence of iterates generated by algorithm [ alg:4_iteradapmin ] .
we assume that there exists a @xmath338 such that @xmath339 whenever @xmath340 .
then for @xmath341 the tikhonov functional @xmath342 is monotonically decreasing , we have @xmath343 for all @xmath340 . according to the assumptions for @xmath340 we have @xmath344 because @xmath317 is a minimizer of @xmath345 . + + summarizing we found a criterion for the iteration sequence @xmath321 to have a weak limit point .
+ adopt the assumptions of lemma [ sat:4_lokalekonvergenz ] .
then the sequence of iterates @xmath321 generated by algorithm [ alg:4_iteradapmin ] has a weakly convergent subsequence , if @xmath303 .
+ lemma [ sat:4_lokalekonvergenz ] shows that the sequence @xmath342 is monotonically decreasing and hence converging , since it is bounded from below .
thus @xmath346 is bounded . because @xmath347 the sequence @xmath321 is bounded , too .
the spaces @xmath330 are reflexive for @xmath303 which implies that @xmath321 has a weakly convergent subsequence .
+ in this subsection we address issues of the implementation of the algorithms [ alg:4_iteradapmin ] and [ alg:4_abstieglintikh ] . for simplicity we set @xmath348 , the closed unit disk in @xmath31 .
+ at first we describe the discretization of @xmath42 and @xmath6 .
let @xmath349 .
for a given discretization step size @xmath350 we define the equally spaced grid [ def:4_diskretisierung ] e^h : = \ { ( r_i , s_j ) ^2 : r_i = ih , s_j = jh , i , j } ^2 . and
@xmath351 by @xmath352 for @xmath353 .
on @xmath354 we define the grid values @xmath355 of @xmath6 by @xmath356 and bilinear functions @xmath357 by _
ij^(r , s ) & = & + & & f^h_ij ( ) + ( f^h_i+1,j ( ) - f^h_i , j ( ) ) + ( f^h_i , j+1 ( ) - f^h_i , j ( ) ) & + + ( f^h_i+1,j+1 ( ) - f^h_i+1,j ( ) - f^h_i , j+1 ( ) + f^h_i , j ( ) ) & ( r , s ) e_ij^h + 0 & . the unique bilinear interpolate @xmath358 of @xmath6 is then given by @xmath359 the set of all piecewise bilinear interpolates is denoted by @xmath360 the following properties of @xmath6 and its interpolate @xmath361 are obvious : + * @xmath362 for all @xmath363 + * @xmath361 ist continuous on every element @xmath364 and differentiable in the interior of @xmath364 + for numerical purposes we cover the unit disk @xmath42 by a subset of the grid @xmath365 \times [ -1,1]$ ] and set @xmath366 for an integer @xmath367 .
for the number of grid points we have @xmath368 and @xmath369 such that number of degrees of freedom increases like @xmath370 . note that in view of ( [ gl:2_geoddgl ] ) it is important that @xmath361 is differentiable on @xmath42 .
this is achieved by extending the gradient of @xmath361 continuously to the edges of @xmath364 .
+ the next step is the discretization of the measure data @xmath20 . in practical applications we of course have only a finite number of tof measurements .
that means the ray transform @xmath371 is given for a finite number of pairs @xmath372 .
this is why we define finite sets @xmath373 of source points @xmath374 and ray directions @xmath375 for @xmath376 .
hence any geodesic curve @xmath377 for which we acquire measure data is characterized by an element @xmath378 of the set @xmath379 where @xmath380 . for the implementation of algorithm [ alg:4_iteradapmin ] we use
then the discrete tikhonov functional @xmath381 @xmath382 the according set of geodesic curves is given by @xmath383 and the initial value problem is solved by a runge - kutta method with stepsize control .
+ the next important ingredient is the implementation of the backprojection operator @xmath254 for given @xmath250 .
we recall that @xmath384 our implementation is similar to that of the backprojection operator @xmath385 in standard 2d computerized tomography , but there are two crucial issues to address : + * the determinant @xmath386 can not be computed , since the geodesic projection @xmath176 as well as the phase function @xmath144 are not explicitly known .
two possible substitutions are to set the determinant equal to @xmath387 or to use the expression for the euclidean geometry @xmath388 what we have done in our computations .
this perfectly fits to our a priori assumption that @xmath6 varies only slightly from @xmath387 .
+ * in standard ct for a given point @xmath169 it is easy to compute the corresponding boundary point @xmath389 such that a line , outgoing from @xmath28 in direction of @xmath11 meets @xmath76 .
for @xmath8 the determination of @xmath390 such that the geodesic @xmath70 meets a given @xmath169 is a difficult task .
+ we developed the following algorithm to compute the backprojection operator @xmath391 for given @xmath199-th iterate @xmath29 in algoritm [ alg:4_iteradapmin ] as it was used in our implementation .
the unit vectors @xmath168 are discretized by @xmath392 for given @xmath393 .
+ [ alg:4_adjoperator ] _ input : _ set of geodesic curves @xmath394 associated with @xmath29 and the corresponding tof measuremensts @xmath395 with @xmath396 , @xmath397 , @xmath398 , reconstruction point @xmath169 . *
set @xmath399 .
* iterate for @xmath400 * * compute @xmath401 and @xmath402 .
* * project the point @xmath76 in direction @xmath401 on the boundary point @xmath403 and determine @xmath404 , which is the nearest neighbor to @xmath405 .
* * determine the unique @xmath406 such that @xmath407 is minimal . * * if @xmath408 , set @xmath409 and go to step ( s1 ) . * * if @xmath410 , then determine @xmath411 such that @xmath412 for ( unique ) @xmath413 . * * if @xmath414 , then * * * check if @xmath415 . if yes then set @xmath416 and go to step ( s1f ) . * * * otherwise interpolate @xmath417 linearly between @xmath418 and @xmath419 and go to step ( s1 ) . * * if @xmath420 , then * * * check if @xmath421 .
if yes then set @xmath422 and go to step ( s1 g ) . * * * otherwise interpolate @xmath417 linearly between @xmath418 and @xmath423 and go to step ( s1 ) .
* approximate the backprojection by @xmath424 _ output : _ backprojection @xmath425 .
+ the algorithm works as follows : at first we project @xmath76 to the boundary in direction @xmath401 in the euclidean sense .
this gives @xmath426 ( s1b ) .
next we choose from evey geodesic @xmath377 starting in @xmath427 that one whose trace has minimal distance from @xmath76 .
the corresponding index of the tangent is denoted by @xmath428 ( s1c ) .
the next step consists of computing the intersection point of @xmath429 with the line @xmath430 yielding @xmath431 .
if the geodesic runs to the left of @xmath76 ( in the sense of @xmath432 ) , i.e. @xmath414 , we increment @xmath433 and repeat the procedure ( s1fi ) .
if the geodesics @xmath429 , @xmath434 are located on different sides of @xmath76 ( in the sense of @xmath432 ) we use linear interpolation to compute @xmath417 ( s1fii ) .
if @xmath435 we proceed in the same way ( s1 g ) . finally the backprojection @xmath436 is computed using the trapezoidal sum .
please note that both concepts , the usage of the trapezoidal sum with respect to the projections @xmath401 as well as linear interpolation , are adopted from the standard backprojection step for fbp algorithm in 2d computerized tomography , see e.g. @xcite .
algorithm [ alg:4_adjoperator ] is emphasized in figure [ fig:4_adjoperator ] . + * if a point @xmath169 is located close to the boundary @xmath98 , then it is possible that there are no neighboring geodesics @xmath437 . in this case
we interpolate to zero .
* it is importand to fix what runs left or right of @xmath76 means .
our choice is with respect to the line @xmath438 but is not necessarily the best choice . * when incrementing or decrementing the index @xmath433 we have to compute modulo @xmath439 . *
the interpolation of @xmath417 can be done in several ways .
here we interpolate linearly with respect to the distance of @xmath76 to the intersection points @xmath440 and @xmath441 . .
the red dotted line is the geodesic curve , which runs exactly throug @xmath76 bit is not in the measured data @xmath442 , @xmath443 . ]
in this section we demonstrate the performance of our method on the basis of several test examples using synthetic data .
we implemented algorithms [ alg:4_iteradapmin ] , [ alg:4_abstieglintikh ] and [ alg:4_adjoperator ] as shown in section 3 . to solve the forward operator , i.e. to compute the ray transform @xmath444 for given @xmath183 and @xmath6 we applied an extrapolation step control method based on the classical runge - kutta method of 4th order .
thereby the tolerance parameter was chosen as @xmath445 , the initial stepsize as @xmath446 and the stopping index was @xmath447 .
since the geodesic curves @xmath377 can be computed independently from each other , the computation of @xmath444 as well as of the synthetic tof data @xmath448 can be parallelized what we have done using 30 cores .
the objective functional @xmath449 is minimized subject of @xmath450 in each iteration step using the steepest descent method given in algorithm [ alg:4_abstieglintikh ] . in the first example
we investigate the behavior of the reconstruction results with respect to a varying regularization parameter @xmath89 and for fixed norm @xmath451 . in every step we chose @xmath452 as a constant .
the exact refractive index @xmath453 is given by the sound speed `` peaks '' as [ sound - peaks ] c(x ) = 1/(x ) = 1 + _ i=1 ^ 3_i(x)_i(x ) with @xmath454 and @xmath455 for given center points @xmath456 , radii @xmath457 and amplitudes @xmath458 , @xmath459 . in our experiments we set q_1 = + , q_2 = - + - , q_3 = + - for the center points , r_1 = , r_2 = , r_3 = for the radii and _ 1 = , _ 2 = - , _ 3 = for the amplitudes . in each iteration step
we chose @xmath460 ( number of detectors ) and @xmath461 ( number of signals per detector ) , such that in every step we have to compute @xmath462 geodesics .
the computation of a full set of geodesics @xmath463 lasts about 30 seconds .
the evaluation of the adjoint @xmath464 lasts about one minute .
we compared results for regularization parameters @xmath465 , @xmath466 . as iteration step size in algorithm [ alg:4_abstieglintikh ]
we used @xmath467 .
the unit square was discretized using the step size @xmath468 yielding @xmath469 reconstruction points .
+ figure [ fig:4_ex1zf ] shows the values @xmath470 for different regularization parameters @xmath471 .
one can see that all functional values at first increase and then decrease ( except at the first one ) to a minimum which is less than the initial value . on the basis of these curves
one is able to determine _ optimal stopping indices _ @xmath472 , @xmath473 , where @xmath474 figure [ fig:4_ex1sgv ] shows @xmath475 for @xmath473 along with the exact solution @xmath476 .
the errors @xmath477 are plotted in figure [ fig:4_ex1sgvdiff ] .
one realizes that for smaller values of @xmath478 there are higher fluctuations in the reconstructions . for all reconstructions the peaks
are well detected , but their quantity is smaller if @xmath478 increases .
for example in the peak @xmath479 we have @xmath480 and the reconstruction with @xmath481 gives @xmath482 , but for @xmath483 we have @xmath484 ( compare figure [ fig:4_ex1sgs ] )
. a bigger @xmath478 leads to a higher weighting of the regularization term which penalizes aberrations from @xmath387 . the weak amplitude at @xmath485 is only detected if @xmath478 is big enough .
otherwise the peak can not be recognized because of the high fluctuations . for different regularization parameters @xmath471 plotted against iteration index @xmath199 .
top left : @xmath486 , top right : @xmath487 , middle left : @xmath488 , middle right : @xmath489 , bottom : @xmath490.,title="fig : " ] for different regularization parameters @xmath471 plotted against iteration index @xmath199 .
top left : @xmath486 , top right : @xmath487 , middle left : @xmath488 , middle right : @xmath489 , bottom : @xmath490.,title="fig : " ] + for different regularization parameters @xmath471 plotted against iteration index @xmath199 .
top left : @xmath486 , top right : @xmath487 , middle left : @xmath488 , middle right : @xmath489 , bottom : @xmath490.,title="fig : " ] for different regularization parameters @xmath471 plotted against iteration index @xmath199 .
top left : @xmath486 , top right : @xmath487 , middle left : @xmath488 , middle right : @xmath489 , bottom : @xmath490.,title="fig : " ] + for different regularization parameters @xmath471 plotted against iteration index @xmath199 .
top left : @xmath486 , top right : @xmath487 , middle left : @xmath488 , middle right : @xmath489 , bottom : @xmath490.,title="fig : " ] and reconstructions @xmath491 for different regularization parameters @xmath471 .
top left : @xmath486 , top right : @xmath487 , middle left : @xmath492 , middle right : @xmath489 , bottom left : @xmath490 , bottom right : original sound speed.,title="fig : " ] and reconstructions @xmath491 for different regularization parameters @xmath471 .
top left : @xmath486 , top right : @xmath487 , middle left : @xmath492 , middle right : @xmath489 , bottom left : @xmath490 , bottom right : original sound speed.,title="fig : " ] + and reconstructions @xmath491 for different regularization parameters @xmath471 .
top left : @xmath486 , top right : @xmath487 , middle left : @xmath492 , middle right : @xmath489 , bottom left : @xmath490 , bottom right : original sound speed.,title="fig : " ] and reconstructions @xmath491 for different regularization parameters @xmath471 .
top left : @xmath486 , top right : @xmath487 , middle left : @xmath492 , middle right : @xmath489 , bottom left : @xmath490 , bottom right : original sound speed.,title="fig : " ] + and reconstructions @xmath491 for different regularization parameters @xmath471 .
top left : @xmath486 , top right : @xmath487 , middle left : @xmath492 , middle right : @xmath489 , bottom left : @xmath490 , bottom right : original sound speed.,title="fig : " ] and reconstructions @xmath491 for different regularization parameters @xmath471 .
top left : @xmath486 , top right : @xmath487 , middle left : @xmath492 , middle right : @xmath489 , bottom left : @xmath490 , bottom right : original sound speed.,title="fig : " ] for different regularization parameters @xmath471 . _
top left : _
@xmath486 , _ top right : _
@xmath487 , _ middle left : _
@xmath488 , _ middle right : _
@xmath489 , _ bottom : _ @xmath490.,title="fig : " ] for different regularization parameters @xmath471 .
_ top left : _
@xmath486 , _ top right : _
@xmath487 , _ middle left : _
@xmath488 , _ middle right : _
@xmath489 , _ bottom : _ @xmath490.,title="fig : " ] + for different regularization parameters @xmath471 . _
top left : _
@xmath486 , _ top right : _
@xmath487 , _ middle left : _
@xmath488 , _ middle right : _
@xmath489 , _ bottom : _ @xmath490.,title="fig : " ] for different regularization parameters @xmath471 .
_ top left : _
@xmath486 , _ top right : _
@xmath487 , _ middle left : _
@xmath488 , _ middle right : _ @xmath489 , _ bottom : _ @xmath490.,title="fig : " ] + for different regularization parameters @xmath471 .
_ top left : _
@xmath486 , _ top right : _
@xmath487 , _ middle left : _
@xmath488 , _ middle right : _
@xmath489 , _ bottom : _ @xmath490.,title="fig : " ] and reconstructions @xmath491 for different regularization parameters @xmath471 .
_ left : _ original , _ middle : _
@xmath486 _ right : _ @xmath490.,title="fig : " ] and reconstructions @xmath491 for different regularization parameters @xmath471 .
_ left : _ original , _ middle : _
@xmath486 _ right : _ @xmath490.,title="fig : " ] and reconstructions @xmath491 for different regularization parameters @xmath471 . _
left : _ original , _ middle : _
@xmath486 _ right : _ @xmath490.,title="fig : " ] for different @xmath493-norms .
_ top left : _
@xmath494 , _ top right : _
@xmath495 , _ middle left : _
@xmath496 , _ middle right : _
@xmath497 , _ bottom left : _
@xmath498 , _ bottom right : _ exact sound speed.,title="fig : " ] for different @xmath493-norms .
_ top left : _
@xmath494 , _ top right : _
@xmath495 , _ middle left : _
@xmath496 , _ middle right : _
@xmath497 , _ bottom left : _
@xmath498 , _ bottom right : _ exact sound speed.,title="fig : " ] + for different @xmath493-norms .
_ top left : _
@xmath494 , _ top right : _
@xmath495 , _ middle left : _
@xmath496 , _ middle right : _
@xmath497 , _ bottom left : _
@xmath498 , _ bottom right : _ exact sound speed.,title="fig : " ] for different @xmath493-norms .
_ top left : _
@xmath494 , _ top right : _
@xmath495 , _ middle left : _
@xmath496 , _ middle right : _
@xmath497 , _ bottom left : _
@xmath498 , _ bottom right : _ exact sound speed.,title="fig : " ] + for different @xmath493-norms .
_ top left : _
@xmath494 , _ top right : _
@xmath495 , _ middle left : _
@xmath496 , _ middle right : _
@xmath497 , _ bottom left : _
@xmath498 , _ bottom right : _ exact sound speed.,title="fig : " ] for different @xmath493-norms .
_ top left : _
@xmath494 , _ top right : _
@xmath495 , _ middle left : _
@xmath496 , _ middle right : _
@xmath497 , _ bottom left : _
@xmath498 , _ bottom right : _ exact sound speed.,title="fig : " ] .
top left : @xmath494 , top right : @xmath495 , middle left : @xmath496 , middle right : @xmath497 , bottom : @xmath498.,title="fig : " ] .
top left : @xmath494 , top right : @xmath495 , middle left : @xmath496 , middle right : @xmath497 , bottom : @xmath498.,title="fig : " ] + .
top left : @xmath494 , top right : @xmath495 , middle left : @xmath496 , middle right : @xmath497 , bottom : @xmath498.,title="fig : " ] .
top left : @xmath494 , top right : @xmath495 , middle left : @xmath496 , middle right : @xmath497 , bottom : @xmath498.,title="fig : " ] + .
top left : @xmath494 , top right : @xmath495 , middle left : @xmath496 , middle right : @xmath497 , bottom : @xmath498.,title="fig : " ] again we consider the exact speed of sound ( [ sound - peaks ] ) but now we calculate reconstructions with different @xmath499-norms .
more explicitly we set @xmath500 in case @xmath494 we implemented the soft threshold method as presented in @xcite with @xmath501 in this case .
for all other @xmath493 we set @xmath502 . in this series of reconstructions we chose @xmath503 and @xmath504 yielding @xmath505 geodesics to be computed in each iteration setp .
the computation of a full set of geodesics lasts about @xmath506 seconds , the evaluation of @xmath391 about 20 seconds . in every iteration step
we make one descent step with step size parameter @xmath507 .
the uinit square was discretized again using @xmath468 .
the optimal stopping indices @xmath508 , @xmath473 were @xmath509 the reconstructions are visualized in figure [ fig:4_ex2peaksvogel ] compared with the exact @xmath510 . as expected the @xmath511 and @xmath512 norms lead to the best reconstructions because @xmath3 and thus @xmath22 is sparse .
particularily the most part of the reconstruction is identical to @xmath387 . for the choices @xmath513 , @xmath514 and @xmath515 one discovers increasing smoothness but also fluctuations of the solution .
it is interesting that the small peak at @xmath516 seems to cause severe artifacts at the boundary @xmath517 for @xmath515 .
this is also emphasized in figure [ fig:4_ex2peaksdraufdiff ] where the errors @xmath477 are plotted .
indeed all peaks are detected correctly ( the maximal error is about @xmath518 ) , but for higher norms the fluctuations in the euclidean areas , i.e. areas where @xmath7 , increase significantly .
particularily the error is as large as the detected peaks .
the artifacts close to the boundary @xmath519 become also obvious when comparing the traces of the geodesics obtained for @xmath511 and @xmath515 to those of the exact solution @xmath520 , see figure [ fig:4_ex2peaksgeodaeten ] .
the left picture shows a good match of the reconstructed and approximated set of geodesics @xmath521 , whereas one clearly recognizes big aberrations in the picture to the right .
+ , the red curves are geodesics of the reconstructions @xmath294 .
left picture : @xmath494 , @xmath522 , right picture : @xmath498 , @xmath523.,title="fig : " ] , the red curves are geodesics of the reconstructions @xmath294 .
left picture : @xmath494 , @xmath522 , right picture : @xmath498 , @xmath523.,title="fig : " ] next we consider a sound speed which is not sparse .
let @xmath524 with @xmath525 and the parameters @xmath526 , @xmath527 .
the riemannian manifold @xmath528 has then constant , positive gaussian curvature @xmath529 .
reconstructions for @xmath499-norms with @xmath511 , @xmath530 can be seen in figure [ fig : constant_curvature ] .
the regularization parameters were @xmath531 and @xmath532 , respectively .
the stopping indices were @xmath533 and @xmath534 , respectively .
as expected the reconstruction for @xmath535 is more accurate then the sparse reconstruction for @xmath536 .
we furthermore realize that at the center the reconstruction deteriorates .
this comes from the specific metric @xmath537 which generates a _ cold spot _ in the center , that means a small region where almost no geodesic curve , i.e. ultrasound signal , intersects .
this fact is clearly visible when we consider the associated geodesic curves for @xmath6 and @xmath538 ( figure [ fig : constant_curvature_geodesics ] ) .
one sees that only few geodesics pass the center of the disk .
+ , _ middle : _ @xmath530 , _ right : _ exact sound speed.,title="fig : " ] , _ middle : _
@xmath530 , _ right : _ exact sound speed.,title="fig : " ] , _ middle : _
@xmath530 , _ right : _ exact sound speed.,title="fig : " ] ( red curves ) and @xmath6 ( black curves ) .
the _ cold spot _ in the center is clearly visible . ] at last we show a numerical test using noise contaminated data @xmath539 .
we perform this test by means of the exact solution @xmath540 with @xmath541 , @xmath542 as in subsection [ c - peaks ] and @xmath543 @xmath544 and the parameters @xmath545 , @xmath546 .
the measure data additionally have been contaminated by unifromly distributed noise @xmath547 , @xmath548 with relative error @xmath549 ( i.e. @xmath550 relative noise ) .
figure [ fig : noise ] shows reconstructions with exact @xmath551 as well as with noisy data @xmath539 .
the parameters for the reconstruction with noisy data are @xmath552 , @xmath553 , @xmath451 and @xmath554 .
+ ( left picture ) from exact data @xmath20 ( middle picture ) and noisy data @xmath539 ( right picture).,title="fig : " ] ( left picture ) from exact data @xmath20 ( middle picture ) and noisy data @xmath539 ( right picture).,title="fig : " ] ( left picture ) from exact data @xmath20 ( middle picture ) and noisy data @xmath539 ( right picture).,title="fig : " ]
in the article we propose a numerical solution scheme for the computation of the refrecative index of a medium from boundary time - of - flight measurements in 2d .
the method relies on the minimization of a tikhonov functional that penalizes aberrations from @xmath7 .
the minimization is done by a steepest descent method where the linearization of the forward operator was achieved by using the old iteration @xmath29 as refractive index to compute the propagation paths .
the ultrasound signals were assumed to propagate along geodesics of the riemannian metric @xmath555 which is due to fermat s principle . we were able to prove that every sequence @xmath321 generated by our iterative scheme has weak limit points .
of course this is a little unsatisfactory
. it would be great if one could prove that these limit points are minimizers of @xmath247 .
one way to do this might be to use the concept of surrogate functionals , see e.g. @xcite .
in fact if we define @xmath556 by @xmath557 then obviously our algorithm [ alg:4_iteradapmin ] reads as @xmath558 if we assume that @xmath559 at least for all @xmath183 close to @xmath6 , then @xmath560 can be interpreted as a ( local ) surrogate functional .
the investigation of convcergence as well as the derivation of the gteaux derivative @xmath561 , @xmath562 is subject of current research . + the numerical experiments show a good performance of the method , also if we have sparse solutions .
+ another result of the article is the explicit representation of the backprojection operator @xmath254 for a non - euclidean geometry as well as its numerical realization .
we showed the analogy to the conventional ( euclidean ) backprojection operator as it is known from 2d computerized tomography .
+ at last we would like to mention that the results of this article do not only affect seismics or phase contrast tof tomography , but also other tomographic problems in inhomogeneous media such as vector and tensor field tomography .
we are indebted to the deutsche forschungsgemeinschaft ( german science foundation , dfg ) which funded this project under schu 1978/7 - 1 . | the article deals with a classical inverse problem : the computation of the refractive index of a medium from ultrasound time - of - flight ( tof ) measurements .
this problem is very popular in seismics but also for tomographic problems in inhomogeneous media .
for example ultrasound vector field tomography needs a priori knowledge of the sound speed .
according to fermat s principle ultrasound signals travel along geodesic curves of a riemannian metric which is associated with the refractive index .
the inverse problem thus consists of determining the index of refraction from integrals along geodesics curves associated with the integrand leading to a nonlinear problem . in this article
we describe a numerical solver for this problem scheme based on an iterative minimization method for an appropriate tikhonov functional .
the outcome of the method is a stable approximation of the sought index of refraction as well as a corresponding set of geodesic curves .
we prove some analytical convergence results for this method and demonstrate its performance by means of several numerical experiments .
another novelty in this article is the explicit representation of the backprojection operator for the ray transform in riemannian geometry and its numerical realization relying on a corresponding phase function that is determined by the metric .
this gives a natural extension of the conventional backprojection from 2d computerized tomography to inhomogeneous geometries .
refractive index , ray transform , riemannian metric , fermat s principle , geodesic curve , backprojection operator , tikhonov functional 45g10 , 53a35 , 58c35 , 65r20 , 65r32 |
with the impressive results obtained in the theory of classical chaos , the interest in the manifestations of chaos in quantum systems has grown over the past years . although there is not a clear difference between chaotic and regular dynamics for quantized hamiltonian dynamics , some insight has been gained with the quantization of conservative models with chaotic classical counterparts . among the peculiar phenomenology that has been found we have the appearance of _ scars _ , which are wavefunctions with enhanced amplitude along the paths of classical unstable periodic orbits @xcite .
further connection between classically unstable orbits and and quantum behavior has been given by gutzwiller s trace formula for the density of states@xcite , which has been applied with some success to many hyperbolic systems @xcite .
a connection has also been found between the nature of the classical dynamics , be it regular or chaotic , and the statistics of the energy spectrum , which obeys the general distributions found in random matrix theory @xcite .
however , for systems that include dissipation , not much work has been done .
this is an important subject of study because , besides its theoretical interest , it is of central importance in several interesting applications ( squids , tunneling phenomena , etc . )
the main problem with such systems is the presence of the dissipative term , which makes the quantization process not feasible by standard methods .
approaches to this problem have been proposed by many authors over the last three decades .
kostin proposes the use of a nonlinear schrdinger equation for the process of quantization @xcite , but it has some problems with the superposition principle and yields results of an unphysical nature .
dekker @xcite uses a canonical quantization method based on complex variables , with some degree of success , but the physical basis of his theory is not very clear .
maximal entropy methods @xcite have also been proposed , using a coupling between quantum and classical degrees of freedom .
this approach ignores the effects of thermally driven fluctuations . in general ,
the only fully successful and physically sound quantization scheme for dissipative systems is the so - called system plus reservoir approach , in which one considers the system coupled to a large reservoir or heath bath . since the entire system plus reservoir may be considered as closed , one can apply to it the standard rules of quantization , paying of course the penalty of dealing with an infinite number of degrees of freedom .
dissipation comes about by the net transfer of energy from the system to the reservoir , and thermal fluctuations appear as a result of the random energy exchange from / to the heat bath .
these approaches ensure the non - violation of commutator rules , and include the appropriate fluctuation dissipation relations .
some common approaches using the system - plus - reservoir picture are the influence - functional methods introduced by feynman and vernon @xcite , that have been used by caldeira and legget to study quantum brownian motion @xcite and dissipative tunneling effects @xcite ; generalized master equations @xcite , used to describe damping phenomena in quantum optics and spin dynamics @xcite ; and quantum langevin equations @xcite , used by ford _ et al . _ to study dissipative quantum tunneling at zero temperature @xcite . in the area of nonlinear ( and classically chaotic ) dissipative models
, there are some results that point to the conclusion that quantum effects act as sources of noise . as a result
, it is expected that the threshold for chaos be lowered _ in a continuous fashion _ upon the introduction of quantum corrections .
this was reported by graham and tel @xcite , who worked in the master equation formalism @xcite to quantize henon s map , and by cerdeira _
@xcite , who built a quantum mapping of wigner functions for a kicked oscillator .
more recently , cohen @xcite studied the quantum dissipative kicked rotator using influence functionals , while dittrich and graham used the master equation formalism for the same purpose @xcite .
liu and schieve @xcite use the quantum langevin equation approach to show that a chaotic attractor is observed in a quantum system whose classical counterpart is non - chaotic . in contradiction to these works , there were some results showing an increase of order ( a raising of the chaoticity thresholds ) for a kicked dissipative oscillator @xcite . in this work
we study the behavior of a periodically kicked harmonic oscillator , with dissipation , and in the presence of quantum and thermal fluctuations .
our interest is mostly to find how the bifurcation points and chaotic thresholds in parameter space change when increasingly large thermal and quantum effects are included .
to do this , we generate a mapping of expectation values , using a quantum langevin approach , that has an immediate analog with its classical limit .
this expectation values mapping ( evm ) constitutes in principle a complete description of the quantum dynamics , based in a discrete set of variables , and has the advantage of exhibiting a clear hierarchy in its terms , introducing from the beginning mean values and a sequence of moments of the fluctuations ( the completeness of the description given by the evm will be shown in appendix a ) . in section
ii we introduce the classical model , a harmonic oscillator with a periodic kick " from a nonlinear potential . in section iii
we will use the langevin equation formalism to obtain a quantal version of the classical map of the preceding section .
the resulting ( truncated ) evm will include both thermal and quantum corrections ; the change in the behavior of the dynamics as function of @xmath0 and @xmath1 is examined . in section iv and v
we show and discuss the results obtained and give some concluding remarks .
starting with the hamiltonian for a kicked oscillator @xmath2 where @xmath3 is the period of the kick , we introduce the dissipative force @xmath4 to obtain the equations of motion in the form @xmath5 from now on we will set the units of mass and time so that @xmath6 and @xmath7 . for times @xmath8 , the kick has no effect so we get the equation @xmath9 whose well known underdamped ( @xmath10 ) solution is given by @xmath11,\end{aligned}\ ] ] where @xmath12 , and @xmath13 and @xmath14 are given in terms of initial conditions . evaluating these constants , and defining an evolution matrix @xmath15 by @xmath16 we can write a discrete linear map corresponding to eqs .
( [ cont1 ] , [ cont2 ] ) @xmath17 where we have set as usual @xmath18 , @xmath19 , with similar definitions for @xmath20 .
we need now to calculate how the kick affects the dynamics , in order to complete the map .
this can be done by integrating eq .
( [ ec_a_integrar ] ) from just before to just after @xmath21
. over this region @xmath22 is continuous and thus we obtain @xmath23 successive application of the mappings ( [ cont1 ] , [ cont2 ] ) and ( [ en1 ] , [ en2 ] ) will give us a map that yields the values for @xmath22 and @xmath20 just after each kick ( dropping from now on the superscripts ) : @xmath24 we have chosen the force acting on the kicks as @xmath25 which gives a potential bounded from below . from now on we will set the units of length so that @xmath26 .
this force is maximum at @xmath27 , so that the closer the harmonic oscillator gets to its rest position , the stronger the repulsive effect of the kick becomes .
we are interested in a region of parameter space such that @xmath28 , and where clear chaotic behavior can be found . after some search ,
we have chosen @xmath29 and @xmath30 . for bifurcation diagrams given in terms of @xmath31
, we find a well defined transition to chaos around @xmath32 .
it should be noticed that there is some coexistence of attractors in this sector of parameter space ; these attractors evolve from order to chaos following feigenbaum scenario ( see fig . ( 1 ) ) . for the fully chaotic region ,
the attractors are hnon - like ( see fig . ( 2 ) ) .
the later map will be quantized in the next section in order to see the temperature and quantum effects on the dynamics .
we use the langevin equation formalism for the process of quantization .
this equation is obtained from the quantization in the heisenberg picture of a closed system given by a relevant degree of freedom @xmath33 , in contact with a large ( in the limit , infinite ) bath of harmonic oscillators @xmath34 .
it is given by @xmath35 where @xmath36 and @xmath37 are operators , @xmath38 is the potential and @xmath37 is a random force that , together with the memory function @xmath39 , depends on the characteristics of the bath .
these oscillators are assumed to be in thermal equilibrium at some initial time @xmath40 .
the memory function @xmath39 , and the expectation value and symmetric correlations of @xmath37 are given by @xcite @xmath41,\\ \gamma(t ) & = & \frac{1}{2}\sum_j m_j \omega_j^2 \cos(\omega_j t ) \theta(t).\end{aligned}\ ] ] where @xmath42 and @xmath43 are the mass and frequency of the @xmath44th oscillator .
these expressions imply a quantum fluctuation - dissipation theorem . it is assumed that no correlation exists between @xmath45 and @xmath33 or @xmath46 .
this implies ignoring some switch - on processes , which arise due to our assumption that the bath _ per se _ is in thermal equilibrium at time @xmath40 . from now on , we will use a bath of equal masses and a white ( _ i .
e. _ , @xmath47independent ) , and continuous frequency distribution , which is the case for constant friction or markovian approximation . in this case ,
@xmath48,\end{aligned}\ ] ] and eq . ( [ lang_eq ] )
reduces to the simpler form @xmath49 the drawback of using the markovian approximation is that it will generate some divergences since we now have contributions from arbitrarily high frequencies in the bath .
we will use the same harmonic oscillator - plus - kick potential indicated in the hamiltonian ( [ h_clas ] ) and eqn . ( [ potencial ] ) . between kicks , the ( underdamped ) solution of eq .
( [ q_eq ] ) is given by @xmath50 where the solution for the non homogeneous part has been obtained from a riccatti equation and we have defined @xmath51 in an analogous way as for the classical case .
the operators @xmath52 and @xmath53 are given in terms of initial conditions , evaluated at some initial time @xmath54 ( not to be confused with @xmath40 ) , and the inhomogeneous term is @xmath55 for the moment @xmath46 we get @xmath56\\ & & + \omega_0e^{-\gamma t}\left[-\mbox{\boldmath$a$}\sin(\omega_0t)+ \mbox{\boldmath$b$}\cos(\omega_0t)\right]+ \mbox{\boldmath$g$}(t,\omega_0,\gamma),\end{aligned}\ ] ] where @xmath57\mbox{\boldmath$f$}(t').\ ] ] now we will take the integration limits to be @xmath58 and @xmath59 , so we can evaluate the initial conditions and get @xmath60 with the same matrix elements @xmath61 defined before . defining the fluctuations as @xmath62 we can now write down the evolution equations for expectation values and fluctuations : @xmath63 since both @xmath64 and @xmath65 have zero expectation value .
hereafter we will use non - bold letters to denote expectation values . in order to find what happens over the kick
, we note that @xmath33 is continuous and @xmath46 has an increment , so @xmath66 since the potential term has an operator argument we will perform a second order taylor expansion .
this is the first truncation we need to carry out in order to close the system of equations . from this
we get @xmath67 the potential and its derivatives can now be evaluated numerically since we have no longer an operator as argument .
however in order to have an evm suitable for numerical work , as in eq .
( [ qmap ] ) , we need to take expectation values of all operators . in order to do this
we have to calculate , among others , the map for the fluctuation @xmath68 so that we can evaluate the above equation for @xmath69 .
we proceed by evaluating @xmath68 and made another truncation to second order in @xmath70 , so that we get a map which is also second order . in an analogous fashion
, we obtain the mappings for the fluctuations @xmath71 and @xmath72 and after some rearrangement of terms we have the 5 expectation value maps for @xmath73 and @xmath74 and for the fluctuations ( some details are given in appendix b ) @xmath75 here the following definitions have been made : @xmath76 and cubic and higher order combinations of @xmath77 , @xmath78 , @xmath79 and @xmath80 have been neglected . in principle , a complete description of the quantum map could be given if one could use all moments @xmath81 .
as an extra approximation , a cutoff frequency has to be introduced in order to calculate expressions which depend on eq .
( [ autocorr ] ) , where the divergences introduced by the use of the markovian approximation show up .
we will now use this set of equations to investigate the effects of quantum and thermal contributions to the dynamics of the system .
the parameters and potential used are the same of the classical map .
in this section we present the results obtained using the maps developed in the previous section .
we analyze the shifts that occur in the threshold of chaos for the system as function of the temperature @xmath0 and @xmath1 .
we also investigate how the cutoff frequency influences this behavior . in fig .
( 3 ) a bifurcation diagram for small values of @xmath1 and @xmath0 is shown ( small compared with the action and energy scales of the problem ) .
there is , as before , some coexistence of attractors .
these are included using simulation of the map over many different initial conditions .
as expected , large segments of this diagram are almost equal to the classical counterpart ( fig .
there are however , four very important differences : first of all , and quite visible in the figure , there are chaotic regions that appear far below the classical threshold of chaos ( @xmath82 ) . one of them , at @xmath83 , is associated to the second bifurcation of the classical diagram .
it shows a loss of stability for the quantum system in a parameter sector close to a point where the classical system has a zero lyapunov exponent .
this behavior is shown in figs .
( 4 ) , which shows the bifurcation diagram , and ( 5 ) , which compares the largest lyapunov exponent for the quantum and classical maps in this region .
another appears for @xmath31 around 2.5 and seems to terminate a stable fixed point branch .
a similar one appears for @xmath84 ) .
second , and also apparent in figs .
( 4 ) and ( 5 ) , is the appearance of a hopf bifurcation just before the chaotic threshold for the quantum map .
these transitions are not allowed for the classical system , since is dissipative and two dimensional . here however , they are preferred over the standard period doubling cascade .
we have verified the appearance of simple non - chaotic cycle attractors , typical of hopf bifurcations , for this transition .
it should be noticed that hopf bifurcations are also present for all other transitions from periodicity to chaos in this map , of which four are visible in fig .
( 3 ) ( @xmath85 , @xmath83 , @xmath86 and @xmath87 . ) the third main difference between the two bifurcation diagrams is given by the fact that , although the system is still dissipative , there seems to be no clear bound for its variables .
here we are showing @xmath73 only for a finite interval , but the numerical simulations generate values of @xmath73 as high as several thousands .
this peculiarity affects some chaotic regions but not all of them . finally , there is no evidence of periodic windows or band mergings inside the chaotic region .
there seems to be no topological changes in the chaotic attractor once the last threshold is crossed .
also , a great loss of detail can be clearly inferred from the bifurcation diagram .
this is of course even more evident in the attractors themselves , as seen in the example given in fig .
this attractor ( at least its projection on the @xmath88 plane ) still resembles the typical attractor for the classical map .
here we are also cropping the figure .
most of these differences are probably associated with the increase in dimensionality of the map , resulting from the use of the evm .
as mentioned , a hopf bifurcation would not be possible in the classical system .
the behavior of the quantum map , for higher values of @xmath1 and/or @xmath0 , is similar to the one already described , except for an expected growth of the chaotic strips .
it is also worth mentioning that the attractor ( at least its projection on the @xmath88 plane ) still resembles the typical attractor for the classical map ( see fig .
it of course becomes diffused in the plane , since it is just a projection .
notice that we are strongly cropping this figure , since values of @xmath74 and @xmath73 spread up to the thousandths .
the changes we have mentioned up to now signal qualitative differences between classical and quantum maps . besides these , we have also studied the evolution of bifurcation points and chaotic thresholds due to changes in @xmath1 and @xmath0 , assuming always @xmath89 , @xmath90 .
we have obtained the bifurcation point for the last hopf transition and the last chaotic threshold for different values of @xmath0 and @xmath1 . in all cases
it is found that both points are shifted towards lower values of @xmath31 with increase of either @xmath0 or @xmath1 , agreeing with previous results @xcite , which deduce this behavior from the reduction of quantization and temperature effects to an effective noise .
this shift seems to have quadratic scaling in both parameters , in difference with the non
integer exponent change given in @xcite .
this is not surprising taking into account the strong truncation in the power series we have performed .
the temperature effects can be seen in fig .
( 7 ) , where we plot the value of the potential parameter @xmath31 for which the systems enters the chaotic regime , as function of the temperature @xmath0 .
the different curves correspond to fixed values of @xmath1 .
the solid line set of curves correspond to the beginning of the hopf bifurcation where @xmath91 , and the dashed ones are for the beginning of fully developed chaos .
note that even when this curves are not as smooth as the first ones , the general tendency is maintained for all of them .
it can be seen how the value of @xmath31 decreases as @xmath0 is raised , that is , keeping everything else constant , the greater the temperature , the earlier chaos appears in the system .
the cutoff frequency has some effect on the curves , but the general behavior is the same independently of the value of @xmath92 .
the influence of the cutoff frequency is more clearly seen in fig .
( 8) , which shows the same two transitions , for fixed @xmath0 and variable @xmath1 , and in fig .
( 9 ) where it is clear that an increase in the cutoff has a similar effect as augmenting the @xmath1 or the temperature , _ i. e. _ , it shifts the thresholds for chaos towards lower values of @xmath31 .
the scaling of the threshold and bifurcation values with @xmath0 is quadratic as mentioned before .
this can be seen in fig .
this is probably due to the truncation to second order the whole evm .
a very similar behavior is obtained when we change the value of @xmath1 . in fig . ( 11 ) , the evolution of bifurcation and threshold points is shown for several fixed values of @xmath0 .
the tendency as before is to have an early onset of chaos for larger values of @xmath1 .
however , a peculiar phenomenon appears here : for small values of the cutoff parameter , this tendency is reversed , and the system gets stabilized when @xmath1 begins to grow , see fig .
this anomalous behavior does not show up when we change @xmath0 .
the kicked harmonic oscillator with dissipation has been quantized by means of the quantum langevin equation formalism , truncating the moment expansion up to second order .
we obtained the evolution maps for the expectation values @xmath73 and @xmath74 and the second order fluctuations @xmath93 , @xmath94 and @xmath95 .
the most important result we find is that chaotic behavior appears in a region of parameter space far below the classical threshold for chaos .
this agrees with the result reported by liu and schieve @xcite , who also found chaotic behavior in a system whose classical version is periodic .
we also find that the change in dimensionality in the map changes the type of bifurcations in the dynamics , as shown by the appearance of hopf bifurcations on the quantized system .
it should be noticed that this is not an artifact of the evm , since any representation of a quantum system via @xmath96numbers ( wavefunctions in configuration space , elements of the density matrix in any given base , etc . )
is infinite dimensional .
therefore , the reduction of a quantum dissipative system to the corresponding classical system plus some noise ignores the possible effects of this change in dimensionality .
once we are in a finite @xmath0 and @xmath1 domain , a continuous displacement of chaotic thresholds and bifurcation points towards lower values of @xmath31 is observed , as @xmath0 and @xmath1 are increased .
this agrees with most the results obtained up to now in this field , and goes to confirm the rule that states that _ quantization of hamiltonian systems eliminates chaos , but dissipative systems become more chaotic when they are quantized_. this is due to the loss of coherence brought in by the dissipative environment . by now
, there are some experimental confirmations of this rule @xcite . here
we should comment that irregularities observed in the curves for the threshold of chaos in figs .
( 7 ) and ( 11 ) are due to the uncertainties incurred in the numerical evaluation of the largest lyapunov exponent for the system .
we are aware , of course , that a second order expansion is not sufficient to reproduce the behavior of the infinite dimensional evm .
hoverer , we doubt that the inclusion or higher order corrections will restore periodic behavior to these new chaotic regions , although it is of course expected that some extra phenomenology will appear with the inclusion of these terms .
we have introduced a cutoff frequency @xmath92 in order to perform some of the integrals .
these cutoff frequency is expected to come from the dimensions of the physical system in a real experiment .
the existence of a cutoff implies a breakdown on the markovian condition , but , as long as @xmath97 , assuming that the system is markovian is not a very strong approximation . in our simulations , we find that this is not a very important parameter except when the obviously inconsistent case when it takes very low values . if we are given the collection of moments @xmath98 @xmath99 being the density matrix of the system , we can construct the function @xmath100 if we fourier - transform the above equation we get @xmath101 solving for @xmath102 if we want to get @xmath103 in a position basis , we can take the left hand side of the above equation and do the following @xmath104 multiply both sides by @xmath105 and integrate over @xmath20 @xmath106 and obtain the desired result @xmath107 this can also be put in the form @xmath108 in order to calculate the 5 maps for @xmath73 , @xmath74 and its three fluctuations , we will need the following terms @xmath109 we make the following definitions @xmath110 the quantities we need are now expressed as @xmath111,\\ f\cdot h_n+h_n\cdot f&=&\frac{1}{\omega_0 ^ 2}\left[-2(\gamma+v''(q_n ) ) i_s\cdot i_s+\omega_0(i_s\cdot i_c+ i_c\cdot i_s)\right],\end{aligned}\ ] ] and the integrals are explicitly given by @xmath112 where @xmath113 the integrals ( [ g1]-[g3 ] ) are evaluated exactly , and give expressions that go like @xmath114 for large @xmath47 . with them , and the given values of @xmath51 , @xmath115 , @xmath3 , @xmath0 and @xmath1 , integrals ( [ i1]-[i3 ] ) are done numerically , including a cutoff factor @xmath116 to control the logarithmic divergence .
a taylor expansion of @xmath117 is needed for @xmath118 .
references electronic address : [email protected] e. j. heller , m. f. crommie , c. p. lutz and d. m. eigler , nature * 369 * , ( 1994 ) 464 .
m. f. crommie , c. p. lutz and d. m. eigler , science * 262 * , ( 1993 ) 218 .
b. eckhardt , g. hose and e. pollak , phys .
a * 39 * , ( 1989 ) 3776 .
e. heller , phys .
53 * , ( 1984 ) 1515 .
m. c. gutzwiller , j. math .
* 8 * , ( 1967 ) 1979 ; * 10 * , ( 1969 ) 1004 ; * 11 * , ( 1970 ) 1791 ; * 12 * , ( 1971 ) 343 .
m. c. gutzwiller , _ chaos in classical and quantum mechanics _
, springer - verlag , new york , 1990 .
b. eckhardt , periodic orbit theory , lecture notes for the international school of physics enrico fermi " on quantum chaos , varena , villa monastero , 23 july 2 august 1991 .
t. brody , j. flores , j. french , p. mello , a. pandey and s. wong , rev .
phys . * 53 * , ( 1981 ) 385 .
m. methta , _ random matrices _ , academic press , new york , 1991 . m. d. kostin , j. chem . phys . * 57 * , ( 1972 ) 3589 .
h. dekker , phys . rev . a * 16 * , ( 1977 ) 2116 .
a. m. kowalski , a. plastino and a. n. proto , phys .
e * 52 * , ( 1995 ) 165 .
feynman and f. l. vernon , annals of physics * 24 * , ( 1963 ) 118 .
a. o. caldeira and a. j. legget , physica * 121a * , ( 1983 ) 587 .
a. o. caldeira and a. j. legget , phys .
* 46 * , ( 1981 ) .
u. weiss , _ quantum dissipative systems _ , world scientific , singapore , 1993 .
f. haake , _ quantum statistics in optics and solid state physics _ , springer tracts in modern physics , vol .
66 , ed . by g. holler , springer - verlag , berlin , 1973 .
h. haken , rev .
phys . , * 47 * , ( 1975 ) 67 . g. w. ford , j. t. lewis and r. f. oconnell , phys . rev .
a * 37 * , ( 1988 ) 4419 .
g. w. ford , j. t. lewis and r. f. oconnel , phys .
a * 128 * , ( 1988 ) 29 .
r. graham and t. tl , z. phys .
b * 60 * , ( 1985 ) 127 . c. w. gardiner , _ quantum noise _ , springer - verlag , berlin , 1991 .
h. a. cerdeira , r. ramaswamy and a. o.caldeira , ictp preprint ic/89/394 d. cohen , j. phys . a : math
* 27 * , ( 1994 ) 4805 .
t. dittrich and r. graham , z. phys .
b * 62 * , ( 1986 ) 515 . w. v. liu and w. c. schieve , phys .
rev . lett . * 78 * , ( 1997 ) 3278 .
m. e. goggin , b. sundaram and p. w.milonni , phys .
a. * 41 * , ( 1990 ) 5705 .
h. ammann , r. gray , i. shvarchuck and n.christensen , phys .
lett . * 80 * , ( 1998 ) 4111 .
bifurcation diagram for the quantized system .
we use the same values of @xmath115 and @xmath3 as in the classical system with a cutoff frequency @xmath122 . here
100 random initial conditions are used , and @xmath123 .
only the central part of the range in @xmath73 is shown for clarity .
comparison of the largest lyapunov exponents as function of @xmath31 .
the dotted line corresponds to the classical case , which undergoes a period doubling bifurcation . in the quantum case ( solid line )
this is replaced by a hopf bifurcation followed by a chaotic strip .
attractor for the quantum system using the same parameters as in fig .
( 3 ) , and a value of @xmath31 of 8 . only the central part of the ranges in both @xmath74 and @xmath73 are shown for clarity .
plot showing the shift of the chaotic thresholds ( dotted line ) and the hopf bifurcation ( solid line ) , with respect to @xmath0 , for two values of the cutoff frequency .
each curve within any of the four bundles corresponds to a different value of @xmath1 .
starting from the upper curve the values of @xmath1 are 0.0005 , 0.002 , 0.004 , 0.006 , 0.008 and 0.01 .
effect of the cutoff frequency on the chaotic threshold near @xmath124 .
values for @xmath0 and @xmath1 are fixed at 0.0001 and 0.0005 , respectively .
beginning of the hopf transition as a function of @xmath0 for @xmath125 .
with different values for the cutoff frecuency , @xmath126 and 125 , with @xmath92 augmenting downwards .
figure 10 .
logarithmic plot of the beginning of the hopf transition as function of @xmath0 .
the thin line has slope 2 .
figure 11 .
plot showing the shift of the chaotic thresholds ( dotted line ) and the hopf bifurcation ( solid line ) , with respect to @xmath1 , for two values of the cutoff frequency .
each curve within any of the four bundles corresponds to a different value of @xmath0 .
starting from the upper curve the values of @xmath0 are 0.0005 , 0.002 , 0.004 , 0.006 , 0.008 and 0.01 .
figure 12 .
anomalous behavior of a chaotic threshold ( upper bundle ) and hopf transition ( lower bundle ) for a low value of the cutoff frequency , as function of @xmath1 . starting from the upper curve
the values of @xmath0 are 0.0005 , 0.002 , 0.004 , 0.006 , 0.008 and 0.01 . | the effects of quantum and thermal corrections on the dynamics of a damped nonlinearly kicked harmonic oscillator are studied .
this is done via the quantum langevin equation formalism working on a truncated moment expansion of the density matrix of the system .
we find that the type of bifurcations present in the system change upon quantization and that chaotic behavior appears for values of the nonlinear parameter that are far below the chaotic threshold for the classical model . upon increase of temperature or planck s
constant , bifurcation points and chaotic thresholds are shifted towards lower values of the nonlinear parameter .
there is also an anomalous reverse behavior for low values of the cutoff frequency . |
the @xmath4 meson , as a @xmath12 bound state , can only interact with the nucleon through the gluon - exchange mechanism . by using the effective field theory method@xcite and lattice qcd @xcite
, it was found that the @xmath4-@xmath5 interaction is attractive .
strong attraction was also found@xcite within a pomeron - quark coupling model of @xmath12-nucleus interactions .
thus it is possible that @xmath4 and a system of nucleons can form nuclear bound states with @xmath13 @xmath14 .
if such @xmath15 nuclear states indeed exist and can be detected , we can get useful information for understanding the role of gluons in nuclei .
the strength of the @xmath4-@xmath5 interaction , parameterized as a yukawa form @xmath16 , deduced from refs.@xcite are rather different from each other . with @xmath17 gev ,
the calculated @xmath4-@xmath5 scattering lengths range from -0.05 fm from using @xmath18 of ref.@xcite to -8.83 fm from @xmath19 of ref.@xcite . to understand quantitatively the role of gluons in hadron interactions , it is important to investigate how the @xmath4-@xmath5 interaction can be extracted from experiments to test these theoretical models .
this is the objective of this work . clearly , this is also a necessary step toward searching for possible nuclear bound states with hidden charm , as investigated in ref.@xcite .
we first consider the @xmath20 reaction . at energies above the @xmath4 production threshold , this reaction is strongly influenced by many possible reaction channels , such as @xmath1 , @xmath21 , @xmath22 , @xmath23 channels , because of the unitarity condition .
to proceed , we cast this complex reaction problem into a manageable coupled - channel model with only @xmath0 , @xmath1 , and @xmath2 channels .
the interaction @xmath24 is chosen to be of a form related to the yukawa form of refs.@xcite . with the @xmath4-@xmath25-@xmath26 coupling constant determined@xcite by the partial decay width of @xmath27
, we can calculate the @xmath28 transition potentials @xmath29 and @xmath30 by using the one-@xmath25 and one-@xmath26 exchange mechanisms , respectively , as illustrated in fig.[fg : pi - rho - exc ] .
the other interactions , which do not connect with the @xmath2 channel directly , will be treated as phenomenological complex potentials with their parameters constrained by reproducing the total cross sections of @xmath0 reactions and @xmath31 reactions in the considered energy region .
we also constrain these potentials by reproducing the total cross section data of @xmath32 which can be related qualitatively to @xmath33 within the vector meson dominance model . to facilitate the experimental determination of the @xmath4-@xmath5 interaction , we will present results showing the dependence of the calculated cross sections of @xmath34 on the potential @xmath24 . with the meson - baryon amplitudes generated from the constructed coupled - channel model described above and the pomeron - exchange model of @xmath35 we had developed in ref.@xcite ,
we then examine the dependence of the @xmath36 cross sections on @xmath24 .
we also follow the suggestion of ref.@xcite to investigate @xmath37 reaction . instead of using the factorization approximation used in ref.@xcite ,
we have performed complete one - loop calculations to account for both the on- and off - shell @xmath2 final state interactions .
in addition , we also include the effects due to the final @xmath7 interaction by using the scattering t - matrix generated from the bonn potential@xcite .
if we use the vector meson dominance hypothesis to convert the incoming photon into @xmath4 and @xmath26 , we can predict the @xmath38 cross sections within the constructed coupled - channel model .
we will see that our predictions also depend on the @xmath4-@xmath5 potential @xmath24 .
we will also present these results which could be tested in the forthcoming experiment@xcite . in section
ii , we present our coupled - channel model for @xmath4 production on the nucleon.the formula for calculating the meson - baryon reaction amplitudes are given in section iii . in section iv , we give the formula for calculating the cross sections of the @xmath20 and @xmath39 reactions .
our results are presented in section v. a summary is given in section vi .
at energies near the @xmath4 production threshold , the @xmath0 reactions involve many meson - baryon channels such as @xmath21 , @xmath40 , @xmath41 etc . within the formulation of ref.@xcite ,
the scattering @xmath42-matrix for such reactions are defined by the following coupled - channel equations @xmath43 where @xmath44 denote the considered channels , @xmath45 is the meson - baryon propagator of channel @xmath46 , and @xmath47 are the interaction potentials .
for investigating the @xmath4-@xmath5 interaction , we will treat @xmath0 and @xmath1 channels explicitly since their interactions with the @xmath2 channel can be calculated , as will be explained later .
we thus set @xmath48 , @xmath49 in eq.([eq : cceq ] ) , where @xmath49 denotes collectively all other channels . by using the standard projection operator
technique@xcite , we can cast eq.([eq : cceq ] ) into @xmath50 where @xmath51 .
the energy - dependent interactions in eq.([eq : cceq - p ] ) are @xmath52v_{x , j}\ , , \label{eq : vij}\end{aligned}\ ] ] where the scattering amplitude @xmath53 in the subspace of the channel @xmath49 is defined by @xmath54 we now turn to defining the interaction potentials @xmath55 in eq.([eq : vij ] ) .
as noticed in refs.@xcite , the decay width @xmath56 of @xmath27 is significant . from the value of @xmath56 listed by the particle data group ( pdg )
, we can determine@xcite the coupling constant @xmath57 of the following interaction lagrangian @xmath58 we find@xcite @xmath59 . by using @xmath60 and the well known@xcite lagrangian @xmath61 and @xmath62
, we can calculate one-@xmath25 ( one-@xmath26 ) exchange transition potential of the @xmath63 ( @xmath64 ) process , as illustrated in fig.[fg : pi - rho - exc ] .
explicitly , their matrix elements can be written ( with the normalizations to be specified in section iii and omitting spin and isospin indices ) as @xmath65 u_{\vec{p}_{n_i } } , \label{eq : pinjn } \\ & & \nonumber \\
< \vec{p}_{j/\psi},\vec{p}_{n_f}| v_{j/\psi n,\rho n } & = & \frac{-1}{(2\pi)^3}\frac{1}{\sqrt{2e_{j/\psi}(\vec{p}_{j/\psi } ) } } \sqrt{\frac{m_n}{e_n(\vec{p}_{n_f})}}\sqrt{\frac{m_n}{e_n(\vec{p}_{n_i } ) } } \frac{1}{\sqrt{2e_{\rho}(\vec{p}_\rho)}}\nonumber \\ & \times&\left(\frac{g_{j/\psi\rho\pi}}{m_{j/\psi}}\frac{f_{\pi nn}}{m_\pi}\right ) \epsilon_{\mu\nu\alpha\beta}p_{\rho}^\mu \varepsilon^\nu(p_\rho ) p_{j/\psi}^\alpha \varepsilon^{\beta}(p_{j\psi } ) \nonumber \\ & & \times \frac{1}{q^2-m_\pi^2 } \bar{u}_{\vec{p}_{n_f } } q\!\!\!/ \gamma^5 u_{\vec{p}_{n_i } } \ , , \label{eq : rhonjn}\end{aligned}\ ] ] where @xmath66 , the coupling constants @xmath67 , @xmath68 , @xmath69 are taken from a dynamical model@xcite of @xmath0 scattering .
all external particles of the matrix element of the above equations are on their mass shell with their momenta defined as @xmath70 and @xmath71^{1/2}$ ] .
we assume that the mesons and baryons in the channel @xmath49 do not contain charmed quarks .
the coupling between @xmath49 and @xmath2 can then be neglected and we can set @xmath72 .
it follows that for the interactions involving @xmath2 , the second term of eq.([eq : vij ] ) vanishes and we have the following simplification , @xmath73 the matrix elements for @xmath30 and @xmath29 have been given in eqs.([eq : pinjn ] ) and ( [ eq : rhonjn ] ) . for @xmath24 ,
we follow refs.@xcite and use the following yukawa form @xmath74 to be consistent with the relativistic expression eqs.([eq : pinjn ] ) and ( [ eq : rhonjn ] ) , we assume that eq.([eq : vjpnjpn ] ) is the non - relativistic limit of a one - scalar meson exchange amplitude in field theory .
we then obtain @xmath75 \frac{v_0}{q^2-\mu^2_0 } \bar{u}_{\vec{p}_{n_f } } u_{\vec{p}_{n_i}}\ , , \label{eq : jpnjpn}\end{aligned}\ ] ] where @xmath76 , and @xmath77 . since @xmath78 , @xmath79 and @xmath80
can not be calculated theoretically , we will determine @xmath81 with @xmath82 in eq.([eq : vij ] ) phenomenologically .
for simplicity , we assume that they all have the following local form in the non - relativistic limit @xmath83 where @xmath84 is unitless such as @xmath85 or @xmath86 . in the strong absorption ( diffractive )
model@xcite , the form of @xmath87 for @xmath88 is similar to the nucleon density .
we will specify @xmath84 later .
we next define @xmath89 including appropriate covariant spin factors , the matrix elements corresponding to the form of eq.([eq : v - fr ] ) are taken to be @xmath90u_{\vec{p}_{n_i}}. \label{eq : rnpn}\end{aligned}\ ] ] we will determine the parameters @xmath91 and @xmath84 phenomenologically by fitting the total cross sections of @xmath0 and @xmath31 , and the @xmath92 cross sections which can be calculated from the predicted @xmath93 amplitudes using the vector meson dominance hypothesis .
in this work we follow the formulation of ref.@xcite within which the scattering @xmath42-matrix is related to the @xmath94-matrix by @xmath95 where @xmath96 and @xmath97 are the total energy and momentum of the state @xmath46 .
the normalizations for the plane wave state @xmath98 and bound state @xmath99 are defined by @xmath100 in our calculations , we need the meson - baryon ( @xmath101 ) matrix elements in a fast moving frame . following the instant form of relativistic quantum mechanics@xcite , we write for the two - particle @xmath102 transition @xmath103 where @xmath104 $ ] and @xmath105 $ ] are the z - components of the spin - isospin quantum numbers for @xmath106 and @xmath107 , respectively .
the total momenta @xmath108 and the energy @xmath109 in the center of mass system are @xmath110^{1/2}\ , .
\label{eq : cm - e}\end{aligned}\ ] ] the relative momentum @xmath111 and jacobian @xmath112 in eq.([eq : cm - lab ] ) are defined by the lorentz boost transformation@xcite @xmath113\ , , \label{eq : rel - k } \\ & & j(\vec{p},\vec{k};\vec{p}_m,\vec{p}_b ) = |\frac{\partial(\vec{p},\vec{k } ) } { \partial(\vec{p}_m,\vec{p}_b)}|^{1/2 } = [ \frac{e_{m}(\vec{k})e_{b}(\vec{k})h_0 } { e_m(\vec{p}_m)e_b(\vec{p}_b)m_0}]^{1/2}\ , , \label{eq : jcob}\end{aligned}\ ] ] with @xmath114 in eq.([eq : cm - lab ] ) , @xmath115 is the scattering operator in the center of mass frame . in the partial - wave representation , we can write @xmath116 where the angle - spin - isospin vector is defined by @xmath117\ , . \nonumber \\ & & \label{eq : s - i - a}\end{aligned}\ ] ] here @xmath118 $ ] and @xmath119 $ ] are the spin - isospin quantum numbers for @xmath106 and @xmath107 , respectively ; @xmath120 is the clebsch - gordon coefficient and @xmath121 the spherical harmonic function , as defined in ref.@xcite . with the same partial - wave expansion eq.([eq : s - i - a ] ) for the meson - baryon potential @xmath81 , the eqs.([eq : cceq - p ] ) and ( [ eq : cm - lab ] ) lead to the following coupled - channel equation @xmath122 where @xmath123 , and the propagator is @xmath124 we use the procedures developed in ref.@xcite to calculate the potential matrix elements @xmath125 from the matrix elements eqs.([eq : pinjn])-([eq : rnpn ] ) . for the calculations on the deuteron target , we need the deuteron bound state @xmath126 moving with a high momentum @xmath127 .
following ref.@xcite , it is defined by @xmath128 where @xmath129 with @xmath130 is the deuteron spin and @xmath131 is the usual deuteron radial wave function in its rest frame .
the jacobian @xmath132 and the relative momentum @xmath133 can be calculated by using the same eqs.([eq : rel - k])-([eq : h0m0 ] ) with the replacement @xmath134 , @xmath135 , @xmath136 , @xmath137 , and @xmath138 .
we also need the @xmath7 amplitude @xmath139 the above matrix element can be calculated from the same equations eqs.([eq : cm - lab])-([eq : s - i - a ] ) with the replacement of @xmath140 .
we generated the @xmath7 partial - wave matrix elements @xmath141 from the bonn potential@xcite .
we need to evaluate the cross sections for the reactions involving three meson - baryon ( @xmath101 ) channels with @xmath142 . with the definitions
eqs.([eq : s - t])-([eq : cm - lab ] ) , the differential cross sections in the center of mass ( cm ) for the @xmath143 reaction can be written as @xmath144\;,\nonumber \\ & & \label{eq : dcrst - mb}\end{aligned}\ ] ] where the matrix element of @xmath145 can be calculated from eqs.([eq : pwa])-([eq : s - i - a ] ) and the solution of eq.([eq : cceq - pw ] ) .
the density of state is defined by @xmath146 the total cross sections @xmath147 of elastic @xmath148 and @xmath149 of @xmath31 can be calculated from using eq.([eq : dcrst - mb ] ) , eq.([eq : pwa ] ) and the solution of the coupled - channel equation eq.([eq : cceq - pw ] ) . the total cross sections @xmath150 can be obtained from the @xmath0 elastic scattering amplitude using the optical theorem . with our definitions
eqs.([eq : s - t])-([eq : cm - lab ] ) , we have @xmath151\end{aligned}\ ] ] where the spin - averaged @xmath0 elastic scattering amplitude is @xmath152 \nonumber \\ & & \end{aligned}\ ] ] and @xmath153 is the on - shell momentum , and @xmath154 .
we now note that with the vector meson dominance hypothesis , we can predict the cross section of @xmath155 within the constructed coupled - channel model .
this is done by writing the amplitude for @xmath156 as @xmath157 where @xmath158 and @xmath159 .
obviously we can get the @xmath160 amplitude from eq.([eq : gn - jn ] ) by interchanging @xmath4 and @xmath161 .
this will allow us to calculate the total cross section @xmath162 of @xmath32 .
clearly , within this model based on the vector meson dominance hypothesis , @xmath162 is closely related to the total cross section of @xmath33 which can not be obtained experimentally , but can be essential in determining the parameters associated with our phenomenological potential @xmath163 . in fig.[fg :
diag ] , we illustrate the mechanisms included in our calculations of the cross sections of @xmath4 production on the deuteron target . since we are mainly interested in the @xmath4-@xmath5 interaction ( fig.[fg : diag].(c ) ) , we will examine how the predicted cross sections depend on the @xmath4-@xmath5 relative momentum in the final @xmath165 state .
this can be done most effectively by considering the following differential cross section in the cm frame of the @xmath166 reaction @xmath167 where @xmath168 is the photon momentum chosen to be in the quantization @xmath169-direction , @xmath170 is the deuteron with momentum @xmath171 , @xmath172 , @xmath173 and @xmath111 are the momenta of the final proton , neutron and @xmath174 , respectively .
@xmath175 is the momentum of @xmath4 in the cm of the @xmath4-@xmath5 subsystem .
the amplitudes @xmath176 , @xmath177 , and @xmath178 are calculated from the impulse ( ( a ) ) , @xmath7 re - scattering ( ( b ) ) , and @xmath4-@xmath5 re - scattering ( ( c ) ) mechanisms shown in fig.[fg : diag ] . in eq.([eq : diffcrsec ] ) , @xmath109 is the invariant mass of @xmath179 and @xmath180 is the invariant mass of the @xmath174-@xmath5 system .
the magnitude of the outgoing proton momentum @xmath173 can be calculated from@xmath181 and @xmath109 by @xmath182 the direction of @xmath173 is specified by @xmath183 with respect to the incident photon momentum @xmath168 .
the other variables @xmath111 for the outgoing @xmath4 and @xmath172 for the outgoing neutron in eq.([eq : diffcrsec ] ) can then be calculated from @xmath184 and @xmath173 as follows : @xmath185 with eqs.([eq : pp])-([eq : momtna ] ) , all kinematic variables for calculating the integrand of eq.([eq : diffcrsec ] ) are completely fixed for a given @xmath183 and @xmath184 . our task is to explore at what @xmath183 , the calculated differential cross section @xmath186 is most sensitive to the amplitude @xmath187 for the @xmath4-@xmath5 re - scattering mechanism ( c ) of fig.[fg : diag ] . in the following three subsections ,
we give formula for evaluating the matrix elements of @xmath188 , @xmath189 , and @xmath190 .
we have evaluated the clebsch - gordon coefficients associated with the isospin quantum numbers in eq.([eq : cm - lab ] ) and thus all isospin indices are suppressed and the amplitudes are on specific charged states specified explicitly as @xmath191 for neutron and @xmath192 for proton etc . the impulse amplitude of eq.([eq : diffcrsec])(fig.[fg : diag](a ) ) can be straightforwardly written as @xmath193 where @xmath194 has been defined by eq.([eq : deut - wf ] ) ( omitting isospin indices ) .
the @xmath195 amplitudes in the above expression are taken from our previous work@xcite @xmath196 \epsilon_\nu(\vec{q},\lambda_\gamma ) u_{m_{n_i}}(\vec{p}_{n_i}))\ , , \label{eq : tot - amp } \nonumber \\ & & \label{eq : photonjpsin},\end{aligned}\ ] ] where @xmath197 and @xmath198 are the contributions from the pomeron exchange and @xmath25 exchange mechanisms , respectively , as shown in the fig.[fg : photon ] .
for the pomeron exchange , the amplitude can be written as : @xmath199 with @xmath200 \right\ } \,,\label{eq : regge - g}\\ \mathcal{t}^{\mu\nu}_\mathbb{p}(t , q ) & = & i 12 \sqrt{4\pi\alpha_{\rm em } } \frac{m_{j/\psi}^2 \beta_c \beta_{c}}{f_{j/\psi } } \frac{1}{m_{j/\psi}^2-t } \left ( \frac{2\mu_0 ^ 2}{2\mu_0 ^ 2 + m_{j/\psi}^2 - t } \right ) f_1(t ) \
{ q\!\!\!/ \ , g^{\mu\nu } - q^\mu \gamma^\nu \}\ , , \label{eq : pom}\end{aligned}\ ] ] where @xmath201 , @xmath202 , @xmath203 with @xmath204 and @xmath205gev@xmath206 , @xmath207 gev , @xmath208 , @xmath209 gev@xmath206 , @xmath210 , and @xmath211 . for the @xmath25 exchange ,
the amplitude @xmath198 is : @xmath212\gamma^5\,,\nonumber\\ \label{eq : opep}\end{aligned}\ ] ] where the @xmath159 , @xmath213 , and @xmath214 mev .
all of the parameters specified above for evaluating eqs.([eq : regge - g])-([eq : opep ] ) were determined in ref.@xcite by fitting the total cross section data of @xmath215 up to the invariant mass @xmath216 gev , as shown in fig.[fg : photonpm ] .
the amplitude @xmath187 in eq.([eq : diffcrsec ] ) ( fig.[fg : diag](c ) ) is : @xmath218 + [ n \leftrightarrow p]\;\;\ ; \ } \label{eq : jpsin}\end{aligned}\ ] ] where @xmath219 , @xmath220 has been defined by eq.([eq : deut - wf ] ) , and @xmath221)-([eq : cceq - pw ] ) .
the first term in the bracket denotes the @xmath215 and @xmath222 , while the second term denotes @xmath223 and @xmath224 . the amplitude @xmath177 of eq.([eq : diffcrsec ] ) ( fig.[fg : diag](b ) )
can be written as : @xmath225\ , , \label{eq : nn}\end{aligned}\ ] ] where @xmath226 , @xmath227 , @xmath228 , @xmath229 has been defined by eq.([eq : deut - wf ] ) , and @xmath230 ) and the @xmath231 partial wave amplitudes generated from the bonn potential@xcite . by replacing @xmath233 by @xmath234 and changing notations appropriately , the formula presented in the previous subsection iv.b can be used to calculate the differential cross section @xmath235 of @xmath232 .
our first task is to determine the parameters of the potentials of the coupled - channel model presented in section ii .
the parameters for the coupling potential @xmath236 with @xmath237 have been specified there .
we determine the parameters of @xmath238 with @xmath239 , as already given in eq .
( [ eq : v - fr ] ) , by fitting the total cross section data @xmath150 of the @xmath0 reaction , @xmath147 of @xmath0 elastic scattering , @xmath149 of @xmath31 , and @xmath162 of @xmath32 .
we achieve this by choosing @xmath240 where @xmath241 fm , @xmath242 fm for all @xmath243 .
the energy dependence of the total cross sections in the energy region near the @xmath4 production threshold ( @xmath244 gev @xmath245 6 gev ) can be fitted qualitatively by setting @xmath246 \label{eq : v0ij}\end{aligned}\ ] ] our fits to @xmath247 are shown in fig.[fg : pin - fit ] .
the small isospin dependence in the data is neglected in our fits . in fig.[fg : rhon - fit ] , we see that the fits to the total cross section data of @xmath248 ( left ) and @xmath249 ( right ) are also reasonable . the resulting parameters @xmath250 , @xmath107 , and @xmath251 are listed in table [ tab : v ] .
note that the forms eqs.([eq : ws])-([eq : v0ij ] ) are purely phenomenological , and our fits in figs.[fg : pin - fit]-[fg : rhon - fit ] are only qualitatively .
but they are sufficient for estimating the order of magnitudes of the @xmath4 production cross sections . for a more quantitative calculation
, we clearly need to improve this phenomenological aspect of our model . with the fits shown in figs.[fg : pin - fit]-[fg : rhon - fit ] ,
all of the parameters for our calculations are completely fixed . in the next few subsections ,
we present our predictions for future experimental determinations of the @xmath4-@xmath5 interaction .
.the parameters for @xmath252 of eq.([eq : v0ij ] ) .
@xmath253 is the total energy in the enter of mass system .
[ cols="^,^,^,^,^,^ " , ] the simplest experiment to determine the @xmath4-@xmath5 interaction within our coupled - channel model is to measure the cross sections of @xmath254 reaction .
we first observe that the coupled - channel effects can change drastically the magnitudes and the energy - dependence of the predicted cross sections .
this is illustrated in the left side of fig.[fg : pin - cc ] . in the absence of coupled - channel effects , the results from the born approximation ( setting @xmath255 ) are the dotted blue curve .
when the interactions associated with the @xmath2 channel , @xmath256 with @xmath237 , are included in solving the coupled - channel equation , we obtain the dashed red curve which is suppressed at high @xmath109 in contrast to the raising behavior of the dotted blue curve from the born approximation .
when the complex potentials @xmath257 with @xmath82 are also included in our full calculations , we obtain the solid black curve .
we see that these complex potentials can reduce drastically the magnitudes of the cross sections .
this is due to the fact that most of the incident pions are absorbed before the @xmath4 production takes place . within our model , this absorption effect is due to the very large imaginary part of @xmath258 .
the coupled - channel effects on @xmath259 are shown in the right side of fig.[fg : pin - cc ] .
we note that the @xmath4-@xmath5 interaction , as defined in eq.([eq : vjpnjpn ] ) , is real and therefore does not have strong absorption effects on @xmath4-@xmath5 scattering . by comparing the differences between the blue dotted curve of the born approximation ( @xmath260 ) and the red dashed curve
, we see that the coupled - channel effects due to @xmath256 with @xmath237 is to increase the cross section at all @xmath109 .
the coupled - channel effects due to @xmath257 with @xmath82 are very small , as can be seen by comparing the red dashed curve and the black solid curve in the right side of fig.[fg : pin - cc ] .
this is due to the fact that the @xmath4-@xmath25-@xmath26 coupling constant @xmath261 , calculated from the partial decay width of @xmath262 , is very small in the calculation of eqs.([eq : pinjn])-([eq : rhonjn ] ) for the one - meson - exchange matrix elements of @xmath263 and @xmath264 .
we next examine the dependence of the @xmath265 cross sections on the strength @xmath46 of the @xmath4-@xmath5 potential @xmath24 . the results from using @xmath266
are compared in the left side of fig.[fg : pin - jpsin ] . the predicted total cross sections at peak positions are about 1.5 nb .
such small cross sections are due to the strong absorption of the incident pion and that the @xmath267 transition potential @xmath268 is weak , as discussed above .
the differences between three results are significant only in the region near @xmath269 gev . on the other hand ,
the predicted @xmath270 cross sections are more sensitive to @xmath46 , as seen in the right side of fig.[fg : pin - jpsin ] .
this suggests that @xmath4-@xmath5 interaction can be more easily determined in the reactions of @xmath4 production on the deuteron target in the kinematic region where the cross sections are sensitive to the @xmath270 re - scattering mechanism ( fig.[fg : diag](c ) ) .
this is what we will examine in the next two subsections .
we only consider the energy region close to the @xmath4 production threshold .
the results presented below and in the next subsection are calculated at the energy 100 mev above the @xmath4 production . as discussed in section iv.d ,
our main task is to use eq.([eq : diffcrsec ] ) to examine the dependence of the differential cross section @xmath272 of this process on the @xmath4-@xmath5 potential @xmath273 .
obviously , we need to identify the region of the outgoing proton angle @xmath274 with respect to the incident photon where the @xmath4-@xmath5 re - scattering mechanism ( ( c ) of fig.[fg : diag ] ) dominates .
we find that this is in the region close to @xmath275 , as can be seen in the left side of fig.[fg : phodpm ] .
we see that in the low @xmath4 momentum , @xmath276 about 200 mev , the contribution from the @xmath4 re - scattering term ( blue dotted curve ) is much larger than that from the other two mechanisms . on the other hand , the impulse term dominates at @xmath277 as seen in the right side of fig.[fg : phodpm ] .
accordingly , three @xmath4-@xmath5 potentials with the strengths @xmath278 , @xmath279 , and @xmath280 obtained by the effective field theory approach and lqcd can be easily tested at @xmath275 , but less possible at @xmath281 .
this is shown in fig.[fg : phodjn ] .
the results presented above clearly indicate that it is highly desirable to have data in the region close to @xmath275 in future experiments in order to be able to determine the @xmath4-@xmath5 interaction . for the calculated @xmath235 of the @xmath282 reaction
, we find that the impulse term dominates at all angles .
furthermore , the contributions from the @xmath4-@xmath5 re - scattering ( fig.[fg : diag](c ) ) are weaker than those of the @xmath7 re - scattering ( fig.[fg : diag](b ) ) . nevertheless , the data from experiments on this reaction can be useful to determine the @xmath4-@xmath5 interaction .
this is illustrated in fig.[fg : pinpid ] where the results from @xmath283 , @xmath279 , and @xmath280 for @xmath275 and @xmath284 are compared . with the vector meson dominance hypothesis
, we can predict the cross section of @xmath155 within the constructed coupled - channel model .
this is done by using the amplitude given in eq.([eq : gn - jn ] ) . for three coupled - channel amplitudes calculated with @xmath286 , @xmath279 , and @xmath280 for @xmath287 ,
our results are compared with that of the pomeron - exchange model ( green dot - dashed curve ) in the left of fig.[fg : phpphd ] .
we note that the coupled - channel model results are expected to be valid only at energies near @xmath4 production threshold where the data are rather uncertain . on the other hand ,
the use of pomeron - exchange model at very low energies is also questionable . our results for @xmath18 are close to the three data points near 4.5 gev . at @xmath288 about 5
gev , we clearly need to improve the model
. the data from the forthcoming experiment@xcite at jlab will be useful to clarify the situation . in the right side of fig.[fg :
phpphd ] , we show that the coupled - channel effects can increase the cross section significantly .
our results here as well as that shown in fig.[fg : pin - cc ] suggest that any attempt to determine the @xmath4-@xmath5 interaction must include the coupled - channel effects , as required by the unitarity condition .
we have developed a coupled - channel model with @xmath0 , @xmath1 and @xmath2 channels to predict the @xmath4 production on the nucleon .
the @xmath4-@xmath5 interaction potential @xmath289 is taken from the calculations using the effective field theory method and lattice qcd .
the @xmath290 transition potentials @xmath291 and @xmath292 are calculated from one-@xmath26-exchange and one-@xmath25-exchange mechanisms with the parameters determined by the decay width of @xmath27 and the previously determined @xmath293 and @xmath294 coupling constants .
the interactions @xmath257 with @xmath295 are treated as phenomenological complex potentials with their parameters constrained by the fits to the data of total cross sections of the @xmath0 reactions , @xmath148 , @xmath31 , and @xmath296 .
the calculated meson - baryon amplitudes are then used to predict the cross sections of @xmath297 and @xmath298 reactions .
the calculations on the deuteron target involve the contributions from the impulse term and the final @xmath7 and @xmath2 scattering .
we have identified the kinematic region where the @xmath4-@xmath5 potentials can be distinguished in @xmath299 , @xmath9 and @xmath300 reactions .
predictions of the dependence of the cross sections of these reactions on the @xmath4-@xmath5 potentials are presented .
our results shown in figs.[fg : pin - jpsin ] , [ fg : phodjn ] , and [ fg : pinpid ] can be used to facilitate the experimental determinations of @xmath4-@xmath5 interactions . within the vector meson dominance model ,
we have also applied the constructed coupled - channel model to predict the @xmath11 cross sections .
our results near the @xmath4 production threshold , as shown in fig.[fg : phpphd ] , are very different from what can be calculated from the conventional pomeron - exchange model which is mainly constrained by the @xmath301 cross section at high energies
. it will be interesting to distinguish these two models in the forthcoming experiment at jefferson laboratory@xcite .
we thank r. machleidt for providing us with the code for generating the @xmath7 amplitudes from the bonn potential .
this work is supported by the u.s .
department of energy , office of nuclear physics division , under contract no .
de - ac02 - 06ch11357 .
this research used resources of the national energy research scientific computing center , which is supported by the office of science of the u.s .
department of energy under contract no .
de - ac02 - 05ch11231 , and resources provided on `` fusion , '' a 320-node computing cluster operated by the laboratory computing resource center at argonne national laboratory .
m. e. peskin , nucl .
b * 156 * , 365 ( 1979 ) .
m. e. luke , a. v. manohar and m. j. savage , phys .
b * 288 * , 355 ( 1992 ) .
s. j. brodsky and g. a. miller , phys .
b * 412 * , 125 ( 1997 ) .
a. b. kaidalov and p. e. volkovitsky , phys .
lett . * 69 * , 3155 ( 1992 ) .
t. kawanai and s. sasaki , phys .
d * 82 * , 091501 ( 2010 ) .
s. j. brodsky , i. a. schmidt and g. f. de teramond , phys .
lett . * 64 * , 1011 ( 1990 ) .
j. -j .
wu and t. -s .
h. lee , phys .
c * 86 * , 065203 ( 2012 ) u. camerini , j. g. learned , r. prepost , c. m. spencer , d. e. wiser , w. ash , r. l. anderson and d. ritson _ et al .
_ , phys .
* 35 * , 483 ( 1975 ) .
b. gittelman , k. m. hanson , d. larson , e. loh , a. silverman and g. theodosiou , phys .
lett . * 35 * , 1616 ( 1975 ) .
r. l. anderson , excess muons and new results in @xmath302 photoproduction .
slac - pub-1471 ( unpublished ) .
m. e. binkley , c. bohler , j. butler , j. p. cumalat , i. gaines , m. gormley , d. harding and r. l. loveless _ et al .
_ , phys . rev
. lett . *
48 * , 73 ( 1982 ) .
b. h. denby , v. k. bharadwaj , d. j. summers , a. m. eisner , r. g. kennett , a. lu , r. j. morrison and m. s. witherell _ et al .
_ , phys .
lett . * 52 * , 795 ( 1984 ) .
r. barate _
[ na14 collaboration ] , z. phys .
c * 33 * , 505 ( 1987 ) .
l. frabetti _ et al . _
[ e687 collaboration ] , phys .
b * 316 * , 197 ( 1993 ) .
s. aid _ et al .
_ [ h1 collaboration ] , nucl .
b * 472 * , 3 ( 1996 ) .
_ et al . _
[ h1 collaboration ] , eur .
j. c * 46 * , 585 ( 2006 ) .
j. breitweg _ et al . _
[ zeus collaboration ] , z. phys .
c * 76 * , 599 ( 1997 ) .
s. chekanov _ et al . _
[ zeus collaboration ] , eur .
j. c * 24 * , 345 ( 2002 ) .
s. chekanov _ et al . _
[ zeus collaboration ] , nucl .
b * 695 * , 3 ( 2004 ) .
m. derrick _ et al .
_ [ zeus collaboration ] , phys .
b * 350 * , 120 ( 1995 ) .
j. beringer et al .
[ particle data group collaboration ] , phys .
d * 86 * , 010001 ( 2012 ) .
a. baldini , v. flamino , w.g .
moorhead and d.r.o .
morrison , landolt - brnstein , _ numerical data and functional relationships in science and technology _ , vol .
12 , total cross sections for reactions of high energy particles , ed . by h. schopper ( springer - verlag , berlin , 1988 ) .
y. .a .
aleksandrov , s. s. baranov , a. s. belousov , n. p. budanov , y. .a .
vazdik , b. b. govorkov , v. v. kim and v. a. kozlov _ et al .
_ , yad . fiz .
* 32 * , 651 ( 1980 ) .
j. ballam , g. b. chadwick , y. eisenberg , e. kogan , k. c. moffeit , p. seyboth , i. o. skillicorn and h. spitzer _ et al .
_ , phys .
d * 7 * , 3150 ( 1973 ) .
j. park , m. davier , i. derado , d. e. c. fries , f. f. liu , r. f. mozley , a. odian and w. p. swanson _ et al .
_ , nucl .
b * 36 * , 404 ( 1972 ) .
meziani , k. hafidi , x. qian , and n. sparveris et al .
, proposal `` near threshold electroproduction of @xmath4 at 11 gev '' , pr12 - 12 - 006(2012 ) , pac39 , jefferson laboratory ( 2012 ) | a coupled - channel model with @xmath0 , @xmath1 and @xmath2 channels is developed to predict the @xmath3 cross sections .
the @xmath4-@xmath5 interaction is parameterized in a form related to what has been predicted by the effective field theory approach and lattice qcd .
the other interactions within the model are constrained by the decay width of @xmath6 and the total cross section data of @xmath0 reactions .
the calculated meson - baryon amplitudes are then used to predict the cross sections of the @xmath4 production on the deuteron target by including the contributions from the impulse term and the one - loop calculations of the final @xmath7 and @xmath2 re - scattering effects .
predictions of the dependence of the cross sections of @xmath8 , @xmath9 , and @xmath10 on the @xmath4-@xmath5 potentials are presented for experimental determinations of the @xmath4-@xmath5 interaction . within the vector meson dominance model ,
we have also applied the constructed coupled - channel model to predict the @xmath11 cross sections near the @xmath4 production threshold . |
a classical theorem of dilworth @xcite states that each finite distributive lattice is isomorphic to the congruence lattice of a finite lattice . since this first result
, the _ congruence lattice representation problem _ has attracted many researchers , and dozens of papers belonging to this topic have been written .
the story of this problem were mile - stoned by huhn @xcite and schmidt @xcite , reached its summit in wehrung @xcite and rika @xcite , and was summarized in grtzer @xcite ; see also czdli @xcite for some additional , recent references . in @xcite
, grtzer started an analogous new topic of lattice theory .
namely , for a lattice @xmath0 , let @xmath3 denote the ordered set of principal congruences of @xmath0 .
a congruence is _ principal _ if it is generated by a pair @xmath4 of elements .
ordered sets and lattices with 0 and 1 are called _
bounded_. clearly , if @xmath0 is a bounded lattice , then @xmath1 is a bounded ordered set . the pioneering theorem in grtzer @xcite states the converse : each bounded ordered set @xmath5 is isomorphic to @xmath1 for an appropriate bounded lattice @xmath0 . actually , the lattice he constructed is of length 5 .
up to isomorphism , he also characterized finite bounded ordered sets as the @xmath1 of finite lattices @xmath0 .
unless otherwise stated , we follow the standard terminology and notation of lattice theory ; see , for example , grtzer @xcite .
our terminology for weak perspectivity is the classical one taken from grtzer @xcite . _
ordered sets _ are nonempty sets equipped with _ orderings _ , that is , with reflexive , transitive , antisymmetric relations .
note that an ordered set is often called a _ partially ordered set _ , which is a rather long expression , or a _ poset _ , which is not tolerated by spell - checkers , or an _ order _ , which has several additional meanings .
motivated by grtzer s theorem mentioned above , our goal is to prove the following theorem .
a set @xmath6 is _ countable _ if it is finite or countably infinite , that is , if @xmath7 .
an ordered set @xmath5 is _ directed _ if each two - element subset of @xmath5 has an upper bound in @xmath5 .
nonempty down - sets of @xmath5 and subsets @xmath8 are called _ order ideals _ and _ principal @xmath9order@xmath10 ideals _ , respectively .
[ thmmain ] [ thmmaina ] an ordered set @xmath11 is isomorphic to @xmath1 for some _ countable _ lattice @xmath0 if and only if @xmath5 is a countable directed ordered set with zero .
[ thmmainb ] if @xmath5 is an ordered set with zero and it is the union of a chain of principal ideals , then there exists a lattice @xmath0 such that @xmath12 .
an alternative way of formulating the condition in part is to say that @xmath13 and there is a cofinal chain in @xmath5 . for a pair @xmath14 of elements ,
the least congruence collapsing @xmath15 and @xmath16 is denoted by @xmath17 or @xmath18 . as it was pointed out in grtzer @xcite
, the rule @xmath19 implies that @xmath1 is always a directed ordered set with zero .
therefore , the first part of the theorem will easily be concluded from the second one .
to compare part of our theorem to grtzer s result , note that a bounded ordered set @xmath5 is always a union of a ( one - element ) chain of principal ideals .
of course , no _ bounded _
lattice @xmath0 can represent @xmath5 by @xmath20 if @xmath5 has no greatest element .
first of all , we need the key idea , illustrated by figure [ fig4 ] , from grtzer @xcite .
second , we feel that without the quasi - coloring technique developed in czdli @xcite , the investigations leading to this paper would have not even begun . as opposed to colorings , the advantage of quasi - colorings is that we have joins ( equivalently , the possibility of generation ) in their range sets .
this allows us to decompose our construction into a sequence of elementary steps .
each step is accompanied by a quasiordering .
if several steps , possibly infinitely many steps , are carried out , then the join of the corresponding quasiorderings gives a satisfactory insight into the construction .
even if it is the `` coloring versions '' of some lemmas that we only use at the end , it is worth allowing their quasi - coloring versions since this way the proofs are simpler and the lemmas become more general .
third , the idea of using appropriate auxiliary structures is taken from czdli @xcite .
their role is to accumulate all the assumptions our induction steps will need .
the rest of the paper is devoted to the proof of theorem [ thmmain ] .
a _ quasiordered set _ is a structure @xmath21 where @xmath22 is a set and @xmath23 is a reflexive , transitive relation on @xmath2 .
quasiordered sets are also called preordered ones . instead of @xmath24 ,
we usually write @xmath25 .
also , we write @xmath26 and @xmath27 for the conjunction of @xmath28 and @xmath29 , and that of @xmath30 and @xmath31 , respectively . if @xmath32 and @xmath33 for all @xmath34 , then @xmath35 is a _ greatest element _ of @xmath2 ; _ least elements _ are defined dually .
they are not necessarily unique ; if they are , then they are denoted by @xmath36 and @xmath37 .
if for all @xmath38 , there exists a @xmath39 such that @xmath40 and @xmath41 , then @xmath21 is a _ directed _ quasiordered set . given @xmath22 , the set of all quasiorderings on @xmath2
is denoted by @xmath42 .
it is a complete lattice with respect to set inclusion . for @xmath43 , the least quasiorder on @xmath2 that includes @xmath6
is denotes by @xmath44 .
we write @xmath45 instead of @xmath46 .
[ fig1 ] ] let @xmath0 be a lattice . for @xmath47 , @xmath48 is called an _ ordered pair _ of @xmath0 if @xmath49 . the set of ordered pairs of @xmath0
is denoted by @xmath50 .
note that we shall often use that @xmath51 holds for sublattices @xmath52 of @xmath0 ; this explains why we work with ordered pairs rather than intervals .
note also that @xmath4 is an ordered pair iff @xmath53 is a quotient ; however , the concept of ordered pairs fits better to previous work with quasi - colorings . by a _ quasi - colored lattice _
we mean a structure @xmath54 where @xmath0 is a lattice , @xmath21 is a quasiordered set , @xmath55 is a surjective map , and for all @xmath56 , * if @xmath57 , then @xmath58 ; * if @xmath58 , then @xmath57 .
this concept is taken from czdli @xcite . prior to @xcite ,
the name coloring " was used for surjective maps onto antichains satisfying in grtzer , lakser , and schmidt @xcite , and for surjective maps onto antichains satisfying in grtzer @xcite . however , in @xcite , @xcite , and @xcite , @xmath59 was defined only for covering pairs @xmath60 . to emphasize that @xmath61 and @xmath62 belong to the ordered set @xmath1 , we usually write @xmath58 rather than @xmath63 .
it follows easily from , , and the surjectivity of @xmath64 that if @xmath65 is a quasi - colored set , then @xmath21 is a directed quasiordered set with least element ; possibly with many least elements .
we say that a quadruple @xmath66 is an _
@xmath67-quadruple _ of @xmath0 if @xmath68 is a six - element sublattice , see figure [ fig1 ] .
if , in addition , @xmath69 and @xmath70 , then we speak of a _ spanning @xmath67-quadruple_. an @xmath67-quadruple of @xmath0 is called a _ strong @xmath67-quadruple _ if it is a spanning one and , for all @xmath71 and @xmath72 , @xmath73 for a subset @xmath6 of @xmath74 , the least lattice congruence including @xmath6 is denoted by @xmath75 . in particular , @xmath76 .
the least and the largest congruence of @xmath0 are denoted by @xmath77 and @xmath78 , respectively .
now , we are in the position to define the key concept we need . in the present paper , by a _ auxiliary structure _ we mean a structure @xmath79 such that the following eight properties hold : * @xmath80 is a quasi - colored lattice ; * the quasiordered set @xmath21 has exactly one least element , @xmath37 , and at most one greatest element ; * @xmath81 and @xmath82 are @xmath83 maps such that @xmath84 and , for all @xmath85 , @xmath86 ; note that we often write @xmath87 and @xmath88 instead of @xmath89 and @xmath90 , respectively ; * for all @xmath91 , @xmath92 ; * if @xmath93 and @xmath94 are distinct elements of @xmath95 , then @xmath96 is an @xmath67-quadruple of @xmath0 ; * if @xmath97 , @xmath98 , and @xmath96 is a spanning @xmath67-quadruple , then it is a strong @xmath67-quadruple of @xmath0 ; * if @xmath0 is a bounded lattice and @xmath99 , then @xmath100 * if @xmath101 and @xmath99 , then @xmath102 .
it follows from that @xmath103 is disjoint from @xmath104 , provided @xmath105 and @xmath85 .
if @xmath21 is a quasiordered set , then @xmath106 is an equivalence relation , and the definition @xmath107\theta_\nu}\leq { [ y]\theta_\nu}\iff x{\mathrel{\leq_\nu}}y$ ] turns the quotient set @xmath108 into an ordered set @xmath109 .
the importance of our auxiliary structures is first shown by the following lemma .
[ impclm ] if @xmath110 in is an auxiliary structure , then the ordered set @xmath1 is isomorphic to @xmath109 .
in particular , if @xmath111 is an ordering , then @xmath1 is isomorphic to the ordered set @xmath21 .
clearly , @xmath112 .
consider the map @xmath113 , defined by @xmath114\theta_\nu}$ ] .
if @xmath115 , then @xmath116\theta_\nu } = { [ \gamma({{\langle x_2,y_2\rangle}})]\theta_\nu}$ ] follows from .
hence , @xmath117 is a map .
it is surjective since so is @xmath64 . finally , it is bijective and an order isomorphism by and .
we say that an auxiliary structure @xmath118 is _ countable _ if @xmath119 and @xmath120 .
next , we give an example .
[ exegy ] let @xmath2 be a set , finite or infinite , such that @xmath121 and @xmath105 .
let us define @xmath122 ; note that @xmath21 is an ordered set ( actually , a modular lattice of length 2 ) .
let @xmath0 be the lattice depicted in figure [ fig2 ] , where @xmath123 is the set @xmath124 . for @xmath125 ,
@xmath126 is defined by the labeling of edges .
note that , in figure [ fig2 ] , we often write @xmath127 and @xmath128 rather than @xmath37 and @xmath36 , because of space consideration .
let @xmath129 for @xmath72 , and let @xmath130 for @xmath131 if @xmath132 .
let @xmath133 . for @xmath134 , we define @xmath135 and @xmath136 .
now , obviously , @xmath118 is an auxiliary structure .
if @xmath137 , then @xmath110 is countable .
[ fig2 ] ] substructures are defined in the natural way ; note that @xmath138 will not be required below .
namely , [ defbeagy ] let @xmath118 and @xmath139 be auxiliary structures .
we say that @xmath110 is a _ substructure _ of @xmath140 if the following hold : @xmath0 is a sublattice of @xmath141 , @xmath142 , @xmath143 , and @xmath144 ; @xmath64 is the restriction of @xmath145 to @xmath50 , @xmath146 is the restriction of @xmath147 to @xmath2 , and @xmath82 is the restriction of @xmath148 to @xmath2 . clearly , if @xmath110 , @xmath140 , and @xmath149 are auxiliary structures such that @xmath110 is a substructure of @xmath140 and @xmath140 is a substructure of @xmath149 , then @xmath110 is a substructure of @xmath149 ; this fact will be used implicitly .
the following lemma indicates how easily but efficiently we can work with auxiliary structures . for an auxiliary structure @xmath118 and an arbitrary ( possibly empty ) set @xmath150
, we define the following objects .
let @xmath151 be the disjoint union @xmath152 , and let @xmath153 .
define @xmath154 by @xmath155 consider the lattice @xmath156 defined by figure [ fig3 ] , where @xmath157 denote the elements of @xmath150 .
the thick dotted lines indicate @xmath158 but not necessarily @xmath159 ; they are edges only if @xmath0 is bounded .
note that all `` new '' lattice elements distinct from @xmath160 and @xmath161 , that is , all elements of @xmath162 , are complements of all `` old '' elements .
extend @xmath146 and @xmath82 to maps @xmath163 by letting @xmath164 and @xmath165 for @xmath166 .
define @xmath167 by @xmath168 by space consideration again , the edge label @xmath128 in figure [ fig3 ] stands for @xmath169 .
finally , let @xmath170 .
the straightforward proof of the following lemma will be omitted .
[ lupstp ] if @xmath110 is an auxiliary structure , then so is @xmath171 .
furthermore , @xmath110 is a substructure of @xmath171 , and if @xmath110 and @xmath150 are countable , then so is @xmath171 .
moreover , if @xmath172 such that @xmath173 and @xmath174 , then @xmath175 is a strong @xmath67-quadruple .
since new bottom and top elements are added , we say that @xmath171 is obtained from @xmath110 by a _ vertical extension _ ; this motivates the triangle aiming upwards in its notation .
[ fig3 ] ] [ fig4 ] ]
the key role in grtzer @xcite is played by the lattice @xmath176 ; see figure [ fig4 ] .
we also need this lattice .
assume that @xmath177 the equivalence of `` spanning '' and `` strong '' in follows from .
we define a structure @xmath178 as follows , and it will take a lot of work to prove that it is an auxiliary structure .
we call @xmath178 a _ horizontal extension _ of @xmath110 ; this explains the horizontal triangle in the notation . by changing the sublattice @xmath179 into an @xmath176 as it is depicted in figure [ fig4 ] ,
that is , by inserting the black - filled elements of figure [ fig4 ] into @xmath0 , we obtain an ordered set denoted by @xmath180 ; see also later for more exact details .
( we will prove that @xmath180 is a lattice and @xmath0 is a sublattice in it . ) the construction of @xmath178 from @xmath110 is illustrated in figure [ fig5 ] .
note that there can be much more elements and in a more complicated way than indicated .
the solid lines represent the covering relation but the dotted lines are not necessarily edges . the new lattice @xmath180 is obtained from @xmath0 by inserting the black - filled elements . note
that while grtzer @xcite constructed a lattice of length 5 , here even the interval , say , @xmath181 $ ] can be of infinite length . from @xmath110 [ fig5 ]
] let @xmath182 . in @xmath183
, we define @xmath184 .
we extend @xmath64 to @xmath185 by @xmath186 the definition of @xmath187 is also illustrated in figure [ fig5 ] , where the edge color @xmath128 stands for @xmath188 .
finally , after letting @xmath189 and @xmath190 , we define @xmath191 [ spalislat ] if @xmath110 satisfies , then @xmath180 is a lattice and @xmath0 is a sublattice of @xmath180 .
first , we describe the ordering of @xmath180 more precisely ; this description is the real definition of @xmath180 .
let @xmath192 here @xmath193 is isomorphic to the lattice @xmath176 , and its `` boundary '' , @xmath194 , to @xmath67 .
the elements of @xmath0 , @xmath195 , and @xmath194 are called _ old _ , _ new _ , and _ boundary _ elements , respectively . for @xmath196 , we define @xmath197 @xmath198 observe that for @xmath199 and @xmath200 , the conjunction of @xmath201 and @xmath202 implies @xmath203 .
hence , it is straightforward to see that @xmath204 is an ordering and @xmath205 is the restriction of @xmath204 to @xmath0 . for @xmath206 , there is a unique least element @xmath207 of @xmath194 such that @xmath208 ( that is , @xmath209 ) .
if @xmath72 , then we let @xmath210 . similarly , for @xmath206 , there is a unique largest element @xmath211 of @xmath194 such that @xmath212 .
again , for @xmath72 , we let @xmath213 . with this notation , is clearly equivalent to @xmath214 next , for @xmath215 , we want to show that @xmath216 and @xmath217 has a join in @xmath180 .
we can assume that @xmath218 has an upper bound @xmath219 in @xmath195 , because otherwise @xmath220 would clearly be the join of @xmath216 and @xmath217 in @xmath180 .
if @xmath219 belonged to @xmath221 , then the principal ideal @xmath222 ( taken in @xmath180 ) would be a chain , and this would contradict @xmath223 .
hence , @xmath224 .
if both @xmath216 and @xmath217 belong to @xmath195 , then @xmath223 gives @xmath225 , @xmath219 and @xmath226 are the only upper bounds of @xmath218 , and @xmath219 is the join of @xmath216 and @xmath217 .
hence , we can assume that @xmath72 .
if @xmath217 also belongs to @xmath0 , then @xmath227 and @xmath228 yields @xmath229 , and @xmath230 is the join of @xmath216 and @xmath217 in @xmath180 since @xmath219 was an arbitrary upper bound of @xmath218 in @xmath195 . therefore , we can assume that @xmath72 and @xmath231 .
it follows from @xmath232 that , for each @xmath233 , @xmath234 has a smallest element ; we denote it by @xmath235 . for @xmath236 ,
we let @xmath237 .
note that , for every @xmath238 , @xmath235 is the smallest element of @xmath239 .
the existence of @xmath219 , mentioned above , implies that @xmath240 .
we assert that @xmath241 is the join of @xmath216 and @xmath217 in @xmath180 .
( note that @xmath242 . )
we can assume @xmath243 since otherwise @xmath244 is the only upper bound of @xmath217 in @xmath0 and @xmath245 is clear .
consider an upper bound @xmath246 of @xmath216 and @xmath217 .
since @xmath243 , we have @xmath247 and @xmath248 . from @xmath215 and @xmath240
, we obtain @xmath249 .
since @xmath250 is a strong @xmath67-quadruple by , the validity of for @xmath110 implies @xmath251 .
this shows that @xmath252 is the join of @xmath216 and @xmath217 in @xmath180 .
the case @xmath47 showed that @xmath253 is a subsemilattice of @xmath254 . for later reference
, we summarize the description of join in a concise form as follows ; note that @xmath223 is not assumed here : @xmath255 we have shown that any two elements of @xmath180 have a join .
although @xmath193 and the construction of @xmath180 are not exactly selfdual , by interchanging the role of @xmath256 and that of @xmath221 , we can easily dualize the argument above . thus
, we conclude that @xmath180 is a lattice and @xmath0 a sublattice of @xmath180 .
the following lemma is due to dilworth @xcite , see also grtzer ( * ? ? ?
* theorem iii.1.2 ) .
[ llajtorja ] if @xmath0 is a lattice and @xmath257 , then the following three conditions are equivalent . @xmath58 ; @xmath258 ; there exists an @xmath259 and there are @xmath260 for @xmath261 and @xmath262 for @xmath263 such that the following equalities and inequalities hold : @xmath264 the situation of lemma [ llajtorja ] is outlined in figure [ fig6 ] ; note that not all elements are depicted , and the elements are not necessarily distinct .
the second half of says that , in terms of grtzer @xcite , @xmath265 is _ weakly _ up or down perspective into @xmath266 ; up for @xmath267 odd and down for @xmath267 even .
besides weak perspectivity , we shall also need a more specific concept ; recall that @xmath268 is _ perspective _ to @xmath269 if there are @xmath270 such that @xmath271 , @xmath272 , and @xmath273 . for @xmath274 [ fig6 ] ] for a quasiordered set @xmath21 and @xmath275
, we say that @xmath93 is a _
join _ of the elements @xmath276 , in notation , @xmath277 , if @xmath278 for all @xmath279 and , for every @xmath280 , the conjunction of @xmath281 for @xmath282 implies @xmath283 .
this concept is used in the next lemma . note that even if a join exists , it need not be unique . [ chainlemma ] if @xmath284 is a quasi - colored lattice and @xmath285 is a finite chain in @xmath0 ,
then @xmath286 let @xmath287 and @xmath288 .
since @xmath289 , yields @xmath278 for all @xmath279 .
next , assume that @xmath280 such that @xmath281 for all @xmath279 . by the surjectivity of @xmath64 ,
there exists a @xmath290 such that @xmath291 .
it follows by that @xmath292 .
since @xmath293 is transitive and collapses the pairs @xmath294 , it collapses @xmath295 .
hence , @xmath296 , and implies @xmath283 .
now , we are in the position to deal with the following lemma .
[ mainlemma ] the structure @xmath178 , which is defined in with assumption , is an auxiliary structure , and @xmath110 is a substructure of @xmath178 .
furthermore , if @xmath110 is countable , then so is @xmath178 .
since we work both in @xmath110 and @xmath178 , relations , operations and maps are often subscripted by the relevant structure . by lemma [ spalislat ]
, @xmath180 is a lattice . obviously , and hold for @xmath178 . since @xmath187 is
an extension of @xmath64 , @xmath297 , @xmath298 , and @xmath0 is a sublattice of @xmath180 , we obtain that and hold in @xmath178 .
let @xmath299 .
since @xmath111 is transitive , @xmath98 , and @xmath300 , we obtain that @xmath301 this clearly implies that holds for @xmath178 .
it follows from that if @xmath302 and @xmath303 , then we have @xmath304 . combining this with , we obtain easily that for all @xmath305 , @xmath306 let @xmath307 denote the congruence of @xmath0 described in . consider the equivalence relation @xmath308 on @xmath180 whose classes ( in other words , blocks ) are the @xmath307-classes , @xmath221 and @xmath309 . based on and its dual , a straightforward argument shows that , for all @xmath310 , @xmath311 iff @xmath312 .
clearly , the intersection of @xmath308 and the ordering of @xmath180 is transitive .
hence , we conclude that @xmath308 is a congruence on @xmath180 . since it is distinct from @xmath313 , @xmath178 satisfies .
next , we prove the converse of .
assume that @xmath305 such that @xmath314 ; we want to show that @xmath315 . since this is clear if @xmath316 , we assume @xmath317 . first , if @xmath47 , then let @xmath318 .
applying to @xmath64 and to @xmath110 , we obtain @xmath319 .
hence @xmath307 , which we used in the previous paragraph , collapses @xmath48 , and @xmath320 .
second , if @xmath321 , then @xmath48 is perspective to @xmath322 or @xmath323 , whence @xmath324 reduces the present case to the previous one .
finally , @xmath325 is excluded since then @xmath48 would be @xmath188-colored .
now , after verifying the converse of , we have proved that for all @xmath305 , @xmath326 next , to prove that @xmath187 satisfies , assume that @xmath327 such that @xmath328 .
let @xmath329 , for @xmath71 .
we have to show @xmath330 . by
, we can assume that @xmath331 .
thus , by , we have @xmath332 .
we can also assume that @xmath333 since otherwise @xmath334 would clearly imply @xmath330 .
thus , @xmath335 . by the construction of @xmath180
, @xmath336 is perspective to some @xmath337 such that @xmath338 , and perspectivity implies @xmath339 .
therefore , we can assume that @xmath340 , because otherwise we could work with @xmath341 and @xmath342 .
according to , we distinguish two cases .
first , assume that @xmath343 .
since @xmath187 extends @xmath64 , we have @xmath344 . applying to @xmath64
, we obtain @xmath345 . using lemma [ llajtorja ] , first in @xmath0 and then , backwards , in @xmath180
, we obtain @xmath346 , which yields @xmath330 .
second , assume that @xmath347 and @xmath348 .
since @xmath349 and @xmath350 by , the argument of the previous paragraph yields that we have @xmath351 and @xmath352 . clearly ( or applying lemma [ llajtorja ] within @xmath353 , we have @xmath354 . hence , transitivity yields @xmath330
consequently , @xmath187 satisfies .
next , to prove that @xmath187 satisfies , assume that @xmath355 such that @xmath356 . our purpose is to show the inequality @xmath328 . by , we can assume @xmath357 , and we can obviously assume @xmath358 .
that is , @xmath359 .
a pair @xmath360 is called _ mixed _ if @xmath361 .
that is , if one of the components is old and the other one is new .
it follows from the construction of @xmath178 and that none of @xmath362 and @xmath363 is mixed . if @xmath362 is a new pair , that is , if @xmath364 , then we can consider an old pair @xmath365 such that @xmath366 and , by perspectivity , @xmath367
hence , we can assume that @xmath362 is an old pair , and similarly for the other pair .
that is , we assume that both @xmath362 and @xmath363 belong to @xmath368 . the starting assumption @xmath356 means that @xmath369 .
this is witnessed by lemma [ llajtorja ] .
let @xmath370 be elements for @xmath371 and @xmath372 that satisfy ; see also figure [ fig6 ] . to ease our terminology , the ordered pairs @xmath373 will be called
_ witness pairs _ ( of the containment @xmath369 ) . since @xmath357 ,
none of the witness pairs generate @xmath313 .
thus , by , @xmath374 take two consecutive witness pairs , @xmath265 and @xmath266 . here @xmath375 .
our next purpose is to show that @xmath376 we assume @xmath377 since trivially holds if these two elements are equal .
hence , @xmath378 also holds .
if both @xmath265 and @xmath266 are old pairs , that is , if they belong to @xmath50 , then yields @xmath379 . from this , we conclude the relation @xmath380 by , applied for @xmath110 , and we obtain the validity of for old witness pairs , because @xmath187 extends @xmath64 .
if both @xmath265 and @xmath266 are new pairs , that is , if they belong to @xmath381 , then and allow only two possibilities : @xmath382 , or @xmath383 and @xmath384 . in both cases , holds . [ old2newcase ] assume first that @xmath267 is odd , that is , @xmath265 is weakly up - perspective into @xmath266 . since @xmath385 ,
being a new element , and @xmath386 are both distinct from @xmath387 , @xmath388 since @xmath389 , @xmath386 is an old element , and @xmath390 is a new one , @xmath391 .
taking @xmath392 and into account , we obtain @xmath393 and @xmath394 . applying the definition of @xmath204 for the elements of the old witness pair and using the `` weak up - perspectivity relations '' from
, we have @xmath395 .
similarly , but also taking @xmath396 into account , we obtain @xmath397 . we claim that @xmath265 is up - perspective to @xmath322 .
we can assume @xmath398 , because otherwise they would be equal , we would have @xmath399 , and the two pairs would be the same .
hence , from @xmath400 , @xmath398 and @xmath401 , we obtain @xmath402 and @xmath403 . since @xmath404 ,
the old pair @xmath265 is up - perspective to the old pair @xmath322 .
hence , @xmath405 .
applying for @xmath110 , we obtain @xmath406 which implies if @xmath267 is odd .
second , let @xmath267 be even . that is , we assume that @xmath265 is weakly down - perspective into @xmath266 .
the dual of the previous argument shows that @xmath407 and @xmath408 .
however , @xmath409 or @xmath410 does not make any difference , and @xmath411 settles for @xmath267 even .
like in case [ old2newcase ] , it suffices to deal with an odd @xmath267 , because an even @xmath267 could be treated dually . since @xmath265 is weakly up - perspective into @xmath266 and @xmath226 is the only old element above @xmath412 , we obtain @xmath413 . we obtain as before .
taking also into account , we obtain that @xmath414 and @xmath386 is one of @xmath415 and @xmath416 .
no matter which one , an argument dual to the one used in case [ old2newcase ] yields @xmath417 and @xmath418 .
hence , @xmath323 is weakly up - perspective into @xmath266 , and we obtain @xmath419 which implies @xmath420 and follows again .
now that we have proved , observe that for @xmath421 and transitivity yield @xmath422 .
hence , lemma [ chainlemma ] implies @xmath423 .
therefore , @xmath178 satisfies , and holds for @xmath178 . next , to prove that @xmath178 satisfies ,
assume that @xmath424 such that @xmath425 and @xmath426 is a spanning @xmath67-quadruple .
we want to show that it is a strong @xmath67-quadruple of @xmath180 .
the treatment for is almost the dual of that for , whence we give the details only for .
since the role of @xmath427 and @xmath428 is symmetric , it suffices to deal with the case @xmath429 ; we want to show @xmath430 . since @xmath425 implies @xmath431 , @xmath0 is a @xmath432-sublattice of @xmath180 , and holds for @xmath110 , we obtain @xmath430 for old elements , that is , for all @xmath72 such that @xmath429 .
hence , we assume that @xmath216 is a new element , that is , @xmath206 .
since @xmath433 is an old element and @xmath434 , we obtain @xmath435 .
hence , @xmath436 .
if we had @xmath437 , then @xmath438 and the description of @xmath204 would imply @xmath439 , which would be a contradiction since holds in @xmath110 .
consequently , @xmath440 .
thus , we have @xmath441 , and we know from @xmath442 and @xmath443 that @xmath444 and @xmath445 .
we also know @xmath446 since @xmath98 .
if we had @xmath447 , then the description of @xmath204 would yield @xmath448 , which would contradict .
hence , @xmath449 , and gives @xmath450 .
therefore , since the spanning @xmath67-quadruple @xmath451 is strong in @xmath110 by and @xmath452 , we conclude @xmath453 , which implies the desired @xmath430 .
consequently , @xmath178 satisfies .
this completes the proof of lemma [ mainlemma ] .
for an ordered set @xmath11 and a subset @xmath454 of @xmath5 , the restriction of the ordering of @xmath5 to @xmath454 will be denoted by @xmath455 .
if each element of @xmath5 has an upper bound in @xmath454 , then @xmath454 is a _ cofinal subset _ of @xmath5 .
the following lemma belongs to the folklore ; having no reference at hand , we will outline its easy proof .
the top elements of these principal ideals form a cofinal chain @xmath457 in @xmath5 .
let @xmath458 and @xmath459 is a well - ordered set@xmath460 . for @xmath461 , let @xmath462 mean that @xmath6 is an order ideal of @xmath463 .
zorn s lemma yields a maximal member @xmath454 in @xmath464 .
clearly , @xmath454 is well - ordered and it is a cofinal subset .
now , we combine the vertical action of lemma [ lupstp ] and the horizontal action of lemma [ mainlemma ] into a single statement . note that the order ideal @xmath2 of @xmath465 in the following lemma is necessarily a directed ordered set .
[ combinlemma ] assume that @xmath118 is an auxiliary structure such that @xmath466 is an order ideal of a bounded ordered set @xmath465 .
@xmath9 in particular , @xmath111 is an ordering and @xmath467.@xmath10 then there exists an auxiliary structure @xmath468 such that @xmath469 is a substructure of @xmath470 .
furthermore , if @xmath110 and @xmath471 are countable , then so is @xmath470 .
we can assume @xmath472 since otherwise @xmath473 would do .
consider the set @xmath474 since every set can be well - ordered , we can also write @xmath475 , where @xmath476 is an ordinal number . in @xmath477 , we define @xmath478 for @xmath479 .
it is an ordering on @xmath471 , because @xmath480 implies that it is antisymmetric .
note that @xmath481 and @xmath482 . for each @xmath483
, we want to define an auxiliary structure @xmath484 such that , for all @xmath485 , the following properties be satisfied : @xmath486 modulo the requirement that @xmath487 should be an auxiliary structure , the equivalence mentioned in is a consequence of .
we define @xmath487 by ( transfinite ) induction as follows .
we define @xmath488 by a vertical extension .
let @xmath489 , let @xmath490 , and let @xmath491 be the auxiliary structure what we obtain from @xmath110 according to lemma [ lupstp ] .
note that , for all @xmath492 , @xmath173 since @xmath493
. hence , by lemma [ lupstp ] , holds for @xmath494 .
assume that @xmath495 is a successor ordinal , that is , @xmath496 , and @xmath497 is already defined and satisfies , , and . since @xmath498 and @xmath499 , we have either @xmath500 , or @xmath501 .
these two possibilities need separate treatments .
first , if @xmath500 , then @xmath502 and we let @xmath503 .
second , let @xmath501 .
we define @xmath487 from @xmath504 by a horizontal extension as follows . with the notation @xmath505
, we obtain from that @xmath506 .
furthermore , the validity of for @xmath504 yields that @xmath507 is a spanning @xmath67-quadruple of @xmath504 . thus , letting @xmath507 and @xmath504 play the role of @xmath508 and @xmath110 in and , respectively , we define @xmath509 as the auxiliary structure @xmath178 taken from lemma [ mainlemma ] . since @xmath510 is a @xmath432-sublattice of @xmath511 , spanning @xmath67-quadruples of @xmath510 are also spanning in @xmath511 .
furthermore , it follows from @xmath512 that @xmath513 .
hence , we conclude that is inherited by @xmath509 from @xmath514 . since all the unions defining @xmath487 are directed unions , @xmath511 is a lattice , and @xmath516 is a quasiordered set . actually , it is an ordered set since @xmath480 . by the same reason ,
@xmath517 , @xmath518 , and @xmath519 are maps .
it is straightforward to check that all of , , hold for @xmath487 ; we only do this for , that is , we verify and , and also for .
assume @xmath520 .
since the unions are directed , there exists an @xmath521 such that @xmath522 , and we have @xmath523 . using that the auxiliary structure @xmath504 satisfies ,
we obtain @xmath524 , that is , @xmath525 . using lemma [ llajtorja ]
, we conclude @xmath526 in the usual way .
this implies @xmath527 .
therefore , @xmath487 satisfies .
similarly , if @xmath527 , then lemma [ llajtorja ] easily implies the existence of an @xmath521 such that @xmath525 and @xmath524 ; for @xmath504 yields @xmath523 ; and we conclude @xmath520 .
hence , @xmath487 satisfies and .
next , for the sake of contradiction , suppose that fails in @xmath487 .
this implies that @xmath528 belongs to @xmath529 , where the join is taken in the congruence lattice of @xmath511 .
since principal congruences are compact , there exists a finite subset @xmath530 such that @xmath528 belongs to @xmath531 .
thus , there exists a finite chain @xmath532 such that , for @xmath533 , @xmath534 .
each of these memberships are witnessed by finitely many `` witness '' elements according to ; see lemma [ llajtorja ]
. taking all these memberships into account , there are only finitely many witness elements all together .
hence , there exists an @xmath521 such that @xmath510 contains all these elements . applying lemma [ llajtorja ] in the converse direction
, we obtain that @xmath535 belongs to @xmath536 , which is a contradiction since @xmath504 satisfies .
consequently , @xmath487 is an auxiliary structure .
clearly , @xmath487 satisfies and since so do the @xmath504 for @xmath521 . if @xmath492 and @xmath537 , then @xmath538 for some ( actually , for every ) @xmath521 .
hence , the satisfaction of for @xmath487 follows the same way as in the successor step since @xmath510 is a @xmath432-sublattice of @xmath511 .
we have seen that @xmath539 is an auxiliary structure for all @xmath540 . letting @xmath495 equal @xmath476
, we obtain the existence part of the lemma . the last sentence of the lemma follows from the construction and basic cardinal arithmetics . in order to prove part of the theorem , assume that @xmath541 is an ordered set with zero and it is the union of a chain of principal ideals . by lemma [ cofinalitylemma ] ,
there exist an ordinal number @xmath476 and a cofinal chain @xmath542 in @xmath5 such that @xmath543 and , for @xmath544 we have @xmath545 .
the cofinality of @xmath454 means that @xmath5 is the union of the principal ideals @xmath546 , @xmath547 .
we let @xmath548 and @xmath549 , where @xmath550 denotes the restriction @xmath551 .
clearly , @xmath552 and @xmath553 , that is , @xmath554 .
note that @xmath555 is not a principal ideal in general since @xmath5 need not be bounded .
assume that @xmath496 is a successor ordinal .
we apply lemma [ combinlemma ] to obtain @xmath487 from @xmath504 .
this is possible since @xmath560 is an order ideal of @xmath561 .
note that lemma [ combinlemma ] does not assert the uniqueness of @xmath470 , and , in principle , it could be a problem later that @xmath487 is not uniquely defined . however , this is not a real problem since we can easily solve it as follows .
let @xmath562 be the smallest _ infinite _ ordinal number such that @xmath563 , let @xmath564 , and let @xmath565 be the smallest ordinal with @xmath566 .
note that @xmath567 is at least the power of continuum but @xmath568 can be finite .
let @xmath569 such that @xmath570 for @xmath571 . also , take a set @xmath572 such that @xmath573 for @xmath574 .
the point is that , after selecting the well - ordered cofinal chain @xmath454 above , we can use the well - ordered index sets @xmath575 and @xmath576 to make every part of our compound construction unique .
namely , when we well - order @xmath457 , defined in , we use the lexicographic ordering of the index set @xmath577 .
when we define lattices , their base sets will be initial subsets of @xmath578 ; a subset @xmath6 of @xmath578 is _ initial _ if , for all @xmath579 , @xmath580 implies @xmath581 . if we have to add new lattice elements , like a new top or @xmath582 , etc .
, then we always add the first one of @xmath578 that has not been used previously .
cardinality arithmetics shows that @xmath578 is never exhausted .
this way , we have made the definition of @xmath487 unique .
if @xmath495 is a limit ordinal , then first we form the union @xmath587 note that @xmath588 . the same way as in the proof of lemma [ combinlemma ] , we obtain that @xmath589 is an auxiliary structure ; the only difference is that now trivially holds in @xmath487 since @xmath590 does not have a largest element . to see this ,
suppose for contradiction that @xmath591 is the largest element of @xmath590 .
then @xmath592 for some @xmath521 . since @xmath495 is a limit ordinal , @xmath593 .
hence @xmath594 , which contradicts @xmath595 .
clearly , @xmath596 is an order ideal in @xmath516 .
thus , applying lemma [ combinlemma ] to this situation , we obtain an auxiliary structure @xmath470 , and we let @xmath597 . obviously , for all @xmath521 , @xmath504 is a substructure of @xmath487 . in order to prove part ,
assume that @xmath0 is a countable lattice .
obviously , we have @xmath600 , and we already mentioned that @xmath1 is always a directed ordered set with 0 , no matter what the size @xmath601 of @xmath0 is .
conversely , let @xmath5 be a directed ordered set with 0 such that @xmath602 .
then there is an ordinal @xmath603 ( where @xmath604 denotes the least infinite ordinal ) such that @xmath605 .
note that @xmath606 is a subset of the set of nonnegative integer numbers . for @xmath607
, there exists a smallest @xmath608 such that @xmath609 and @xmath610 ; we let @xmath611 .
this defines a binary operation on @xmath5 ; it need not be a semilattice operation .
let @xmath612 .
for @xmath613 , let @xmath614 .
a trivial induction shows that @xmath615 is an upper bound of @xmath616 , for all @xmath617 , and @xmath618 for all @xmath613 .
hence , the principal ideals @xmath619 form a chain @xmath620 , and @xmath5 is the union of these principal ideals
. therefore , part of the theorem yields a lattice @xmath0 such that @xmath5 is isomorphic to @xmath1 .
since the @xmath619 are countable and there are countably many of them , and since all the lemmas we used in the proof of part of the theorem preserve the property `` countable '' , @xmath0 is countable . | for a lattice @xmath0 , let @xmath1 denote the ordered set of principal congruences of @xmath0 . in a pioneering paper , g. grtzer characterized the ordered sets @xmath1 of finite lattices @xmath0 ; here we do the same for countable lattices .
he also showed that each bounded ordered set @xmath2 is isomorphic to @xmath1 of a bounded lattice @xmath0 .
we prove a related statement : if an ordered set @xmath2 with least element is the union of a chain of principal ideals , then @xmath2 is isomorphic to @xmath1 of some lattice @xmath0 . |
as vividly described in many popular science writings @xcite , a distant observer ( * _ o _ * ) sees a body falling towards a bh moving slower and slower , becoming darker and darker , and is eventually frozen near the event horizon of the bh , due to extremely strong time - dilation effect caused by the spacetime singularity at the event horizon . following several textbooks @xcite ,
let s first show why the community and general readers have acquired this view .
consider a test particle free - falls towards a schwarzschild bh ( non - spinning and non - charged ) of mass @xmath0 from the last circular stable orbit of the bh , @xmath1 ( here we take the natural unit , i.e. , @xmath2 ) .
will cause the particle plunge inwards , and for radius greater than @xmath3 a distant observer and an in - falling ( with the test particle ) observer see almost the same phenomena , if the particles free - falls from a larger radius . ]
the event horizon of the bh is located at @xmath4 . for an in - falling ( with the test particle ) observer ( * _ o _ * ) ,
the in - falling velocity measured is given by @xmath5 , where @xmath6 is the proper time measured by * _
o _ * and @xmath7 is the energy per unit mass of the test particle .
it is clear that the test particle is seen to pass through the event horizon with high speed .
however for * _ o _ * the in - falling velocity of the test particle is given by @xmath8 , where @xmath9 is called coordinate time , i.e. , the time measured by * _
o_*. clearly when @xmath10 , @xmath11 , i.e. , the test particle is `` frozen '' near the event horizon of the bh , as seen by * _
o_*. for this reason an astrophysical bh formed by in - falling matter has also been called a `` frozen '' star .
the situation is depicted in figure 1 ( left ) , where `` proper velocity '' and `` coordinate velocity '' are measured by * _ o _ * and * _ o _ * respectively . introducing the `` cycloid parameter '' @xmath12 by @xmath13 , the times taken for the test particle to reach different locations from the event horizon are @xmath14 , and @xmath15 , where @xmath16 $ ] and @xmath17 , for * _ o _ * and * _ o _ * , respectively , as shown in figure 1 ( right ) .
it can be shown easily that @xmath9 increases exponentially as the test particle approaches to the event horizon , as given by @xmath18 , where @xmath19 when @xmath20 . clearly the time measured by * _ o _ * is finite when the test particle reaches the event horizon .
however for * _ o _ * it takes infinite time for the test particle to reach exactly the location of the event horizon . in other words ,
* _ o _ * will never see the test particle falling into the event horizon of the black hole .
it is therefore legitimate to ask the question whether in real astrophysical settings bhs can ever been formed and grow with time , since all real astrophysical bhs can only be formed and grown by matter collapsing into a `` singularity '' . , the `` coordinate velocity '' is exactly equal to zero , i.e.
, the test particle is `` frozen '' near the event horizon of the bh . *
right : * the times taken for a test particle falling towards a bh of mass @xmath0 , starting from @xmath21 . here
`` proper time '' ( dotted line ) and `` coordinate time '' ( solid line ) are measured by * _ o _ * and * _ o _ * respectively . the test particle is seen by * _ o _ * to pass through the event horizon within a finite time .
however the `` coordinate time '' approaches to infinity as the test particle moves asymptotically to the event horizon , i.e. , the test particle will never move across the event horizon of the bh , as seen by * _
o_*.,scaledwidth=68.0% ] , the `` coordinate velocity '' is exactly equal to zero , i.e. , the test particle is `` frozen '' near the event horizon of the bh . *
right : * the times taken for a test particle falling towards a bh of mass @xmath0 , starting from @xmath21 . here
`` proper time '' ( dotted line ) and `` coordinate time '' ( solid line ) are measured by * _ o _ * and * _ o _ * respectively .
the test particle is seen by * _ o _ * to pass through the event horizon within a finite time .
however the `` coordinate time '' approaches to infinity as the test particle moves asymptotically to the event horizon , i.e. , the test particle will never move across the event horizon of the bh , as seen by * _
o_*.,scaledwidth=70.0% ] two possible answers have been proposed so far .
the first one is that since * _ o _ * indeed has observed the test particle falling through the event horizon , then in reality ( for * _ o _ * ) matter indeed has fallen into the bh . however since * _ o _ * has no way to communicate with * _ o _ * once * _ o _ * crosses the event horizon , * _ o _ * has no way to ` know ' if the test particle has fallen into the bh .
the second answer is to invoke quantum effects .
it has been argued that some kind of quantum radiation of the accumulated matter just outside the event horizon @xcite , similar to hawking radiation , will eventually bring the matter into the bh , as seen by * _ o_*.
unfortunately it has been realized that the time scale involved for this to work is far beyond the hubble time , and thus this does not answer the question in the real world @xcite .
apparently * _ o _ * can not be satisfied with either answer . in desperation ,
* _ o _ * may take the altitude of `` who cares ? ''
when the test particle is sufficiently close to the event horizon , the redshift is so large that practically no signals from the test particle can be seen by * _ o _ * and apparently the test particle has no way of turning back , therefore the `` frozen star '' does appear `` black '' and is an infinitely deep `` hole '' . for practical purposes * _ o _ * may still call it a `` bh '' , whose total mass is also increases by the in - falling matter .
apparently this is the view taken by most people in the astrophysical community and general public , as demonstrated in many well known text books @xcite and popular science writings @xcite .
however , this view is incorrect , because in real astrophysical settings * _ o _ * ( in the schwarzschild coordinates ) does observe matter falling into the bh , as we shall show as follows .
and @xmath22 are @xmath23 and @xmath24 , respectively . * ( a ) * the shell is sufficiently close to , but still completely outside the bh s horizon , such that @xmath25 and @xmath26 .
the solution of zone * iii * is the schwarzschild metric of the mass @xmath27 , as given in equation ( 2 ) . since @xmath28 is not in zone * iii * , there is no singularity in zone * iii*. the solution of zone * ii * is given in equation ( 3 ) . since @xmath29 is not in zone * ii * ( see text for the definition of @xmath30 ) , there is also no singularity in zone * ii*. the solution of zone * i * is given in equation ( 4 ) .
however , in this case , @xmath31 is a singularity .
therefore the only singularity in this system is the location of the horizon of the pre - existing bh . *
( b ) * part of the shell is inside @xmath31 . in this case ,
zone * i * does no longer exist .
the solutions for zone * iii * and zone * ii * take the same forms . there is still no singularity in zone * iii * , as expected . since @xmath32 , in zone * ii * there is a singularity at @xmath33 , which is the apparent horizon @xcite of the new bh with a mass of @xmath30 , as a consequence of the growth of the bh .
it is important to note that in both cases the location of @xmath34 is only an apparent horizon @xcite , but neither a singularity nor the event horizon before the shell is inside @xmath28.,scaledwidth=86.0% ] and @xmath22 are @xmath23 and @xmath24 , respectively . *
( a ) * the shell is sufficiently close to , but still completely outside the bh s horizon , such that @xmath25 and @xmath26 .
the solution of zone * iii * is the schwarzschild metric of the mass @xmath27 , as given in equation ( 2 ) . since @xmath28 is not in zone
* iii * , there is no singularity in zone * iii*. the solution of zone * ii * is given in equation ( 3 ) . since @xmath29 is not in zone
* ii * ( see text for the definition of @xmath30 ) , there is also no singularity in zone * ii*. the solution of zone * i * is given in equation ( 4 ) . however , in this case , @xmath31 is a singularity .
therefore the only singularity in this system is the location of the horizon of the pre - existing bh . *
( b ) * part of the shell is inside @xmath31 . in this case ,
zone * i * does no longer exist . the solutions for zone * iii * and zone * ii * take the same forms .
there is still no singularity in zone * iii * , as expected . since @xmath32 , in zone
* ii * there is a singularity at @xmath33 , which is the apparent horizon @xcite of the new bh with a mass of @xmath30 , as a consequence of the growth of the bh .
it is important to note that in both cases the location of @xmath34 is only an apparent horizon @xcite , but neither a singularity nor the event horizon before the shell is inside @xmath28.,scaledwidth=86.0% ] for a static and pressureless spherical mass shell around a bh as shown in fig . 2 ( a ) , we assume the metric is of the form , @xmath35 for zone * iii * , according to the birkhoff theorem @xcite the solution is the schwarzschild metric for a mass @xmath36 , where @xmath23 and @xmath24 refer to the masses of the central bh and the shell respectively
. therefore , @xmath37 where @xmath38 is the schwarzschild radius of @xmath39 .
therefore in zone * iii * ( @xmath40 ) , there is no singularity , even in the schwarzschild coordinates . for zone *
ii * and zone * i * , the solutions can be found by applying the continuity conditions for the metric at the inner and outer boundaries of the shell .
for zone * ii * , @xmath41 where the density of matter inside the shell is assumed to have the form @xmath42 ( so that the mass between any two radii inside the shell is given by @xmath43 ) , @xmath44 ( @xmath45 ) , @xmath46 , and @xmath47 . when @xmath48 , @xmath49
. therefore in zone * ii * ( @xmath50 ) , there is also no singularity , even in the schwarzschild coordinates .
note that this metric is quite different from the schwarzschild metric , meaning that the birkhoff theorem can not be applied to zone * ii*. for zone * i * , @xmath51 where @xmath52 is a singularity .
note that this metric is also different from the schwarzschild metric , because the metric becomes time - dependent when the shell starts moving .
this means that the birkhoff theorem also can not be applied to zone * i * directly .
it can be shown easily that in the limit of a thin shell , i.e. , @xmath53 , @xmath54 , where @xmath55 is the radius of the thin shell .
making a transformation @xmath56 in equation ( 1 ) , the metric for zone * i * becomes , @xmath57 which is exactly the schwarzschild metric , but for a slower clock ( since @xmath58 ) compared to the solution of the same bh without the enclosing shell ; once the shell starts moving , the clock must be re - calibrated .
this can be considered as the generalized birkhoff theorem .
however in most astrophysical settings , either @xmath59 and @xmath60 , or @xmath61 ; in either case , @xmath62 , thus the original birkhoff theorem is still sufficiently accurate .
therefore for the case when the shell is still completely outside the pre - existing bh , the only singularity ( when @xmath63 ) in the schwarzschild coordinates is at @xmath64 , just like as if the shell did not exist .
however , the metric everywhere ( exterior to , inside and interior to the shell ) is influenced by the existence of the shell .
once @xmath65 as depicted in fig
. 2 ( b ) , zone * i * does no longer exists .
however in zone * ii * , @xmath66 and @xmath33 is the location of the only singularity ( when @xmath63 ) . in this case
a temporary bh is formed with mass @xmath30 and an apparent horizon @xcite located at @xmath33 .
once @xmath67 , i.e. , the shell is completely within the temporary bh , the final state of a new bh is reached with mass @xmath27 and an event horizon located at @xmath28 .
we note that our solutions are consistent to the graphical illustration ( but without mathematical description ) of the space - time diagram for a spherical shell around a pre - existing bh given in ref .
@xcite . due to the page limit of the proceedings , further details of the above solutions and their possible applications to other astrophysical problems will be presented elsewhere @xcite .
now we turn to the question if the time measured by * _ o _ * for the shell falling into a bh is finite or infinite .
first we need to clarify which bh we are talking about , since a temporary bh will be formed during the process of a shell falling to a pre - existing bh .
apparently the final bh with mass of @xmath68 should be what we are concerned with .
then the final event horizon is at @xmath69 .
the important insight we have learnt from the exact solutions for the shell around a pre - existing bh is that when @xmath70 , the location of @xmath71 is only the location of an apparent horizon @xcite , neither a singularity nor an event horizon . only _ after _
@xmath22 is smaller than @xmath28 , @xmath71 will become the event horizon of the final bh and the static clock at @xmath71 will be stopped .
assuming there is some matter very far outside the shell , i.e. , the system is not placed in a complete vacuum , the static clock at @xmath71 will never be stopped _ before _
the shell is completely inside @xmath71 ; even the mere existence of * _ o _ * can satisfy this condition .
therefore within a finite time , * _ o _ * will certainly see the shell crossing @xmath71 , which would be the final event horizon of the final bh .
next we give some numerical examples . in the case
that the mass of the shell ( @xmath72 ) is much smaller than the mass of the pre - existing bh ( @xmath0 ) , the in - falling motion of the shell should be very close to that of a test particle making a free - fall towards a bh .
the only difference is that the mass of the bh will be increased once the shell reaches the horizon of the pre - existing bh .
this means that a test particle will be frozen to the horizon of the pre - existing bh , but the shell will not be frozen at the apparent horizon @xcite at @xmath71 when the shell crosses the apparent horizon ( which is not a real horizon , since in zone * iii * there is no singularity ) . in figure
( 1 ) ( right ) , the location of the apparent horizon @xcite for a shell with @xmath73 is marked .
therefore the coordinate time interval for a test particle crossing the apparent horizon can be found easily from this plot .
although several times longer than the proper time interval , the coordinate time interval ( called `` in - falling time '' hereafter ) , is still `` finite '' for all practical physical settings as shown in the following . given @xmath74 ( such that @xmath73 ) , the in - falling time is about 2 millisecond or 2.3 days for a 10 solar masses bh or a supermassive bh of @xmath75 solar masses , completely negligible in the cosmic history . even if @xmath76 is decreased by a factor of @xmath77 , the in - falling time
is only doubled .
it is interesting to note that the above in - falling time scales are easily accessible to current astronomy instruments ; perhaps the processes of matter falling into bhs can be indeed recorded by modern telescopes .
let s now address briefly on the question if it is practical to see a spaceship free - falling into a bh starting from the last stable circular orbit , such as that in the center of the milky way .
assuming the mass of the spaceship is @xmath78 kg , thus @xmath79 , and consequently @xmath80 minutes as seen from the earth ; the corresponding time for the clock on the spaceship is about 3 minutes .
the duration of this event seems to be just adequate for the instruments on the spaceship to conduct some experiments during the in - fall and report the results back to the earth .
we have obtained the exact ( albeit static ) solutions for a spherically symmetric and pressureless mass shell around a pre - existing bh .
the metric inside or interior to the shell is different from the schwarzschild metric of the enclosed mass , meaning that the well - known birkhoff theorem can only be applied to the exterior of a spherically symmetric mass .
consequently the interior metric is influenced by the outside mass , different from the gauss theorem for newton s inverse - square law of gravity .
this is contrary to some ( mis)understanding of the birkhoff theorem , e.g. , in ref .
@xcite . for the region between the shell and the pre - existing bh
, the metric can be transformed to the schwarzschild metric for a slower clock ; we call this generalized birkhoff theorem .
however in most astrophysical settings the original birkhoff theorem is still sufficiently accurate .
another important insight from the solutions is that there does not exist a singularity nor event horizon in the shell , i.e. , the shell can not be frozen at the event horizon of the final bh with the total mass of the pre - existing bh and the shell
. therefore * _ o _ * can indeed observe matter falling into a bh .
we also show that for all practical astrophysical settings the in - falling time recorded by * _ o _ * is sufficiently short that future astrophysical instruments may be able to follow the whole process of matter falling into bhs .
however , more precise calculations of the in - falling time of a shell crossing the final event horizon of the final bh can only be done with the exact _ dynamical _ solution of a shell _ falling _ towards a bh .
nevertheless much physical insights have been provided by our solutions to the static shell around a bh .
since a single `` frozen star '' appears in - different from a single `` bh '' , as far as observational consequences are concerned , one may nevertheless still ask the question : why bother ? one interesting thought experiment is to imagine what would happen if two `` frozen stars '' or two bhs merge together .
it has been proposed that electromagnetic waves will be produced when two `` frozen stars '' merge together @xcite , if the in - falling matter is ordinary matter rather than dark matter , but no electromagnetic waves should be produced when two `` real '' bare bhs ( no matter falling onto the bhs ) merge together
. however gravitational waves will be produced in both cases .
thus our conclusion is testable by future observations , for example , future x - ray and gravitational wave telescopes observing the same merging events in centers of distant galaxies .
snz makes the following acknowledgments . the basic idea for this paper
was sparked during a simulating discussion in may 2007 with my former student ms .
sumin tang , who is now studying in astronomy department , harvard university ; she also derived some formula in the first section and wrote the code producing figure 1 .
richard lieu of university of alabama in huntsville is acknowledged for making valuable comments to the first draft of the manuscript made in june 2007 and many follow - up discussions on this problem .
i appreciate the open - mind altitude of the organizers of this conference , mostly drs .
dong lai , yefei yuan and xiang - dong li , for allowing me to make this presentation , which stimulated many heated debates and discussions .
many participants of this conference , especially drs .
kinwah wu , kazuo makishima , neil gehrels , masaruare shibata and ramesh narayan are thanked for interesting discussions .
i thank partial funding supports by the ministry of education of china , directional research project of the chinese academy of sciences under project no .
kjcx2-yw - t03 , and by the national natural science foundation of china ( nsfc ) under project no.10521001 , 10733010 & 10725313 . | it has been well known that in the point of view of a distant observer , all in - falling matter to a black hole ( bh ) will be eventually stalled and `` frozen '' just outside the event horizon of the bh , although an in - falling observer will see the matter falling straight through the event horizon .
thus in this `` frozen star '' scenario , as distant observers , we could never observe matter falling into a bh , neither could we see any `` real '' bh other than primordial ones , since all other bhs are believed to be formed by matter falling towards singularity . here
we first obtain the exact solution for a pressureless mass shell around a pre - existing bh .
the metrics inside and interior to the shell are all different from the schwarzschild metric of the enclosed mass , meaning that the well - known birkhoff theorem can only be applied to the exterior of a spherically symmetric mass .
the metric interior to the shell can be transformed to the schwarzschild metric for a slower clock which is dependent of the location and mass of the shell ; we call this generalized birkhoff theorem .
another result is that there does not exist a singularity nor event horizon in the shell .
therefore the `` frozen star '' scenario is incorrect .
we also show that for all practical astrophysical settings the in - falling time recorded by an external observer is sufficiently short that future astrophysical instruments may be able to follow the whole process of matter falling into bhs .
the distant observer could not distinguish between a `` real '' bh and a `` frozen star '' , until two such objects merge together .
it has been proposed that electromagnetic waves will be produced when two `` frozen stars '' merge together , but not true when two `` real '' bare bhs merge together .
however gravitational waves will be produced in both cases .
thus our solution is testable by future high sensitivity astronomical observations .
address = physics department and center for astrophysics , tsinghua university , beijing 100084 , china address = physics department and center for astrophysics , tsinghua university , beijing 100084 , china |
the o i triplet lines at 77715 @xmath3 ( 3s @xmath14s@xmath153p @xmath14p , multiplet 1 ) are known to play important roles in stellar spectroscopy , since they are sufficiently strong to be visible in stars of wide temperature range ( spectral types from k to b ) and provide us with possibilities of studying not only oxygen abundances but also line - broadening parameters ( such as turbulences or projected rotational velocities ) . making use of this merit
, our group has been exploited these lines for investigating the behavior of microturbulent velocities in a
f stars ( takeda 1992 ) as well as the o abundances of a f dwarfs ( takeda & sadakane 1997 ; takeda et al .
2008 ) , g k giants ( takeda , kawanomoto , & sadakane 1998 ; takeda et al .
2015 ) , and m 67 stars from the turn - off point through the red - giant branch ( takeda & honda 2015 ) .
this time we pay our attention to b - type stars , while expecting a usefulness of applying this o i triplet also to this high - temperature regime .
studying surface oxygen abundances in b - type stars ( which have comparatively high @xmath10 and @xmath16 ) is especially significant in connection with rapid rotators , since a deep mixing caused by the meridional circulation may salvage on - cycled material from the interior to produce more or less an underabundance of o at the surface , similarly to an overabundance of he due to mixing of h@xmath17he processed product . since rapidly - rotating b - type stars showing he - enhancement
are actually observed ( e.g. , lyubimkov et al . 2004 ) , o - deficiency may be another indicator for the existence of efficient rotation - induced mixing , and thus observational information on the oxygen abundances of a large number of b - type stars with various rotational velocities can make an important touchstone for the prediction from the stellar evolution theory ( e.g. , georgy et al .
2013 ) .
on the observational side , oxygen abundances of b - type stars have so far been studied mainly by using o ii lines in the blue region ( see , e.g. , gies & lambert 1992 ; kilian 1992 ; cunha & lambert 1994 ; korotin , andrievsky , & luck 1999 ; simn - daz 2010 ; nieva & przybilla 2012 ; lyubimkov et al .
2013 ) , although these ionized oxygen lines are useful because they are numerous in number , their application is exclusively limited to early b stars , since the strengths of these lines quickly fall off below @xmath18 k. this is a shortcoming from a viewpoint of continuation / comparison with lower temperature stars ( a , f , and g types ) , for which neutral o i lines have to be used . besides , since these previous studies using o ii lines are mostly directed to sharp - lined b stars , prospect for applying them to rapid rotators is not clear . meanwhile , spectroscopic investigations on the o abundances of b - type stars using neutral oxygen lines have been less common and tend to be biased to either late - b stars or slow rotators .
takeda et al .
( 2010 ; hereinafter referred to as paper i ) carried out a non - lte analysis on the o ( and ne ) abundances of sharp - lined b stars ( @xmath19 030 km s@xmath2 and 10000 k @xmath20 k ) by using o i 61568 lines ( multiplet 10 ) , which however are inapplicable to rapid rotators because of their weakness .
niemczura , morel , and aerts s ( 2009 ) lte study on the abundances of oxygen ( along with other elements ) based on weaker o i lines ) , since they stated that o i 77715 lines were excluded because of the strong non - lte effect . ] is limited to b6b9.5 stars .
similarly , only late - b ( b5b9 ) stars were targeted in hempel and holweger s ( 2003 ) determination of non - lte oxygen abundances based on o i 7771 - 5 lines . among the previous abundance studies of b - type stars , most notable in connection with our present interest
would be the recent one of nieva and przybilla ( 2012 ) , who determined the abundance of o ( along with those of he , c , n , ne , mg , si , and fe ) for 20 sharp - lined mid - to - early b stars ( 16000 k @xmath21 33000 k ) by using the state - of - the - art non - lte technique based on high - s / n and high - resolution spectra . while their analysis mainly depends on a number of o ii lines , they also employed o i lines ( e , g .
, o i features at 61568 @xmath3 , 77715 @xmath3 , and 8446 @xmath3 ) for stars with @xmath22 k , in order to check that o i / o ii ionization equilibrium is accomplished .
although their work is regarded as an important benchmark , only apparently slow rotators ( @xmath23 km s@xmath2 ) were treated and rapid rotators were out of their consideration .
this situation motivated us to investigate a possibility of applying o i 77715 lines to b stars with much wider range stellar properties ; i.e. , @xmath0 1000030000 k corresponding to late b through early b type , slow as well as rapid rotators covering @xmath24 200300 km @xmath2 .
+ is this o i triplet feature still usable for early b - type stars ( @xmath0 2000030000 k ) of various rotational velocities , despite that its strength should be considerably weakened as most o atoms are ionized at such high @xmath10 range ?
+ how is the behavior of its strength in terms of the stellar parameters ; e.g. , @xmath10 , surface gravity ( @xmath11 ) , microturbulence ( @xmath25 ) , and oxygen abundance ? is it suitable for abundance determination ? + how much important is the non - lte effect regarding this strong o i triplet ? does the required non - lte abundance correction depend on the atmospheric parameters ?
toward answering these questions , we first conducted non - lte calculations on an extensive parameter grid in order to understand how the o i 77715 strengths of b stars behave in terms of various key parameters , and then carried out an observational study of non - lte oxygen abundance determinations for 34 selected late - b through early - b type stars with diversified @xmath26 values based on the spectra newly obtained at nishi - harima astronomical observatory .
the purpose of this paper is to report the consequence of this investigation .
the remainder of this article is organized as follows : the observational data of our program stars are described in section 2 , and their spectroscopic parameters are derived in section 3 .
we explain our non - lte calculations in section 4 , and the procedures of abundance determination ( including profile fitting , equivalent width evaluation , and error estimation ) in section 5 .
the resulting behavior of non - lte correction and the trend of oxygen abundances of b - type stars are discussed in section 6 , followed by section 7 where the conclusions are summarized .
the list of our 34 targets is presented in table 1 , all of which are apparently bright ( @xmath27 mag ) and selected from the extensive @xmath26 compilation of 1092 northern b - type stars published by abt , levato , and grosso ( 2002 ) , where b supergiants and be stars were excluded from our selection .
our sample includes stars of diversified @xmath26 values ( from slow rotators to rapid rotators up to @xmath28 km s@xmath2 ) .
these program stars are plotted on the @xmath29 vs. @xmath30 diagram ( theoretical hr diagram ) in figure 1 , where girardi et al.s ( 2000 ) theoretical evolutionary tracks corresponding to different stellar masses are also depicted .
we can see from this figure that the masses of our sample stars are in the range between @xmath31 and @xmath32 .
spectroscopic observations of these stars were carried out in the summer season of 2015 ( may through september ; observation date for each star is given in table 1 ) by using the medium and low - dispersion longslit spectrograph ( malls ; cf .
ozaki & tokimasa 2005 ) installed on the nasmyth platform of the 2 m nayuta telescope at nishi - harima astronomical observatory ( nhao ) .
equipped with a 2k@xmath332 k ccd detector ( 13.5 @xmath34 m pixel ) , malls can record a spectrum covering @xmath12 400 @xmath3 ( 76008000 @xmath3 ) in the medium - resolution mode with the resolving power of @xmath35 . while the exposure time of one frame was typically a few minutes , several spectral frames were co - added to improve the signal - to - noise ratio .
the reduction of the spectra ( bias subtraction , flat - fielding , spectrum extraction , wavelength calibration , co - adding of frames to improve s / n , continuum normalization ) was performed by using the `` noao.onedspec '' package of the software iraf in a standard manner .
the s / n ratios of the resulting spectra were estimated to be around @xmath12 200300 in most cases ( except for hr 0779 , for which s /
n is only @xmath12 6070 ) , which turned out to be sufficient for our purposes .
our final spectra in the 77007850 @xmath3 region for each of the 34 stars are shown in figure 2 .
the effective temperature and the surface gravity of each program star were determined from the colors of strmgren s @xmath36 photometric system with the help of napiwotzki , scnberner , and wenske s ( 1993 ) uvbybetanew program , http://www.astro.le.ac.uk/~rn38/uvbybeta.html@xmath37 . ] where the observational data of @xmath38 , @xmath39 , @xmath40 , and @xmath41 were taken from hauck and mermilliod ( 1998 ) via the simbad database .
the resulting @xmath10 and @xmath11 are summarized in table 1 . the model atmosphere for each star
was then constructed by two - dimensionally interpolating kurucz s ( 1993 ) atlas9 model grid in terms of @xmath10 and @xmath11 , where we exclusively applied the solar - metallicity models . as to typical errors of these parameters
, we may estimate @xmath42 in @xmath10 and @xmath43 dex in @xmath11 for most of our sample stars ( i.e. , 32 stars at 10000 k @xmath21 24000 k ) according to napiwotzki et al .
( 1993 ; cf . section 5 therein ) .
we should note , however , that parameters derived from napiwotzki et al.s code would not be reliable any more for stars at @xmath44 k , as pointed out by nieva ( 2013 ) based on the results of nieva and przybilla ( 2012 ) who used non - lte ionization equilibria of various species along with the profiles of balmer lines to determine the atmospheric parameters .
accordingly , the parameters of the relevant two stars ( hr 8238 and hr 8797 ; cf .
table 1 ) having particularily high @xmath10 ( 25231 k and 27853 k , respectively ) should be viewed with caution and their errors may be larger than the typical amounts mentioned above . given that hr 779 ( hd 16582 ) and hr 8238 ( hd 205021 ) in our sample were also studied by nieva and przybilla ( 2012 ) , it is interesting to compare our color - based @xmath10 and @xmath11 with their spectroscopically determined values which may be more reliable .
the results of ( @xmath45 , @xmath46 ) are ( 21747/21250 , 3.63/3.80 ) for hr 779 and ( 25231/27000 , 3,63/3.95 ) for hr 8238 .
we can see a reasonable consistency within the expected errors for the former b2 star ( @xmath47 k ) , while the discrepancy is larger for the latter b0.5 star ( @xmath44 k ) which may be attributed to the reason mentioned above ( i.e. , outside of the reliability limit of napiwotzki et al.s calibration ) .
regarding the microturbulence , which is necessary for abundance determination , we had to assign an appropriate value since it was not possible to establish this parameter based on our spectra . here
, we divide our targets into two classes of @xmath48 k and @xmath49 k , and tentatively assume @xmath50 km s@xmath2 for the former and @xmath51 km s@xmath2 for the latter , based on the discussion in takeda et al .
( 2014 ; cf . section 3 therein ) and paper i ( cf . section 3 therein ) , respectively . admittedly , this is a rather rough choice ; but it appears to be reasonable as will be discussed in subsection 6.2 .
in order to evaluate the non - lte effect on the strength of o i 77715 triplet , we carried out non - lte calculations for oxygen on an extensive grid of solar - metallicity model atmospheres resulting from combinations of eleven @xmath10 values ( 9000 , 10000 , 12000 , 14000 , 16000 , 18000 , 20000 , 22000 , 24000 , 26000 , 28000 k ) and four @xmath11 values ( 3.0 , 3.5 , 4.0 , 4.5 ) , models at @xmath52 k presumably because of an instability problem , models of only three gravities ( @xmath11 = 3.5 , 4.0 , and 4.5 ) could be used for @xmath10 = 28000 k. ] which sufficiently cover the parameter ranges of our program stars . the statistical equilibrium calculations for oxygen were done in the same manner as in paper i ( cf .
subsection 4.1 therein ) .
our non - lte calculation program is based on the code described in takeda ( 1991 ) .
the atomic model of oxygen is the same as that adopted by takeda ( 2003 ; cf .
subsection 2.1 therein ) , which is based on the atomic data given in kurucz and bell ( 1995 ) .
other various fixed collisional and radiative rates were evaluated as described in takeda ( 1992 ) .
the non - lte departure coefficients [ @xmath53 applied to each star ( to be used for non - lte abundance determination in section 5 ) were then derived by two - dimensionally interpolating this grid in terms of @xmath10 and @xmath11 , as was done for model atmospheres . toward understanding the behavior of the non - lte effect to be discussed in subsection 6.1 , theoretical
non - lte and lte equivalent widths ( @xmath54 and @xmath55 ) of the whole triplet were computed on this grid of models for combinations of four oxygen abundances ( [ o / h ] [ @xmath56(o ) @xmath57(o ) ] is the differential logarithmic oxygen abundance relative to the sun , where @xmath58 is defined in the usual normalization of @xmath58(h ) = 12.00 and we employed @xmath59(o ) = 8.93 in order to maintain consistency with the similar non - lte grid computed by takeda ( 2003 ) for fgk stars .
note that this reference solar abundance ( 8.93 ) , which was taken from anders and grevesse ( 1989 ) and adopted as the standard value in the atlas9 program , is appreciably higher by asplund et al.s ( 2009 ) recently updated value ( 8.69 ) based on their 3d hydrodynamical solar model photosphere .
however , it is closer to the solar non - lte oxygen abundance ( 8.81 ) derived by takeda and honda ( 2005 ) with the same o i 77715 lines ( see the discussion in subsection 5.1 of paper i ) . ]
= @xmath60 , @xmath61 , 0.0 , and + 0.3 ) and three microturbulences ( @xmath25 = 1 , 3 , and 5 km s@xmath2 ) by using kurucz s ( 1993 ) width9 program ) and the nlte line source function ratio ( @xmath62 ) resulting from our non - lte calculation code , respectively ( cf .
subsection 4.2 in takeda 1991 for more details ) . besides
, since the original width9 program can handle only a single isolated line , a spectrum synthesis technique was incorporated for calculating the total equivalent width of a multi - component feature ( such as o i 77715 ) . ] with the atomic data presented in table 2 , where @xmath63 was evaluated by directly integrating the synthesized profiles of triplet lines over the wavelength region between 7770.44 @xmath3 and 7776.94 @xmath3 ( i.e. , 1.5 @xmath3 margin both shortward of the 7771.94 line and longward of the 7775.39 line ) .
we also evaluated the corresponding non - lte abundance corrections ( @xmath64 ) defined as @xmath65 , where @xmath66 and @xmath67 are the abundances derived by inversely analyzing @xmath54 in non - lte and lte , respectively .
the resulting @xmath63 and @xmath64 values are summarized in the electronic data tables ( file name : ncor7773_xi?.dat ) available as on - line materials .
the behaviors of the computed @xmath63 and @xmath68 in terms of the atmospheric parameters are shown in figure 3 for selected representative cases , on which we will focus later in subsection 6.1 .
as in paper i , our strategy of determining the abundance and related quantities ( e.g. , non - lte correction , uncertainties due to ambiguities of atmospheric parameters ) consists of the following consecutive steps : ( i ) derivation of provisional abundance solution by applying the spectrum - fitting technique , ( ii ) evaluation of the equivalent width ( @xmath63 ) by using this fitting - based abundance , and ( iii ) analysis of such derived @xmath63 with the standard parameters and changing the conditions ( perturbed parameters , lte and non - lte ) .
we first searched for the solutions for the oxygen abundance ( @xmath58 ) , projected rotational velocity ( @xmath26 ) , and radial velocity ( @xmath69 ) such as those accomplishing the best fit ( minimizing @xmath70 residuals ) between the theoretical and observed spectrum in the @xmath12 77657780 @xmath3 region , while applying the automatic fitting algorithm ( takeda 1995 ) . in this preparatory step ,
we used the model atmosphere ( cf .
section 3 ) and the non - lte departure coefficients ( section 4 ) prepared for each star but adopted the same microturbulence of @xmath71 km s@xmath2 for computing the non - lte theoretical spectrum .
regarding the macroscopic line - broadening function to be convolved with the intrinsic spectrum , we included the instrumental broadening ( assumed to be the gaussian function ) corresponding to @xmath72 and the rotational broadening with the limb - darkening coefficient of @xmath73 ( estimated from fig .
17.6 of gray 2005 ) how the theoretical spectrum for the converged solutions fits well with the observed spectrum is demonstrated in figure 4 .
residuals , we sometimes masked some regions showing features irrelevant to stellar spectra , such as spurious spectrum defect , as highlighted in green in this figure . ]
the resulting @xmath26 values are presented in table 1 , which are also compared with those given in abt et al.s ( 2002 ) catalogue in figure 5 .
next , we inversely computed the equivalent width ( @xmath63 ) for the whole o i 77715 triplet ( as done in section 4 ) by using the abundance solution ( derived from the fitting analysis ) along with the same model atmosphere / parameters .
based on such evaluated @xmath63 value , the non - lte abundance ( @xmath66 ) as well as lte abundance ( @xmath67 ) were freshly computed by applying the microturbulence assigned to each star ( 1 or 3 km s@xmath2 depending on @xmath10 ; cf .
section 3 ) , from which the non - lte correction ( @xmath64 ) was further derived .
these @xmath63 , @xmath66 , and @xmath64 values are also given in table 1 .
we also estimated the errors in @xmath66 caused by ambiguities involved in the standard atmospheric parameters ( presented in table 1 ) . since their typical uncertainties are @xmath74 in @xmath10 , @xmath75 dex in @xmath11 , and @xmath76 km s@xmath2
( for @xmath48 k ) or @xmath77 km s@xmath2 ( for @xmath49 k ) in @xmath25 ( cf . section 3 ) . six kinds of abundance variations ( @xmath78 , @xmath79 , @xmath80 , @xmath81 , @xmath82 , and @xmath83 ) were derived by repeating the analysis on the @xmath63 values while perturbing the standard atmospheric parameters interchangeably by these amounts . finally , the root - sum - square of these perturbations , @xmath84 , was regarded as the abundance uncertainty ( due to combined errors in @xmath10 , @xmath11 , and @xmath25 ) , where @xmath85 , @xmath86 , and @xmath87 are defined as @xmath88 , @xmath89 , and @xmath90 , respectively .
the resulting abundances ( @xmath66 ) , equivalent widths ( @xmath63 ) , non - lte corrections ( @xmath64 ) , and abundance variations ( @xmath68 s ) in response to parameter changes , are graphically displayed as functions of @xmath10 in figure 6 , which we will discuss in subsection 6.2 .
before discussing the results of our analysis derived in section 5 for the program stars , we first review how the strength of o i 77715 triplet and its non - lte effect behaves in terms of the atmospheric parameters .
the following characteristics can be summarized by inspection of figure 3 : + generally , @xmath63 declines with an increase in @xmath10 .
while its change is gradual at @xmath7 k , a rather abrupt drop of @xmath63 occurs at @xmath9 k where oxygen is predominantly ionized and very few neutral oxygen atoms remain .
+ naturally , @xmath63 increases with an increase in [ o / h ] and @xmath25 , though it is almost @xmath25-independent at @xmath9 k where the line is weak in the linear part of the curve of growth . + the behavior of @xmath63 in response to changing @xmath11 is somewhat complicated ( figure 3c ) .
while the line strength slightly grows with a lowering of @xmath11 at @xmath7 k ( because of the increased non - lte effect ; note that this tendency is seen only in @xmath54 but not in @xmath55 ) , the trend is reversed at @xmath9 k where @xmath63 markedly weakens with a decrease in @xmath11 because the effect of enhanced ionization becomes more important at the condition of lower - density .
+ the non - lte effect always acts in the direction of intensifying the line strength ( @xmath91 ) , which makes the non - lte abundance correction ( @xmath64 ) always negative . +
the extent of @xmath92 has a peak at @xmath0 1500020000 k and tends to be larger for higher [ o / h ] , smaller @xmath25 , and lower @xmath11 , which can be interpreted that the non - lte effect becomes more significant when the line - forming depth is higher or in the lower density region .
we now discuss figure 6 .
the equivalent widths ( @xmath63 ) derived in subsection 5.2 show a decreasing trend with an increase in @xmath10 ( figure 6a ) , which is reasonably consistent with the theoretically predicted trend for @xmath54 ( cf .
figure 3a c ) .
the @xmath92 values are in the range of @xmath12 0.61.7 dex ( figure 6b ) , which are so large that these corrections are indispensable for reliable determination of oxygen abundances of b - type stars based on the o i 77715 lines . comparing these results with those of hempel and holweger ( 2003 ) , who determined the non - lte oxygen abundances of b5b9 stars
( @xmath0 900017000 k ) based on the same o i triplet , we see that their @xmath92 values ( @xmath12 0.31.2 ; cf .
their table 5 ) are considerably smaller than ours .
we suspect that they underestimated the importance of the non - lte effect , which may explain the significantly supersolar [ o / h]@xmath93 results ( by @xmath12 + 0.30.6 dex ) they obtained for many of their sample stars .
figure 6c suggests that the oxygen abundances derived for the 34 b - type stars distribute around @xmath12 8.78.8 without any systematic @xmath10-dependence .
actually the average abundance over all 34 stars makes @xmath94 ( with standard deviation @xmath95 ) .
if three stars with @xmath58(o ) @xmath96 ( hr 6779 , 7236 , 8238 ) showing appreciable deviations are excluded , we have @xmath97 ( @xmath98 ) .
this means that the surface oxygen abundances of these b - type stars are mostly near to the solar composition ( see footnote 5 ) , which is in fairly good agreement with the previous studies for b - type stars ; e.g. , @xmath99 derived in paper i based on o i 61568 lines or @xmath100 by lyubimkov et al .
( 2013 ) based on o ii lines ( see also table 6 and table 7 therein ) .
regarding the abundance errors caused by uncertainties in atmospheric parameters , @xmath87 ( @xmath12 0.10.2 dex ) is predominantly important over @xmath85 ( @xmath101 ) or @xmath86 ( @xmath101 ) at @xmath7 k , to which most of the program stars are relevant ( cf .
figure 6d f ) .
accordingly , the extents of @xmath102 ( error bars attached to the symbols in figure 6c ) are essentially determined by @xmath87 and on the order of @xmath12 0.1 - 0.2 dex ( except for the two stars of @xmath9 k ) , which is consistent with the standard deviation of our abundance results .
it may appropriate here to comment on the assigned values of @xmath25 mentioned in section 3 . since the abundances derived for given microturbulences of @xmath103 km s@xmath2 ( @xmath48 k ) and @xmath104 km s@xmath2 ( @xmath49 k ) do not show any systematic trend in terms of @xmath10
, this may imply that our choice is reasonably justified .
if a constant @xmath25 were assigned for all stars , we would have obtained a spurious @xmath10-dependence .
km s@xmath2 ( cf .
subsection 5.1 ) showed a tendency of appreciable underabundance in late b - type stars . ]
the oxygen abundances of 34 stars are plotted against @xmath26 as well as @xmath29 in figure 7 , where we can not see any systematic dependence .
this means that no appreciable o abundance anomaly is detected in our sample stars at least within the precision of @xmath105 0.10.2 dex . according to georgy et
al.s ( 2013 ) calculation , while no change in the surface composition is produced at all during the main sequence period for the case of non - rotating stars , rapidly rotating stars may show some abundance peculiarities caused by the rotational mixing of nuclear processed product . for example , regarding the cases of solar - metallicity ( @xmath106 ) models , the changes of he@xmath107o abundances at the end of the main sequence for 4 @xmath108 star is @xmath109 ( @xmath110 = 0.5 ) and @xmath111 ( @xmath110 = 0.95 ) , while those for 9 @xmath108 star is @xmath112 ( @xmath110 = 0.5 ) and @xmath113 ( @xmath110 = 0.95 ) , where @xmath114 is the angular rotational velocity corresponding to the critical break - up .
however , even if such an abundance change as large as @xmath12 20% exists , it corresponds to only @xmath115 dex in the logarithmic scale .
thus , it is no wonder that we could not detect such a subtle amount , given that the precision of our abundance determination is @xmath6 dex . in this sense
, we may state that our result is more or less consistent with the recent stellar evolution calculations based on the canonical mixing theory . finally , we turn to the question which motivated this investigation ( cf .
section 1 ) : is the o i 77715 triplet practically useful for oxygen abundance determination of b - type stars ?
our answer is as follows : + yes , it is surely useful as far as stars of @xmath7 k are concerned ( given that the non - lte effect is properly taken into account ) , since its strength is still sufficiently large ( @xmath116 m@xmath3 ) and abundances are not seriously affected by uncertainties in the atmospheric parameters .
+ for example , if we consider a case of @xmath117 km s@xmath2 and @xmath118 m@xmath3 , the fwhm and the central depth of the merged triplet feature would be @xmath119 @xmath3 and @xmath120% , respectively , which is measurable without difficulty on a spectrum with s / n ratios of a few hundred .
+ in contrast , regarding very early b stars of @xmath9 k ( corresponding to @xmath12 b0 ) , the strength of this triplet quickly drops away and its sensitivity to ambiguities of @xmath10 or @xmath11 becomes seriously large , which makes it unsuitable as a reliable abundance indicator .
+ we should keep in mind , however , even for the well - behaved case of @xmath7 k stars , the precision of abundance derivation would be on the order of @xmath6 dex , which is mainly restricted by uncertainties in the choice of microturbulence .
accordingly , very high - precision should not be expected .
+ in this sense , detection of surface oxygen abundance anomaly based on these o i lines would be feasible only if the extent of peculiarity is sufficiently large ( e.g. , by @xmath121 dex ) , such as the case where non - canonical deep mixing is relevant . the slight abundance change ( up to @xmath115 dex ) predicted by the current canonical rotational mixing theory
can hardly be detected .
in order to examine whether the o i triplet lines at 77715 @xmath3 can be used as a reliable oxygen abundance indicator for b - type stars of high @xmath10 including rapid rotators , we carried out oxygen abundance determinations for selected 34 b - type stars based on the spectra ( @xmath72 ) obtained by the 2 m nayuta telescope at nishi - harima astronomical observatory .
regarding the atmospheric parameters , @xmath10 and @xmath11 were determined from @xmath36 colors and the microturbulence was assumed to be @xmath103 km s@xmath2 ( @xmath48 k ) and @xmath104 km s@xmath2 ( @xmath49 k ) by consulting previous studies .
the non - lte effect was properly taken into consideration based on our non - lte calculations carried out for an extensive grid of models .
we first determined the provisional abundance and @xmath26 by applying the synthetic spectrum - fitting analysis to the 77715 feature , and then inversely determined ( from the tentative abundance ) the equivalent width of the whole triplet to be used for further evaluations of non - lte / lte abundances , non - lte correction , and errors due to ambiguities of atmospheric parameters .
it was found that the extent of the ( negative ) non - lte correction is appreciably large to be @xmath122 0.61.7 dex and its consideration is indispensable .
the resulting non - lte oxygen abundances turned out to be nearly solar ( @xmath123 8.78.8 with @xmath124 dex ) without any clear dependence upon rotation as well as luminosity ( or mass ) , which is consistent with the results of published observational studies and stellar evolution calculations with the canonical mixing theory .
we concluded that this triplet is a useful o abundance indicator ( with a precision of @xmath6 dex ) up to @xmath7 k , since its total equivalent width is sufficiently large ( @xmath8 m@xmath3 ) .
in contrast , it is not adequate for abundance determination for stars at @xmath9 k , where its strength rapidly turns down and its sensitivity to @xmath10 or @xmath11 becomes considerably large .
abt , a. a. , levato , h. , & grosso , m. 2002 , apj , 573 , 359 anders , e. , & grevesse , n. 1989 , geochim .
acta , 53 , 197 arenou , f. , grenon , m. , & gmez , a. 1992 , a&a , 258 , 104 asplund , m. , grevesse , n. , sauval , a. j. , & scott , p. 2009
, ara&a , 47 , 481 cunha , k. , & lambert , d. l. 1994 , apj , 426 , 170 esa 1997 , the hipparcos and tycho catalogues , esa sp-1200 , available from nasa - adc or cds in a machine - readable form ( file name : hip_main.dat ) flower , p. j. 1996 , apj , 469 , 355 girardi , l. , bressan , a. , bertelli , g. , & chiosi , c. 2000 , a&as , 141 , 371 georgy , c. , ekstrm , s. , granada , a. , meynet , g. , mowlavi , n. , eggenberger , p. , & maeder , a. 2013 , a&a , 553 , a24 gies , d. r. , & lambert , d. l. 1992 , apj , 387 , 673 gray , d. f. 2005 , the observation and analysis of stellar photospheres , 3rd ed .
( cambridge : cambridge university press ) hauck , b. , & mermilliod , m. 1998 , a&as , 129 , 431 hempel , m. , & holweger , h. 2003 , a&a , 408 , 1065 kilian , j. 1992 , a&a , 262 , 171 korotin , s. a. , andrievsky , s. m. , & luck , r. e. 1999 , a&a , 351 , 168 kurucz , r. l. 1993 , kurucz cd - rom , no . 13 ( harvard - smithsonian center for astrophysics ) kurucz , r. l. , & bell , b. 1995 , kurucz cd - rom , no .
23 ( harvard - smithsonian center for astrophysics ) lyubimkov , l. s. , lambert , d. l. , poklad , d. b. , rachkovskaya , t. m. , & rostopchin , s. i. 2013 , mnras , 428 , 3497 lyubimkov , l. s. , rostopchin , s. i. , & lambert , d. l. 2004 , mnras , 351 , 745 napiwotzki , r. , schnberner , d. , & wenske , v. 1993 , a&a , 268 , 653 niemczura , e. , morel , t. , aerts , c. 2009 , a&a , 506 , 213 nieva , m .- f .
2013 , a&a , 550 , a26 nieva , m .- f .
, & przybilla , n. 2012 , a&a , 539 , a143 ozaki , s. , & tokimasa , n. 2005 , annu . rep .
nishi - harima astron .
15 , 15 ( in japanese ) simn - daz , s. 2010 , a&a , 510 , a22 takeda , y. 1991 , a&a , 242 , 455 takeda , y. 1992 , pasj , 44 , 309 takeda , y. 1995 , pasj , 47 , 287 takeda , y. 2003 , a&a , 402 , 343 takeda , y. , han , i. , kang , d .-
, lee , b .- c .
, & kim , k .- m .
2008 , jkas , 41 , 83 takeda , y. , & honda , s. 2005 , pasj , 57 , 65 takeda , y. , & honda , s. 2015 , pasj , 67 , 25 takeda , y. , kambe , e. , sadakane , k. , & masuda , s. 2010 .
pasj , 62 , 1239 ( paper
i ) takeda , y. , kawanomoto , s. , & ohishi , n. 2014 , pasj , 66 , 23 takeda , y. , kawanomoto , s. , & sadakane , k. 1998 , pasj , 50 , 97 takeda , y. , & sadakane , k. 1997 , pasj , 49 , 367 takeda , y. , sato , b. , omiya , m. , & harakawa , h. 2015 , pasj , 67 , 24 ccccc rcrc rccchr # & hd # & desig . &
sp.type & @xmath125 & @xmath10 & @xmath11 & @xmath126 & @xmath25 & @xmath63 & @xmath66 & @xmath64 & obs .
hr 7437 & hd 184606 & 9 vul & b8 iiin & 5.00 & 12703 & 3.52 & 239.1 & 1 & 515 & 8.80 & @xmath1271.40 & 150731 + hr 7106 & hd 174638 & @xmath41 lyr & b8.5 ib - ii & 3.52 & 17124 & 3.92 & 218.4 & 3 & 422 & 8.83 & @xmath1271.14 & 150612 + hr 0708 & hd 015130 & @xmath128 cet & a0 v & 4.88 & 9764 & 3.48 & 204.5 & 1 & 541 & 8.59 & @xmath1271.10 & 150926 + hr 6779 & hd 166014 & @xmath129 her & b9.5 iii & 3.84 & 9657 & 3.24 & 172.4 & 1 & 475 & 8.33 & @xmath1271.05 & 150501 + hr 7739 & hd 192685 & @xmath130 & b3 v & 4.79 & 17855 & 3.94 & 170.2 & 3 & 317 & 8.63 & @xmath1270.88 & 150815 + hr 7121 & hd 175191 & @xmath131 sgr & b2 v & 2.05 & 19012 & 4.27 & 166.4 & 3 & 283 & 8.68 & @xmath1270.74 & 150731 + hr 8634 & hd 214923 & @xmath132 peg & b8 v & 3.41 & 11182 & 3.65 & 148.3 & 1 & 528 & 8.70 & @xmath1271.21 & 150927 + hr 7236 & hd 177756 & @xmath133 aql & b9 vn & 3.43 & 11534 & 4.14 & 142.5 & 1 & 350 & 8.30 & @xmath1270.72 & 150731 + hr 7528 & hd 186882 & @xmath68 cyg & b9.5 iv+ .
& 2.86 & 10098 & 3.67 & 139.0 & 1 & 518 & 8.59 & @xmath1271.05 & 150926 + hr 8418 & hd 209819 & @xmath134 aqr & b8 v & 4.29 & 12138 & 4.26 & 130.8 & 1 & 437 & 8.65 & @xmath1270.93 & 150927 + hr 8781 & hd 218045 & @xmath135 peg & b9 iii & 2.49 & 9643 & 3.52 & 129.0 & 1 & 618 & 8.82 & @xmath1271.12 & 150927 + hr 7906 & hd 196867 & @xmath135 del & b9 iv & 3.77 & 10794 & 3.89 & 119.5 & 1 & 497 & 8.62 & @xmath1271.03 & 150926 + hr 7963 & hd 198183 & @xmath133 cyg & b5
v & 4.53 & 13982 & 3.89 & 110.5 & 1 & 389 & 8.63 & @xmath1271.11 & 150927 + hr 7306 & hd 180554 & 1 vul & b4 iv & 4.76 & 15474 & 3.78 & 94.2 & 3 & 421 & 8.68 & @xmath1271.10 & 150731 + hr 8522 & hd 212097 & 32 peg & b9 iii & 4.78 & 11119 & 3.44 & 77.8 & 1 & 599 & 8.86 & @xmath1271.36 & 150927 + hr 8301 & hd 206672 & @xmath136 cyg & b3 iii & 4.69 & 17191 & 3.55 & 71.4 & 3 & 405 & 8.73 & @xmath1271.22 & 150927 + hr 7178 & hd 176437 & @xmath137 lyr & b9 iii & 3.25 & 9638 & 2.73 & 71.0 & 1 & 670 & 8.84 & @xmath1271.43 & 150612 + hr 7447 & hd 184930 & @xmath134 aql & b5 iii & 4.36 & 13400 & 3.69 & 67.3 & 1 & 497 & 8.84 & @xmath1271.37 & 150731 + hr 8579 & hd 213420 & 6 lac & b2 iv & 4.52 & 19885 & 3.67 & 64.3 & 3 & 343 & 8.79 & @xmath1271.08 & 150927 + hr 7592 & hd 188260 & 13 vul & b9.5 iii & 4.57 & 10425 & 3.89 & 57.6 & 1 & 585 & 8.86 & @xmath1271.07 & 150815 + hr 7372 & hd 182568 & 2 cyg & b3 iv & 4.99 & 18937 & 4.28 & 56.2 & 3 & 272 & 8.65 & @xmath1270.71 & 150801 + hr 0718 & hd 015318 & @xmath138 cet & b9 iii & 4.30 & 10438 & 4.09 & 55.6 & 1 & 498 & 8.65 & @xmath1270.91 & 150926 + hr 7852 & hd 195810 & @xmath139 del & b6 iii & 4.03 & 13611 & 3.68 & 51.4 & 1 & 475 & 8.80 & @xmath1271.35 & 150815 + hr 8335 & hd 207330 & @xmath140 cyg & b3 iii & 4.23 & 17953 & 3.23 & 46.1 & 3 & 509 & 8.94 & @xmath1271.64 & 150927 + hr 6787 & hd 166182 & 102 her & b2 iv & 4.37 & 20048 & 3.45 & 41.8 & 3 & 347 & 8.77 & @xmath1271.15 & 150501 + hr 7298 & hd 180163 & @xmath141 lyr & b2.5 iv & 4.43 & 15974 & 3.19 & 41.0 & 3 & 535 & 8.85 & @xmath1271.61 & 150801 + hr 7039 & hd 173300 & @xmath142 sgr & b8 iii & 3.17 & 12263 & 3.66 & 39.3 & 1 & 567 & 8.92 & @xmath1271.37 & 150815 + hr 7613 & hd 188892 & 22 cyg & b5 iv & 4.95 & 14008 & 3.38 & 33.0 & 1 & 487 & 8.80 & @xmath1271.53 & 150815 + hr 7710 & hd 191692 & @xmath143 aql & b9.5 iii & 3.24 & 10410 & 3.71 & 32.9 & 1 & 585 & 8.82 & @xmath1271.14 & 150815 + hr 8797 & hd 218376 & 1 cas & b0.5 iii & 4.84 & 27853 & 3.64 & 32.5 & 3 & 42 & 8.86 & @xmath1270.63 & 150927 + hr 8238 & hd 205021 & @xmath41 cep & b0.5 iiis & 3.23 & 25231 & 3.63 & 30.3 & 3 & 75 & 8.36 & @xmath1270.60 & 150927 + hr 7426 & hd 184171 & 8 cyg & b3 iv & 4.74 & 15858 & 3.54 & 27.8 & 3 & 434 & 8.69 & @xmath1271.23 & 150801 + hr 7773 & hd 193432 & @xmath144 cap & b9 iv & 4.77 & 10180 & 3.91 & 23.2 & 1 & 595 & 8.88 & @xmath1271.03 & 150801 + hr 0779 & hd 016582 & @xmath68 cet & b2 iv & 4.08 & 21747 & 3.63 & 15.0 & 3 & 281 & 8.74 & @xmath1270.93 & 150926 + note .
+ in columns 1 through 8 are given the hr number , hd number , star designation , spectral type , apparent visual magnitude ( in mag ) , effective temperature ( in k ) , logarithmic surface gravity ( in cm s@xmath145 ) , and projected rotational velocity ( in km s@xmath2 ; derived from our fitting analysis ) .
columns 9 through 12 show the results of abundance analysis : assigned microturbulence ( cf .
section 3 ) , total equivalent width of the o i 77715 lines ( in m@xmath3 ) , non - lte o abundance ( in dex ; expressed in the usual normalization of h=12.00 ) , and non - lte correction ( in dex ) .
column 13 gives the observation date ( for example , 150907 means 2015 september 7 ) .
the objects are arranged in the descending order of @xmath126 . note .
+ the first four columns are self - explanatory . presented in columns 57
are the damping parameters : gammar is the radiation damping width ( s@xmath2 ) [ @xmath146 , gammas is the stark damping width ( s@xmath2 ) per electron density ( @xmath147 ) at @xmath148 k [ @xmath149 , and gammaw is the van der waals damping width ( s@xmath2 ) per hydrogen density ( @xmath147 ) at @xmath148 k [ @xmath150 .
+ these data were taken from the compilation of kurucz and bell ( 1995 ) as long as available , while the parenthesized damping parameters are the default values computed by the width9 program . | oxygen abundances of 34 b - type stars in the effective temperature range of @xmath0 1000028000 k with diversified rotational velocities ( @xmath1 0250 km s@xmath2 ) were determined from the o i triplet lines at 77715 @xmath3 , with an aim to examine whether this o i feature can be a reliable abundance indicator for such high - temperature stars including rapid rotators .
it revealed that the required non - lte abundance correction is distinctly large ( ranging from @xmath4 dex to @xmath5 dex ) and its consideration is indispensable . on the condition that the non - lte effect is taken into account ,
this triplet is a useful o abundance indicator ( with a precision of @xmath6 dex ) up to @xmath7 k , since its total equivalent width is sufficiently large ( @xmath8 m@xmath3 ) .
in contrast , it is not adequate for abundance derivation for stars at @xmath9 k , where its strength rapidly drops down toward a hardly detectable level ( except for sharp - lined stars ) and its sensitivity to @xmath10 or @xmath11 becomes considerably large .
the resulting non - lte oxygen abundances turned out to be almost normal ( i.e. , near - solar around @xmath12 8.78.8 within @xmath13 0.2 dex ) for most stars without any dependence upon projected rotational velocity as well as luminosity ( or mass ) , which is consistent with the prediction of recent stellar evolution calculations . |
a thermodynamic system may undergo a phase transition when temperature and/or pressure vary .
for the strong interacting qcd system , it is known that the hadronic matter will change to partonic phase at very high temperature and density @xcite .
the data from ultra - relativistic nuclear - nuclear collisions at rhic has shown that the partonic phase of qcd matter
quark - gluon plasma ( qgp ) has been formed @xcite .
one of the main goals of current ultra - relativistic heavy - ion collision experiments is to map the qcd phase diagram @xcite .
the critical point at the qcd phase boundary is particular interested , as the large fluctuations are expected .
the sensitive probes of qcd critical point are the high cumulants of the conserved charges , i.e. , net baryon number , net electric charge , and net strangeness @xcite .
the net baryon number fluctuations have been studied using lattice qcd simulations and qcd effective models @xcite . however , due to the difficulties of the lattice calculations and model estimations , the high cumulants of the net baryon number are still not final @xcite .
the universality of the critical fluctuations allows us to study the relevant cumulants in a simple system , where the high cumulants can be precisely obtained .
it is known that the qcd critical point , which terminates the first order phase transition line , is in the same universality class as the 3-dimensional ising model @xcite .
so the high cumulants of order parameter and energy in the vicinity of the critical point of the 3-dimensional ising model are called for .
they should be a good reference for various relevant calculations . in the chiral limit , the chiral phase transition for 2-flavor qcd belongs to the same universality class as the 3-dimensional @xmath3 spin model @xcite .
the singular behavior of the net baryon number fluctuations is expected to be governed by the universal @xmath3 symmetry group @xcite .
owing to the staggered fermions in lattice calculations , the ( 2 + 1)-flavor chiral phase transition may fall into the 3-dimensional @xmath2 universality class @xcite . in this paper
, we present the high cumulants of the order parameter and energy in the vicinity of the critical temperature from the 3-dimensional ising model , and energy from the 3-dimensional @xmath2 and @xmath3 spin models , respectively .
the simulation is performed in a finite - size system .
the obtained results should be instructive to a finite system formed in ultra - relativistic heavy - ion collisions .
the paper is organized as follows , firstly , the cumulants of the order parameter and energy from the @xmath4 spin models are derived in section ii .
then , their relations to the net baryon number fluctuations in the qcd phase transitions are discussed . in section iii ,
the high cumulants from the 3-dimensional ising , @xmath2 and @xmath3 spin models are presented and discussed .
finally , the generic features of the cumulants from the three models are summarized in section iv .
the @xmath4-invariant nonlinear @xmath5-models ( @xmath4 spin models ) are defined as , @xmath6 where @xmath7 is the hamiltonian , @xmath8 is an interaction energy between nearest - neighbor spins @xmath9 , and @xmath10 is the external magnetic field . @xmath8 and @xmath10 are both reduced quantities which already contain a factor @xmath11 .
@xmath12 is a unit vector of @xmath13-components at site @xmath14 of a @xmath15-dimensional hyper - cubic lattice .
it is usually decomposed into the longitudinal ( parallel to the magnetic field @xmath10 ) and the transverse component @xmath16 where @xmath17 . for the 3-dimensional ising , @xmath2 , and @xmath3 spin models , @xmath15 = 3 , and @xmath13 = 1 , 2 and 4 , respectively . the ( reduced ) free energy per unit volume is @xmath18 where @xmath19 is the partition function .
@xmath20 is the energy of a spin configuration , @xmath21 is the lattice average of the longitudinal spin components , @xmath22 is the volume , and @xmath23 is the number of lattice points of each direction . as we known , the cumulants of the order parameter are the derivatives of the free energy density ( eq . ) with respect to @xmath24 .
we can get the cumulants of the order parameter from the generating function @xcite , @xmath25 so the first , second , third , fourth and sixth order cumulants of the order parameter are as follows , @xmath26 where @xmath27 , and @xmath28 is the magnetization ( order parameter ) of the system . at vanishing external magnetic field , due to
the spatial rotation symmetry of the @xmath4 groups , such defined order parameter is zero . in the case ,
an approximated order parameter definition is suggested as , @xmath29 @xcite .
on the other hand , the cumulants of the energy are the derivatives of the free energy density with respect to the temperature @xmath30 .
the generating function is @xmath31 so the first , second , third , fourth and sixth order cumulants of the energy are as follows , @xmath32 where @xmath33 . in the vicinity of the critical point , the free energy density ( eq.([free energy density ] ) )
can be decomposed into two parts , the regular and singular parts .
the critical related fluctuations are determined by the singular part .
it has the scaling form @xmath34 here @xmath35 and @xmath36 are reduced temperature and magnetic field , @xmath37 and @xmath38 are the normalized parameters .
@xmath39 is the critical temperature . @xmath40 and @xmath41 are universal critical exponents . in our simulation
, we set @xmath42 and choose the approximate critical temperatures , @xmath43 @xcite , @xmath44 @xcite , and @xmath45 @xcite for the 3-dimensional ising , @xmath2 and @xmath3 spin models , respectively . in order to map the result of the 3-dimensional ising model to that of the qcd ,
the following linear ansatz is suggested @xcite , @xmath46 @xmath47 and @xmath48 are the temperature and chemical potential at the qcd critical point , respectively . @xmath49 and @xmath50 are two parameters determined by qcd .
the baryon - baryon correlation length diverges with the exponent @xmath40 and exponent @xmath41 when the critical point is approached along the @xmath51-direction and @xmath52-direction , respectively @xcite .
the cumulants of the net baryon number are the derivatives of the qcd free energy density with respect to @xmath53 . in the vicinity of the critical point , it is the combination of the derivatives with respect to @xmath51 and @xmath52 in the 3-dimensional ising model .
since @xmath41 is larger than @xmath40 @xcite , the critical behavior of the net baryon number fluctuations is mainly controlled by the derivatives with respect to @xmath52 , i.e. , the fluctuations of the order parameter in the 3-dimensional ising model .
the singular part of the free energy density for the chiral phase transition is suggested as @xcite @xmath54 where @xmath55 and @xmath56 are the universal critical exponents of the 3-dimensional @xmath3 spin model .
@xmath57 is the scaling function .
the reduced temperature @xmath51 and external field @xmath52 are expressed as follows @xmath58 here @xmath59 is the critical temperature in the chiral limit .
@xmath60 is a parameter determined by qcd @xcite .
the net baryon number susceptibility is the derivative of free energy density with respective to the chemical potential @xmath61 . from eqs . and
, we can get the derivatives of the free energy density with respect to @xmath30 , to chemical potential @xmath62 at @xmath63 , and to @xmath64 respectively , @xmath65 @xmath66 @xmath67 where @xmath68 is even in eq . .
comparing eqs . and with eq .
, we can see that the scaling form of derivatives of the free energy density with respect to @xmath30 and @xmath61 are equivalent .
so the net baryon number fluctuations are related to the derivatives of the free energy density with respect to @xmath30 in the 3-dimensional @xmath3 spin model .
the @xmath68-th order cumulant of the energy from the 3-dimensional @xmath3 spin model is relevant to the @xmath69-th ( or @xmath68-th ) order cumulant of the net baryon number at @xmath70 ( or @xmath71 ) in the chiral phase transition .
the cumulants of the order parameter from the 3-dimensional @xmath2 and @xmath3 spin models are the derivatives of the free energy density with respect to the external field .
their critical behavior has been presented and discussed in ref .
the monte carlo simulations of the 3-dimensional ising , @xmath2 , and @xmath3 spin models are performed by the wolff algorithm with helical boundary conditions @xcite .
we start simulation without magnetic field , and then modify the algorithm to include a magnetic field @xmath72 and @xmath73 @xcite .
as we know , the valid region of system size of the finite - size scaling varies with magnetic field and observable .
the typical size of an observable at a given magnetic field is determined by the saturation of size dependence , as shown in ref .
the typical system sizes for each kind of cumulants at a given magnetic field and model are listed in table i. .[comparison]the typical system size for each kind of cumulant at a given external field and model [ cols="<,<,<,<,<",options="header " , ] in order to compare the basic structure of the cumulants with and without external fields , each cumulant of the energy at vanishing and non - vanishing external fields is plotted in an identical sub - figure , and rescaled to unity by its maximum or minimum ( except for the first order cumulant of the energy from @xmath2 and @xmath3 spin models ) , as shown in fig . 1 , fig . 4 , and fig . 5 . for the ising model ,
the cumulants of the order parameter at vanishing and non - vanishing external fields are quite different , which may be caused by the definitions as discussed in section ii below eq .
( [ cumulants of order parameter ] ) .
so their cumulants at non - vanishing and vanishing external fields are presented in fig . 2 and fig . 3 , respectively .
( color online ) .
@xmath74 ( a ) , @xmath75 ( b ) , @xmath76 ( c ) and @xmath77 ( d ) at @xmath78 from the 3-dimensional ising model.,scaledwidth=45.0% ] the cumulants of the energy , i.e. , @xmath74 , @xmath75 , @xmath76 and @xmath77 , from the 3-dimensional ising model at @xmath78 are shown in fig .
1(a ) to 1(d ) , respectively .
we can see that the basic features of the cumulants , i.e. , the patterns of the fluctuations , are not influenced by the magnitude of the external field . with the appearing or increasing of the external field , the whole critical region
is amplified and shifted toward the high temperature side . in the vicinity of the critical temperature
, @xmath74 has a maximum peak .
@xmath75 oscillates and its sign changes from positive to negative when temperature increases and passes the critical one .
@xmath76 has two positive maximums locating at the two sides of @xmath59 , respectively .
the minimum between them is negative .
in contrast to the two positive maximums of @xmath76 , @xmath77 has two negative minimums and one positive maximum at the critical temperature region .
the sign change for the cumulants of the energy starts at the third cumulant .
it happens twice in the fourth and sixth order cumulants .
( color online ) .
@xmath79 ( a ) , @xmath80 ( b ) , @xmath81 ( c ) , @xmath82 ( d ) at @xmath83 from the 3-dimensional ising model.,scaledwidth=45.0% ] the cumulants of the order parameter at non - vanishing external fields , i.e. , @xmath79 , @xmath80 , @xmath81 and @xmath82 , from the 3-dimensional ising model are presented in fig .
2(a ) to 2(d ) , respectively .
we can see that the influences of the external field are similar to those as discussed in fig . 1 . in the vicinity of @xmath59
, @xmath79 shows the same peak structure as that for the energy .
@xmath80 has a negative valley and no sign change in the critical region .
@xmath81 shows a obvious positive maximum and a very small negative minimum when the temperature increases and passes the critical one .
@xmath82 oscillates from positive to negative , and the negative valley is more obvious than that in @xmath81 . here
the sign change starts at the fourth order cumulant in fig .
( color online ) .
@xmath79 ( a ) , @xmath80 , @xmath81 and @xmath82 ( b ) at @xmath84 from the 3-dimensional ising model.,scaledwidth=45.0% ] the high cumulants of the order parameter from the 3-dimensional ising model at vanishing external field are shown in fig .
3(a ) to 3(b ) , respectively .
@xmath79 in fig .
3(a ) shows a narrow and sharp peak at the critical temperature region . from fig .
3(b ) , we can see that , both @xmath80 and @xmath81 oscillate , but the former changes from negative to positive , while the latter changes from positive to negative with the increasing temperature .
the generic structure of @xmath82 is similar to that of @xmath76 , having two positive maximum locating at the two sides of @xmath59 and a negative minimum between them .
so the generic structure of the high cumulants at non - vanishing external field is quite different from that at vanishing external field .
the sign change in the former case appears in the fourth order cumulant , while the third one in the later case . comparing the energy fluctuations in fig . 1 with the order parameter fluctuations in fig . 2 and fig . 3
, we can see that the generic structure of the same order cumulant of the energy is different from that of the order parameter , except for the second order one .
the way of the sign change of @xmath80 at @xmath84 as shown in fig .
3(b ) is consistent with the expectation from the effective model @xcite .
the generic structure of @xmath81 at non - vanishing external field as shown in fig .
2(c ) is very similar to that from the @xmath5-model @xcite .
( color online ) .
@xmath85 ( a ) , @xmath74 ( b ) , @xmath75 ( c ) and @xmath76 ( d ) at @xmath78 from the 3-dimensional @xmath3 spin model.,scaledwidth=45.0% ] the cumulants of the energy , i.e. , @xmath85 , @xmath74 , @xmath75 and @xmath76 , from the 3-dimensional @xmath3 spin model at @xmath78 are presented in fig .
4(a ) to 4(d ) , respectively .
again , the external field shows the similar influences as discussed above .
@xmath85 increases with the temperature .
@xmath74 has a peak and is positive in the whole critical temperature region .
@xmath75 oscillates and changes from positive to negative with the increasing temperature .
@xmath76 has two maximums and a minimum between them .
the behavior of @xmath74 , @xmath75 and @xmath76 is similar to that of the energy fluctuations from the 3-dimensional ising model , cf . fig .
1(a ) to 1(c ) . as we discussed in section
ii , at vanishing baryon chemical potential , @xmath86 , @xmath87 , @xmath88 and @xmath89 in the chiral phase transition are corresponding to @xmath85 , @xmath74 , @xmath75 and @xmath76 from the 3-dimensional @xmath3 spin model .
the positive peak of @xmath74 is consistent with @xmath87 from the calculations of lattice qcd @xcite and the estimations of the pnjl @xcite and pqm models @xcite .
the sign change of @xmath75 is also observed in @xmath88 from the pqm model @xcite .
( color online ) .
@xmath85 ( a ) , @xmath74 ( b ) , @xmath75 ( c ) and @xmath76 ( d ) at @xmath78 from the 3-dimensional @xmath2 spin model.,scaledwidth=45.0% ] the cumulants of the energy , i.e. , @xmath85 , @xmath74 , @xmath75 and @xmath76 , from the 3-dimensional @xmath2 spin model at @xmath78 are presented in fig .
5(a ) to 5(d ) , respectively .
we can see that each sub - figure in fig .
5 is similar to that in fig . 4 .
each order cumulant of the energy from the 3-dimensional @xmath2 and @xmath3 spin models shows qualitatively similar temperature dependence in the vicinity of the critical temperature .
in this paper we perform the simulations of the 3-dimensional ising , @xmath2 and @xmath3 spin models at a given system size at three different external fields @xmath78 .
the critical behavior of the high cumulants of the order parameter and energy in the 3-dimensional ising model , and the cumulants of the energy in the 3-dimensional @xmath2 and @xmath3 spin models is presented , respectively .
we find that the external field does not influence the generic structure of the cumulants , except the cumulants of the order parameter from the 3-dimensional ising model .
but it widens the temperature region of the critical fluctuations . for the 3-dimensional ising model ,
the generic structure of the high cumulants of the energy are different from that of the order parameter . for the energy fluctuations ,
the sign change starts at the third order cumulant no matter with or without external field .
so does the order parameter at vanishing external field . at non - vanishing external field ,
the first sign change of the order parameter fluctuations appears at the fourth order cumulant .
the common feature is that the higher the order of the cumulant , the more complicated the fluctuation pattern is . the critical behavior of the third order cumulant of the order parameter at vanishing external field , and the fourth order cumulant of the order parameter are consistent with the expectations of the effective model and the @xmath5-model , respectively .
for the 3-dimensional @xmath2 and @xmath3 spin models , the behavior of the second to fourth order cumulants of the energy is similar to that from the 3-dimensional ising model .
the sign change also starts at the third order cumulant .
the net baryon number fluctuations based on the @xmath3 spin model are qualitatively consistent with the calculations from the lattice qcd , and expectations from the pnjl and the pqm models .
our results also show that at vanishing chemical potential , the sixth order cumulant of the net baryon number is necessary in order to observe a sign change in the chiral phase transition .
99 k. fukushima and t. hatsuda , rep .
. phys . * 74 * , 014001 ( 2011 ) .
m. gyulassy , l. mclerran , nucl . phys . *
a750 * , 30 ( 2005 ) ; j. adams _ et al . _
( star collaboration ) , nucl .
* a757 * , 102 ( 2005 ) . h. caines - for the star collaboration , arxiv:0906.0305 . y. hatta and m.a .
stephanov , phys .
lett . * 91 * , 102003 ( 2003 ) .
v. koch , arxiv:0810.2520 .
stephanov , phys .
102 * , 032301 ( 2009 ) .
m. asakawa , s. ejiri , and m. kitazawa , phys .
lett . * 103 * , 262301 ( 2009 ) .
stephanov , phys .
lett . * 107 * , 052301 ( 2011 ) .
b. friman , f. karsch , k. redlich , v. skokov , eur .
j. c * 71 * , 1694 ( 2011 ) .
et al . _ ,
d * 79 * , 074505 ( 2009 ) .
gavai , s. gupta , phys .
b * 696 * , 459 ( 2011 ) .
liu and y.l .
wu , phys .
d * 81 * , 014028 ( 2010 ) .
s. chatterjee and k.a .
mohan , arxiv:1201.3352 .
f. karsch and k. redlich , phys .
d * 84 * , 051504 ( 2011 ) .
s. borsnyi(for the wuppertal - budapest collaboration ) , arxiv:1210.6901 .
p. de forcrand and o. philipsen , phys .
lett . * 105 * , 152001 ( 2010 ) .
m. stephanov , k. rajagopal , and e. shuryak , phys .
lett . * 81 * , 4816 ( 1998 ) .
m. asakawa , j. phys .
g : nucl . part .
phys . * 36 * , 064042 ( 2009 ) .
pisarski and f. wilczek , phys .
d * 29 * , 338 ( 1984 ) .
y. hatta and t. ikeda , phys .
d * 67 * , 014028 ( 2003 ) . c. bernard _
et al . _ ,
d * 61 * , 054503 ( 2000 ) .
et al . _ ,
d * 80 * , 094505 ( 2009 ) .
o. kaczmarek _
d * 83 * , 014504 ( 2011 ) .
see for instance s.l .
lauritzen , thiele pioneer in statistics , oxford university press , 2002 .
talapov and h.w.j .
blte , j. phys . a : math
* 29 * , 5727 ( 1996 ) .
ballesteros , l.a .
fernndez , v. marn - mayor , a. muoz sudupe , phys .
b * 387 * , 125 ( 1996 ) .
rehr and n.d .
mermin , phys .
rev . a * 8 * , 472 ( 1973 ) .
n. b wilding , j. phys . : condens .
matter * 9 * , 585 ( 1997 ) . c. nonaka and m. asakawa , phys
c * 71 * , 044904 ( 2005 ) .
j. garca , j.a .
gonzalo , phys .
a * 326 * , 464 ( 2003 ) .
pan xue s talk at cpod11 , http://conf.ccnu.edu.cn/ cpod2011/. u. wolff , phys .
62 * , 361 ( 1989 ) .
i. dimitrovi , p. hasenfratz , j. nager and f. niedermayer , nucl .
b350 * , 893(1991 ) .
york , g.d .
moore , arxiv:1106.2535 . | considering different universality classes of the qcd phase transitions , we perform the monte carlo simulations of the 3-dimensional @xmath0 models at vanishing and non - vanishing external field , respectively .
interesting high cumulants of the order parameter and energy from @xmath1 ( ising ) spin model , and the cumulants of the energy from @xmath2 and @xmath3 spin models are presented .
the generic features of the cumulants are discussed .
they are instructive to the high cumulants of the net baryon number in the qcd phase transitions . |
heretofore , there have been some researches on the encoding of codes on algebraic curves , although they are fewer than researches on the decoding of codes . heegard _
@xcite proposed an encoding for linear codes with nontrivial automorphism groups by using grbner bases for modules over polynomial rings , which was applied by chen _
@xcite proposed another encoding for codes on curves , based on the linear combination of extended reed
solomon ( rs ) codes by the work of yaghoobian _ et al .
_ @xcite . in this research
, we propose a novel encoding scheme for various algebraic codes ; this scheme is considered to be the natural generalization of the well - known encoding for rs codes .
we first establish a simple but non - systematic encoding that employs two - dimensional ( 2-d ) inverse discrete fourier transforms ( idft ) and that generalizes the encoding for rs codes by using one - dimensional idft ( that is , the mattson
solomon polynomial ) .
since the syndromes correspond to the discrete fourier transform ( dft ) , we also obtain a concise decoding via berlekamp massey sakata ( bms ) algorithm .
next , we establish systematic encoding in the sense of the separation of given information and generated redundant in a resulting code - word . this second method of encoding employs a grbner basis and its 2-d linear feedback shift - register and corresponds to the euclidean division by the generator polynomial in the case of rs codes .
both the methods often employ the enlargement of the finite - field arrays to the entire plane by the elements of grbner bases , typically , the defining equation of the algebraic curves . as a more essential idea of our encoding and decoding scheme , we can mention the following duality for substitution @xmath0 then , the rational points having any zero are exceptional ; however , they can be treated similarly to the case of lengthened rs codes as shown in section [ treatment ] .
let @xmath1 denote the set of non - negative integers .
let @xmath2 denote a non - singular c@xmath3 algebraic curves over @xmath4 for @xmath5 with @xmath6 and gcd@xmath7 .
then , the genus of @xmath2 is given by @xmath8 , and @xmath2 has only one @xmath9-rational point at infinity @xmath10 .
we fix a primitive element @xmath11 of @xmath9 .
let @xmath12 denote a set of @xmath9-rational points of the form @xmath13 , i.e. , non - zero coordinates .
we construct codes of symbol - field @xmath9 on @xmath14 s in @xmath15 ; @xmath9-rational points whose coordinates include zero are considered in section [ treatment ] .
we define a subset @xmath16 of @xmath17 as @xmath18 where @xmath19 is equal to the pole order @xmath20 of @xmath21 at @xmath10 . in this study
, we consider codes on algebraic curves @xmath22 where @xmath23 with monomials @xmath24 for @xmath13 . for simplicity , we assume @xmath25 ; then , we obtain @xmath26 . _ elementary encoding : _ the condition @xmath27 in is equivalent to the ordinary linear system @xmath28_{0\le h
< n,\,0\le l < n - k } = \mathbf{0},\ ] ] where @xmath29 for @xmath30 with order @xmath31 .
an encoding method for @xmath32 is the use of the generator matrix @xmath33 $ ] , where @xmath34 is the @xmath35 identity matrix and @xmath36 is obtained from @xmath37 $ ] by the row transform of @xmath38 $ ] and , if needed , the order - changing of @xmath15 .
then , we can systematically encode information symbols @xmath39 to a code - word @xmath40 .
however , this requires the multiplication of ( @xmath41 ) matrix @xmath36 .
thus , our goal should be to eliminate the matrix - multiplication from the encoding algorithm .
on the other hand , with regard to the computing of syndromes , the situation is similar but better than the above encoding because of .
we suppose that an error - vector @xmath42 has occurred during the transmission of @xmath43 and we have received a word @xmath44 .
then , the syndrome decoding requires the @xmath45 values of syndrome @xmath46 $ ] , which agree with @xmath47 for our expression of @xmath32 .
this generalizes the syndrome - calculation for rs codes by the substitution of the roots of the generator polynomial .
hence , we consider an effective encoding method based on our definition of @xmath32 .
in this section , we provide the example of a hermitian code over @xmath48 with defining equation @xmath49 of genus @xmath50 , the minimal pole order ( first non - gap ) @xmath51 , and 24 @xmath9-rational points of @xmath52 and finite .
the primitive element @xmath11 is fixed to satisfy @xmath53 , and the non - zero element @xmath54 ( @xmath55 ) is simply denoted as @xmath56 ( resp .
zero as @xmath57 ) .
note that @xmath58 .
we represent 24 @xmath59-rational points as monomials @xmath60 , which correspond to the shaded boxes of ( c ) in fig .
[ fourier encoding ] .
let @xmath61 be the support of a grbner basis of the ideal @xmath62\,|\,f(p_{h})=0,\,p_{h}\in\mathcal{p}\}$ ] , i.e. , @xmath63 corresponds to the set of monomial representatives of @xmath64/i_\mathcal{p}$ ] , where @xmath64=k[x , y]/(y^{3}+y - x^{4})$ ] .
then , we have @xmath65 . for hermitian codes on non - zero coordinates
, @xmath63 agrees with @xmath66 ; in general , @xmath63 is its subset .
we arrange the information symbols @xmath67 on @xmath68 , and then obtain @xmath69 by considering @xmath70 if @xmath30 , as ( a ) in fig .
[ fourier encoding ] .
furthermore , the grbner basis of @xmath71 extends @xmath69 into @xmath67 for @xmath72 with @xmath73 ; for hermitian codes , @xmath74 from the defining equation , as shown in fig .
[ fourier encoding](b ) , where @xmath75 if @xmath76 .
to encode @xmath67 , we perform the 2-d idft for @xmath67 : @xmath77 then , by substituting @xmath13 into @xmath78 , we have @xmath79 [ lemma of fourier ] we have @xmath80 if there is no @xmath81 with @xmath13
. moreover , the transform defines an injective linear map and a code - word @xmath82 .
we omit the proof and discuss rs - code case in section [ motivation ] . the received word @xmath83 is viewed as @xmath84 for @xmath85 and @xmath86 if there is @xmath81 with @xmath87 ; otherwise @xmath88 ( cf .
[ fourier encoding](d ) ) .
the syndromes from the received word @xmath89 can be obtained by the substitution of @xmath90 for @xmath30 into @xmath91 , as described in section [ codes on algebraic curves ] . in our framework , it is convenient , as shown in fig .
[ fourier encoding](e ) , to substitute the entire @xmath92 , which can be considered as the dft @xmath93 .
then , we have @xmath94 through the error polynomial @xmath95 and the extended information symbols @xmath67 because of our encoding and the 2-d fourier inversion formula @xmath96 we notice that the values @xmath97 are not yet known for @xmath98 since @xmath99 outside @xmath16 . to obtain and subtract all syndrome - values @xmath97 for @xmath73 from @xmath100 , we run the bms algorithm to calculate the grbner basis of the ideal @xmath101 , where @xmath102 denotes the set of error locations .
since the grbner basis provides the 2-d linear recurrence formula for syndromes , we can extend @xmath103 to the entire plane , where the array ( f ) represents the result . finally , as illustrated in fig .
[ fourier encoding](g ) , the information @xmath104 ( and its extension @xmath105 ) is obtained by ( e ) minus ( f ) .
thus , dft is utilized for both the encoding and computing of syndromes ; the decoding consists of two steps , i.e. , this dft and bms algorithm to remove the syndrome - values , without chien search and error - evaluator formula .
since the conventional rs codes are usually encoded systematically , it is natural to consider effective systematic encoding for codes on algebraic curves . however , while the roots of the generator polynomial can be considered a subset of locations in rs code - words , it does not hold in general for our @xmath32 since actually @xmath106 . in this section
, we apply theorem [ lemma of fourier ] and its argument to this problem and obtain a satisfactory solution . as preliminaries , we choose @xmath107 and @xmath108 so that @xmath109 , @xmath110 , and @xmath111 ; @xmath107 is the redundant - point set and is considered to satisfy @xmath112 without loss of generality , and @xmath108 is the information - point set .
we calculate the grbner basis of the ideal @xmath113 in advance . in the example of fig .
[ systematic ] , the shaded boxes in ( h ) indicate @xmath108 , and we obtain the grbner basis by the bms algorithm as follows : @xmath114 where , for example , the leftmost array represents the polynomial @xmath115 . for simplicity ,
we assume that the support of the grbner basis for @xmath113 is generic @xcite , that is , the support corresponds to @xmath116 ; this assumption is not very strong since the generic support has the probability @xmath117
. then we represent @xmath118 information symbols @xmath119 as shown in fig .
[ systematic](h ) . to generate the redundant part of the code - word ,
we first compute its dft @xmath120 by @xmath121 and then , we extend @xmath122 ( nine values in fig .
[ systematic](i ) ) on the support into @xmath123 on the entire plane by the grbner basis ( @xmath124 ) , or more precisely , by its recursive formula in section [ proof of theorem ] .
if we perform idft for the negative @xmath125 of the extended array , the redundant part can be obtained since the resulting values on @xmath108 are zero by theorem [ lemma of fourier ] and their dft ( i.e. , syndrome ) agrees with @xmath126 . if we perform idft for the subtraction @xmath127 ( fig .
[ systematic](j ) ) , i.e. , we compute @xmath128 then @xmath129 is a code - word in @xmath32 since @xmath130 for @xmath30 .
moreover , it is systematic , as observed at fig .
[ systematic](k ) , and in fact we have @xmath131 for @xmath132 since the idft of @xmath133 vanishes at @xmath108 by theorem [ lemma of fourier ] .
while the error - value estimation was performed by using the idft of ( n ) in @xcite , it is efficiently incorporated into our procedure . in the encoding of section [ encoding 2-d ] , each procedure of encoding and decoding contains either the dft or idft .
although in this section , each step of encoding and decoding requires both the dft and idft , we can use only one dft calculator for all transforms in practical circuits for the encoder and decoder .
recall that the systematic matrix - encoding described in section [ codes on algebraic curves ] requires multiplications of the @xmath41 matrix ; our method requires the calculators of the 2-d feedback shift - registers and memory - elements for at most @xmath134 coefficients , which correspond to the @xmath45 coefficients of the generator polynomial for rs codes .
our encoding and decoding scheme is widely applied to various algebraic codes such as 2-d cyclic codes and hyperbolic cascaded rs ( hcrs ) codes ; for these codes , the encodings in sections 3 and 4 are similarly performed except for the total order in the bms algorithm . here
, we deal with the systematic encoding of hcrs codes . in this section
let @xmath135 .
a hcrs code @xcite over @xmath4 is defined as @xmath136 where @xmath137 .
then , the minimum distance @xmath138 of @xmath32 is bounded as @xmath139 . in fig .
[ hcrs ] , @xmath140 over @xmath59 is demonstrated for four - error correction .
the non - systematic encoding is equal to the idft of ( p ) . to encode systematically ,
we first compute a grbner basis of an ideal @xmath141 , where @xmath142/(x^{q-1}-1,y^{q-1}-1,)$ ] , with respect to a total order @xmath143 @xmath144 the elements of the basis that is needed for the extension in the systematic encoding are shown below .
@xmath145 then , the values after dft on @xmath16 in ( q ) are extended by the recurrence formula similar to .
thus , the idft of ( r ) , where ( r ) equals ( q ) minus the extended array , is a systematic code - word ( s ) . to decode a received word ( t ) from the channel ,
we perform , for syndrome values on @xmath16 in ( u ) , the bms algorithm with respect to the total order ( @xmath146 ) @xcite . in the case of our example , the error - locator polynomials are expressed as follows . @xmath147
the recurrence similar to by the above basis extends the syndrome values on @xmath16 to the entire plane , as fig .
[ hcrs](v ) . finally , the idft of @xmath148 in fig .
[ hcrs ] provides the correct transmitted word fig .
[ hcrs](w ) .
recall the encoding for rs codes by euclidean division : @xmath149 where @xmath150 is an information polynomial with @xmath151 for @xmath152 ; @xmath153 , the generator polynomial ; @xmath154 , the remainder of the division with quotient @xmath155 .
then , it is apparent that @xmath43 from @xmath156 is a code - word of the rs code @xmath157 with @xmath158 and @xmath159 .
this method is _ systematic _ , i.e. , @xmath160 for @xmath161 .
if we have received a polynomial @xmath162 containing an error polynomial @xmath163 in the channel , the values of syndromes @xmath164 can be computed as @xmath165 by substituting the roots of @xmath166 into @xmath167 .
we notice that @xmath168 can be also considered to be the dft of @xmath169 .
thus , we obtain another encoding method ( _ non - systematic _ ) by the idft @xmath170 .
then , @xmath171 is another code - word of @xmath32 since @xmath172 it is possible to systematically encode by using an alternative procedure . from , we obtain @xmath173 for @xmath152 .
moreover , we define array @xmath174 inductively by @xmath175 where @xmath176 . then , it follows that @xmath177 with @xmath178 not only for @xmath179 but also for @xmath161 .
thus , the idft @xmath180 for @xmath174 is observed to agree with @xmath181 of @xmath154 by using fourier inversion formula ; @xmath182 again indicates the encoding , and moreover we obtain two ways of calculating @xmath154 , i.e. , a commutative diagram .
thus , we obtain two encoding methods for rs codes , which we have generalized .
we generate @xmath183 for @xmath73 recursively from @xmath184 as in the encoding at section [ encoding 2-d ] and [ systematic encoding ] , which is stated here more precisely .
it may be assumed that we have the support @xmath185 and that each grbner basis consists of @xmath186 elements @xmath187 , where @xmath188 with @xmath189 and @xmath190 .
then the recurrence for @xmath191 is defined as follows : @xmath192 the resulting values do not depend on the order of the generation because of the property of grbner bases .
with regard to the systematic encoding in section 4 , we can treat locations including zero in a manner similar to the case of non - zero locations .
there are three @xmath59-rational points @xmath193 , @xmath194 , and @xmath195 for our example of hermitian codes , which are denoted by the shaded boxes in the top row of fig .
[ zero](x ) .
although we can not compute the dft for three information symbols at these locations , we note that if an error @xmath196 has occurred on , e.g. , @xmath193 , then the syndrome values are all @xmath57 except for @xmath197 at @xmath198 since they are computed by the substitution of @xmath193 into @xmath199 .
we also note that the idft of @xmath200 equals an @xmath201 all-@xmath197 array .
thus the analogue of dft is obtained for information 7 at @xmath193 as only 7 at @xmath198 .
similarly , the analogue is obtained for information 5 at @xmath194 as only @xmath202 $ ] in the first row of @xmath63 , which also equals the one - dimensional ( 1-d ) dft for @xmath203 $ ] ; for information 2 at @xmath195 , the analogue is obtained as @xmath204 $ ] at the top , which is also the 1-d dft for @xmath205 $ ] .
hence , we obtain the analogue of dft for the array ( x ) as the array ( y ) by summing and obtain the redundant part of the code - word ( aa ) by the idft of the extended array ( z ) .
notice that the idft of @xmath206 such as in section 4 provides the sum of the parts of the code - word on @xmath63 and the idft of the analogue arrays of dft , i.e. , all-@xmath207 array , all-@xmath208 array except @xmath209^{t}$ ] in the third column , which is the idft of @xmath210 , and all-@xmath208 except @xmath211^{t}$ ] in the seventh column .
thus , we have completed the treatment of locations including zero components .
this work was partly supported by the academic frontier project for future data storage materials research by the japanese ministry of education , culture , sports , science and technology ( 19992008 ) , toyota physical and chemical research institute , and storage research consortium ( src ) .
k. saints , c. heegard , `` algebraic - geometric codes and multidimensional cyclic codes : a unified theory and algorithms for decoding using grbner bases , '' _ ieee trans .
inf . theory , _
vol.41 , no.6 , pp.17331751 , nov . 1995 .
chen , c .- c .
lu , `` a serial - in - serial - out hardware architecture for systematic encoding of hermitian codes via grbner bases , '' _ ieee trans .
communications , _ vol.52 , no.8 , pp.13221332 , aug . 2004 .
h. matsui , s. mita , `` footprint of polynomial ideal and its application to decoder for algebraic - geometric codes , '' _ proc .
int . symp .
information theory and its applications ( isita ) , _
pp.14731478 , oct .
2004 .
s. sakata , h. e. jensen , t. hholdt , generalized berlekamp massey decoding of algebraic geometric code up to half the feng rao bound , " _ ieee trans .
vol.41 , no.6 , part i , pp.17621768 , nov . 1995 . | we propose a novel encoding scheme for algebraic codes such as codes on algebraic curves , multidimensional cyclic codes , and hyperbolic cascaded reed solomon codes and present numerical examples .
we employ the recurrence from the grbner basis of the locator ideal for a set of rational points and the two - dimensional inverse discrete fourier transform .
we generalize the functioning of the generator polynomial for reed solomon codes and develop systematic encoding for various algebraic codes . |
the search for a light higgs boson and the study of gauge boson pair production @xcite are important physics goals of the lep2 upgrade at cern .
if kinematically accessible to lep2 , the higgs boson will be primarily produced via the bjorken process @xmath3@xcite .
an irreducible physics background to this processes is the production of @xmath4 boson pairs , which is most troublesome when @xmath5 .
the clear separation of these two processes is required for a convincing higgs boson discovery or lower bound mass limit .
kinematics as well as characteristic angular correlations will be useful in this separation .
the study of @xmath1 boson pair production is also interesting in its own right@xcite , as it involves vertices coupling three gauge bosons ( especially @xmath6 ) , vertices which to date are poorly - probed experimentally .
these couplings may be probed with varying degrees of sophistication by testing the predictions for the total cross section , production angle distribution , and correlations among the vector boson decay products .
hence for all of these processes , @xmath0 and @xmath7 , we are keenly interested in the angular correlations among the final state particles . to disentangle these correlations
it is very useful to understand the spin correlations of the heavy particles produced .
these spin correlations as well as the subsequent correlations of the decay products are the focus of this paper . until recently
, most spin - related studies were carried out within the framework provided by the zero momentum frame helicity basis ( see , for example refs .
@xcite ) . for particles which are ultra - relativistic , such as at the next generation of linear colliders , this is appropriate .
however , for particles which are only moderately relativistic , there is no reason to expect that the helicity basis will produce the best description of the physics involved .
indeed , for the case of @xmath8 production at low energy @xmath9 and @xmath10 colliders , it was shown in refs .
@xcite that the helicity basis is far from optimal . in this paper we use the generalized spin basis introduced in ref .
@xcite to describe @xmath11 , @xmath12 , and @xmath13 production and decay at lep2 energies .
in particular , we write the polarized production amplitudes in terms of the most general @xmath14-conserving choice of spin axes , and from these expressions determine the choice of axes which leads to the most distinctive correlations among the final state particles .
we argue that it is advantageous at lep2 energies to tune the choice of spin axis to the experimental issue being investigated , rather than to simply use the helicity basis and try to unravel the contributions from what are in many cases nearly - equally populated spin states . for all three processes ,
@xmath0 and @xmath7 , much effort in the literature has been devoted to the analysis of various anomalous trilinear couplings@xcite , employing , of course , the helicity basis . rather than attempt to rework all of those studies in terms of the generalized spin - basis , we have chosen to concentrate on the picture within the standard model .
it is clear from our results , however , that such a study would be worthwhile , as we expect different spin bases to be optimal for the extraction / limitation of different anomalous couplings .
such an analysis would be a major extension of this work .
the remainder of this paper is organized as follows . after a brief discussion of our notation and conventions in sec .
[ xibasis ] , we consider the @xmath15 system within the generalized spin basis framework ( sec . [ zhandzz ] ) . for each process
, we consider in detail the polarized production cross sections , and construct spin bases suggested by the expressions for the amplitudes .
we consider the decay of polarized @xmath4 s , and then link the production and decay to study the correlations among the decay products .
we are able to construct a basis in which these correlations differ significantly for the @xmath11 and @xmath12 cases , whereas in the helicity basis the correlations are virtually non - existent . sec .
[ wwsection ] contains the corresponding analysis for the @xmath16 system . in this case , matters are less clear - cut , and we present several suggestions for bases which may be useful under different circumstances . finally , sec .
[ conc ] contains our conclusions .
explicit expressions for the polarization vectors we employ for the massive vector bosons in the generalized spin basis appear in the appendix .
to describe the polarized production cross sections for @xmath11 , @xmath12 , and @xmath13 discussed in this paper , we employ the generic spin basis introduced by parke and shadmi in ref .
@xcite , as illustrated in fig .
[ xidef ] . for @xmath12 and @xmath13 ,
this is the most general basis which conserves @xmath14 .
we label the two particles produced in the collision by @xmath17 and @xmath18 .
the zero momentum frame ( zmf ) production angle @xmath19 is defined as the angle between the electron and @xmath17 directions .
the spin states for the first particle are defined in the its rest frame , where we decompose the @xmath17 spin along the direction @xmath20 , which makes an angle @xmath21 with the @xmath18 momentum in the clockwise direction .
likewise , the @xmath18 spin states are defined in its rest frame along the direction @xmath22 , which makes the _ same _ angle @xmath21 with the @xmath17 momentum , also in the clockwise direction .
we denote the two transverse polarization states by @xmath23 and @xmath24 and the state by 0 . throughout this paper
we use the terms `` transverse '' and `` longitudinal '' to refer to directions relative to the spin axis , _ not _ the direction of motion of the particle .
a generic vector boson spin will be designated by @xmath25 .
if we sum over all of the polarizations of the vector boson(s ) , then the dependence on @xmath21 drops out of the result . note
that although the higgs boson is a scalar , we may still define a spin axis for it as if it were a vector .
the spin zero character of the higgs will be reflected in a lack of any dependence on the choice of this axis . within the generic spin basis framework ,
specific spin bases are defined by stating the relationship between @xmath21 and @xmath19 ( and any other relevant event parameters ) .
for example , the ubiquitous helicity basis is defined by fixing .
[ helicitydef ] in this case , the spins are defined along the directions of motion of the particles as seen in the zmf .
another interesting basis is the beamline basis @xcite , which is defined by = ; = - .
[ bmldef ] here @xmath26 is the zmf speed of @xmath17 . in this basis , the spin axis for @xmath17 is the electron direction .
furthermore , if @xmath17 and @xmath18 have identical mass , then the spin axis for @xmath18 is the positron direction . later in this paper
, we will encounter additional bases , whose definitions are inspired by the form of the matrix elements for the processes under consideration . except for the fermion masses , which we set equal to zero ,
all input masses and coupling constants used in the computations presented in this paper are the central values as reported in the 1996 review of particle properties@xcite .
furthermore , we also neglect the coupling between the electron and higgs .
if the higgs is light enough to be observed at lep2 , then its production will be dominated by the process @xmath28 , for which the largest background is @xmath29 .
this suggests the potential for difficulties should the higgs mass lie too close to the mass of the @xmath4 . in particular , at @xmath30 ,
the tree - level cross section for @xmath11 is only 0.5 pb when @xmath31 , compared to 1.2 pb for @xmath12 . as noted by brown @xcite , this situation improves if @xmath32-tagging is employed , as the higgs decays mainly to @xmath33 whereas the @xmath4 decays to this final state only 15% of the time . rather than
rely on @xmath32-tagging , kunszt and stirling @xcite have considered the distribution of the angle that the decay leptons in @xmath34 .
make with the direction of the beam in the zmf .
they find a significant difference in this distribution in the @xmath11 and @xmath12 cases .
finally , summers @xcite has added the effect of polarizing the beams on the distribution of ref .
@xcite . in this section
we examine the @xmath11 and @xmath12 processes within the generalized spin basis framework .
we begin in sec .
[ polzhprod ] by analyzing the polarized production amplitudes for the process @xmath35 in great detail .
the expressions for these amplitudes will show us how to construct a basis in which only transversely - polarized @xmath4 s are produced in association with the higgs . in sec .
[ polzzprod ] we present polarized production amplitudes for @xmath36 . after describing the decay of polarized @xmath4 s ( sec .
[ polzdec ] ) , we combine the production and decay information to discuss the angular correlations among the decay products in the @xmath15 system ( sec .
[ zzcorrel ] ) .
our analysis shows that the angular variable suggested by kleiss and stirling @xcite , while close to optimal , may be improved upon , especially as the @xmath11 pair is produced further and further above threshold .
in particular , we find that in the basis constructed in sec .
[ polzhprod ] , certain decay angular distributions are very different for @xmath11 versus @xmath12 . in contrast , these same distributions are nearly featureless in the helicity basis .
we conclude this part of the paper in sect . [ pseudo ] with a brief look at how the production of a pseudoscalar higgs would differ from the standard model higgs .
although we focus on the case where @xmath37 , our results are more general , since the existence of our optimized basis for @xmath28 depends only on that process being kinematically allowed .
it is also potentially useful at the tevatron , because the amplitude for @xmath38 has the same spin structure as the amplitude for @xmath35 . to describe the process @xmath35
, we take particle @xmath17 in fig .
[ xidef ] to be the @xmath4 , and particle @xmath18 to be the higgs .
in general , the higgs and @xmath4 have different masses ( @xmath39 and @xmath40 respectively ) , leading to the following connection between the center - of - mass energy @xmath41 and speed @xmath42 of the @xmath4 boson : = . [ betaz ] rather than use eq .
( [ betaz ] ) to eliminate one of @xmath43 or @xmath42 from our expressions , it is more convenient to use both quantities . if we neglect the electron mass , there is but a single diagram for @xmath35 .
it leads to the differential cross section = g_f^2 m_w^2 & & m_z^3 ( s+m_z^2-m_h^2 ) ^2 & & \ { ^2 2_w [ _ l^(,,)]^2 + 4 ^4 _ w [ _ r^(,,)]^2 } [ zhsigma ] where @xmath44 is the fermi coupling constant , @xmath45 is the mass of the @xmath1 boson , @xmath46 is the weinberg angle , and @xmath47 is the usual relativistic factor , @xmath48 .
the two terms in the curly brackets come from the two possible chiralities of the initial electron line .
all of the spin information is contained in the spin functions , @xmath49 , which are given by _
l^ ( , , ) = _ r^ ( , , ) = 1 [ zhtranss ] and _ l^0 ( , , ) = _ r^0 ( , , ) = - . [ zhlongs ] summing over the three possible spins of the @xmath4 we obtain the total ( unpolarized ) differential cross section _ d^ = g_f^2 m_w^2 & & m_z^3 ( s+m_z^2-m_h^2 ) ^2 & & \ { 1 - 4 ^ 2_w+8 ^ 4_w } \ { 2(1-^2 ) + ^2 ^ 2}. [ zhsigmatotal ] eq .
( [ zhsigmatotal ] ) is independent of the spin axis angle @xmath21 , as it must be .
since , in general , at each value of @xmath19 we are allowed to choose a different value of @xmath21 , it is useful to plot these amplitudes in the @xmath50-@xmath51 plane . consequently , we define the quantity f^ ( , , ) , [ zhfraction - def ] which is the fraction of the total cross section coming from the spin state @xmath25 . in fig .
[ zh - prodcontours ] we have plotted @xmath52 for a collider energy @xmath30 and higgs mass @xmath53 .
these plots illustrate the features of the amplitude discussed below , and are an indispensable aid when the amplitude becomes complicated , as in the @xmath12 and @xmath13 cases . for @xmath54
, we recover the usual helicity basis expressions from eqs .
( [ zhsigma])([zhlongs ] ) : _ l^ ( , , ) = _ r^ ( , , ) = & & 1 ( 1 - ) ; _ l^0 ( , , ) = _ r^0 ( , , ) = & & . at high energy ,
the transverse amplitudes die off , leaving only the longitudinal amplitude .
an examination of eq .
( [ zhtranss ] ) reveals that it is not possible to make both of the transverse spin functions @xmath55 and @xmath56 vanish simultaneously .
consequently , it is impossible to make either of the transverse amplitudes of eq .
( [ zhsigma ] ) vanish .
there is , however , a zero in the longitudinal contribution , eq .
( [ zhlongs ] ) , which results from choosing = ; = , [ zhideal ] @xmath57 .
the basis corresponding to this choice , the basis , will turn out to be very useful in the study of @xmath11 events .
the nonvanishing spin functions in the basis are _
l^(,,^-1 ( ) ) & = & _ r^(,,^-1())&= & 1 .
these functions are completely flat in @xmath50 at threshold , and become proportional to @xmath58 for @xmath59 .
note that the existence of this basis does not depend on either the machine energy or the higgs mass : eqs .
( [ betaz ] ) and ( [ zhideal ] ) remain well - defined so long as @xmath60 . in particular , for @xmath61 , we have 1 ; 0 , [ zhhighe ] that is , @xmath62 .
this is clearly _ not _ the helicity basis .
the basis is also the basis in which the @xmath63 and @xmath64 components are each maximized . for completeness
, we mention the basis , defined by @xmath65 .
as its name suggests , the basis maximizes the fraction of @xmath4 s .
this basis is potentially useful for large values of @xmath66 since the fraction of @xmath4 s with increasing energy approaches unity faster than in the helicity basis . in table
[ zh - breakdowns ] , we list the contributions to the total amplitude for each of the three possible @xmath4 spins as measured in the helicity , beamline , , and bases .
we employ a center - of - mass energy @xmath67 and a higgs mass @xmath53 , the value which makes separation of @xmath11 and @xmath12 events the most difficult .
note that in the helicity basis , the three spin states are populated nearly equally . as we shall see below
, this feature makes the helicity basis a very poor choice for studying the @xmath15 system . because the higgs mass is still unknown
, we might worry that the predictions for the spin compositions would depend strongly on @xmath39 , rendering this discussion pointless .
for any particular machine energy and higgs mass , eq .
( [ betaz ] ) tells us the appropriate value of @xmath42 to use . in fig .
[ zhbetaplot ] we have plotted the fractional contribution of each of the three spin components as a function of @xmath42 for the helicity , beamline , , and bases .
we see that in the basis , the spin composition is remarkably flat as a function of @xmath42 , except near @xmath68 .
in fact , for @xmath69 , @xmath70 when @xmath67 .
so , for the values of @xmath39 and @xmath41 accessible at lep , the spin composition in the basis is effectively constant .
contrast this to the beamline basis , which actually coincides with the basis at @xmath71 and the helicity basis for @xmath72 ! in this case ,
the spin composition is not particularly stable . a second reason for preferring a basis in which the spin composition is stable with respect to changes in @xmath42 relates to the effects of initial state radiation ( isr ) .
although we have neglected these effects in our computations , it is apparent what should happen qualitatively : some of the observed events will be produced from @xmath73 pairs with a center - of - mass energy lower than the machine energy .
therefore , the smaller the dependence on @xmath42 , the less sensitive the breakdowns will be to the effects of isr . finally , in fig .
[ zh - cth ] we compare the production angular distributions of the polarized cross sections in the helicity and bases .
we see that in the helicity basis , the identity of the dominant component depends on @xmath50 , whereas in the basis , the ratio of the two non - vanishing amplitudes is essentially constant .
we now turn to the process @xmath36 . with our comparison to @xmath11 production in mind , we identify the @xmath4 which decays leptonically with @xmath17 and the @xmath4 which decays to @xmath33 with @xmath18 in fig .
[ xidef ] .
the only diagrams for @xmath12 production are @xmath74-channel electron exchange diagrams , one for each ordering of the two @xmath4 s . for an initial fermion pair with left - handed chirality we have & & = g_f^2 m_w^4 ^2 ^2 , [ prodzzleft ]
while the result for right - handed chirality reads & & = g_f^2 m_w^4 ^2 ^2 .
[ prodzzright ] these expressions contain the common zmf speed of the two @xmath4 s , which is connected to the center - of - mass energy by = .
[ plainbeta ] we have not combined the pairs of terms appearing in the square brackets in order to make manifest the similarities to the @xmath1-pair amplitudes presented in sec . [ wwsection ] .
all of the spin information is contained in the spin functions , @xmath75 , which may be written as _
l^ ( , , ) = 2_1(,,)_2 ( , , ) [ funcab ] _
l^ ( , , ) = 2_3(,,)_4 ( , , ) [ funcaa ] _
l^ ( , , ) = _ l^0 ( , , ) = [ funcaoob ] _
l^00 ( , , ) = 2(1 - 2+^2 ) -4_3(,,)_4 ( , , ) [ funcoo ] where _ 1 ( , , ) & = & ( 1++ ) - ( + ) _ 2 ( , , ) & = & ( - ) + ( + ) _ 3 ( , , ) & = & ( - ) + _ 4 ( , , ) & = & - ( - ) .
[ gfuns ] since the replacement @xmath76 has the effect of interchanging the @xmath23 and @xmath24 states , expressions for the remaining spin combinations may be obtained from the relation _
l^| ( , , ) = _ l^-,-| ( , , ) .
[ leftother ] the total differential cross section is remarkably simple . summing over the spins of both @xmath4 s and including both fermion chiralities we obtain _
, |,c d_c^| = & & g_f^2 m_w^4 ( ^42_w + 16 ^ 8 _ w ) & & ( 1-^2 ) .
[ zztotal ] of the nine spin configurations , there are only six which are independent , since @xmath14 invariance forces the equality of the @xmath77 and @xmath78 components , as well as the equality of the @xmath79 and @xmath80 components @xcite .
therefore , for the rest of the paper we will refer to @xmath81 $ ] , @xmath82 $ ] , and @xmath83 $ ] as single components .
one of the ways which the @xmath84 and @xmath85 violation present in the weak interactions manifests itself as the different prefactors in eqs .
( [ prodzzleft ] ) and ( [ prodzzright ] ) .
the spin functions , on the other hand , obey these symmetries , and are connected by _ r^| ( , , ) = _ l^-,-| ( , , ) . [ leftright ] because of the somewhat complicated form of the amplitudes , it is instructive to plot the amplitudes for each of the six independent spin configurations in the @xmath50-@xmath51 plane . in fig .
[ zz - prodcontours ] , we have plotted the quantity f^| ( , , ) [ fraction - def ] ( _ i.e. _ the fraction of the total amplitude in a given spin configuration ) for a machine center - of - mass energy of 192 gev .
visible in the plots is the fact that the @xmath86 contribution is exactly twice the @xmath81 $ ] contribution for all values of @xmath19 and @xmath21 [ see eqs .
( [ prodzzleft ] ) , ( [ funcaa ] ) and ( [ funcoo ] ) ] . also noteworthy are the broad minima in the @xmath87 , @xmath88 , @xmath86 , and @xmath81 $ ] components in the vicinity of the diagonal . the @xmath82 $ ] and @xmath89 $ ] contributions have corresponding maxima in this region . in table
[ zz - breakdowns ] we give the fraction of the total amplitude coming from each of the six independent spin configurations in the four bases tabulated for @xmath11 . in two of the bases ( helicity and ) , the high degree of symmetry present in eqs .
( [ prodzzleft ] ) and ( [ prodzzright ] ) resulting from the two identical final state particles manifests as the equalities @xmath90 and @xmath82=[(0+)+(-0)]$ ] .
it is important to note that we _ define _ the basis in terms of eqs .
( [ betaz ] ) and ( [ zhideal ] ) , even when we are dealing with @xmath12 production .
thus , the spin breakdowns in this basis depend on the higgs mass via the value assigned to @xmath66 . for @xmath91 ( the interesting region ) ,
the curve traced out by the basis in the @xmath50-@xmath51 plane lies near the diagonal , an area in the @xmath50-@xmath51 plane where the contributions from the various spins are relatively stable .
thus , it is not surprising to find that for @xmath91 and @xmath67 , there are no significant deviations from the entries for the basis in table [ zz - breakdowns ] . in fig .
[ zzbetaplot ] we have have plotted the fraction of the total amplitude in each of the spin components using the helicity and bases as a function of @xmath26 . in the helicity basis , we find that the @xmath92 and @xmath93 contributions are always equal , as are the @xmath94 $ ] and @xmath95 $ ] contributions . consequently , if we look at the population of the spin states for one of the @xmath4 s while summing over the other @xmath4 , we find that the @xmath23 and @xmath24 states are populated equally .
furthermore , at threshold , this inclusive breakdown is precisely @xmath96 for each spin , and this changes relatively slowly with increasing @xmath26 .
the situation is considerably better in the basis , where no two of the six spin contributions are equal , and where there is a reasonable difference in the inclusive contributions .
while we have used @xmath97 ( _ i.e. _ @xmath31 ) in preparing these plots , we have verified that for the range accessible to lep , the dependence on the higgs mass is inconsequential . in fig .
[ zz - cth ] we compare the angular distributions of the polarized production cross sections in the helicity and bases at @xmath67 . in the helicity basis there is a complicated interplay among the six spin components , whereas in the basis the two dominant components appear in approximately the same ratio , independent of @xmath50 .
the @xmath98 coupling violates both parity and flavor universality .
thus , the angular distributions for the decay of polarized @xmath4 bosons are forward - backward asymmetric , and depend on which fermions appear in the final state . neglecting the mass of the final state fermions , ) and ( [ zl ] ) .
these effects are greatest for @xmath32-quarks , where they are less than a few percent .
] we obtain the following angular distributions in the rest frame of the decaying @xmath4 : = 38[zt ] for the transverse polarizations and = 34 ^2[zl ] for the polarization .
these distributions have been normalized to unit area by inclusion of the partial width @xmath99 for the decay @xmath100 .
we take @xmath101 to be the angle between the direction of motion of the fermion and the spin axis as seen in the @xmath4 rest frame , and absorb the dependence on the couplings between left- and right - handed chirality fermions to the @xmath4 into the factor @xmath102 . for convenience ,
we have collected the values of @xmath102 in table [ zdecaytable ] .
the corresponding distributions are plotted in fig .
[ zdecaydist ] .
unfortunately , the decays with the most distinct distributions , @xmath103 are invisible . furthermore ,
charge and flavor identification for decays to quarks ( which appear as two jets ) is virtually impossible , except perhaps for @xmath32 quarks . for @xmath104 , we have a fairly large overlap between the @xmath63 and @xmath64 distributions .
all of these features lead us to conclude that it may not be possible to distinguish between the @xmath63 and @xmath64 polarizations .
however , as we shall see below , the difference between the @xmath105 and @xmath106 states is sufficient for separating @xmath11 and @xmath12 , provided that the appropriate spin basis is used .
naturally , since the higgs is a scalar , its decay angle distributions will be completely flat .
in this section we will put everything together to discuss the correlations among the @xmath15 decay products .
for concreteness , we take the final state to be @xmath107 , the @xmath108 coming from a decaying @xmath4 , and the @xmath33 from the higgs or other @xmath4 .
the results we present do not change very much if we consider a @xmath109 final state : since it is not possible to tell which light jet is the @xmath110 and which is the @xmath111 , all distributions in the @xmath4 decay angle @xmath112 defined below would be symmetric about @xmath113 . on the other hand ,
@xmath114 is only slightly larger than @xmath115 , leading to small asymmetries between positive and negative values of @xmath116 . in order to display the advantages of using the basis defined by eq .
( [ zhideal ] ) , we consider a scatter plot in the decay angles @xmath112 and @xmath117 , defined as follows . on the @xmath118 side of the event ,
let @xmath112 be the angle between the @xmath119 emission direction and the @xmath4 spin axis in the @xmath4 rest frame . on the @xmath120 side of the event ,
let @xmath117 be the angle between the @xmath32 emission direction jets , then the distributions presented here should be folded about @xmath121 . ] and the @xmath122 spin axis in the @xmath122 rest frame . in fig .
[ zz - zhdecaycontours ] we present this distribution for @xmath11 and @xmath12 production for the helicity and bases for @xmath123 and @xmath31 .
the difference between the two plots is dramatic : in the helicity basis , the two distributions are nearly flat , while there is significant structure when the basis is used .
this difference is underscored by the plots in fig .
[ zzvszh1d ] , which compare the 1-dimensional distributions in @xmath116 and @xmath124 .
clearly , the angular information contained in an analysis employing the basis provides a useful handle in the separation of @xmath11 and @xmath12 events , even when @xmath5 . in contrast , the helicity angles are virtually useless in this respect !
kunszt and stirling @xcite have noted significant angular correlations in the zmf between the direction of the final state muons and the beam direction .
effectively this is the beamline basis . at threshold ,
the beamline basis coincides with the optimal ( ) basis , and the zmf coincides with the @xmath4 rest frame . at lep2 , @xmath125 , _
i.e. _ not that far above threshold
. thus , even though they do nt use the optimal basis , kunszt and stirling do see non - trivial angular correlations .
the largest possible correlations , however , are obtained using the muon direction relative to the basis spin axis in the @xmath4 rest frame . furthermore , the gap between their distribution and the optimal one widens as we move further above threshold .
an issue which may be raised concerns the effect of finite detector resolution on the ability of the experimentalists to accurately measure the decay angles we have proposed , especially since they have been defined in the rest frames of the parent particles , not the lab frame .
to get an estimate of the size of these effects , we introduced gaussian energy smearing with a width set by @xmath126 into the energies of the charged leptons and the @xmath32 quarks , and then compared the decay angles reconstructed solely from the smeared momenta to the actual decay angles .
we take the root mean square deviation of this difference to be an estimate of the error introduced by the smearing .
since the definition of @xmath21 depends both on the speed of the parent boson as well as the production angle , the detailed values of our error estimate will vary with the machine energy , choice of basis , and the process ( @xmath11 or @xmath12 ) in which the bosons were produced .
typical values at @xmath30 and @xmath31 are @xmath127 , for all three bases . to see the effects of this angular smearing ,
an experimentalist would need to use bins of approximately this size or smaller .
the total number of events observed will probably not be large enough to warrant such a large number of bins in the @xmath128 plots .
although a more detailed study using the full - blown detector simulation is clearly needed , it would appear that it is possible to measure the indicated angles with sufficient precision .
the amplitudes for the processes @xmath129 can be written as a sum over the spin states of the intermediate massive particles of the production times the decay amplitudes for each spin component . in the classical limit
there is no interference between these amplitudes .
however , in general , these amplitudes do interfere .
the size of these interference terms can be used as a measure of how well the classical result for a particular basis describes the full process : if the interference terms are small , then the classical picture captures nearly all of the physics . on the other hand ,
if the interference terms are large , a significant amount of the physics is omitted from the classical picture . to quantify these considerations , at each kinematic point for a given process
, we may define the quantity @xmath130 , the value of the interference terms for a given spin basis divided by the total @xmath131 matrix element squared . for a process containing @xmath132 independent intermediate spin configurations
, @xmath130 can range from @xmath133 ( total destructive interference ) to @xmath134 ( total constructive interference ) .
a plot of @xmath135 displays the relative importance of the various values of @xmath130 .
clearly , in the ideal case where classical intuition captures all of the physics , this distribution will be a delta function centered at @xmath136 .
[ zhinf ] compares the helicity , beamline ant bases for @xmath11 production and decay , whereas fig .
[ zzinf ] does the same for @xmath12 . for the signal ( @xmath11 )
, we see that the peaking in @xmath130 is about an order of magnitude greater in the basis than in the beamline basis .
the helicity basis shows essentially no peaking . for the background ( @xmath12 ) ,
the difference is not as dramatic , but it is still clear that the basis gives the smallest interference terms , and hence the superior description .
although the higgs boson of the standard model is unambiguously @xmath14-even , for the sake of comparison , it is useful to consider how the spin correlations differ for a @xmath14-odd higgs @xmath137 ( commonly referred to as a pseudoscalar ) , a feature of two - doublet higgs models .
although a tree - level @xmath138 coupling is forbidden , such a coupling can be generated at the one - loop level @xcite .
following ref .
@xcite , we write the ( effective ) @xmath138 vertex as - igm_z k_1^ k_2^ _ , where @xmath139 and @xmath140 are the 4-momenta of the two @xmath4 s , @xmath141 is a dimensionless coupling constant , and @xmath142 is the mass scale at which this vertex is generated .
this coupling leads to the differential cross section = g_f^2 m_w^2 & & ( m_z ) ^4 ( s+m_z^2-m_h^2 ) ^2 & & \ { ^2 2_w [ _ l^(,)]^2 + 4 ^4 _ w [ _ r^(,)]^2 } .
[ zasigma ] in this case the spin functions are independent of energy : _ l^ ( , , ) = _ r^ ( , , ) = & & 1 ( ) ; _ l^0 ( , ) = _ r^0 ( , ) = & & [ zaspinfun ] thus , the total ( unpolarized ) differential cross section is _ d^ = g_f^2 m_w^2 & & ( m_z ) ^4 ( s+m_z^2-m_h^2 ) ^2 & & ( 1 - 4 ^ 2_w+8 ^ 4_w ) ( 2 - ^2 ) , [ zasigmatotal ] in agreement with ref . @xcite .
it is clear from the especially simple form of the spin functions , eq .
( [ zaspinfun ] ) , that the optimal basis for @xmath143 is the helicity basis , independent of the machine energy .
only the helicity basis has the property that one of the three amplitudes ( the longitudinal one ) vanishes . because the @xmath21 dependence of the @xmath11 and @xmath143 amplitudes differs so greatly , a measurement of the @xmath4 spin composition of a higgs signal for different choices of @xmath21 could potentially differentiate between the two cases .
in particular , it would be useful to measure the fraction of longitudinally polarized @xmath4 s in a sample of @xmath4-higgs candidates in both the helicity and bases .
for the signal events , a scalar higgs should show no longitudinal @xmath4 s in the basis , while for a pseudoscalar higgs this would occur only in the helicity basis .
with the crossing of the @xmath1-pair production threshold at lep2 , it is now possible to probe the electroweak sector of the standard model in new ways .
the tree - level diagrams for @xmath145 contain triple gauge boson vertices ( @xmath146 and @xmath147 ) , allowing direct studies of the yang - mills structure of the theory .
the precise form of these couplings predicted by the @xmath148 nature of the underlying theory is reflected in the various angular distributions associated with the @xmath13 events @xcite . if we allow the strengths of the existing @xmath146 and @xmath147 couplings to vary , or if we introduce new ones , then not only will the differential distribution in the production angle @xmath19 be affected , but the shares of the total cross section coming from the different spin states of the @xmath1 s will also vary .
this second effect will show up in the distributions in the angles between the decay products and suitably defined spin axes . as mentioned in the introduction , a complete reworking of the anomalous trilinear couplings analysis using our generalized spin basis , while worthwhile , would be a major extension of this work .
nevertheless , the general means of proceeding should be clear : for each anomalous coupling a computation of the polarized differential cross sections as a function of @xmath21 is required . by examining the functional form of these cross sections ( with the aid of contour plots in the @xmath50-@xmath51 plane ) , a basis may be constructed which highlights the deviations introduced by the anomalous coupling .
the decay angular distributions in this basis then encode the consequences of the new physics being investigated . in the standard model ,
three diagrams contribute to the total @xmath149 cross section .
we regroup these three diagrams into three contributions as follows .
first , we note that the two diagrams involving an @xmath43-channel photon or @xmath4 boson contribute for both possible chiralities of the initial fermion line .
thus , the first two contributions to the total cross section will be referred to as the @xmath150 and @xmath151 pieces .
is trivial in the sense that it shows up as a ( @xmath1 ) spin - independent prefactor : no choice of spin basis for the @xmath1 s can make these two contributions easier or harder to separate .
note that the contribution from the photon alone may be determined from eqs .
( [ prodwwleft ] ) and ( [ prodwwright ] ) by taking the @xmath4 mass to infinity , and dropping the last term of eq .
( [ prodwwleft ] ) . ]
the third diagram involves the @xmath74-channel exchange of a neutrino and contributes only when the initial fermion line is left - handed .
it will form the third contribution , which we will refer to as the neutrino piece . before presenting the differential cross sections for polarized production ,
let us consider the relative size of these three contributions to the total cross section . in table
[ diagcontribs ] , we have tabulated the values for each piece at tree - level for a machine energy of 192 gev . we see that the @xmath151 contribution is insignificant .
furthermore , there is a non - trivial interference term , which we will denote by @xmath152 , between the @xmath150 and neutrino contributions .
contribution and the left - handed pieces . ] in fig .
[ infzplot ] , we have plotted the squares of the three contributions as well as the interference term @xmath152 as a function of the zero momentum frame production angle @xmath19 , which is taken to be the angle between the momentum of the electron and the @xmath153 . although it is tempting to attribute the well - known peaking of the cross section in the forward direction to the @xmath74-channel pole in the neutrino diagram , fig .
[ infzplot ] indicates that this is not the case .
the square of the neutrino diagram is not only well - behaved near @xmath154 , but even _ decreases _ slightly in that region .
instead , the shape of the differential cross section is dominated by the interference between the @xmath150 and neutrino pieces , which changes sign from @xmath64 to @xmath63 for increasing @xmath50 .
these general features become even more pronounced at higher center - of - mass energies . to describe the polarized production cross sections for @xmath155 , we choose to identify the @xmath153 with particle @xmath17 and the @xmath156 with particle @xmath18 in fig .
[ xidef ] .
we now present the differential cross sections for polarized @xmath157 production . when the initial fermion pair has left - handed chirality , we have & & = ^2 & &
\ { 2(m_w / m_z)^2 - ( 1-^2)^2_w & & - 2 _ l^| ( , , ) } ^2 .
[ prodwwleft ] here @xmath25 ( @xmath158 ) is the spin of the @xmath153 ( @xmath156 ) .
the first term in the curly brackets is the @xmath159 contribution , while the second term comes from the neutrino .
for the other initial fermion chirality ( right - handed ) , we have simply = g_f^2 m_w^2 ^4_w ^2 .
[ prodwwright ] all of the spin information in contained in the @xmath160 , which are the same spin functions defined in connection with @xmath12 production , namely eqs .
( [ funcab])([leftright ] ) .
naturally , we should re - interpret @xmath26 as the common zmf speed of the two @xmath1 s .
the presence of the neutrino contribution to eq .
( [ prodwwleft ] ) makes the expression for the total differential cross section somewhat messy : _ , |,c d_c^| = & & & & & & ^3\ { 16 + ^2 } + g_f^2 m_w^2 & & \ { 2 + 12 ^ 2 ^ 2 ^ 2 + 2 ^ 2(1-^2)^2 } - & & & & \ { ( 1+^2)[4+^2 ^ 2 ^ 2 ] + 2(1-^2)[^2 ^ 2- 2(1-^2 ) ] } .
[ wwtotal ] the successive terms in eq .
( [ wwtotal ] ) come from the square of the @xmath159 contributions ( summed over both @xmath73 chiralities ) , the square of the neutrino contribution , and the interference term @xmath152 . as in the @xmath12 case ,
the individual spin - dependent amplitudes are complicated .
therefore , in fig . [ prodcontours ] , we have plotted the fractional contribution to each of the six independent spin configurations for a center - of - mass energy of 192 gev . from these plots
, we see that the @xmath161 $ ] and @xmath86 contributions never dominate the total amplitude at 192 gev , while the @xmath162 $ ] or @xmath163 contributions can be made large with the proper choice of @xmath21 . from the contour plots and the expressions for the spin components we have identified four interesting spin bases : helicity , , beamline , and 00b , which we have studied in some detail . in fig .
[ wwsolns ] , we have plotted the connection between @xmath21 and @xmath19 at 192 gev for these four bases . in tables
[ ww - breakdowns][infz - breakdowns ] , we indicate the fraction of the total cross section , @xmath159 , @xmath164 and @xmath152 contributions broken down into the six independent spin configurations for an @xmath73 collider running at 192 gev .
we begin our survey with the helicity basis .
this basis has one noteworthy feature : the @xmath163 and @xmath165 components of both @xmath159 contributions vanish at all @xmath41 .
this is not true for the neutrino diagram ; therefore , a measurement of these spin components is a direct measurement of the neutrino contribution . at @xmath166
gev the sum of these two spin components is 56% of the total cross section .
another useful basis is the basis , which is defined by choosing @xmath21 so that the @xmath92 component is as large as possible . in principle
, it is a straightforward exercise to derive an analytic expression connecting @xmath21 and @xmath19 by differentiating @xmath167 with respect to @xmath21 and setting the result equal to zero . in practice , however , such an expression would be too complicated to be illuminating .
furthermore , it turns out that we must allow negative values of @xmath168 in order to get the largest possible contribution .
this is equivalent examining both the @xmath163 and @xmath93 amplitudes , and taking the larger of the two at each point . selecting @xmath169 whenever the @xmath93 amplitude is chosen
converts it to a @xmath92 amplitude , according to eq .
( [ leftother ] ) . thus
, it is expedient to use a numerical procedure to determine @xmath21 in this basis : the results are plotted in fig .
[ wwsolns ] for @xmath123 .
the discontinuity seen in the @xmath51 versus @xmath50 curve for this basis near @xmath170 is caused by the presence of two competing maxima . at @xmath171
the @xmath92 component is 64% of the total in this basis . in the beamline basis ,
defined by eq .
( [ bmldef ] ) , the direction of the spin axis for the @xmath153 ( @xmath156 ) viewed in its rest frame coincides with the direction of the electron ( positron ) in that frame . with the choice ( [ bmldef ] ) for @xmath21 , the functions @xmath172 and @xmath173 vanish .
this means that the neutrino does not contribute to the @xmath174 $ ] , @xmath95 $ ] , and @xmath165 spin configurations @xcite .
unfortunately , the fraction of the total in these spin configurations is small , less than 5% at @xmath166 gev .
however , the @xmath175 $ ] component is 80% of the total at this @xmath41 .
finally , we come to the 00b basis , which is defined by choosing the value of @xmath21 which maximizes the fraction of the amplitude coming from the @xmath94 $ ] component .
once again the maximization condition leads to a complicated expression ( it is a quartic equation for @xmath176 ) .
therefore , we present the numerically derived solution in fig .
[ wwsolns ] . in this basis , the @xmath175 $ ] component is more than 92% of the total at @xmath177 gev . for these last three bases , , beamline , and 00b , there are large negative interference terms between the @xmath164 and @xmath150 contributions in all of the dominant spin components .
[ wwcomposcth ] shows the angular dependence of the polarized production cross sections .
it is clear from these plots that in the helicity basis , there is a complicated structure , with four different amplitudes being the largest , depending on the value of @xmath50 . on the other hand , in the basis , the @xmath87 component dominates at all angles .
likewise , in the beamline and 00b bases , the @xmath82 $ ] component dominates at all angles .
finally , because the relative contributions of the polarized cross sections depend upon the center - of - mass energy , we display this dependence in the range from threshold to 210 gev in fig .
[ wwcomposroots ] . as mentioned near the end of sec .
[ polzhprod ] , the effect of including initial state radiation in the results is to produce events at lower @xmath41 . from fig .
[ wwcomposroots ] , we see that the addition of such events would slightly increase the fraction of the dominant spin component in the beamline and 00b bases and slightly decrease the dominant component in the helicity and bases .
the 00b basis is slightly less sensitive to small shifts in @xmath41 than the beamline basis , while the basis is slightly less sensitive to small shifts in @xmath41 than the helicity basis .
the presence of maximal parity violation in the coupling between the @xmath1 boson and fermions results in distinctly different angular distributions for the @xmath1 decay products for all three @xmath1 polarizations .
in fact , the decay distributions in the rest frame of the decaying @xmath153 ( or @xmath156 ) are the same as for @xmath4 decay [ _ i.e. _ eqs .
( [ zt ] ) and ( [ zl ] ) ] with @xmath178 for all possible final states , leptonic and hadronic .
thus , they follow the @xmath179 curves presented in fig .
[ zdecaydist ] .
note , however , in the case of @xmath180 , it is more convenient to use the distribution in the _ antilepton_angle @xmath181 : this distribution may be generated by replacing @xmath128 with @xmath182 in eqs .
( [ zt ] ) and ( [ zl ] ) . because the distributions for transversely - polarized @xmath1 s are narrower than the one for ly - polarized @xmath1 s , it is easier to identify the @xmath63 and @xmath64 polarization states than the @xmath106 state . to demonstrate this ,
suppose we `` tag '' the spin of the parent @xmath153 based on the value of @xmath128 as follows : the spin is taken to be @xmath23 if @xmath183 , @xmath24 if @xmath184 , and @xmath106 if @xmath185 .
the probabilities that these assignments are correct may be computed from eqs .
( [ zt ] ) and ( [ zl ] ) , yielding ( 7 + 4y+y^2 ) , & & ( 3-y^2 ) , & & .
[ wtagging ] no matter what value is chosen for @xmath186 , the probability of correct identification in the central region is always less than 50% . on the other hand ,
the correct identification rate for the @xmath63 and @xmath64 states in their respective regions can be of order 75% .
we now consider correlations between the decay products of the @xmath157 pair .
for the sake of concreteness , let us consider the case where the @xmath153 decays to @xmath187 and the @xmath156 decays to @xmath188 .
let @xmath101 be the angle between the @xmath189 emission direction and the @xmath153 spin axis in the @xmath153 rest frame , and @xmath181 be the angle between the @xmath190 emission direction and the @xmath156 spin axis in the @xmath156 rest frame . for simplicity
we are assuming one can determine the direction of the two neutrinos .
if one of the @xmath1-bosons decays into two jets , then the correlations of the down - quark jet are the same as that of the charged lepton .
since determining which jet is the down - type jet is problematic , the distributions given here will have to be folded such that they are symmetric when @xmath191 .
the first set of correlations we wish to discuss may be conveniently displayed as a scatter plot in @xmath128 and @xmath192 . in fig .
[ wwdecaycontours ] we have plotted the predictions for this distribution using the helicity , , beamline , and 00b bases . on one hand ,
the 00b plot is very nearly the pure distribution for the decay of the @xmath82 $ ] spin state .
however , the broadness of the decay distributions for the longitudinal @xmath1 s cause the events to be rather spread out in this plot . on the other hand
, there is a nice contrast between the maximum and minimum values in the plot .
the functional dependence is less simple , however , as fully three of the six spin configurations contribute at or above the @xmath193 level .
although a precision measurement of the complete 2-dimensional distribution requires a large number of events , it is still possible to perform some interesting tests with fewer events .
for example , if @xmath14 is conserved , then the scatter plot will be symmetric about the line @xmath194 .
so a measure of the asymmetry in the number of events on either side of this line provides a simple test of @xmath14 symmetry . in fig .
[ wwinf ] , we have plotted the distribution of the interference terms @xmath135 ( as described in sec . [ zzcorrel ] ) for @xmath157 production and decay using the helicity , , beamline , and 00b bases . none of the bases under consideration ( nor any basis which we have found ) performs particularly well in this respect .
the basis contains the smallest interference terms on average , but there is still a non - negligible component away from @xmath136 .
this bump / peak at positive @xmath130 for all spin bases is caused by the strong azimuthal correlations associated with the pure @xmath195 coupling of the @xmath1-boson to its decay products .
a similar structure appears in the @xmath130 plots for @xmath36 in the situation where both @xmath4 s decay to @xmath179 .
although there is no obvious spin basis which is ideally suited to the study of @xmath1 pair production at lep , the basis is clearly better than the helicity basis in many ways .
the same can be said of the 00b basis compared to the beamline basis . a detailed study of the correlations in more than one basis is required to disentangle the spin correlations in this process .
if the higgs mass turns out to be in the vicinity of the @xmath4 mass , then the basis , in which the @xmath4 produced in association with he higgs is purely in the @xmath105 polarization states , provides a useful handle with which to distinguish between @xmath11 and @xmath12 events , based upon the angular distributions of the decay products .
in this situation , the helicity basis supplies no useful information . even if the higgs mass is significantly different from the @xmath4 mass , the basis still exists and leads to distinctive angular correlations among the decay products .
a comparison between the and helicity descriptions of the data is useful for distinguishing a scalar from pseudoscalar higgs . for the study of the @xmath13 events there is no ideal basis .
however , the basis is better than helicity , and the 00b basis is better than the beamline basis at describing the correlations . not only is there a greater fraction of the total cross section concentrated in the dominant components in these preferred bases , but the interference distributions are somewhat narrower .
the lack of an ideal basis implies that study of the @xmath13 system is best carried out using more than one basis , depending upon which quantity is being tested .
furthermore , simply checking to see that the correlations change in the correct manner as the basis is varied tests the standard model in ways which can not be accomplished with the helicity basis alone .
the fermi national accelerator laboratory ( fnal ) is operated by universities research association , inc .
, under contract de - ac02 - 76cho3000 with the u.s .
department of energy .
high energy physics research at mcgill university is supported in part by the natural science and engineering research council of canada .
we would like to thank peter fisher and philip bambade for illuminating discussions .
gm would like to thank the fnal theory group for their kind hospitality during visits to initiate and wrap - up this work .
sp would like to thank the aspen center for physics where a portion of this work was completed .
following kuijf@xcite , we consider a massive vector boson of momentum @xmath196 and mass @xmath197 such that @xmath198 .
let @xmath199 be the spin vector associated with this vector boson such that @xmath200 and @xmath201 .
we need the three polarization vectors @xmath202 such that in the vector boson rest frame the boson has spin projection @xmath203 with rest to the spatial part of the spin vector @xmath199 .
these three polarization vectors are conveniently written in terms of the two vectors v_1 = ( v+ms ) v_2 = ( v - ms ) .
note @xmath204 , @xmath205 and @xmath206 .
the decay products of a transverse vector boson are directly correlated with @xmath207 and @xmath208 , not @xmath196 and @xmath199 . using the spinor notation of mangano and parke@xcite , where |u(v_1 ) ( 1_5 ) ( 1_5 ) v(v_1 ) , we have _ ^ & = & ^ + _ 0^ & = & ^ - ^ ( = v_1 ^ -v_2^ ) .
the phase factors have been chosen such that @xmath209 is real , @xmath210 , and @xmath211 if we interchange @xmath207 and @xmath208 .
our convention is that these polarization vectors are for outgoing vector bosons whereas for incoming vector bosons we use @xmath212 .
the helicity basis is obtained by choosing the spatial part of the spin vector @xmath199 to be in the same direction as the spatial part of the momentum vector @xmath196 of the vector boson .
for example , if v = m ( 1 , ) using an obvious notation then choose v_1 = 12 m ( 1 + ) ( 1 , ) v_2 = 12 m ( 1- ) ( 1 , - ) .
bjorken , proc .
summer institute on particle physics , slac report 198 ( 1976 ) ; j. ellis , m.k .
gaillard , and d.v .
nanopoulos , nucl . phys . *
b106 * , 292 ( 1976 ) ; b.w .
lee , c. quigg , and h.b .
thacker , phys . rev . *
d16 * , 1519 ( 1977 ) ; b.l .
ioffe and v.a .
khoze , sov .
* 9 * , 50 ( 1978 ) .
gaemers and g.j .
gounaris , z. phys . *
c1 * , 259 ( 1979 ) ; g. gounaris , j. layssac , g. moultaka , and f.m .
renard , int . j. mod
a8 * , 3285 ( 1993 ) ; m. bilenky , j.l .
kneur , f.m .
renard , and d. schildknecht , nucl . phys . *
b409 * , 22 ( 1993 ) ; k. hagiwara , t. hatsukano , s. ishihara and r. szalapski , nucl . phys . *
b496 * , 66 ( 1997 ) . for a recent review of trilinear gauge boson couplings
see : h. aihara , _ et .
al . _ , in : _ electroweak physics and beyond the standard model _ , world scientific publishing co. , singapore , eds .
t. barklow , s. dawson , h. haber , and j. siegrist , hep - ph/9503425 .
ccddddc & spin & helicity & beamline & & & + & configuration & basis & basis & basis & basis & + & @xmath213 & 35.6% & 3.5% & 0.0% & 51.8% & + & @xmath214 & 32.2% & 55.4% & 57.4% & 24.1% & + & @xmath215 & 32.2% & 41.2% & 42.6% & 24.1% & + ccddddc & spin & helicity & beamline & & & + & configuration & basis & basis & basis & basis & + & @xmath86 & 11.7% & 5.6% & 0.0% & 0.1% & + & @xmath87 & 22.6% & 4.2% & 0.3% & 25.7% & + & @xmath88 & 22.6% & 2.4% & 0.2% & 25.7% & + & @xmath216 & 5.8% & 2.8% & 0.0% & 0.0% & + & @xmath217 & 18.7% & 54.6% & 64.1% & 24.2% & + & @xmath218 & 18.7% & 30.4% & 35.4% & 24.2% & + & @xmath219 & 30.4% & 48.1% & 49.7% & 24.4% & + & @xmath220 & 34.8% & 32.9% & 32.4% & 37.8% & + & @xmath221 & 34.8% & 19.0% & 17.9% & 37.8% & + ccccdcc & & fermion &
@xmath102 & value & & + & & @xmath222 & @xmath223 & 0.834 & & + & & @xmath224 & @xmath225 & 0.968 & & + & & @xmath226 & @xmath227 & 0.574 & & + & & @xmath164 & @xmath228 & 1.000 & & + & & @xmath229 & @xmath230 & 0.500 & & + cccddcc & & diagram(s ) & contribution & fraction & & + & & @xmath151 & 0.2 pb & 1.0% & & + & & @xmath150 & 13.8 pb & 70.6% & & + & & @xmath164 & 36.7 pb & 188.0% & & + & & inf[@xmath150,@xmath164 ] & @xmath6431.1 pb & @xmath64159.6% & & + & & total & 19.5 pb & & & + ccddddc & spin & helicity & @xmath87 & beamline & @xmath217 & + & configuration & basis & maximized & basis & maximized & + & @xmath86 & 8.9% & 10.4% & 11.6% & 2.2% & + & @xmath87 & 9.9% & 64.6% & 2.1% & 3.6% & + & @xmath88 & 46.2% & 2.9% & 0.6% & 0.6% & + & @xmath216 & 4.0% & 7.5% & 3.6% & 0.6% & + & @xmath217 & 9.9% & 0.1% & 79.7% & 92.5% & + & @xmath218 & 21.1% & 14.6% & 2.4% & 0.6% & + ccddddc & spin & helicity & @xmath87 & beamline & @xmath217 & + & configuration & basis & maximized & basis & maximized & + & @xmath86 & 26.8% & 11.0% & 1.9% & 2.1% & + & @xmath87 & 0.0% & 16.5% & 30.6% & 42.1% & + & @xmath88 & 0.0% & 1.2% & 0.6% & 0.6% & + & @xmath216 & 3.6% & 1.9% & 3.6% & 7.3% & + & @xmath217 & 20.6% & 36.6% & 32.5% & 18.6% & + & @xmath218 & 20.6% & 4.4% & 2.4% & 0.8% & + ccddddc & spin & helicity & @xmath87 & beamline & @xmath217 & + & configuration & basis & maximized & basis & maximized & + & @xmath86 & 43.2% & 6.3% & 10.8% & 0.1% & + & @xmath87 & 9.9% & 110.2% & 24.5% & 63.9% & + & @xmath88 & 46.2% & 5.6% & 0.0% & 0.0% & + & @xmath216 & 7.5% & 7.6% & 0.0% & 6.0% & + & @xmath217 & 32.0% & 36.5% & 152.7% & 118.0% & + & @xmath218 & 49.2% & 21.8% & 0.0% & 0.8% & + ccddddc & spin & helicity & @xmath87 & beamline & @xmath217 & + & configuration & basis & maximized & basis & maximized & + & @xmath86 & @xmath6461.1% & @xmath646.9% & @xmath641.1% & @xmath640.0% & + & @xmath87 & 0.0% & @xmath6462.2% & @xmath6453.1% & @xmath64102.5% & + & @xmath88 & 0.0% & @xmath643.9% & 0.0% & 0.0% & + & @xmath216 & @xmath647.1% & @xmath642.0% & 0.0% & @xmath6412.7% & + & @xmath217 & @xmath6442.8% & @xmath6473.0% & @xmath64105.4% & @xmath6444.1% & + & @xmath218 & @xmath6448.7% & @xmath6411.6% & 0.0% & @xmath640.3% & + | we apply a generalized spin - basis analysis to associated higgs production and gauge boson pair production at lep .
this framework allows us to identify a choice of spin axes for the processes @xmath0 which leads to strikingly different correlations among the decay products , even well above threshold .
this spin basis optimizes the difference in the angular correlations for these two processes . in contrast , the same distributions display little contrast when the helicity basis is used .
we also apply this technique to the case of @xmath1 boson pair production . 00b@xmath2$]-maximized # 1#2 # 1#2 # 1 # 1 # 1#2 |
in recent years , there has been some activity about approximate solutions of the euler and navier - stokes ( ns ) equations , viewed as tools to infer accurate _ a posteriori _ estimates on the exact solutions .
we mention , in particular : the works by chernyshenko et al .
@xcite , dashti and robinson @xcite , robinson and sadowski @xcite , and our papers @xcite @xcite @xcite .
the present work seats within the same research area ; here we consider the incompressible euler / ns equations = - ( u u ) + u + f , [ eul ] where : @xmath13 is the divergence free velocity field ; the space variables @xmath14 belong to the torus @xmath0 ( and yield the derivatives @xmath15 ) ; @xmath16 is the laplacian ; @xmath17 ( @xmath18 ) ; @xmath19 is the leray projection onto the space of divergence free vector fields ; @xmath20 is the viscosity coefficient , so that @xmath21 in the euler case and @xmath22 in the ns case ; @xmath23 is the leray projected density of external forces .
the dimension @xmath24 is arbitrary in the general setting of the paper , but we put @xmath25 in a final application . the functional setting that we consider for eq .
( [ eul ] ) relies on the sobolev spaces ^n _ ( ) ^n _ : = \ { v : ^d | v = 0 , v = 0 , ^n v ^2 ( ) } , with @xmath26 indicating the mean over @xmath0 ; for any real @xmath27 , the above space is equipped with the inner product @xmath28 and with the corresponding norm @xmath29 .
one of the main issues in this setting is the behavior of the bilinear map ( v , w ) : = - ( v w ) in the above mentioned sobolev spaces .
it is well known that there are positive constants @xmath30 and @xmath31 fulfilling the `` basic inequality '' ( v , w ) _
n k_n v _ n w _
n+1 , [ bbasineq ] and the so - called `` kato inequality '' | ( v , w ) | w _
n | g_n v _
n w ^2_n ; [ kkatineq ] fully quantitative upper and lower bounds on @xmath32 and @xmath33 were derived in our previous works @xcite @xcite , for reasons related to the present setting and described more precisely in the sequel . independently of the problem
to estimate @xmath32 and @xmath33 , the above two inequalities play a major role in the very interesting paper @xcite on approximate euler / ns solutions and _ a posteriori _ estimates on exact solutions .
to give an idea of the framework of @xcite we describe a result therein , using notations closer to our setting .
consider the euler / ns equation ( [ eul ] ) with a specified initial condition @xmath34 ; let @xmath35 be an approximate solution of this cauchy problem .
given @xmath2 ( and assuming suitable regularity for @xmath36 ) , let @xmath5 possess the differential error estimator @xmath37 , the datum error estimator @xmath38 and the growth estimators @xmath39 ; this means that , for @xmath40 , ( t + ( ) - - f ) ( t)_n _ n(t ) , ( 0 ) - u_0 _ n _
n , ( t ) _ n _ n(t ) , ( t ) _ n+1 _
n+1(t ) ( with @xmath41 , etc . ) . according to @xcite , eq .
( [ eul ] ) with datum @xmath42 has an exact ( strong , @xmath43-valued ) solution @xmath4 on a time interval @xmath44 , if @xmath45 ( with the estimators for @xmath5 ) fulfills the inequality _
n + _ 0^ dt _
n(t ) < 1 g_n e^. [ criter ] the present work aims to refine , to some extent , the approach of @xcite and to apply it to get fully quantitative estimates on the exact solution of the euler / ns cauchy problem on @xmath0 , with some specific initial datum .
our main result can be described as follows : assuming suitable regularity for @xmath36 , and intending @xmath46 as above , suppose there is a function @xmath47 , with @xmath48 $ ] , fulfilling the _ control inequalities _ - _
n + ( g_n _ n + k_n _
+ g_n ^2_n + _
n(0 ) _ n [ ciq ] ( with @xmath49 the right upper dini derivative , see section [ prel ] ) .
then , the solution @xmath4 of the euler / ns equation ( [ eul ] ) with initial datum @xmath42 exists ( in a classical sense ) on the time interval @xmath50 , and its distance from the approximate solution admits the bound u(t ) - ( t ) _ n _ n(t ) .
[ buap ] some features distinguishing our approach from @xcite are the following ones .
\(i ) differently from ( [ criter ] ) , our control inequalities ( [ ciq ] ) depend explicitly on @xmath20 and thus could allow a more accurate analysis of the influence of viscosity on the regularity of the euler / ns solutions .
\(ii ) our approach promises better lower bounds on the time of existence of @xmath4 .
for example , for @xmath51 and under specific assumptions illustrated in the paper , the inequalities ( [ ciq ] ) have solutions @xmath6 with @xmath52 , implying the global nature of the ns solution @xmath4 ; on the contrary , if @xmath53 or @xmath54 are nonzero the inequality ( [ criter ] ) can not have a solution with very large @xmath45 , since the right hand side is bounded by @xmath55 and thus vanishes for @xmath56 .
\(iii ) in @xcite there is not an explicit bound on the distance between @xmath4 and @xmath5 , such as ( [ buap ] ) ( however , our analysis yielding ( [ buap ] ) is greatly indebted to @xcite and , in a sense , it mainly refines and completes a chain of inequalities for @xmath57 appearing therein ) .
\(iv ) the constants @xmath32 and @xmath33 in the inequalities ( [ bbasineq ] ) ( [ kkatineq ] ) are not evaluated in @xcite .
on the contrary , here we have at hand our previous results @xcite @xcite on these constants ; thus , in specific applications , we can implement the control inequalities ( [ ciq ] ) and their outcome ( [ buap ] ) in a fully quantitative way . as an example of our approach , in the final part of the paper we consider the euler / ns equations on @xmath58 with a specific initial datum @xmath42 . independently of the approach developed here , this datum has been already considered in an interesting paper by behr , ne@xmath59as and wu @xcite , where it is indicated as the origin of a possible blow - up for the euler equations . however , in the cited work the blow - up is conjectured on the grounds of a merely `` experimental '' analysis of a finite number of terms in the power series @xmath60 solving formally the euler cauchy problem . in the present work ,
dealing with the initial datum of @xcite both for @xmath61 and for @xmath51 , a different approach to the cauchy problem is developed using the familiar galerkin approximation ( with a convenient set of fourier modes ) , combined with our general setting for approximate solutions based on the control inequalities ( [ ciq ] ) ; in this case , the sobolev order is @xmath62 , @xmath5 is the galerkin solution and we use for it the required estimators , to be substituted in the control inequalities ( [ ciq ] ) ( with the values for the constants @xmath63 and @xmath64 obtained in @xcite @xcite ) .
we search for a solution @xmath65 fulfilling eqs .
( [ ciq ] ) as equalities ( i.e. , with @xmath66 replaced by @xmath67 ) ; this gives rise to an ordinary cauchy problem for @xmath65 , which is solved very easily and reliably by numerical means .
admittedly , our computations are preliminary : they were performed using mathematica on a pc , with a fairly small set of @xmath68 fourier modes for the galerkin approximation ; we plan to develop the same approach with more powerful computational tools in a subsequent work . in a few words ,
our results are as follows : in the case @xmath61 , the solution @xmath65 of the control equations ( [ ciq ] ) exists on a finite time interval @xmath69 ( after which it blows up ) ; so , we can grant existence for the euler cauchy problem on the interval @xmath69 , where we also have the estimate ( [ buap ] ) on the @xmath70 distance between the exact solution @xmath4 and the galerkin approximate solution @xmath5 .
( unfortunately , @xmath3 is less than the blow - up time suggested in @xcite for the euler cauchy problem , so we can not disprove the conjecture of the cited paper ; the situation could change using many more galerkin modes , which is our aim for the future ) . for @xmath71 ,
the situation is similar : @xmath65 blows up in a finite time @xmath3 , and we can grant existence for the ns cauchy problem only up to @xmath3 . on the contrary , for @xmath72 , our approach grants global existence for the ns cauchy problem ( and a bound of the type ( [ buap ] ) on the full interval @xmath73 ) . to conclude this introduction ,
let us describe the organization of the paper .
in section [ prel ] we present some preliminaries : these concern mainly the sobolev spaces on @xmath0 , in view of their applications to the euler / ns equations ( [ eul ] ) . in section [ caupro ]
we define formally the euler / ns cauchy problem , the general notion of approximate solution for this problem and the related error estimators . in section [ secmain ] , that contains the main theoretical results of the paper , we develop the general framework yielding the control inequalities ( [ ciq ] ) , and prove the estimate ( [ buap ] ) on the distance between the exact solution @xmath4 of the euler / ns cauchy problem and an approximate solution @xmath5 ( here we also give more details on the connections of the present work with @xcite ) . in section [ analyt ]
we present some analytical solutions of the control inequalities ( [ ciq ] ) , under specific assumptions for their estimators and supposing , for simplicity , that the external forcing in ( [ eul ] ) is zero ; as anticipated , in certain cases our analytical solutions for the control inequalities are global , thus ensuring global existence for the euler / ns cauchy problem . in section [ galegen ]
we describe the general galerkin method for ( [ eul ] ) ; in particular , we give error and growth estimators for the galerkin approximate solutions , to be used with our control inequalities ( [ ciq ] ) . in section [ caunecas ]
we consider the galerkin method with the initial datum of @xcite , both for @xmath61 an for @xmath51 . in appendix [ appekapl ]
we review some comparison lemmas of the aplygin type about differential inequalities ; these are employed in section [ secmain ] in relation to the control inequalities . in appendix
[ appegal ] , for completeness we report the proof of an essentially known statement on the galerkin approximants for the euler / ns cauchy problem .
* dini derivatives . * consider a function : [ 0,t ) , t ( t ) ( with @xmath74 ) .
the _ right _ , lower and upper dini derivatives of @xmath75 at any point @xmath76 are , respectively , ( t_0 ) : = _ h 0^+ ( t_0 + h ) - ( t_0 ) h ; [ dlower ] ( t_0 ) : = _ h 0^+ ( t_0 + h ) - ( t_0 ) h . [ dupper ] of course , ( t_0 ) ( t_0 ) ; furthermore , the opposite function @xmath77 is such that ( t_0 ) = - d^+ d t(t_0 ) . the _ left _ , lower and upper dini derivatives @xmath78 , @xmath79 are defined similarly , with @xmath80 ; however , left derivatives are not used in this paper .
of course , all dini derivatives coincide with the usual derivative if this exists . * sobolev spaces of vector fields on the torus ; laplacian , leray projection , and so on .
* we work in any space dimension @xmath81 ( using @xmath82 as indices in @xmath83 ) . for @xmath84 in @xmath85
we put @xmath86 and write @xmath87 for the complex conjugate @xmath24-tuple @xmath88 ; @xmath85 carries the inner product @xmath89 and the norm @xmath90 .
we often restrict the previous operations to @xmath91 .
we consider the torus @xmath0 , i.e. , the product of @xmath24 copies of @xmath92 ; a point of @xmath0 is generically written as @xmath14 . in the sequel we often refer to the space ( ) of the real distributions on @xmath0 , and to the space ( ) : = \ { v = ( v_r)_r=1, ...
,d | v_r d } .
elements of @xmath93 can be interpreted as `` generalized functions @xmath94 '' ; in the sequel , we call them ( distributional ) vector fields on @xmath0 . @xmath95 and @xmath93 will be equipped with their weak topologies . for more details on distributions ( and on the function spaces mentioned in the sequel ) we refer , e.g. , to @xcite . using distributional derivatives
, we can give a meaning to several differential operators acting on vector fields , e.g. the laplacian @xmath96 and the divergence @xmath97 .
any real distribution @xmath98 has a mean @xmath99 ( the right hand side in this definition indicates the action of @xmath100 on the constant test function @xmath101 ) . passing to a vector field @xmath102 ,
we can define componentwise the mean @xmath103 .
each vector field @xmath102 has a unique ( weakly convergent ) fourier series expansion v = _
k v_k e_k , v_k ^d , e_k(x ) : = 1 ( 2 ) ^d/2 e^i k x ; [ fs ] of course , the @xmath104-th component of @xmath105 is @xmath106 ( i.e. , it equals the action of @xmath107 on the test function @xmath108 ) . due to the reality of @xmath109 , the fourier coefficients have the property @xmath110 for all @xmath111 ; one has @xmath112 . in the sequel
we often refer to the space of zero mean vector fields , of the divergence free ( or solenoidal ) vector fields and to their intersection ; these are , respectively , : = \ { v | v = 0 } , : = \ { v | v = 0 } , : = .elements of @xmath113 and @xmath114 are characterized , respectively , by the conditions @xmath115 and @xmath116 for all @xmath117 ; handling with the fourier components of vector fields in @xmath113 , it is convenient to put _ 0 : = \{0 } .
of course , for each @xmath102 one has @xmath118 ; this suggests to define , for any @xmath119 , ^n : , v ^n .
we denote with @xmath120 the space of square integrable vector fields @xmath121 ; this is a real hilbert space with the inner product @xmath122 @xmath123 and the norm @xmath124 . for any @xmath119 ,
let us consider the sobolev space ^n _ ( ) ^n _ : = \ { v | ^n v
^2 ( ) } = @xmath125 this is a real hilbert space with the inner product and the norm v | w _ n : = ^n v | ^n w _
l^2 = _ k _ 0 |k|^2 n w_k , v _ n : = = . clearly , @xmath126 implies @xmath127 and @xmath128 . in this paper
, we mainly fix the attention on the divergence free sobolev space ^n _ ( ) ^n _ : = ^n _ = \ { v ^n _
| k v_k = 0 } ( @xmath119 ) ; this is a closed subspace of @xmath129 , and thus a real hilbert space with the restriction of @xmath130 . for @xmath131 let us consider the space @xmath132 of vector fields @xmath121 of class @xmath133 in the ordinary sense
; if @xmath134 , equip this with the norm v _ c^m : = _ _ x |_s_1 ... s _ u_r(x)| .
there is a well known sobolev imbedding ^n _ ^m , _
n , [ imb ] involving suitable positive constants @xmath135 . for arbitrary @xmath119
, we have ^n+2 _ = ^n _ , ^n+2 _ = ^n _ , [ defa]and @xmath136 is continuous between the above spaces ; furthermore , v | v _
n = - v ^2_n+1 - v ^2_n [ inelap ] ( all the above statements are made evident by the fourier representations ) . by definition ,
the _ leray projection _ is the map : , v v ; here @xmath137 is the orthogonal projection of @xmath85 onto @xmath138 ( given explicitly by @xmath139 and @xmath140 , for @xmath141 and @xmath142 ) . for each real @xmath27
, we have a continuous linear map ^n _ : ^n _ ^n _ ( which , in fact , is the orthogonal projection of @xmath143 onto the subspace @xmath43 ) . * the fundamental bilinear map in the euler / ns equations . * in this subsection we assume @xmath144 . with this condition
, we have a continuous bilinear map @xmath145 , @xmath146 , where @xmath147 is the vector field on @xmath0 with components @xmath148 . by composition with the leray projection @xmath19
, we obtain a continuous bilinear map : ^n _ ^n+1 _ ^n _ , ( v , w ) ( v , w ) : = - ( v w ) .
this appears in the euler / ns equations , and is referred in the sequel as the fundamental bilinear map for such equations . as is known , for all @xmath149 and @xmath150 of fourier components @xmath105 and @xmath151 , @xmath152 has fourier components @xmath153 and ( v , w)_k = - i ( 2 ) ^d/2 _ h _ 0 v_h ( k - h ) _
k w_k - h [ repp ] for all @xmath141 , where @xmath137 is the already mentioned projection of @xmath85 onto @xmath154 .
the continuity of @xmath155 is equivalent to the existence of a constant @xmath156 such that ( v , w ) _
n w _ n+1 ; [ basineq ] we refer to this as the _ basic inequality _ about @xmath155 . with the stronger assumption @xmath157
, it is known that there is a constant @xmath158 such that | ( v , w ) | w _
n w ^2_n ; [ katineq ] we call this the _ kato inequality _ , since it originates from kato s seminal paper @xcite ( for completeness , we mention that @xmath159 for @xmath160 as above ) . in our previous works
@xcite @xcite , we derived upper and lower bounds for the sharp constants in the above inequalities ; throughout this paper , @xmath32 and @xmath33 are any two constants fulfilling eqs .
( [ basineq ] ) ( [ katineq ] ) .
from here to the end of section [ galegen ] we fix any space dimension @xmath81 , we consider the sobolev spaces of vector fields on @xmath0 , and we choose a real number @xmath27 such that n ( d 2 + 1 , + ) . [ propen ] * euler and ns equations : the cauchy problem .
* let us choose a `` viscosity coefficient '' [ 0,+ ) , [ vis]a `` forcing '' f c([0,+ ) , ^n _ ) [ for]and an initial datum u_0 ^n+2 _ .
[ in ] * definition .
* the cauchy problem for the ( incompressible ) fluid with viscosity @xmath20 , initial datum @xmath42 and forcing @xmath161 is the following : u c([0 , t ) , ^n+2 _ ) c^1([0,t ) , ^n _ ) [ cau ] @xmath162 ( with @xmath163 $ ] , depending on @xmath4 ) .
as usually , we speak of the `` euler cauchy problem '' if @xmath61 , and of the `` ns cauchy problem '' if @xmath164 .
* remark . *
( i ) the map @xmath165 sends continuously @xmath166 into @xmath43 , while @xmath136 sends continuously @xmath167 into @xmath43 ; this explains the appearing of @xmath167 in the previous definition , at least in the ns case . in the euler case @xmath61 , @xmath136 is absent from ( [ cau ] ) ; so , in the previous definition and in the subsequent theoretical developments one could systematically replace @xmath167 with @xmath166 ; this is not done just to avoid tedious distinctions between the euler and the ns case .
( ii ) consider any function @xmath168 . by the sobolev imbedding ( [ imb ] )
, one has @xmath169 .
[ obprp ] the following results are well known , and reported for completeness
. * proposition .
* for any @xmath170 and @xmath42 as above , ( i)-(iii ) hold .
\(i ) the cauchy problem ( [ cau ] ) has a unique maximal ( i.e. , nonextendable ) solution , hereafter denote with @xmath4 , with a suitable domain @xmath171 ( @xmath172 ) .
all the other solutions of ( [ cau ] ) are restrictions of @xmath4 .
( ii ) @xmath4 has the property ( [ obprp ] ) .
furthermore , if @xmath51 and the forcing @xmath173 is @xmath174 from @xmath175 to @xmath91 , the function @xmath176 is @xmath174 as well from @xmath177 to @xmath91 .
( iii ) for any @xmath178 consider the vorticity matrix @xmath179 , of elements @xmath180 ( @xmath181 ) ; let @xmath182 . if @xmath183 , one has _
t t^- ( t ) _ c^0 = + ; [ vor ] this implies _
t t^- u(t ) _
n = + . [ uninf ] * proof . *
( i ) see @xcite .
( ii ) the property ( [ obprp ] ) holds because @xmath184 . for @xmath51 and @xmath161 of class @xmath174 on @xmath175 , the same smoothness property is granted for @xmath4 by theorem 6.1 of @xcite ( with the domain @xmath185 considered therein replaced by @xmath0 ) .
( iii ) let @xmath186 .
( [ vor ] ) is the celebrated beale - kato - majda blow - up criterion @xcite ( see also @xcite ) . to prove ( [ uninf ] )
we note that , for each @xmath178 , ( t ) _
c^0 u(t ) _ c^1 u(t ) _
n ; [ notobs ] here the first inequality follows from the definition of @xmath179 in terms of the derivatives @xmath187 , and the second one from the sobolev imbedding ( [ imb ] ) ( with @xmath188 ) .
the relations ( [ vor ] ) and ( [ notobs ] ) immediately give eq .
( [ uninf ] ) .
* approximate solutions . *
our treatment uses systematically the present terminology .
[ defapp ] * definition .
* an approximate solution of the problem ( [ cau ] ) is any map @xmath189 ( with @xmath190 $ ] ) .
given such a function , we stipulate ( i ) ( ii ) .
\(i ) the differential error of @xmath5 is e ( ) : = d d t - - ( , ) - f c([0 , ) , ^n _ ) ; the datum error is ( 0 ) - u_0 ^n+2 _ .
( ii ) let @xmath191 .
a _ differential error estimator of order @xmath192 _ for @xmath5 is a function _
m c([0 , ) , [ 0,+ ) ) e()(t ) _ m _
m(t ) . let @xmath193 , @xmath194 .
a _ datum error estimator of order @xmath192 _ for @xmath5 is a real number _
m [ 0,+ ) ( 0 ) - u_0 _ m _ m ; a _ growth estimator of order @xmath192 _ for @xmath5 is a function _ m c([0 , ) , [ 0,+ ) ) ( t ) _ m _
[ din ] in particular the function @xmath195 , the number @xmath196 and the function @xmath197 will be called the _ tautological estimators _ of order @xmath192 for the differential error , the datum error and the growth of @xmath5 .
* assumptions and notations . * throughout this section we fix a viscosity coefficient , a forcing and an initial datum as in eqs .
( [ vis])([in ] ) ( recalling the condition ( [ propen ] ) @xmath198 ) .
we consider for the cauchy problem ( [ cau ] ) an approximate solution c([0 , ) , ^n+2 _ ) c^1([0 , ) , ^n _ ) ; in the sequel @xmath199 and @xmath38 are differential error and datum error estimators of order @xmath27 for @xmath5 , while @xmath200 are growth estimators of orders @xmath201 ( see definition [ defapp ] ) .
finally , we denote with u c([0 , ) , ^n+2 _ ) c^1([0 , ) , ^n _ ) the maximal solution of ( [ cau ] ) ( typically unknown , as well as @xmath202 ) . * some lemmas .
* for the sake of brevity , we put : = ( , ) , w : = u - c([0 , ) , ^n+2 _ ) c^1([0 , ) , ^n _ ) . [ tw ] furthermore , we introduce the function _
n : [ 0 , ) [ 0,+ ) , _
n(t ) : = w(t ) _
n ; [ defwn ] this is clearly continuous , and @xmath203 in a neighborhood of any instant @xmath204 such that @xmath205 . in the sequel we often consider the right , upper dini derivative @xmath206 ( see eq .
( [ dupper ] ) ) , that is just the ordinary derivative at any @xmath207 with @xmath205 . the forthcoming two lemmas review and partially refine some results of @xcite ( after adaptation to our slightly different setting ;
for example , dini derivatives are not even mentioned in @xcite ) .
* lemma . *
one has = w + ( , w ) + ( w , ) + ( w , w ) - e ( ) , [ ethen ] w(0 ) = u_0 - ( 0 ) .
[ ew0 ] * proof .
* by definition of the differential error @xmath208 , we have = + ( , ) + f + e ( ) ; [ wehaveuno ] of course , @xmath209 whence , writing @xmath210 , @xmath211 = + w + ( , ) + ( , w ) + ( w , ) + ( w , w ) + f .
[ eqlast ] subtracting eq .
( [ wehaveuno ] ) from ( [ eqlast ] ) we obtain eq .
( [ ethen ] ) ; eq .
( [ ew0 ] ) is obvious .
[ lemwn ] * lemma . *
consider the above mentioned estimators @xmath212 , and the function @xmath213 defined in eq .
( [ defwn ] ) . then - _
n + ( g_n _ n + k_n _
n + g_n ^2_n + _
n , [ ver1 ] _
n(0 ) _ n .
[ ver2 ] * proof .
* we proceed in three steps .
verification of eq .
( [ ver1 ] ) at a time @xmath207 such that @xmath205 . _ in this case , @xmath214 for all @xmath215 in some interval @xmath216 ; @xmath213 is nonzero and @xmath203 on @xmath217 . in the same interval
, we have : on the other hand , we have the following inequalities : w | w _ n - w ^2_n = - ^2_n ; [ fol1 ] ( , w)|w_n | ( , w)|w_n | g_n _ n w ^2_n g_n _
n ^2_n ; ( w , w)|w_n | ( w , w)|w_n | g_n w ^3_n = g_n ^3_n ; - e ( ) | w _ n |e ( ) | w _
n| e ( ) _ n w _
n _ n ; [ fol4 ] inserting the relations ( [ fol1])([fol4 ] ) into eq .
( [ wehave ] ) , we obtain the inequality ( [ ver1 ] ) at all times @xmath218 and , in particular , at time @xmath207 .
verification of eq .
( [ ver1 ] ) at a time @xmath207 such that @xmath219 .
_ at this instant we have @xmath220 , and the relation ( [ ver1 ] ) to be proved becomes ( t_0 ) _
n(t_0 ) . [ toprove ]
to go on we recall that , given a @xmath203 function @xmath221 with values in a banach space @xmath222 equipped with the norm @xmath223 , we have @xmath224 at all times , including instants when @xmath225 vanishes ( see , e.g. , @xcite and references therein ) ; if this result is applied to the function @xmath100 with norm @xmath226 , we get ( t_0 ) ( t_0 ) _ n . [ now1 ] now , we evaluate @xmath227 via eq .
( [ ethen ] ) , taking into account that @xmath219 ; this gives ( t_0 ) _
n = e()(t_0 ) _ n _ n(t_0 ) , [ now2 ] and eqs .
( [ now1 ] ) ( [ now2 ] ) yield the thesis ( [ toprove ] ) .
verification of eq .
( [ ver2 ] ) .
_ in fact , the definitions of @xmath228 and @xmath53 give to go on , we need a comparison lemma of the aplygin type .
this is a variant of some known results @xcite @xcite : for more details , see appendix [ appekapl ] .
[ lemtesto ] * lemma .
* let @xmath229 $ ] ; consider a function @xmath230 , @xmath231 possessing partial derivative @xmath232 ; furthermore , let @xmath233 .
suppose that @xmath234 are functions such that ( t ) ( ( t),t ) , ( 0 ) s_0 , ( t ) ( ( t),t ) , ( 0 ) s_0 . then ( t ) ( t ) .
[ eqtesto ] hereafter , interesting conclusions will arise from the combination of the above comparison result with lemma [ lemwn ] . * the control inequalities and the main theorem . *
we begin with a definition , which is followed by the main result of the section .
* definition . *
consider a function @xmath47 , with @xmath48 $ ] .
this function is said to fulfill the _ control inequalities _
( with respect to the estimators @xmath235 ) if - _
n + ( g_n _ n + k_n _
n + g_n ^2_n + _
n , [ cont1 ] _
n(0 ) _ n . [ cont2 ] [ main ] * proposition .
* suppose there is a function @xmath47 fulfilling the control inequalities , and consider the maximal solution @xmath4 of the euler / ns cauchy problem ( [ cau ] ) .
then , the existence time @xmath236 of @xmath4 is such that t ; [ tta ] furthermore , u(t ) - ( t ) _ n _ n(t ) .
[ furth ] * proof .
* let us employ the following notation , already used in the previous subsections : @xmath237 , @xmath238 and @xmath239 ( see eqs .
( [ tw ] ) ( [ defwn ] ) ; @xmath100 and @xmath213 have domain @xmath240 ) .
we further define @xmath241 ( and note that @xmath242 implies @xmath243 ) . for the moment we have not yet proved ( [ tta ] ) , so we do not know whether @xmath244 equals @xmath3 , or not .
we proceed in three steps .
u(t ) - ( t ) _ n _ n(t ) . [ urw ]
consider the functions @xmath213 and @xmath6 on @xmath245 .
due to lemma [ lemwn ] and to the control inequalities , these fulfill the relations therefore , by lemma [ lemtesto ] we have @xmath246 for @xmath247 ; this is just the relation ( [ urw ] ) .
one has the relation ( [ tta ] ) @xmath248_. to prove this , we assume @xmath249 and try to infer a contradiction . to this purpose
we note that , according to ( [ uninf ] ) , @xmath250 on the other hand , recalling eq .
( [ urw ] ) we can write @xmath251 @xmath252 @xmath253 @xmath254 for @xmath178 ; from here and from the continuity of @xmath255 , @xmath6 on @xmath256 $ ] we get _ t [ 0,t ) u(t ) _ n _ t ( ( t ) _
n + _ n(t ) ) < + . [ wehave2 ] the results ( [ uninf ] ) and ( [ wehave2 ] ) are contradictory .
conclusion of the proof_. eq .
( [ tta ] ) proved in step 2 implies @xmath257 ; now , the inequality ( [ urw ] ) of step 1 coincides with the thesis ( [ furth ] ) .
of course , the control inequalities ( [ cont1 ] ) ( [ cont2 ] ) are fulfilled by a function @xmath258 such that @xmath259 $ ] , and * definition . * eqs .
( [ econt1 ] ) ( [ econt2 ] ) are referred in the sequel as the the _ control equations _ , or the _ control cauchy problem _ for @xmath6 .
throughout this section we consider again the euler / ns cauchy problem ( [ cau ] ) , for a given viscosity @xmath260 , external forcing @xmath161 and initial datum @xmath42 as in ( [ vis])([in ] ) . after choosing an approximate solution @xmath5 for ( [ cau ] ) and determining the corresponding estimators ,
one is faced with the problem of solving the control inequalities ( [ cont1 ] ) ( [ cont2 ] ) , or the control equations ( [ econt1 ] ) ( [ econt2 ] ) , for the unknown real function @xmath261 . in many applications , such as the one of the next section on the galerkin approximate solutions , a numerical treatment of the control equations is recommended ( and the result is generally reliable : ( [ econt1 ] ) ( [ econt2 ] ) is a typically nonstiff cauchy problem ) . however , an analytic approach to the control equations / inequalities has its own interest , both for theoretical reasons and for building simple user - ready criteria . in this section
we propose an analytic approach for special forms of the approximate solution and/or its estimators . for simplicity , throughout the section we assume zero external forcing : f(t ) : = 0 ; however , many results presented in the section could be extended to the case of nonzero @xmath161 , with suitable assumptions on this function .
we begin by considering the approximate solution @xmath262 ; this choice seems to be very trivial but , in spite of this , it can be used to obtain nontrivial estimates on the time of existence of the exact solution of the euler / ns cauchy problem . in the ns case @xmath51 , these estimates include a condition for global existence under a fully quantitative norm bound on the initial datum . the second case considered in the section
is much more general : the approximate solution is unspecified , even though a certain form is assumed for its estimators .
* results from the zero approximate solution .
* let us choose for ( [ cau ] ) the approximate solution : [ 0,+ ) ^n+2 _ , ( t ) : = 0 .
[ lemzero ] * lemma . *
( i ) the differential and datum errors of the zero approximate solution are e()(t ) = 0 , ( 0 ) - u_0 = - u_0 ; consequently , this approximate solution has the differential and datum estimators _
n(t ) : = 0 , _ n : = u_0 _ n , [ estia ] and the growth estimators
_ n(t ) : = _ n+1(t ) : = 0 .
[ estib ] the control equations ( [ econt1 ] ) ( [ econt2 ] ) with the above estimators take the form = - _ n + g_n _
n(0 ) = u_0 _ n
. [ estic ] ( ii ) the cauchy problem ( [ estic ] ) has the solution @xmath263 , determined as follows : : = \ { + & , + - ( 1 - ) & + & .
[ ta ] ( intending @xmath264 if @xmath265 ) ; _ n(t ) : = u_0 _ n e^-t 1 - g_n u_0 _
n e_(t ) , [ ern ] e_(t ) : = \ { & , + t & . [ enu ] ( note that @xmath266 ) . * proof . *
( i ) obvious .
\(ii ) the cauchy problem ( [ estic ] ) is solved by the quadrature formula _
^ d r g_n r^2 - r = t ; the integral in the left hand side equals @xmath267 if @xmath268 , and @xmath269 if @xmath61 ; some elementary manipulations yield for the solution @xmath6 the expression ( [ ta])-([enu ] ) . from the previous lemma and from the main theorem on approximate solutions ( proposition [ main ] ) ,
here applied with @xmath270 and @xmath6 as in ( [ ern ] ) , we obtain the following result .
[ prozero ] * proposition . *
consider the cauchy problem ( [ cau ] ) for the euler / ns equations with zero external forcing , and any datum @xmath271 ; let @xmath272 be the maximal solution .
define @xmath3 and @xmath273 as in eqs .
( [ ta ] ) ( [ enu ] ) ; then t , u(t ) _
[ cobg ] in particular , due to ( [ ta ] ) , ; [ coglob ] in this case @xmath4 is global and , if @xmath51 , it decays exponentially .
* other sufficient conditions for global existence .
* consider again the cauchy problem ( [ cau ] ) with zero external forcing and any datum @xmath271 ; let @xmath272 be the maximal solution .
an obvious implication of proposition [ prozero ] is the following .
[ coro ] * corollary . *
assume u(t_1 ) _
[ assume ] then : t = + , u(t ) _
n , [ tes0 ] with @xmath273 as in eq .
( [ enu ] ) . * proof . *
the function @xmath274 is the maximal solution of the cauchy problem with datum @xmath275 _ specified at time @xmath276 _ , rather than at time @xmath277 . by eqs .
( [ cobg ] ) ( [ coglob ] ) , with an obvious shift in time , we obtain the thesis ( [ tes0 ] ) .
a consequence of the previous result , of more practical use , is the following .
[ coro2 ] * corollary .
* let @xmath189 be any approximate solution of ( [ cau ] ) , with estimators @xmath212 ; assume that the corresponding control inequalities ( [ cont1 ] ) ( [ cont2 ] ) have a solution @xmath258 , with @xmath48 $ ] .
finally , assume that ( _ n + _
n)(t_1 ) . [ iprn ] then : t = + ,
u(t ) _ n .
[ tes1 ] * proof .
* by proposition [ main ] we have @xmath7 for all @xmath8 and , in particular , for @xmath278 ; we further write @xmath279 , which implies u(t_1 ) _ n ( _ n + _
n)(t_1 ) . [ furth1 ]
( [ furth1 ] ) and the assumption ( [ iprn ] ) gives the inequality @xmath280 which has the form ( [ assume ] ) . by the previous corollary , we have eq . ( [ tes0 ] ) and this result , combined with ( [ furth1 ] ) , gives the thesis ( [ tes1 ] ) . * a general result , under conditions of exponential decay ( @xmath281 ) or boundedness ( @xmath282 ) for the approximate solution . * in this paragraph we exhibit a solution of the control equations ( [ econt1 ] ) ( [ econt2 ] ) , holding for _ any _ approximate solution whose estimators have certain features . such features are described in the forthcoming eq .
( [ dec1 ] ) ; these indicate that the norms of orders @xmath27 , @xmath283 of @xmath5 behave like @xmath284 , while the @xmath27-th norm of the differential error behaves like @xmath285 .
( in the ns case these are conditions of exponential decay , while in the euler case they simply indicate the boundedness of @xmath5 and its error . )
the assumption that @xmath5 behaves like @xmath284 corresponds to a rather typical behavior of the approximate solutions under zero external forcing ; for example , this behavior occurs in the case of the galerkin approximate solutions discussed in the next section ( see lemma [ rem64 ] and the subsequent remark ) .
the differential error of @xmath5 typically behaves like @xmath285 when @xmath5 is bounded by @xmath284 and the differential error depends only on the quadratic function @xmath286 ; again , this situation occurs in the example of the galerkin approximate solutions with no forcing ( see lemma [ lem67 ] ) .
other approximation methods suggested for the ns equations ( @xmath51 ) yield approximate solutions with a behavior as in ( [ dec1 ] ) ; as an example , this happens if one assumes no external forcing and takes for @xmath5 the truncation to any order of the power series solution introduced by sinai @xcite .
[ lemdec ] * lemma . * for any @xmath260 , consider for ( [ cau ] ) an approximate solution @xmath287 . assume this to have growth and differential error estimators of the forms _
n(t ) : = d_n e^-t , _
n+1(t ) : = d_n+1 e^-t , _
n(t ) : = e_n e^-2 t , [ dec1 ] with @xmath288 , @xmath289 ; furthermore , assume this to have any datum error estimator @xmath38 . from the above constants ,
let us define _ n : = 1 2 ( g_n d_n + k_n d_n+1 ) ; [ deden ] furthermore , assume the `` error bound '' g_n e_n < ^2_n [ erbound ] and define w^_n : = _
[ dewn ] finally , consider the function @xmath290 defined by ( [ enu ] ) , and put _
n : [ 0,+ ) [ 1,+ ) , _
n(t ) : = e^ .
[ etn ] then , ( i ) ( ii ) hold .
\(i ) the control equations ( [ econt1 ] ) ( [ econt2 ] ) with the above estimators take the form = - _ n + 2 f_n e^-t _
n + g_n _ n^2 + e_n e^-2 t , _
[ esticc ] ( ii ) the cauchy problem ( [ esticc ] ) has the solution @xmath263 , determined as follows : : = \ { + & , + - ( 1 - ) & , + & . [ tb ] _
n(t ) : = 1g_n w^+_n ( w^-_n + g_n _ n ) _
n(t ) - w^-_n ( w^+_n + g_n _ n ) ( w^+_n + g_n _ n ) - ( w^-_n + g_n _ n ) _
n(t ) e^-t .
[ eqrn ] * proof . *
( i ) obvious .
\(ii ) we write the unknown solution @xmath6 of ( [ esticc ] ) as _
n(t ) = _ n(t ) e^-t where @xmath291 ( and @xmath3 ) are to be found .
( [ esticc ] ) is equivalent to the cauchy problem = ( e_n + 2 _ n _ n + g_n _
n^2 ) e^-t , _
n , [ estizz ] with @xmath292 as in ( [ deden ] ) ; this is solved by the quadrature formula _ _
n^_n(t ) = _ 0^t d s e^-s .
[ estiqq ] on the other hand _ 0^t d s e^-s = e_(t ) ; [ imp1 ] furthermore @xmath293 , which implies _ _
n^_n(t ) d z e_n + 2 _ n z + g_n z^2 = - 1 w^+_n - w^-_n ^z=_n(t)_z=_n . [ imp2 ]
inserting eqs .
( [ imp1 ] ) ( [ imp2 ] ) into eq .
( [ estiqq ] ) we easily obtain an explicit expression for @xmath294 which implies eq .
( [ eqrn ] ) for @xmath295 .
the interval @xmath50 where @xmath296 is well defined is also easily determined , by the same manipulation employed to make explicit @xmath294 . from the previous lemma and from the main theorem on approximate solutions ( proposition [ main ] ) , we obtain the following result .
[ prodec ] * proposition . *
consider the euler / ns cauchy problem ( [ cau ] ) , and an approximate solution @xmath287 ; assume this to have growth and differential error estimators fulfilling all conditions in lemma [ lemdec ] , and define @xmath3 , @xmath6 via eqs .
( [ tb ] ) ( [ eqrn ] ) of the same lemma .
now , consider the maximal solution @xmath272 of ( [ cau ] ) ; then t , u(t ) - ( t ) _
( in particular , @xmath4 is global under the conditions giving @xmath52 in eq .
( [ tb ] ) . )
* one could easily derive a variant of lemma [ lemdec ] ( and proposition [ prodec ] ) dealing with the limit case @xmath297 , where @xmath298 .
the expressions for @xmath3 and @xmath6 in this case coincide with the ones derivable from eqs .
( [ tb ] ) ( [ eqrn ] ) in the limit @xmath299 .
the results in lemma [ lemzero ] ( and proposition [ prozero ] ) could also be derived from lemma [ lemdec ] ( and proposition [ prodec ] ) , in the limit case @xmath300 and @xmath301 .
throughout this section , we consider a set @xmath302 with the following features : _ 0 , , k - k .
[ modesg ] hereafter we write @xmath303 for the linear subspace of @xmath113 made of the sums @xmath304 ( which is , in fact , contained in @xmath305 ) .
the forthcoming definitions follow closely the setting proposed in our previous work @xcite for the galerkin method . * galerkin subspaces and projections . *
we define them as follows . * definition . *
the galerkin subspace and projection corresponding to @xmath306 are , respectively : : = e_k _ k g = \ { _ k v_k e_k | v_k ^d , = v_-k , k v_k = 0 } ; [ hg ] : , v = _ k _ 0 v_k e_k v : = _ k v_k e_k . [ pg ] it is clear that ^ _ ; ( ) = ; ^m _ , ( ^m _ ) = .
the following result is useful in the sequel
. * lemma .
* let @xmath307 , @xmath308 and @xmath309
. then , ( 1 - ) v _ m , || : = _ k _ 0 * proof .
* we have @xmath310 ; thus , ( 1 - ) v ^2_m = _ k _ 0 g
|k|^2 m | v_k |^2 = _ k _ 0 g |k|^2 p | k @xmath311 * galerkin approximate solutions .
* let us be given @xmath260 , @xmath312 and @xmath313 . * definition .
* the _ galerkin approximate solution _ of the euler / ns equations corresponding to the datum @xmath42 , to the forcing @xmath161 and to the set of modes @xmath302 is the maximal ( i.e. , nonextendable ) solution @xmath314 of the following cauchy problem , in the finite dimensional space @xmath315 : [ galer ] @xmath316 eq . ( [ galer ] ) describes the cauchy problem for an ode in the finite - dimensional vector space @xmath315 ; this relies on the continuous function @xmath317 , @xmath318 , which is smooth with respect to the variable @xmath109 .
so , the standard theory of odes in finite dimensional spaces grants existence and uniqueness for the solution of ( [ galer ] ) .
the following facts are known ( see , e.g. , @xcite ) ; a proof suitable for the present setting is given , for completeness , in appendix [ appegal ] .
[ rem64 ] * lemma . *
( i ) for any @xmath260 , @xmath312 and @xmath313 , the maximal solution of problem ( [ galer ] ) is in fact global : = + . [ tglob ] ( ii ) if the forcing @xmath161 is identically zero , one has ( t ) _ l^2 \ { = u_0 _ l^2 & + u_0 _ l^2 e^-t & . [ stacit ] ( iii )
let @xmath260 .
for any @xmath312 one has ( t ) _
l^2 ( u_0 _ l^2 + _ 0^t d s e^s f(s ) _
l^2 ) e^-t .
[ comb1 ] in particular , let j : = _ 0^+ ds e^s f(s ) _
l^2 < + ; [ comb2 ] then , ( t ) _ l^2 ( u_0 _ l^2 + j ) e^-t . [ comb3 ] * remark .
* let @xmath260 . if @xmath319 or , more generally , if @xmath320 , the previous lemma gives a bound of the type @xmath321 const .
this fact , with the equivalence of all norms on the finite dimensional space @xmath315 , implies the following : for any real @xmath192 , ( t ) _
m u_m e^-t , [ teorest ] with a suitable constant @xmath322 ( depending on the initial datum and on the forcing ) .
* definition . *
( i ) for all @xmath141 , @xmath323 and @xmath324 are the fourier components of the forcing and the initial datum : @xmath325 , @xmath326 .
\(ii ) for @xmath327 , @xmath328 are the fourier components of @xmath329 : ( t ) = _ k g _
k(t ) e_k .(note the relations @xmath330 , @xmath331 , and the analogous relations for @xmath332 , @xmath333 . )
let us review a well known fact .
* proposition .
* under the correspondence @xmath334 , eq .
( [ galer ] ) for @xmath329 is equivalent to the following problem : find a family of functions @xmath335 ( @xmath327 ) such that = - |k|^2 _ k - i ( 2 ) ^d/2 _ h g [ _ h ( k - h ) ] _ k _ k - h + f_k , _
k(0 ) = u_0 k [ galecomp ] ( intending @xmath336 if @xmath337 ; recall that @xmath137 is the orthogonal projection of @xmath85 onto @xmath338 ) . a family of functions @xmath339 ( @xmath327 ) fulfilling eqs .
( [ galecomp ] ) automatically possesses the properties @xmath340 and @xmath341 .
* proof . * clearly , eq .
( [ galer ] ) is equivalent to @xmath342 making explicit the above fourier components via eq .
( [ repp ] ) , we obtain eqs .
( [ galecomp ] ) .
let us show that eqs .
( [ galecomp ] ) imply @xmath340 and @xmath341 .
the first implication is proved noting that the functions @xmath343 and @xmath344 are solutions of the same cauchy problem .
the second implication follows noting that , for each @xmath327 , @xmath345 and @xmath346 ; the solution of this cauchy problem is identically zero . * the galerkin solutions in the general framework for approximate solutions . * from now on we stick to the framework of the previous sections , i.e. : we consider the sobolev spaces of orders @xmath347 , for a fixed @xmath348 ; as in ( [ vis])([in ] ) , we choose a viscosity @xmath260 , a forcing @xmath349 and an initial datum @xmath271 ; we consider the euler / ns cauchy problem ( [ cau ] ) .
having fixed a set of modes @xmath302 as in ( [ modesg ] ) , we regard the corresponding galerkin solution @xmath329 as an approximate solution of the euler / ns cauchy problem ( [ cau ] ) ; our aim is to apply the general theory of the previous sections with @xmath350 . the desired application requires to give growth and error estimators for @xmath329 .
of course , we have the tautological growth estimators _
m(t ) : = ( t ) _ m = , to be used with @xmath351 or @xmath352 ; these are employed systematically in the sequel .
let us pass to the errors of @xmath329 and their estimators .
[ lem67 ] * lemma . *
( i ) the galerkin solution @xmath329 has the datum error ( 0 ) - u_0 = -(1 - ) u_0 = - _ k g u_0 k e_k [ ovv1 ] and its tautological estimator _
n : = ( 0 ) - u_0 _ n = .
[ ovv2 ] there is a rougher estimator ( 0 ) - u_0 _
n _n p , _n p : = u_0 _ p
|g|^p - n , [ ovv3 ] where @xmath353 is any real number such that @xmath354 , @xmath355 .
\(ii ) the differential error of @xmath329 is e ( ) = - ( 1 - ) ( , ) - ( 1 - ) f ; [ prov1 ] the fourier representation of the above summands is ( 1 - )
f = _ k _ 0 g f_k e_k ; [ prov2 ] ( 1 - ) ( , ) = _ k dg p_k e_k , [ prov3 ] @xmath356 \lp_k \ga_{k - h}~.\ ] ] ( in the above : @xmath357 ; @xmath358 is the usual set - theoretical difference ; again , we intend @xmath359 if @xmath360 . )
\(iii ) the above terms in @xmath361 have norms ( 1 - ) f _
n = , [ norfa ] ( 1 - ) ( , ) _ n = ; [ nora ] these admit the bounds ( 1 - ) f _
n , [ norfb ] ( 1 - ) ( , ) _
n _ q _ q+1 [ norb ] where : @xmath353 is any real number such that @xmath354 and @xmath362 ; @xmath363 is any real number such that @xmath364 ; @xmath365 is such that @xmath366 for all @xmath367 , @xmath368 ( of course @xmath369 , and similarly for @xmath370 ) .
thus , a differential error estimator @xmath371 of order @xmath27 for @xmath329 is obtained setting _
n : = + .
[ estim ] * proof . *
( i ) eqs .
( [ ovv1 ] ) ( [ ovv2 ] ) are obvious .
( [ ovv3 ] ) follows writing @xmath372 and using the inequality ( [ egmp ] ) .
\(ii ) we have @xmath373 where the first equality is just the definition of the differential error , and the second follows from ( [ galer ] ) ; so , eq . ( [ prov1 ] )
is proved .
the fourier representation ( [ prov2 ] ) for @xmath374 follows immediately from the definition of @xmath375 . to derive the representation ( [ prov3 ] ) for @xmath376 ,
let us consider the fourier components of @xmath329 that we indicate in any case with @xmath333 , intending @xmath377 for @xmath378 .
( [ repp ] ) gives @xmath379 where , for any @xmath117 , @xmath380 is defined following eq .
( [ prov3 ] ) . on the other hand : for each @xmath141 , @xmath380 is a sum over @xmath381 containing terms of the form @xmath382 ; if @xmath383 , for all @xmath381 we have @xmath337 ( since @xmath384 would imply @xmath385 ) ; @xmath386 implies @xmath387 . due to the above considerations we have @xmath388 for @xmath383 , whence ( , ) = _ k ( g + g ) \{0 } p_k e_k .
application of @xmath389 removes from the above sum the terms with @xmath327 ; thus ( 1 - ) ( , ) = _ k ( g + g ) ( g \{0 } ) p_k e_k , in agreement with eq .
( [ prov3 ] ) .
\(iii ) eqs .
( [ norfa ] ) ( [ nora ] ) are straightforward consequences of the fourier representations ( [ prov2 ] ) ( [ prov3 ] ) .
( [ norfb ] ) follows immediately from ( [ egmp ] ) . to infer eq .
( [ norb ] ) , we note that ( [ egmp ] ) and the inequality involving @xmath390 imply @xmath391 finally , eq . ( [ prov1 ] ) implies e ( ) ( 1 - ) ( , ) _
n + ( 1 - ) f _ n ; binding the two summands in the right hand side via eqs .
( [ nora ] ) or ( [ norb ] ) , ( [ norfa ] ) or ( [ norfb ] ) we obtain the estimator @xmath371 in ( [ estim ] ) . * remarks . *
( i ) the exact expression ( [ nora ] ) for @xmath392 involves a sum over the _ finite _ set @xmath393 ; for each @xmath394 , the term @xmath380 in the above sum is itself a finite sum , that can be computed explicitly from the fourier coefficients @xmath395 of the galerkin approximate solution ( assuming these ones to be known , say , from the numerical integration of the system ( [ galecomp ] ) ) .
however , when @xmath302 is large the set @xmath393 is typically very large , and this can make too expensive the computation of the sum over @xmath393 . in these cases
one can use for @xmath392 the bound ( [ norb ] ) ; this requires computation of @xmath369 and of @xmath370 , both of them less expensive since these sums are over @xmath302 , rather than @xmath393 .
( ii ) we are aware that , if the galerkin equations ( [ galecomp ] ) are solved numerically by some standard method for odes , one does not obtain the exact solutions @xmath396 ( @xmath327 ) but , rather , some approximants whose distance from the @xmath333 s should be estimated on the grounds of the employed numerical scheme . in the application presented in the next section ,
relying on a relatively small set @xmath302 of modes , the intrinsic error in the numerical integration of ( [ galecomp ] ) has been regarded as negligible ; the situation would be different using a much larger set of modes .
* a preliminary . * in this section we frequently report the results of computations performed with mathematica .
a formula like @xmath397 must be intended as follows : computation of the real number @xmath104 via mathematica produces as an output @xmath398 , followed by other digits not reported for brevity . * setting up the problem . * throughout this section , we work in space dimension d = 3 , with any viscosity @xmath260 .
we consider the euler / ns equations on @xmath58 with no external forcing , in the sobolev framework of order @xmath62 .
so , the cauchy problem ( [ cau ] ) takes the form the initial datum @xmath42 in @xmath399 ( in fact , in @xmath400 for any real @xmath192 ) is chosen of this form : for @xmath61 , this initial datum has been considered by behr , ne@xmath59as and wu @xcite as the origin of a possible blow - up for the euler equations . more precisely , the authors of the cited work try a solution @xmath4 of the euler equations in the form of a power series @xmath401 , where the zero order term corresponds to the initial datum , and @xmath402 are determined recursively . by a means of @xmath403 program
, the authors compute the terms @xmath404 for @xmath405 , allowing to construct the partial sums @xmath406 up to @xmath407 .
a merely `` experimental '' analysis of these partial sums and of their @xmath70 norms brings the authors to conjecture that the exact solution @xmath408 of the euler cauchy problem blows up in @xmath70 for @xmath409 , for some @xmath410 . for subsequent use
, we mention that eq .
( [ unec ] ) implies u_0 _ m = for any real @xmath192 .
this formula gives , for example , u_0 _ 1 = 77.15 ... , u_0 _ 2 = 109.1 ... , [ u03 ] @xmath411 * introducing our approach . * in this section we propose a different approach to the euler cauchy problem of @xcite , that we apply as well to the ns case ; so , we consider the problem ( [ caun ] ) ( [ unec ] ) , for arbitrary @xmath260 .
we refer to the general setting developed in the present paper , using the spaces @xmath43 , @xmath412 , @xmath167 with n = 3 .
this requires , amongst else , the numerical values of two constants @xmath63 , @xmath64 fulfilling for @xmath62 the `` basic inequality '' ( [ basineq ] ) and the `` kato inequality '' ( [ katineq ] ) . due to the computations in @xcite
@xcite , these can be taken as follows : k_3 = 0.323 , g_3 = 0.438 .
[ k3g3 ] to illustrate our setting , we start with a very elementary result . * a simple sufficient condition for global existence ( and exponential decay ) . *
according to proposition [ prozero ] , the solution @xmath4 of the ns cauchy problem ( [ caun ] ) ( [ unec ] ) is global and exponentially decaying for @xmath413 , if g_3 u_0 _ 3 = 67.58 ... [ glosimp ] ( in the last passage , we have used the numerical values in ( [ u03 ] ) ( [ k3g3 ] ) for @xmath64 and @xmath414 ) ( , discussed in @xcite ; this can be written as @xmath415 , where @xmath42 is an arbitrary initial datum in @xmath416 . with the datum @xmath42 in ( [ unec ] ) ( and the value of @xmath417 in ( [ u03 ] ) )
, we conclude that @xcite ensures global existence in @xmath416 for @xmath418 . by a known regularity theorem about ns equations
@xcite , this result of global existence in @xmath416 also implies global existence in @xmath70 ; however , the condition @xmath419 arising from the @xmath420 setting of @xcite is manifestly weaker than the condition ( [ glosimp ] ) . ] ) . as shown hereafter , a more refined application of our setting for approximate solutions significantly improves the above condition : in fact , in the next pages , combining this setting with the galerkin method we will be able to infer global existence for @xmath72 .
* going on in our approach : galerkin approximants .
* the idea developed in the sequel , both in the euler case ( @xmath61 ) and in the ns case ( @xmath164 ) , is the following : to compute numerically the galerkin approximate solution @xmath329 for a suitable set of modes @xmath302 ; to construct for it error and growth estimators , in the sobolev norms of orders @xmath421 or @xmath422 ; to solve numerically the control equations ( [ econt1 ] ) ( [ econt2 ] ) for an unknown function @xmath423 .
after finding the solution of this control problem , we can grant on theoretical grounds that the solution @xmath4 of the euler / ns cauchy problem ( [ caun ] ) exists at least up to time @xmath3 , and that @xmath424 for @xmath8 .
our computation has been performed using mathematica on a pc , with the relatively small set of @xmath68 modes the results we present here are somehow provisional ; we plan to attack the problem by more powerful numerical tools in a future work , using for @xmath302 a larger set of modes . * a sketch of the operations to be done . *
the list of such operations is the following : \(i ) first of all , one chooses a value @xmath260 for the viscosity , and a finite time interval @xmath425 for the numerical computation of the galerkin solution .
\(ii ) the galerkin solution @xmath329 for the set of modes @xmath302 in ( [ defg ] ) and for the initial datum @xmath42 is found numerically on the chosen time interval @xmath426 .
more precisely , one solves numerically the system of differential equations ( [ galecomp ] ) for the fourier components @xmath427 of @xmath329 , with the initial conditions @xmath332 corresponding to eq .
( [ unec ] ) .
we recall ( see lemma [ rem64 ] ) that the general theory of the galerkin solutions for zero external forcing ensures the global existence of @xmath329 ( and its exponential decay , if @xmath51 ) ; thus , from a theoretical viewpoint there is no obstruction to the computation of @xmath329 on any finite interval @xmath425 .
\(iii ) we apply to @xmath329 on @xmath425 the framework of the present paper for the approximate euler / ns solutions using ( we repeat it ) the spaces @xmath43 , @xmath166 and @xmath167 with @xmath428 .
\(iv ) some important characters in our approach are the norms _
n(t ) : = ( t ) _
n = ( n=3,4 ) , [ d34 ] which can be obtained from the numerical values @xmath396 of the fourier components .
further , we need a datum error estimator @xmath429 and a differential error estimator @xmath430 . on the other hand
the datum error is zero in this case , since the initial condition @xmath42 is in the subspace @xmath315 spanned by the chosen set ( [ defg ] ) of modes ; thus , we can take _ 3 = 0
.as for the differential error estimator , we take the precise expression coming from eqs .
( [ nora ] ) ( [ estim ] ) ( taking into account that the forcing @xmath161 is zero in this case ) ; these yield the expression _
3(t ) = [ nora3 ] where @xmath431 and @xmath432 \lp_k \ga_{k - h}(t)$ ] , as in eq .
( [ prov3 ] ) .
the computation of @xmath433 following eq .
( [ nora3 ] ) is rather expensive : the set @xmath393 consists of @xmath434 modes and , for each one of them , one must perform the nontrivial computation of @xmath435 .
( for these reasons , a computation with a set of modes much larger than the @xmath302 in ( [ defg ] ) would suggest to replace this @xmath430 with a rougher estimator , coming from eqs .
( [ norb ] ) ( [ estim ] ) ) .
\(v ) now we consider the control equations ( [ econt1 ] ) ( [ econt2 ] ) , taking the form = - _ 3 + ( g_3 _ 3 + k_3 _ 4 ) _ 3 + g_3 ^2_3 + _ 3 , [ econt13 ] _
3(0 ) = 0 , [ econt23 ] with @xmath436 as in eq.([k3g3 ] ) and @xmath437 as in eqs .
( [ d34 ] ) ( [ nora3 ] ) .
the unknown is a function @xmath438 , with @xmath439 ; this is determined numerically ( with a package for self - adaptive integration , allowing to detect a possible blow - up of @xmath65 ; in this case , @xmath3 is the blow - up time ) .
once we have @xmath65 , the general theory allows to state the following .
\(a ) the maximal solution @xmath4 of the euler / ns cauchy problem ( [ caun ] ) ( [ unec ] ) has a domain containing @xmath50 , and u(t ) - ( t ) _ 3 _ 3(t ) [ ubr3 ] ( see proposition [ main ] ) .
( b ) suppose @xmath51 , and ( _ 3 + _
3)(t_1 ) .
[ iprn3 ] then , the solution @xmath4 of the ns cauchy problem ( [ caun ] ) is global , and u(t ) _ 3 , [ tes ] with @xmath273 as in eq .
( [ enu ] ) ( see corollary [ coro2 ] ) .
hereafter , we report the results obtained from the above scheme for some choices of @xmath20 .
* case @xmath282 * .
first of all , the galerkin equations ( [ galecomp ] ) have been solved for the set of modes ( [ defg ] ) on a time interval of length @xmath440 .
( this required , approximately , @xmath441 seconds of cpu time on our machine . ) in figures 1a , 1b and 1c we report , as examples , the graphs of @xmath442 for @xmath443 , in the cases @xmath444 , @xmath445 and @xmath446 , respectively .
( of course , we could consider many alternative choices , such as plotting the norms @xmath442 for other modes , or the real parts of the components @xmath447 ( @xmath448 ) , or the imaginary parts of the same components . ) as a supplementary information , we mention that the graphs of @xmath442 are identical in the three cases @xmath449 ) .
after computing numerically all the @xmath450 ( for @xmath327 ) , one obtains from eqs .
( [ d34 ] ) and ( [ nora3 ] ) the norms @xmath451 ( @xmath452 ) and the differential error estimator @xmath433 , for @xmath443 ; figures 1d and 1e report the graphs of @xmath453 and @xmath433 ( to give some more detail , we mention that the functions @xmath453 , @xmath454 and @xmath433 have been computed for a set of sample values of @xmath443 , and then interpolated by means of the mathematica algorithms .
the computation of @xmath433 at each sample value @xmath215 is rather expensive , for it involves a sum on the large set @xmath393 ; the cpu time is about @xmath441 seconds for each @xmath215 and , for this reason , we have used only @xmath455 sample values ) .
the final step is the numerical solution of the cauchy problem ( [ econt1 ] ) ( [ econt2 ] ) for the unknown function @xmath456 ( a task performed almost instantaneously by our pc ) .
the mathematica self - adaptive routines for odes indicate a divergence of @xmath456 for @xmath457 , with @xmath458 .
figure 1f gives the graph of @xmath456 for @xmath459 .
the conclusion of these computations is the following : the solution @xmath4 of the euler cauchy problem ( [ caun ] ) ( [ unec ] ) has a domain containing @xmath69 , and its distance from @xmath329 is bounded by @xmath65 on this interval . * some remarks on the @xmath282 case .
* we have already mentioned that the blow - up of @xmath408 is conjectured in @xcite for @xmath460 .
the basis of this conjecture is an experimental analysis of the first 35 terms in the power series ( in time ) solving formally the cauchy problem ; in principle , this analysis does not even prove existence of @xmath408 at _ any _ specified time @xmath461 .
our lower bound @xmath458 for the interval of existence of @xmath4 is about @xmath462 of the blow - up time @xmath463 suggested by @xcite , but it relies on an analytic theory for approximate solutions , their errors , etc . , summarized by proposition [ main ] ; in this sense , it is theoretically grounded .
perhaps , our method could give a sensibly larger lower bound @xmath3 on the time of existence , when implemented with a set @xmath302 of galerkin modes much larger than ( [ defg ] ) .
alternatively , one could apply our theoretical framework using as an approximate solution the partial sum @xmath464 , with @xmath407 as in @xcite or with a larger @xmath465 .
both tasks require much more expensive numerical computations ( to be done with more appropriate hardware and software ) : we plan to do this elsewhere .
one can not exclude that an attack with more powerful devices could finally result into a theoretically grounded lower bound @xmath3 larger than the suspected blow - up time @xmath463 of @xcite ; however , at present this is just a hope . * cases @xmath466 and @xmath467*. we use again the galerkin solution @xmath329 with @xmath302 as in ( [ defg ] ) .
due to the positivity of @xmath20 , all components @xmath468 of the galerkin solution decay exponentially for @xmath413 ( recall lemma [ rem64 ] ) ; the same happens of the norms @xmath453 , @xmath454 and of the differential error estimator @xmath433 , which are essential objects for our purposes .
the system ( [ galecomp ] ) for the galerkin components @xmath450 has been solved numerically on a time interval of length @xmath469 ( which required a cpu time of about @xmath441 seconds for @xmath470 , and @xmath471 seconds for @xmath472 ) .
subsequently , @xmath453 , @xmath454 and @xmath433 have been computed from the components @xmath450 and eqs .
( [ d34 ] ) ( [ nora3 ] ) ( indeed , some interpolation has been done as in the case @xmath61 ; as in that case , for the computation of @xmath433 we have used only 30 sample values of @xmath215 in @xmath473 , with a cpu time of about @xmath441 seconds for each one ) . for both the above values of @xmath20
, the final step has been the ( very fast ) numerical solution of the cauchy problem ( [ econt1 ] ) ( [ econt2 ] ) for the unknown function @xmath456 . according to the mathematica routines for odes ,
@xmath456 diverges for @xmath457 , where @xmath474 for @xmath470 , and @xmath475 for @xmath472 .
we repeat that these results grant existence on a domain @xmath476 for the solution @xmath4 of the ns cauchy problem ( [ caun ] ) ( [ unec ] ) , and the bound ( [ ubr3 ] ) on this interval . as examples , in figures 2a-2f we have reported some details on computations for @xmath472 .
more precisely , figures 2a , 2b and 2c are the graphs of @xmath442 for @xmath477 , in the cases @xmath444 , @xmath445 and @xmath446 , respectively .
( in fact , the graphs of @xmath442 for @xmath478 and @xmath479 are identical to the graph of the case @xmath480 . )
figures 2d and 2e are the graphs of @xmath453 and @xmath433 , for @xmath477 .
figure 2f gives the graph of @xmath456 that , as anticipated , diverges for @xmath481 .
* case @xmath482*. again , we have used the galerkin solution @xmath329 with @xmath302 as in ( [ defg ] ) .
all the components @xmath468 , the norms @xmath453 , @xmath454 and the differential error estimator @xmath433 decay exponentially for @xmath413 .
the system ( [ galecomp ] ) for the galerkin components @xmath450 has been solved numerically on a time interval of length @xmath469 ( which required a cpu time of about @xmath471 seconds ) .
the forthcoming figures 3a,3b,3c are the graphs of @xmath442 for @xmath477 , @xmath444 , @xmath445 and @xmath446 , respectively .
subsequently , the norms @xmath483 ( @xmath452 ) and the error @xmath433 have been computed from the components @xmath450 and eqs .
( [ d34 ] ) ( [ nora3 ] ) ( making some interpolations , as in the previous cases ) .
figures 2d and 2e are the graphs of @xmath453 and @xmath433 , for @xmath477 . the final step has been the ( very fast ) numerical solution of the cauchy problem ( [ econt1 ] ) ( [ econt2 ] ) for the unknown function @xmath456 . differently from all the previous cases , the numerical solution @xmath456 determined by mathematica is defined on the whole interval @xmath473 ; its graph is reported in figure 3f which suggests , via some extrapolation , that @xmath456 should be defined for all @xmath484 , with @xmath485 for @xmath413 ; of course , this would imply global existence for the solution @xmath4 of the ns cauchy problem ( [ caun ] ) ( [ unec ] ) .
however , global existence of @xmath4 can even be inferred without extrapolating the behavior of @xmath65 outside the interval @xmath473 .
in fact , global existence is granted if the condition ( [ iprn3 ] ) @xmath486 holds at any instant @xmath487 ; in the present case @xmath488 , and the numerical computation performed in the interval @xmath473 shows that ( [ iprn3 ] ) is satisfied for any @xmath489 . in conclusion
, we can take for granted that we have global existence for the solution @xmath4 of the ns cauchy problem .
of course , @xmath490 is bounded by the numerically computed function @xmath456 , for @xmath477 . after choosing a @xmath491
, we obtain as well a bound of the form ( [ tes0 ] ) for @xmath492 , which also implies @xmath492 to vanish exponentially for @xmath413 .
for example , let us choose @xmath493 .
we have @xmath494 , @xmath495 ; with the already known value @xmath496 , eq .
( [ tes0 ] ) gives u(t ) _ 3 0.0614 e^-8 ( t-0.9 ) .
[ espbound ] ( here the first inequality follows directly from ( [ tes0 ] ) , recalling that @xmath497 ; the second inequality is obvious . ) * summary of the previous results , and final comments . * our method to treat the
galerkin approximant @xmath329 , with @xmath302 as in ( [ defg ] ) , has given the following results about the euler / ns cauchy problem ( [ caun ] ) ( [ unec ] ) .
\a ) @xmath498 : we can grant existence of the solution @xmath4 of ( [ caun ] ) ( [ unec ] ) on an interval containing @xmath69 , with @xmath499 , respectively , for these three choices of @xmath20 . we have @xmath500 for @xmath459 , where @xmath65 is computed numerically solving ( [ econt1 ] ) ( [ econt2 ] ) ( for @xmath61 and @xmath472 , the graph of @xmath65 is reported in figures 1f and 2f ) .
\b ) @xmath501 : we can grant global existence for the solution @xmath4 of ( [ caun ] ) ( [ unec ] ) .
for @xmath477 we have a bound @xmath424 , with @xmath65 obtained again from the numerical solution of ( [ econt1 ] ) ( [ econt2 ] ) ; the graph of this function is reported in figure 3f .
for @xmath502 we have a bound of the form ( [ espbound ] ) on @xmath492 , decaying exponentially for large @xmath215 . by extrapolation , we are led to conjecture that results similar to ( a ) should be obtained for all @xmath503 , while results similar to ( b ) should be obtained for all @xmath504 , for some @xmath505 .
( from a qualitative viewpoint , this is just the behavior described by lemma [ lemdec ] on the analytical solution of the control inequality . 0.4 cm 0.4 cm 0.4 cm 0.4 cm 0.4 cm 0.4 cm 0.4 cm 0.4 cm 0.4 cm
let @xmath229 $ ] ; suppose we are given c([0 , ) , ) , ( s , t ) ( s , t ) s c([0 , ) , ) ; s_0 . under the above assumptions , the cauchy problem ( t ) = ( ( t),t ) , ( 0 ) = s_0 .
[ caus ] has a unique maximal ( i.e. , nonextendable ) solution @xmath506 .
( of course @xmath507 $ ] ; later on we give conditions under which @xmath508 . )
the following is a known result , of the aplygin type : see @xcite or @xcite .
[ lem1 ] * lemma .
* suppose there is a function @xmath509 such that ( t ) ( ( t),t ) , ( 0 ) s_0 [ ipotesiw ] ( with @xmath510 the right lower dini derivative , see eq .
( [ dlower ] ) ) . then ( t ) ( t ) .
a straightforward variant of the previous result is the following .
[ lem2 ] * lemma .
* suppose there is a function @xmath511 such that ( t ) ( ( t),t ) , ( 0 ) s_0 , [ ipotesir ] ( with @xmath512 the right upper dini derivative , see eq .
( [ dupper ] ) ) . then ( t ) ( t ) .
[ so0 ] * proof .
* we consider the function @xmath513 , recalling that @xmath514 ; from ( [ ipotesir ] ) we infer ( t ) -((t),t ) , -(0 ) -s_0 . on the other hand ,
the function @xmath515 is the maximal solution of the cauchy problem ( t ) = -((t),t ) , -(0 ) = -s_0 .
therefore , lemma [ lem1 ] with @xmath516 replaced by @xmath513 ( and other obvious substitutions ) gives @xmath517 yielding the thesis ( [ so0 ] ) .
[ lem3 ] * lemma .
* suppose there are functions @xmath234 fulfilling eqs .
( [ ipotesiw ] ) , ( [ ipotesir ] ) respectively . then = , ( t ) ( t ) ( t ) . [ so ] * proof . * from lemmas [ lem1 ] and [ lem2 ]
we infer ( t ) ( t ) ( t ) ; [ so1 ] we claim that this inequality implies = .
[ so2 ] in fact , whenever @xmath518 is finite , the nonextendability of @xmath519 beyond this time implies that @xmath519 is unbounded in any left neighborhood of @xmath520 ; on the other hand , if it were @xmath521 , eq .
( [ so1 ] ) would ensure the boundedness of @xmath519 on @xmath522 .
now , the combination of ( [ so1 ] ) and ( [ so2 ] ) gives the thesis ( [ so ] ) .
if we forget the solution @xmath519 of the cauchy problem ( [ caus ] ) , the inequality ( [ so ] ) becomes @xmath523 for @xmath524 ; this is just the statement in eq .
( [ eqtesto ] ) of lemma [ lemtesto ] , which is thus recognized as a weaker version of lemma [ lem3 ] .
let @xmath260 , @xmath312 and @xmath313 . in the finite dimensional space @xmath315 , we consider the maximal solution @xmath329 of the cauchy problem ( [ galer ] ) , here reproduced for the reader s convenience : @xmath525 @xmath316 for the moment , the maximal time of existence @xmath526 is not known ; one of our aims is to show that @xmath527 .
let us proceed to prove this claim and all the other statements of lemma [ rem64 ] ; our arguments are divided in several steps .
_ step 1 ( `` energy balance equation '' ) . everywhere on @xmath528
, one has _
^2_l^2 = 2 | _ l^2 + 2 f | _ l^2 .
[ ebal ] in fact , ^2_l^2 = 2 | _ l^2 [ eb1 ] @xmath529 on the other hand , ( , ) | _ l^2 = ( , ) | _ l^2 = 0 ; [ eb2 ] in the first passage above , the projection @xmath375 on @xmath315 is omitted since we are taking the inner product with @xmath530 ; the second passage reflects the well known identity @xmath531 , holding for all zero mean , divergence free vector fields @xmath160 on @xmath0 sufficiently regular to give meaning to the indicated inner product ( see , e.g. , @xcite , lemma 2.3 ) . eqs .
( [ eb1 ] ) ( [ eb2 ] ) yield the thesis ( [ ebal ] ) .
_ step 2 . everywhere on @xmath532
, one has _
^2_l^2 - 2 ^2_l^2 + 2 f _ l^2 _
[ tps ] this follows from the energy equation ( [ ebal ] ) , combined with the inequalities | _ l^2 - ^ 2_l^2 , f | _ l^2 f _ l^2 _ l^2 .
the first of these relations follows using ( [ inelap ] ) , with @xmath533 ; the second one is an application of the cauchy - schwarz inequality for @xmath534 .
_ step 3 . everywhere on @xmath532
, one has _
_ l^2 - _ l^2 + f _ l^2 , [ tpr ] _ with @xmath49 the upper dini derivative .
_ let us first prove eq .
( [ tpr ] ) at a time @xmath535 such that @xmath536 . in this case ,
@xmath537 for all @xmath215 in an interval @xmath217 containing @xmath207 ; in this interval the function @xmath538 is @xmath203 , and we have @xmath539 the last passage following from ( [ tps ] ) . now , consider an instant @xmath207 such that @xmath540 .
due to a general result already mentioned ( see the comments before eq .
( [ now1 ] ) ) , we can write @xmath541 on the other hand , eq .
( [ galer ] ) and the assumption @xmath542 give @xmath543 , so @xmath544 this is just the thesis ( [ tpr ] ) , at the instant under consideration . _ step 4 .
one has _ ( t ) _
l^2 ( u_0 _ l^2 + _ 0^t d s e^s f(s ) _ l^2 ) e^-t .
[ jute ] let us consider the continuous function @xmath545 ; due to step 3 and to @xmath546 , we have ( t ) _ l^2 ( ( t ) _ l^2 , t ) , ( t ) _
l^2 = u_0 _ l^2 where : [ 0,+ ) , ( s , t ) : = - s + f(t ) _
l^2 . on the other hand , the cauchy problem ( t ) = ( ( t ) , t ) , ( 0 ) = u_0 _ l^2 has the global solution ( t ) : = ( u_0 _ l^2 + _ 0^t d s e^s f(s ) _
l^2 ) e^-t .
so , by the comparison results reviewed in the previous appendix ( see , in particular , lemma [ lem2 ] ) , we have @xmath547 this is just the thesis ( [ jute ] ) .
one has @xmath527 ( which justifies statement ( [ tglob ] ) in the lemma under proof ) . _
if it were @xmath548 , we would have @xmath549 ; this would be contradicted by eq .
( [ jute ] ) , implying @xmath550 @xmath551 @xmath552 .
proof of statement ( [ stacit ] ) in the lemma , under the assumption of zero forcing .
_ we use the results of steps 1 and 4 , putting therein @xmath527 . for @xmath319 and @xmath61 , the energy balance equation ( [ ebal ] ) takes the form @xmath553 ; thus , @xmath554 for all @xmath484 . for @xmath319 and @xmath51 , eq.([jute ] ) gives @xmath555 again for @xmath484 .
the above two equations correspond to the content of ( [ stacit ] ) .
proof of statements ( [ comb1])-([comb3 ] ) in the lemma .
_ statement ( [ comb1 ] ) is just eq .
( [ jute ] ) , with @xmath556 .
having proved ( [ comb1 ] ) , statements ( [ comb2 ] ) ( [ comb3 ] ) are obvious
. 0.7 cm * acknowledgments .
* this work was partly supported by indam , infn and by miur , prin 2008 research project `` geometrical methods in the theory of nonlinear waves and applications '' .
99 j. t.beale , t. kato , a. majda , _
remarks on the breakdown of smooth solutions for the 3d euler equations _ , commun .
phys . * 94 * , 61 - 66 ( 1984 ) .
e. behr , j. ne@xmath59as , h. wu , _ on blow - up of solution for euler equations _
, m2an : math .
35 * 229 - 238 ( 2001 ) .
chernyshenko , p. constantin , j.c .
robinson , e.s .
titi , _ a posteriori regularity of the three - dimensional navier - stokes equations from numerical computations _ , j. math .
* 48 * , 065204/10 ( 2007 ) .
m. dashti , j.c .
robinson , _ an a posteriori condition on the numerical approximation of the navier - stokes equations for the existence of a strong solution _ , siam j. numer .
46 * , 3136 - 3150 ( 2008 ) .
h. fujita , t. kato , _ on the navier - stokes initial value problem .
i _ , arch .
rational mech .
* 16 * , 269315 ( 1964 ) .
t.kato , _ nonstationary flows of viscous and ideal fluids in @xmath557
_ , j.funct.anal . *
9 * , 296 - 305 ( 1972 ) .
t. kato , _ quasi - linear equations of evolution , with applications to partial differential equations _ , in `` spectral theory and differential equations '' , proceedings of the dundee symposium , lecture notes in mathematics * 448 * , 23 - 70 ( 1975 ) .
h. kozono , y. taniuchi , _ limiting case of the sobolev inequality in bmo , with application to the euler equations _ ,
phys . * 214 * , 191200 ( 2000 ) .
v. laksmikantham , s. leela , `` differential and integral inequalities '' , volume i , academic press , new york ( 1969 ) .
lemari - rieusset , `` recent developments in the navier - stokes problem '' , chapman & hall , boca raton ( 2002 ) .
mitrinovic , j.e .
pecaric , a.m. fink , `` inequalities involving functions and their integrals and derivatives '' , kluwer , dordrecht ( 1991 ) . c. morosi , l. pizzocchero , _ on approximate solutions of semilinear evolution equations _ , rev .
* 16 * , 383 - 420 ( 2004 ) . c. morosi
, l. pizzocchero , _ on approximate solutions of semilinear evolution equations ii .
generalizations , and applications to navier - stokes equations _ , rev .
phys . * 20 * , 625 - 706 ( 2008 ) . c. morosi , l. pizzocchero , _ an @xmath420 setting for the navier - stokes equations : quantitative estimates _
, nonlinear anal . *
74 * , 2398 - 2414 ( 2011 ) . c. morosi , l. pizzocchero , _ on the constants in a basic inequality for the euler and ns equations _ , arxiv : 1007.4412v2 [ mathap ] ( 2010 ) . c. morosi , l. pizzocchero , _ on the constants in a kato inequality for the euler and ns equations _
, arxiv : 1009.2051v1 [ mathap ] ( 2010 ) . to appear in commun .
pure appl .
t. petry , _ on the stability of the abramov transfer for differential - algebraic equations of index 1 _ , siam j. numer .
35 * , 201216 ( 1998 ) .
robinson , w. sadowski , _ numerical verification of regularity in the three - dimensional navier - stokes equations for bounded sets of initial data _ , asymptot .
59 * , 39 - 50 ( 2008 ) .
y. sinai , _ power series for solutions of the 3d navier - stokes system on @xmath557 _ , j. stat .
phys . * 121 * , 779803 ( 2005 ) .
r. temam , `` navier - stokes equations and nonlinear functional analysis '' , siam , philadelphia ( 1983 ) . | we consider the incompressible euler or navier - stokes ( ns ) equations on a torus @xmath0 , in the functional setting of the sobolev spaces @xmath1 of divergence free , zero mean vector fields on @xmath0 , for @xmath2 .
we present a general theory of approximate solutions for the euler / ns cauchy problem ; this allows to infer a lower bound @xmath3 on the time of existence of the exact solution @xmath4 analyzing _ a posteriori _ any approximate solution @xmath5 , and also to construct a function @xmath6 such that @xmath7 for all @xmath8 . both @xmath3 and @xmath6
are determined solving suitable `` control inequalities '' , depending on the error of @xmath5 ; the fully quantitative implementation of this scheme depends on some previous estimates of ours on the euler / ns quadratic nonlinearity @xcite @xcite . to keep in touch with the existing literature on the subject ,
our results are compared with a setting for approximate euler / ns solutions proposed in @xcite . as a first application of the present framework
, we consider the galerkin approximate solutions of the euler / ns cauchy problem , with a specific initial datum considered in @xcite : in this case our methods allow , amongst else , to prove global existence for the ns cauchy problem when the viscosity is above an explicitly given bound .
p c 1**t**o^1 a 1 addtoresetequationsection on approximate solutions of the incompressible euler and navier - stokes equations carlo morosi@xmath9 , livio pizzocchero@xmath10 ( ) + @xmath11 dipartimento di matematica , politecnico di milano , + p.za l. da vinci 32 , i-20133 milano , italy + e mail : [email protected] + @xmath12 dipartimento di matematica , universit di milano + via c. saldini 50 , i-20133 milano , italy + and istituto nazionale di fisica nucleare , sezione di milano , italy + e mail : [email protected] * keywords : * navier - stokes equations , existence and regularity theory , theoretical approximation .
* ams 2000 subject classifications : * 35q30 , 76d03 , 76d05 . |
while it is clear that the majority of massive early - type galaxies formed the bulk of their stars at @xmath3 , the precise nature and duration of star formation in these galaxies has proven difficult to constrain .
the difficulty arises from the fact that the main sequence turnoff point varies little with age at late times , implying that the integrated spectral energy distribution varies little with age at late times . for old stellar systems such as early - type galaxies ,
it has become common to instead rely on elemental abundance ratios to probe the star formation histories at early times ( e.g. , * ? ? ?
* ; * ? ? ?
the classic example is the ratio of @xmath0 to iron - peak elements , [ @xmath0/fe]@xmath4log(x / fe)@xmath5log(x / fe)@xmath6 .
the @xmath0 elements form mostly in massive stars , which evolve on short timescales , while iron - peak elements form mostly in type ia supernovae ( sne ) , which occur on longer timescales .
the [ @xmath0/fe ] ratio is thus a sensitive probe of sf on @xmath7 yr timescales ( e.g. , * ? ? ?
* ; * ? ? ?
one of the principal difficulties in using [ @xmath0/fe ] to measure a sf timescale is the unknown type ia sne delay time distribution .
other difficulties include the possibility that the overall ia rate varies in some systematic way with galaxy properties , the possibility of selective mass - loss such that fe is preferentially lost from the system , and potential variation in the initial mass function @xcite .
a promising alternative chronometer is ba , which is believed to form predominately within the envelopes of asymptotic giant branch stars via @xmath1-process neutron captures @xcite .
the precise mass - dependent yields of ba are presently not well - known , but there is hope that progress on this topic can be made in the near - term ( in contrast to the delay time distribution for ia s , which appears to be governed largely by the initial conditions of stellar binarity ) .
the prospects for measuring ba in the integrated light spectra of galaxies are favorable owing to the strong transitions of ba@xmath8ii in the blue spectral region .
sr is another neutron capture element with strong transitions in the blue . like ba
, sr is predominantly produced by the @xmath1-process , at least in the solar system . these two elements probe two of the three @xmath1-process peaks , with sr belonging to the first ( along with y and zr ) , and ba belonging to the second ( along with la , ce , pr , and nd ) . while the nucleosynthetic origin of ba is relatively secure ,
the same can not be said for the elements in the first @xmath1-process peak @xcite .
indeed , current chemical evolution models of the galaxy are unable to reproduce the observed behavior of sr - y - zr at low metallicity without appealing to exotic and/or ad hoc nucleosynthetic sites ( e.g. , * ? ? ?
* ; * ? ? ?
perhaps the [ sr / ba ] ratios in metal - rich massive galaxies will provide some insight into this problem .
in this letter we employ our new stellar population synthesis ( sps ) model to interpret high quality optical spectra of early - type galaxies in order to constrain the abundances of sr and ba . in sections
[ s : model ] and [ s : data ] we describe the model and data , and in section [ s : res ] we present our results .
the sps model used herein was developed in ( * ? ? ?
* cvd12 ) .
the model adopts standard sps techniques , including constructing libraries of isochrones and stellar spectra .
empirical spectra form the core of the model .
the empirical stars are of approximately solar metallicity and have solar abundance patterns .
we computed a large grid of model stellar atmospheres and spectra in order to construct response functions ( i.e. , the relative change in the spectrum of a star due to a change in the abundance of a single element ) .
the model atmospheres and spectra were computed with the atlas model atmosphere and spectrum synthesis package @xcite , ported to linux by @xcite .
specifically , we use the atlas12 code , computing new atmospheres for each change in abundance pattern .
the line list was provided by r. kurucz , including linelists for tio and h@xmath9o , amongst other molecules .
the spectral response functions were applied to the empirical stellar spectra in order to create models with arbitrary abundance patterns .
the model allows for arbitrary variation in the initial mass function ( imf ) and spans ages from @xmath10 gyr .
see cvd12 for further details regarding the model .
we follow @xcite in fitting the model to data . in its present form
the model contains 27 free parameters , including the redshift and velocity dispersion , a two - part power - law imf , two population ages , four nuisance parameters , and the abundances of c , n , na , mg , si , ca , ti , v , cr , mn , fe , co , ni , sr , y , and ba , and o , ne , s are varied in lock - step . these parameters are fit to the data via a markov chain monte carlo fitting technique . the data and models are split into four wavelength intervals ( defined in the following section ) and , within each interval the spectra are normalized by a high - order polynomial ( with degree @xmath11 where @xmath12 ; see * ? ? ?
* for details ) .
we have masked the spectral regions surrounding the h@xmath0 , [ n@xmath8ii ] , [ s@xmath8ii ] , h@xmath13 , [ o@xmath8iii ] , and [ n@xmath8i ] emission lines .
the effective resolution of the data varies from @xmath14 owing to intrinsic doppler broadening .
both because of this low resolution and the fact that in our model the synthetic spectra are only used differentially , several uncertainties associated with abundance analysis at high resolution ( including the adopted microturbulent velocity , treatment of convection , non - lte effects , and hyperfine splitting of certain levels ) are mitigated to some extent herein .
moreover , @xcite has demonstrated that the strong sr@xmath8ii line at 4077 suffers from negligible non - lte corrections at [ fe / h]@xmath15 .
of course when working at low resolution one is more dependent on the fidelity of the model than the case where equivalent widths of individual unsaturated lines are directly measured from the data . as an example of what can be measured from low resolution spectra , in figure [ fig : response ] we show the change in the model spectrum due to a 0.3 dex increase in the [ sr / fe ] and [ ba / fe ] abundance ratios ( holding [ fe / h ] fixed to the solar value ) .
the reference model is for an age of 13 gyr , with a galactic @xcite imf , and for solar metallicity .
the model was convolved with a velocity dispersion of @xmath16 .
notice that even at this low resolution , at least two sr@xmath8ii lines and two ba@xmath8ii lines produce a change in the spectrum of @xmath17% per 0.3 dex change in abundance . as we will demonstrate in section [ s : res ] , the existence of multiple , well - separated sr and ba lines allows us to test the robustness of our results in a relatively model - independent fashion .
in this letter we consider two samples of early - type galaxies . the first sample of 34 early - type galaxies ( and the nuclear bulge of m31 ) was presented in @xcite .
the sample was selected from the sauron survey of nearby galaxies @xcite .
high s / n spectra of these objects were obtained with the low resolution imaging spectrometer ( lris ) on the keck i telescope .
the wavelength coverage spans the range @xmath18@xmath19 with a gap at @xmath20@xmath21 .
four wavelength intervals were defined , @xmath22 , @xmath23 , @xmath24 , and @xmath25 for fitting models to data .
the wavelength choices reflect the limitations of the model and data wavelength coverage and regions of severe telluric corrections .
the s / n at 5000 varies from @xmath26 @xmath27 to @xmath28 @xmath27 .
spectra were extracted within @xmath29 , with a weighting intended to mimic a circular aperture , where @xmath30 is the effective radius .
these data have already been compared to the cvd12 models in ( * ? ? ?
* see the appendix of that paper for a comparison of the best - fit models and data for each of the 35 lris galaxies ) . in that earlier work
, we did not consider variation in sr and ba , and only briefly commented on the abundances of the other elements considered in the fit .
we found that the ages , [ fe / h ] and [ mg / fe ] values derived with our full spectrum fitting technique agreed with conventional lick index - based techniques .
the second source of data is provided by the sloan digital sky survey ( sdss ; * ? ? ?
* ) . following the methodology outlined in @xcite , we have selected a sample of galaxies within a narrow redshift interval of @xmath31 , with
no detected emission in h@xmath0 nor in [ o@xmath8ii ] , and with concentrated , de vaucouleurs - like light profiles .
the sdss obtains spectra with fibers that have 3 diameters .
the typical s / n of sdss spectra in our sample is modest , of order 20 @xmath27 , and so we have chosen to stack the spectra in seven bins of velocity dispersion with mean dispersions in each bin of 88 , 112 , 138 , 167 , 203 , 246 , and 300@xmath32 .
for the smallest sigma bin the sdss fiber samples the inner @xmath33 , while for the most massive bin the fiber samples the inner @xmath34 .
we have verified that our results do not change if we select galaxies within each bin such that the fiber samples the same fraction of @xmath30 .
each spectrum was continuum - normalized and convolved to an effective dispersion of @xmath16 before stacking , and each spectrum contributed equally to the stack .
the resulting s / n of the stacked spectra at 5000 ranges from @xmath35 @xmath27 to @xmath36 @xmath27 . as with the lris data ,
four wavelength intervals were defined for the sdss stacks , @xmath37 , @xmath38 , @xmath39 , and @xmath40 for fitting models to data .
these stacked data will be the focus of future work aimed at measuring the detailed abundance patterns as a function of galaxy properties . in this work
we limit our attention to the neutron capture elements sr and ba . as a first look at the data , in figure [ fig : indx ] we show spectral indices for both the sdss stacked spectra and the individual lris galaxies .
the mg_b _ index is sensitive to mg , while the other indices are sensitive to ba and sr lines ( with central wavelengths indicated by the index name ) .
the br and sr indices are defined in @xcite , while the mg_b _ index is defined in ( * ? ? ?
* no effort was made to place the latter on the lick index scale ) .
for this figure , the lris data were convolved to a common dispersion of @xmath16 in order to afford a direct comparison with the sdss stacked spectra .
the lris data do not cover the 6142 spectral region .
overall the agreement between the lris and sdss samples is encouraging , given the very different observational setups and reduction techniques .
it is however difficult to interpret spectral indices because an index is in general sensitive not only to the feature of interest but also to features contributing to the sidebands ( i.e. , the index measures a feature strength relative to a pseudocontinuum ) .
this is especially true when the feature of interest is weak compared to other features .
full spectrum fitting such as the kind performed in the present work does not suffer from this drawback .
before showing the derived abundance trends , in figure [ fig : spec ] we explore the quality of the fit around the ba@xmath8ii lines for the stacked sdss spectrum in the second highest @xmath41 bin . in the figure we compare two
model fits , one in which [ ba / fe]@xmath42 and another in which the ba abundance is allowed to vary in the fit ( with a resulting best - fit value of [ ba / fe]@xmath43 ) .
we also mark the locations of the strongest ba@xmath8ii lines in this spectral region .
it is noteworthy that the residuals decrease in _ all _ spectral regions sensitive to ba .
this is an important test of the results to follow because it is also evident from this figure that there are numerous places where the residuals exceed the formal s / n limits of the data ( demarcated by the grey band ) .
if there were only one strong ba line then it would have been difficult to argue that a region of high residual required non - solar [ ba / fe ] values .
we return to this point later in this section .
our main result is shown in figure [ fig : basr ] . in this figure
we show the best - fit [ ba / fe ] and [ sr / fe ] abundance ratios as a function of [ mg / fe ] .
we include both the stacked sdss spectra and the lris spectra of individual galaxies .
the formal statistical errors from the sdss stacks are very small owing to the very high s / n of the spectra
. it is encouraging that the lris data and sdss stacks follow the same general trends of decreasing [ ba / fe ] and constant [ sr / fe ] with increasing [ mg / fe ] .
such low values of [ ba / fe ] are not uncommon amongst metal - poor halo stars in the galaxy @xcite . indeed , below [ fe / h]@xmath44 essentially all halo stars have [ ba / fe]@xmath45 , with many stars having ba abundances as low as [ ba / fe]@xmath46 @xcite .
such stars are also @xmath0 enhanced , and have [ sr / ba]@xmath47 , again broadly consistent with the mean abundance ratios of the massive early - type galaxies ( although with very different overall metallicities )
. taken at face value , the [ ba / fe ] versus [ mg / fe ] trends are consistent with the standard interpretation of [ mg / fe ] probing the sf timescale in these galaxies .
a shorter timescale should lead to higher [ mg / fe ] and lower [ ba / fe ] values , as observed .
although not shown , the sdss data also display a strong correlation between [ mg / ba ] and @xmath41 .
the derived [ fe / h ] values range from @xmath48 to @xmath49 over the sdss early - type sample , and so the trends observed in figure [ fig : basr ] are unlikely to be due to metallicity - dependent yields .
other explanations for the [ mg / fe]@xmath50 relation have been proposed , including a varying imf , varying sn ia rate , and selective mass - loss of sn ia ejecta @xcite .
the [ mg / ba ] data allow us to rule out the explanations that are particular to the ia s , since the ia s influence neither mg nor ba , but it can not discriminate between a varying sf timescale and a varying high - mass imf for the observed variation in [ mg / fe ] with @xmath41 .
the near constancy of [ sr / fe ] is puzzling .
it has been suggested that sr may have multiple nucleosynthetic origins , including the @xmath51-process , both the weak and main @xmath1-processes , and perhaps even hypernovae @xcite . as [ mg / fe ] ( and presumably the sf timescale ) varies , then perhaps the main nucleosynthetic site varies in such a way as to keep a rough constancy in [ sr / fe ] . with detailed chemical evolution models constrained to match the observed [ mg / fe ] and [ ba / fe ] trends
, it may be possible to place interesting constraints on the origin(s ) of sr in these massive galaxies .
we turn now to several tests that have been performed to assess the robustness of these results . in figure
[ fig : sys ] we show the same abundance trends as before , for only the sdss stacked spectra , and for a variety of permutations to the fiducial model .
we consider a model in which the imf is held fixed to the galactic @xcite form , a model in which only the blue ( @xmath52 ) spectral region is fit , and a model in which the sdss stacks are not smoothed to a common resolution of @xmath53 .
the latter is a particularly important test because much more information is available in the low dispersion bins , which effectively have higher spectral resolution , compared to the high dispersion bins .
relatedly , when the data are not broadened to @xmath53 , the associated response functions ( figure [ fig : response ] ) become much stronger .
the derived abundance ratios vary between these model permutations by @xmath54 dex , but the overall qualitative trends are clearly robust to these details . we have also masked each of the two strongest ba@xmath8ii lines ( one at a time ) and refit the models to the data .
the resulting [ ba / fe ] versus [ mg / fe ] trends are unaltered , although in the case of masking the ba@xmath8ii 4555 line the overall [ ba / fe ] abundances shift lower by @xmath55 dex .
this highlights that while trends appear to be robust in our analysis , there are zero - point uncertainties at the @xmath56 dex level .
a concern is that the features we are trying to measure are much weaker than other metal lines . in fact , as shown in figure [ fig : spec ] the expected ba line strengths are similar to the typical systematic residual in the spectra _
after _ subtraction of our best - fit model !
a robust test of our result is to randomly shuffle the central wavelengths of the strongest ba@xmath8ii lines and recompute @xmath57 around the shuffled lines . in practice , this is achieved by altering the response function for ba ( see figure 1 ) by assigning random central wavelengths in the interval @xmath58@xmath59 while keeping the relative strengths and widths of the lines the same .
we have performed this test @xmath60 times for each of the sdss stacked spectra . as an example of the results for the highest @xmath41 bin , the residuals decrease around the
strongest shuffled ba@xmath8ii line in only @xmath61% of the shuffled models when using the best - fit [ ba / fe ] value shown in figure [ fig : basr ] . moreover , in _
none _ of the shuffled models do the residuals decrease in both of the two strongest shuffled ba@xmath8ii lines . in the four highest @xmath41 bins , where the best - fit [ ba / fe ] ratios are @xmath62 , the shuffled models are never able to reduce the residuals in all three of the strongest ba@xmath8ii lines .
it is therefore highly unlikely that the derived [ ba / fe ] abundances are the result of a chance alignment between the strong ba@xmath8ii lines and spectral regions where the residuals happen to be high .
the elemental abundances derived herein should provide novel constraints on the duration of star formation in early - type galaxies once chemical evolution models that include yields from the @xmath1- and @xmath51-processes are employed .
the [ mg / ba ] ratio is particularly promising because the derived timescale will not depend on the uncertain delay time distribution or overall rate of type ia sne , in contrast to timescales derived from the [ mg / fe ] ratio .
in future work , we will present detailed abundance patterns for @xmath63 elements , which promise to provide further insight into the early formation histories of these now dormant galaxies .
some of the data presented herein were obtained at the w.m .
keck observatory , which is operated as a scientific partnership among the california institute of technology , the university of california and the national aeronautics and space administration .
the observatory was made possible by the generous financial support of the w.m .
keck foundation .
funding for the sdss and sdss - ii has been provided by the alfred p. sloan foundation , the participating institutions , the national science foundation , the u.s .
department of energy , the national aeronautics and space administration , the japanese monbukagakusho , the max planck society , and the higher education funding council for england .
the sdss web site is http://www.sdss.org/. the sdss is managed by the astrophysical research consortium for the participating institutions .
the participating institutions are the american museum of natural history , astrophysical institute potsdam , university of basel , university of cambridge , case western reserve university , university of chicago , drexel university , fermilab , the institute for advanced study , the japan participation group , johns hopkins university , the joint institute for nuclear astrophysics , the kavli institute for particle astrophysics and cosmology , the korean scientist group , the chinese academy of sciences ( lamost ) , los alamos national laboratory , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , ohio state university , university of pittsburgh , university of portsmouth , princeton university , the united states naval observatory , and the university of washington . | the detailed abundance patterns of the stars within galaxies provide a unique window into the history of star formation ( sf ) at early times .
two widely used ` chronometers ' include the @xmath0 and iron - peak elements , which are created on short and long timescales , respectively .
these two clocks have been widely used to estimate sf timescales from moderate - resolution spectra of early - type galaxies .
elements formed via @xmath1-process neutron captures ( e.g. , sr and ba ) comprise a third type of chronometer , as the site of the main @xmath1-process is believed to be intermediate and low - mass asymptotic giant branch stars .
the [ @xmath0/ba ] ratio in particular should provide a powerful new constraint on the sf histories of galaxies , in part because it is insensitive to the uncertain distribution of type ia sne detonation times and the overall ia rate . here
we present new measurements of the abundance of sr and ba in nearby early - type galaxies by applying stellar population synthesis tools to high s / n optical spectra .
we find a strong anti - correlation between [ mg / fe ] and [ ba / fe ] , and a strong positive correlation between [ mg / ba ] and galaxy velocity dispersion .
these trends are consistent with the idea that more massive galaxies formed their stars on shorter timescales compared to less massive galaxies , and rule out several other proposed explanations for the observed super - solar [ mg / fe ] values in massive galaxies . in contrast , [ sr / fe]@xmath2 , with no strong variation across the sample .
it is difficult to interpret the sr trends without detailed chemical evolution models owing to the multiplicity of proposed nucleosynthetic sites for sr . |
spiral galaxies present two faces to observers . when seen in blue light , these galaxies look disordered , with spurs , branches and feather - like structures seen even in grand design galaxies , while k - band images show that these galaxies possess smooth , usually two - armed spiral patterns .
we study here the dynamical interaction between the population i component seen in blue light and the older population ii component revealed in the longer wavelengths , specifically with the intent of understanding how the interaction between the gas and the stars affects the time evolution of spiral structure in the stellar disk .
the idea that the stellar spiral structure is not a material phenomenon , but a roughly time - steady density wave rotating at a pattern speed was developed in the seminal work of lin & shu ( 1964 ; henceforth ls64 ) .
toomre ( 1964 ) gave the minimum radial stellar velocity disperion required to suppress the onset of local instabilities in a razor - thin stellar disk stabilized by rotation and the velocity dispersion of stars on large and small scales respectively .
much of the early numerical work of the study of spiral structure utilized n - body simulations to study the growth of spiral instabilities in the collionsionless components of galaxies .
transient spiral structure was evinced in the early n - body simulations of hohl ( 1970 ) , as well as in the later works of sellwood & carlberg ( 1984 ; henceforth sc84 ) .
toomre ( 1981 ) showed that a shearing disk with a supply of leading waves will swing amplify these leading disturbances to trailing disturbances , and was one of the earliest works to emphasize the role of cosmological accretion of gas .
tidal disturbances or minor mergers can torque the stellar disk and create transient spiral structure ( hernquist & mihos 1995 ) . some of these early influential works noted that the dissipative nature of gas may provide a self - regulatory influence in the combined gas - disk system ( lin & shu 1964 ; bertin & lin 1996 ) , with dissipation in the gas layer serving to continually regenerate the shearing spiral instability ( goldreich & lynden - bell 1965 ) .
contemporary numerical investigations of spiral structure have focused mainly on the response of a gas disk when driven by a spiral potential , to varying levels of physical detail and numerical sophistication .
kim & ostriker ( 2002 ) carried out local magnetohydrodynamic ( mhd ) simulations to study the development of spiral substructure in a gas disk forced by a spiral potential .
chakrabarti , laughlin & shu ( 2003 ) ( henceforth cls ) studied the development of spiral substructure due to ultraharmonic resonances using global simulations which incorporate the self - gravity of the gas and found that inter - arm features of long azimuthal extent occurred at resonance locations , which they called `` branches '' .
cls also found that when the gas disk is forced by a time - steady , but nonlinear spiral potential ( which is composed of many fourier components ) , overlap of resonances produced a highly time - dependent response in the gas which they speculated may signal the onset of turbulence in the purely hydrodynamic regime .
shetty & ostriker ( 2006 ; henceforth so06 ) carried out global mhd simulations and found that the magneto - jeans instability is effective at forming spiral - substructure in low - shear ( i.e. , near the spiral arms ) regions when the magnetic field is large .
the stabilizing effect of magnetic pressure allowed these simulations to be carried out for longer times than those of cls .
yanez et al.s ( 2007 ) hydrodynamic simulations also evince the role of ultraharmonic resonances in the development of spiral sub - structure in the gaseous component .
a significant work on the analysis of spiral structure in self - gravitating disks is that of laughlin et al .
( 1998 ) - these authors rigorously show that the primary unstable mode is endemic to the disk and not due to the numerical implementation .
one of the essential points evinced by these simulations of the response of a self - gravitating gas disk to a time - steady stellar potential is that low @xmath0 gas disks display a very time - dependent response and continue to do so over many dynamical times ( where we refer to the toomre q parameter , i.e. , @xmath1 , where @xmath2 is the sound speed of the relevant component ; toomre 1964 ) .
this is in marked contrast to purely stellar disks which may exhibit transient spiral structure ( if they are low @xmath0 disks ) but the spiral patterns disappear after a few dynamical times .
block et al .
( 1994 ; 1996 ) found that many spiral galaxies have a grand design structure in infrared light ( elmegreen & elmegreen 1984 ; rix & zaritsky 1995 ) while appearing quite flocculent in blue light .
they interpret this discrepancy in morphology to mean that the population i and population ii components are dynamically decoupled .
recent multi - wavelength observations have begun to address in detail the long term evolution of spiral structure in nearby galaxies .
notably , kendall et al . (
2008 ) infer from analysis of high spatial resolution multi - wavelength photometry of m81 ( from the b - band to irac bands ) a phase shift between the gas and the stars that is suggestive of a long - lived spiral structure in the stellar disk .
global torques may have a role in fueling star formation over many gyrs if they are able to transport the large amounts of mass in extended hi disks ( wong & blitz 2002 ) to the inner regions .
these issues require a @xmath3 analysis that treats the interplay between gas and stars rather than treating the response of the gas to an imposed stellar potential as has been studied previously ( cls ; so06 , among other previous papers ) . in this paper
, we employ high resolution three - dimensional simulations performed with the smoothed particle hydrodynamics ( sph ) code gadget-2 ( springel 2005 ) to simulate the evolution of @xmath4 in the stellar disks of galaxies that are evolving in relative isolation .
the recent analysis of the chemical and star formation history of the milky way by colavitti et al .
( 2008 ) suggests that the milky way has not undergone recent ( i.e. , after @xmath5 ) significant merger events .
we study here the problem of the generation and evolution of spiral structure in an isolated galaxy with gas , stars , and a dark matter halo , where these components are free to interact dynamically with each other . specifically , by including not only the stars with their gravitational forces and velocity dispersions , but also the cold interstellar gas with its self - gravity , pressure , its ability to cool radiatively and dissipate energy through shocks , and form new stars , we focus our study on the dynamical coupling between the gas and the stars and the torques that the gas is able to exert on the stars , an effect that has not been analyzed in previous studies which have focused on the effect of the stars ( whether live or modeled through a potential ) on the gas ( hernquist & mihos 1995 ; cls ; s006 ; kim & ostriker 2007 ) . we also comment on the coupling with the live halo and its role as a sink of angular momentum , as has been noted in previous studies of bar instabilities ( athanassoula 2003 ; martinez - valpuesta et al .
2006 ; berentzen et al . 2007 ) .
we find that the dissipative nature of the gas leads to @xmath6 , @xmath7 azimuthal features , with the gas response leading the stellar response , which allows for a phase shift between these components such that the gas is able torque the stars .
it is this angular momentum transport interior to co - rotation that allows for the stellar disk to be unstable to spiral structure for a longer time ( many revolutions ) compared to simulations that do not include the dissipative gaseous component .
the paper is organized as follows . in 2 , we review the simulation methodology , specifically the initial equilibrium structure of the disks and implementation of radiative cooling , viscosity , and star formation prescriptions for turning gas into stars . in 3 , we give our results for simulations that treat the collionsionless components only , and in 4 , we describe results from simulations that include the time evolution of gas , stars , and a dark matter halo .
we discuss our results in the context of observations in 5 , and we conclude in 6 .
the appendices are devoted to a detailed exploration of simulation parameters and assumptions that can affect the gas - star angular momentum transport , specifically the artificial bulk viscosity and equation of state for the gas .
we employ the parallel treesph code gadget-2 ( springel 2005 ) to perform simulations of isolated disk galaxies .
gadget-2 uses an n - body method to follow the evolution of the collionsionless components , and sph to follow the gaseous component .
energy and entropy are both conserved in regions free of dissipation .
the simulations reported here ( unless otherwise noted ) have gravitational softening lengths of @xmath8 for the gas and stars , and @xmath9 for the halo .
the number of gas , stellar , and halo particles are @xmath10 , @xmath11 and @xmath12 respectively for our fiducial case . the total angular momentum of these simulations is conserved to 0.1 % over 3 gyr for a time integration error tolerance accuracy of 0.025 . in appendix
a , we give some results of a few cases where we vary the artificial bulk viscosity and the equation of state - these simulations are performed at lower resolution and their specifications are noted in the appendix .
the implementation of viscosity in gadget-2 is described in detail in springel et al .
( 2001 ) and springel ( 2005 ) , so we briefly review the results here .
gadget-2 uses a bulk artificial viscosity which is the shear - reduced version ( balsara 1995 ; steinmetz 1996 ) of the standard monaghan and gingold ( 1983 ) artificial viscosity .
specifically , the form of the artificial viscosity experienced between particles @xmath13 and @xmath14 is taken to be given by : @xmath15 where @xmath16 is the arithmetic mean of the sph density estimator ( equation 5 in springel 2005 ) of particles @xmath13 and @xmath14 , @xmath2 and @xmath17 are the sound speeds , and @xmath18 and @xmath19 are the relative projected velocities . here
, we adopt a fiducial value of 0.75 for the bulk artificial viscosity @xmath20 parameter .
this value has been empirically recommended by springel et al .
( 2001 ) by comparing the results of the shock tube test to other numerical implementations . in discontinuities that develop in the gaseous flow ,
the entropy generation and width of the shock is proportional to this parameter .
we vary this parameter in appendix a and find that angular momentum transport from the gas to the stars is inhibited for very low values of the artificial bulk viscosity parameter . for low @xmath20 values ,
the effect of dissipation in shocks is prohibited such that the gaseous component becomes more akin to the collionsionless component and the stars no longer exhibit a significant time lag or azimuthal phase shift in their response .
radiative cooling and heating are implemented in gadget-2 in a manner similar to katz , weinberg & hernquist ( 1996 ) ( henceforth kwh96 ) .
the code implements radiative cooling for an optically thin gas of helium and hydrogen in collisional ionization equilibrium with a user - specified ( time - dependent ) ultraviolent background .
radiative cooling allows the gas to dissipate energy and form dense , cold clumps , which as we describe below are prone to star formation .
for the densities resolved in these simulations , radiative cooling due to two - body processes is important , namely , collisional excitation of neutral hydrogen ( @xmath21 ) and singly ionized helium ( @xmath22 ) , collisional ionization of @xmath23 and @xmath22 , standard recombination of @xmath24 , and @xmath25 , dielectric recombination of @xmath26 and free - free emission .
the adopted cooling rates are given in table 1 of kwh96 .
the heating rate from photoionization can be computed once the intensity of the uv background is specified .
since this heating rate is very low inside disks in the local universe , we do not make use of this heating rate from a uv background .
the version of gadget-2 that we employ uses a sub - resolution model for energy injection from supernovae that is proportional to the star formation rate ( springel & hernquist 2003 ) . in this approach ,
the ism is taken to be composed of cold clouds from which stars can form , which are embedded in a pressure - confining hot phase .
the effective equation of state for this two - component gas ( eq .
24 in springel et al .
2001 ) is taken to be 0.25 for our fiducial cases .
in essence , employing such a equation of state for the star forming gas allows one to transition from the isothermal case to a stiffer than isothermal equation of state , where in the latter case energy injection from supernovae ( which is proportional to the star formation rate ) allows one to evolve gas - rich disks for many dynamical times .
we vary the equation of state in appendix b and study its effect on the gas - star angular momentum transport .
we discuss in this paper both simulations where star formation is turned on and those where star formation is turned off . for simulations where star formation is turned on ,
we employ a density dependent kennicutt - schmitdt algorithm , to convert gas to stars in these simulations .
the star formation rate is given by : @xmath27 where @xmath28 is the total gas density , @xmath29 for a salpeter imf , and the quantity @xmath30 is given by : @xmath31 where @xmath32 is fixed by the multiphase model and is the threshold density above which the gas is assumed to be thermally unstable to the onset of a two - phase medium composed of cold star forming clouds embedded in a hot , pressure - confining phase ( sh03 ) .
we take @xmath33 , the gas consumption timescale , to be 4.5 gyr and 8 gyr which are intermediate values between that adopted by sh03 and springel et al .
( 2005 ) and agree reasonably well ( to within a factor of 2 - 3 ) with gas consumption time scales cited by kennicutt ( 1998 ) .
we aim to highlight in this paper general trends in the dynamics that are seen across simulations .
due to the large dynamic range in global galaxy simulations , star formation is currently treated in a prescriptive way . to minimize dependence of our results on our choice of a star formation prescription
, we also include results from simulations where the star formation is turned off . for the simulation where star formation is turned off , there is no energy injection from supernovae ( as sh03 s model takes the energy injection rate to be proportional to the star formation rate ) , and the gas remains isothermal at @xmath34 .
we discuss in a future paper results from various star formation prescriptions ( where star formation is mediated by other factors such as shocks , magnetic fields and turbulence ) for turning gas into stars . [ cols="<,<,>,>,>,>,<",options="header " , ]
the initial conditions that we choose for the disks represent milky way type galaxies and are similar to those described in cox et al .
( 2006 ) and springel et al .
namely , we generate a stable isolated disk galaxy that is initially in equilibrium .
the extended dark matter halo is taken to have a hernquist ( 1990 ) profile , with a concentration parameter of 10 .
we model the disk components of gas and stars as exponential disks , with the fiducial case having a scale length of 3 kpc .
the primary simulation that we focus on here has a virial velocity of @xmath35 , with a rotation curve and mass similar to the milky way ( @xmath36 and a stellar mass of @xmath37 initially ) .
we do not include a bulge in the simulations reported here ( there is no qualitative difference as far as our main results are concerned ) , and take the baryonic gas mass fraction to be 10 % for the fiducial case , with the gas distributed initially in an axisymmetric exponential disk having the same scale length as the stars .
( in retrospect , inclusion of an extended gas disk as seen in observations by wong & blitz ( 2002 ) would have been more empirically suited ) values of these parameters for the other simulations are reported in table 1 .
these values yield an initial toomre q parameter for the stars that is slightly in excess of unity in the inner regions , and of order unity for the gas .
we have performed resolution studies to confirm that simulations with the given number of particles and softening lengths yield converged results .
the primary simulations that we analyze here are the collisionless spngnb simulation , its gaseous analogue ( spgnsfr ) , and its gaseous , star - forming analogues ( spgks and spgks8 ) . for a given cooling curve and star formation prescription , there are two aspects of the gas physics that can affect the gas - star dynamical coupling in these simulations , namely the artificial bulk viscosity and the equation of state for the gas
. these aspects of the simulation methodology and their effect on angular momentum transport and treated in appendix a and b respectively .
fourier component of the projected stellar density at three time snaps for the spngnb simulation . ] for the stars ( a ) and the halo ( b ) , for the spngnb simulation for @xmath38 and @xmath39,title="fig : " ] for the stars ( a ) and the halo ( b ) , for the spngnb simulation for @xmath38 and @xmath39,title="fig : " ] we depict in figure 1 the projected ( i.e. , we take our stellar density in x , y , z and project it on the z - axis ) stellar density ( @xmath40 ) out to 2 gyr for the spngnb simulation .
this stellar disk is only somewhat susceptible to spiral structure as it is initialized with @xmath41 in the inner parts of the disk , with @xmath42 in the outer regions .
a more quantitative representation of the spiral structure is given in figure 2 , where we have plotted the @xmath43 fourier component of the stellar surface density .
we obtain this by taking the amplitude of the real and imaginary components of @xmath44 , where @xmath45 is given by : @xmath46 figure 2 shows that that the power in the @xmath43 disturbance ( relative to the m=0 axisymmetric mode ) varies by a factor of @xmath47 in magnitude peak - to - peak in the inner regions of the disk ( @xmath48 ) over the timescale of a gyr .
the spiral patterns disappear within an orbital time due to disk heating , an effect also seen by sc84 .
rybicki ( 1971 ) estimated that the relaxation timescale is comparable to the orbital time for purely stellar razor - thin disks . the swing amplification mechanism
( toomre 1981 ) is effective for low values of @xmath0 - as the spiral patterns swing from leading to trailing in a shearing disk , they are amplified across co - rotation . in the absence of dissipation , this increases the velocity dispersion of the stars secularly until the disk is no longer susceptible to this amplification mechanism .
as noted earlier , the total angular momentum of the simulations is conserved .
however , different galaxy components are free to exchange angular momenta .
we show in figure 3a the angular momentum ( in arbitrary units ) of the stars and dark matter halo particles within @xmath49 and @xmath50 ( figure 3b ) .
the stars inside co - rotation ( which is at @xmath51 ) lose angular momentum to the halo ; of particular note is the steady decline of angular momentum in the stellar disk .
the massive slowly rotating halo is then a net sink of angular momentum , as has been found in previous simulations of bar instabilities with live halos ( athanassoula 2003 ; martinez - valpuesta et al .
2006 ; berentzen et al . 2007 ) .
the basic idea that energy and angular momentum can be absorbed by the outer part of a galaxy if it is rotating slower than a given mode was discussed in the early studies of angular momentum transport in galaxies ( lynden - bell & kalnajs 1972 ) .
we show in figure 4 the time variation of the pattern speed , computed following tremaine & weinberg ( 1984 ) .
this method derives the pattern speed on the assumption of a well - defined pattern speed and the tracer ( @xmath40 ) obeying the continuity equation .
this gives a pattern speed in terms of the projected surface densities ( @xmath52 ) , which we have computed in the center of mass frame , and velocities ( @xmath53 ) of the stars : @xmath54 which we have expressed the pattern speed here for an edge - on disk ; the weight function @xmath55 can be chosen arbitrarily .
for this simulation , we find that the pattern speed is on average @xmath56 ( for @xmath57 ) , which places co - rotation at @xmath58 , and is in agreement with expectations from modal theory as we discuss later .
the early - time ( @xmath59 ) pattern speed is high ( which would imply a co - rotation radius close to the very inner regions of the disk ) , as the @xmath43 spiral mode is not yet the dominant mode in most of the disk . for this collisionless simulation
, there is a 97 % change in the pattern speed out to 3 gyr , and a 38 % change for @xmath60 . as we discuss later ,
simulations that include a gaseous component show less variance in their pattern speeds as a function of time .
we now proceed to describing our results from simulations that include a gaseous component .
we focus our analysis on angular momentum transport from the gas to the stars .
the torque exerted by the self - gravitating gas disk on the stellar disk is given by - @xmath61 where it is specifically the asymmetric components that contribute to this integral .
we depict in figure 5 the projected stellar density from the spgnsfr simulation , which contains 10% of the disk mass in the gaseous component .
star formation is turned off in this simulation .
as is clear , the stellar disk is unstable to spiral structure for a longer time when there is gas , and @xmath62 spiral arms are seen in the gaseous component ( figure 6 ) .
figure 7 shows the time variation of the power in the @xmath43 fourier mode , and can be contrasted with figure 2 . in this case , when gas is included in the simulation , the @xmath43 disturbance is roughly time - steady - there is less than a factor of three varation in the power peak - to - peak .
fourier component of the projected stellar density at three time snaps for the spgnsfr simulation . ] for ( a ) the gas , ( b ) the stars ; overplotted is @xmath63 for the stars for the spngnb simulation as a reference point for a case without gas ( in this case the stars lose 7 % of their angular momentum inside of @xmath64 out to 2 gyr in contrast to the 35 % loss in the purely collisionless case ) , and ( c ) the halo for the spgnsfr simulation within @xmath38 and @xmath39,title="fig : " ] for ( a ) the gas , ( b ) the stars ; overplotted is @xmath63 for the stars for the spngnb simulation as a reference point for a case without gas ( in this case the stars lose 7 % of their angular momentum inside of @xmath64 out to 2 gyr in contrast to the 35 % loss in the purely collisionless case ) , and ( c ) the halo for the spgnsfr simulation within @xmath38 and @xmath39,title="fig : " ] for ( a ) the gas , ( b ) the stars ; overplotted is @xmath63 for the stars for the spngnb simulation as a reference point for a case without gas ( in this case the stars lose 7 % of their angular momentum inside of @xmath64 out to 2 gyr in contrast to the 35 % loss in the purely collisionless case ) , and ( c ) the halo for the spgnsfr simulation within @xmath38 and @xmath39,title="fig : " ] figure 8 depicts the evolution of angular momentum for the gas , stars , and halo interior and exterior to co - rotation .
angular momentum transport from the gas to the stars interior to co - rotation is able to torque the stellar disk and serves to self - regulate the disk structure .
figure 8b shows the angular momentum of the stars is considerably more time - steady than its purely collisionless analogue ( figure 3a ) , as the stars are gaining the angular momentum that is lost by the gas ( figure 8a ) .
there continues to be a slight net transfer to the live halo .
the sharp azimuthal response of the gas and the smooth response of the stars is shown in figure 9 for the 250 myr snapshot - this also shows that the gas leads the stars , with a phase shift of @xmath65 close to co - rotation .
given that the potential traces the density , figure 9 also shows why the gas is able to exert torques on the stars
@xmath66 , as the gas is compressed and shocks , it forms the sharp azimuthal features shown in figure 9 .
the response of the gas to a time - steady forcing potential in the nonlinear regime was studied by shu , milione & roberts ( 1973 ) ( henceforth smr ) .
smr found that the response of the gas can be quite nonlinear even when the amplitude of the stellar spiral forcing is a small fraction of the axisymmetric field .
the response of the gas is much sharper than that of the stars as the response of some galactic component to a forcing potential is proportional to the inverse square of its dispersive speed .
smr also discussed in detail how the breadth of the zone of high gas compression depends on whether the doppler shifted phase velocity of the stellar density wave ( or the component of the circular motion that is perpendicular to the wave front ) is greater or less than the sound speed of the gaseous component , which results in narrow and broad zones of gas compression respectively . in this simulation , where star formation is turned off ,
there is no energy injection from supernovae .
this allows the sound speed of the gas to be low enough to satisfy the criterion of smr , and can be contrasted with the cases discussed later in this paper where star formation is turned on along with an injection of energy prescribed by the model of sh03 which results in broad gaseous arms , particularly in the outer parts of the galaxy .
time - dependent numerical calculations by cls of the response of the gaseous disk when forced by a spiral stellar potential displayed similar sharp azimuthal features . ) . ]
figure 10 displays the time variation of the pattern speed for the spgnsfr simulation .
the average pattern speed of this simulation is @xmath67 , which places co - rotation at @xmath68 , and is in agreement with analytic expectations of the pattern speed from modal theory , i.e. , a few scale lengths ( bertin et al . 1998 ) and from observational comparisons ( lin et al . 1969 ; shu et al .
there is no inner lindblad resonance , and the outer lindblad resonance is at @xmath69 .
we note that in contrast to the spngnb collisionless simulation , this simulation exhibits less variance in its pattern speed , with a maximum difference of 21 % out to @xmath70 .
it is interesting to note that the average radial velocities of the gas ( at times when the spiral structure is well established ) are of order @xmath71% of the sound speed , which would cause gas to flow radially inwards @xmath72 on the time scale of a gyr .
this effect can not be neglected on the scale of the spiral arms , and the radial velocities in these simulations are in rough agreement with observational estimates ( wong et al .
radial inflow may also be produced in the outskirts of isolated galaxies due to the torque on the gas from new stars that exist along spiral arms far outside the optical radius as seen in recent galex images ( thilker et al .
we investigate the large scale transport of gas from extended hi disks as a possible mechanism for fueling star formation over many gyr in a forthcoming paper . .
] figures 11 and 12 show the projected stellar and gas surface densities for the spgks simulation , where star formation is prescribed via the kennicutt - schmidt relation with a gas consumption time scale of 4.5 gyr . although the results are qualitatively similar to the spgnsfr case , the spiral structure in the gas is not as sharp due to energy injection from star formation as prescribed in the multiphase model of sh03 ( which may be expected from the work of smr as we have discussed in 4 ) , and the stellar disk is also not as unstable to spiral structure over 2 gyr .
figure 14 shows the time evolution of the @xmath43 fourier component of the stellar surface density , which is also in qualitative agreement with the spgnsfr simulation , although the variation in power over a gyr is somewhat larger .
the time evolution of the angular momentum ( figures 15 ) of the stars is not as time steady as in the spgnsfr simulation - over the period of 2 gyr , the angular momentum of the stars declines by about 20 % , in contrast to the 7 % decrease for the spgnsfr simulation , but both of these are below the 30% stellar angular momentum loss seen in the spngnb simulation that did not include gas
. of note is the projected density of the new stars as shown in figure 13 , which is more akin to the sharper response of the gas shown in figure 12 .
figure 16 depicts the density vs azimuth at 7 kpc for the spgks simulation and is similar to the spgnsfr case , though with a slightly broader azimuthal response in the gaseous component .
the time variation of the pattern speed is shown in figure 17 , with an average pattern speed of @xmath73 , which places co - rotation at @xmath74 .
there is a maximum variance of 27 % in the pattern speed for the spgks simulation out to 3 gyr .
fourier component of the projected stellar density at three time snaps for the spgks simulation . ] for ( a ) the gas , ( b ) the stars shown along with the comparison to the purely collisionless spngnb case ( in this case the stars lose 38 % of their angular momentum inside of @xmath64 out to 3 gyr in contrast to the 43 % loss seen in the purely collisionless case ) , and ( c ) the halo for the spgks ( in my notation spgks ) simulation within @xmath38 and @xmath39,title="fig : " ] for ( a ) the gas , ( b ) the stars shown along with the comparison to the purely collisionless spngnb case ( in this case the stars lose 38 % of their angular momentum inside of @xmath64 out to 3 gyr in contrast to the 43 % loss seen in the purely collisionless case ) , and ( c ) the halo for the spgks ( in my notation spgks ) simulation within @xmath38 and @xmath39,title="fig : " ] for ( a ) the gas , ( b ) the stars shown along with the comparison to the purely collisionless spngnb case ( in this case the stars lose 38 % of their angular momentum inside of @xmath64 out to 3 gyr in contrast to the 43 % loss seen in the purely collisionless case ) , and ( c ) the halo for the spgks ( in my notation spgks ) simulation within @xmath38 and @xmath39,title="fig : " ] gas consumption time scales in nearby spiral galaxies are inferred to be on the order of a few gyr ( kennicutt 1983 ; wong & blitz 2002 ; crosthwaite & turner 2007 ) , with a range between 1 - 10 gyr ( wong & blitz 2002 ) .
the values are uncertain both due to use of various tracers to infer the rate of star formation and hence the gas depletion time scales , as well as due to the necessary use of globally averaged measures which recent works have been able to address observationally ( see e.g. , kennicutt et al .
our use of a density dependent star formation prescription is simplistic ; star formation in real galaxies may be mediated magnetically ( shu et al .
2007 ) and or through turbulence ( krumholz & mckee 2005 ) , and may be understood in a simple way in terms of the pressure ( blitz & rosolowsky 2006 ) .
nonetheless , use of this standard star formation prescription is fairly pervasive ( sh03 ; springel et al .
2005 ; cox et al .
2006 ) and does agree with globally averaged measures of the star formation rate in a diverse range of galaxy types . . ]
we also include the spgks8 case , with a gas consumption time scale of 8 gyr .
figures 18 - 20 display the projected images stellar , gas , and new stars respectively , which are similar to the spgks simulation , but with a slightly greater spiral disturbance seen in figure 18 at @xmath75 , which is also reflected in the late time @xmath43 fourier component shown in figure 21 .
the somewhat longer ( by @xmath76 ) gas consumption time scale for this simulation allows the gaseous component to couple dynamically to the stellar disk for a slightly longer time .
there is likely to be continued accretion of gas from the intergalactic medium , which may serve to replenish the spiral instability over even longer timescales , a point that we discuss qualitatively in 5 .
finally , figure 24 display the time variation of the pattern speed for the spgks8 simulation , which is on average @xmath77 and varies by @xmath78% over 3 gyr .
fourier component of the projected stellar density at three time snaps for the spgks8 simulation . ] for ( a ) the gas , ( b ) the stars overplotted with the reference spngnb case ( in this case the stars lose 23 % of their angular momentum inside of @xmath64 out to 3 gyr in contrast to the 43 % loss seen in the purely collisionless case ) , and ( c ) the halo for the spgks8 simulation within @xmath38 and @xmath39,title="fig : " ] for ( a ) the gas , ( b ) the stars overplotted with the reference spngnb case ( in this case the stars lose 23 % of their angular momentum inside of @xmath64 out to 3 gyr in contrast to the 43 % loss seen in the purely collisionless case ) , and ( c ) the halo for the spgks8 simulation within @xmath38 and @xmath39,title="fig : " ] for ( a ) the gas , ( b ) the stars overplotted with the reference spngnb case ( in this case the stars lose 23 % of their angular momentum inside of @xmath64 out to 3 gyr in contrast to the 43 % loss seen in the purely collisionless case ) , and ( c ) the halo for the spgks8 simulation within @xmath38 and @xmath39,title="fig : " ]
a number of papers have discussed the role of gas in providing a dissipative and therefore a saturation mechanism ( roberts & shu 1972 ; lubow , balbus & cowie 1986 ; bertin et al .
our time - dependent simulations treat the gas - star interplay self - consistently .
the gaseous component is able to cool radiatively and dissipate energy , and thereby remain cool enough to be prone to the spiral instability ; the azimuthal phase shift between the gas and the stars allows for the donation of angular momentum from the gas to the stellar disk .
we find that this expropriation of angular momentum is a critical factor in the long - term persistence of spiral structure in the stellar disks of galaxies . specifically , the torques that the gaseous spiral sub - structure ( those features that are not azimuthally symmetric ) exert on the stellar disk transfer angular momentum and therefore lead to a roughly time - steady spiral structure in the stellar disk .
the inclusion of a cold gaseous component also reduces the effective q parameter of the two - fluid system ( jog & solomon 1984 ; bertin et al .
1998 ; cls03 ) .
springel et al .
( 2005 ) elucidated the effects of the springel & hernquist ( 2003 ) multiphase model by performing a parameter space survey ; due to the energy injection from supernovae that is implicit in this model , extremely high gas fractions are not unstable to spiral structure .
we discuss this point further in appendix b. kendall et al .
( 2008 ) recently produced a composite surface density map of m81 using @xmath79 irac 3.6 @xmath80 , 4.5 @xmath80 , 8 @xmath80 , optical , b , v , i and 2mass @xmath81 band data .
they extract the amplitude and phase of the density from the map , and find that the response of the gas , which is traced by dust emission in the 8 @xmath80 irac band , is phase - shifted in azimuth relative to the stellar density wave .
our analysis of the simulations here is consistent with their findings , with azimuthal offsets between the gas and stars of order one radian . in a future paper , we perform radiative transfer calculations through the simulations to calculate simulated images as done previously by chakrabarti & whitney ( 2008 ) for ulirgs in the continuum bands .
our gas surface density images ( figure 6 for example ) show a much sharper response as seen in b - band images of spirals , while the stellar surface density images ( figure 5 for example ) show a much smoother response as seen in the k band images of local spirals ( block et al .
1996 ; elmegreen & elmegreen 1984 ; rix & zaritszky 1995 ) .
a major caveat to our simulations here is that we have not included the cosmological infall of gas , which can potentially serve to replenish the supply of gas over gyr timescales as the pre - existing supply of gas is steadily exhausted by star formation and transported inwards by gravitational torques .
colavitti et al .
( 2008 ) have recently analyzed the chemical evolution of a spiral galaxy in a cosmological context and find a gas infall prescription that reproduces observed chemical properties .
such a prescription may also be used to study the long term dynamics of spiral structure in a more cosmologically motivated way than we have done here .
as we have studied galaxies evolving in isolation which can not make use of the continued accretion of gas to replenish the critical supply of a cold dissipative component , as gas is tranported inwards or consumed by star formation , we very likely find a lower limit ( @xmath70 ) for the lifetime of spiral structure in disk galaxies .
this time scale may well be extended by fresh infall of gas .
gas - rich mergers will certainly affect the stability of disks ; however , detailed knowledge of their frequency , mass ratio ( i.e. , minor or major mergers ) , and final dynamical impact can not be obtained without performing high resolution hydrodynamical cosmological simulations .
recent observational works estimate that a third of lirgs ( @xmath82 ) are produced by major mergers at @xmath5 , while the rest are morphologically akin to spirals ( elbaz et al .
the morphological variation of normal galaxies with redshift ( due to the influence of other factors such as mergers ) is another issue that we have not treated here .
our approach here has been to study the simple case of a galaxy evolving in isolation and to understand specifically the gas - star dynamical coupling that is a critical factor in the long term persistence of spiral structure in stellar disks . due to the relative simplicity of our approach
, we have been able to study aspects of the gas physics that impact the angular momentum transfer from the gas to the stars , such as the effect of star formation ( by studying simulations where star formation is turned off , i.e. , in 4 , as well as simulations where star formation is turned on and the gas consumption timescale is varied , i.e. , in 4.1 ) , as well as the effect of the artificial bulk viscosity ( appendix a ) , and the equation of state ( appendix b ) .
@xmath83 we find that angular momentum exchange between the self - gravitating gas disk and the stars leads to relatively long - lived spiral structure in the stellar disk , with spiral structure lifetimes of order a few gyr .
spiral structure is washed out , when it is incipient , in simulations that do not include a cold , dissipative gaseous component .
@xmath83 the gas disk is able torque the stellar disk and donate angular momentum interior to co - rotation .
the gas is able to cool radiatively and dissipate energy - leading to a time lag ( or azimuthal phase shift on the order of @xmath84 close to co - rotation ) between the response of the gas and the response of the stars .
even though the gas mass is less than the stellar mass , the gas can exert appreciable torques ( and donate non - negligible angular momentum ) as the azimuthal response which is highly asymmetric ( or azimuthal gradient of the potential ) is much larger than that of the stars .
over timescales of many gyr , there is radial inflow of gas which ultimately limits the donation from the gas ( if it is not replenished due to cosmological infall ) ; star formation ( depending on gas consumption time - scales ) will convert the cold , dissipative component to stars which heat up and have no cooling vent .
@xmath83 the primary ( specifically the @xmath43 ) fourier components of the stellar surface density are roughly time - steady in simulations that include a gaseous component , with variation in surface density of @xmath85 over a gyr .
the pattern speed of the @xmath43 component is more time steady in simulations that include gas than those that do not .
@xmath83 the halo is a net sink of angular momentum , as has been seen in previous studies of bar instabilities with live halos .
i am grateful to frank shu for insightful discussions on the long - term persistence of spiral structure .
i thank t.j .
cox for advice on performing gadget simulations and providing a script to plot surface density outputs .
i also thank c.k .
chan , mark reid , phil chang , and ramesh narayan for detailed comments on the manuscript , and doug lin , martin weinberg , lars hernquist , and dusan keres for helpful discussions .
we discuss here the consequence of varying the artificial bulk viscosity parameter and its effect on angular momentum transfer from the gas to the stars .
figure 23 ( a - c ) depicts the angular momentum transfer for the spgnsfr simulation with the bulk viscosity parameter set to zero .
this is similar to the collisionless case , and results in the angular momentum of the stars steadily declining , as there is no transfer from the gas due to a lack of phase shift between the gaseous and stellar components .
the standard value for the artificial bulk viscosity that we have used in the paper is 0.75 ; this case is shown in figures 24 ( a - c ) .
figures 25 ( a - c ) depict the spgnsfr case with a artificial bulk viscosity of unity , which results in a very similar behavior of angular momentum transfer as our standard case .
the zero viscosity simulation ( spgnsfr0 ) shows a 24 % loss in the stellar angular momentum out to 1.4 gyr inside of 7 kpc ( which is comparable to the 30 % loss in stellar angular momentum for the reference spngnb purely collisionless case ) , while the standard viscosity case ( spgnsfr ) and the spgnsfr1 both experience a 6 % loss out to 1.4 gyr interior to 7 kpc . in summary ,
the effect of lowering the artificial bulk viscosity is to approximate the purely collisionless problem , thereby removing the phase shift in azimuth that is needed for the gas to exert a torque on the stellar component . aside from the effect of viscosity
, gas particles also exert pressure this will not affect the angular momentum transport directly , but will affect the propensity of a gas disk to be prone to spiral structure .
we discuss the effect of pressure ( i.e. , assumptions about the equation of state ) in appendix b. for ( a ) the gas , ( b ) the stars overplotted with the reference spngnb case which has a 30 % loss in angular momentum in contrast to the 6 % loss seen in this standard viscosity case , and ( c ) the halo for the spgnsfr simulation within @xmath38 and @xmath39 .
this is the standard case with the bulk artificial viscosity set to 0.75.,title="fig : " ] for ( a ) the gas , ( b ) the stars overplotted with the reference spngnb case which has a 30 % loss in angular momentum in contrast to the 6 % loss seen in this standard viscosity case , and ( c ) the halo for the spgnsfr simulation within @xmath38 and @xmath39 .
this is the standard case with the bulk artificial viscosity set to 0.75.,title="fig : " ] for ( a ) the gas , ( b ) the stars overplotted with the reference spngnb case which has a 30 % loss in angular momentum in contrast to the 6 % loss seen in this standard viscosity case , and ( c ) the halo for the spgnsfr simulation within @xmath38 and @xmath39 .
this is the standard case with the bulk artificial viscosity set to 0.75.,title="fig : " ]
another aspect of the gas physics that affects the stability of the gas disk , and indirectly impacts the angular momentum transport to the stellar disk is the assumed equation of state for the gas . as discussed by springel et al .
( 2005 ) , the equation of state , when it is taken to be greater than zero which is the isothermal limit , models the effect of supernova feedback by over - pressurizing regions relative to an isothermal equation of state .
in the isothermal limit , the pressure in compressed regions increases in proportion to the density .
the multiphase model mediates that for @xmath86 , the gas is thermally unstable and stars can form ; the energy injection rate from supernovae in this model is proportional to the star formation rate .
the compressed gas then would have an effective equation of state that is stiffer than isothermal if it retains this feedback energy .
one can expect that simulations performed with equations of state that are stiffer than isothermal are somewhat less prone to a spiral instability as the effective sound speed has been increased .
we vary the equation of state from 0 - 1 in figures 26(a - c ) and find that the simulation performed with the isothermal equation of state ( eqs0lr ) shows the lowest decrease in stellar angular momentum , while the full multiphase model ( eqs1lr ) has a 25 % decrease in the stellar angular momentum out to 1.4 gyr .
however , this relative decease 18 % versus 25 % , as the equation of state varies from isothermal to the fully pressurized multiphase model is a small enough change to suggest that assumptions about the equation of state for the gas only slightly affect the angular momentum transport to the stellar disk , for a gas fraction of 10 % .
springel et al .
( 2005 ) have performed a comprehensive parameter study in terms of the gas fraction and the equation of state to study the effects these parameters have on the gaseous disk .
they find that for the full multiphase model ( stiff equation of state ) even pure gas disks are stable to fragmentation , with an increase in gas fraction lowering the susceptibility of the gas disk to spiral structure . in the multiphase model ,
increasing the gas fraction increases the star formation ( as mediated by the adopted kennicutt - schmidt algorithm ) and thereby the supernova feedback energy ( which is proportional to the star formation rate ) that is deposited into the gas .
this can cause the effective temperature of the gas to be dynamically hotter than the stars . on the other exterme ,
extremely high gas fraction disks that are evolved with an isothermal equation of state undergo violent fragmentation .
thus , extremely high gas fraction disks ( whether they are evolved with an isothermal equation of state or full multiphase model ) will not be effective as angular momentum donors to the stellar disk .
99 athanassoula , e . ,
2003 , mnras , 341 , 1179a bertin , g . & lin , c.c . , 1999 , spiral structure in galaxies : a density wave theory , cambridge ma mit press , isbn0262023962 balsara , d.s . , 1995 , journal of computational physics , vol .
2 , p. 357 - 372
berentzen , i . , et al . , 2007 , apj , 666 , 189b block , d.l.,bertin , g.,et al . , 1994 , a&a,288,365b chakrabarti ,
laughlin , g .
, & shu , f.h . , 2003 , apj , 596 , 220c ( cls ) colavitti , e.,et al . , 2008 , a&a,483,401c elmegreen , d.m . & elmegreen , b.g.,1984,apjs,54,127e lin , c.c . & shu , f.h . , 1964 ,
apj , 140 , 646l lin , c.c . , yuan , c. & shu , f.h .
, 1969 , apj , 155 , 721 l goldreich , p . &
lynden - bell , d. , 1965 , mnras , 130 , 125 g hernquist , l. & mihos , c. , 1995 , apj , 448 , 41 h hohl , f . , 1970 , nastr.343,h kendall , s . , et al . , 2008 , mnras , 387 , 1007k kennicutt , r.c . 1983 ,
apj , 272 , 54 kim , w.t . & ostriker , e.c.,2002,apj,570,132k kim , w.t . & ostriker , e.c.,2007,apj,660,1232k laughlin , g.,korchagin , v . &
adams , f.c.,1998,apj,504,945l lubow , s.h .
1983 , apj , 309 , 496l lynden - bell , d .
& kalnajs , a.j .
, 1972 , mnras,157 , 1l martinez - valpuesta , i . , et al . , 2006 ,
apj , 637 , 214 m rix , h . & zaritsky , d . , 1995 ,
apj , 447 , 82r roberts , w.w . & shu , f.h . , 1982 , apl , 12 , 49 r sellwood , j . & carlberg , r.g . , 1984 , apj , 282 , 61s shetty , r . & ostriker , e.c . , 2006 ,
apj , 647 , 997s ( s006 ) shu , f.h . ,
et al . , 1971 , apj , 166 , 465s shu , f.h . ,
milione , v. & roberts , w. , 1973 , apj , 183 , 819s ( smr ) shu , f.h . , et al . , 2007 , apj , 662l , 75s springel , v . , et al . , 2001 ,
new astronomy , vol .
6 , issue 2 , p. 79 - 117 .
springel , v . , 2005 ,
mnras , 364 , 1105s thilker , d.a . , et al .
, 2007 , apjs , 173 , 538 t toomre , a. , 1964 , apj , 139 , 1217 t toomre , a. , 1981 , the structure and evolution of normal galaxies ; proceedings of the advanced study institute , cambridge , england , august 3 - 15 , 1980 .
( a82 - 11951 02 - 90 ) cambridge and new york , cambridge university press , 1981 , p. 111 - 136
tremaine , s . & weinberg , m.d . , 1984 ,
apj , 282l , 5 t wong , t . & blitz , l . , 2002 ,
apj , 569 , 157w wong , t . , blitz , l .
& bosma , a. , 2004 , apj , 605 , 183 w yanez , m . , norman , m . , et al .
, 2007 , arxiv:0710.1331 | we study the dynamical evolution of spiral structure in the stellar disks of isolated galaxies using high resolution smoothed particle hydrodynamics ( sph ) simulations that treat the evolution of gas , stars , and dark matter self - consistently .
we focus this study on the question of self - excited spiral structure in the stellar disk and investigate the dynamical coupling between the cold , dissipative gaseous component and the stellar component .
we find that angular momentum transport from the gas to the stars inside of corotation leads to a roughly time - steady spiral structure in the stellar disk .
to make this point clear , we contrast these results with otherwise identical simulations that do not include a cold gaseous component that is able to cool radiatively and dissipate energy , and find that spiral structure , when it is incipient , dies out more rapidly in simulations that do not include gas .
we also employ a standard star formation prescription to convert gas into stars and find that our results hold for typical gas consumption time scales that are in accord with the kennicutt - schmidt relation .
we therefore attribute the long - lived roughly time steady spiral structure in the stellar disk to the dynamical coupling between the gas and the stars and the resultant torques that the self - gravitating gaseous disk is able to exert on the stars due to an azimuthal phase shift between the collisionless and dissipative components .
galaxies : formation galaxies spiral stars : formation |
the periodic anderson model ( pam ) represents probably the simplest way of modeling some of the rich physics found in lanthanides and actinides@xcite .
one believes that most of the physical properties typical for these materials originate from the interplay between the incompletely filled @xmath1 or @xmath2 shells which contain almost localized electrons , and a broad conduction band of @xmath3 , @xmath4 and @xmath5-electrons . in the periodic anderson model ,
this is simulated by an atomic - like level with strong on - site coulomb interaction which hybridizes with an uncorrelated conduction band .
usually the model is further simplified by assuming both the atomic- and conduction states to be non - degenerate .
the on - site coulomb interaction within the @xmath0-states has to be considered as the largest energy scale of the system since the electrons occupying these states are less screened than the conduction electrons . in the pam ,
the coulomb interaction is further necessary to ensure that the @xmath0-levels are not fully occupied , thus resembling the situation found in the lanthanides and actinides . due to the incompletely filled @xmath0-levels ,
local moments could be formed at every lattice site .
it has been a long - standing question whether these moments order magnetically or whether the local moments are screened by conduction band electrons ( kondo screening)@xcite . in this paper
we will focus on those parameters where the @xmath0 electron density is non - integer ( intermediate valence regime ) . in this parameter regime
, we find a stable ferromagnetic phase and investigate its properties .
kondo screening has been the subject of extensive investigations@xcite . besides few exact statements@xcite ,
several approximate or numerical methods have been applied to explore the possibility of ferromagnetism in the pam , e. g. hartree - fock calculations@xcite , spectral density approach ( sda)@xcite , slave - boson techniques@xcite and dynamic mean - field theory ( dmft)@xcite .
the first methods have severe limitations concerning the description of the low - energy physics .
the dmft - based method promise an improvement at this point .
however , there are still only few results available on ferromagnetism in the pam within dynamical mean - field theory .
the antiferromagnetic phase of the pam has been discussed in more detail@xcite .
it seems to be widely accepted that close to the symmetric parameter conditions , i. e. half - filling and symmetric dos , antiferromagnetism has to be expected .
this can also be concluded from the following : the pam can be mapped onto an effective hubbard model@xcite . in the vicinity of the symmetric point
the effective hubbard model will also be close to half - filling . in that case
the hubbard model is commonly expected to show antiferromagnetic ordering@xcite , therefore the pam will also have a tendency towards antiferromagnetic ordering as was indeed found@xcite .
ferromagnetism is expected further away from this symmetric point .
a ferromagnetic phase was found for the pam both in the kondo regime@xcite and in the intermediate valence region@xcite . whereas in the former case
, the driving force towards ferromagnetism can be understood via an effective heisenberg - like coupling of the @xmath0-spins ( rkky interaction)@xcite , such a picture is not easily applied to the iv regime due to the non - integer density of @xmath0-electrons and their effective itineracy . in this paper
, we will investigate ferromagnetism in the pam using dynamical mean - field theory@xcite .
this theory is based on the assumption of a @xmath6-independent self - energy , which becomes exact in the limit of infinite dimensions ( @xmath7)@xcite . as pointed out in reference , in this limit the lattice coherence and the exchange effects due to the pauli principle are preserved , contrary to other approaches based on @xmath8 expansion ( @xmath9 being the degeneracy of the model)@xcite .
especially the exchange effects should be vital for ferromagnetism , therefore the dmft seems a method of choice . in the next section , the dmft together with the _ modified perturbation theory _ ( mpt ) are introduced .
the dmft will lead to a mapping of the pam onto a single - impurity model , which then will be solved by applying the mpt .
the results concerning zero temperature as well as finite temperatures will be presented and discussed in section [ sec : results - discussion ] .
the periodic anderson model is defined by its hamiltonian @xmath10 @xmath11 ( @xmath12 ) and @xmath13 ( @xmath14 ) are the annihilation and creation operators for an electron in a non - degenerate conduction band state ( localized @xmath0-state ) , and @xmath15 is the occupation number operator for the @xmath0-states .
the dispersion @xmath16 describes the propagation of free , i. e. unhybridized conduction electrons , @xmath17 is the position of the free @xmath0-level relative to the center of mass of the conduction band density of states .
the hybridization @xmath18 is taken as a real , @xmath6-independent constant , and finally @xmath19 is the coulomb repulsion between two @xmath0-electrons on the same lattice site
. the quantity of interest will be the @xmath0-electron green function @xmath20 to determine this function , we employ the dynamical mean - field theory ( dmft)@xcite .
it was shown that in the limit @xmath21 , the self - energy @xmath22 becomes purely local , i. e.@xmath6-independent@xcite . in this case , the self - energy of the pam is equivalent to the self - energy of a properly defined single - impurity anderson model ( siam)@xcite . the latter has to be defined by the so - called self - consistency condition @xmath23 instead of the usual definition @xmath24 for the pure siam . using perturbation theory it has been shown in references that for a three - dimensional system , the local approximation ( equivalent to the limit @xmath7 ) provides already for satisfactory results .
now one is left with the problem to solve the siam defined by equation ( [ eq : selfconsistency ] ) .
here we use the _ modified perturbation theory _
( mpt ) which can be understood as an improvement of the ipt scheme introduced in references .
this method was presented in more detail elsewhere@xcite , so we will restrict ourselves to a short summary here . starting point is the following ansatz for the self - energy@xcite : @xmath25 @xmath26 and @xmath27 are introduced as parameters to be determined later . @xmath28
is the second - order contribution to perturbation theory around the hartree - fock solution@xcite .
the hartree - fock solution introduces another free parameter , namely the chemical potential within this calculation : @xmath29 . _ a priori _ it is not clear whether this should be equal to the chemical potential in the full ( dmft - mpt ) calculation , or whether e. g. it should be determined such that the electron density on the impurity site of the siam is equal both for the hartree - fock and dmft - mpt calculation . in reference and other papers@xcite
yet another condition was used to determine @xmath29 .
there , the luttinger theorem@xcite or equivalently the friedel sum rule@xcite was forced to hold .
since these theorems are applicable only for @xmath30 , this limits the calculations to zero temperature . in order to access finite temperatures
, we used the condition of identical electron densities for the hartree - fock and the full calculation ( @xmath31 ) . with this choice
, the friedel sum rule is still fulfilled in a large parameter region as could be shown for the pure siam@xcite .
a more detailed analysis of the different possibilities to determine @xmath29 is found in reference .
next , the remaining parameters @xmath26 and @xmath27 have to be determined . instead of using the `` atomic '' limit of @xmath32 as was done e. g. in references we make use of the moments of the spectral density @xmath33_-, ... ,h]_- } _ { \text{$n$-fold commutator } } , f_{\sigma}^{\dagger } ] _
+ \rangle \\ a^{(f)}_{\sigma}(e)=-\frac{1}{\pi}\im g^{(f)}_{ii\sigma}(e+i0^+)\nonumber\end{gathered}\ ] ] where @xmath34_-$ ] ( @xmath34_+$ ] ) denotes the commutator ( anticommutator ) of the operators @xmath35 and @xmath36 and @xmath37 denotes the imaginary part of @xmath38 .
as indicated , these can be calculated on two different ways , therefore conditions ( sum rules ) can be derived . to determine @xmath26 and @xmath27 , the first four moments @xmath39 have to be used since the @xmath40- and @xmath41-moments are fulfilled for any @xmath26 and @xmath27 . in the @xmath42-moment
a higher - order correlation function that we will call _ bandshift _ @xmath43 is introduced : @xmath44 with the fermi function @xmath45 .
now , the solution of the siam by the mpt has to be integrated into the dmft - self - consistency loop@xcite : starting with a guessed value of @xmath46 , equations ( [ eq : greenfunction ] ) and ( [ eq : selfconsistency ] ) are evaluated , and then the new mpt self - energy ( [ eq : ansatz ] ) is calculated for the appropriate siam .
this procedure is iterated until a self - consistent solution is found . within this formalism ,
the spontaneous symmetry - breaking necessary for ferromagnetic solutions can be introduced via spin - asymmetric starting values for the self - consistency cycle . in general
, it is not necessary to introduce a symmetry - breaking zero - field . at this point , let us comment on the quality of our approximative method : being based on perturbation theory , the mpt is expected to give reliable results for small interaction strengths . for the symmetric siam , it was even shown@xcite that the perturbative expansion is essentially equivalent to the exact bethe - ansatz solution@xcite and therefore also valid in the strong coupling regime .
for asymmetric parameters , we improve on the perturbation theory by enforcing the correct high - energy expansion for the self - energy and equivalently the green functions .
this is archieved by fulfilling the first four sum rules defined by the spectal moments ( [ eq : moments ] ) which leads to the correct determination of the spin - dependent positions and weights of the charge excitations up the order @xmath47 in accordance with reference @xcite . from this
, we derive our proposition that the mpt can also be reasonably applied to a pam in the intermediate - to - strong coupling regime .
although for the high - energy ( high - temperature ) behaviour this follows from the discussion above , the quality of the special low - energy properties of the pam within the mpt is not known _ a priori_. to estimate the significance of the results for the intermediate coupling strengths , comparison with exact or numerically exact methods is necessary .
for example , one can apply the mpt to a siam without the context of the dmft and compare with the exactly known properties of that model@xcite .
this comparison was done in reference and the results can be summarized as follows : the charge excitations are at the proper positions for symmetric as well as asymmetic parameters . concerning the low - energy behaviour , there is a qualitative agreement but quantitative discrepancies : the friedel sum rule is fulfilled in a large parameter space off the symmetric point , but the kondo temperature does not follow an exponential law .
a power law is found instead .
to summarize , the mpt does qualitatively include essential parts of the kondo physics , but is prone to deviations concerning energy ( temperature ) scales .
further tests of the mpt can be done by comparing results for the pam with different methods based on dmft , as e. g. the numerically exact quantum monte carlo ( qmc)@xcite , exact diagonalization ( ed)@xcite or numerical renormalization group theory ( nrg)@xcite . comparisons of this kind for the paramagnetic pam have been published in reference for qmc and reference for ed calculations .
also the results of reference can easily be compared with above - mentioned references .
again , the qualitative features compare well , however temperature scales concerning low - energy properties can not be reproduced quantitatively .
also , the @xmath18-dependence seems to be overestimated by the mpt when compared to nrg@xcite .
we do believe that these shortcomings of the mpt do not inhibit the analysis of the ferromagnetic properties of the pam . in the discussion of our results
, we will substanciate this claim . in this paper
, we use the following system parameters : the conduction band is described by a semi - elliptic density of states of unit width centered at @xmath48 , thus defining the energy scale .
the position of the @xmath0-level , @xmath17 will be given relatively to the center of gravity of the conduction band .
the hybridization is taken to be constant . unless otherwise noted , we chose @xmath49 .
the coulomb interaction strength @xmath19 is set to @xmath50 for most calculations , thus representing the largest energy in the system .
the total number of electrons @xmath51 , @xmath17 and the temperature @xmath52 will be varied .
the latter will be given in @xmath53 within the energy scale defined by the width of the conduction band .
before discussing the results concerning ferromagnetism in the pam , let us recall some results for the paramagnetic case as they will be important for the following .
the density of states ( dos ) consists of the two charge excitation peaks approximately at @xmath17 and @xmath54 .
in addition , a sharp resonance close to the chemical potential is induced by the correlations , the kondo resonance . due to lattice coherence ,
the kondo resonance is split by the coherence gap in case of symmetric parameters .
another result for the paramagnetic pam found within the dmft - mpt scheme is the existence of local kondo singlets for large hybridization strengths@xcite .
more detailed reviews on the paramagnetic pam can be found in references . in figure
[ fig : phdia ] , the region of ferromagnetic order is plotted in the @xmath17 vs. @xmath51 phase diagram .
we have restricted our evaluation to the so - called intermediate valence region ( iv region ) by positioning the @xmath0-level within the conduction band which extends from @xmath55 to @xmath56 .
this results in a @xmath0-occupancy @xmath57 smaller than unity .
previous investigations have shown that ferromagnetism can exist in this parameter regime@xcite .
the opposite case is the so - called kondo regime , which is obtained by taking @xmath17 well below the lower conduction band edge , and @xmath19 sufficiently large so that @xmath54 is clearly above the conduction band .
this leads to a nearly half - filled @xmath0-level . as discussed in the introduction
, a half - filled @xmath0-level should introduce a tendency towards antiferromagnetism which we have not yet investigated . in the following we will therefore concentrate on ferromagnetism in the intermediate valence region , but we will also be noting some trends on how the ferromagnetic phase would continue into the kondo regime . as can be seen in figure [ fig : phdia ] , the region of ferromagnetic order is quite large , determined by an upper bound for @xmath17 and upper and lower bounds for @xmath51 .
as will be shown below , the ferromagnetic region continues into the kondo regime . in the left panel of figure [ fig : m_vef ] , the @xmath0- ( @xmath3- ) magnetization ,
@xmath58 is plotted as function of @xmath0-level position @xmath17 .
the parameters are as in figure [ fig : phdia ] , the band occupation is given by @xmath59 .
ferromagnetic order breaks down for @xmath17 approaching the center of gravity of the conduction band as it is commonly expected@xcite . within our numerical accuracy ,
this quantum phase transition is of second order . in the other limit ,
i. e. @xmath17 below the lower edge of the conduction band ( @xmath60 ) , the ferromagnetic solution remains stable .
we confirmed that down to @xmath61 the magnetization behaves as one would extrapolate from figure [ fig : m_vef ] .
so , the ferromagnetic phase plotted in figure [ fig : phdia ] continues down into the kondo regime and connects to the ferromagnetic phase found in reference .
there , the authors found a stable ferromagnetic phase for @xmath62 .
the right picture of figure [ fig : m_vef ] shows the magnetization as function of the hybridization strength for @xmath63 , the remaining parameters as in the left panel .
the ferromagnetic order is destroyed by large hybridizations .
this can be understood in terms of local singlet formation .
if the conduction electrons screen the magnetic moment of the localized electrons , as discussed in detail in reference , magnetic ordering will not be possible any more . comparison with nrg results@xcite seem to suggest that the @xmath18-dependence of certain `` low - energy '' quantities is over - estimated by the mpt , as e. g. the size of the coherence gap .
this implies the possibility that the critical @xmath18 is underestimated since the formation of local kondo singlets is due to the low - energy physics of the pam . for @xmath64
the ferromagnetic phase is stable , for numerical reasons no calculations for @xmath65 have been performed . for @xmath32 ,
the @xmath0-level and the conduction band decouple completely .
the @xmath0-level basically corresponds to the zero - bandwidth hubbard model@xcite in which ferromagnetism is not stable . in figure
[ fig : m_vef ] , the thin lines represent the conduction band polarization @xmath66 . in general , it is found to be antiparallel to @xmath67 . however ,
as will be shown below , a parallel alignment of @xmath67 and @xmath66 is also possible for certain parameters . in figure
[ fig : uc ] , the @xmath19-dependence is examined .
the inset shows that ferromagnetism sets in at a critical @xmath68 and saturates quickly afterwards .
the absolute value of @xmath68 varies strongly with @xmath17 : on shifting @xmath17 below the conduction band into the kondo regime , @xmath69 seems to saturate at a value of @xmath70 .
however , there is no true saturation , because when @xmath17 is low enough so that the upper charge excitation is close to or even below the chemical potential , the physics of the system will change again . on increasing @xmath17 , @xmath68 increases as well and exhibits a sharp upturn at @xmath71 . in reference ,
a ferromagnetic phase was found in the kondo regime of the pam for interaction strengths smaller than our @xmath72 .
as for the hybridization , the mpt seems to misjudge the energy scale .
we will argue below that the kondo scale is decisive for ferromagnetism in the kondo regime@xcite , but not in the intermediate valence regime . in figure
[ fig : m_n ] , both the @xmath0- and @xmath3-magnetization are plotted as function of the electron density for various values of @xmath17 below half - filling ( @xmath73 ) . as already seen in the phase diagram , the system is ferromagnetic in a range of values of @xmath51 below half - filling @xmath74 .
both the lower and upper boundary of the ferromagnetic region depend on @xmath17 .
the lower boundary is in fact determined by the number of @xmath0-electrons @xmath57 , which itself depends on @xmath17 and @xmath51 .
the phase transition occurs for each @xmath17 at that specific electron density , where @xmath57 drops below @xmath75 .
approaching the half - filled system , no ferromagnetism is found .
this is similar to the findings in reference .
however as discussed above , antiferromagnetic order is to be expected@xcite in this case .
for @xmath76 the ferromagnetic phase can be separated into two regions of different properties . in the low - density regime
, the @xmath0-polarization tends to saturation , the @xmath3-polarization is positive . on the contrary , in the high - density region ( @xmath77 )
, the @xmath0-polarization is almost constant , below saturation , and the conduction band polarization is of opposite sign ( antiferromagnetic coupling between @xmath0- and @xmath3-band ) .
for more clarity , we will call the first case parallel , the latter case antiparallel coupling of @xmath0- and @xmath3-electrons .
this latter case of antiparallel coupling is generally expected for the pam due to its close relation to the kondo lattice model via the schrieffer - wolff transformation@xcite .
this transformation maps the pam onto a model with an antiferromagnetic interband - spin exchange .
one condition for the schrieffer - wolff transformation is that the @xmath0-level should be half - filled , which is not met here .
however , for large total occupation number , @xmath57 is closer to unity than for small @xmath51
. therefore the finding of parallel coupling for low , and antiparallel coupling for high electron densities is no contradiction to the schrieffer - wolff transformation . in figures [ fig : qdos_para ] and [ fig :
qdos_ferro ] , the densities of states are plotted for various band occupations with @xmath63 , @xmath49 , @xmath50 and zero temperature .
the respective left panel corresponds to a projection onto @xmath0- , the right panel onto @xmath3-states .
the upper charge excitation , located approximately at @xmath78 is not visible in these figures .
the arrows indicate the respective position of the chemical potential . to get more insight into the complicated structure of the dos , we start by explaining figure [ fig : qdos_para ] , where we have forced the system to be paramagnetic . in figure
[ fig : qdos_ferro ] , the stable ferromagnetic dos are shown and will be explained below . starting with the @xmath0-dos in figure [ fig : qdos_para ]
, one observes three structures : the first is the lower charge excitation located roughly at the lower band edge . with increasing electron density , this peak shifts to lower energies .
the next structure is the kondo resonance which is in the vicinity of the chemical potential @xmath79 . due to the asymmetry of the parameters , i. e. the two charge excitation peaks
are not symmetric around the chemical potential and the band occupation is well below @xmath80 , the kondo resonance is not exactly at @xmath81 but slightly above .
this is in agreement with previous findings@xcite and was also observed in the nrg method@xcite and the single - impurity anderson model@xcite . with increased electron density , the kondo resonance moves towards higher energies , this is clearly related to the according shift of the chemical potential @xmath79 .
it is reasonable to expect the kondo resonance to vanish for decreasing @xmath57 . within the mpt ,
the kondo resonance remains stable for all parameters discussed in this paper .
only for much reduced @xmath57 , where @xmath17 is located well above @xmath79 , the kondo resonance disappears . as already mentioned , there is the `` coherence gap '' in the center of the kondo resonance for symmetric parameters . due to the asymmetry of the parameters in figure [ fig : qdos_para ]
, this gap is closed . only for the @xmath82 case ( upper panel of figure [ fig : qdos_para ] ) , there is a slight dip in the @xmath0-dos visible .
this complies with the findings of reference .
the third feature in the @xmath0-dos is a broad structure representing states induced by the hybridization with the conduction band .
a fourth feature is of course also present in the dos , although not plotted in figures [ fig : qdos_para ] and [ fig : qdos_ferro ] : the upper charge excitation which is located roughly at @xmath83 and has a fwhm ( full width at half maximum ) of @xmath84 .
all these structures have their counterpart in the conduction - band dos shown in the right panel of figure [ fig : qdos_para ] .
the charge excitation induces via the hybridization states in the @xmath3-dos .
where the kondo resonance is visible in the @xmath0-dos , the @xmath3-dos shows also some anomalies .
the original conduction band can also be clearly recognized .
let us now discuss how the dos are modified in the ferromagnetic phase . in figure
[ fig : qdos_ferro ] the ( stable ) ferromagnetic dos are plotted for the same parameters as in figure [ fig : qdos_para ] .
the solid lines represent the spin-@xmath85 , the dotted lines the spin-@xmath86 dos . for the lowest panel ( @xmath87 ) ,
the system is paramagnetic . for the ferromagnetic solutions ,
the following picture emerges : firstly , the lower charge excitation is fully polarized , i. e. it consists purely of spin-@xmath85 states .
correspondingly , the upper charge excitation , which is not visible , exists only in the spin-@xmath86 dos .
some interesting changes occur to the kondo resonance : whereas in the spin-@xmath85 channel it disappears almost completely , there is still a peaked structure visible in the spin-@xmath86 channel .
contrary to what one would expect for the kondo resonance , its shift towards higher energies on increasing @xmath51 is stronger than the according shift of @xmath79 .
it is therefore not located at @xmath79 any more .
although this structure clearly develops from the kondo resonance , its physical interpretation ( kondo screening ) can not so easily be transfered .
finally there are some substantial changes in the @xmath3-dos as well : due to the full polarization of the lower charge excitation , the hybridization - induced @xmath3-states in the same energy range are also purely of majority spin .
the feature that corresponds to the kondo resonance is , for similar reasons , only visible in the spin-@xmath86-@xmath3-dos .
finally , in the spin-@xmath85 dos , a new gap appears which separates the lower charge excitation from the remainder of the bare band . due to this new gap ,
the system becomes a semi - metal around quarter - filling ( @xmath88 ) meaning that the spin-@xmath85-dos vanishes at @xmath79 , so only spin-@xmath86-electrons contribute to electrical current .
the new gap , which we call hybridization gap is not equivalent to the previously discussed coherence gap although the latter is of course also induced by the hybridization .
we will explain the differences between these gaps below .
this gap is also present in the @xmath0-dos , but not visible in the plotted figures because of an artificial broadening of the lower charge excitation which is necessary for numerical reasons . from these observations
the change of sign of the conduction band polarization can be understood : in the case of low electron density , the chemical potential lies within the lower charge excitation .
as explained above , due to the hybridization there are only spin-@xmath85 conduction band states available in the occupied region of the dos , the conduction band polarization is necessarily parallel to the @xmath0-polarization .
but on the other hand , with increasing electron density , @xmath79 is increasing as well .
the chemical potential now has to cross the hybridization gap which exists only in the spin-@xmath85 dos . as a result , there are more occupied spin-@xmath86 than spin-@xmath85 states and the conduction band polarization is antiparallel to the @xmath0-magnetization . let us now discuss the difference between the well - known coherence gap and the above - described hybridization gap .
the coherence gap is found for a symmetric pam with on - site hybridization@xcite .
contrary to the single - impurity case , a gap develops in the center of the kondo resonance located at @xmath81 . shifting away from symmetric parameters ,
this coherence gap gets suppressed , although a dip within the kondo resonance is still visible for a large parameter region@xcite .
so why is the new gap observed in figure [ fig : qdos_ferro ] not this coherence gap ?
this question is answered by the fact that the new gap does not shift with @xmath79 on increasing the electron density .
although there is no proper kondo resonance in the case of ferromagnetic order , those structures that clearly develop from it do shift with increasing @xmath51 . within this structure , even a dip which is the relic of the coherence gap is visible for @xmath82 .
the position of the new gap is neither related to @xmath79 nor the relics of the kondo resonance , it is rather located at @xmath17 . and what is the origin of the new gap ?
since its position is determined by the value of @xmath17 and its width scales with @xmath18 , it reminds of the gap found e. g. within the sda@xcite or alloy analogy@xcite , but also for the @xmath89 case of the pam which is exactly diagonalizable@xcite .
this indicates that this gap is due to level - repulsion between the lower @xmath0-peak and the conduction band , as the coherence gap can be interpreted as level - repulsion between a virtual @xmath0-level at the fermi energy and the conduction band .
to distinguish between these gaps , we call the former `` hybridization '' and the latter `` coherence '' gap . at first sight
, it surprises that this gap was not found in any paramagnetic dmft - based calculation@xcite .
this is in our opinion simply due to damping effects caused by the correlation , i. e. a finite imaginary part of the self - energy . in case of saturated ferromagnetism ,
@xmath90 vanishes and the corresponding spin-@xmath85 dos should resemble the @xmath89 dos including the hybridization gap at @xmath17 . in the spin-@xmath86- or also the paramagnetic dos , all structures are broadened due to the finite damping .
this effectively closes the hybridization gap for these dos .
this proposal can be supported by investigating the temperature - dependence of the dos . on increasing the temperature , the magnetization will decrease .
the system is pushed away from saturation , damping effects occur also for spin-@xmath85 electrons . in figure
[ fig : qdos_ferro_temp ] , the dos for @xmath91 are plotted for various temperatures ( the remaining parameters as in figure [ fig : qdos_ferro ] ) . for these parameters ,
the curie temperature is @xmath92 .
one can see how the gap closes on approaching @xmath93 .
more complicated is the case of @xmath94 in figure [ fig : qdos_ferro ] . on the one hand , ferromagnetic saturation
is not reached ( cf .
figure [ fig : m_n ] ) and the hybridization gap is partly closed by damping . and
on the other hand , the chemical potential is located in the same energy range , further complicating the situation . in the following
we will discuss the temperature dependence in more detail .
figure [ fig : m_t ] shows magnetization curves for @xmath91 and three different values of @xmath95 ( from top to bottom ) .
the lowest line ( @xmath96 ) shows a clear second - order phase transition .
the middle case corresponds to the data of figure [ fig : qdos_ferro_temp ] .
the phase transition at @xmath92 is within numerical accuracy of second order .
the demagnetization process is dominated by a transfer of spectral weight from the minority- to the majority - spin spectrum of the lower charge excitation peak as can be seen in figure [ fig : qdos_ferro_temp ] .
however , for the upper charge excitation , plotted in figure [ fig : qd_t_upper ] for the same parameters as figure [ fig : qdos_ferro_temp ] , not only the reverse transfer of spectral weight occurs but also a spin - dependent bandshift can be observed . when decreasing @xmath17 towards the lower band edge of the conduction band , the phase transitions become sharper , and
finally , first order phase transitions are found for @xmath97 within numerical accuracy .
this is similar to findings for the hubbard model@xcite , where phase transitions were found to be of second order for lower , and of first order for higher electron densities . in the pam
, the density of correlated electrons , @xmath57 , increases on lowering @xmath17 , therefore the behaviour is analogous to the findings for the hubbard model . as discussed in reference , we can not exclude that this is an artifact of our method . in figure
[ fig : tc_ef ] , the curie temperature is plotted as function of the @xmath0-level position .
the trends are very similar to the @xmath30-magnetization as seen in figure [ fig : m_vef ] . for @xmath17 approaching the upper critical value , both @xmath93 and @xmath98 vanish . in the opposite direction ,
both quantities keep increasing , however , with diminishing slope .
the curie temperature as function of band occupation is shown in figure [ fig : tc_n ] for @xmath63 and @xmath96 .
as already mentioned in the introduction , the magnetism in the kondo limit of the pam can be understood in terms of an effective rkky interaction@xcite between the @xmath0-spins . however , in the intermediate - valence regime under investigation here , the effective rkky exchange should be suppressed due to the non - integer @xmath0-electron filling and the effective itineracy of the @xmath0-electrons as indicated by the broader @xmath0-peaks in the dos .
we believe that instead of the interband - rkky exchange , a `` narrow - band '' mechanism is the main driving force towards ferromagnetism in the iv regime .
the effective itineracy of the @xmath0-electrons due to the hybridization in combination with the strong on - site coulomb interaction leads to the strong similarities to the well - known hubbard model@xcite .
recently the existence of ferromagnetism was shown within the hubbard model@xcite .
one important condition for it is the shape of the free , i. e. @xmath89 dos entering the model .
a highly asymmetric dos with a divergence or sharp peak close to the band edge enhances the possibility of ferromagnetic order@xcite . turning back to the periodic anderson model we note that this condition favoring ferromagnetic order is excellently met for the hybridization - broadened @xmath0-level in combination with the hybridization - induced @xmath0-states within the conduction band . our proposal ,
that ferromagnetism in the iv regime of the pam originates from an intraband mechanism is supported by several findings : \i ) whereas in the kondo regime , small values of @xmath19 can already lead to a ferromagnetic phase , in the iv regime , significantly larger values of @xmath19 are necessary .
\ii ) the particular low - energy physics , that are decisive for ferromagnetism in the kondo regime , have no significant influence on the magnetism in the iv regime .
\iii ) the critical @xmath51 , where ferromagnetism breaks down , is in fact determined by a critical number of correlated @xmath0-electrons , which is of similar magnitude as for the hubbard model .
\iv ) the polarization of the conduction band seems to have almost no influence on the magnetic properties in the iv regime .
let us discuss these points in more detail : the first point follows directly from figure [ fig : uc ] and reference . in the kondo regime , where the low - energy physics determine the magnetic properties@xcite
, the mpt probably overestimates the absolute value of @xmath68 .
however , the sharp increase of @xmath69 for increasing @xmath17 indicates the importance of a strong - coupling mechanism in the iv regime .
the second point can be made clear by comparing the mpt results with those obtained by two other methods , namely the spectral density approximation ( sda)@xcite and the modified alloy analogy ( maa)@xcite . whereas all three methods do fulfill the first four moment - sum rules ( [ eq : moments ] ) important for the strong - coupling limit , the mpt surely represents a major improvement over the other two methods concerning the low - energy physics .
the sda and the maa completely neglect the kondo physics from the very beginning while the mpt includes , as discussed in section [ sec : dmft - modif - pert ] , at least qualitatively the special low - energy properties of the pam .
now , if the low - energy physics have a major influence on ferromagnetism in the iv regime of the pam , one would expect a dramatic change of the properties of the ferromagnetic phase .
however , in all three methods , the ferromagnetic phase turns out to be very similar , key features as e. g. the phase diagram , but also the unusual conduction band polarization are present in all approximations .
this clearly indicates the minor importance of the special low - energy physics of the pam when describing ferromagnetism in the iv regime .
the contrary seems to be the case for the kondo regime : here the sda does not give a stable ferromagnetic phase@xcite , whereas within the mpt , the ferromagnetic solution remains stable in the kondo regime and the parameters leading to ferromagnetism agree very well with the results of reference if one ignores the already discussed overestimation of @xmath68 .
the third point can be clarified by pointing out that for an rkky mechanism , the density of the conduction band electrons should have a major influence on the effective exchange integrals , and therefore also on phase boundaries . however , as discussed in the context of figure [ fig : m_n ] , the lower phase boundary as function of @xmath51 is governed by the number of @xmath0-electrons and is independent of @xmath99 .
this is the expected behaviour for an intraband mechanism as proposed by us .
it is clear that for the kondo regime with fixed @xmath100 , the situation is different .
finally let us discuss the influence of the conduction band polarization .
it is clear that it is decisive for an rkky mechanism .
however , our results indicate that the magnitude and the unusual behaviour of @xmath66 have no direct effect on the stability of ferromagnetism in the iv regime ( cf
. figures [ fig : m_vef ] , [ fig : m_n ] and [ fig : tc_n ] ) .
this is a clear indication that an rkky mechanism is suppressed here . to summarize the discussion ,
we believe that two different mechanism are to be considered when discussing ferromagnetism in the pam : in the kondo regime , an effective rkky exchange ( interband mechanism ) is mainly responsible for ferromagnetism , in the iv regime , the rkky exchange is suppressed and an intraband mechanism similar to the one found for ferromagnetism in the single - band hubbard model should be of major importance . from this proposal
it follows that the bandshift ( [ eq : band - shift ] ) , which is introduced by the fourth moment of the spectral density , should also be decisive for ferromagnetism in the pam in the iv regime as it was shown to be in the hubbard model@xcite .
the low - energy physics , however , are vital in the kondo regime since the rkky exchange is believed to origin in the formation of kondo screening clouds@xcite but less important in the iv regime
. therefore the shortcomings of the mpt in reproducing the low - energy scale of the pam are nonrelevant for the present investigation .
of course , both the rkky and the intraband mechanism are always present , only each respective importance varies .
whereas in the kondo limit , the effective @xmath0-electron itineracy becomes negligible , in the iv regime , strong - coupling `` hubbard - like @xmath0-band '' picture is dominant .
this explains the smooth transitions found between the to scenarios ( cf.figures [ fig : m_vef ] , [ fig : uc ] and [ fig : tc_ef ] ) .
in this paper we have presented a detailed investigation of ferromagnetism in the intermediate valence regime of the periodic anderson model .
we applied the dynamical mean - field theory in combination with the modified perturbation theory .
we have shortly discussed the quality of the mpt based on a comparison with exact results for the siam and numerically exact results for the paramagnetic pam .
although the mpt can not recover the correct exponential energy scale for the low - energy physics , the qualitative features of the low - energy physics emerge properly . and
the high - energy features are reproduced with much higher credibility .
we have established the phase diagram as function of the total electron density and the position of the @xmath0-level .
furthermore , an upper bound of the hybridization strength was found , above which no magnetic ordering seems to be possible due to local kondo singlet formation@xcite .
the critical interaction strength @xmath68 was found to vary strongly when moving from the intermediate valence to the kondo regime .
the main contribution to the magnetic moment originates from the localized @xmath0-electrons .
however , the conduction band is , due to the hybridization , also polarized . as function of electron density , the conduction band polarization changes sign . for higher electron densities , the conduction - band polarization is antiparallel to the @xmath0-level magnetization @xmath67 as one would expect from the schrieffer - wolff transformation . for low densities ,
this transformation is not applicable . here , the conduction band polarization is parallel to @xmath67 .
the densities of states in the ferromagnetic state show rather complicated structures . besides relics of the kondo resonance and the dip corresponding to the coherence gap ,
another gap is present which we named hybridization gap .
this gap is closed when approaching the paramagnetic state .
investigating the temperature dependence , we find phase transitions of second and first order . we also have presented the curie temperatures as function of @xmath0-level position and total electron density . concerning ferromagnetism in the kondo and intermediate valence regime , two different pictures
emerge : whereas in the kondo regime , an kondo - screening induced rkky exchange leads to the ordering of the @xmath0-spins already for relatively low interactions strengths@xcite , in the intermediate - valence region , another mechanism becomes more important : due to the effective itineracy of the @xmath0-electrons , the @xmath0-levels now represent an strongly correlated narrow band , which by itself can lead to ferromagnetism as known for the hubbard model .
d. meyer , diploma thesis humboldt - universitt zu berlin may 1997 .
( http://tfk.physik.hu-berlin.de/
although a ferromagnetic solution can be found within the sda for @xmath17 below the conduction band , its free energy is then much larger than the one of the paramagnetic solution . | the ferromagnetic phase diagram of the periodic anderson model is calculated using dynamical mean - field theory in combination with the modified perturbation theory .
concentrating on the intermediate valence regime , the phase boundaries are established as function of the total electron density , the position of the atomic level and the hybridization strength . the main contribution to
the magnetic moment stems from the @xmath0-electrons .
the conduction band polarization is , depending on the system parameters either parallel or antiparallel to the @xmath0-magnetization . by investigating the densities of states
, one observes that the change of sign of the conduction band polarization is closely connected to the hybridization gap , which is only apparent in the case of almost complete polarization of the @xmath0-electrons .
finite - temperature calculations are also performed , the curie temperature as function of electron density and @xmath0-level position are determined . in the intermediate - valence regime , the phase transitions are found to be of second order . |
polyelectrolytes ( pes ) , such as dna molecules and many synthetic polymers , are macromolecules carrying a large number of charged groups which are ionizable in water . the presence of long - ranged electrostatic interaction and the confinement of charges on chains , together with the degree of freedom of chain conformation , make this kind of system very complicated and unpredictable .
the behavior of pes is greatly influenced by ions in the solutions , owing to the screening effect and ionic redistribution .
for example , addition of salt in the solutions , especially multivalent salt , can change drastically the properties of pes @xcite .
when salt is added , the multivalent counterions dissociated from the salt can compete with the monovalent ones in the solutions and condense onto pe chains , which eventually leads the system to a series of phase transition .
the condensation of multivalent counterions is a process of both energetic and entropic favor because a @xmath0-valent counterion ( @xmath1 ) attracting to a pe chain is @xmath0 times stronger than a monovalent one and the release of @xmath0 monovalent counterions into the solution , due to the condensation of the @xmath0-valent counterion , also increases the system entropy @xcite .
a phase separation takes place when salt concentration is higher than a critical value .
the chains collapse and aggregate in the solution , resulting in a precipitation from the solution .
when the salt concentration is further increased and goes beyond a second critical value , the precipitated pes redissolve back into the solution . the system returns to a homogeneous phase @xcite .
the phase transition occurred at the first critical salt concentration is a result of ionic screening and bridging of multivalent counterions bound near pes @xcite .
the transition occurred at the second one exists several explanations @xcite .
one of them attributes the resolubilization of pes to the occurrence of charge inversion of pes @xcite .
it argues that the condensed multivalent counterions can overcompensate the chain charge when salt concentration is high , and the repulsive force between chain segments due to the charge of inverted sign at that moment causes the segregation and resolubilization of chains .
another theory considers the bjerrum association between ions @xcite . by using a two - state model of chain
, it predicts the possibility of either sign of the effective chain charge under different conditions .
these theories usually over - simplifies the situations and are constructed based upon certain preassumed chain conformation and ion distribution .
further verification is thus needed . to understand the systems , it is important to study chain conformation and ion distributions around a chain in solutions .
we have investigated the first issue , the effect of chain stiffness on pe chain conformation , in our previous paper @xcite . in this work ,
we focus on the second issue : the ion distribution .
to investigate ion distributions directly from experiments is difficult .
computer simulation , on the contrary , can do it easily because the positions of atoms are explicitely given in the study .
this issue can be investigated on different scales @xcite .
an all - atom model , which simulates molecules with detailed chemical composition and structure , are usually applied in the study of small systems for short simulation time , owing to high computational cost related to today s computer power @xcite .
space and time scales are both restricted . a coarse - grained model , which reduces the complexity of the molecular structure by grouping atoms , on the other hands , can decrease significantly the computational cost .
researchers can hence investigate larger systems evolving over a longer simulation time . using coarse - grained models ,
scientists have investigated ion distribution and effective interaction and free energy of dna molecule systems @xcite .
the chains were modeled by an infinitively long post through periodic boundary condition .
this model preserved the dna helical structure and can be used to study local structural transitions , such as the b - to - z transition , in addition to the ion distribution .
however , it fails to characterize the global chain conformation , as what happened in dna condensation when polyvalent ions are added into the solution . to study the collapse and the reexpansion of single pe chains ,
bead - spring chain models are usually employed @xcite .
this kind of simple models allows people to investigate the behavior of pe systems in a large space and time scale . by exploiting it
, researchers have demonstrated that various conformational transitions of pe chains take places under different conditions .
the models could be hence good candidates in the study of ion distribution related to the chain conformations .
liu and muthukumar showed that a chain can be collapsed in a salt - free solution when coulomb coupling strength is strong @xcite .
their study reveals the relation between counterion condensation and the formation of local dipole moment .
hsiao and luijten studied how pe chains are collapsed in multivalent salt solutions @xcite .
they found that the chains are charge overcompensated at their very surface when the amount of the multivalent counterions of salt is greater than the amount needed to neutralize the bare chain charge @xcite .
counterion condensation on a pe chain has also been studied by several groups @xcite . however , these studies aimed to understand the collapse transition of chain happened in salt - free or low - salt conditions .
the information concerning the unfolding transition of chain provoked by high concentration of counterions was missing . in this work ,
we plane to study the ion distribution around a pe chain in solutions over a wide range of salt concentration , so wide that single chains can subsequently undergo collapse and reexpansion transitions , using a bead - spring chain model
. the focus will be the exploitation of the ion distribution at different salt concentrations and the relation with the chain conformation .
it is known that the structure of collapsed pe chains depends sensitively on the chain rigidity @xcite .
studies also showed that the rigidity determines the transition of chain conformation to be continuous or discrete @xcite .
it is , therefore , very relevant to know how ion distributions are influenced by the chain rigidity in the problem of pe condensation .
ou and muthukumar observed that the number of the condensed monovalent counterions on a chain decreases in a salt - free solution when chain stiffness increases @xcite .
sarragua and pais studied pe chains collapse by trivalent salt in different circumstances @xcite .
they found that the density of trivalent ions around stiff chains is smaller than around flexible ones .
moreover , flexible chains are compacted into smaller dimensions and the ion correlation is strong due to small volume of the confinement . nonetheless , the situation for high salt concentrations where chain reexpansion occurs was not discussed .
this paper is organized as follows .
we describe our model and the simulation setup in section [ sec_model_setup ] .
the results and discussions are given in section [ sec_results ] .
after a short revisit of the dependence of chain size and morphology on the salt concentrations for flexible , semiflexible and rigid chains in section [ subsec_chain_size ] , we investigate various ion distributions in pe solutions .
the topics include the integrated ion distribution in the radial direction of a chain ( section [ subsec_ion_dist_in_chain_radial_direction ] ) , the charge distribution ( section [ subsec_charge_dist_in_chain_radial_direction ] ) , the ion condensation and the effective chain charge ( section [ subsec_ion_condensation_&_effective_charge ] ) , the condensed ion distribution along a chain backbone ( section [ subsec_dist_cond_ion_along_chain ] ) , and the radial distribution functions ( section [ subsec_rdf ] ) . by investigating these topics
, we can make connection between ion distributions and chain conformations and understand the role of chain stiffness in different salt regions .
moreover , the results obtained from the rigid - chain case provide a direct verification of the manning condensation theory by simulations .
we give our conclusions in section [ subsec_conclusions ] .
our system consists of a pe chain and tetravalent salt molecules , placed in a cubic box with periodic boundary condition .
the chain dissociates into a polyion and many counterions .
the polyion is modeled by a bead - spring chain and comprises 48 beads .
each bead carries a negative unit charge @xmath2 .
the counterions are modeled by charged spheres , carrying a @xmath3 charge each .
the tetravalent salt dissociates into tetravalent cation ( counterions ) and monovalent anions ( coions ) .
these ions are also modeled by charged spheres .
solvent molecules are not explicitly incorporated in the study . nonetheless , we consider the entire solvent as a dielectric medium with dielectric constant equal to @xmath4 .
we consider two kinds of non - bonded interactions : the excluded volume interaction and the coulomb interaction , applied to all the beads on the chain and spheres in the system , and two kinds of bonded interactions : the connectivity and the chain rigidity , applied solely to the chain .
the excluded volume is modeled by a purely repulsive lennard - jones ( lj ) potential @xmath5 + \varepsilon_{lj } & \mbox{for } r \leq 2^{1/6 } \sigma\\ 0 & \mbox{for } r > 2^{1/6 } \sigma \end{array}\right.\]]where @xmath6 is the distance , @xmath7 is the interaction strength and @xmath8 represents the diameter of a particle .
we assume identical @xmath7 and @xmath8 for the ion spheres and monomer beads .
the coulomb interaction between two charged ions or beads is given by @xmath9where @xmath10 and @xmath11 are charge valences , @xmath12 is the boltzmann constant , @xmath13 is the temperature , and @xmath14 is the bjerrum length ( @xmath15 is the vacuum permittivity and @xmath4 the solvent dielectric constant ) .
the connectivity of two consecutive beads on the chain is described by a finitely extensible nonlinear elastic potential @xmath16where @xmath17 is the length of a bond , @xmath18 is the maximum bond extention , and @xmath19 is the spring constant .
the chain rigidity is represented by a harmonic angle potential , @xmath20where @xmath21 is the bond angle between two consecutive bonds and @xmath22 is the force constant .
in addition to the above interactions , two more forces are exerted on each particle .
they are ( 1 ) a dissipative force , which takes into account the friction force due to the particle moving through the solvent , and ( 2 ) a stochastic force , which models the thermal collisions by solvent molecules .
the equation of motion , also known as the langevin equation , reads as @xmath23where @xmath24 is the mass of particle , @xmath25 is the friction coefficient , and @xmath26 is the stochastic force . we have assumed identical @xmath24 and @xmath25 for all the particles .
the system reaches an equilibrium state with the temperature given by the fluctuation - dissipation theorem : @xmath27where @xmath28 and @xmath29 are the kronecker and the dirac delta functions , respectively .
we set @xmath30 , @xmath31 , @xmath32 and @xmath33 .
@xmath25 is set to @xmath34 ( where @xmath35 ) to mimic aqueous solutions .
these setups give a linear charge density of chain corresponding to a value typically for many polyelectrolyte systems , such as sodium polystyrene sulfonate .
we varied @xmath22 over a large range from 0 to @xmath36 to study the effect of chain stiffness .
the values cover three typical chain stiffness : flexible chains , semiflexible chains , and rigid chains ( see explanation in section [ subsec_chain_size ] ) . the ( 4:1)-salt concentration @xmath37 is varied from @xmath38 to @xmath39 whereas the monomer concentration @xmath40 is fixed at @xmath41 .
the size of the simulation box is @xmath42 , which is large enough to avoid chain overlapping under periodic boundary condition .
therefore , the studied case is a dilute pe solution .
it has been shown in experiments @xcite and simulations @xcite that pes show similar condensation and decondensation behavior in the dilute region .
the largest system in our study contains totally 7776 charged particles .
we list all the values of @xmath37 investigated and the corresponding numbers of tetravalent counterions and monovalent coions in table [ tab : ion_no ] for reader s reference .
.values of @xmath37 and the corresponding numbers of tetravalent counterions and monovalent coions in the simulation box [ cols="^,^,^",options="header " , ] we perform langevin dynamics simulations with the integrating time step equal to @xmath43 . the coulomb interaction is calculated by the technique of ewald sum .
an equilibration phase takes about @xmath44 time steps of run , followed by a production phase cumulating data for @xmath44 to @xmath45 time steps . in order to sample sufficiently the data ,
several independent runs are performed at each salt concentrations , starting with different initial configurations . to simplify the notation , in the rest of the text , length , energy , and time
will be measured in unit of @xmath8 , @xmath46 , and @xmath47 , respectively .
therefore , concentration will be described in unit of @xmath48 and @xmath49 in unit of @xmath50 , and so on .
in this section , we shortly revisit some of the major findings obtained in our previous work @xcite , how the chain size and morphology depend on the chain stiffness in multivalent salt solutions , to give readers a thorough picture of the system .
the detailed discussions can be found in the paper @xcite .
starting from the next section , we will study ion distributions and show that they depend strongly on the size and the morphology of pe chains in solutions .
instead by using the radius of gyration , we characterize here the chain size by the hydrodynamic radius , @xmath51 , which is the stokes radius of a polymer moving through a solution @xcite .
@xmath51 is calculated by the equation @xmath52 .
the results , @xmath51 vs. @xmath37 , for different chain stiffness @xmath49 are presented in fig .
[ f : rhvcs ] . as a function of tetravalent salt concentration @xmath37 .
snapshots of some typical chain structures at the points ( i ) to ( vii ) are shown by the side of the figure.,scaledwidth=70.0% ] we can see that the curves show different variations against @xmath37 , depending on the chain stiffness .
these variations can be classified into three types .
the first type is attributed to the case of small @xmath49 such as @xmath53 .
the chain is said _
flexible _ and @xmath51 in the semi - log plot looks similar to a tilted - l curve .
the second type is addressed to intermediate @xmath49 , such as @xmath54 .
the chain is called _ semiflexible _ , for which the @xmath51 curve takes a shape similar to the letter u. the third type is ascribed to large @xmath49 like @xmath55 .
the chain is very stiff and there is basically no variation against @xmath37 ; we are in the case of _ rigid _ chains .
snapshots for these three types of variations are shown by the side of the figure . for a flexible chain ( snapshots ( i ) to ( iii ) ) , the chain size decreases gradually as @xmath37 increases up to @xmath56 ; the chain undergoes a continuous coil - to - globule transition .
the chain size then turns to increase slowly when @xmath57 , indicating a swelling transition .
@xmath56 is found equal to @xmath58 , the equivalence point . at this particular salt concentration ,
the added tetravalent counterions are in charge equivalence with the bare charge of the pe chain .
we observed that the chain displays a disordered , globule structure at @xmath56 . for a semiflexible chain ( snapshots ( iv ) to ( vi ) ) , two sharp transitions take place at low and at high salt concentrations , respectively .
they are , in turn , the coil - to - globule and the globule - to - coil transitions .
it has been shown that these two transitions occur in a discrete manner @xcite , in contrast to the continuous transitions happened for a flexible chain . in the region between the two salt concentrations ,
the chain is in a collapsed state and exhibits a compact ordered structure , such as a toroidal structure or a folded - chain structure . outside the salt region
, the chain structure becomes extended .
this kind of phenomena is called `` reentrant condensation '' and has attracted much attention in the study of dna condensation @xcite .
for the case of a rigid chain ( snapshot ( vii ) ) , the chain is so stiff that it can not be folded by the condensed multivalent counterions .
the chain stays elongated , no matter how many salts are added in the solution .
in addition to the chain size and morphology , which are global properties of chain in solutions , we studied also local properties such as the average bond length .
this information tells us how the bond length changes in different conditions , which is directly related to the line charge density of a pe chain , and is important , later on , in the comparison with the manning condensation theory .
the average bond length @xmath59 as a function of @xmath37 is shown in fig . [
f : blvcs ] . as a function of @xmath37 for different chain stiffness @xmath49 .
the values of @xmath49 are indicated in the figure .
the data for @xmath60 are also plotted in the dashed box on the left - top side of the figure.,scaledwidth=70.0% ] the chains of different stiffness show similar variational behavior .
@xmath59 decreases with increasing @xmath37 when @xmath61 .
it then stays basically at a constant for @xmath57 .
the maximum reduction of @xmath59 is less than @xmath62 .
we remark that this decreasing in @xmath59 is a result of the condensation of multivalent counterions on a pe chain , which form triplets with monomers ( as we will see in section [ subsec_rdf ] ) and hence pull closer the distance between two neighboring monomers via the coulomb interaction . in the region @xmath57 , although the chains are charge overcompensated ( cf . fig .
[ f : tube_rc3 ] ) , the extra number of the condensed ions can not pull closer the monomer - monomer distance anymore due to the excluded volume effect .
@xmath59 is thus saturated at a constant value .
combining with the information that @xmath51 increases when @xmath57 , we know that the bond length is not responsible for the global chain reexpansion in this region . after knowing the morphology of pes in different salt and chain stiffness conditions , we study now different kinds of ion distributions and their relation with the chain morphology .
we focus firstly on the ion distribution in chain radial direction . a tube region around a chain
is defined as the union of spheres of radius @xmath63 , centered at each monomer center .
the average number of ions inside this tube region is calculated for each ion species . by varying @xmath63
, we are able to investigate the integrated ion distribution @xmath64 around the chain radially .
the results for monovalent counterions , tetravalent counterions , and coions are shown in fig .
[ f : iondist3 ] .
( a ) for monovalent counterions ( ( + 1)-ion ) , tetravalent counterions ( ( + 4)-ion ) , and coions ( ( -1)-ion ) , around a chain at @xmath65 ( a ) , @xmath66 ( b ) , and @xmath67 ( c ) .
the values of chain stiffness @xmath49 are indicated in the figure .
, title="fig:",scaledwidth=50.0% ] ( b ) for monovalent counterions ( ( + 1)-ion ) , tetravalent counterions ( ( + 4)-ion ) , and coions ( ( -1)-ion ) , around a chain at @xmath65 ( a ) , @xmath66 ( b ) , and @xmath67 ( c ) .
the values of chain stiffness @xmath49 are indicated in the figure . , title="fig:",scaledwidth=50.0% ] ( c ) for monovalent counterions ( ( + 1)-ion ) , tetravalent counterions ( ( + 4)-ion ) , and coions ( ( -1)-ion ) , around a chain at @xmath65 ( a ) , @xmath66 ( b ) , and @xmath67 ( c ) .
the values of chain stiffness @xmath49 are indicated in the figure . ,
title="fig:",scaledwidth=50.0% ] fig .
[ f : iondist3](a ) shows typical ion distributions in the low salt region , @xmath68 . manning condensation theory @xcite states that in salt - free solutions , if the value of @xmath69 , defined as the ratio of the bjerrum length @xmath70 to the charge distance on the chain backbone @xmath71 , is greater than unity , the condensed counterions will neutralize the fraction @xmath72 of charge on the rigid pe chain . according to this theory , there should be about @xmath73 of the monovalent counterions , or equivalently 30 monovalent counterions , condensing on the pe chain at @xmath74 in our simulation condition ( @xmath75 and @xmath76 ; thus , @xmath77 ) . in this figure
, we see only about 20 monovalent counterions condensing on the chain ( inside the tube region with @xmath78 ) .
this is because the tetravalent counterions added in the solution compete with the monovalent ones and replace part of the condensed monovalent counterions on the chain .
therefore , the number of condensed monovalent counterions is smaller than that in the sat - free solution .
moreover , we observed that when @xmath79 , the integrated number @xmath64 for tetravalent counterions is a constant .
this result indicates that nearly all the tetravalent counterions condense on the chain in this low salt region .
[ f : iondist3 ] , panels ( b ) and ( c ) , show typical ion distributions in the middle ( @xmath80 ) and in the high ( @xmath57 ) salt regions .
we saw that @xmath64 for monovalent counterions is almost zero inside the tube @xmath81 .
thus , the condensed ions are now tetravalent in majority . for tetravalent counterions , @xmath64 is no more a constant when @xmath79 .
there are hence non - condensed tetravalent counterions presented in the bulk solution .
form the above three figures , we can see that the ion distributions depend strongly on the chain morphology .
for example , at low @xmath37 ( fig .
[ f : iondist3](a ) ) , the morphology of the flexible chain ( @xmath53 ) is an extended coil , whereas the semiflexible chain ( @xmath54 ) is elongated and rodlike , resembling the rigid chain ( @xmath55 ) .
therefore , the ion distribution curves are nearly identical for the two cases @xmath54 and @xmath55 .
similar results are found at high @xmath37 ( fig .
[ f : iondist3](c ) ) where the chains reenter into the solution . in that case , the reentered semiflexible chains is as extended as a rod , different from the coil structure of a flexible chain . as a consequence , the @xmath64 curve is distinguishable to that for the flexible chain , but similar to the rigid chain . in the mid - salt region
[ f : iondist3](b ) ) , the flexible and the semiflexible chains are collapsed , due to the condensation of tetravalent counterions , but the collapsed chain structures are different : a disordered globule structure for the former chains and an ordered structure for the latter .
@xmath64 curves are therefore non - identical for the three chain stiffness .
the ion distributions for the semiflexible chain ( @xmath54 ) were further investigated in detail in the middle salt region because in this region the chains can collapse into one of the two typical structures : toroid and hairpin structures .
we observed that ( in fig .
[ f : iondist - semi](a ) ) toroidal chains and hairpin chains show similar ion distributions .
( a ) for the semiflexible chains ( @xmath54 ) ( a ) at @xmath82 where the chains are of toroid and of hairpin structures , ( b ) at @xmath83 where the chains are in an extended ( ext . ) state and in a collapsed ( coll . ) state.,title="fig:",scaledwidth=50.0% ] ( b ) for the semiflexible chains ( @xmath54 ) ( a ) at @xmath82 where the chains are of toroid and of hairpin structures , ( b ) at @xmath83 where the chains are in an extended ( ext . ) state and in a collapsed ( coll . ) state.,title="fig:",scaledwidth=50.0% ] the reason is that the condensed tetravalent counterions bridge between chain segments in these two cases to force the chains to form toroid or hairpin .
consequently , the ion distributions look similar along the viewpoint of the chain radial direction .
moreover , when the salt concentration is close to the globule - to - coil transition , our previous study showed that the semiflexible chains can alter between three structures : the toroid , the hairpin and the extended - rod structures @xcite . since the toroid and the hairpin structures give a similar distribution of ions around a chain , it is relevant to verify whether or not the @xmath64 curve for the extended - rod chain structure is also distinguishable to the collapsed chain structures ( toroid and hairpin ) .
the results are presented in fig .
[ f : iondist - semi](b ) and significant differences are observed .
we found that the more the chain morphology is opened or extended , the larger the @xmath64 will be .
this is because more space is available for ions to condense or to approach to a chain when the chain has an opened structure .
for example , in fig .
[ f : iondist3](b ) , the elongated chains ( @xmath55 ) have a larger value of @xmath64 than the toroid or hairpin chains ( @xmath54 ) , and than the chains collapsed into disordered globule structure ( @xmath53 ) .
similar results have been also observed in figs .
[ f : iondist3](a ) , [ f : iondist3](c ) and [ f : iondist - semi](b ) .
the above evidences show that ion distributions around a chain are determined by chain morphology , while chain morphology depends on chain stiffness and salt concentration . using the information obtained in the previous section , by summing up all of the charges inside a tube region around a chain , we studied the integrated charge distribution @xmath84 .
the results are shown in fig .
[ f : q_m4 + ] for the flexible , the semiflexible , and the rigid chains , over a broad range of salt concentration . for the rigid ( @xmath55 ) , the semiflexible ( @xmath54 ) , and the flexible ( @xmath53 ) chains .
the number ( @xmath85 ) near each curve is the salt concentration @xmath37.,scaledwidth=70.0% ] we observed that when @xmath86 , @xmath84 is a monotonic increasing function of @xmath63 with the value increasing from @xmath87 , the bare charge of chain , to @xmath38 , due to the electro - neutrality of the system . in this salt region ,
the value of @xmath84 increases with @xmath37 at a given @xmath63 .
when the system is at the equivalence point @xmath56 , the curve @xmath84 turns out to be zero outside the region of chain excluded volume ( @xmath88 ) .
this result tells us an important information that the condensed ions completely neutralize the chains such that the total charge inside the tube region is zero . when @xmath89 , @xmath84 starts to show a positive peak near the chain surface at @xmath90 .
the peak decays rapidly to zero .
the result indicates the occurrence of charge overcompensation ; that is , there is an excessive number of tetravalent counterions condensing on the chain surface than needed to neutralize it .
a charge overshooting thus takes place .
the higher the salt concentration , the more pronounced the peak will be .
moreover , we found that the rigid and the semiflexible chains are more overcharged than the flexible chain .
this is because the opened chain structure provides larger space to condense multivalent counterions , leading to a more pronounced @xmath84 peak . in the last two sections ,
the integrated ion and charge distributions were studied as a function of @xmath63 .
these figures contain detailed information at any given tube radius .
nonetheless , in a general discussion , people usually concern about the number of the condensed ions on a chain and pay more attention on the effective chain charge resulting from the ion condensation . to study these topics , one needs to know firstly the region of ion condensation .
however , there exists no clear boundary to separate the condensation region from the non - condensation one .
we remark that the condensation region is simply a concept , which can be only qualitatively described . in this study
, we define the condensation region to be the tube region around a chain with @xmath63 equal to the bjerrum length @xmath70 ( @xmath91 ) .
this definition is based upon the manning condensation theory @xcite .
the results for the three different chain stiffness are shown in fig .
[ f : tube_rc3 ] , panels ( a)(d ) , where the numbers of the condensed monovalent counterions ( @xmath92 ) , the condensed tetravalent counterions ( @xmath93 ) , and the condensed coions ( @xmath94 ) , together with the effective charge of chain inside the condensation region ( @xmath95 ) , are presented as a function of salt concentration .
( a ) , ( b ) @xmath93 , ( c ) @xmath94 , and ( d ) @xmath95 as a function of @xmath37 .
the values of chain stiffness @xmath49 are indicated in the figures.,title="fig:",scaledwidth=45.0% ] ( b ) , ( b ) @xmath93 , ( c ) @xmath94 , and ( d ) @xmath95 as a function of @xmath37 .
the values of chain stiffness @xmath49 are indicated in the figures.,title="fig:",scaledwidth=45.0% ] ( c ) , ( b ) @xmath93 , ( c ) @xmath94 , and ( d ) @xmath95 as a function of @xmath37 .
the values of chain stiffness @xmath49 are indicated in the figures.,title="fig:",scaledwidth=45.0% ] ( d ) , ( b ) @xmath93 , ( c ) @xmath94 , and ( d ) @xmath95 as a function of @xmath37 .
the values of chain stiffness @xmath49 are indicated in the figures.,title="fig:",scaledwidth=45.0% ] focus firstly on the lowest salt concentration @xmath65 in the figures . at this @xmath37 ,
2 tetravalent counterions and 8 monovalent coions are added into the simulation box ( refer to table [ tab : ion_no ] ) .
the results , @xmath96 and @xmath97 , show that the tetravalent counterions entirely condense on the chain whereas the coions stay all in the bulk solution . manning condensation theory
predicts that the condensation of counterions on a rigid chain in a salt - free solution should result in an effective chain charge density equal to @xmath98 ( in charge unit ) @xcite . in our simulations , @xmath99 monomers constitute a chain and the charge distance on a chain is @xmath100 ( see fig . [
f : blvcs ] ) .
there are hence @xmath101 , about 30 monovalent counterions neutralizing the chain at @xmath74 , according to the theory .
since 2 condensed tetravalent counterions could repel 8 monovalent counterions originally condensed on the chain , back into the bulk solution , we expect a reduction of the number of the condensed monovalent counterion , @xmath92 , to a number of about 22 .
this is exactly what we observed in fig .
[ f : tube_rc3](a ) for the rigid ( @xmath55 ) and the semiflexible ( @xmath54 ) chains .
please notice that @xmath92 for the flexible chain ( @xmath53 ) is larger than 22 .
it suggests a stronger ion condensation .
focus next on the low salt region @xmath102 .
we can see that @xmath93 increases with @xmath37 .
in fact , @xmath93 is equal to the number of tetravalent counterions added into the simulation box .
it shows a full condensation of these ions on the chain . in this salt region ,
the @xmath93 curves overlap with each other , regardless of the chain stiffness . for condensed coions
, we found @xmath97 .
it suggests that the chains repel the coions .
this fact can be examined in fig .
[ f : tube_rc3](d ) where the effective chain charge @xmath95 indeed takes a negative value ( when @xmath103 ) .
therefore , the coion condensation is unfavored because of the coulomb repulsion .
concerning the condensation of monovalent counterions on a chain , we found that the number decreases with increasing @xmath37 and hits the minimum value , zero , at @xmath104 . at the same moment , @xmath95 increases and crosses eventually the zero point .
focus thirdly on the high salt region @xmath57 .
we observed that @xmath93 continuously increases with @xmath56 but splits into different curves , depending on the chain stiffness .
these splitting curves are upper bounded by the curve coming from the rigid chain , while lower bounded by the one coming from the flexible chain . for the semiflexible chain
, @xmath93 follows at beginning the lower - bounded curve and then transfers gradually to the upper - bounded curve .
once reexpanded in the solutions ( @xmath105 ) , the chain looks similar to a rigid rod and the curve overlaps with the upper - bounded curve .
please notice that @xmath93 is larger than the number needed to neutralize the bare chain charge .
charge overcompensation hence takes place and the chain charge turns to be positive .
we note that in this salt region , the tetravalent counterions are not always condensed on the chain ; in other words , fraction of them stay in the bulk solution .
since the chain charge is effectively positive , the coions can be attracted toward the chain but not the monovalent counterions .
for this reason , @xmath94 starts to increase with @xmath37 ( see fig .
[ f : tube_rc3](c ) ) but @xmath92 stays at zero ( see fig .
[ f : tube_rc3](b ) ) .
we observed that @xmath94 behaves similarly to @xmath93 , which split into curves depending on the chain stiffness .
this behavior is directly associated to that of the @xmath93 because coion condensation necessitates the help of bridging by excessively condensed tetravalent counterions . for the effective chain charge
, we did observe that @xmath95 becomes positive when @xmath57 .
nevertheless , the @xmath95 curve does not trivially follow the @xmath93 curve .
it shows convex - down behavior at some salt concentration .
the decrease of @xmath95 tells us that the condensing rate for coions exceeds 4 times of that for tetravalent counterions beyond the salt concentration .
again , @xmath95 splits into different curves when @xmath57 with the upper bound being the case of the rigid chain and the lower bound being the case of the flexible chain .
@xmath95 for the semiflexible chain deviates from the upper bound in the mid - salt regions because drastic change of chain conformation occurs in this region , very distinguishable to the morphology of the rigid chain .
finally , a trend was observed when @xmath37 is very high : @xmath95 tends to zero , no matter of which stiffness the chain is .
the charged chains are effectively neutralized .
the results obtained in this section demonstrate again that it is the chain morphology deciding the ion condensation and the effective chain charge .
the chain stiffness plays a role , behind curtains , to determine the chain morphology at a given salt concentration . in the previous three sections
, we have investigated the ion distribution and condensation in the radial direction away from a chain axis . however , the pertinent information concerning the distribution in the perpendicular direction , _
i.e. _ , in the direction of the chain axis , is still missing . in order to fill up this gap of information ,
we investigate here how condensed ions distribute along a chain axis .
the distribution is calculated by the following way .
we firstly index the monomers sequentially , starting by 1 , from one end of the chain to the other .
we then assign each condensed ion ( inside the condensation region @xmath81 ) to a monomer , to which from the ion the distance is the shortest .
the average numbers of the condensed ions assigned to every monomer are thirdly calculated for each ion species .
the results give the ion distribution along the chain . fig .
[ f : ionpos_s0 ] shows the distribution of the condensed monovalent counterions in a salt free solution .
each curve denotes one chain stiffness . ,
@xmath106 and @xmath107 .
in order to show the importance of ion excluded volume , we plotted , in addition , the case of @xmath55 with the ion size being reduced to half of the original one , marked by @xmath108 .
, scaledwidth=70.0% ] we observed that the distributions show plateau structure in the central part of the chains , while exhibiting a rapid decrease when the monomers approach to the two chain ends .
it shows that the counterions are more likely to condense on the central part than on the two ends .
when condensing onto the central part , a counterion is attracted by the monomers from both sides of the chain , while it is attracted only by one side of the chain if the condensation occurs near the chain ends .
the coulomb attraction is therefore stronger for the former case , and as a consequence , the condensed counterions are more numerous in the chain center .
the results are in agreement with the works done by other groups @xcite .
moreover , we found that the value at the plateau is higher for the flexible chain than for the rigid chain , which shows that monovalent counterions condense more strongly when the chain is flexible .
we know that a flexible chain displays a zig - zagged structure when counterions condense onto it and becomes less extensive compared to a semiflexible chain and a rigid chain @xcite . consequently , the line charge density on a flexible chain is effectively higher , which can lead to a stronger ion condensation . to examine it , we calculated the average angle @xmath109 between two adjacent bonds for the three cases of chain stiffness , and the results are @xmath110 , @xmath111 , and @xmath112 for the flexible ( @xmath53 ) , the semiflexible ( @xmath54 ) , and the rigid ( @xmath55 ) chains , respectively
. the effective charge distance in the main direction of the zig - zagged chain axis thus can be estimated by @xmath113 . to adapt it for the chains with flexibility , we modify the manning condensation theory by replacing the charge distance on the chains with the effective charge distance .
the modified theory predicts a fraction of charge neutralization on the chains equal to @xmath114 , which yields 0.665 , 0.622 , and 0.617 condensed monovalent counterions per monomer for the three studied stiffness , respectively .
our simulations show consistent trend of behavior with the theory . nonetheless , the numbers are slightly smaller than the predictions .
this is because the manning theory was derived under the assumption of pointlike ions condensing onto an infinite cylindrical chain .
our modeling ions have excluded volume and the pe chain has finite chain length ; both of the effects can decrease the number of the condensed ions . to verify it ,
identical simulations have been performed for the rigid chain ( @xmath55 ) except that the counterions were set to be half of the original ion size . the result is also plotted in fig .
[ f : ionpos_s0 ] in dash - dotted curve ( denoted by @xmath108 ) for comparison .
we observed that the number of the condensed ions increases , which shows the importance of ion excluded volume on ion condensation . here , the distance of closest approach between the monomer and the counterion is also a relevant parameter .
the smaller the ion size , the closer the distance , and therefore , the stronger the condensation .
we next concentrate on the systems with added tetravalent salt .
since the tetravalent counterions play a crucial role in determination of chain morphology by substituting the condensed monovalent counterions on a chain , we study here how the condensed tetravalent counterions distribute on a chain . fig .
[ f : ionpos_4+](a ) shows such distribution on the rigid chain ( @xmath55 ) . ( a ) .
the value of @xmath115 can be read near the corresponding curve .
, title="fig : " ] ( b ) .
the value of @xmath115 can be read near the corresponding curve .
, title="fig : " ] ( c ) .
the value of @xmath115 can be read near the corresponding curve .
, title="fig : " ] at the low salt concentration @xmath116 , the ions distribute uniformly on the chain backbone except near the chain ends .
the distribution profile looks similarly to that for monovalent counterions in salt - free solution .
this is because many monovalent counterions still condense on the chain ( cf .
[ f : tube_rc3](a ) ) and collide with the condensed tetravalent counterions .
the condensed tetravalent ions thus diffuse on the chain , resulting in a uniform distribution . as @xmath37 increases
, the condensation of tetravalent counterions ensures a rapid decrease of the number of the condensed monovalent counterions .
eventually , the tetravalent ions become the majority on the chain and the strong electrostatic repulsion between them induces the localization of these ions at some specific positions along the one - dimensional chain .
this is why at @xmath117 and @xmath118 , several peaks appear on the distribution profile .
the number of the peaks corresponds to the number of the tetravalent ions condensed on the chain , which is exactly the amount of these ions added in the simulation box ( cf .
table [ tab : ion_no ] ) .
when @xmath37 is increased over the equivalence point @xmath119 , the number of the condensed tetravalent counterions exceeds the one needed to neutralize the chain backbone charge .
charge overcompensation takes place .
the ion localization is now disturbed by these extra ions .
the disturbing shows its effect starting from the central part of the chain and propagating to the chain ends with increasing the salt concentration .
finally , the distribution becomes uniform again .
we found that , different to the cases at low salt concentrations , the tetravalent ions favor to condense on the chain ends when @xmath37 is high .
this is because near the chain ends , there are more spaces available to accommodate these over - crowded ions .
[ f : ionpos_4+](b ) shows the distribution on the semiflexible chain ( @xmath54 ) .
when @xmath120 or @xmath121 , the pe chain displays an extended structure .
similar to the rigid chain , the condensed tetravalent ions distribute uniformly on the chain backbone . in between the salt concentrations , the chain forms ordered collapsed structure and the distribution displays some specific profile .
for example , at @xmath122 , the distribution shows a symmetric profile with two humps .
this profile results from a particular hairpin structure in which the length of the two branches of the hairpin is not equal ( cf . snapshots in fig .
[ f : snapshots ] ) .
each hump in the distribution reflects the drastic change of chain geometry at an eccentric bending point of the hairpin .
we note that the appearance of the two humps do not mean two bendings in one hairpin .
it is a result of ensemble average in which the probability of the bending point close to one of the two chain ends is equal when the system is in equilibrium .
therefore , the distribution is a symmetric function with respect to the middle point . at @xmath117 and @xmath58 ,
the number of the condensed ions is sufficiently numerous to bind the chain , against the chain stiffness , to form a hairpin with two equal branches ( cf .
[ f : snapshots ] ) .
a condensed counterion is localized at the bending point , which results in the peak at the middle point of the chain . for @xmath123 ,
the number of the condensed tetravalent ions is excess .
a loop is formed near the hairpin head in order to create more room to accommodate these extra ions .
the chain morphology thus looks similar to a tennis racquet .
the more the excess number of the ions , the larger the loop will be formed and the distribution shows a broader and higher hump near the middle of the chain .
we remark that when @xmath124 , toroid is also a favored structure of the chain .
the ion distribution on toroidal structure is similar to the one on a hairpin ( or a racquet ) , which displays a hump in the middle . to help readers to obtain the whole pictures of the system in mind , we present in fig . [
f : snapshots ] typical snapshots of the semiflexible chain together with the condensed tetravalent counterions in the middle range of salt concentration . ) .
the numbers below the pictures are the salt concentrations ( @xmath125 ) . ]
figure [ f : ionpos_4+](c ) shows the distribution on the flexible pe ( @xmath53 ) .
we have learned that the flexible chain is collapsed into a disordered structure when the salt concentration is intermediate .
the condensed tetravalent ions can not be localized at some specific positions of the chain .
the distribution hence shows a uniform distribution . by comparing fig .
[ f : ionpos_4+](a ) , ( b ) and ( c ) , we noticed that the number of the condensed tetravalent counterions is less numerous for the flexible chain than for the semiflexible and the rigid chains when @xmath57 .
the reason is that the flexible chain occupies a smaller volume space compared to the other two chain stiffness ( cf .
[ f : rhvcs ] ) .
consequently , smaller room is available to accommodate the strongly - repelled tetravalent ions , leading to fewer condensed tetravalent counterions .
it is worth noticing that in the condensation of the monovalent counterions at @xmath60 ( cf .
[ f : ionpos_s0 ] ) , the number of condensed monovalent counterions is more numerous for the flexible chain , which shows an opposite trend to this case .
in addition to the ion distributions in the radial and in the axial directions discussed in the above sections , we study also the radial distribution function @xmath126 between monomers and tetravalent counterions .
@xmath126 is defined as the averaged density function @xmath127 of the tetravalent counterions around a monomer dividing the mean density of the tetravalent counterions in the whole system , @xmath37 .
it is a correlation function , which provides the information of organization of the tetravalent counterions surrounding an individual monomer .
the results at different @xmath37 are presented in fig .
[ f : rdf_m4 + ] for the three chain stiffness , @xmath53 , @xmath106 , and @xmath107 . ( a ) for ( a ) the rigid chain , ( b ) the semiflexible chain , and ( c ) the flexible chain at different @xmath37 .
the value of @xmath115 is indicated in the figures.,title="fig : " ] ( b ) for ( a ) the rigid chain , ( b ) the semiflexible chain , and ( c ) the flexible chain at different @xmath37 .
the value of @xmath115 is indicated in the figures.,title="fig : " ] ( c ) for ( a ) the rigid chain , ( b ) the semiflexible chain , and ( c ) the flexible chain at different @xmath37 .
the value of @xmath115 is indicated in the figures.,title="fig : " ] fig . [
f : rdf_m4+](a ) shows @xmath126 for the rigid chain ( @xmath55 ) .
the distribution functions show a series of peaks located subsequently at @xmath128 , @xmath129 , @xmath130 , @xmath131 , @xmath132 , .... since the chain conformation is a straight line , the results indicate that the condensed tetravalent counterions sit on the clefts between two adjacent monomers and form triplets with monomers , as sketched in the cartoon picture ( inside the figure ) , because the peak positions match well the distances calculated from the picture .
similar triplet formation has been reported by other group using monovalent counterions , instead @xcite .
when @xmath133 , the distribution functions overlap with each other .
this is because nearly all of the tetravalent counterions condense on the chain in this salt region . as a result ,
the correlations between monomers and tetravalent counterions are identical . when @xmath134 , the curves move downward with increasing @xmath37 .
the correlation decreases , owing to the increase of the non - condensed tetravalent counterions presented in the bulk solutions .
@xmath126 for the semiflexible chain ( @xmath54 ) displays subsequent peaks too , as shown in fig .
[ f : rdf_m4+](b ) . however , these peaks are broader and less pronounced , in comparison with the previous case , because thermal fluctuations can take effect on chain conformation for this chain stiffness .
since the chain collapses in the intermediate salt region , the size of the chain reduces , leading to a faster decay of @xmath126 against @xmath6 .
@xmath126 for the flexible chain ( @xmath53 ) shows only one peak at @xmath135 ( see fig .
[ f : rdf_m4+](c ) ) .
this peak marks the close contact of the counterions with a monomer .
it is more pronounced than the first peak for the semiflexible and the rigid chains , suggesting a stronger correlation . following the peak ,
a broad shoulder appears on the right - hand side .
the shoulder narrows down to a finite region when the salt concentration is increased from @xmath136 to @xmath58 .
this behavior reflects two facts : ( 1 ) the nature of random arrangement of tetravalent counterions surrounding a monomer when the chain is flexible , ( 2 ) the reduction of chain size in the intermediate salt region .
using langevin dynamics simulations , we have studied ion and charge distributions in pe solutions from various points of view . the discussed topics focused on the effects of the two parameters : the salt concentration and the chain stiffness .
the salt concentration ran over a wide range of value in which the reentrant condensation takes place , and the chain stiffness covers the three types of behavior discussed in section [ subsec_chain_size ] .
the results showed that the ion and the charge distributions depend strongly on the morphology of a chain , while the morphology is determined by both of the two parameters . when the salt concentration is low and can not induce collapse of chain , the ion distribution shows similar profile for flexible , semiflexible and rigid chains , because these chains all display extended structures . in the middle salt region , the chain charge is nearly neutralized by the condensed multivalent counterions .
the chains show drastic conformational change .
the flexible chains are collapsed into disordered globule structures whereas the semiflexible chains are collapsed into ordered structures such as toroid and hairpin . for the rigid chains , the chain conformation remains rodlike .
since the chain morphologies are different for these three kinds of chain stiffness , the ion distributions show significant difference .
a general rule is that the larger the space that a chain structure occupies , the more the ions can condense on a chain .
therefore , the number of the condensed ions and the effective chain charge for the semiflexible chains are upper- and lower - bounded by those curves for the flexible and the rigid chains , respectively . in the high - salt region ,
the condensed multivalent counterions overcompensate the chain charge on the chain surface .
the flexible and the semiflexible chains reexpand in the solutions .
the latter chain displays a rodlike structure , more extended than the former one which displays a coil structure .
therefore , the semiflexible and the rigid chains show similar ion distributions , distinguishable to the flexible chain .
the study of the condensed ion distribution in the chain axial direction tells us other information . at low or at high salt concentrations ,
the condensed ions distribute uniformly on the chain backbone ( because the chain is not collapsed ) , whereas at intermediate salt concentrations , the distribution showed specific profile which has direct connection with the structure of the collapsed chains .
finally , in the study of the radial distribution function , we demonstrated that a condensed multivalent counterion favors to form a triplet with two adjacent monomers .
consequently , the radial distribution of multivalent counterions shows a series of peaks away from a monomer . before the closure of this paper
, we describe more about how flexible chains and semiflexible chains get reexpanded in a high - salt region . according to fig .
[ f : tube_rc3 ] , the number of the condensed multivalent counterions increases monotonically when the salt concentration is increased .
we know that flexible chains and semiflexible chains are both collapsed and charge - neutralized at the equivalence point @xmath56 .
further condensing the multivalent counterions leads to the arising of the repulsive forces , including the coulomb repulsion and the excluded volume interaction , in the interior of the collapsed chain structures .
the collapsed chains are therefore getting looser and becomes more and more unstable . for a flexible chain ,
the chain unfolds gradually into a coil structure .
similar results have been reported recently in the simulation study of other flexible charged systems , such as dendrimers @xcite and spherical polyelectrolyte brushes @xcite . in these studies ,
the dendrimers and the pe brushes were collapsed by multivalent salt and oppositely charged linear pes , respectively . when the concentration of these condensing agents is higher than the equivalence point , the collapsed structures reswell due to the excluded volume and the repulsive interactions of the condensing agents excessively adsorbed on the systems .
the reswelling goes in a continuous way @xcite . for a semiflexible chain , there exists , in addition , a restoring force , owing to the intrinsic chain stiffness , which plays a crucial role in the chain reexpansion .
the binding of chain segments in a collapsed chain structure decreases when the excess number of the condensed counterions increases .
once the restoring force can overcome the binding force , the chain reexpands in a sudden and drastic way .
a first - order transition takes place @xcite .
the detailed information provided in this study gives valuable insight to the problem of dna condensation and can help researchers to develop new models to explain these phenomena more quantitatively .
this material is based upon work supported by the national science council , the republic of china , under contract no .
nsc 95 - 2112-m-007 - 025-my2 and nsc 97 - 2112-m-007 - 007-my3 .
computing resources are supported by the national center for high - performance computing under the project `` taiwan knowledge innovation national grid '' . | ion distributions in dilute polyelectrolyte solutions are studied by means of langevin dynamics simulations .
we show that the distributions depend on the conformation of a chain while the conformation is determined by the chain stiffness and the salt concentration .
we observe that the monovalent counterions originally condensed on a chain can be replaced by the multivalent ones dissociated from the added salt due to strong electrostatic interaction .
these newly condensed ions give an important impact on the chain structure . at low and at high salt concentrations , the conformation of a semiflexible chain is rodlike .
the ion distributions show similarity to those for a rigid chain , but difference to those for a flexible chain whose conformation is a coil . in the mid - salt region ,
the flexible chain and the semiflexible chain collapse but the collapsed chain structures are , respectively , disordered and ordered structures . the ion distributions hence show different profiles for these three chain stiffness with the curves for the semiflexible chain lying between those for the flexible and the rigid chains .
the number of the condensed multivalent counterions , as well as the effective chain charge , also shows similar behavior , demonstrating a direct connection with the chain morphology .
moreover , we find that the condensed multivalent counterions form triplets with two adjacent monomers and are localized on the chain axis at intermediate salt concentration when the chain stiffness is semiflexible or rigid .
the microscopic information obtained here provides valuable insight to the phenomena of dna condensation and is very useful for researchers to develop new models . |
present experimental information available in the field of heavy - ion collisions allows for the systematic investigation of strongly interacting hadronic matter under extreme conditions .
high statistics accumulated in various experiments facilitates reconstruction of hadronic resonances via their decay products . at cern - sps experiments
several resonances have been identified so far : @xmath2 mesons are detected by na49 collaboration via @xmath3 decay @xcite and by na50 via @xmath4 channel @xcite , @xmath5 and @xmath1 are seen by na49 in their dominant decay modes @xcite , the @xmath6 and @xmath7 mesons are measured by na50 in dimuon channel @xcite .
already the first preliminary data were surprising .
total averaged @xmath8 meson multiplicity extracted from na49 data was found to be significantly smaller than that obtained from the extrapolation of preliminary na50 data .
another interesting observation is that preliminary na49 data on @xmath5 multiplicity reveal significant suppression of the yield per participating nucleon when compared to inelastic proton - proton collisions .
this contradicts to systematics of other strange particle abundance at sps energies . suggested explanation of the @xmath8-meson puzzle @xcite was that @xmath8-mesons decaying inside a fireball can disappear from the @xmath9 mass peak due to rescattering and absorption of secondary kaons in the surrounding medium .
such mechanism has been quantitatively studied in @xcite where suppression at the level 4060% has been obtained from simulation using rqmd code .
similar mechanism could also account for @xmath5 suppression . in @xcite 75%
suppression for @xmath2 and 50% for @xmath5 particles have been obtained within urqmd model .
therefore the rescattering of resonance decay products can indeed explain in part some experimental observations .
however , besides the lack of quantitative account of the difference in the extrapolated @xmath8 meson yields , other caveat has been issued in @xcite : since the vacuum width of @xmath1 mesons is more than 3 times larger than the @xmath5 width , @xmath1 mesons decay inside a fireball more frequently and should be therefore suppressed even stronger than @xmath5 .
the preliminary data of na49 @xcite do not show such a suppression for @xmath10 . in this paper
we discuss to what extent the in - medium modifications of the resonance decay width can effect the resonance suppression and help to accommodate empirical information .
in our study we consider a two - stage picture of the hadronic phase of heavy - ion collision at sps energies which can be characterized by the notions of chemical and thermal freeze - outs @xcite . at the initial stage of the hadronic phase we assume the temperature @xmath11 to be close to the qcd phase transition temperature @xmath12 mev .
then the system expands up to the point , when numbers of different kinds of particles freeze in a chemical freeze - out .
thermodynamical parameters of this stage could be obtained by fitting the final total hadron multiplicities .
typical temperature is found to be @xmath13 mev @xcite . during the second stage of collision ,
elastic scattering processes change momentum distributions of hadrons in accord with a decreasing temperature until they cease and distributions freeze in .
the freeze - out temperature can be extracted from a simultaneous fit to the single - particle @xmath14-spectra of different particles supplemented by the analysis of particle correlation data .
typical temperatures are found to be @xmath15 mev @xcite .
the key assumption behind the decay product rescattering ( dpr ) mechanism of @xcite is that the final resonance distribution is formed at some stage between the chemical and thermal freeze - out . at this moment , mean free paths of the resonance , @xmath16 , pions , kaons and nucleons , @xmath17 , and the characteristic size of the fireball , @xmath18 , should satisfy [ mfp ] _ , _ k,_nr _ r. this implies that upon this stage resonances stream freely out from the fireball . for their decay products , pions and kaons , on the other hand , the surrounding medium is still opaque .
therefore , if a resonance decays inside a fireball in hadronic channel , the decay products can be rescattered or absorbed , so that the resonance remains unobserved experimentally .
it is now obvious that the possible increase of a resonance decay width in medium will enlarge the probability of its decay inside a fireball enhancing thereby the dpr mechanism . besides the total decay width of the resonance and the typical size of the fireball ,
one of the crucial parameters controlling the efficiency of the dpr mechanism is , the life time of the fireball after the resonance freeze - out , @xmath19 .
we denote the phase - space distribution of the resonance @xmath20 in the center - of - mass system of two colliding nuclei at the moment of its thermal freeze - out as @xmath21 .
then the primary momentum distribution of the resonance is [ integr ] _ 0^(r)(p ) = 1, = _
f^(r ) ( , ) ( ) . where integration goes over the fireball volume within a freeze - out hyper - surface @xmath22 . in the absence of any in - medium modification of the resonance and rescattering of its decay products , the shape of the momentum distribution _ observed _ in the decay channel @xmath23 is given by @xmath24 , where @xmath25 and @xmath26 are the partial and total decay widths of the resonance , respectively .
accordingly , in the experimental analysis , the resonance distribution would be reconstructed from momentum distribution of decay products multiplied by the corresponding inverse branching ratio .
taking into account modifications of partial and total widths of the resonance and the rescattering of decay products in medium we write expression for the observed resonance distributions in the decay channel @xmath23 in the following form _
j^(r)(p ) & = & ( ) + _ 0^t
j^(r ) * ( t ) p^(r)_j ( t ) p_rec .
[ ratrescat ] here [ pbx ] d(t)&= & , is the probability that resonance will fly for a time @xmath27 starting from a position @xmath28 where it suffers the last interaction .
@xmath29 is the total width of a moving particles with energy @xmath30 , where @xmath31 is a resonance mass . here and below asterisk
denotes in - medium values of the quantities , which is determined by the current local temperature and density .
the probability of the decay products to leave fireball without any rescattering is @xmath32 . for the explicit form of @xmath32
we refer to ref .
@xcite , where it is written for the case of the two - particle decay . the quantity @xmath33 stands for the probability to identify a resonance from non - rescattered decay product .
without in - medium effects we put @xmath34 assuming an ideal detector .
if spectra of resonance decay products differ in medium from the vacuum ones , the momenta of secondary particles will change on the way out from the fireball even without being rescattered , since in this case a fireball serves as a potential well .
this effect can be especially strong for strange resonances since the properties of kaons in the final state are strongly modified in nuclear matter or / and isospin asymmetrical pion gas @xcite . for a rough estimation of the maximal suppression effect we put in this case @xmath35 .
the time scale @xmath36 in ( [ ratrescat ] ) is the time spent by the resonance in medium .
it is given by @xmath37 , where @xmath38 is the time during which the resonance with a velocity @xmath39 flies from the position @xmath28 till the border of the fireball . finally , in ( [ ratrescat ] ) we integrate over all initial resonance positions @xmath40 with the phase space distribution ( [ integr ] ) .
in the case when the partial width of decay channel @xmath23 does not change in medium @xmath41 , decay products are not rescattered @xmath42 and the total width @xmath43 , we have [ ratmu ] _ j^(r)(p ) & = & ( ) + _ tot^(r ) 0 _ 0^ dt d(t ) .
this expression corresponds , e.g. , to the case when resonances are reconstructed in the leptonic channels . note that for @xmath44 we have @xmath45 , for @xmath46 we have @xmath47 and if @xmath48 then @xmath49 . after the general considerations we specify the model of the fireball expansion , which is used in our numerical calculations .
we assume a homogeneous spherical fireball with the constant density and temperature profiles . the primordial resonance momentum distribution [ dist ] f_r(,)=,is determined by the temperature @xmath50 , flow velocity profile @xmath51 and radius @xmath52 .
the fireball expands @xmath53 , its density evolves as @xmath54 and the temperature decreases as @xmath55 as expected for relativistic pion gas , cf .
time of flight of a resonance inside a fireball starting from a position @xmath28 till the border is given by _
r ( , ) & = & ( + [ tau ] & -&(-v_fr_0 ) ) ( ^2-v_f^2)^-1 , which is valid for @xmath56 and @xmath57 . in the case
@xmath58 we put @xmath59 .
parameters @xmath52 , @xmath50 , @xmath60 , serve as input for numerical evaluations below .
the particular realization of ( [ ratrescat ] ) is a rather crude approximation .
however , the final results are found @xcite to be rather insensitive to the details of hydrodynamical evolution of a fireball , being determined mainly by the values of @xmath61 , @xmath62 , and @xmath60 .
first we discuss numerical results obtained in @xcite for @xmath63 meson yields reconstructed via @xmath9 and @xmath64 channels in central pb+pb collisions at 158 gev / n sps energy .
being dominantly an @xmath65 state , the @xmath8 mesons can decouple quite early , since its interaction with non - strange matter is suppressed according to ozi rule .
the mean free path @xmath66 of @xmath8 mesons in hadron gas is estimated in @xcite .
comparison with mean free paths of pions and kaons , @xmath67 , from @xcite gives @xmath68 for temperatures @xmath69 .
therefore condition ( [ mfp ] ) can be satisfied . for a comparison with experimental data
we define the following suppression factors [ etamt](m_t)=_k(m_t)/_(m_t ) , _
k,(m_t)=<^()_k,(p)>_y/<^()_0(p ) > _ y , where @xmath70 means the integration over experimental rapidity interval . the apparent distribution of @xmath8 mesons in the hadronic channel ( @xmath71 ) , @xmath72 ,
is calculated according to ( [ ratrescat ] ) . in the leptonic channel ,
@xmath73 we use ( [ ratmu ] ) .
first , we evaluate suppression factor ( [ etamt ] ) without any modification of @xmath8 properties in medium . in this case
we have @xmath74 and @xmath75 .
we use several combinations of input parameters @xmath76 chosen consistently with the @xmath8-meson freeze - out temperature @xmath50 which we vary between @xmath77 and @xmath78 : ( @xmath79 ) ( 150 mev , 20 fm , 0.5 ) , ( @xmath80 ) ( 160 mev , 15 fm , 0.46 ) , ( @xmath81 ) ( 170 mev , 10 fm , 0.41 ) .
size of the fireball @xmath52 at the moment of @xmath8 freeze - out has to be comparable with @xmath8 meson mean free path @xmath82 .
the above values for different temperatures are taken according to estimations of @xmath66 in @xcite .
flow velocities are adjusted to reproduce the slope of @xmath63 meson @xmath14 distribution measured by the na50 collaboration : @xmath83 mev .
the fireball life time is determined by equation @xmath84 calculated for cases @xmath85@xmath86 without medium effects .
solid lines are calculated with @xmath87 mev and dash lines correspond to @xmath88 mev .
+ _ right plane _ : thick lines show the ratio @xmath89 calculated with account for in - medium modification of meson properties associated with a decreasing kaon mass @xmath90 mev results for different parameters sets @xmath85@xmath86 are depicted by solid , dash and dotted lines , respectively .
thin lines illustrate the suppression factors in the muon @xmath91 and kaon @xmath92 channels calculated for parameter set ( @xmath79).,title="fig:",width=226 ] calculated for cases @xmath85@xmath86 without medium effects .
solid lines are calculated with @xmath87 mev and dash lines correspond to @xmath88 mev .
+ _ right plane _ : thick lines show the ratio @xmath89 calculated with account for in - medium modification of meson properties associated with a decreasing kaon mass @xmath90 mev results for different parameters sets @xmath85@xmath86 are depicted by solid , dash and dotted lines , respectively .
thin lines illustrate the suppression factors in the muon @xmath91 and kaon @xmath92 channels calculated for parameter set ( @xmath79).,title="fig:",width=207 ] to reproduce results of rqmd calculations described in ref .
@xcite we take freeze - out temperature @xmath87 mev , the kaon mean free path @xmath93 with @xmath94 fm and @xmath95 .
the temperature corresponds to the lowest limit allowed by the analysis @xcite .
the results are shown in fig .
[ fig : phi ] ( left panel ) by solid lines .
the limiting scenario considered in @xcite , when the freeze - out volume is determined by the last kaon interactions , can be reproduced with @xmath88 mev .
this case is shown by dash lines .
the solid lines in the left panel of fig .
[ fig : phi ] we take as a reference point for our further investigation of in - medium effects .
let us now consider the case when the @xmath8 meson width increases strongly in hadronic medium .
we simulate this effect by decreasing the kaon mass , which can result e.g. from rescattering of kaons on pions through @xmath96 and heavier kaonic resonances @xcite .
the modification of kaon properties in medium prevent the @xmath8 meson reconstruction even when kaons can leave a fireball without a hard rescattering . leaving the fireball , kaons have to come back to their vacuum mass shell . due to the energy conservation
the momenta of kaons change and , thereby , the invariant mass and momentum of the pair . for this effect we account by putting @xmath97 . then the both suppression factors @xmath92 and @xmath91 depend only on the total @xmath8 meson width .
the @xmath8 total width reads as @xmath98 with @xmath99 .
the vacuum widths of @xmath100 and @xmath101 decay processes are equal to @xmath102 mev and @xmath103 mev .
we do not consider here in - medium change of the @xmath104 width , which effects weakly the resulting suppression factor @xcite .
we restrict ourselves to rather conservative modification of kaon masses .
initially , at the @xmath8 freeze - out moment we put @xmath90 mev , and then it decreases linearly with the decreasing fireball density @xmath105 .
it corresponds to the initial @xmath8 width @xmath106 mev . in this case
@xmath107 and at small @xmath108 region @xmath109 can be suppressed up to 4060% .
thus , for a given freeze - out temperature @xmath50 we have to readjust the flow velocity @xmath110 to reproduce the slope of the @xmath14 distribution measured in dimuon channel by na50 @xcite . for our three sets of parameters specified above
we obtain new flow velocities : ( @xmath79 ) @xmath111 , ( @xmath80 ) @xmath112 , ( @xmath81 ) @xmath113 .
the results for a relative suppression of hadronic and leptonic channels are presented in the right plane of fig .
[ fig : phi ] .
thick solid , dashed and dotted lines drawn for cases ( @xmath79)-(@xmath81 ) should be compared with the solid lines in the left plane .
we observe that the increase of the @xmath8 width results in the overall increase of the suppression effect by about 20% .
the suppression factors for leptonic and hadronic channels , separately , are shown in fig .
[ fig : phi ] as well .
the increase of the @xmath8-meson width provides a strong suppression of @xmath114 .
however , since the dimoun channel is also suppressed the resulting ratio @xmath115 remains on the level @xmath116 for small @xmath108 .
let us now consider closer dependence of the suppression factor on the fireball parameters @xmath19 and @xmath52 .
first , we simplify the time dependence of the total width .
we use a linear in density interpolation between the initial value at the freeze - out moment , @xmath117 and the vacuum value [ gam ] _
tot^(r ) * ( t)= _ tot^(r)0 + ( _ in^(r)-_tot^(r ) 0)(r_0/r(t))^3 .
this approximation works still well and can reproduce the right panel of fig .
[ fig : phi ] , using @xmath118 mev within 510% accuracy . we will vary parameters @xmath52 and @xmath19 keeping other parameter fixed : @xmath119 mev , @xmath120 , and @xmath35 . in fig .
[ fig : map ] we present the contour plot of the value @xmath121 as a function the dimensionless variables @xmath122 and @xmath123 , calculated for @xmath124 , 20 , 30 mev .
the thin solid line shows the level @xmath125 for the vacuum width @xmath126 mev .
we see that for the vacuum width the 50% suppression can be reached only for rather large @xmath127 fm . from this plot
we can also read out the minimal combinations of parameters @xmath128 required to attain a certain degree of suppression .
for instance , @xmath129 will be reached ( we pick out the points closest to the origin on the lines labeled with 0.2 ) for @xmath124 mev if ( @xmath130 fm , @xmath131 fm ) , for @xmath132 mev if ( @xmath133 fm , @xmath131 fm ) , and for @xmath134 mev if ( @xmath135 fm , @xmath136 fm ) . have we fixed the size of a fireball to be about 20 fm , the level @xmath137 can be arrived at @xmath13863 , 39 , 26 fm for @xmath13910 , 20 and 30 mev , respectively . in ( [ res ] ) for @xmath8 , @xmath5 and @xmath1 particles on the plane @xmath140@xmath141 .
different line styles correspond to the different initial resonance widths @xmath142 in ( [ gam ] ) .
thick solid , dashed and dotted lines are drawn , respectively , for @xmath124 , 20 , 30 mev , @xmath143 , 60 , 120 mev , and @xmath144 , 70 , 80 mev .
the thin solid lines are calculated with the vacuum width of the resonance and stand for the level @xmath145 .
all calculations are done for @xmath146 mev , @xmath120 , and @xmath35.,title="fig:",width=200 ] in ( [ res ] ) for @xmath8 , @xmath5 and @xmath1 particles on the plane @xmath140@xmath141 .
different line styles correspond to the different initial resonance widths @xmath142 in ( [ gam ] ) .
thick solid , dashed and dotted lines are drawn , respectively , for @xmath124 , 20 , 30 mev , @xmath143 , 60 , 120 mev , and @xmath144 , 70 , 80 mev .
the thin solid lines are calculated with the vacuum width of the resonance and stand for the level @xmath145 .
all calculations are done for @xmath146 mev , @xmath120 , and @xmath35.,title="fig:",width=200 ] in ( [ res ] ) for @xmath8 , @xmath5 and @xmath1 particles on the plane @xmath140@xmath141 .
different line styles correspond to the different initial resonance widths @xmath142 in ( [ gam ] ) .
thick solid , dashed and dotted lines are drawn , respectively , for @xmath124 , 20 , 30 mev , @xmath143 , 60 , 120 mev , and @xmath144 , 70 , 80 mev .
the thin solid lines are calculated with the vacuum width of the resonance and stand for the level @xmath145 .
all calculations are done for @xmath146 mev , @xmath120 , and @xmath35.,title="fig:",width=200 ] we turn now to the @xmath5 and @xmath1 resonances .
their mean free paths in the hot hadronic matter have not been studied so far , unfortunately .
there is also no additional mechanism like ozi for @xmath8 mesons which would suppress the @xmath147 and @xmath96 interaction with surrounding pions , kaons , and nucleons .
@xmath148 scattering has contributions from s - channel hyperon exchange processes , particularly the reaction @xmath149 should be operative due to the s - wave @xmath150 coupling . therefore we can expect that the mean free paths of @xmath147 and @xmath96 are comparable with those of pions and kaons . hence their final momentum distribution should be formed rather close to the common break - up of the fireball . in fig .
[ fig : map ] we depict the contour plots of the quantity [ res ] ^(r)(m_t)=<_hadr.^(r)(p)>_y/<^(r)_0(p)>_y evaluated for @xmath151 , @xmath146 mev , @xmath120 .
the apparent distribution in the dominant hadronic channel @xmath152 is calculated according to ( [ ratrescat ] ) with @xmath153 and depends only on the total width of the resonance .
the time dependence of the width is given by ( [ gam ] ) with @xmath154 varying in the broad range separately for each resonance .
the last panel in fig .
[ fig : map ] shows us that to have for @xmath96 mesons the suppression factor @xmath155 , we should have @xmath1560.61 .
this translates for @xmath1575080 mev into @xmath1582.4 fm .
hence the @xmath96 mesons should indeed be produced close to the fireball break up . using these values for @xmath147 we find from the middle plane in fig .
[ fig : map ] that the ratio @xmath159 becomes less than 0.5 only if @xmath160 mev .
this is a very large value .
although there are indications that @xmath5 properties suffer a strong modifications in nuclear medium acquiring the total width of the order 100 mev at the normal nuclear matter density @xcite , calculations have been done so far for cold nuclear matter only . at high temperatures these effects might be reduced .
alternatively we can suggest that @xmath5 hyperons decouple from the fireball at some earlier stage than @xmath96 mesons .
then we obtain the following estimates for the time between @xmath147 freeze - out and the fireball breakup : @xmath161 fm for @xmath16230 mev and @xmath163 fm for the vacuum width .
finally we do not exclude a possibility that the formation of @xmath164 baryonic state @xmath5 in heavy - ion collisions is already primarily suppressed .
the purpose of the presented work is to find out to what extent the dpr suppression mechanisms enhanced by the in - medium modification of resonance width is able to accommodate experimental data .
we investigate dependence of the decay - product rescattering mechanism of the resonance production suppression on the size of the hadronic fireball and on the time between resonance freeze - out and fireball break - up .
possible modification of the resonance width in medium is shown to enhance the suppression .
the model is applied to the production of @xmath2 , @xmath5 and @xmath1 particles at sps energies .
we conclude that in the case of @xmath2 mesons the discrepancy between @xmath63 momentum distributions observed in @xmath9 ( na49 ) and leptonic ( na50 ) channels can be explained if at the moment of the @xmath8 freeze - out the @xmath8 total decay width is about 30 mev and after freeze - out the fireball lives about 20 fm / c till the break - up .
the observed ( na49 ) signals of @xmath1 meson production indicate that @xmath96 mesons escape from the fireball at the moment very close to the breakup ( not further than 2.0 - 2.4 fm / c ) . otherwise their apparent distribution would be strongly suppressed .
we can obtain the attenuation of the @xmath5 on the level of 50% assuming either a very large width of the resonance @xmath165 mev at the freeze - out moment close to fireball break - up or that @xmath147 particles leave the fireball at least 69 fm / c before breakup .
we conclude that the resonance production in heavy ion collisions can serve as an indicator of the fireball dynamics at the last stage of heavy - ion collision .
to draw precise quantitative conclusions careful investigation of the modification of resonance properties in hadronic medium is necessary .
99 afanasiev s v 2000 , 59 .
willis n for na50 1999 , 534c ; markert c 2000 phd thesis ; contribution to this proceedings .
sammer t 2001 phd thesis ; na49 note # 260 .
friese v for na49 2001 contribution to the conference quark matter 2001 , stony brook , january 2001 quintas c for na50 2001 contribution to this proceedings .
shuryak e v 1999 , 119 .
johnson s c , jacak b v and drees a 2001 , 645 .
soff s 2001 , 421 torrieri g and rafelski j 2001 , 239 .
braun - munzinger p , heppe i and stachel j 1999 , 15 ; cleymans j and redlich k 1998 , 5284 ; + cleymans j , oeschler h and redlich k 1999 , 1663 .
tomik b , wiedemann u a and heinz u 1999 e - print : nucl - th/9907096 .
filip p and kolomeitsev e e 2001 , 054905 .
voskresensky d n 1994 _ jetp _ * 78 * , 793 .
brown g e , ko c m and kubodera k 1992 , 301 ; kolomeitsev e e and kmpfer b , voskresensky d n 1996 _ int.j .
_ * e5 * , 313 ; kolomeitsev e e and voskresensky d n 1995 _ phys .
nucl . _ * 58 * , 2082 ; ko c m 2001 , 327 .
lutz m f m and korpa c l 2001 eprint : nucl - th/0105067 .
lutz m f m , contribution to this proceedings .
haglin k 1995 , 719 .
haglin k l and pratt s 1994 , 255 .
weldon h a 1992 , 431 .
shuryak e v and thorsson v 1992 , 739 . | production of @xmath0 and @xmath1 resonances at the final stage of a heavy - ion collision is considered .
it is shown that original momentum distributions and abundance of resonances formed during the process of heavy ion collision may differ significantly from their measured spectra and yields .
reconstruction probability of resonances decaying inside the fireball can be strongly suppressed because of interactions of their hadronic decay products in the fireball medium .
we investigate dependence of the degree of the suppression on the fireball size , dynamics , and the resonance decay width in medium .
quantitative results are presented for lead - lead collisions at 158 gev sps beam energy . |
temporal meson correlation functions in euclidean time can be related to meson spectral functions by analytic continuation . furthermore using the maximum entropy method one can extract the spectral functions from meson correlators calculated on a finite lattice @xcite .
the method was applied at zero @xcite as well as non - zero temperature @xcite . in such an analysis one
usually uses all time slices in the correlation function not just the long distance part .
therefore one is sensitive to short distance physics which may be affected by lattice artifacts .
the lattice distortion of the correlators at short distances translates into lattice artifacts in the spectral functions for large energies .
so far only wilson gauge action and wilson ( clover ) fermion action have been used to study meson spectral functions .
the lattice artifacts seen in numerical simulations for this combination of actions turns out to be quite different from what is expected in the free theory @xcite .
the purpose of this paper is to study cutoff effects of meson spectral functions using different gauge and fermion actions .
we consider the temporal correlators of meson currents @xmath1 with @xmath2 and @xmath3 for pseudo - scalar and vector channels . for temperature
@xmath4 the spectral function is related to the meson correlator by the integral relation @xmath5 gev .
the vertical line and band show the mass of the ground and first excited states .
, title="fig:",width=230 ] gev .
the vertical line and band show the mass of the ground and first excited states .
, title="fig:",width=230 ] although this relation is valid _ a priori _ only in the continuum , it has been shown recently that in the limit of a non - interacting theory it holds on the lattice @xcite as well .
we use quenched domain wall fermon correlators on dbw2 gauge configurations from the rbc collaboration .
furthermore , we calculate tree - level clover fermion correlators on quenched 1-loop tadpole - improved symanzik lattices and on partially - quenched 2 + 1 flavor asqtad fermion lattices from the milc collaboration . to extract the spectral function we use the maximum entropy method .
we also parametrize the spectral functions as a sum of delta functions @xmath6 and determine the parameters @xmath7 and @xmath8 using constrained curve fitting techniques@xcite which allows for better systematic control of the results .
we start the discussion of our numerical results with the case of domain wall fermions .
here we performed calculations at three different lattice spacings , @xmath9 gev and valance quark mass around the strange quark mass .
for the coarsest lattice we see a two peak structure in the spectral functions with the second peak located around 2gev .
the positions of the peaks are reasonably reproduced using a two exponential fit . in fig .
[ dwf2.0_fig ] we show our results for finer lattices , @xmath10 gev for the pseudo - scalar and vector spectral functions .
we see a three peak structure , with the first excited state being close to the position of the physical excited states of the @xmath11 and @xmath12 mesons .
the third peak is likely to be a lattice artifact .
we also performed a simple two exponential fit of the corresponding meson correlators .
the results of the fit are shown as the vertical line and as a band .
the ground state mass from the fit ( statisitcal error @xmath13% not shown ) is in reasonable agreement with the mem analysis while for first excited state there is larger statistical uncertainty indicated by the band . to see whether or not the second and third peak correspond to physical resonances or lattice artifacts , we calculated the spectral functions at @xmath14 gev .
the results are shown in fig .
[ dwf3.0_fig ] which shows that the ground state peak did not shift ( apart from a small shift due to slightly larger quark mass ) while the second and the third peaks have moved roughly as @xmath15 .
this indicates that these peaks are lattice artifacts .
similar results were found for the wilson action @xcite . in ref .
@xcite they were interpreted as bound states of wilson doublers . for dwf , however , such an interpretation is more difficuilt to apply . for this lattice spacing
we have also performed a constrained fit with three exponentials .
the ground state mass shown as the vertical line in fig .
[ dwf3.0_fig ] was obtained from a simple uncostrained fit which was then used as a prior for the constrained fit .
the result for the second and third peak from the constrained fit is shown as bands in fig .
[ dwf3.0_fig ] .
reasonable agreement between the mem and constrained curve fits is found .
the vertical bands indicate the error on the peak location as determined from a constrained fit , title="fig:",width=230 ] gev .
the vertical bands indicate the error on the peak location as determined from a constrained fit , title="fig:",width=230 ] one may wonder whether the second and third peaks are artifacts of quenching .
absent dynamical quarks , all hadrons are absolutely stable ( i.e. have zero width ) no matter how large their mass is .
if dynamical quarks are present , hadrons may decay , and the higher the mass the larger the width .
therefore higher lying resonances merge into a continuum . to check this we performed calculations in quenched and partially - quenched qcd using the clover action .
more precisely we used gauge configurations with symanzik action generated by the milc collaboration as well the dynamical 2 + 1 asqtad configurations . in both cases
the lattice spacing was @xmath16 .
we use the tree - level value for the clover coefficient , @xmath17 , and @xmath18 which corresponds to @xmath19 .
results for the spectral functions are shown in fig .
[ wilson_fig ] .
as one can see , the spectral functions for quenched and partially quenched qcd are very similar , indicating that for quark masses around the strange quark mass effects due to quenching are quite small .
in addition we have also calculated meson spectral functions wit hyp smeared clover action in @xmath20 partially - quenched qcd , i.e. we have applied hyp smearing on the dynamical configuations before calculating the meson correlators .
we have performed calculations at several quark masses . from extrapolation of
the pseudo - scalar mass squared to zero we determine the critical hopping parameter to be @xmath21 ( statistical error only ) .
the purpose of calculation with hyp smearing was to see whether the smoothing of the gauge fields removes the artificial peaks , and the corresponding results are shown in fig .
[ wilson_fig ] .
as one can see from the figure , the lattice artifacts are reduced but not removed completely by hyp smearing . however , as a by - product we observe that number of exceptional configurations as well as the number of cg in the fermions matrix inversions are greatly reduced . in fact
we did not observe any exceptional configurations in a sample of 59 configurations down to quark masses @xmath22 ( @xmath23 is the strange quark mass ) .
this work has been authored under the contract de - ac02 - 98ch10886 with the u.s .
department of energy .
tb was partly supported by the ldrd program of brookhaven national laboratory , project no .
we thank the rbc collaboration for the dwf correlators .
the program for the mem analysis was provided by ines wetzorke to whom we are very grateful .
the calculations were performed on the ibm - sp supercomputer at nersc using gauge configurations from the milc collaboration and the milc code for which we are grateful .
99 y. nakahara _
et al . _ ,
d * 60 * ( 1999 ) 091503 ; m. asakawa _
et al . _ ,
prog . part .
* 46 * ( 2001 ) 459 yamazaki , _ et al._. , phys .
d65 ( 2002 ) 014501 f. karsch , _ et al .
b530 * ( 2002 ) 147 t. umeda _
et al . _ , hep - lat/0211003 m. asakawa and t. hatsuda , phys.rev.lett . 92 ( 2004 ) 012001 s. datta , _ et al . _ ,
d69 ( 2004 ) 094507 f. karsch , _ et al .
d * 68 * ( 2003 ) 014504 g.p .
et al._. , nucl.phys.proc.suppl .
106 ( 2002 ) 12 - 20 | we study the lattice spacing dependence of meson spectral functions calculated in quenched qcd with domain wall fermions as well as clover wilson fermions in quenched and partially - quenched qcd .
we conclude that for lattice spacing @xmath0 gev all excited states appearing in the spectral functions are lattice artifacts . |
in this short paper at first i shall try to show links between studies of cosmic rays ( crs ) and studies of close - by compact objects ( in particular neutron stars nss ) .
the reason for an existence of such relation is obvious : both phenomena ( crs and nss ) have the same origin
supernova ( sn ) explosions .
then i discuss in more details the population synthesis of close - by nss .
the analysis of the population of these sources makes us to conclude that the solar neighbourhood ( by that i mean a region about few hundred of parsecs around the sun ) in enriched with young nss .
it is a natural consequence of the existence of the gould belt local structure formed by massive stars .
the main message i want to deliver in this contribution is the following .
we are living in a region of the milky way enriched with massive stars in comparison with a typical place at 8 kpc from the center of the galaxy . due to this fact sn rate in @xmath0 600 pc around the sun during the last few tens of myrs is enhanced .
it results also in the ebhanced number of near - by young compact objects .
now we know about two tens of young ( age @xmath1 4 myrs ) close ( distance @xmath1 1 kpc ) nss .
this local ( in space and time ) increase of sn rate can be important for modeling of cr flux on the earth .
galactic crs are genetically connected with sn : crs are accelerated in sn remnants ( see other materials in this volume , especially contributions by michal ostrowski , vladimir ptuskin and john wefel ) . the local cr density is inevitably dependent on the rate of sn in the galaxy in general and in the solar proximity in particular .
cr flux can be increased if the sun appears in a region of the galaxy with enhanced sn rate , or if by chance several sn explosions in a row happen in the vicinity of the sun .
for example shaviv ( 2004 ) suggested , that passages of the sun through galactic spiral arms are accompanied by increase of cr flux on earth ( at this conference some results were presented in the talk by smadar levi ) .
as crs can have influence on earth climate ( in particular on cloud formation ) such passages can be responsible for the ( quasi)periodic appearence of ice age epochs on our planet .
however , i would like to note , that even local quasiperiodic fluctuations of starformation rate ( and correspondent changes in sn rate ) can give a similar effect .
such fluctuations were found by vanbeveren and de donder ( 2003 ) in their modeling of star formation in the solar vicinity . another possibility to enhance sn rate
is connected with the formation of structures like the gould belt ( see description of the belt below ) .
all these effects can work together , and therefore modeling of the cr flux history becomes a very complicated business because of a `` turbulent history '' of star formation in @xmath0 1 kpc around the sun .
for a long time local sn were considered as one of candidates for a catastrophic event responsible for mass extinction of species .
still sn potentially can play a `` positive '' role in the history of life on earth .
an interesting possibility of sn influence on bioworld of our planet can be connected with the hypothesis discussed by tsarev ( 1999 ; it is also described in an easier available paper by chernavskii 2000 ) . in the original version
the author suggested that chirality symmetry breaking existing in the bioorganic world can be connected with strong neutrino burst due to a close sn which happened in the early stages of the history of the earth .
however , a close - by sn ( close enough to povide a strong neutrino flux ) can just destroy terrestrial bioorganics . in that sense it is possible to suggest a more promising hypothesis which includes so called `` choked gamma - ray bursts '' ( meszaros , waxman 2001 ) . if a jet is unable to penetrate through the stellar envelope , then only a neutrino beam will break - out . in that case
it is possible both : to increase the neutrino flux , and to decrease negative influence of hard radiation of a sn on the biosphere .
recent local sn ( and related close - by ns ) can strongly influence not only the flux of cr but the spectrum as well .
erlykin and wolfendale ( 1997 ) proposed that a knee in the spectrum is due to a single source . according to their estimates there
should exist a near - by relatively young sn remnant .
an estimated age of such a hypothetical object was found by these authors to be equal to 80 - 100 kyr , and its distance to 230 - 350 pc .
they suggested that the monogem nebula ( and related to it psr 0656 + 14 ) can be such a source which produce the knee in the spectrum ( see also the contribution by jorg hoerandel in this volume for a more detailed discussion of the knee ) .
how probable an appearence of such an object is ?
if the sun is sutuated in a typical region of the galaxy at 8 kpc from its center then the probability of appearence of a sn remnant of necessary age at necessary distance is very low .
however , we are living in a region of enhanced sn rate , and there are many young compact objects around us . in the following section
i am going to discuss this population and it origin in more details .
the solar proximity is rich in young nss .
some of them represent a new class of sources : radioquiet x - ray dim isolated nss ( see a review in treves et al .
the reason of such enrichment is an existence of a structure formed by massive stars the gould belt .
in the following subsection we discuss it in some details .
then we present new results on a population synthesis of young nss , and finally we discuss a convenient way to represent this population .
population of near - by young nss is presented in the table ( reproduced from popov , turolla 2004 ) .
we included there 20 objects .
thirteen of them are detected as thermal emitters , i.e. as _
coolers_. they are seven radioquiet rosat nss ( aka the magnificent seven
m7 ) , geminga and geminga - like source 3eg j1835 + 5918 , and four normal radiopulsars ( among which the famous vela pulsar , and psr b0656 + 14 which is probably connected with the single source , see above ) .
seven close - by nss which are not detected as thermal emitters ( shown at the bottom of the table ) are normal radiopulsars with ages 2 - 4 myrs .
the gould belt is a structure consisting of clusters of massive stars .
about 2/3 of massive stars inside 600 pc around the sun belong to the belt ( see a recent paper by grenier 2004 for main references ) .
the gould belt radius is about 300 - 500 pc .
it is inclined at @xmath018 - 22@xmath2 respect to the galactic plane .
the sun is situated not far from the center of that disc - like structure .
the age of the belt is not well estimated , 30 - 60 myrs can be a good guess .
the origin of this system is unclear .
one of the possibility is an impact of a high velocity cloud with the galactic disk .
it is important to note , that the sun and the belt are not genetically related to each other . due to the presence of the belt
the rate of sn around us ( say in a few hundred parsecs ) during last several tens of million years is higher than it is at an average place at a solar distance from the galactic center .
the enhancement is about 3 - 4 times over the galactic average .
grenier ( 2003 ) estimates a sn rate in the belt as 20 - 27 per myr during the last few myrs . because of that there should be a local overabundance of young nss which can appear as hot cooling objects , as gamma - ray sources etc .
( in priciple in addition to nss there should be young bhs in the solar neighbourhood , but these objects are much more elusive ; it would be very interesting to detect them , see prokhorov , popov 2002 ) .
for example , such young nss can form a subpopulation among egret unidentified sources .
the existence of the local bubble can be also linked with the belt ( berghofer , breitwerdt 2002 ) .
one of the most powerful methods of investigating properties of some population of sources is a log n log s distribution .
it is an integral distribution , which shows the number of sources with a flux larger than a given value .
it is useful to remind two limiting cases .
if sources are distributited homogeniously and isotropicaly then the slope of log n
log s is equal to -3/2 : the number of sources grows as the cube of a distance , but the observed flux decreases as the square . in the case of a disk - like system
the slope is -1 ( the number of objects is growing only as a square of a distance ) .
if sources are located at the same distance ( for example , sources in a given far - away galaxy ) then log n log s shows distribution of sources in luminosity .
popov et al . ( 2003 ) calculated log n log s distribution of cooling nss in the solar vicinity , which can be observed by rosat and other x - ray missions .
the results are shown in the fig .
[ lognlogs ] . & & & & & & + source name & period & cr@xmath3 & @xmath4 & d & age@xmath5 & refs + & s & cts / s & @xmath6 s / s & kpc & myrs & + & & & & & & + magnificent seven & & & & & & + rx j1856.5 - 3754 & & 3.64 & &0.117@xmath7&@xmath8 & [ 1,2 ] + rx j0720.4 - 3125 & 8.37 & 1.69 & @xmath9 & & & [ 1,3 ] + rx j1308.6 + 2127 & 10.3 & 0.29 & & & & [ 1.4 ] + rx j1605.3 + 3249 & & 0.88 & & & & [ 1 ] + rx j0806.4 - 4123 & 11.37 & 0.38 & & & & [ 1,5 ] + rx j0420.0 - 5022 & 3.45 & 0.11 & & & & [ 1,11 ] + rx j2143.7 + 0654 & & 0.18 & & & & [ 6 ] + & & & & & & + geminga type & & & & & & + psr b0633 + 17 & 0.237 & 0.54@xmath10 & 10.97&0.16@xmath7&0.34 & [ 7 ] + 3eg j1835 + 5918 & & 0.015 & & & & [ 8 ] + & & & & & & + thermal . emit .
psrs & & & & & & + psr b0833 - 45 & 0.089 & 3.4@xmath10 & 124.88 & 0.294@xmath7 & 0.01 & [ 7,9,10 ] + psr b0656 + 14 & 0.385 & 1.92@xmath10 & 55.01 & 0.762@xmath11 & 0.11 & [ 7,10 ] + [ 1ex ] psr b1055 - 52 & 0.197 & 0.35@xmath10 & 5.83 & @xmath12 & 0.54 & [ 7,10 ] + [ 1ex ] psr b1929 + 10 & 0.227 & 0.012@xmath10 & 1.16 & 0.33@xmath7 & 3.1 & [ 7,10 ] + & & & & & & + other psrs & & & & & & + psr j0056 + 4756 & 0.472 & & 3.57 & 0.998@xmath11 & 2.1 & [ 10 ] + psr j0454 + 5543 & 0.341 & & 2.37 & 0.793@xmath11 & 2.3 & [ 10 ] + psr j1918 + 1541 & 0.371 & & 2.54 & 0.684@xmath11 & 2.3 & [ 10 ] + psr j2048 - 1616 & 1.962 & & 10.96 & 0.639@xmath11 & 2.8 & [ 10 ] + psr j1848 - 1952 & 4.308 & & 23.31 & 0.956@xmath11 & 2.9 & [ 10 ] + psr j0837 + 0610 & 1.274 & & 6.8 & 0.722@xmath11 & 3.0 & [ 10 ] + psr j1908 + 0734 & 0.212 & & 0.82 & 0.584@xmath11 & 4.1 & [ 10 ] + + + + + + + + + + + + the main output of these calculations can be formulated as follows : observed cooling isolated nss originated in the gould belt .
this result does not depend on the uncertainties of the population synthesis model .
i want to note , that according to our calculations there should be about several tens more unidetified isolated nss with fluxes detectable by rosat .
these objects are hiding in crowded regions close to the galactic plane ( and to the plane of the gould belt ) .
in addition there should be many nss with ages above @xmath0 1 myr , which now are too cold now to be detected .
a total expected number of nss depending on their ages and distances can be easily shown on the age - distance diagram .
if all uncertainties like initial spatial distribution of nss , their kick velocity distribution , characteristics of emission ( atmospheric effects etc . ) , mass spectrum of nss are fixed , then log n - log s distribution of close - by nss can be a powerful tool to put constraints on models of thermal evolution of nss . to illustrate the population of young close - by nss it is convenient to use the age - distance diagram ( popov 2004 ) .
if we are speaking about observations of thermal emission of young cooling nss then most of their properties depend on their age , and their detectability in x - rays obviously strongly depends on the distance ( not only because of the flux dilution , but also because of the strong interstellar absorption of soft x - rays ) . in that sense it is useful to plot an age - distance diagram ( add ) for these objects .
there are several reasons to introduce adds .
first an add easily illustrates distributions of sources in age and distance .
then it is possible to plot `` expectation lines '' for an abundance of sources of different types and compare them with data ( i.e. line which show how many sources of smaller than a given value we can expect to find at distances smaller than a given one ) .
finally since the observability of sources depends mainly on their ages and distances , it is possible to illustrate observational limits . in the fig .
[ agedist ] an add for close - by young nss is presented .
these objects can be observed in soft x - rays due to their thermal emission .
the thermal evolution strongly depends on the mass . according to different models ( see kaminker et al . 2002 , blaschke et al . 2004 and references there in ) a ns with a mass @xmath13@xmath14 remains hot ( @xmath15@xmath16 k , or equivalently @xmath0 50100 ev ) up to @xmath0 0.51 myr .
it corresponds to a luminosity @xmath17 erg s@xmath18 .
usually nss with low masses ( 11.3 @xmath19 ) cool down slower , while with higher masses faster . to plot an add it is necessary to fix maximal values for age and distance for selection of sources .
i choose 4.25 myrs as a limiting age for selection of observed sources .
this is the time after which even a low - mass nss cools down to @xmath20 k and becomes nearly undetectable in x - rays ( see kaminker et al .
2002 , popov et al .
a limiting distance is taken to be equal to 1 kpc ( because of the absorption it is difficult to detect in x - rays an isolated ns at larger distance unless the star is very hot ) .
there are 20 sources of different nature which are supposed to be young close - by nss ( age @xmath1 4.25 myr , distance @xmath1 1 kpc ) ( see popov et al .
2004 and the table above ) : the m7 , geminga and the geminga - like source rx j1836.2 + 5925 , four psrs with detected thermal emission ( vela , psr 0656 + 14 , psr 1055 - 52 , psr 1929 + 10 ) , and seven psrs without detection of thermal emission .
not for all of these sources there are good estimates of ages and/or distances ( especially for the m7 ) . in the figure there are data points for 13 objects for which such estimates exist ( distance to psr b1055 - 52 is uncertain , and we accept it to be 1 kpc ) . note , that there are also two psrs with distance @xmath11 kpc and with ages in between 4.25 and 5 myr ( psr b0823 + 26 , psr b0943 + 10 ) and psr j0834 - 60 with distance 0.49 kpc and unknown age .
these three objects are not included into the figures .
also psr b1822 - 09 with age 0.23 myr and distance 1 kpc is not plotted .
two types of objects are distinguished on the graph : detected ones ( shown as black diamonds ) and undetected due to the thermal x - ray emission ( remember , there are at least seven additional sources six from the m7 and one geminga - like object for which there are no definite determinations of age or / and distance ) . in the picture
the `` expectation '' lines are shown with an inclusion of the dynamical evolution .
i.e. here nss movements in the galactic potential are accurately calculated . to plot them we use the same model of progenitor distribution and sn rate as in ( popov et al .
2003 ) . at birth nss
obtain kick velocities in accordance with arzoumanian et al .
( 2002 ) model , and move in the galactic potential .
five `` expectation '' lines ( solid ) are plotted for 1 , 4 , 13 , 20 and 100 sources .
for example each point on the line for 20 sources shows us values of age ( @xmath21 ) and distance ( @xmath22 ) such that according to our model we expect to see 20 sources with ages less than @xmath21 at distances less than @xmath22 .
the dotted line represents the `` visibility '' line .
the idea of adding such a line is to show the maximal distance for a given age ( or vice versa a maximal age for a given distance ) at which a hot ( i.e. low - mass ) ns can be detected . to calculate this line i take the cooling curve from ( kaminker et al .
2002 ) for @xmath23@xmath24 ( in the model of these authors cooling of nss with @xmath25 is nearly mass - independent ) .
such curves were used for example in ( popov et al .
as soon as a cooling curve is fixed then the age determines the luminosity of the object .
the limiting unabsorbed flux is assumed to be @xmath26 erg @xmath27 s@xmath18 . according to webpimms it corresponds to @xmath28 rosat psps counts per second for @xmath29 @xmath27 and a blackbody spectrum with @xmath30 ev , or to @xmath31 rosat psps counts per second for @xmath32 @xmath27 and a blackbody spectrum with @xmath33 ev .
the latter values corresponds to the dimmest source among the m7
rx j0420.0 - 5022 ; the former to possibly detectable hot far away objects . without any doubt such a simplified approach underestimates the absorption at large distances .
so the age at which a ns is still observable is overestimated , but for distances @xmath1 1 kpc and ages @xmath1 1 myr it should not be a dramatic effect . the fact that the line for 20 sources lies below @xmath01/2 of the observed points tells us that according to our model not all nss are observed .
for example , at r=1 kpc we expect to see 20 sources with ages @xmath00.5 myr , but in reality we observe 20 sources with age @xmath14.25 myr .
it means that for a constant ns formation rate we see only about 10% of them .
one has to bear in mind that for seven sources ages or / and distances are unknown , but they are suspected to be young and close , this is why here we discuss 20 and not 13 sources .
their inclusion or exclusion changes our conclusions quantitatively , but not qualitatively .
six of the m7 sources which are not plotted on the graphs due to lack of good distance measurement should populate the left bottom corner of the graphs , since according to models of ns thermal evolution ( see for example kaminker et al .
2002 , blaschke et al .
2004 ) their ages are expected to be @xmath1 1 myr , and their distances should be @xmath1 500 pc .
nss can easily escape detection as _
coolers _ simply because they cooled down in @xmath0 1 myr .
a fraction of psrs can be undetected due to beaming ( however , they may be observed as egret sources ) .
psr b1929 + 10 lies above the `` visibility line '' .
however it is unclear if x - ray emission of this faint ( 0.012 rosat cts s@xmath18 ) source is due to non - thermal mechanism or not ( becker & trumper 1997 ) .
now this object is not widely accepted as a _ cooler _ and usually it is not plotted on the @xmath34 ( temperature - age ) graphs with cooling curves ( see for example kaminker et al .
2002 and blaschke et al .
2004 ) . surprisingly fluctuations in the part of the plot with small number of nss ( bottom left ) are not large : four detected _ coolers _ lie just below the line for four sources .
probably we see nearly all low - massive nss with ages @xmath1 1 myr at @xmath35 300400 pc .
it should be noted that we are very lucky to have such a young and close object as the vela pulsar .
if one believes in the mass spectrum with small fraction of nss with @xmath36 then inevitably one has to conclude that vela can not be a massive ns .
it can be important in selection of cooling models ( see blaschke et al .
if in a model vela is explained only by a cooling curve for @xmath36 then the model may be questionable .
in fact , having such a young and massive ns so close is very improbable .
a similar conclusion can be made for geminga and psr 0656 + 14 .
there is a deficit of sources below the `` visibility '' line .
these sources could be already detected as dim x - ray sources by rosat but were not identified as isolated nss .
the limit for the number of isolated nss from the bsc ( bright source catalogue ) is about 100 sources at rosat count rate @xmath370.05 cts s@xmath18 ( see rutledge et al .
however we do not expect to see so many young isolated nss due to their thermal emission .
an expected number of _ coolers _ observable by rosat is about 40 objects with ages @xmath38 1 myr inside @xmath391 kpc ( see the `` visibility '' line in the fig.[agedist ] in comparison with the line for 20 sources , for example ) . as it was shown by popov et al .
( 2003 ) most of unidentified _ coolers _ in rosat data are expected to be located at low galactic latitudes in crowded regions . according to the classical picture all nss
were assumed to be born as psrs more or less similar to crab .
depending on the initial parameters they were assumed to be active for @xmath40 yrs . in the last years this picture was significantly changed . without any doubts
we see a deficit of young psrs in the solar vicinity in comparison with an expected value of young nss from the gould belt and the beaming factor can not be the only reason for this deficit .
this deficit can be explained if one assumes that a significant part of nss do not pass through the radio pulsar stage or that this stage is extremely short for them ( long periods of nss from the magnificent seven prove it ) . of course fluctuations ( in time and space ) of ns production rate can be important as far as statistics is not that high .
i conclude this subsection with a note that an add can be the useful tool for an illustration of the properties of close - by nss .
its modifications can be applied to other types of sources .
for example an addition of the third axis ( for @xmath41 or @xmath42 for example ) can be useful in discussing the population of radiopulsars .
i want to stress once again that the region of @xmath0 few hundred parsecs around the sun is characterized by an enhanced sn rate .
this is due to the gould belt .
the belt ( as a structure formed by bright stars ) was well known since 19th century . in addition now we observe young close - by nss which , as it was shown by population synthesis calculations , originated in the belt .
some of egret unidentified sources also are suspected to be young nss born in the belt .
1 arzoumanian , z. , chernoff , d.f . , cordes , j.m .
`` the velocity distribution of isolated radio pulsars '' , apj 568 , 289 .
becker , w. , trumper , j. ( 1997 ) .
`` the x - ray luminosity of rotation - powered neutron stars '' , a&a 326 , 682 .
berghofer , t.w . ,
breitschwerdt , d. ( 2002 ) .
`` the origin of the young stellar population in the solar neighborhood a link to the formation of the local bubble ? '' , a&a 390 , 299 .
blaschke , d. , grigorian , h. , voskresensky , d. ( 2004 ) .
`` cooling of neutron stars .
hadronic model . '' , a&a ( in press ) [ astro - ph/0403170 ] .
chernavskii , d.s .
( 2000 ) , `` the origin of life and thinking from the modern physics point of view '' , physics uspekhi 170 , 157 .
dermer , c.d .
`` gamma ray bursts , sepernovae , and cosmic ray origin '' , in : proc . of iscra 13th
course , eds . m. shapiro , t. stanev and j. wefel , world scientific publ .
erlykin , a.d . ,
wolfendeil , a.w .
( 1997 ) . `` a single source of cosmic rays '' , j. phys .
g 23 , 979 .
grenier , i.a .
`` unidentified egret sources in the galaxy '' , astro - ph/0303498 .
grenier , i.a .
( 2004 ) . `` the gould belt , star formation , and the local interstellar medium '' , astro - ph/0409096 .
haberl f. , zavlin v. ( 2002 ) .
`` xmm - newton observations of the isolated neutron star rx j0806.4 - 4123 '' , a&a 391 , 571 .
haberl , f. et al .
`` the isolated neutron star x - ray pulsars rx j0420.0 - 5022 and rx j0806.4 - 4123 : new x - ray and optical observations '' , a&a 424 , 635 .
hambaryan , v. , hasinger , g. , schwope , a. d. , schulz , n. s. ( 2001 ) .
`` discovery of 5.16 s pulsations from the isolated neutron star rbs 1223 '' , a&a 381 , 98 .
hobbs , g. manchester , r. teoh , a. hobbs , m. ( 2003 ) .
`` the atnf pulsar catalogue '' , in : proc . of iau symp
`` young neutron stars and their environment '' [ astro - ph/0309219 ] .
kaminker , a.d . , yakovlev , d.g .
, gnedin , o.y .
( 2002 ) . `` three types of cooling superfluid neutron stars : theory and observations '' , a&a 383 , 1076 .
kaplan , d.l . ,
van kerkwijk , m.h . ,
anderson , j. ( 2002 ) .
`` the parallax and proper motion of rx j1856.5 - 3754 revisited '' , apj , 571 , 447 .
meszaros , p. , waxman , e. ( 2001 ) .
`` tev neutrinos from successful and choked gamma - ray bursts '' , phys .
87 , 171102 .
mirabal , n. , halpern , j.p .
( 2001 ) . `` a neutron star identification for the high - energy gamma - ray source 3eg j1835 + 5918 detected in the rosat all - sky survey '' , apj 547 , l137 .
pavlov , g. et al .
`` the x - ray spectrum of the vela pulsar resolved with the chandra x - ray observatory '' , apj 552 , l129 .
pohl , m. , perrot , c. , grenier , i. , digel , s. ( 2003 ) .
`` the imprint of gould s belt on the local cosmic - ray electron spectrum '' , a&a 409 , 581 .
popov , s.b .
, turolla , r. , prokhorov , m.e . ,
colpi , m. , treves , a. ( 2003 ) .
`` young close - by neutron stars : the gould belt vs. the galactic disc '' , astro - ph/0305599 .
popov , s.b .
, turolla , r. , prokhorov , m.e . ,
colpi , m. , treves , a. ( 2004 ) .
`` young compact objects in the solar vicinity '' , in : proc . of iscra 13th course , eds . m. shapiro , t. stanev and j. wefel , world scientific publ .
popov , s.b .
, turolla , r. ( 2004 ) .
`` isolated neutron stars : an astrophysical perspective '' , in : proc . of
the nato advanced research workshop on superdense qcd matter and compact stars ( in press ) [ astro - ph/0312369 ] .
popov , s.b .
`` age - distance diagram for close - by young neutron stars '' , astro - ph/0407370 .
prokhorov , m.e . ,
popov , s.b .
`` close young isolated black holes '' , astronomy letters 28 .
rutledge , r.e . , fox , d.w .
, bogosavljevic , m. , mahabal , a. ( 2003 ) .
`` a limit on the number of isolated neutron stars detected in the rosat bright source catalogue '' , astro - ph/0302107 .
schwope , a.d . ,
hasinger , g. , schwarz , r. , haberl , f. , schmidt , m. ( 1999 ) .
`` the isolated neutron star candidate rbs 1223 ( 1rxs j130848.6 + 212708 ) '' , a&a 341 , l51 .
shaviv , n. ( 2004 ) .
`` cosmic ray diffusion in the milky way : model , measurement and terrestrial effects '' , in : proc . of iscra 13th course , eds . m. shapiro , t. stanev and j. wefel , world scientific publ . , p.113 .
treves , a. , turolla , r. , zane , s. , colpi , m. ( 2000 ) .
`` isolated neutron stars : accretors and coolers '' , pasp 112 , 297 .
tsarev , v.a .
`` chiral effects of neutrino after supernova explosions '' , bulletin of the lebedev physics institute 2 , 22 .
vanbeveren , d. , de donder , e. ( 2003 ) .
`` the effect of the process of star formation on the temporal evolution of the wr and o - type star populations in the solar neighbourhood '' , astro - ph/0309104 .
zampieri , l. et al .
`` 1rxs j214303.7 + 065419/rbs 1774 : a new isolated neutron star candidate '' , a&a 378 , l5 .
zane , s. et al .
`` timing analysis of the isolated neutron star rx j0720.4 - 3125 '' , mnras 334 , 345 . | i discuss young close - by compact objects , recent supernovae in the solar neighbourhood , and point to their connection with cosmic ray studies .
especially the role of the gould belt is underlined .
neutron stars , evolution , supernovae , cosmic rays |
extreme horizontal branch stars ( ehbs ) are hot ( t@xmath320000 k ) , evolved stars of low initial mass ( 0.72 m@xmath2 ) burning helium in their core @xcite .
they have lost most of the hydrogen envelope during their evolution , to the point that the external layer is too thin to sustain the hydrogen burning shell .
thus , after the exhaustion of helium , they are expected to evolve directly to the white dwarf cooling sequence ( agb - manqu stars * ? ? ?
they are found both in the galactic field and in globular clusters ( gcs ) , and the comprehension of their formation mechanisms is required to understand the late stages of the evolution of low - mass stars .
they are also responsible for the ultraviolet emission of elliptical galaxies @xcite .
the causes of the heavy mass loss required to form an ehb star are still under debate , and this is one of the foremost grey points of low - mass stellar evolution theory .
dynamical interactions inside binary systems are considered a major channel to produce an ehb star @xcite , as confirmed by the high frequency of close binaries among ehbs in the galactic field ( e.g. , * ? ? ?
however , it is unclear if the same mechanisms are at work in the denser environment of gcs , where close ehb binaries are very rare @xcite .
the old age of the cluster stellar population was proposed as an explanation @xcite , but an alternative scenario has become more popular in the last decade : ehbs could be the progeny of second - generation , single , helium - enriched stars @xcite .
the star m5865 in ngc6752 is so far the only spectroscopically - confirmed ehb close binary in a gc @xcite .
later after discovery , @xcite showed that the presence of a g- or k - type main sequence ( ms ) companion had previously been proposed for this star by @xcite , to explain its red color and a faint mgib triplet in its spectrum . a ehb star with a close companion of this kind has never been observed among the more than 180 ehb binaries discovered to date in the galactic field @xcite .
in fact , their close companions are all either compact remnants ( white dwarfs or neutron stars ) , or very low - mass ms stars ( [email protected] m@xmath2 , e.g. , * ? ? ? * ; * ? ? ?
more massive ms companions have been found only in wide pairs ( e.g. , * ? ? ? * ) or hierarchical triple systems @xcite .
indeed , m5865 could be such a rare triple system , where the ms star in a wide orbit does not interfere with the evolution of the inner , close pair .
alternatively , the ms star could also be a foreground object physically unrelated to the ehb star , a common fact in the crowded environment of gcs .
we analyzed the available data of m5865 to identify the role of the ms star in the system .
the data sets consisted of @xmath5 ) twenty - one high - resolution spectra ; @xmath6 ) five intermediate - resolution spectra ; @xmath7 ) composite ground - based optical images ; @xmath8 ) more than 800 frames of time - series photometry in the v band .
we estimated the mass of the ms star ( m@xmath9 ) , comparing the position of the star in the color - magnitude diagram ( cmd ) with synthetic loci representative of a ehb star plus a ms companion .
we empirically determined the starting point of the curve , i.e. the position of a single cluster hb star ( @xmath10 , @xmath11 ) at the temperature of the hot component of the system ( t@xmath12 k , * ? ? ? * ) .
we fitted a @xmath13t@xmath14 relation to the hotter part of the horizontal branch ( hb ) of ngc6752 , matching photometric colors ( obtained with wfi at the mpg / eso telescope , * ? ? ? * ) with spectroscopic temperature measurements @xcite .
we excluded all the stars with `` anomalous '' spectroscopic results , in particular the eight ehbs with high spectroscopic mass , and we were left with 23 stars .
we thus derived @xmath15 .
the error was estimated from the scatter of the points with respect to the fitted curve , as also done for the magnitude obtained in the next step .
we then fitted a fourth - order polynomial to the hb in the cmd ( dashed curve in fig.[f_mass ] ) , and from the color @xmath10 we derived @xmath16 .
given the expected photometric properties of the ehb component alone , we calculated the resulting magnitude and color when the flux of a ms star between m@xmath9=0.45 and 0.78 m@xmath2 is added .
this was taken from yale - yonsei theoretical isochrones @xcite for metallicity z=0.0004 and age=12.5 gyr @xcite , assuming @xmath17=13.38 , e(@xmath18)=0.04 @xcite , and a standard reddening law @xcite .
we thus obtained the curve shown in fig .
[ f_mass ] , with color and magnitude increasing with m@xmath9 .
the point closest to the observed position returned m@xmath19m@xmath2 .
the error was estimated from the variations of the result when the starting point was shifted by 1@xmath20 in either magnitude or color ( assuming an error of 0.05 mag in distance modulus ) .
the photometric error of the observed position was considered also , but it resulted negligible ( 0.006 and 0.012 mag in @xmath21 and ( @xmath22 ) , respectively ) .
the mgib triplet , signature of the ms star , was observed in five intermediate - resolution ( r=4000 ) spectra collected with fors2 at paranal observatory between 2002 , june 11 and 14 , with the 1400v grism and a 0.5@xmath23-wide slit .
we remand the reader to @xcite for more details about the observations and the data reduction .
we obtained a spectroscopic estimate of m@xmath9 by fitting this feature with synthetic spectra .
we shifted the spectra to laboratory wavelength and co - added them .
flux - calibrated synthetic spectra were calculated with the spectrum lte spectral synthesis code @xcite , fed with model atmospheres obtained by interpolating on the @xcite grid .
the temperature and gravity of the ehb star were taken from our previous measurements @xcite .
we calculated five models with t@xmath24 between 3800 and 5800 k , in steps of 500 k. their gravities and masses were taken from the same models used before .
we interpolated the three hottest models at a step of 100 k to produce a finer grid in the range t@xmath24=48005800
k. the stellar radii were estimated as @xmath25 , where @xmath26 is the surface gravity and @xmath27 is the mass , assuming 0.5 m@xmath2 for the ehb star .
we then scaled the fluxes for the different radii of the two components , and finally co - added them .
we compared each model with the observed spectrum , calculated the @xmath28 statistics , and fitted its dependence on t@xmath24 with a third - order polynomial .
we thus found a minimum @xmath28 at t@xmath29 k , corresponding to an early k - type ms star with m@xmath30 m@xmath2 .
the fit is shown in fig .
[ f_mass ] .
the errors were estimated from the statistical behavior of the @xmath28 function .
they are likely underestimated , because only random noise of the observed spectrum is taken into account , while systematic effects , such as errors in the continuum placement or the exact magnesium abundance , can occur .
the two independent approaches returned very similar results .
accounting for possible systematics , we derive m@[email protected] m@xmath2 . according to the employed models ,
the cool star is therefore an early k - type ms object with t@xmath3120 k and @xmath320.1 .
seventeen high - resolution ( r=18000 ) spectra centered on the h@xmath33 line were collected between june and september 2007 in service mode at paranal observatory , with the flames - giraffe spectrograph and the h7a setup .
this data set was complemented with four more epochs , collected on 2005 june 29 , with the same instrument and setup @xcite .
we reduced the data by means of the cpl - based eso pipeline .
we checked the wavelength calibration analyzing the extracted and calibrated lamp fibers , and we found only small random deviations from the laboratory wavelengths ( 0.3 km s@xmath34 rms ) .
we extracted the science spectra with an optimum algorithm @xcite , and subtracted the smoothed average of ten fibers allocated to the sky background .
some spectra presented a wide emission feature on the red wing of the h@xmath33 line , as a result of contamination from a nearby lamp fiber .
we fitted it with a gaussian profile and removed it .
we repeated the procedure varying the constraints ( continuum definition and fitted data points ) , to estimate the uncertainty thus introduced on the rvs .
this was negligible ( @xmath35 km s@xmath34 ) in all but two spectra , where it was of the order of 2.5 km s@xmath34 . to measure the rvs , we cross - correlated @xcite the h@xmath33 line with its wings ( 48404880 ) with a synthetic spectrum with parameters matching those of the ehb star , drawn from the library of @xcite .
we corrected the observed rvs to heliocentric values .
the spectra of about 60 hot stars were contemporarily observed by the multi - object giraffe spectrograph .
we estimated and removed zero - point offsets between the frames forcing the average rv of the 40 best targets ( in terms of spectral quality and measurement accuracy , and showing no relevant rv variation ) to be constant among the epochs ( see * ? ? ?
* for more details ) .
we estimated the associated uncertainty from the related error - on - the - mean .
we analyzed the periodograms of the data calculated with several algorithms , such as the analysis of variance ( anova , * ? ? ?
* shown as an example in fig .
[ f_anova ] ) , the lomb - scargle algorithm @xcite , the @xcite fourier analysis , the data - compensated discrete fourier transform @xcite , and the fourier analysis of light curves of @xcite .
a prominent peak at p=1.61 days dominates all the periodograms .
the corresponding phased rv curve is shown in fig .
[ f_phase ] .
this solution is robust , because it is stable against the exclusion of up to five of the most uncertain measurements , and none of the secondary peaks returned an equally satisfactory solution , in terms of rms deviation and @xmath36-statistics of the fit . despite the high significance of the rv variations ( 4.6@xmath20 ) ,
the semiamplitude of the rv curve ( [email protected] km s@xmath34 ) is small compared to the typical variations of close ehb systems @xcite .
this suggests a nearly face - on orbital plane . unfortunately , m5865 is outside the field of view of all hst archive images .
we therefore stacked our thirteen best - seeing , ground - based , high - spatial resolution frames in search of hints of a ms star projected along the line of sight .
no physical companion should be detected , as even a wide binary system would be unresolved at the cluster distance ( 4.2 kpc , * ? ? ?
the data were collected with magic at the clay telescope at las campanas observatory on may 9 , 2006 , with a spatial resolution of @xmath37 per pixel .
the effective seeing of the composite image is 0.41@xmath23 , and m5865 has exactly the same full - width - at - half - maximum as other isolated stars in this frame . any blended star should be closer than 0.16@xmath23 , or a larger point - spread - function would have been detected .
m5865 is located at a distance to the center of twice the cluster half - mass radius @xcite , where the crowding conditions are not severe .
the density of stars within 0.5@xmath38 from m5865 in the wfi photometric catalog @xcite , corrected for the catalog s completeness ( 95% for stars with @xmath39 ) , is @xmath00.028 stars arcmin@xmath40 .
hence , the probability of a chance aligment with separation @xmath41 is 0.2% . as discussed in
[ ss_lr ] , the ms star is rv variable .
the fraction of ms binary stars in the cluster is extremely low ( @xmath421% in the external regions , * ? ? ?
* ) , and the joint probability of having a blended ms foreground star that is also a binary is negligible ( @xmath43 ) .
the ms star must therefore be part of the m5865 system .
we used the fors2 spectra presented in
[ ss_mass ] to measure and compare the rv of the h@xmath33 line and of the mgib feature .
@xcite showed that these data are affected by rv zero - point offsets up to 1015 km s@xmath34 , and they are suitable only for measurements of rv variations , with precision of 34 km s@xmath34 .
we therefore measured the rv variations ( @xmath44 ) in the spectra with respect to the first spectrum of the series , and we corrected them as detailed in @xcite .
we found rv variations of the mgib triplet , significant at the 3.8@xmath20 level .
we then derived the absolute rv of the reference spectrum ( rv@xmath45 ) , cross - correlating it with the best - fit synthetic template used for the spectroscopic estimate of m@xmath9 .
the combination of @xmath44 and rv@xmath45 returned the absolute rv in each spectrum , but all these measurements were offset by the zero - point systematic of the reference spectrum .
hence , we phased the rvs of the h@xmath33 line with the ephemeris obtained in sect.[ss_hr ] , then we compared them with the best - fit rv curve ( black curve in fig .
[ f_phase ] ) , and we minimized the @xmath28 statistics to find the zero - point correction . the third object of a hierarchical triple system must show a low - amplitude , low - frequency rv curve unrelated to the orbital period of the inner pair .
the rv variations of the mgib triplet are not compatible with this scenario , because they are too large ( comparable to those of the ehb star ) in a temporal interval of only 3 days , and they are in rough antiphase with the variations of the h@xmath33 line on the same spectra .
unfortunately , the latter are not significant when compared to the large errors @xcite , and the measurements are too few to independently derive the periodicity of the ms star . however , when we phase the rv of the ms star with the solution previously obtained from the high - resolution spectra , we find that the best - fit sinusoidal curve is in antiphase with that of the ehb star , as shown in fig.[f_phase ] .
thus , the behavior of the mgib triplet demonstrates that the ms star is the close companion of the hotter component . an ehb star with a close ms companion
should show light variability @xcite .
stars deformed by tidal forces show ellipsoidal variations faster than the orbital period by a factor of two , with maxima at the phases of maximum and minimum rv . on the other hand , a star in rotationally locked orbit exhibits reflection effects on the heated hemisphere , synchronous with the orbital period and with maximum brightness at phase @xmath46=0.25 before the minimum rv .
more than 800 frames in the @xmath21 band were acquired with imacs at the 6.5 m baade telescope at las campanas observatory , during three consecutive nights ( 2007 july 7 to 9 , * ? ? ? * ) .
the instrument was used at f/4.3 and binned 2@xmath472 on the chip , for a resulting scale of 0.22@xmath23 per pixel .
exposure times varied between 30s and 40s .
the weather was good during the first and third night , with average seeing of 0.7@xmath23 and 1.1@xmath23 , respectively .
the second night was affected by strong wind , and the seeing varied between 1@xmath23 and 1.8@xmath23 .
we performed the photometry using standard image subtraction routines as implemented in isis2 @xcite .
we first constructed a high s / n ( reference ) image from the 50 best - seeing images .
this was convolved with a kernel to match the seeing of each input frame , with a background variable at the second order , and subtracted to it after astrometric alignment .
aperture photometry was then performed on the difference image by means of our own customized software .
we masked out bad or saturated pixels , and all stars within 20 pixels from such bad features were not further considered .
we fixed the aperture radius to the full - width - at - half - maximum of the input image , and the inner and outer radius of the sky annulus to 20 and 35 pixels , respectively .
we then subtracted from the light curves the median photometric zero point between each frame and the first one of the series .
we considered only the brightest stars in its calculation , and we applied a 3@xmath20-clipping algorithm to exclude outliers .
we eventually removed a tiny residual trend , of the order of a few mmag , forcing the average instrumental magnitude of 43 stars within 1@xmath38 from the target to be constant among the frames .
we also applied a 3@xmath20-clipping selection to clean the light curve from outliers , and excluded the data points affected by large errors .
we thus ended up with 585 measurements spanning a temporal interval of about 2.5 days .
the resulting periodogram is shown in fig.[f_imacs ] .
it is characterized by wide and poorly - significant peaks , as a consequence of the poor temporal sampling . however , the two dominant peaks match extremely well the orbital period and its first overtone ( [email protected] and 0.81 days ) .
the corresponding light variation is tiny ( of the order of 2 mmag peak - to - peak ) , a fact is consistent with a high inclination of the orbital plane along the line of sight .
the position of the maximum at phase @xmath48=0.75 , as shown in fig .
[ f_phase ] , excludes ellipsoidal variations with [email protected] days , and it suggests that the dominant variability is due to reflection effects ( [email protected] days ) on the surface of the ehb star .
this fact is new , because in known systems the ms object is too cold , or too far away , to noticeably heat the ehb companion .
clearly , better time series are required to study the irradiation from this star in more detail .
our study reveals that the close companion of the ehb star m5865 is a ms star of mass m@xmath49m@xmath2 .
this discovery does not increase the low estimate of the close binary fraction in the cluster ( @xmath04% , * ? ? ?
* ) , because this binary had already been identified . however , it marks a striking difference with field studies , where close binaries are very common but such a system has never been observed .
the theoretical models predict the formation of few such objects , resulting from a first common - envelope ( ce ) stage , although low - mass companions are highly preferred @xcite . in an old stellar population , these systems could be more common than ehbs with a compact companion , products of a second - ce phase @xcite . hence , it could not be completely surprising that the first ehb close binary studied in a gc is such a unique pair .
nevertheless , their absence in the field is harder to explain because , while @xcite shows that first - ce products could be a factor of two less frequent at higher metallicity , @xcite finds no clear decrease of the first - ce channel efficiency at younger ages .
m5865-like systems have so far been overlooked in the literature due to the lack of observational counterparts , and theoretical models should study their formation mechanism in more details .
it is also possible that they have remained undetected in the field due to selection effects such as the known `` gk '' bias @xcite . in this case
, the discovery of m5865 would indicate that the studies of field ehb stars are incomplete , because a class of systems has so far been neglected .
cmb , sv , and mc acknowledge support from proyectos fondecyt regular 1150060 , 1130721 , and 1141141 .
mc acknowledges support by the ministry of economy , development , and tourism s programa iniciativa cientfica milenio , through grant ic210009 , awarded to the millenium institute of astrophysics ( mas ) by proyecto basal pfb-06/2007 .
mm acknowledges support from fct through the grant and sfrh / bdp/71230/2010 .
alard , c. , & lupton , r. h. 1998 , apj , 503 , 325 alard , c. 2000 , a&as , 144 , 363 bloomfield , p. 1976
, wiley series in probability and mathematical statistics , new york : wiley ( 1976 ) cardelli , j. a. , clayton , g. c. , mathis , j. s. 1989 , apj , 345 , 245 catelan , m. , prieto , g. e. , zoccali , m. , et al .
2008 , asp conf .
series , 392 , 347 chung , c. , yoon , s .- j .
, & lee , y .- w . 2011 , apj , 740 , 45 dantona , f. , caloi , v. , montalbn , j. , ventura , p. , & gratton , r. 2002 , a&a 395 , 69 dalessandro , e. , salaris , m. , ferraro , f. r. , et al . 2011 ,
mnras , 410 , 694 ferraz - mello , s. 1981 , aj , 86 , 619 faulkner , j. 1972 , a&a , 173 , 401 geier , s. , classen , l. , brnner , p. , et al .
2012 , asp conf .
series , 452 , 153 gray , r. o. , corbally , c. j. 1994 , aj , 107 , 742 greggio , l. , & renzini , a. 1990 , apj , 364 , 35 han , z. , podsiadlowski , p. , maxted , p. f. l. , marsh , t. r. , & ivanova , n. 2002 , mnras , 336 , 449 han , z. , podsiadlowski , p. , maxted , p. f. l. , & marsh , t. r. 2003 , mnras , 341 , 669 han , z. , podsiadlowski , p. , & lynas - gray , a. e. 2007 , mnras , 380 , 1098 han , z. 2008 , a&a , 484 , l31 harris , a. w. , young , j. w. , bowell , e. , et al .
1989 , icarus , 77 , 171 harris , w. e. 1996 , aj , 112 , 1487 , dec 2010 web version heber , u. 1986 , a&a , 155 , 33 heber , u. , moehler , s. , napiwotzki , r. , thejll , p. , & green , e. m. 2002 , a&a , 383 , 938 heber , u. , drechsel , h. , stensen , r. , et al .
2004 , a&a , 420 , 251 horne , k. 1986 , pasp , 98 , 609 kawka , a. , vennes , s. , otoole , s. , et al .
2015 , mnras , 450 , 3514 kurucz , r. l. 1993 , cd - rom 13 , 18 http://kurucz.harvard.edu lomb , n. r. 1976 , ap&ss , 39 , 447 maxted , p. f. l. , heber , u. , marsh , t. r. , & north , r. c. 2011 , mnras , 326 , 1391 milone , a. p. , piotto , g. , king , i. r. , et al .
2010 , apj , 709 , 1183 momany , y. , piotto , g. , recio - blanco , a. , et al .
2002 , apj , 576 , l65 moni bidin , c. , moehler , s. , piotto , g. , et al .
2006 , a&a , 451 , 499 moni bidin , c. , moehler , s. , piotto , g. , momany , y. , & recio - blanco , a. 2007 , a&a , 474 , 505 moni bidin , c. , catelan , m. , & altmann , m. 2008 a&a , 480 , l1 moni bidin , c. , moehler , s. , piotto , g. , momany , y. , & recio - blanco , a. 2009 , a&a , 498 , 737 moni bidin , c. , & piotto , g. 2010 , ap&ss , 329 , 19 moni bidin , c. , villanova , s. , piotto , g. , & momany , y. 2011 , a&a , 528 , a127 morales - rueda , l. , maxted , p. f. l. , marsh , t. r. , north , r. c. , & heber , u. 2003 , mnras , 338 , 752 munari , u. , sordo , r. , castelli , f. , & zwitter , t. 2005 , a&a , 442 , 1127 stensen , r. h. 2006 , balt .
astr . , 15 , 85 scargle , j. d. 1982 , apj , 263 , 835 schwarzenberg - czerny , a. 1996 , apj , 460 , l107 spada , f.,demarque , p. , kim , y .- c .
, & sills , a. 2013 , apj , 776 , 87 tonry , j. , & davis , m. 1979 , aj , 84 , 1511 vandenberg , d. a. , brogaard , k. , leaman , r. , & casagrande , l. 2013 , apj , 775 , 134 vos , j. , stensen , r. h. , nmeth , p. , et al .
2013 , a&a , 559 , 54 wade , r. , barlow , b. , liss , s. , & stark , m. 2014 , asp .
series 481 , 311 yungelson , l. r. , & tutukov , a. v. 2005 , astron . | dynamical interactions in binary systems are thought to play a major role in the formation of extreme horizontal branch stars ( ehbs ) in the galactic field . however , it is still unclear if the same mechanisms are at work in globular clusters , where ehbs are predominantly single stars . here
we report on the discovery of a unique close binary system ( period @xmath01.61 days ) in the globular cluster ngc6752 , comprising an ehb and a main - sequence companion of [email protected] m@xmath2 .
such a system has no counterpart among nearly two hundred known ehb binaries in the galactic field .
its discovery suggests that either field studies are incomplete , missing this type of systems possibly because of selection effects , or that a particular ehb formation mechanism is active in clusters but not in the field . |
the standard model ( sm ) of particle physics has been the most successful phenomenological theory specially after the discovery of its last missing piece , the higgs boson at the large hadron collider ( lhc ) back in 2012 with subsequent null results for beyond standard model ( bsm ) searches . however , the sm fails to address several observed phenomena as well as theoretical questions .
for example , it fails to explain the sub - ev neutrino mass @xcite , the origin of parity violation in weak interactions and the origin of three fermion families .
the first two questions can be naturally addressed within the framework of the left - right symmetric model ( lrsm ) @xcite , one of the most widely studied bsm frameworks .
these models not only explain tiny neutrino masses naturally through seesaw mechanism but also give rise to an effective parity violating sm at low energy through spontaneous breaking of a parity preserving symmetry at high scale .
the conventional lrsm based on the gauge group @xmath3 can be enhanced to a more general lrsm based on the gauge group @xmath4 ( or in short @xmath5 ) .
the advantage of such an up gradation of the gauge symmetry is the ability of the latter in providing an explanation to the origin of three fermion families of the sm in addition to having other generic features of the lrsm .
in such a model , the number of three fermion generations is no longer a choice , but a necessity in order to cancel chiral anomalies . in such models , where the usual lepton and quark representations are enlarged from a fundamental of @xmath6 in the sm to a fundamental of @xmath7
, the number of generations must be equal to the number of colors in order to cancel the anomalies @xcite .
this is in contrast with the sm or the usual lrsm where the gauge anomalies are canceled within each fermion generation separately .
one can also build such a sequential @xmath5 model by including additional chiral fermions .
but since such a model does not explain the origin of three families from the anomaly cancellation point of view and contain non - minimal chiral fermion content , we stick to discussing a special type of non - sequential @xmath5 model here .
there have been a few works @xcite recently done within the framework of such @xmath5 models with different motivations . particularly from the origin of neutrino mass point of view , the work @xcite considered a scalar sector comprising of bitriplets plus sextets which gives rise to tiny neutrino masses through canonical type i @xcite and type ii @xcite seesaw .
another recent work @xcite studied a specific @xmath5 model with bitriplet and triplet scalar fields that can explain tiny neutrino masses through inverse @xcite and linear seesaw mechanism @xcite .
the earlier work @xcite considered effective higher dimensional operators to explain fermion masses in @xmath5 models while the recent work @xcite studied the model and several of its variants from lhc phenomenology point of view .
here , we simply consider another possible way of generating fermion masses in @xmath5 models through the universal seesaw mechanism @xcite where all fermions acquire their masses through a common seesaw mechanism . incorporating additional vector like fermion pairs corresponding to each fermion generation , we show that the correct fermion mass spectrum can be generated in such a model with a scalar sector where all of them transform as fundamentals under @xmath8 without the need of bi - fundamental and sextet scalars shown in @xcite for the implementation of different seesaw mechanism for neutrino masses .
we also discuss the possibilities of light neutral fermions apart from sub - ev active neutrinos , their role in neutrinoless double beta decay @xmath9 and different possible symmetry breaking chains of the gauge symmetry @xmath4 to that of the sm .
we show that for tev scale @xmath2 gauge bosons , the right handed neutrinos are constrained to lie around the kev mass regime having interesting consequences for @xmath10 .
we find that although the pure heavy neutrino contribution to @xmath10 remains suppressed compared to the one from light neutrinos , the heavy - light neutrino mixing which can be quite large in this model without any fine - tuning , gives a large contribution to @xmath10 keeping it within experimental reach .
this letter is organized as follows . in section [
sec : model ] we briefly discuss the model with the details of the particle spectrum , fermion masses via universal seesaw and gauge boson masses . in sections [ sec:0nubb ] and [ sec:0nubb2 ]
we discuss the contributions to @xmath10 from purely light ( heavy ) neutrinos and heavy - light neutrino mixing respectively .
finally we discuss about different possible symmetry breaking chains in section [ sec : symbreak ] and then conclude in section [ sec : conclude ] .
the usual fermions transform under @xmath0 as @xmath11 q_{m l } & = \begin{pmatrix } d_{\alpha l } \\
u_{\alpha l } \\ j^{-q-1/3}_{\alpha
l } \end{pmatrix}\ , , & q_{m r } = \begin{pmatrix } d_{\alpha r } \\
u_{\alpha r } \\
j^{-q-1/3}_{\alpha r } \end{pmatrix}\ , , \nonumber \\[1 mm ] q_{3
l } & = \begin{pmatrix } u_{3 l } \\
d_{3 l } \\
j^{q+2/3}_{3 l } \end{pmatrix}\ , , & q_{3 r } = \begin{pmatrix } u_{3 r } \\
u_{3 r } \\
j^{q+2/3}_{3 r } \end{pmatrix } \,.\end{aligned}\ ] ] with a=1,2,3 whereas m = 1,2 .
the transformation of the fields under the gauge symmetry @xmath0 are given in table [ table1 ] .particle content of the model [ cols="^,^",options="header " , ] [ table1 ] assuming @xmath12 , if the neutral components of the scalar fields acquire their vacuum expectation value ( vev ) as @xmath13 @xmath14 the yukawa lagrangian can be written as @xmath15_{ma } \left ( \overline{q}_{ml } \phi_l d_{ar } + \overline{q}_{mr } \phi_r d_{al } \right ) - \left[m_d\right]_{ab } \overline{d}_{al } d_{br } \nonumber \\ & \hspace*{-0.5cm}\quad - \left[y_u\right]_{ma } \left ( \overline{q}_{ml } \chi_l u_{ar } + \overline{q}_{mr } \chi_r u_{al } \right ) - \left[m_u\right]_{ab } \overline{u}_{al } u_{br } \nonumber \\ & \hspace*{-0.5cm}\quad - \left[y^\prime_d\right]_{3a } \left ( \overline{q}_{3l } \chi^{*}_l d_{ar } + \overline{q}_{3r } \chi^{*}_r d_{al } \right ) \nonumber \\ & \hspace*{-0.5cm}\quad - \left[y^\prime_u\right]_{3a } \left ( \overline{q}_{3l } \phi^{*}_l u_{ar } + \overline{q}_{3r } \phi^{*}_r u_{al } \right ) \nonumber \\ & \hspace*{-0.5cm}\quad - \left[y_e\right]_{ab } \left ( \overline{\psi}_{al } \chi^{*}_l e_{br } + \overline{\psi}_{mr } \chi^{*}_r e_{bl } \right ) - \left[m_e\right]_{ab } \overline{e}_{al } e_{br } \nonumber \\ & \hspace*{-0.5cm}\quad - \left[y_n\right]_{ab } \left ( \overline{\psi}_{al } \phi^{*}_l n_{br } + \overline{\psi}_{mr } \phi^{*}_r n_{bl } \right ) - \left[m_{lr}\right]_{ab } \overline{n}_{al } n_{br } \nonumber \\ & \hspace*{-0.5cm}\quad - \left[m_{ll}\right]_{ab } n_{al } n_{bl } - \left[m_{rr}\right]_{ab } n_{ar } n_{br } + \text{h.c . }
\end{aligned}\ ] ] model . ] after integrating out the heavy fermions , we can write down the effective yukawa terms for charged fermions of the standard model as follows @xmath16 similarly , the heavy neutral singlet fields @xmath17 can be integrated out to generate the effective mass matrix of neutrinos @xmath18 which contains a dirac mass term and two majorana mass terms . the effective dirac mass as well as majorana mass terms are given by @xmath19 there are additional neutrino leptons @xmath20 which acquire dirac and majorana masses similar to @xmath21 shown above .
they are given by @xmath22 model . ] model . ] the origin of the dirac masses can be understood from the mass diagrams shown in figure [ feyn1 ] whereas the majorana mass diagrams are given in figure [ feyn2 ] . apart from these ,
the light neutrinos also receive non - leading contribution to their masses from the diagram shown in figure [ feyn3 ] .
the contribution of this diagram can be written as @xmath23 here we are assuming equality of left and right sector yukawa couplings as well as masses @xmath24 .
the approximate scale of light neutrino mass matrix @xmath25 has to be less than @xmath26 ev which puts limit on model parameters @xmath27 .
for example , if @xmath28 then @xmath29gev to keep @xmath30 ev .
further lowering the scale of @xmath31 will involve more fine - tuning in the yukawa coupling @xmath32 . in the limit of tiny @xmath33 , the light neutrino mass solely originates from @xmath25 given in equation and the mixing between heavy and light neutrinos
can be neglected .
in such a case , the light neutral lepton mass matrix can be written in the basis @xmath34 as @xmath35 as a result of this particular structure of the mass matrix , both the mass eigenvalues for left - handed and right - handed neutrinos are proportional to each other .
the two mixing matrices are related as @xmath36 where @xmath37 is the pontecorvo - maki - nakagawa - sakata ( pmns ) leptonic mixing matrix . for a representative sets of input model parameters like @xmath38gev , @xmath39tev , and @xmath40ev ,
the right - handed neutrino masses lie in the range of kev scale .
such light kev scale right handed neutrinos can also have very interesting implications for cosmology @xcite .
we leave such a detailed study of this model from cosmology point of view to an upcoming work .
the relevant kinetic terms leading to gauge boson masses are given by @xmath41 where the factor @xmath42 is defined as @xmath43 using respective vev s for scalar fields shown in equation we can derive the gauge boson masses for the present model . in the gauge boson spectrum , we have * one massless photon @xmath44 , * four neutral gauge bosons @xmath45 , @xmath46 , * four charged gauge bosons @xmath47 , * four gauge bosons with charge @xmath48 , @xmath49 , * four gauge bosons with charge @xmath50 , @xmath51 .
we find that the light sub - ev scale left - handed neutrinos ( @xmath52 ) and the heavy right - handed neutrinos ( @xmath53 ) with kev scale masses can give sizable contributions to neutrinoless double beta decay .
since the bitriplet scalar is absent in the present left - right symmetric @xmath5 model , there are no dirac mass term for light neutrinos at tree level .
however one can eventually generate majorana masses for @xmath52 and @xmath54 through universal seesaw , see fig.[feyn2 ] .
such majorana nature of neutrinos violate lepton number by two units and thus , contributes to @xmath1 decay .
the feynman diagram for @xmath1 decay is depicted in fig.[feyn4 ] due to exchange of left - handed as well as right - handed neutrinos .
the corresponding feynman amplitudes due to exchange of left - handed and right - handed neutrinos are given by @xmath55 decay diagrams due to exchange of light left - handed neutrinos with left - handed charged currents and right - handed neutrinos with right - handed charged currents.,scaledwidth=51.0% ] the inverse half - life for a given isotope for @xmath1 decay- due to exchange of left - handed light neutrinos via left - handed currents , and right - handed neutrinos via right - handed currents is given by @xmath56^{-1 } \!=\ !
g_{01 } \left(|\mathcal{m}_\nu \eta_{\nu}|^2 + |\mathcal{m}_n \eta_n |^2 \right),\end{aligned}\ ] ] where @xmath57 is @xmath1 phase space factor , @xmath58 correspond to the nuclear matrix elements ( nme ) and @xmath59 is the corresponding dimensionless particle physics parameter . since we have @xmath60mev masses of right - handed neutrinos in the present model and satisfying @xmath61 where @xmath62 being the neutrino virtually momentum around @xmath63mev , the nmes for right - handed neutrinos and left - handed neutrinos are same i.e , @xmath64 . ' '' '' + [ 10pt ] @xmath65 & @xmath66 ' '' ''
+ [ 10pt ] @xmath67 & @xmath68 ' '' '' + [ 10pt ] the standard mechanism for neutrinoless double beta decay due to exchange of light left - handed neutrinos via left - handed currents gives dimensionless particle physics parameter as , @xmath69 here , @xmath70 is the electron mass .
the effective mass parameter for standard mechanism is explicitly given by @xmath71 in the present left - right symmetric @xmath5 model , we found that the right - handed neutrino mass lies around a few kev ( for tev scale @xmath72 ) which is much less than its momentum , @xmath73 . under this condition
, the propagator simplifies in a similar way way as for the light neutrino exchange , @xmath74 this results dimensionless particle physics parameter @xmath75 due to exchange of right - handed neutrinos via right - handed currents as , @xmath76 where the proportionality relation between @xmath75 and @xmath77 appears at the last step due to the proportionality between heavy and light neutrino mass matrix discussed above .
after a little simplification , the effective mass parameter due to exchange of right - handed neutrinos can be expressed as , @xmath78 it is clear from eq.([eq : numass_rel ] ) that both light and heavy neutrino mass eigenvalues are proportional to each other as @xmath79 thus , one can express @xmath72 mass as @xmath80 as a result , the effective majorana mass parameter with @xmath81 and @xmath82 is modified to @xmath83 comparing this with the light neutrino contribution @xmath84 , it is straightforward to estimate that the heavy neutrino contribution is suppressed by a factor of @xmath85 compared to the light neutrino contribution .
the total contribution to inverse half - life for neutrinoless double beta decay for a given isotope in the present left - right symmetric @xmath5 model due to exchange of left - handed as well as right - handed neutrinos is given by @xmath56^{-1 } \!=\ !
\left| \frac{\mathcal{m}_\nu}{m_e}\right|^2 @xmath86 for normal hierarchical ( nh ) pattern of light neutrinos we consider the following mass structures for left - handed and right - handed neutrinos , @xmath87 where @xmath88 is fixed around few kev range , as a result of choosing the @xmath72 mass scale at a few tev . the analytic form for effective mass parameters due to exchange of right - handed neutrinos
is given by @xmath89 similarly for inverse hierarchical ( ih ) pattern of light neutrinos , the masses for light left - handed and right - handed neutrinos are fixed as , @xmath90 where we fixed the heaviest right - handed neutrino mass @xmath91 around few kev .
the effective mass parameters due to exchange of right - handed neutrinos is given by @xmath92 we have generated effective mass ( left - panel ) and half - life ( right - pannel ) with the variation of lightest neutrino mass , @xmath93 for nh and @xmath94 for ih as shown in fig.[plot1 ] .
the green and red lines are for nh and ih pattern of light neutrinos applicable for both left- and right - panel .
the horizontal dashed line is for bound from @xmath1 experiments while the vertical dashed lines along with brown shaded regions are excluded from planck limit .
the present bound on half - life is @xmath95 yr at 90% c.l . from kamland - zen @xcite .
it is found that the new physics contributions to @xmath1 is very much suppressed and the standard mechanism due to exchange of light neutrinos are dominant .
the total contribution to @xmath1 is similar to the light neutrino contribution , saturating the experimental bound only near the quasi - degenerate regime , clearly visible from the plots shown in figure [ plot1 ] .
though the purely heavy neutrino contribution to @xmath10 remains suppressed , there can be sizable contributions from heavy - light neutrino mixing diagrams .
this heavy - light neutrino mixing can also contribute to light neutrino masses through a type i seesaw formula which was ignored in the above discussion for simplicity .
such an assumption is valid for negligible heavy - light neutrino mixing . however ,
if we go beyond this simple assumption or equivalently consider @xmath96 , then the dirac as well as majorana mass matrices for @xmath52 and @xmath97 can be written as @xmath98 thus , the neutral lepton mass matrix in the basis @xmath99 as @xmath100 in the limit @xmath101 , the type - i seesaw contribution to light neutrino mass is given by @xmath102 and the light - heavy neutrino mixing is proportional to @xmath103 . with @xmath104gev and @xmath105 around few tev , we find that light - heavy neutrino mixing is large of the order of @xmath106 .
this large value of light - heavy neutrino mixing where heavy neutrinos are fixed at few kev scale , can contribute to @xmath1 significantly .
decay diagrams due to purely left - handed charge current interaction and with the exchange of @xmath52 and @xmath107.,scaledwidth=51.0% ] the new physics contributions to @xmath1 decay arising from purely left - handed currents due to exchange of kev scale right - handed neutrinos results in the following effective mass parameter , @xmath108 here @xmath109 is in kev range and @xmath110 is the light - heavy neutrino mixing . since @xmath111 , the effective mass parameter for @xmath1 is estimated to be @xmath112ev which is of the order of @xmath113ev , saturating the kamland - zen @xcite bound .
decay diagrams due to mediation of one @xmath114 and one @xmath72 which also involves light - heavy neutrino mixing.,scaledwidth=51.0% ] the effective mass parameters due to the @xmath116 mediated diagrams ( known as @xmath117 diagrams ) shown in figure [ feyn4 ] are given by @xmath118 with @xmath119gev , @xmath120tev and @xmath121 , the effective mass parameters due to these @xmath117 diagrams are found to be around sub - ev which can translated to a life - time of @xmath122 yrs .
this value is very close to experimental bound for @xmath123 isotope .
it is interesting that , such large observable heavy - light neutrino mixing arises naturally in the model without any fine - tuning of the yukawa couplings involved .
although we consider one contribution at a time in the above discussions , in general one has to include all the contributions and at the same time keeping the light neutrino mass and mixing in the allowed range .
we intend to perform a detailed study , considering the most general neutrino mass formula @xmath124 to an upcoming work .
depending on the scale of different vev s mentioned above , the gauge symmetry of the model can be broken down to that of the standard model through several possible symmetry breaking chains .
they are summarized pictorially in figure [ fig2 ] .
the relevant scalar potential for the model can be written as @xmath125 in the scalar potential written above , the discrete left - right symmetry is assumed which ensures the equality of left and right sector couplings . however , as shown in earlier works @xcite in the context of usual lrsm with universal seesaw that the scalar potential of such a model with exact discrete left - right symmetry is too restrictive and gives to either parity preserving @xmath126 solution or a solution with @xmath127 at tree level . while the first one is not phenomenologically acceptable the latter solution can be acceptable if a non - zero vev @xmath128 can be generated through radiative corrections @xcite .
while it may naturally explain the smallness of @xmath129 compared to @xmath130 , it will constrain the parameter space significantly @xcite . another way of achieving a parity breaking vacuum
is to consider softly broken discrete left - right symmetry by considering different mass terms for the left and right sector scalars @xcite . as it was pointed out by the authors of @xcite , such a model which respects the discrete left - right symmetry everywhere except in the scalar mass terms ,
preserve the _ naturalness _ of the left - right symmetry in spite of radiative corrections .
another interesting way is to achieve parity breaking vacuum is to decouple the scale of parity breaking and gauge symmetry breaking by introducing a parity odd singlet scalar @xcite .
while we do not perform a detailed analysis of different possible symmetry breaking chains and their constraints on the parameter space of the model , we outline them pictorially in the cartoon shown in figure [ fig2 ] . as can be seen from figure [ fig2 ] , there are seven different symmetry breaking chains through which the gauge symmetry of the model @xmath4 can be broken down to that of the sm as summarized below .
* one step breaking : the vev s
satisfy @xmath131 in this case . *
two step breaking : the vev s satisfy either @xmath132 or @xmath133 or @xmath134 .
the usual @xmath135 model presumes an intermediate stage in the first case while the usual lrsm or @xmath136 symmetry arises an intermediate symmetry in the second case . in the third case
, the @xmath137 symmetry assumes an intermediate stage .
the phenomenology of such asymmetric lrsm was discussed recently by @xcite . *
three step breaking : this is possible in three different ways when the vev s
satisfy @xmath138 or @xmath139 or @xmath140 .
one can have both the usual lrsm or @xmath136 or asymmetric lrsm ( @xmath141 or @xmath137 ) or the usual @xmath135 model as an intermediate stage .
gauge symmetry to the standard model gauge symmetry @xmath142 . ]
all these different symmetry breaking chains can not only provide a different phase transition history in cosmology but also give rise to different particle spectra including gauge bosons as well as neutral fermions which could be tested in different experiments .
we have demonstrated a class of left - right symmetric model with extended gauge group @xmath0 with a universal seesaw mechanism for fermion masses and mixing and the implications for neutrinoless double beta ( @xmath1 ) decay .
the novel feature of the model is that masses and mixing for left - handed and right - handed neutrinos are exactly determined by oscillation parameters and lightest neutrino mass .
this forces the heavy neutrino masses to lie in kev regime if the @xmath72 mass is fixed at a few tev .
we show that for such a case , the heavy neutrino contribution to @xmath143 remains suppressed compared to the usual light neutrino contribution . we also show that for such a tev scale model , the heavy - light neutrino mixing can be quite large and can contribute substantially to @xmath10 diagrams , keeping it within experimental reach . in the end
we have discussed the scalar potential and possible symmetry breaking patterns that can be allowed for spontaneous breaking of the 3331 gauge symmetry to that of the standard model .
the authors would like to thank the organizers of the _ indo - us bilateral workshop on understanding the origin of the invisible sector : from neutrinos to dark matter and dark energy _ during november 16 - 18 , 2016 at the school of physics , university of hyderabad , india where this work was initiated .
17 s. fukuda et al .
( super - kamiokande ) , phys .
lett . * 86 * , 5656 ( 2001 ) , hep - ex/0103033 ; q. r. ahmad et al .
( sno ) , phys .
* 89 * , 011301 ( 2002 ) , nucl - ex/0204008 ; phys
lett . * 89 * , 011302 ( 2002 ) , nucl - ex/0204009 ; j. n. bahcall and c. pena - garay , new j. phys .
* 6 * , 63 ( 2004 ) , hep - ph/0404061 ; k. nakamura et al .
, j. phys . * g37 * , 075021 ( 2010 ) .
k. abe et al .
[ t2k collaboration ] , phys .
107 * , 041801 ( 2011 ) , [ arxiv:1106.2822 [ hep - ex ] ] .
y. abe et al . , phys .
108 * , 131801 ( 2012 ) , [ arxiv:1112.6353 [ hep - ex ] ] .
f. p. an et al .
[ daya - bay collaboration ] , phys .
lett . * 108 * , 171803 ( 2012 ) , [ arxiv:1203.1669 [ hep - ex ] ] .
j. k. ahn et al .
[ reno collaboration ] , phys .
lett . * 108 * , 191802 ( 2012 ) , [ arxiv:1204.0626][hep - ex ] ] .
p. adamson et al .
( minos ) , phys.rev.lett . * 110 * , 171801 ( 2013 ) .
j. c. pati and a. salam , phys .
* d10 * , 275 ( 1974 ) ; g. senjanovic and r. n. mohapatra , phys .
* d12 * , 1502 ( 1975 ) ; r. n. mohapatra and r. e. marshak , phys .
* 44 * , 1316 ( 1980 ) ; j. f. gunion , j. grifols , a. mendez , b. kayser and f. i. olness , phys .
* d40 * , 1546 ( 1989 ) ; n. g. deshpande , j. f. gunion , b. kayser and f. i. olness , phys .
* d44 * , 837 ( 1991 ) .
r. n. mohapatra and j. c. pati , phys .
* d11 * , 2558 ( 1975 ) ; m. singer , j. w. f. valle and j. schechter , phys .
* d22 * , 738 ( 1980 ) ; f. pisano and v. pleitez , phys . rev .
* d46 * , 410 ( 1992 ) ; p. h. frampton , phys .
lett . * 69 * , 2889 ( 1992 ) ; r. foot , o. f. hernandez , f. pisano and v. pleitez , phys . rev . *
d47 * , 4158 ( 1993 ) .
m. reig , j. w. f. valle and c. a. vaquera - araujo , arxiv:1611.02066 [ hep - ph ] .
e. t. franco and v. pleitez , arxiv:1611.06568 [ hep - ph ] .
m. reig , j. w. f. valle and c. a. vaquera - araujo , arxiv:1611.04571 [ hep - ph ] .
a. g. dias , c. a. de s.pires and p.
s. rodrigues da silva , phys .
d * 82 * , 035013 ( 2010 ) ; [ arxiv:1003.3260 [ hep - ph ] ] .
d. t. huong and p. v. dong , phys . rev . *
d93 * , 095019 ( 2016 ) .
p. minkowski , phys .
b67 * , 421 ( 1977 ) ; m. gell - mann , p. ramond , and r. slansky ( 1980 ) , print-80 - 0576 ( cern ) ; t. yanagida ( 1979 ) , in proceedings of the workshop on the baryon number of the universe and unified theories , tsukuba , japan , 13 - 14 feb 1979 ; r. n. mohapatra and g. senjanovic , phys .
lett * 44 * , 912 ( 1980 ) ; j. schechter and j. w. f. valle , phys . rev . * d22 * , 2227 ( 1980 ) .
r. n. mohapatra and g. senjanovic , phys . rev . * d23 * , 165 ( 1981 ) .
g. lazarides , q. shafi and c wetterich , nucl .
b181 * , 287 ( 1981 ) ; c. wetterich , nucl .
* b187 * , 343 ( 1981 ) ; j. schechter and j. w. f. valle , phys .
* d25 * , 774 ( 1982 ) ; b. brahmachari and r. n. mohapatra , phys . rev . * d58 * , 015001 ( 1998 ) ; r. n. mohapatra , nucl .
* 138 * , 257 ( 2005 ) ; s. antusch and s. f. king , phys .
b597 * , ( 2 ) , 199 ( 2004 ) .
r. n. mohapatra and j. w. f. valle , phys . rev .
* d34 * , 1642 ( 1986 ) ; m. gonzalez - garcia and j. w. f. valle , phys
* b216 * , 360 ( 1989 ) ; m. e. catano , r. martinez and f. ochoa , phys . rev . *
d86 * , 073015 ( 2012 ) .
e. k. akhmedov , m. lindner , e. schnapka and j. w. f. valle , phys . lett . *
b368 * , 270 ( 1996 ) ; e. k. akhmedov , m. lindner , e. schnapka and j. w. f. valle , phys . rev . * d53 * , 2752 ( 1996 ) ; m. malinsky , j. romao and j. w. f. valle , phys .
rev . lett . * 95 * , 161801 ( 2005 ) .
b. brahmachari , e. ma and u. sarkar , phys .
. lett . * 91 * , 011801 ( 2003 ) .
a. davidson and k. c. wali , phys .
59 * , 393 ( 1987 ) ; r. n. mohapatra and y. zhang , jhep * 1406 * , 072 ( 2014 ) . k. s. babu and r. n. mohapatra , phys .
* 62 * , 1079 ( 1989 ) ; p. -h .
gu and m. lindner , phys . lett .
* b698 * , 40 ( 2011 ) .
s. patra , phys .
d * 87 * , no . 1
, 015002 ( 2013 ) doi:10.1103/physrevd.87.015002 [ arxiv:1212.0612 [ hep - ph ] ] .
s. b. dev , r. n. mohapatra and y. zhang , jhep * 1602 * , 186 ( 2016 ) doi:10.1007/jhep02(2016)186 [ arxiv:1512.08507 [ hep - ph ] ] . f. f. deppisch , c. hati , s. patra , p. pritimita and u. sarkar , phys .
b * 757 * , 223 ( 2016 ) doi:10.1016/j.physletb.2016.03.081 [ arxiv:1601.00952 [ hep - ph ] ] . f. f. deppisch , c. hati , s. patra , p. pritimita and u. sarkar , arxiv:1701.02107 [ hep - ph ] . m. c. gonzalez - garcia , m. maltoni , j. salvado and t. schwetz , jhep * 1212 * , 123 ( 2012 ) [ arxiv:1209.3023 [ hep - ph ] ] .
a. r. ade _ et al .
_ [ planck collaboration ] , astron .
astrophys . * 594 * , a13 ( 2016 ) [ arxiv:1502.01589 [ astro-ph.co ] ] .
s. mertens [ katrin collaboration ] , phys .
procedia * 61 * , 267 ( 2015 ) .
_ et al . _
[ kamland - zen collaboration ] , phys .
lett . *
117 * , no . 8 ,
082503 ( 2016 ) addendum : [ phys .
lett . * 117 * , no .
10 , 109903 ( 2016 ) ] doi:10.1103/physrevlett.117.109903 , 10.1103/physrevlett.117.082503 [ arxiv:1605.02889 [ hep - ex ] ] .
d. borah , phys . rev . *
d94 * , 075024 ( 2016 ) .
a. kobakhidze and a. spencer - smith , jhep * 1308 * , 036 ( 2013 ) .
d. chang , r. n. mohapatra and m. k. parida , phys .
* 52 * , 1072 ( 1984 ) . | we consider a class of left - right symmetric model with enlarged gauge group @xmath0 without having scalar bitriplet . in the absence of scalar bitriplet
, there is no dirac mass term for fermions including usual quarks and leptons .
we introduce new isosinglet vector - like fermions so that all the fermions get their masses through a universal seesaw mechanism .
we extend our discussion to neutrino mass and its implications in neutrinoless double beta decay ( @xmath1 ) .
we show that for tev scale @xmath2 gauge bosons , the heavy - light neutrino mixing contributes dominantly to @xmath1 that can be observed at ongoing experiments . towards the end
we also comment on different possible symmetry breaking patterns of this enlarged gauge symmetry to that of the standard model . |
in condensed matter physics , the task of obtaining different mechanical properties of materials , simulated atomistically with a large number of atoms under _ ab initio _ methods , is an almost prohibitive one , in terms of computational effort with the current computer architectures .
it might even at times be impossible .
because of this , producing a `` classical '' interatomic potential as a substitute for the genuine quantum - mechanical interaction of the particles is highly desirable .
the usual procedure is to fit some empirical interatomic potential function , depending on @xmath0 parameters , requiring either agreement with certain macroscopic properties ( structural , thermodynamical , etc . ) or simply agreement between the predicted and observed energies and atomic forces .
a standard algorithm based on force information is the force matching method @xcite . in this work we present a methodology for fitting interatomic potentials to _ ab initio _ data , using the particle swarm optimization ( pso ) algorithm @xcite .
the objective function to be minimized is the total prediction error in the energies for the configurations provided , thus the algorithm does not require any information besides the atomic positions for each configuration and their corresponding _ ab initio _ energies . in particular it does not require the atomic forces , as in other fitting procedures such as force matching methods .
we implemented two families of interatomic potentials , pair potentials and embedded atom potentials . among the former , we tested the well - known lennard - jones potential @xcite , given by @xmath1,\ ] ] and the 6-parameters `` generic '' potential as implemented in moldy @xcite , @xmath2 from the family of embedded atom potentials @xcite , having the general form @xmath3 we implemented the sutton - chen potential , where the pair functions and the embedding function are given by @xmath4
the particle swarm optimization ( pso ) algorithm is based on the idea of distributing the search procedure among a large number of `` agents '' , which act independently of each other .
each agent moves through the search space with a simple dynamics , reacting to fictitious forces drawing it towards its own _ current best _ solution and the _ global best _
solution for the whole swarm . in this way ,
when an agent finds a better solution than the current global best , it becomes the new global best and all the other agents react instantly , the swarm is directed towards the new solution .
for a set of @xmath5 particles represented by their positions @xmath6 , the velocity for the @xmath7-th particle and the @xmath8-th step is @xmath9 and the position is given by @xmath10 we employed the following choice of pso parameters : @xmath11=0.7 , @xmath12=1.4 and @xmath13=1.4 , after a few trial convergence runs .
for a potential function where we wish to find the parameters @xmath14 from a set of positions @xmath15 and energies @xmath16 satisfying the relation @xmath17 with @xmath18 we can define an objective function which is just the total prediction squared error , of the form @xmath19 and then for the set of parameters @xmath20 that correctly fit the potential @xmath21 we have @xmath22
. then the problem may be solved numerically with the pso algorithm minimising the function @xmath23 .
we have included some improvements on the pso implementation , particular to our problem .
for instance , we perturbed the swarm every time the procedure gets stuck in a minimum for @xmath24 steps ( @xmath24 proportional to the number of parameters @xmath25 in the potential , usually @xmath26 ) , completely randomizing their positions .
on the other hand , we exploit the fact that for several families of potentials there is a scale parameter for the interatomic distance , let us call it @xmath27 , such that the potential depends on @xmath28 only through @xmath29 .
this is the case for the @xmath27 parameter in the lennard - jones potentials , for the @xmath30 , @xmath31 , @xmath32 , @xmath33 and @xmath34 parameters in the generic potential from moldy , and also for the @xmath35 parameter in the sutton - chen variant of the embedded atom potentials .
this distance scale parameter can be constrained to be between the minimum observed distance and a multiple of this value ( typically 10 times ) , which considerably reduces the search space .
parallelization was achieved simply by distributing the pso particles evenly among the different processors using the message passing interface ( mpi ) framework , at each step sharing the global best between all processors .
in order to test the consistency of our procedure , we randomly generate @xcite a set of 20 configurations and we compute their energy according to the standard lennard - jones parameters for argon , @xmath36 = 0.0103048 ev and @xmath27 = 3.41 .
the resulting set has a standard deviation of energy of 0.41063 ev .
then , with the information of positions and energies ( in a parallel run using 64 cores and 500 pso particles ) , the time needed to find the minimum prediction error was 212.6 s. we can see that the algorithm converge quickly for each parameter , recovering their exact values at 1300 steps ( the prediction error reached is below 10@xmath37 mev / atom ) . coordinate for the case of a lennard - jones potential as a function of optimization step . ]
coordinate for the case of a lennard - jones potential as a function of optimization step . ] for the 6-parameters pair potential using the same set of positions and energies obtained for the previous lennard - jones test , the time needed to find the minimum prediction error was 3159.9 s , again using 64 cores and 500 pso particles . in this case
the error for the converged set of parameters falls below 8@xmath3810@xmath39 mev / atom at 9000 steps .
we repeated the same approach for the embedded atom potential , this time using the standard sutton - chen parameters for copper , @xmath36=0.0123820 ev , @xmath35=3.61 , @xmath5=9 , @xmath40=6 and @xmath30=39.432 .
we used 4 configurations as input , and we stopped the minimization procedure after 193015 steps ( execution time was 23 hours with 64 cores and 800 pso particles ) , when we reached a prediction error of about 0.8 mev / atom and the following fitted parameters : @xmath36=0.0145749 ev , @xmath35=3.5834 , @xmath5=8.82683 , @xmath40=5.67465 , and @xmath30=37.028 .
in order to test our procedure on a more realist scenario and assess the quality of the fitted potentials we performed _
ab initio _ microcanonical molecular dynamics simulations of copper at different temperatures ( covering its solid , liquid and superheated phases ) .
all molecular dynamics calculations were performed using density functional theory ( dft ) as implemented in vasp @xcite .
we used perdew - burke - ernzerhof ( pbe ) generalized gradient approximation ( gga ) pseudopotentials @xcite with an energy cutoff of 204.9 ev and @xmath8-point expansion around the @xmath41 point only . from these simulations , we generated 13229 different atomic configurations with their respective energies , mixed from solid , liquid and superheated state simulations . among them
we chose a subset of 30 with maximum standard deviation of the energy ( namely 0.24 ev / atom ) , in order to increase the transferability of the fitted potential .
these configurations were used as input to the fitting procedure .
we found the sutton - chen potential parameters presented in table [ tbl_cu ] , with a prediction error of 5.19 mev / atom .
.sutton - chen potential parameters for cu , fitted from ab initio data . [ cols="^,^,^,^,^,^",options="header " , ] we tested these parameters by performing classical molecular dynamics simulations using the lpmd @xcite code , with a 4x4x4 fcc simulation cell ( 256 atoms ) . fig .
[ fig_cu_gdr ] shows the radial distribution function @xmath42 produced by our fitted cu potential for liquid at @xmath43=1500 k. it reproduces exactly all features ( positions of minima and maxima , heights of the maxima ) found in a previous ab initio fitting study @xcite . for liquid copper at @xmath43=1500 k. ]
[ fig_cu_msd ] shows the mean square displacement for liquid at @xmath43=1500 k. from this we obtained a diffusion coefficient @xmath31=0.276924 @xmath44/ps , lower than the experimental value reported by meyer @xcite , 0.45 @xmath44/ps at @xmath43=1520 k. = 1500 k. ] the quality of the potential in reproducing thermal properties was assessed by computing the melting point , using the microcanonical z - method @xcite . in this method , for constant volume
the @xmath45 curve is drawn by performing different molecular dynamics simulations at different initial kinetic energies ( in every simulation the system starts with the ideal crystalline configuration ) .
the discontinuity in the isochore signals the melting point .
[ fig_cu_z ] shows the isochoric curve for different energies around the melting point , where the lowest point of the rightmost branch correspond to an upper estimate of the melting temperature @xmath46 , in our case approx .
1700 k ( the experimental value is @xmath46=1356.6 k ) .
the highest point is the critical superheating temperature @xmath47 , around @xmath48 2020 k. for comparison we also included the isochoric curve calculated with the potential parameters by sutton and chen , which gives @xmath46 around 2000 k for the same system size and number of simulation steps .
we have shown that it is possible to use a parallel algorithm based on particle swarm optimization to fit interatomic potentials to _ ab initio _ energies only .
our procedure has been tested by fitting both pair potentials and embedded atom potentials , up to a prediction error of the order of 1 mev / atom , using between 5 and 30 different configurations .
the implementation code is parallelized using message passing interface ( mpi ) libraries .
we demonstrated the capabilities of our method by fitting a set of sutton - chen parameters for copper using _ ab initio _ data from three thermodynamic phases .
this fitted potential is able to reproduce the radial distribution function , although it underestimates the diffusion coefficient for liquid copper at @xmath43=1500 k ( with respect to experimental data ) .
it also yields a better prediction of the melting point than the standard sutton - chen parameters . | we present a methodology for fitting interatomic potentials to ab initio data , using the particle swarm optimization ( pso ) algorithm , needing only a set of positions and energies as input .
the prediction error of energies associated with the fitted parameters can be close to 1 mev / atom or lower , for reference energies having a standard deviation of about 0.5 ev / atom .
we tested our method by fitting a sutton - chen potential for copper from _ ab initio _ data , which is able to recover structural and dynamical properties , and obtain a better agreement of the predicted melting point versus the experimental value , as compared to the prediction of the standard sutton - chen parameters . |
at center of mass energies above @xmath0 gev , total hadronic cross sections @xmath1 show a continuous rise which persists up to the highest energies explored so far and which is consistent with a gentle powerlike growth , @xmath2 however , at asymptotically high @xmath3 , hadronic cross sections satisfy the unitarity or froissart bound , @xmath4 where @xmath5 is the smallest mass in the theory .
thus , the power law growth ( [ eq1 ] ) , parametrized by the soft pomeron intercept @xmath6 , can be valid only in a preasymptotic , though possibly large , kinematical regime .
so far , this growth is not understood in the context of qcd . indeed ,
perturbation theory in the lo bfkl framework does result in a powerlike growth @xmath7 but the bfkl intercept @xmath8 is much larger than the experimentally observed one . inspecting at large rapidity @xmath9 the analytically known dependence of the bfkl density @xmath10 of gluons of momentum @xmath11 at impact parameter @xmath12 originating from a gluon of momentum @xmath13 at initial rapidity @xmath14 , two different perturbative ( non - unitary ) growth mechanisms can be identified in the bfkl calculation : 1 . for @xmath15 , one finds @xmath16 . for bfkl ,
this is the _ dominant growth mechanism _ , resulting in ( [ eq2 ] ) .
2 . for @xmath17
, one finds @xmath18 . for bfkl ,
this is a _ subleading growth mechanism_. here , we review arguments that including `` gluon saturation '' , i.e. parton density effects , in the lo perturbative qcd calculation allows to regulate the dominant growth in density , thus taming in comparison to bfkl the non - unitary powerlike growth to @xmath19 .
this provides a perspective for a perturbative calculation of the soft pomeron intercept .
for an explicit discussion of gluon saturation effects , we take recourse to a toy model of a hadronic collision . at initial rapidity @xmath14 , we consider a perturbative @xmath20-dipole with quarks at transverse positions @xmath21 and @xmath22 colliding with a hadronic target . the dipole scattering probability is @xmath23 [ we also write @xmath24 , and the total cross section is obtained by integrating over the impact parameter , @xmath25 . to calculate the rapidity dependence of the cross section , one may work in the target rest frame where @xmath26 corresponds to boosting the projectile .
this boost generates additional gluons which in the large @xmath27-limit translate into additional dipoles at different transverse positions .
pictorially : @xmath28 calculating the interaction of this boosted dipole wave function with the hadronic target and taking the derivative with respect to the phase space opened up by the boost , one regains the balitsky - kovchegov equation @xcite : @xmath29 \ , .\label{eq3}\end{aligned}\ ] ] this non - linear evolution equation differs from the linear bfkl equation by the addition of the last term @xmath30 only .
the physics contained in this modification of the bfkl equation is that of a _
double counting correction_. namely , for a projectile consisting of two dipoles @xmath31 and @xmath32 , the scattering probability is the sum of the individual scattering probabilities minus the probability that both dipoles scatter : @xmath33 the remaining fourth term @xmath23 in ( [ eq3 ] ) is a virtual correction required by the proper normalization of the projectile wavefunction @xmath34 . in conclusion ,
the bk evolution equation ( [ eq3 ] ) can be viewed in the target rest frame as bfkl evolution of the boosted projectile but supplemented by the double counting correction for the interaction of the evolved projectile wavefunction with the target .
the swelling of the bk dipole projectile wavefunction in impact parameter space is thus known from bfkl .
the density of dipoles of size @xmath35 at @xmath12 obtained by boosting the initial dipole of size @xmath36 to rapidity @xmath37 takes the form @xmath38\ , .
\label{eq5}\end{aligned}\ ] ] to determine the growth of the total cross section from this density , we consider the smallest dipole for which the scattering probability @xmath39 is of order one for @xmath40 : @xmath41 if after boosting the density of dipoles at @xmath12 becomes unity , then the scattering probability is of order unity as well .
this establishes the non - unitary growth @xcite @xmath42\ , .
\label{eq6}\ ] ] from ( [ eq5 ] ) , one finds for the bk intercept @xmath43 $ ] .
this expression is parametrically correct but numerically questionable since ( [ eq5 ] ) is based on a saddle point approximation valid for sufficiently small impact parameter only .
still , the numerical value @xmath44 is compatible with a numerical result of g. salam @xcite who finds a non - unitary growth of total cross sections with @xmath45 in a simulation which encodes essentially the same physics as ( [ eq3 ] ) .
the above derivation was given in the target rest frame where the evolution resides in the projectile wavefunction .
the same result can be obtained @xcite within weigert s formulation where the evolution is ascribed to the target .
this latter argument can be viewed as establishing the non - unitary growth of the hadronic cross section for a coloured scattering probe .
moreover , it was shown @xcite that the total cross section violates unitarity for a colourless projectile as soon as it does so for a coloured one .
these three arguments are independent of each other but consistent with each other and with ref .
they establish that while the bk equation ensures _
saturation _ of the scattering probability locally in impact parameter space ( i.e. @xmath46 for all @xmath12 ) , it violates _ unitarization _ of the total cross section @xcite .
in the target rest frame , the only density effect included in the bk equation ( [ eq3 ] ) is the nonlinear double counting correction in ( [ eq4 ] ) .
this correction sets in as soon as the _ number _ of gluons in the projectile state becomes large , even if the gluonic _ density _ may still be small .
in addition , as soon as the partonic density in the projectile reaches a critical value of order @xmath47 , one expects wavefunction saturation effects which further tame the growth of density in the projectile and are not included in ( [ eq3 ] ) .
the rapidity scales for these different effects are:@xcite @xmath48 at rapidity @xmath49 , the bfkl equation violates unitarity due to the growth in density . from ( [ eq6 ] )
one finds that at rapidity @xmath50 , the bk equation violates unitarity for a ( small ) dipole of initial size @xmath36 incident on a sufficiently extended target of size @xmath51 .
this violation is due to the growth of the projectile in transverse size .
the applicability of the bk evolution , however , crucially depends on the nature of the target .
if the target is thick enough , so that the scattering probability is parametrically larger than @xmath52 [ for a large nucleus of atomic number @xmath53 , it is @xmath54 , and if the target is wide enough , so that saturation occurs before the projectile radius swells beyond that of the target , then there is an intermediate regime @xmath55 in which bk applies .
however , if the target is a nucleon , neither one of these conditions is satisfied .
wavefunction saturation effects should set in when the interaction probability of a dipole of size @xmath35 with another dipole of similar size is of order unity , @xmath56 . taking the density growth exponential in rapidity , @xmath57
, this indicates that saturation effects beyond those included in the bk equation become relevant for @xmath58 .
it is reasonable to expect that projectile wavefunction saturation effects further diminish the intercept parameter of total cross sections beyond the exponent of the bk equation .
the reason is that a tamed growth of the density in the projectile due to dipole interactions will result in a tamed growth in impact parameter .
this suggests the hierarchy @xmath59 thus , one is led naturally to conjecture that a non - linear generalization of the bfkl equation which goes beyond bk by taking into account wave function saturation effects ( `` pomeron loops '' ) provides a perturbative framework for the calculation of the soft pomeron intercept , @xmath60 .
however , this requires that the perturbative growth mechanism occurs in the hadron on a scale smaller than the confinement scale at which the froissart bound can be expected to arise as a consequence of non - perturbative physics . as explained in ref.@xcite ,
a picture of the proton as consisting of three small black discs associated with the valence quarks is consistent with what is known about hadronic cross sections and allows for this possibility .
9 i. balitsky nucl
* b463 * ( 1996 ) 99 ; + yu . kovchegov phys . rev .
* d60 * ( 1999 ) 034008 .
a. kovner and u. a. wiedemann , arxiv : hep - ph/0112140 , phys .
d in press .
a. kovner and u. a. wiedemann , arxiv : hep - ph/0204277 , phys .
salam , nucl.phys.*b461 * ( 1996 ) 512 - 538 .
a. kovner and u. a. wiedemann , arxiv : hep - ph/0207335 . | we show that the inclusion of parton density effects in the perturbative small - x evolution reduces the strength of the powerlike growth of total hadronic cross sections . |
during the last two decades the amount of information transmitted via optical fiber communication systems has increased enormously .
high performance of modern communication lines is provided by an optical fiber @xcite , where the information is encoded and transmitted as a sequence of light pulses . despite the very high data rate achieved today ( @xmath0 50 tb / s over single fiber )
, there is urgent demand for even more capacity of the fiber line , originating from the needs of telecommunications and the internet .
different approaches are being pursued to address the capacity problem , such as improving the transmission properties of the optical fiber , lying additional parallel cables and using the information coding schemes going beyond the binary format , which is currently employed . among the above mentioned approaches using the extended alphabet , where a bound state of two or more optical solitons called _ soliton molecule _
serves as a new carrier of information unit @xcite , seems to be most appealing because it employs already existing optical fiber lines , and therefore is beyond the competition from the viewpoint of cost - effectiveness .
it is appropriate to recall that solitons are self - localized wave packets that can propagate along wave - guides preserving their shape and velocity , and exhibit particle - like collisions with each - other .
bright solitons emerge from the fine balance between the dispersive broadening and nonlinear self - focusing of the wave packet .
when the optical soliton was theoretically predicted @xcite and experimentally observed @xcite , there was a promising idea that it can be used as information carrier in optical fiber communication systems @xcite due to its exceptional robustness against perturbations . besides
, the nonlinearity of the optical fiber , considered to be nuisance in linear systems , has been used as an advantage in this case as it provides the soliton s self - healing property . however , some other detrimental effects like four - wave mixing and gordon - haus timing jitter @xcite have imposed difficulties for the progress in this direction
. later the concept of dispersion - management ( dm ) and dm soliton was put forward ( for a recent review see @xcite ) , which allowed to suppress these adverse effects , and eventually has led to realization of several commercial soliton based optical fiber communication systems .
an interesting chronicle of the growth of the telecommunications industry , including the fiber optic systems , is given in ref .
@xcite .
recent progress in using solitons for information transfer is linked to the observation , that solitons in dm fibers can form stable bound states called soliton molecules , and realizing their potential for enhancing the capacity of the communication system via extension of the coding alphabet @xcite .
the binding mechanism of solitons in the molecule was proposed in @xcite .
the proof - of - principle demonstration of the data transmission via optical fiber using the extended alphabet : logical _ zero _ , _ one _ ( single soliton ) , _ two _ ( two - soliton molecule ) and _ three _ ( three - soliton molecule ) , was recently reported in @xcite .
the practical implementation of this novel approach requires extensive research on the existence regimes , stability , mutual interactions and propagation dynamics of soliton molecules in dm fibers . in this work
we study the existence regimes and dynamics of soliton molecules in dm fibers by analytical and numerical means .
first we develop a variational approach ( va ) to find the stationary shape of the molecule and equilibrium separation between solitons in the molecule . at this stage
we obtain two coupled ordinary differential equations ( ode ) for the temporal separation between solitons and chirp parameter , which describes the fast dynamics of the molecule within a unit cell of the dm fiber .
although the derived ode system is capable of describing the propagation of the molecule for arbitrary distance , long - haul transmission of dm solitons and molecules is convenient to explore using the averaged nonlinear schrdinger equation ( nlse ) .
the averaged nlse can help to specify the existence regimes of solitons and molecules , elucidate the physical mechanism by which they loose energy , and eventually disintegrate in conservative dm fibers .
the paper is organized as follows . in the next sec .
ii we introduce the nlse which governs the pulse propagation in dm fibers . here
we also present the parameters of the dm fiber used in our calculations . in sec .
iii we develop the va for the fast dynamics of solitons and molecules and compare its predictions with the results of pde simulations .
then we derive in sec .
iv the averaged nlse and determine its coefficients using the va .
here we also perform the analysis of the pulse propagation in the nlse with inverted parabolic potential . in sec .
v we summarize our findings .
propagation of optical pulses in fibers with inhomogeneous parameters is described by the following nonlinear schrdinger equation @xmath1 where @xmath2)$ ] , @xmath3 $ ] , and @xmath4 $ ] are respectively , the complex envelope of the electric field , the propagation distance , and the retarded time .
the coefficients @xmath5 $ ] , @xmath6 $ ] , and @xmath7 $ ] represent the fiber s group velocity dispersion ( gvd ) , nonlinearity and gain / loss parameter , respectively . here and below in rectangular brackets
@xmath8 $ ] we show the physical unit of the corresponding variable . for qualitative analysis and numerical simulations
it is convenient to reduce the eq .
( [ nlse0 ] ) into dimensionless form . at first
we eliminate the gain / loss term via new variable @xmath9)$ ] , following ref .
@xcite @xmath10,\ ] ] where @xmath11 is dimensionless constant .
the new complex function @xmath12 satisfies the equation @xmath13 where @xmath14 is the fiber s effective nonlinearity .
now we convert the eq .
( [ nlse1 ] ) into standard form with constant nonlinearity by introducing the new coordinate @xmath15 $ ] defined by @xmath16 @xmath17 where @xmath18 [ w @xmath19 s@xmath20 is the fiber s effective dispersion , characterizing both the fiber s gvd and nonlinearity .
the original parameters @xmath21 , @xmath22 and @xmath23 in eq .
( [ nlse0 ] ) are periodic functions of propagation distance with common period @xmath24 , while the effective dispersion @xmath25 in eq .
( [ nlse2 ] ) is a periodic function with period defined by @xmath26.\ ] ] to obtain the final equation we introduce dimensionless variables @xmath27 where @xmath28 is the characteristic time scale equal to pulse duration of the laser source @xmath29 . in terms of these variables
the dimensionless governing equation acquires the form @xmath30 where @xmath31 represents the fiber s dimensionless effective dispersion . to show the results in dimensional variables we solve the equation @xmath32 with respect to @xmath33 [ m ] , for given dimensionless propagation distance @xmath34 , the period @xmath35 [ 1/w ] , and known nonlinear map function @xmath36 [ @xmath37 .
the original time @xmath38 [ s ] and field amplitude @xmath39 @xmath40 $ ] are restored via eq.([dim ] ) .
note that @xmath41 corresponds to @xmath42 , in accordance with eq .
( [ lprime ] ) . in the absence of gain
/ loss term in eq .
( [ nlse0 ] ) , i.e. @xmath43 @xmath44 \ , dz } \right)^{1/2 } = 1.\ ] ] the dimensionless pulse energy is @xmath45 [ j ] - is the original pulse energy . in the following sections we employ the dm map parameters , corresponding to the setup of ref .
@xcite , for laser wavelength @xmath46 nm * @xmath47 = @xmath48 4.259 [ ps@xmath49/km ] , @xmath50=1.7 [ 1/(w @xmath19 km ) ] , @xmath51 = 22 [ m ] ( positive gvd fiber ) , * @xmath52 = @xmath53 5.159 [ ps@xmath49/km ] , @xmath54=1.7 [ 1/(w @xmath19 km ) ] , @xmath55 = 24 [ m ] ( negative gvd fiber ) .
the period of the original dm map is @xmath56 = 46 [ m ] .
for the effective dispersion in eq .
( [ nlse2 ] ) the period is @xmath57 = 0.078 [ 1/w ] . the path averaged dispersion and nonlinearity are equal to * @xmath58 = @xmath53 0.655 [ ps@xmath49/km ] , @xmath59 = 1.7 [ 1/(w @xmath19 km ) ] . the characteristic time , length and energy scales are given below * @xmath60 [ ps ] , @xmath61 [ ps ] ( pulse duration ) , * @xmath62 [ km ] ( dispersion length ) , * @xmath63 = @xmath53 3.796 , @xmath64 = 0.522 , * @xmath65 = @xmath48 3.135 , @xmath66 = 0.478 , * @xmath67 = @xmath53 0.482 , @xmath68 = 1 .
figure [ fig1 ] shows the dispersion profiles for the original and reduced governing equations . ) .
right panel : for reduced dimensionless eq . ( [ nlse3 ] ) .
note that here the coefficient of nonlinearity is equal to one and has not been shown in eq .
( [ nlse3 ] ) and figure.,title="fig:",width=302,height=151 ] ) .
right panel : for reduced dimensionless eq . ( [ nlse3 ] )
. note that here the coefficient of nonlinearity is equal to one and has not been shown in eq .
( [ nlse3 ] ) and figure.,title="fig:",width=302,height=151 ]
propagation of a dm soliton is characterized by fast variation of the pulse shape within each period of the dm map and slow variations on longer distances when observed stroboscopically , i.e. once per dispersion period .
the periodic dynamics is compromised by different imperfections of the fiber , continuous emission of linear waves by the soliton and non - perfect initial shape of the dm soliton injected into the fiber . to compensate for the power loss and other distortions , in practice
the pulses are regenerated by amplifiers which are equidistantly installed along the fiber line .
the variational approach , initially developed for description of optical soliton propagation in homogeneous fibers @xcite , later was successfully applied to dm solitons @xcite and antisymmetric solitons in dm fibers @xcite .
the advantages and disadvantages of the va as compared to other methods of exploring dm solitons are discussed in @xcite .
below we elaborate the va for two- and three - soliton molecules in dm fibers , based on the reduced nlse ( [ nlse3 ] ) .
( [ nlse3 ] ) can be obtained from the following lagrangian density ( for tidier notations we again use small letters instead of capitals ) @xmath69 to derive the va equations we consider the shape of the soliton molecule in general form @xcite @xmath70 where @xmath71 are the chirp parameter and phase , @xmath72 is the pulse duration ( proportional to the separation between solitons in the molecule ) , @xmath73 is the real function which represents the stationary shape of the soliton molecule .
the energy of the pulse is ( subscript zero in eq .
( [ energy ] ) is dropped ) @xmath74 substitution of the trial function ( [ ansatz ] ) into eq .
( [ lagr ] ) yields the lagrangian density @xmath75 the averaged lagrangian is obtained by integrating the last expression over the reduced time variable @xmath76 with a few integral constants , determined solely by the pulse shape @xmath73 .
the euler - lagrange equations with respect to variational parameters @xmath77 give rise to a coupled set of ode s @xmath78 where @xmath79 we adopt the following trial functions @xmath73 to specify the shapes of the pulses and define the corresponding parameters @xmath80 the fixed point ( @xmath81 ) of the coupled system ( [ tau])-([alpha ] ) , which can be found using the nijhof s method for va models @xcite , defines the stationary shapes of the pulses with given energy @xmath82 and dm map function @xmath83 . in fig .
[ fig2 ] we illustrate the shapes of the two - soliton and three - soliton molecules , which are found by solving the va system ( [ tau])-([alpha ] ) and compare with the corresponding results of the nijhof s method applied to the original eq .
( [ nlse3 ] ) . ) , while blue dashed lines are the prediction of va equations ( [ tau])-([alpha ] ) .
appreciable deviations of the two waveforms are seen only in far tails , where the pulse amplitude is less than a percent of the maximum value .
the parameters for the 2-soliton molecule : @xmath84 ( in original units @xmath85 pj ) , @xmath86 , @xmath87 . for the 3-soliton molecule : @xmath88 ( @xmath89 pj ) , @xmath90 , @xmath91 .
, title="fig:",width=302,height=226 ] ) , while blue dashed lines are the prediction of va equations ( [ tau])-([alpha ] ) .
appreciable deviations of the two waveforms are seen only in far tails , where the pulse amplitude is less than a percent of the maximum value .
the parameters for the 2-soliton molecule : @xmath84 ( in original units @xmath85 pj ) , @xmath86 , @xmath87 . for the 3-soliton molecule : @xmath88 ( @xmath89 pj ) , @xmath90 , @xmath91 .
, title="fig:",width=302,height=226 ] in fig .
[ fig3 ] we show the phase trajectory for the periodic solution of the va system ( [ tau])-([alpha ] ) and temporal center - of - mass positions of solitons in the two - soliton molecule while it propagates along the fiber .
the center - of - mass position of the right / left soliton is calculated from the solution of the va system ( [ tau])-([alpha ] ) as @xmath92 , while for the nlse ( [ nlse3 ] ) the corresponding formula is @xmath93 where in actual calculations we use the temporal domain s half length as the upper limit of the integration .
, @xmath94 corresponds to the periodic solution of the system ( [ tau])-([alpha ] ) for a two - soliton molecule with energy @xmath95 , shown in fig.[fig2 ] .
right panel : the solution of the nlse ( [ nlse3 ] ) with the initial pulse shape predicted by the va shows that solitons in the two - soliton molecule perform oscillatory motion periodically reducing and increasing the temporal separation between their center - of - mass positions ( top and bottom curves ) .
the evolution of the chirp parameter ( middle curve ) shows qualitative agreement with the va.,title="fig:",width=302,height=226 ] , @xmath94 corresponds to the periodic solution of the system ( [ tau])-([alpha ] ) for a two - soliton molecule with energy @xmath95 , shown in fig.[fig2 ] .
right panel : the solution of the nlse ( [ nlse3 ] ) with the initial pulse shape predicted by the va shows that solitons in the two - soliton molecule perform oscillatory motion periodically reducing and increasing the temporal separation between their center - of - mass positions ( top and bottom curves ) .
the evolution of the chirp parameter ( middle curve ) shows qualitative agreement with the va.,title="fig:",width=302,height=226 ] for evaluation of the pulse width and chirp parameter from the solution of eq .
( [ nlse3 ] ) we employ the following expressions @xcite . @xmath96
averaged equation for soliton molecules in dm fibers ----------------------------------------------------
long - haul propagation of soliton molecules in dm fibers is convenient to study using the averaged nlse .
the averaging procedure was developed in @xcite .
below we use the approach based on the scaling arguments proposed in ref . @xcite and look for the solution of eq .
( [ nlse3 ] ) in the form ( capital letters changed to lower - case ) @xmath97 where @xmath98 is the chirp parameter . inserting this into the governing eq .
( [ nlse3 ] ) we obtain @xmath99 by introducing the new time ( @xmath100 ) and amplitude ( @xmath101 ) functions @xmath102 the last equation can be reduced into the form @xmath103 when the following relations are satisfied @xmath104 the nlse with a parabolic potential results from eq .
( [ eq4 ] ) @xmath105 the amplitude function is linked to the pulse width as @xmath106 , which can be readily verified from eq .
( [ eq5 ] ) .
the constant @xmath107 is specified by the selected trial function . now applying the averaging procedure to eq .
( [ eq6 ] ) we obtain in leading order @xmath108 where @xmath109 is the slowly varying core of the dm soliton .
the quantities averaged over one dispersion period are defined as @xmath110 , @xmath111 , @xmath112 \ , \tau^2(z ) \right>$ ] .
the averaging is performed as @xmath113 ( for the reduced map period @xmath114 ) . in numerical implementation of the averaging procedure
we employ the periodic solution of va eqs .
( [ tau])-([alpha ] ) .
in particular from eq .
( [ alpha ] ) it follows that @xmath115 , with constants @xmath116 given by eq .
( [ c1c2 ] ) . for the dm map parameters specified in sec .
[ params ] all coefficients of eq .
( [ averaged ] ) appear to be positive ( @xmath117 , @xmath118 , @xmath119 ) , therefore we have gotten the nlse with anomalous dispersion , focusing nonlinearity and inverted parabolic potential .
this equation is formally similar to the quantum mechanical equation for a wave packet , evolving under the effective potential @xmath120 the stationary solution @xmath121 can be found from the initial value problem with respect to variable @xmath100 @xmath122 and suitable initial conditions @xmath123 and @xmath124 .
it as appropriate to mention that eq .
( [ averaged ] ) with additional gain / loss term was previously considered also in other contexts , such as the nonlinear compression of chirped optical solitary waves @xcite , and with regard to integrability issues @xcite . to study the evolution of pulses governed by eq .
( [ averaged ] ) we develop the va .
the corresponding lagrangian density is @xmath125 the trial function will be of the form @xmath126 where @xmath127 , @xmath72 , @xmath128 , @xmath129 , @xmath98 , @xmath130 are variational parameters , designating the pulse amplitude , width , center of mass position , velocity , chirp and phase , respectively .
the auxiliary function @xmath131 is introduced for convenience and defines the type of the pulse . specifically for a single soliton @xmath132 , for two - soliton molecule @xmath133 and for three - soliton molecule @xmath134 .
it is instructive to start with considering the existence and dynamics of a single soliton on top of an inverted parabolic potential .
the integration @xmath135 using the trial function ( [ ansatz1 ] ) with @xmath132 gives rise to the following effective lagrangian @xmath136 where the pulse energy @xmath137 is conserved .
now applying the euler - lagrange equations with respect to the variational parameters , we get the ode system for the pulse width and its center - of - mass position @xmath138 the system for the two - soliton and three - soliton molecules will be similar , with only re - scaled energy coefficient , @xmath139 for the former case and @xmath140 for the latter case ( see the coefficient @xmath141 in eqs.([2sol])-([3sol ] ) ) .
as can be seen from this system , the center - of - mass and internal dynamics of the soliton are decoupled .
this is due to the property of a parabolic potential and is the manifestation of the ehrenfest s theorem ( its validity for the nonlinear schrdinger equation with a linear and parabolic potentials was proved in ref .
@xcite ) . in other types of potentials
these two degrees of freedom are coupled .
in fact the equation for the center - of - mass is the harmonic oscillator equation with a purely imaginary frequency @xmath142 , since in practical situations @xmath143 and @xmath144 .
therefore , the center - of - mass of the soliton is unstable against sliding down an inverted parabola with exponentially increasing distance from the origin @xmath145 the equation for the pulse width ( [ tau1 ] ) is similar to the equation of motion for a unit mass particle in the anharmonic potential @xmath146 which is depicted in fig .
the minimum of this potential is found from the solution of the quartic equation @xmath147 and corresponds to the stationary width of the soliton . for a given pulse energy @xmath82
this defines the shape of the soliton according to the ansatz ( [ ansatz1 ] ) . at some critical energy @xmath148 the local minimum in this
potential disappears , which means that eq .
( [ averaged ] ) does not support solitons and molecules with energy below @xmath148 .
the value of the critical energy can be found from the condition that eq .
( [ quartic ] ) has a real solution , which takes place if @xmath149 , or in terms of energy @xmath150 in fact the shape of the potential @xmath151 gives the evidence that solitons and molecules in the system are meta - stable . in the same figure [ fig4 ]
we compare the pulse profiles obtained by solution of the stationary state eq .
( [ ivp ] ) with the prediction of va .
the parameter @xmath152 for eq .
( [ ivp ] ) is obtained using the shooting method , where the pulse energy @xmath153 , @xmath154 are used . as can be seen from this figure ,
the deviation between the two wave profiles is notable at far tails of the pulse , where the field intensity significantly decreases .
the wavy tails on the numerically exact pulse profile , found from solution of eq .
( [ ivp ] ) indicate the existence of waves reflected from ( and partially transmitted through ) the borders of the effective potential ( [ potential ] ) .
the waves escaping the effective potential contribute to the continuous outflow of energy from the soliton propagating along the fiber , and that will be the fundamental source of its instability in conservative dm fibers . ) in the effective particle model of va for two values of the pulse energy .
the pulse with greater energy forms deeper potential well . the stationary amplitude and width of the soliton for @xmath155 , according to va are @xmath156 and @xmath157 respectively .
the existence of a local maximum and decay at large @xmath158 is the evidence of meta - stability of the stationary state .
right panel : comparison between the pulse shapes ( in logarithmic scale ) obtained by solving the stationary state problem ( [ ivp ] ) for @xmath159 and prediction of va for the same pulse energy .
the wavy tails are due to the interference of waves reflecting from the borders of the effective potential ( [ potential ] ) .
parameters @xmath160 , @xmath161 , @xmath162 are obtained using the va eqs .
( [ tau])-([alpha ] ) for a single dm soliton with energy @xmath163 ( in original units @xmath164 pj ) and fiber data of sec .
[ params ] .
, title="fig:",width=302,height=226 ] ) in the effective particle model of va for two values of the pulse energy .
the pulse with greater energy forms deeper potential well . the stationary amplitude and width of the soliton for @xmath155 , according to va are @xmath156 and @xmath157 respectively .
the existence of a local maximum and decay at large @xmath158 is the evidence of meta - stability of the stationary state .
right panel : comparison between the pulse shapes ( in logarithmic scale ) obtained by solving the stationary state problem ( [ ivp ] ) for @xmath159 and prediction of va for the same pulse energy .
the wavy tails are due to the interference of waves reflecting from the borders of the effective potential ( [ potential ] ) .
parameters @xmath160 , @xmath161 , @xmath162 are obtained using the va eqs .
( [ tau])-([alpha ] ) for a single dm soliton with energy @xmath163 ( in original units @xmath164 pj ) and fiber data of sec .
[ params ] . , title="fig:",width=302,height=226 ] the pulse in the effective potential " picture , shown in fig .
[ fig5 ] , is helpful for the analysis of instability issues .
the first aspect to be noted is that , the height and width of the effective potential barrier in both directions from the pulse are finite , and depends on the intensity of the pulse itself . therefore a continuous and nonlinearly progressing energy outflow from the soliton takes place via the tunneling effect , whose rate can be estimated by means of the semiclassical wkb method as done for the matter - wave analogue of this problem in ref .
these authors also have shown that the rate of energy outflow ( number of particles for condensates ) nonlinearly increases , eventually leading to disintegration of the soliton .
the second aspect is the instability of the center - of - mass position of the soliton against sliding down an inverted parabola . according to eq .
( [ com ] ) any small departure of the center - of - mass position from the origin ( the top of the inverted parabola ) in either direction will exponentially grow . ) are shown by red solid lines .
the corresponding effective potentials according to eq .
( [ potential ] ) are shown by blue dashed lines .
the parameters for the single soliton are the same as in previous figure , while for the two - soliton molecule the parameters are @xmath165 , @xmath166 , @xmath167 , @xmath168 ( in original units @xmath169 pj ) . , title="fig:",width=302,height=226 ] )
are shown by red solid lines .
the corresponding effective potentials according to eq .
( [ potential ] ) are shown by blue dashed lines .
the parameters for the single soliton are the same as in previous figure , while for the two - soliton molecule the parameters are @xmath165 , @xmath166 , @xmath167 , @xmath168 ( in original units @xmath169 pj ) .
, title="fig:",width=302,height=226 ] to verify the above conclusions following from our model , we performed numerical simulations of the pulse propagation governed by eq .
( [ averaged ] ) .
when we introduce the wave profile predicted by va , which slightly differs from the solution of eq .
( [ ivp ] ) , as initial condition to eq .
( [ averaged ] ) , the pulse quickly adjusts itself by performing damped oscillations of its amplitude , and then continuously decay both in the energy and amplitude , as shown in fig .
the frequency of oscillations of the amplitude well agrees with the prediction of va , when we expand the potential ( [ pot1 ] ) near its minimum @xmath170 and estimate the corresponding frequency and period @xmath171 , @xmath172 .
however , the va does not take into account the dissipative effects . in numerical simulations
we use the absorbing boundary technique @xcite to prevent the interference of the soliton with the linear waves , otherwise reflected from the integration domain boundaries . to calculate the energy outflow from the soliton , we monitor the amount of energy in the central part of the domain ( @xmath173 $ ] in fig .
[ fig6 ] ) where the bulk of the pulse is confined .
the energy loss rate @xmath174 is characteristic for the fiber parameters and initial pulse power , which define the coefficients of the averaged nlse ( [ averaged ] ) .
the instability of the center - of - mass position of the soliton is demonstrated in the middle panel of fig .
if the calculation is performed with centering of the soliton , when at each step its center - of - mass is held at the origin , we observe complete disintegration of the pulse due to the energy loss , as shown in the right panel of this figure .
the critical energy , at which the pulse disintegrates , according to the inset of the left panel is @xmath175 .
this numerical finding is in good agreement with the prediction of va eq .
( [ ecrit ] ) @xmath176 .
in fact the above mentioned mechanism of energy loss sets the limit to attainable robustness of dm solitons and molecules in the given setup .
the instability of the center - of - mass of the soliton will make it difficult to held the pulse in the middle of its time slot . ) with absorbing boundary conditions .
the initial wave is a gaussian pulse with the amplitude @xmath156 and width @xmath157 , as predicted by va .
middle panel : the soliton slightly displaced from the top of the inverted parabola ( by amount @xmath177 ) slides down with increasing velocity in the corresponding direction .
this instability develops even due to a numerical noise , when the soliton is placed exactly at the origin @xmath178 .
right panel : when calculation is performed with centering of the pulse , a complete disintegration can be observed at sufficiently long propagation distance.,title="fig:",width=264,height=226 ] ) with absorbing boundary conditions .
the initial wave is a gaussian pulse with the amplitude @xmath156 and width @xmath157 , as predicted by va .
middle panel : the soliton slightly displaced from the top of the inverted parabola ( by amount @xmath177 ) slides down with increasing velocity in the corresponding direction .
this instability develops even due to a numerical noise , when the soliton is placed exactly at the origin @xmath178 .
right panel : when calculation is performed with centering of the pulse , a complete disintegration can be observed at sufficiently long propagation distance.,title="fig:",width=188,height=226 ] ) with absorbing boundary conditions .
the initial wave is a gaussian pulse with the amplitude @xmath156 and width @xmath157 , as predicted by va .
middle panel : the soliton slightly displaced from the top of the inverted parabola ( by amount @xmath177 ) slides down with increasing velocity in the corresponding direction .
this instability develops even due to a numerical noise , when the soliton is placed exactly at the origin @xmath178 .
right panel : when calculation is performed with centering of the pulse , a complete disintegration can be observed at sufficiently long propagation distance.,title="fig:",width=188,height=226 ] in fig .
[ fig7 ] we demonstrate the propagation of two solitons governed by eq .
( [ averaged ] ) .
when two in - phase solitons are initially placed at a separation exceeding some critical value , they move apart with increasing velocity under the effect of the expulsive parabolic potential . if the solitons are placed at a smaller temporal distance , they collide few times and merge together , and later the combined pulse develops instability , analogous to the single soliton case and slides down the inverted parabola .
similar behavior of two - soliton states were reported in the dissipative counterpart of eq .
( [ averaged ] ) in ref .
an additional fact relevant to coupled soliton propagation in this system is that , the out - of - phase solitons always repel each - other and diverge from the origin in accelerated manner . ) move apart .
middle panel : when solitons are placed at a distance ( @xmath179 ) which is less than some critical value , they collide few times and merge .
right panel : long distance behavior of the coalescent pulse ( of the middle panel ) is similar to the single soliton case , as its center - of - mass develops instability and slides down the inverted parabola.,title="fig:",width=226,height=226 ] ) move apart .
middle panel : when solitons are placed at a distance ( @xmath179 ) which is less than some critical value , they collide few times and merge .
right panel : long distance behavior of the coalescent pulse ( of the middle panel ) is similar to the single soliton case , as its center - of - mass develops instability and slides down the inverted parabola.,title="fig:",width=226,height=226 ] ) move apart .
middle panel : when solitons are placed at a distance ( @xmath179 ) which is less than some critical value , they collide few times and merge .
right panel : long distance behavior of the coalescent pulse ( of the middle panel ) is similar to the single soliton case , as its center - of - mass develops instability and slides down the inverted parabola.,title="fig:",width=226,height=226 ] the main result of our study of pulse propagation governed by the averaged nlse ( [ averaged ] ) is that , it does not support truly stable solitons and molecules as the original dm eq .
( [ nlse3 ] ) does .
rather it shows , that solitons and molecules in the conservative dm fiber are meta - stable , both in terms of energy and temporal position .
in addition , the model of averaged nlse allows to find the existence regimes and identify the limits of stability of solitons and molecules in dm fibers .
the inclusion of dissipation and gain into the model may change the results to some degree .
this will be a subject for separate study .
we have developed a variational approximation which successfully describes the propagation of soliton molecules in dm fibers .
the pulse shapes for a two - soliton molecule and three - soliton molecule , predicted by va are shown to be sufficiently close to the numerically exact shapes found by solution of the original dm nls equation .
then we have studied the dynamics of solitons and molecules in the averaged nlse corresponding to the selected dm fiber system .
the approach of averaged nlse allows to identify the regimes of existence of solitons and molecules in the original dm system , and reveal the fundamental source of instability of soliton propagation in dm fibers , which is linked to continuous outflow of energy from the pulse due to the wave tunneling phenomenon .
the model also predicts the instability of the temporal position of the pulse within its time slot .
all calculations are performed using the parameters of the existing dm fiber setup @xcite .
the model may provide guidance in further studies of the properties of soliton molecules in dm fibers .
we thank f. mitschke , a. hause and e. n. tsoy for valuable discussions . u.a.k . and
. are grateful to the department of physics of the kfupm for the hospitality during their stay .
this work is supported by the research grant uaeu - nrf 2011 and kfupm research projects rg1214 - 1 and rg1214 - 2 .
s. k. turitsyn , i. gabitov , e. w. laedke , v. k. mezentsev , s. l. musher , e. g. shapiro , t. schafer , and k. h. spatschek , opt . commun . * 151 * , 117 ( 1998 ) .
c. pare and p. -a .
belanger , opt .
* 168 * , 103 ( 1999 ) . | the existence regimes and dynamics of soliton molecules in dispersion - managed ( dm ) optical fibers have been studied .
initially we develop a variational approximation ( va ) for description of periodic dynamics of a soliton molecule within each unit cell of the dispersion map .
the obtained system of coupled equations for the pulse width and chirp allows to find the parameters of dm soliton molecules for the given dispersion map and pulse energy .
then by means of a scaling transformation and averaging procedure we reduce the original nonlinear schrdinger equation ( nlse ) with piecewise - constant periodic dispersion to its counterpart with constant coefficients and additional parabolic potential .
the obtained averaged nlse with expulsive potential can explain the essential features of solitons and soliton molecules in dm fibers related to their energy loss during propagation . also , the model of averaged nlse predicts the instability of the temporal position of the soliton , which may lead to difficulty in holding the pulse in the middle of its time slot .
all numerical simulations are performed using the parameters of the existing dm fiber setup , and illustrated via pertinent examples . |
colored scalars are frequent harbingers of new physics . for example , they are inherent to any theory of matter unification @xcite .
phenomenologically , colored scalars that can couple directly to matter fields are of particular interest .
this feature makes them very appealing candidates for collider physics and precision flavor studies .
for example , some have been suggested as possible explanations of the enhanced forward - backward asymmetry in @xmath6 production as measured at the tevatron @xcite or the @xmath7 anomaly @xcite .
more recently , it has been pointed out that some could also contribute significantly to the unexpectedly large cp asymmetry in the decays @xmath8 @xcite . moreover , in a framework of simple one - particle extensions of the standard model ( sm ) , these states can establish a unique link between collider and planck scale physics via proton decay @xcite .
they might even help induce tiny neutrino masses through loop effects @xcite .
it has been demonstrated that light colored scalars help improve unification of gauge couplings in the grand unified theory ( gut ) setting @xcite .
it could thus be possible to establish a firm connection between the colored scalar mass scale and observable proton decay signatures in the minimal matter unification scenarios .
moreover , the colored states often reside in the same representation of the gut group as the fields responsible for the electroweak symmetry breaking ( ewsb ) .
this could allow one to infer the strength of yukawa couplings of colored states to matter .
these , on the other hand , are subject to nontrivial constraints originating from low - energy flavor phenomenology and the requirement of viable fermion masses .
what one can end up with is a highly predictive class of simple models that connect matter stability to low - energy phenomenology @xcite .
a generic feature of massive scalars in extensions of the sm is that they couple to the sm higgs boson . above the ewsb scale
, one can write marginal
the so - called higgs portal " @xcite operators of the form @xmath9 , where @xmath10 is the sm higgs doublet , @xmath11 is the new scalar weak ( and color ) multiplet and all possible color and weak contractions are assumed .
after ewsb this will induce corrections to the masses of weak @xmath11 components ( @xmath12 ) and couplings of the form @xmath13 , where @xmath14 is the physical higgs boson . at the one - loop level ,
the presence of such interactions can affect higgs production and decays as measured at the lhc @xcite .
in fact preliminary data on the higgs - like resonance with a mass of @xmath15gev show an excess of events in the loop - induced @xmath0 decay channel with respect to the sm prediction at the @xmath16 level , while other , tree - level dominated higgs decays to vector bosons show good consistency with the sm expectations @xcite .
if it were to persist at larger significance , such a deviation would point towards the existence of new degrees of freedom at the ew scale that couple to the higgs boson .
motivated by these observations , in the present study , we focus on the contributions of colored scalars to loop - induced production and decay channels of the higgs boson through interactions of the form @xmath9 ( a similar analysis based on earlier experimental results can be found in @xcite ) . in particular , we study fields with direct couplings to the sm matter fields that can thus play an interesting role in other collider and flavor observables .
we single out two fields which accommodate the observed @xmath0 enhancement while remaining in good agreement with the other measured decay channels .
interestingly enough , these fields turn out to be sextets and octets of color .
this precludes them from having leptoquark - like couplings to the ordinary matter .
we then proceed to show how and where these particular states appear in simple scenarios of matter unification .
we revisit gauge coupling unification in two full - fledged models based on the @xmath5 gauge group @xcite to demonstrate the connection between the lightness of these states and observable proton decay signatures .
one of the models can also address the @xmath7 anomaly , albeit through the presence of an additional light colored scalar .
the remainder of the paper is structured as follows . in section [ htogammagamma ]
we discuss a possible explanation of the @xmath0 excess with colored scalar fields .
we analyze the impact of all the relevant experimental inputs and specify the numerical procedure used to generate our results .
we also provide predictions for @xmath4 and di - higgs production using the best - case scenarios and associated parameters .
the fields that can help accommodate the @xmath0 excess are shown to appear naturally in matter unification scenarios in section [ color_scalars ] .
there we also correlate the lightness of these states with observable proton decay signatures .
finally , we conclude in section [ conclusions ] .
we study the influence of colored scalar fields coupling to the sm higgs doublet through @xmath9 interactions on higgs production and decay signatures at the lhc .
if sufficiently light , such states can significantly modify the loop - induced @xmath17 , @xmath18 , @xmath19 and @xmath0 processes , while leaving other , tree - level dominated higgs production and decay channels sm - like . after specifying the color and ew representations , each scenario ( @xmath11 )
can be parametrized in terms of the relevant higgs couplings and masses of the weak components @xmath12 of the multiplet .
different possible weak contractions of the @xmath20 terms will in general induce different relative contributions to these couplings and masses . however , severe experimental constraints coming mainly from the @xmath21 parameter require an approximate custodial symmetry to be active in the ew symmetric scalar potential ( c.f .
it turns out that in this limit , one can without loss of generality consider a single interaction term of the form @xmath22 where we have written out the summed over weak ( @xmath23 ) and color ( @xmath24 ) indices explicitly , and the dots denote further terms in the ew broken phase expansion of the higgs fields .
furthermore , in the custodial limit , @xmath25 are almost degenerate and we will use @xmath26 to denote their common mass . then , the partial decay width for @xmath0 at one loop is given by @xcite , @xmath27 where @xmath28 is the fermi constant , @xmath29 the fine structure constant , @xmath30 and @xmath31 the weinberg angle . also , @xmath32 for @xmath33 , while relevant one - loop functions are given by @xmath34 the first and the second term in eq .
are the sm one - loop contributions from the @xmath35 and the top quark , respectively . for @xmath36 ,
their numerical values are @xmath37 and @xmath38 . finally , @xmath39 is the dimension of the color representation of @xmath11 , and @xmath40 the electric charges of weak @xmath12 components .
analogously , the parton level @xmath17 cross section at partonic c.m.s .
energy @xmath41 reads @xcite @xmath42 @xmath43 where @xmath44 is the strong coupling constant and @xmath45 is the index of the color representation @xmath46 of @xmath11 .
the only color representations we consider are triplet , sextet and octet as these are the ones that can contract directly with sm matter fields in @xmath47 space to yield a singlet .
we accordingly use @xmath48 , @xmath49 and @xmath50 .
our goal is to confront available higgs signal strength data with possible contributions from light colored scalars .
parameters used to fit the data are @xmath51 and @xmath26 . following @xcite we define individual channel signal rates , normalized to their respective sm values as @xmath52 where the labels @xmath53 and @xmath54 denote the relevant higgs production and detection channels .
we furthermore denote the reported experimental values and variances of @xmath55 by @xmath56 and @xmath57 , respectively ( listed in table [ tab:2 ] ) . a global @xmath58
is then defined as @xmath59 where we neglect correlations among the various terms as they are not supplied by the experimental collaborations . as pointed out in @xcite ,
theoretical uncertainties are only relevant for @xmath60 , where they amount to a relative error of @xmath61 . to obtain the @xmath62 values in eq . , we add this contribution in quadrature with the experimental errors ( @xmath63 ) for each observable . for each scenario ( @xmath11 )
, we determine the minimum of the @xmath58 ( @xmath64 ) , and define the 68% ( @xmath65 ) and 95% ( @xmath16 ) best - fit regions as solutions to @xmath66 , where @xmath67 are set by the appropriate cumulative distribution function .
the relevant observables , which can be significantly affected by colored scalar contributions , are the recently measured lhc higgs production rates in the @xmath68 , @xmath69 , @xmath70 and @xmath71 channels @xcite listed in table [ tab:2 ] . in our fit
we combine the independent measurements for each channel using weighted average .
the @xmath71 channel is the only one with a significant contribution from vector boson fusion ( vbf ) .
the expected signal strength in this channel can be parametrized as @xcite @xmath72 where @xmath73 denotes the hadronic @xmath17 cross section and the ratio @xmath74 @xcite remains almost constant when going from @xmath75tev to @xmath76tev c.m.s .
energy at the lhc .
since vbf remains sm like in our scenarios , we can also identify @xmath77 .
all other channels are completely dominated by @xmath17 production alone .
.data used in the analysis .
[ cols="<,<,<",options="header " , ] we notice two different behaviors in the interesting range of parameters . obviously , to enhance @xmath0 , a negative @xmath51 is preferred , but this tends to lower @xmath17 , affecting all the measured channels .
the goal then would be to have as small @xmath45 as possible and as big @xmath78 as possible , and to have a small negative @xmath51 .
it turns out , however , that none of the considered scalars can accommodate all the data in this way .
the best candidate from the list would be the @xmath79 and we present its @xmath58 plot in the @xmath80 plane in fig .
[ fig:1 ] .
clearly , @xmath79 can fit @xmath17 well , but fails to enhance @xmath0 .
this is typical for most of the scalars listed in table [ tab:1 ] , as can be seen from the last column which shows separate contributions for all observables in the fit . for the sake of the argument , we also plot the scalar @xmath81 example , which fits the data perfectly but does not couple directly to the sm fermions .
another possibility is to allow @xmath82 to be negative and large , so that contribution of the scalar in the loop is twice the contribution of the top quark but with opposite sign . in this case @xmath17
will again have the same value as in the sm . this particular scenario to keep @xmath17 at the sm level has been observed in refs .
@xcite . to accomplish this
, we need large @xmath45 since @xmath51 is constrained by perturbativity arguments .
it turns out that we have two good candidates listed in table [ tab:1 ] that can accomplish the task .
these are the color sextet @xmath2 and color octet @xmath1 .
the corresponding @xmath58 plots for these states are presented in fig .
[ fig:2 ] .
note that the optimal parameter space in both cases is very narrow .
this practically fixes the allowed @xmath83 for a given value of @xmath84 and vice versa .
we obtain @xmath85 and @xmath86 for @xmath2 and @xmath1 scenarios , respectively . in this best - fit region ,
the sextet ( octet ) masses above @xmath87gev ( @xmath88gev ) quickly lead to non - perturbative values of the coupling @xmath89 . on the other hand ,
the remaining interesting range of masses is not yet completely excluded by direct searches at the lhc as recently emphasized in @xcite . in particular the mass window between @xmath90 gev and @xmath91 gev is allowed by present searches for colored scalars at the lhc .
furthermore , current experimental analyses assume that such scalars are narrow and decay @xmath92 to pairs of jets , none of which is required or predicted by our fit to the higgs data ( for a recent discussion on light colored resonances escaping present experimental searches c.f .
@xcite ) .
plot for @xmath79 state where @xmath93 .
black strip is a region in parameter space with the minimum @xmath58 .
two other regions are the @xmath65 and @xmath16 .
( right ) @xmath58 plot for @xmath81 state where @xmath94.,title="fig:"]@xmath95 plot for @xmath79 state where @xmath93 .
black strip is a region in parameter space with the minimum @xmath58 .
two other regions are the @xmath65 and @xmath16 .
( right ) @xmath58 plot for @xmath81 state where @xmath94.,title="fig : " ] plot for @xmath2 state contribution where @xmath94 .
( right ) @xmath58 plot for @xmath1 state contribution where @xmath96.,title="fig:"]@xmath95 plot for @xmath2 state contribution where @xmath94 .
( right ) @xmath58 plot for @xmath1 state contribution where @xmath96.,title="fig : " ] another important issue to address in these scenarios , given the large required @xmath83 couplings , concerns vacuum stability and perturbativity of the scalar potential .
we focus here on the @xmath1 state for concreteness , while similar conclusions hold for the @xmath2 as well . a recent study of vacuum decay constraints on colored scalars coupling to the higgs @xcite included the @xmath1 representation and found that in the low energy effective theory comprising the sm and the color octet weak doublet state , and for the interesting range of color octet masses as extracted from our analysis , vacuum ( meta)stability constraints can be satisfied , provided a quartic term of the form @xmath97 is present and the couplings satisfy @xmath98 , where @xmath99 is the sm higgs quartic coupling @xmath100 and @xmath101
gev is the ew condensate .
while this shows that all involved quartic couplings can be perturbative at the ew scale , this is not necessarily true up to arbitrary high scales ( @xmath102 ) .
we study the issue using renormalization group equations ( rges ) . for concreteness
we consider the two relevant @xmath11 quartics in eqs . and .
state and the sm higgs c.f .
@xcite . ] taking into account also the effects of the top yukawa ( @xmath103 ) interaction , the rges for @xmath99 , @xmath104 and @xmath105 at one loop are given by @xmath106 we observe that the positive beta function of @xmath107 necessarily drives this coupling to large positive values at high scales . in particular , for the parameter range preferred by the fit to the higgs data and satisfying the vacuum metastability constraint , @xmath107 develops a landau pole already at scales @xmath108 tev .
such a scenario can thus only represent a low energy effective theory , which needs to be extended below @xmath109 .
one possibility without introducing new light degrees of freedom is to utilize the couplings of @xmath11 to matter fields . in particular , couplings to quarks of the form @xmath110
can induce negative contributions to the beta functions of @xmath111 and @xmath112 proportional to @xmath113 and @xmath114 , respectively . here ,
@xmath115 , @xmath116 refer to left- and right - handed quark fields , respectively .
if large enough , such contributions can , in principle , stabilize the color octet quartics .
however , @xmath117 also induce mixing among the various possible terms in the most general scalar potential involving the @xmath118 state @xcite and its complete rge study is clearly beyond the scope of the present paper .
finally we note that the presence of @xmath117 might have interesting phenomenological consequences in both flavor observables as well as for collider signatures of the color octet scalar ( c.f .
@xcite ) . a detailed study of the implications of sizable @xmath117 on the uv behavior of the @xmath118 scalar potential and the associated phenomenology is in progress .
in the sm , the @xmath4 decay is generated at the loop level in a very similar way as @xmath0 .
being extremely suppressed , this channel has not yet been explored by the experimental collaborations .
however , once it is measured , it will be an extremely useful test of possible explanations of the @xmath0 anomaly since , in general , these two rates are related .
the partial decay width for @xmath4 including colored scalars contribution is given by @xcite @xmath119 where @xmath120 and the coupling of @xmath121 to @xmath122 boson is given in units of @xmath123 , that is @xmath124 . here
@xmath125 represents the value of the weak isospin of @xmath12 and in eq .
we sum over all @xmath54 within the given weak multiplet @xmath11 .
the relevant one - loop functions are defined as @xmath126 @xmath127 @xmath128 where@xmath129 @xmath130 @xmath131 the sm contributions to @xmath132 induced by the @xmath35 boson and the top quark are proportional to @xmath133 and @xmath134 , respectively . clearly , the sm result is dominated by the @xmath35 boson contribution .
we calculate @xmath135 for the best - fit values of coupling @xmath51 and mass @xmath26 for the scalars that can accommodate current higgs data . for the @xmath1 state we predict
@xmath136 throughout the interesting range of scalar masses and couplings .
similarly , for the @xmath137 state we get @xmath138 , again independent of the scalar mass in the interesting range of parameters .
we observe that both scenarios predict a mild suppression of the @xmath4 decay rate and that a @xmath139 measurement of the relevant branching fraction could start probing the scalar color sextet explanation of the @xmath140 excess , something possibly in reach of the 14tev lhc with several hundred @xmath141 @xcite .
another interesting related process is the di - higgs production via gluon fusion , i.e. , @xmath18 , which is again loop suppressed in the sm and thus potentially sensitive to non standard contributions .
we accordingly turn our attention to this process .
the di - higgs production can be significantly affected in the presence of light colored fields .
this is especially true for the regime we explore where @xmath142 is large ( @xmath143 ) and colored state mass @xmath26 is relatively small @xcite .
we accordingly evaluate @xmath144 for the best - fit values of coupling @xmath51 and mass @xmath26 for both scalars that accommodate current higgs data in a satisfactory manner .
we use the parton level loop - induced top quark contribution towards @xmath18 amplitude @xcite as well as the loop - induced colored scalar contribution @xcite to evaluate total cross - sections @xmath145 and @xmath146 that are relevant for the current lhc energy reach .
we integrate the parton level @xmath18 cross sections using the lo mstw2008 parton distribution functions @xcite with fixed factorization and normalization scales of @xmath147 . at the 8tev c.m.s .
energy lhc we obtain for the @xmath1 and @xmath137 states an enhancement of @xmath148 and @xmath149 respectively , for the best fit regions of couplings and almost independent of the scalar masses in their interesting range .
we have checked that these values also hold for the 14tev lhc energy .
we observe that both scenarios predict significant enhancement of the total cross - section for di - higgs production via gluon fusion with respect to the sm predictions .
the signals and discovery strategies within the regime of enhanced di - higgs production at lhc have been discussed extensively in ref .
@xcite , where the relevant rates for various final states can be found for both 8tev and 14tev c.m.s . energy lhc .
our analysis singles out two particular colored scalars@xmath1 and @xmath2as potential candidates that can consistently address the observed enhancement in @xmath0 decay channel if light enough . here
we want to comment on matter unification scenarios that predict these scalars to be in a required mass range .
note that although both scalars couple directly to matter they do not mediate proton decay .
color octet state is an appealing source of new physics .
for example , it is the only scalar beside the higgs doublet that can be consistently coupled to the sm quarks within the minimal flavor violation framework @xcite .
octet production at the lhc and relevant electroweak constraint on its mass and yukawa couplings in that particular context have been extensively studied @xcite .
more recently , there have been numerous studies of the color octet influences on higgs physics in view of lhc data and potential signals @xcite .
note that in all these instances the presence of light octet state is simply assumed .
there are , however , viable unifying models that predict existence of a light color octet state and correlate it with proton decay signatures @xcite .
we discuss two particular models based on the @xmath5 gauge group @xcite in what follows .
the first model @xcite uses one @xmath150-dimensional ( @xmath151 ) and one @xmath152-dimensional ( @xmath153 ) scalar representation to accommodate charged fermion masses @xcite and a set of extra fermion fields from one @xmath154-dimensional ( @xmath155 ) representation @xcite to accommodate neutrino masses .
the phenomenology of light octet state that resides in the @xmath152-dimensional scalar representation within this particular context has been addressed in ref .
the second model uses three scalar representations@xmath151 , @xmath156 and @xmath153to accommodate all fermion masses @xcite .
these two models , beside the usual matter representations , also use one @xmath154-dimensional scalar representation ( @xmath157 ) to break @xmath5 down to @xmath158 . in both scenarios
an upper bound on octet mass is correlated with the observable partial proton decay lifetimes . in view of the latest experimental data on partial
proton decays modes @xcite , in particular on @xmath159 , we update these predictions for both models to demonstrate this correlation .
we show in fig .
[ fig:3 ] a viable unification for the model with one @xmath150-dimensional and one @xmath152-dimensional scalar representation and one @xmath154-dimensional matter representation .
numerical procedure that is used to establish unification of gauge couplings and implement proton decay constraints is described in detail in ref .
@xcite . here
, we outline the most important points that lead to results presented in fig .
[ fig:3 ] .
the scalar fields are denoted as @xmath160 , @xmath161 and @xmath162 .
the extra fermions in @xmath163 are related through a following set of mass relations @xcite @xmath164 here , @xmath165 is a dimensionless free parameter that describes the mass splitting between masses of @xmath166 and @xmath167 .
we accordingly present our findings in a @xmath165 vs. @xmath168 plane , where @xmath168 represents the scale of gauge coupling unification .
@xmath168 is maximized through numerical procedure that varies scalar and fermion masses , in accordance with mass splitting constraints of eq .
, in the following ranges : @xmath90gev@xmath169 , @xmath170gev@xmath171 and @xmath172gev@xmath173 @xcite .
solid lines in fig .
[ fig:3 ] , going from top to bottom , correspond to @xmath174gev , @xmath175gev , @xmath176tev and @xmath177tev .
horizontal dashed line is due to a constraint imposed by experimental results on proton decay through @xmath178 on unification case for @xmath174gev . note that the difference between constraints on @xmath174gev case and @xmath177tev case is practically negligible as it is dominated by the difference in values of the appropriate unified gauge coupling at the gut scale for these two cases .
the most important point is that all unification scenarios below the dashed line in fig .
[ fig:3 ] are excluded by experimental limits on @xmath178 .
the proton decay signature through @xmath178 channel is derived assuming that the yukawa matrices for matter fields are symmetric .
this assumption allows for light @xmath179 as it prevents @xmath179 to couple to a quark - quark pair @xcite . that , on the other hand , renders @xmath179 innocuous as far as proton decay constraints are concerned .
the same assumption removes dependence on unitary redefinitions of quark and lepton fields from proton decay operators induced through tree - level exchange of heavy gauge bosons .
note that the mass of @xmath180 needs to be below @xmath181gev if it is to explain the @xmath182 anomaly of muon through perturbative yukawa couplings @xcite .
we take it to be @xmath183gev to generate fig .
[ fig:3 ] .
if @xmath179 mass is closer to @xmath181gev the allowed gut scale would be slightly raised with respect to what is shown in fig .
[ fig:3 ] .
one should also worry that such a light colored state that resides in the same representation of @xmath5 as the octet field could spoil results for the satisfactory enhancement of @xmath184 signal .
for example , if @xmath185 is to couple to the higgs field with the same strength as the octet components one would need to simultaneous fit data with both fields to address the viability of such scenario .
in the most general case one would have a situation where the triplet and the octet have different couplings to the higgs boson and different masses . to generate results shown in fig .
[ fig:3 ] we update some of input parameters with regard to what is used in ref .
@xcite to produce partial decay width for @xmath178 .
we use @xmath186 @xcite , @xmath187years @xcite and @xmath188gev@xmath189 @xcite . here
@xmath190 is the relevant nucleon matrix element .
-dimensional and @xmath152-dimensional scalar representations and an extra set of fermions in @xmath154-dimensional representation in @xmath165 vs. @xmath168 plane at the one - loop level for central values of low - energy observables .
solid lines , going from top to bottom , correspond to @xmath174gev , @xmath175gev , @xmath176tev and @xmath177tev .
horizontal dashed line is due to the constraint imposed by experimental limit on partial proton decay lifetime through @xmath178 on @xmath174gev case.,width=264 ] the predicted proton lifetime for @xmath159 due to gauge mediation is at most a factor of 5 above the current experimental limit for the @xmath174gev case while the proton lifetime due to scalar mediation is already at the present limit . the gut scale in fig . [ fig:3 ]
is maximized by imposing a lower bound on proton decay mediating scalars , i.e. , @xmath191 , @xmath192 and @xmath193 , to be @xmath170gev .
it has been recently shown @xcite that the scalar exchange dominated proton decay in the models with @xmath150- and @xmath152-dimensional scalar representations with symmetric yukawa couplings to matter fields constrains the mass of @xmath191 through experimental data on @xmath194 channel to be above @xmath195 and @xmath196 in the most and least conservative case , respectively .
we use @xmath197years @xcite , while the vevs for @xmath150- and @xmath152-dimensional representations@xmath198 and @xmath199satisfy @xmath200 .
this conclusively shows that @xmath170gev is a reliable lower bound on the mass of proton decay mediating scalars for one - loop unification considerations .
we present in fig .
[ fig:4 ] a viable unification for the model with @xmath150- , @xmath201- and @xmath152-dimensional scalar representations in a @xmath202 vs. @xmath168 plane at the one - loop level for central values of low - energy observables .
we use @xmath90gev@xmath203 , @xmath170gev@xmath204 , @xmath172gev@xmath173 , where @xmath205 . in this case
the predicted proton partial lifetime for @xmath159 channel due to gauge mediation is at most a factor of 26 above the current experimental limit for @xmath206gev .
the proton lifetime due to scalar mediation , on the other hand , is at the present limit . - , @xmath201- and @xmath152-dimensional scalar representations in @xmath202 vs. @xmath168 plane at the one - loop level for central values of low - energy observables .
horizontal dashed line represents a limit due to the constraint imposed by experimental limit on partial proton decay lifetime through @xmath178.,width=264 ] the relevant coupling of the octet to the - would - be higgs field @xmath14 originates from the following set of @xmath5 contractions : @xmath207 and @xmath208 .
the couplings of the neutral and charged component of the octet @xmath209 to the - would - be sm higgs , under the assumption that the sm doublet primarily originates from @xmath150-dimensional representation , are @xmath210 and @xmath211 , respectively .
any mixing between the doublets in @xmath150- and @xmath152-dimensional representation can be easily accounted for . in any case , we need to go to the limit @xmath212 . to reproduce the setup used in section [ htogammagamma ] where
the @xmath0 excess is accounted for via custodial symmetric color octet loops one needs to assume that @xmath213 is much smaller than @xmath214 .
the color sextet @xmath215 resides in @xmath216- and @xmath217-dimensional representations of @xmath5 @xcite .
it has been shown that it can provide for gauge coupling unification within an @xmath5 framework @xcite .
it is , however , difficult to connect its lightness to proton decay or correlate its yukawa couplings with the origin of masses for matter fields without additional assumptions in that particular setup .
this seems , instead , to require an @xmath218 embedding scheme .
namely , the color sextet is part of @xmath219- and @xmath220-dimensional representations of @xmath218 .
both representations are frequently used in model building with the former one being crucial in explaining observed fermion masses .
for example , colored scalars , which are part of @xmath219-dimensional representation of @xmath218 and are light enough to be accessible at the lhc , have recently been proposed in refs .
in fact , there already exists a viable setup with an intermediate - scale sextet @xcite .
it might thus be feasible to have very light @xmath215 scalar that would originate from an @xmath218 model .
we leave it to future studies .
we have considered the impact of light colored scalars that can couple directly to matter fields on the recently observed @xmath0 excess . we find two viable scenarios where two states@xmath1 and @xmath2can individually influence the excess in a constructive way and remain in excellent agreement with all available data .
the colored states in question should have a substantial coupling to the sm higgs of order one and a mass of order @xmath3gev ( or below ) in order to explain the data .
in particular , perturbativity arguments require that the sextet and octet masses should be below @xmath87gev and @xmath88gev , respectively .
the best fit values for the colored scalar masses and couplings to the higgs are used to generate predictions for @xmath4 and di - higgs production .
we find moderate suppression of the partial decay width @xmath4 with regard to the sm value .
the di - higgs production , on the other hand , is enhanced by at least a factor of a hundred with respect to the sm prediction at both the 8 tev and 14 tev lhc .
we subsequently study extensions of the sm where these states naturally appear with primary focus on matter unification models based on @xmath5 gauge group .
it is shown that two simple models correlate light color octet mass with observable proton decay .
for the color octet mass of @xmath88gev the predicted partial proton decay lifetime through @xmath159 channel is a factor of 5 ( 25 ) above the current limit for the model with extra fermions ( scalars ) in @xmath154-dimensional ( @xmath201-dimensional ) representation . in the model with extra fermions
it is also possible to accommodate the @xmath7 anomaly , albeit through a presence of an additional light colored scalar . in conclusion
, the presented setup relates high scale matter unification and matter stability to higgs physics via the effects of light colored scalar states .
j. c. pati and a. salam , phys .
d * 10 * , 275 ( 1974 ) [ erratum - ibid .
d * 11 * , 703 ( 1975 ) ] .
j. shu , t. m. p. tait and k. wang , phys .
d * 81 * , 034012 ( 2010 ) [ arxiv:0911.3237 [ hep - ph ] ] ; i. dorsner , s. fajfer , j. f. kamenik and n. kosnik , phys .
d * 81 * , 055009 ( 2010 ) [ arxiv:0912.0972 [ hep - ph ] ] ; a. arhrib , r. benbrik and c. -h .
chen , phys .
d * 82 * , 034034 ( 2010 ) [ arxiv:0911.4875 [ hep - ph ] ] .
s. m. barr and x. calmet , arxiv:1203.5694 [ hep - ph ] .
p. fileviez perez and m. b. wise , phys .
d * 80 * , 053006 ( 2009 ) [ arxiv:0906.2950 [ hep - ph ] ] .
i. dorsner and p. fileviez perez , nucl .
b * 723 * , 53 ( 2005 ) [ hep - ph/0504276 ] .
i. dorsner and p. fileviez perez , phys .
b * 642 * , 248 ( 2006 ) [ hep - ph/0606062 ] .
i. dorsner and i. mocioiu , nucl .
b * 796 * , 123 ( 2008 ) [ arxiv:0708.3332 [ hep - ph ] ] .
p. fileviez perez , h. iminniyaz and g. rodrigo , phys .
d * 78 * , 015013 ( 2008 ) [ arxiv:0803.4156 [ hep - ph ] ] .
k. m. patel and p. sharma , jhep * 1104 * , 085 ( 2011 ) [ arxiv:1102.4736 [ hep - ph ] ] .
i. dorsner , s. fajfer , j. f. kamenik and n. kosnik , phys .
d * 81 * , 055009 ( 2010 ) [ arxiv:0912.0972 [ hep - ph ] ] .
d. c. stone and p. uttayarat , jhep * 1201 * , 096 ( 2012 ) [ arxiv:1111.2050 [ hep - ph ] ] .
s. bertolini , l. di luzio and m. malinsky , phys .
d * 85 * , 095014 ( 2012 ) [ arxiv:1202.0807 [ hep - ph ] ] .
s. babu and r. n. mohapatra , arxiv:1206.5701 [ hep - ph ] .
k. s. babu and r. n. mohapatra , arxiv:1207.5771 [ hep - ph ] .
i. dorsner , s. fajfer , j. f. kamenik and n. kosnik , phys .
d * 82 * , 094015 ( 2010 ) [ arxiv:1007.2604 [ hep - ph ] ] .
i. dorsner , j. drobnak , s. fajfer , j. f. kamenik and n. kosnik , jhep * 1111 * , 002 ( 2011 ) [ arxiv:1107.5393 [ hep - ph ] ] .
b. patt and f. wilczek , hep - ph/0605188 .
w. -f .
chang , j. n. ng and j. m. s. wu , arxiv:1206.5047 [ hep - ph ] .
j. r. espinosa , c. grojean , m. muhlleitner and m. trott , arxiv:1207.1717 [ hep - ph ] .
giardino , k. kannike , m. raidal and a. strumia , arxiv:1207.1347 [ hep - ph ] . t. corbett , o. j. p. eboli , j. gonzalez - fraile and m. c. gonzalez - garcia , arxiv:1207.1344 [ hep - ph ] ; d. carmi , a. falkowski , e. kuflik , t. volansky and j. zupan , arxiv:1207.1718 [ hep - ph ] .
b. batell , s. gori and l. -t .
wang , jhep * 1206 * , 172 ( 2012 ) [ arxiv:1112.5180 [ hep - ph ] ] .
h. georgi and s. l. glashow , phys .
* 32 * ( 1974 ) 438 .
a. djouadi , phys .
* 459 * , 1 ( 2008 ) [ hep - ph/0503173 ] .
r. barbieri , a. pomarol , r. rattazzi and a. strumia , nucl .
b * 703 * , 127 ( 2004 ) [ hep - ph/0405040 ] .
s. dittmaier _ et al .
_ [ lhc higgs cross section working group collaboration ] , arxiv:1101.0593 [ hep - ph ] .
b. a. dobrescu , g. d. kribs and a. martin , phys .
d * 85 * , 074031 ( 2012 ) [ arxiv:1112.2208 [ hep - ph ] ] .
g. d. kribs and a. martin , arxiv:1207.4496 [ hep - ph ] .
g. m. tavares and m. schmaltz , phys .
d * 84 * , 054008 ( 2011 ) [ arxiv:1107.0978 [ hep - ph ] ] ; j. drobnak , j. f. kamenik and j. zupan , arxiv:1205.4721 [ hep - ph ] .
m. reece , arxiv:1208.1765 [ hep - ph ] .
a. v. manohar and m. b. wise , phys .
d * 74 * , 035009 ( 2006 ) [ hep - ph/0606172 ] .
m. gerbush , t. j. khoo , d. j. phalen , a. pierce and d. tucker - smith , phys . rev .
d * 77 * , 095003 ( 2008 ) [ arxiv:0710.3133 [ hep - ph ] ] .
m. carena , i. low and c. e. m. wagner , arxiv:1206.1082 [ hep - ph ] .
j. s. gainer , w. -y .
keung , i. low and p. schwaller , arxiv:1112.1405 [ hep - ph ] .
e. w. n. glover and j. j. van der bij , nucl .
b * 309 * , 282 ( 1988 ) .
a. d. martin , w. j. stirling , r. s. thorne and g. watt , eur .
j. c * 63 * , 189 ( 2009 ) [ arxiv:0901.0002 [ hep - ph ] ] .
m. i. gresham and m. b. wise , phys .
d * 76 * , 075003 ( 2007 ) [ arxiv:0706.0909 [ hep - ph ] ] .
j. m. arnold and b. fornal , phys .
d * 85 * , 055020 ( 2012 ) [ arxiv:1112.0003 [ hep - ph ] ] .
y. bai , j. fan and j. l. hewett , arxiv:1112.1964 [ hep - ph ] .
k. kumar , r. vega - morales and f.
yu , arxiv:1205.4244 [ hep - ph ] .
p. fileviez perez , phys .
b * 654 * , 189 ( 2007 ) [ hep - ph/0702287 ] .
h. georgi and c. jarlskog , phys .
b * 86 * , 297 ( 1979 ) .
b. bajc and g. senjanovic , jhep * 0708 * , 014 ( 2007 ) [ hep - ph/0612029 ] .
m. miura , pos ichep * 2010 * , 408 ( 2010 ) .
h. nishino _ et al . _
[ super - kamiokande collaboration ] , phys .
d * 85 * , 112001 ( 2012 ) [ arxiv:1203.4030 [ hep - ex ] ] .
i. dorsner , s. fajfer and n. kosnik , arxiv:1204.0674 [ hep - ph ] .
j. beringer _
( particle data group ) , phys .
d * 86 * , 010001 ( 2012 ) . | we consider the impact of colored scalars that can couple directly to matter fields on the recently measured @xmath0 excess . among all possible candidates only scalar states transforming as @xmath1 and @xmath2 under the standard model gauge group can individually accommodate the excess and remain in agreement with all available data .
current experimental constraints require such colored states to have an order one coupling to the standard model higgs and a mass below @xmath3gev .
we use the best fit values to predict the correlated effect in @xmath4 and di - higgs production .
we furthermore discuss where and how these states appear in extensions of the standard model with primary focus on scenarios of matter unification .
we revisit two simple @xmath5 setups to show that these two full - fledged models not only accommodate a light color octet state but correlate its mass with observable partial proton decay lifetimes . |
the concept of hierarchy has been used for systems described with networks for a significant time and for different purposes .
it has been considered in context of elections @xcite , investigation of road network geometry @xcite and neural networks @xcite .
more recently , the concept of hierarchy has been used for social systems @xcite , artificial neural networks @xcite , financial markets @xcite and communication networks @xcite .
it has also been used on a more theoretical level , to distinguish and explain specific network properties @xcite .
while the concept of hierarchy has been extensively used , it is continuously lacking a definition .
there are papers that attempt to remedy that by introducing more formal definition and measures @xcite .
the paper by corominas - murtra @xcite introduces definition of causal graphs and perfect hierarchical structure , and introduces measure of hierarchy that is based on similarity of a given graph to a perfect hierarchy .
the paper by mones et .
@xcite distinguishes between three types of hierarchy : order ( as in @xcite ) , nested ( as in @xcite ) and flow ( as in that paper and @xcite ) .
the order hierarchy is a simple ordering of elements of a set along one dimensional axis .
the nested hierarchy is tied to the community structure and multi - scale organization of the network .
the flow hierarchy is basically a causal structure , with ( using tree analogy ) the `` root '' being the origin of all signals or flows , and `` leaves '' being simply receivers . in this paper , we focus on the flow hierarchy . as pointed out in @xcite order hierarchy can be seen as simple flow hierarchy , and nested hierarchy can be represented by a flow hierarchy . in our work , we intend not to make another definition of hierarchy , but to measure a single aspect of it
its depth .
we introduce two depth measures for directed networks and discuss their behavior and differences between them .
we also show how the existence of cycles can be taken into account and how they influence depth measures .
we define depth measures for directed networks , that are composed of vertices connected by arcs ( directed edges ) . in general
, they are applicable to any directed network , even if it contains cycles .
the depth is defined in the context of flow hierarchies , and is therefore most meaningful for networks that have a certain overall direction of all arcs .
the _ rooted depth _ is a value defined for each vertex in a network , and is the distance of that vertex from a root .
use of this definition requires a fixed root vertex .
let us start simple and first consider a tree , where root is well defined .
we are considering directed networks and consider tree where arcs are all directed from root and towards leaves .
the root is a vertex with only outgoing arcs , while leaves have only incoming arcs .
each vertex @xmath0 is at certain depth @xmath1 in the network , which we define as the distance from root to the vertex @xmath2 .
this means that all vertices are organized into distinct depth levels .
+ let us now generalize to any graph with a specified root , which we will call _ rooted network_. the root is the only vertex that has only outgoing arcs , and no incoming arcs ( @xmath3 ) .
leaves are vertices without outgoing arcs , but some incoming arcs .
the definition remains same without any changes , the distance being the length of shortest path . because of the existence of multiple paths of different lengths , it is possible however , that an arc connects higher depth vertex with lower depth vertex ( figure [ ksjh_fig_simplehierarchy ] ) .
because the definition uses shortest path , it effectively ignores any arcs that are not part of a shortest path .
this definition works for networks with loops and effectively ignores them .
it is easy to understand , as at least part of the loop has to be not a shortest path from root to somewhere , so it simply is ignored .
+ if we define any vertex that has no incoming arcs as a root , we may treat any directed graph as a sum of several rooted networks ( figure [ ksjh_fig_acyclic ] ) .
each root vertex has a part of network reachable form it , forming its own _ rooted component _ , which may encompass whole network , but it not necessarily does .
given multiple roots , a given vertex @xmath0 has several possibly different depths , depending which root we will look at .
reasonable solution is to average the depths from different roots to obtain a single depth value for a vertex .
additional issue comes from fact , that existence of loops means that no roots may exist in the whole network . in such case , the definition does not work , unless we perform loop collapse ( see below ) . + the _ relative depth _ is defined as a value @xmath1 assigned to each vertex @xmath0 , such that any arc @xmath4 always connects from a vertex with lower depth @xmath1 to a vertex with higher depth @xmath5 ( @xmath6 ) .
this basically means that we treat the arcs as a `` has depth lesser than '' relation on the set of vertices of the graph , and thus define the depth levels for vertices .
such definition is ambiguous , and to give precise values , we add that the difference must be minimal but at least @xmath7 ( @xmath8 ) .
depth values assigned according to such definition correspond to difference between vertex depths of @xmath0 and @xmath9 being equal to longest directed path between them , provided such path exists ( figure [ ksjh_fig_alternative ] ) . in absence of the path , the difference can be any value .
note that since the values are derived from relations , there is no set `` zero level '' and the values can be set relative to an arbitrary reference value . for convenience ,
it is reasonable to set vertex with least depth to @xmath10 , thus recovering non - negative depth levels like in the rooted depth definition .
the definition of relative depth sets a strict inequality relations between vertices that have directed paths between them , meaning that they can indirectly interact in some way if the network is treated as a topology of interactions .
the definition works only for acyclic graphs .
any directed loop is unresolvable , just like an inequality @xmath11 does not have any solution .
+ we can extend the definition to graph containing cycles if we force all vertices belonging to a given cycle to have the same depth value .
this means each cycle is at fixed depth level , regardless of the directed arcs it contains .
this is however a logical rule
if all vertices can influence all other vertices in the set , all should be at same depth .
effectively , from the point of view of the definition , a cycle is a single complex vertex .
we define _ loop collapse _ to be a process , where all cycles are substituted with a complex single vertices ( see figure [ ksjh_fig_collapse ] ) .
note that the loop collapse is virtual
the vertices are not really collapsed , they are only treated so for the single purpose of assigning the depth level .
all vertices that are component of a collapsed loop are assigned same depth .
if all loops are collapsed , then we recover a directed acyclic graph , and relative depth definition can be used to calculate depth levels .
note that it is possible to use loop collapse independently from relative depth definition , for example using it alongside rooted depth definition , in which case it changes the depth levels .
+ both definitions attempt to measure the vertical size of the hierarchy or a directed network . despite aiming for the same ,
they significantly differ from each other on several points.// the _ information reduction _ is present in both cases .
rooted depth uses only shortest path , which means that it effectively ignores any arcs that are not part of a shortest path , which essentially reduces the whole network to a spanning tree of shortest paths .
the relative depth on the other hand effectively ignores arcs that are not part of a longest path .
relative depth definition essentially takes into account only arcs that are part of transitive reduction of a relation defined by arcs ( which by definition of relative depth is transitive ) .
all `` shortcuts '' are ignored , often including the shortest paths .
+ both definitions are different in the aspect of _
locality_. rooted depth is defined globally and does not describe the local situation consistently depth levels viewed from local perspective of single vertex may be completely not related to the graph structure .
the only rule is that outgoing arcs of given vertex @xmath0 always connect to vertices with depth no larger than @xmath12 and incoming arcs are form vertices with depths no smaller than @xmath13 .
the relation between depth levels does not otherwise reflect arc structure , but the depth level of @xmath0 accurately describes how close the vertex is to the root .
moreover the root is always at known , fixed depth level @xmath10 . on the other hand relative depth
is defined locally and as such is relevant to local structure .
arcs always go from lesser to greater depth levels , although the actual values are effect of global topology .
the depth values do not reflect how close to the eventual root ( the `` top '' ) is the vertex .
a vertex at depth @xmath14 could be connected directly from the root . moreover ,
the depth of a root is not fixed , and if several roots exist , may be found at any depth .
overall , the rooted depth describes the global position more accurately , but fails locally .
the relative depth vice versa .
+ _ vulnerability _ of both definitions are also different .
rooted depth requires existence of root vertices and can not work without them , while the definition of relative depth does nt work in presence of directed loops . in both cases
the loop collapse allows them to work , but nonexistence of root is rare for a network that would not be almost wholly loop - collapsed into single vertex ( at least for random graphs ) .
both definitions of depth are volatile in regard to addition or removal of arcs .
a single removal or addition could change depth levels of significant parts of the whole network . adding new links can only decrease depth levels for rooted depth , and increase them for relative depth , and vice versa .
+ the rooted depth can take into account the _ multiple aspects _ of depth .
if multiple roots are present , then the depth levels relative to each root can be treated separately , not averaged into a single number , or alternatively averaged with certain weights depending on the root or its rooted component .
if the depth is considered in the context of a hierarchy , this can differentiate between positions within different hierarchies .
relative depth can not distinguish multiple aspects , as it has to be defined with a single number for each vertex .
it is however _ consistent _ , as that number is relatively well defined , without multiple aspects and issues tied to averaging .
overall , both definitions have their strong and weak points . as they describe the complex topology by simple numbers , they reduce information in different ways , which yields different properties .
the choice of depth measure will therefore depend on the aspects of the system being investigated and what one wishes to describe with it .
we have investigated the depth measure values in directed random networks .
we investigated a directed erdos - renyi graph , where probability of an arc existing between an ordered pair of vertices @xmath4 is constant @xmath15 .
opposite arcs ( e.g. @xmath4 and @xmath16 ) can exist and their existence is decided independently .
note that the parameter @xmath17 is a directed degree
@xmath18 means that on average vertices have @xmath19 outgoing arcs as well as @xmath19 incoming arcs .
if we looked at those links as undirected the total degree would be twice as high .
+ first , we attempt to investigate most general properties of the depth depending on the network parameters .
the average depth @xmath20 is a very simple measure , allowing us to look across different network parameters .
we are interested in the total depth of the network however , not necessarily the average value of @xmath21 over all vertices .
similar to physical depth , being distance from surface to bottom , not some average over a volume , we decide to measure @xmath20 as average od depth level over leaves in the network .
for rooted depth and single - root network , this means @xmath22 where @xmath23 is the number of leaves , the sum over @xmath0 is over all leaves and @xmath2 is the path length from the root @xmath24 to leaf @xmath0 .
if there is more than one root , the depth level of vertices is average over all roots _ it has directed paths from _ , which means @xmath25 where @xmath23 is the number of leaves , the sum over @xmath0 is over all leaves , @xmath26 is number of roots that have path to leaf @xmath0 , the sum over @xmath24 is over all roots that have path to leaf @xmath0 and @xmath2 is the path length from the root @xmath24 to leaf @xmath0 .
the averaging od depth level for each vertex makes sense when we are investigating that single vertex .
if we are looking at whole network however , it is more reasonable to look at pairs root - leaf and averaging over that .
this would yield a definition @xmath27 where @xmath28 is number of pairs root - leaf connected by directed path , the sum is over all pairs @xmath29 that are connected by a path and @xmath2 is the lengh of that path .
equations [ ksjh_eq_depth1 ] and [ ksjh_eq_depth1pair ] define slightly different values . in the first equation
, if vertex is under small number of roots , the depth levels generated by them are contributing more to the total compared to a vertex under large number of roots , which averages over more different depths .
if all vertices have directed paths from all roots , then both equations are equivalent as @xmath30 and therefore it can be extracted from under the sum over @xmath0 and @xmath31 is all pairs have paths between them , thus recovering equation [ ksjh_eq_depth1pair ] . as we are calculating a global property
, we decided to use the definition described by equation [ ksjh_eq_depth1pair ] as more appropriate on the global level .
+ the dependence of average _ rooted depth _ @xmath20 on average node degree @xmath17 is shown at figure [ ksjh_fig_erdos1 ] .
note that in case where no roots or leaves could be found , we assumed depth equal to 0 for practical reasons .
we explain the results at figure [ ksjh_fig_erdos1 ] as follows .
below @xmath32 , the network is not percolated and consist of single vertices ( @xmath33 ) and small clusters ( responsible for increasing value of @xmath20 with @xmath17 ) . loops are very rare . at @xmath32
the network becomes percolated ( note that this is directed percolation ) , creating the giant cluster .
this corresponds with the maximum of the average depth level .
the average depth is tied ot the diameter of the giant cluster as well as the amount of still present disconnected clusters . above @xmath32 adding more arcs means creating loops , since the network is already percolated .
loops introduced randomly cause decrease of the component diameter and thus depth , which corresponds to the first drop of @xmath20 after the peak . increasing number of loops also may cause situations where no roots or no leaves exist in the network , causing the average depth to drop towards @xmath10 ( the assumed value for network without roots or leaves ) .
this starts to happen only at high @xmath17 and is responsible for second drop of the @xmath34 ( around @xmath35 for @xmath36 ) .
figure [ ksjh_fig_topology ] shows how example network looks like at different @xmath17 .
+ it is notable that the shape of the dependence of network depth on density shown on figure [ ksjh_fig_erdos1 ] is very similar to global reaching centrality measure introduced by mones et.al .
@xcite and calculated for directed random graphs @xcite .
the similarity can be explained , by both depth and grc being measures for hierarchies .
below percolation threshold , the graph consists of many small clusters , which yield very low average depth , at the same moment giving very low grc due to narrow local reaching centrality distributions . significantly above percolation threshold
, the presence of many loops mean that the network has large strongly connected component , that has relatively low depth as well as again producing narrow lrc distribution and in effect low grc . at percolation threshold ,
the giant connected component is critically sparsely connected , meaning no loops and in effect a high depth ( long shortest paths ) .
on other hand , without loops the vertices have a broad distribution of lrc , resulting in high grc .
both measures achieve high values for a large , extended directed network without loops , which explains the correlation .
+ because the rooted depth can not be resolved in case there are no roots or leaves in the network , and also to have better comparison between rooted depth and further relative depth , we decided to also investigate rooted depth while using the loop - collapse .
the results are somewhat different and found at figure [ ksjh_fig_erdos2 ] .
the main differenc is the plateau after the peak , with value @xmath37 .
this corresponds to a network with a dense `` core '' containing many interleaving loops , and thus collapsed to a single complex vertex , with single vertex `` roots '' and `` leaves '' sticking out of this core . without the loop - collapse ,
the loops were partially converted to `` vertical '' and assigned some differing depths ( differently from perspective of each root ) , the denser the core , the smaller depths . with loop - collapse ,
the value of @xmath20 for network with the dense core is fixed at @xmath38 , resulting in plateau .
note that since loop - collapse virtually removes all loops , there are always roots and leaves and in worst case the sole remaining complex vertex is both root and leaf and thus has @xmath33 .
also note that all the values ( @xmath20 , @xmath17 , @xmath39 ) are calculated for the real , full network .
loop - collapse is used only to calculate depth levels .
+ the results are similar , although there are differences .
below @xmath32 , the behavior is the same , as lack of loops means loop - collapse does nt do anything . after the maximum
, the average depth @xmath20 declines much faster , because addition of loops causes loop - collapse , thus decreasing the effective size of the giant component much quicker .
the value then stabilizes at @xmath37 , because the typical topology of loop - collapsed network at this point consists of single complex network representing the majority of original vertices , with few single roots and leaves directly attached to it . as
the @xmath17 becomes even higher , the network starts to entirely collapse to single complex network , causing @xmath20 to approach @xmath10 for very high @xmath17 .
this explanation is reinforced by the size of the loop - collapsed network , that is very close to original size below percolation threshold @xmath32 , and then exponentially decreases towards single vertex ( figure [ ksjh_fig_size1 ] ) .
the measurement of _ relative depth _ can be performed similarly .
since we need to use loop - collapse , we always recover roots and leaves .
if our relative depth levels are scaled so that the lowest value is @xmath10 ( similarly to how roots in rooted depth have depth @xmath10 ) , we can calculate average depth of leaves ( the `` bottom '' ) and treat that as the depth of the network @xmath40 where @xmath41 is average over leaves .
note however , that unlike in rooted depth , roots ( the `` surface '' ) can be at different depth levels themselves .
we can thus calculate the depth of whole network as difference between average depth of leaves and roots ( the `` bottom '' and `` surface '' ) @xmath42 .
the first equation ( [ ksjh_eq_depth2r ] ) describes the depth of the network globally , taking into account the extreme value ( for roots ) .
the second equation ( [ ksjh_eq_depth2 ] ) is a more locally relevant , showing what is the likely depth in a part of the network .
it could be compared to rooted depth where we only take `` highest '' root into account and normal rooted depth that considers all roots .
both equations yield similar results , presented in figure [ ksjh_fig_erdos3 ] .
in general , it can be seen that relative depth behaves similarly to rooted depth .
for small @xmath17 the value is low .
the peak is around percolation threshold , corresponding to the maximally stretched treelike structure of the giant component , and then the decline of average depth correspond to the gradual increase in density and number of loops , which flattens the network when if comes to depth ( remember we use loop - collapse ) .
the one difference between rooted and relative depth discrenible on the plots is the absence of the sudden drop for higher @xmath43 .
this can be explained by the fact , that when we could nt find either roots or leaves in rooted depth calculation , we assumed total depth equal to @xmath10 . if we look at figure [ ksjh_fig_topology ] , we see that in the `` core with roots and leaves '' stadium , loss of last root effectively makes the value @xmath10 ( the assumed value ) , despite one arc earlier having depth @xmath38 or even more .
this accounts for the fast drop of depth for rooted depth not using loop - collapse .
since relative depth always uses loop collapse , the `` core with roots and leaves '' would simply change into depth @xmath7 structure , where either the `` core '' would assume role of leaf / leaves ( loss of last leaf ) or role of root ( loss of last root ) .
the transition is smooth , thus no sudden drops on the graphs for relative depth could be found . trying to describe the behavior of the depth measures for different network parameters we also investigated the height of the @xmath20 peak for different network sizes and different definitions .
figure [ ksjh_fig_peaks ] shows results for rooted depth of networks without loop - collapse , with loop - collapse and for relative depth .
we found power dependence od rooted depth on size , with exponent @xmath44 to @xmath45 ) .
we expect these values to be related to the diameter ( maximum shortest path length ) of the percolation cluster in random graph .
the relative depth peaks also behave as a power of network size , albeit with significantly lower exponent @xmath46 .
we have defined two depth measures for flow hierarchies , generalized to any type of directed network . _ rooted depth _ is defined as shortest path from one of network s roots , while _ relative depth _ is defined through relations between vertices and is effectively equal the longest path from the root .
the differences between these measures are discussed rooted depth ignores arcs that are not shortest paths , while relative depth ignores arcs that are not longest paths .
rooted depth has more global meaning , while relative one is more meaningful locally .
we have investigated the two measures on a random erdos - renyi networks , showing how they behave on purely random network , thus allowing better understanding of values obtained in other types of networks .
both measures behave similarly to a global reaching centrality measure of how hierarchical the network is . +
the research leading to these results has received funding from the european union seventh framework programme ( fp7/2007 - 2013 ) under grant agreement no 317534 ( the sophocles project ) .
00 a. shimbel , communication in a hierarchical network , bulletin of mathematical biophysics 14 ( 1952 ) 141 - 151 .
a. okabe , h. yomono , statistical methods for evaluating the geometrical hierarchy of a network , geographical analysis 20 ( 1988 ) 122 - 139 .
willcox , understanding hierarchical neural network behaviour : a renormalization group approach , journal of physics a 24 ( 1991 ) 2655 - 2664 .
e. bonabeau , g. theraulaz , j .- l .
deneubourg , phase diagram of a model of self - organizing hierarchies , physica a 217 ( 1995 ) 373 - 392 .
sousa , d. stauffer , reinvestigation of self - organizing social hierarchies , international journal of modern physics c 11 ( 2000 ) 1063 - 1066 .
d. stauffer , phase transition in hierarchy model of bonabeau , international journal of modern physics c 14 ( 2003 ) 237 - 239 .
gallos , self - organizing social hierarchies on scale - free networks , international journal of modern physics c 16 ( 2005 ) 1329 - 1336 .
s. valverde , r.v .
sole , self - organization versus hierarchy in open - source social networks , physical review e 76 ( 2007 ) 046118 .
y. yamashita , j. tani , emergence of functional hierarchy in a multiple timescale neural network model : a humanoid robot experiment , plos computational biology 4 issue 11 ( 2008 ) 1000220 m. tumminello , f. lillo , r.n .
mantegna , correlation , hierarchies , and networks in financial markets , journal of economic behavior & organization 75 ( 2010 ) 4058 .
y. wang , m. iliofotou , m. faloutsos , b. wu , analyzing interaction communication networks in enterprises and identifying hierarchies , proceedings of the 2011 ieee 1st international network science workshop , nsw 2011 ( 2011 ) , 17 - 24 ( art .
6004653 ) e. ravasz , a .-
barabasi , hierarchical organization in complex networks , physical review e 67 ( 2003 ) 026112 .
a. trusina , s. maslov , p. minnhagen , k. sneppen , hierarchy measures in complex networks , physical review letters 92 ( 2004 ) 178702 .
e. mones , l. vicsek , t. vicsek , hierarchy measure for complex networks , plos one 7(3 ) ( 2012 ) e33799 .
doi:10.1371/journal.pone.0033799 b. corominas - murtra , c. rodriguez - caso , j. goni , measuring the hierarchy of feedforward networks , chaos 21 ( 2011 ) 016108 .
e. mones , hierarchy in directed random networks , physical review e 87 ( 2013 ) 022817 . | we explore depth measures for flow hierarchy in directed networks .
we define two measures rooted depth and relative depth , and discuss differences between them .
we investigate how the two measures behave in random erdos - renyi graphs of different sizes and densities and explain obtained results .
hierarchy , complex networks , depth , directed graph , random graphs |
despite the great success of the standard model ( sm ) , it offers no explanation of the origin of fermion flavor , or of the existence of three families .
the flavor physics of the third family is particularly mysterious for the smallness of the mixing angles and the huge hierarchy in masses . in this letter
, we consider a model @xcite based on the gauge symmetry g= @xmath21 .
the third generation of fermions ( top quark , @xmath22 , bottom quark , @xmath23 , tau lepton , @xmath24 , and its neutrino , @xmath25 ) experiences a new gauge interaction , instead of the usual weak interaction advocated by the sm . on the contrary
, the first and second generations only feel the weak interaction supposedly equivalent to the sm case .
the new gauge dynamics is attributed to the @xmath26 symmetry under which the left - handed fermions of the third generation transform in the fundamental representation ( doublets ) , while they remain to be singlets under the @xmath27 symmetry . on the other hand , the left - handed fermions of the first and second generation
transform as doublets under the @xmath27 group and singlets under the @xmath26 group .
the @xmath28 group is the sm hypercharge group .
the right - handed fermions only transform under the @xmath29 group as assigned by the sm . finally the qcd interactions and the color symmetry @xmath30
are the same as that in the sm .
the spontaneous symmetry - breaking of the group @xmath31 is accomplished by introducing two scalar matrix fields @xmath32 and @xmath33 , which are assigned as @xmath34 and @xmath35 .
( the scalar doublet @xmath33 carries equivalent quantum numbers as the sm higgs doublet . )
we assume that the first stage of symmetry breaking is accomplished through the @xmath36 field by acquiring a vacuum expectation value @xmath37 , i.e. , @xmath38 the second stage is through the scalar @xmath39 field by acquiring a vacuum expectation value @xmath40 , so @xmath41 , where @xmath40 is at the same order as the sm symmetry - breaking scale .
because of this pattern of symmetry breaking , the gauge couplings are related to the @xmath42 gauge coupling @xmath43 by the relation @xmath44 we can define @xmath45 where @xmath46 is the usual weak mixing angle and @xmath47 is a new parameter of the model . to derive the mass eigenstates and physical masses of the gauge bosons , we need to diagonalize their mass matrices . for @xmath48 ( equivalently @xmath49 )
, we require @xmath50 ( which implies @xmath51 ) so that the perturbation theory is valid .
similarly , for @xmath52 , we require @xmath53 . for simplicity
, we focus on the region where @xmath54 is much larger than 1 , and ignore the corrections which are suppressed by higher powers of @xmath55 . to the order @xmath55 , the light gauge boson masses are found to be @xmath56 where @xmath57 . while for the heavy gauge bosons , one finds @xmath58 up to this order , the heavy gauge bosons are degenerate in mass because they do not mix with the hypercharge gauge boson field .
the first and second generation fermions acquire their masses through the yukawa interactions to the @xmath39 doublet field , while the third generation fermion masses have to be generated through higher dimension operators .
given the fermion mass matrices , one can derive their physical masses by diagonalizing the mass matrices using bilinear unitary transformations .
for example , in terms of the fermion mass eigenstates , we have the interaction terms : @xmath59 \left ( \begin{array}{c } d_l \\ s_l \\ b_l \end{array }
\right ) w_\mu ^{+}\,+h.c.,\end{aligned}\ ] ] @xmath60 \left ( \begin{array}{c } d_l \\
s_l \\ b_l \end{array } \right ) z_\mu \ , , \label{qu2}\ ] ] where @xmath61 and @xmath62 are unitary matrices needed to diagonalize the up and down quark matrices , respectively , and @xmath63 for the charged - current interactions in the quark sector , one observes that in the case of ignoring the new physics effect , quark mixing is described by a unitary matrix @xmath64 which is identified as the usual cabibbo - kobayashi - maskawa ( ckm ) mixing matrix . with the inclusion of new physics
, the mixing acquires an additional contribution proportional to @xmath65 , with @xmath66 therefore , we would expect the extracted values of the ckm matrix elements to be slightly modified due to the new contributions of the model . in this model
, lepton mixing is an exciting possibility .
however , the almost vanishing partial decay widths of @xmath67 and @xmath68 have severely suppressed any possible mixing between the first and second family leptons .
since the lepton number violation processes involving the third family are not as well constrained , and the mixing strength between leptons may be directly related to their masses , we assume in this paper that lepton flavor mixing is only allowed between the second and the third family .
the strength of the lepton flavor mixing in this model is parametrized by @xmath69 .
the minimal mixing corresponds to @xmath70 , and a maximal mixing implies @xmath71 .
furthermore , flavor mixing can exist in the neutrino sector despited that neutrinos are massless , which may induce an interesting effect to the neutrino oscillation phenomena .
as to be shown later , quark mixing must be present in the model , although lepton mixing is merely a possibility . hence , flavor - changing neutral currents ( fcncs ) are unavoidable in the quark sector . in the following ,
we discuss the effect of the new physics predicted by this model to low - energy experiments , as well as the constraints derived from the present data on the parameter space of the model .
following ref .
@xcite , we calculate the changes in the relevant physical observables relative to their sm values to leading order in @xmath55 .
namely , any observable @xmath72 can be written as @xmath73 where @xmath74 is the sm prediction including the one - loop sm correction , and @xmath75 represents the new physics effect to leading order in @xmath55 . using the most recent lep and slc measurements @xcite ( including the total width of the @xmath1 boson @xmath76 , @xmath77 , @xmath78 , @xmath79 , the vector @xmath80 and the axial - vector @xmath81 couplings of the electron , the ratios @xmath82 , @xmath83 , @xmath84 , @xmath85 , @xmath86 , @xmath87 , @xmath88 , @xmath89 , the hadronic cross section @xmath90 , @xmath91 , @xmath92 , @xmath93 , @xmath94 , and @xmath95 ) , and the sm predictions of the above observables @xcite , we find that the precision @xmath1-pole data requires the heavy gauge boson mass to be larger than about @xmath96 tev at the @xmath3 level , and the parameter @xmath97 to be larger than 20 for @xmath98 . in figure 1 ( solid curve )
we show the minimal @xmath99 mass as a function of @xmath100 at the @xmath3 level , assuming no mixing in the lepton sector .
the @xmath1-pole observable that imposes the most stringent constraint on the model is @xmath79 , which is the ratio of the partial decay widths of @xmath1 boson into the @xmath101 and the hadronic modes .
the measurement of @xmath76 also plays an important role secondary to @xmath79 , especially for small @xmath100 , due to the high precision of data .
we find that @xmath97 can be as small as 20 for the smallest allowed value of @xmath100 ( @xmath102 ) , and it increases as @xmath100 increases .
furthermore , the quantity @xmath103 is constrained by data to be less than about @xmath104 for a large range of @xmath100 . as also shown in figure 1 ( dotted curve ) , without including the leptonic observables from the @xmath1-pole data , the bounds on @xmath105 and @xmath97 is about 900 gev and 5 , respectively .
( in this case , the important constraint is coming from the measurement of @xmath91 . )
if a mixing between @xmath5 and @xmath4 is allowed , we also need to include the limit on the branching ratio of @xmath106 from lep / slc data @xcite . in figure 1 ( dashed curve )
, we depict the new constraints on @xmath100 and @xmath105 for @xmath107 .
we find that the lower limit on the heavy mass is reduced to @xmath108 tev .
the reason for this slightly lower bound is due to the reduced non - universal effect in @xmath79 .
we find that for the smallest allowed value of @xmath109 , the value of @xmath97 can be as low as 20 , which is similar to the case of @xmath110 .
the reason is that for small values of @xmath100 ( @xmath111 ) , the measurement @xmath76 , which is independent of the mixing angle @xmath112 , plays the important role in constraining the parameter @xmath97 .
again , the quantity @xmath103 is found to be less than about @xmath113 .
even though the @xmath1-pole data already impose significant constraints , this model has a rich structure that can be further examined at much lower energy scales . in the following sections
, we would like to examine those constraints obtained from the low - energy hadronic , leptonic , and semi - leptonic data .
we will concentrate on the very low - energy regime , i.e. , physics at zero - momentum transfer , and examine whether the parameters of the model can be better constrained than those imposed by lep and slc data . for clarity
, we shall separately discuss below the effects from the lepton and quark sectors to lepton number violation phenomena , as well as to kaon and bottom physics .
it is easy to see that if the third family lepton does not mix with the first family lepton , then the partial decay width of @xmath114 and @xmath115 will not be modified .
however , if the lepton mixing involves the third family , then the lifetime of the @xmath24 lepton will be modified .
the only modification to the total decay width , to the order @xmath55 , is coming from the partial decay width @xmath116 .
we find @xmath117 where @xmath118 is a phase factor @xcite given by @xmath119 after comparing with data @xcite , we find that the above ratio constrains the parameter @xmath120 . also the lepton number violation process @xmath121 provides the constraint @xmath122 consistent with the above measurement .
therefore , the above two measurements give a stronger constraint than the @xmath1-pole data for @xmath123 . on the other hand , giving its present experimental limit , the decay process @xmath124 has not yet played a significant role .
nevertheless , if the above discussed decay widths can be measured to a better accuracy in future experiments , they can further test the proposed model .
hereafter , when discussing the predictions of our model to low - energy processes , we will consider @xmath125 for @xmath126 , and @xmath127 for @xmath107 . to completely describe the interactions of gauge bosons and quarks , it requires two mixing matrices @xmath61 and @xmath62 because both the up- and down - type quarks are massive .
as noted in eq .
( [ qu2 ] ) that the neutral - current mixing matrices ( @xmath128 and @xmath129 ) are related to the charged - current mixing matrix ( @xmath130 ) .
because of the experimental evidence of the ckm matrix in charged currents , fcncs must occur in this model between the interaction of quarks and gauge bosons , and can be realized in three different ways : ( i ) in the down - quark sector only , ( ii ) in the up - quark sector only , and ( iii ) in both sectors .
all these three possibilities have to confront the existing low - energy data . in the following , we investigate these three possibilities , separately .
first , we consider the case that only down - type quarks can mix , so that @xmath131 , and @xmath132 .
this implies that fcnc processes are completely determined by the ckm matrix @xmath133 , in addition to @xmath134 and @xmath97 .
since the matrix elements of @xmath135 are naturally small , we generally do not expect large effects in fcnc processes . under the assumption that the @xmath4 lepton does not mix with the electron @xmath43 ,
the decay width of @xmath136 is not modified at tree level .
however , @xmath137 will be modified through the fcnc interaction @xmath138 . the expected branching ratio , normalized to the predicted branching ratio for @xmath139 , can be written as @xmath140 for @xmath141 , we find that @xmath142 is a few times larger than the sm prediction . in the above result we summed over all neutrino flavors , therefore , the @xmath112 dependence cancels .
furthermore , since the above observables do not depend on the parameter @xmath100 , they can directly constrain the parameter @xmath97 of the model .
the effect to the @xmath143-@xmath144 mixing is of the same order as the sm prediction , which can only be useful if the long distance contribution can be better understood theoretically .
this model also predicts non - standard effects for bottom physics . the important process to consider is @xmath145-@xmath146 mixing where new effect is expected to occur at tree level . after substituting all the relevant variables by their numerical values , we find @xmath147 given the constraint on @xmath97 ( @xmath148 ) imposed by the @xmath1-pole data , without lepton flavor mixing ( i.e. , @xmath149 )
, the new contribution can be a few times of the sm predictions . in the case
that @xmath4 and @xmath5 mix with the maximal strength ( i.e. , @xmath150 and @xmath151 ) , the new contribution to @xmath145-@xmath146 mixing is expected to be less than @xmath152 .
furthermore , our model predicts tree - level contribution to the process @xmath153 , whose branching ratio , when normalized by @xmath154 , is given as @xmath155 where @xmath156 and @xmath157 @xcite .
the typical size of the predicted branching ratio is listed in table 1 . another interesting process to consider is the decay @xmath158 . at tree level ,
the decay rate is given by @xmath159 using the values @xmath160 , @xmath109 , @xmath161 gev , and @xmath162 gev @xcite , we find that the current data on the branching ratios of @xmath163 does not impose a better constraint on the model than that by the @xmath1-pole data .
since this model also predicts the non - sm decay modes , such as @xmath164 and @xmath165 , with comparable branching ratios , they should be measured to test the model prediction on the lepton flavor mixing dynamics ( i.e. , @xmath112 dependence ) . as a summary to this scenario , in table 1
we tabulate the predictions of our model for various processes in two cases : ( i ) @xmath126 and @xmath141 , ( ii ) @xmath107 and @xmath166 . for both cases we set @xmath109
. under the scenario that only up - type quarks can mix , there will be no non - standard effect present in the hadronic decays of @xmath10 , @xmath167 and @xmath11 mesons .
this is because in the pure hadronic interaction , the new physics effect is only expected in processes that @xmath168 vanishes , i.e. , the number of @xmath23 quark must be the same in the initial and final states of a reaction .
it is easy to see that new effects in the charged - current semi - leptonic decays are only expected in the @xmath23-quark system . as an example
, it can modify the decay branching ratio of @xmath169 and @xmath170 to be @xmath171 where all the three lepton ( including neutrino ) flavors are included .
( note that the @xmath112 dependence cancels . ) however , with @xmath172 , imposed by the @xmath1-pole data , we do not expect the new physics effects to exceed @xmath173 .
furthermore , the unitarity condition of the ckm matrix will be modified , but its change is extremely small for the values of @xmath97 that agree with @xmath1-pole data . at one - loop level ,
the @xmath143-@xmath144 , @xmath8-@xmath9 mixing and the decay width of @xmath174 can be modified , but the one - loop effects are generally small and do not exceed the level of @xmath175 compared with the sm prediction . although fcnc decay of charm meson is expected to be modified at tree level , the non - standard effect is again small because of the natural suppression imposed by the fcnc couplings ( which are the products of ckm matrix elements ) .
the only suspected new effect in the @xmath23-quark system is through the @xmath176 decay , which proceeds through @xmath177 .
we find the expected maximal deviation in the ratio @xmath178 is less than @xmath179 for @xmath125 .
it vanishes for the maximal lepton mixing scenario , i.e. for @xmath107 , and the partial decay width of @xmath176 into the @xmath180 has to be measured to test this scenario . under the general mixing scenario ,
the model requires two additional free ( real ) parameters , which can be chosen to be any two of the 31 , 32 and 33 elements ( i.e. , @xmath181,@xmath182 and @xmath183 ) of the @xmath62 mixing matrix , although additional phases can be introduced to generate a new source of cp violation .
( note that the unitarity condition of the matrix @xmath62 requires @xmath184 . )
it turns out that depending on the values of the parameters , sizable effects in various fcnc processes are expected . in table 2
, we summarize the results under this scenario by giving the constraints on the mixing parameters , extracted from various experiments .
similar to the discussions given for the other two scenarios , the general mixing scenario also allows lepton number violation processes , such as @xmath185 and @xmath164 .
since their branching ratios are of the same order as those for the @xmath186 mode , they can be very useful for further testing the model .
finally , we note that in the sm neither @xmath61 nor @xmath62 can be separately determined , and only the ckm matrix @xmath133 , which is the product of @xmath187 and @xmath62 , can be measured experimentally .
however , in this model , the elements in the third column of the @xmath188 mixing matrices can be determined , and can be further constrained by including the other low - energy data . in conclusion
, we have shown that the current low energy data ( including the @xmath1-pole data ) still allows the third family fermions to have different gauge interactions from the first and the second family fermions , though the mass of the heavy gauge bosons in this case has to be larger than 1.7 tev . fortunately ,
as summarized in tables 1 and 2 , many of the predicted new effects of this model can be further tested at the tau / charm , kaon and bottom factories .
e.m . would like to thank k. hagiwara , y. okada , for useful discussion and comments .
he also thanks the matsumae international foundation for offering the fellowship .
part of this work was supported by the u.s .
nsf under grant phy-9802564 .
.predictions of various decay rates and mixing in the sm and this model under the @xmath189-quark mixing scenario .
case i : @xmath110 , @xmath141 , @xmath109 .
case ii : @xmath107 , @xmath166 , @xmath109 . [ cols="^,^,^,^,^",options="header " , ] a.j .
buras and r. fleischer , hep - ph/9704376 ; a.j .
buras , hep - ph/9711217 ; g. buchalla and a.j .
buras , nucl .
phys . * b400 * , 225 ( 1993 ) ; a. ali , c. greub and t. mannel , desy-93 - 016 ; g. buchalla and a.j .
buras , m.e .
lautenbacher , rev .
* 68 * , 1125 ( 1996 ) . | we consider a model in which the third family fermions are subjected to an @xmath0 interaction different from the first two family fermions . constrained by the @xmath1-pole data ,
the heavy gauge boson mass is bounded from below to be about @xmath2 tev at the @xmath3 level . in this model ,
the flavor mixing between @xmath4 and @xmath5 can be so large that @xmath6 and @xmath7 provide a better constraint than the lep / slc data in some region of parameter space .
furthermore , flavor - changing neutral currents are unavoidable in the quark sector of the model .
significant effects to the @xmath8-@xmath9 mixing and the rare decays of the @xmath10 and @xmath11 mesons , such as @xmath12 , @xmath13 , @xmath14 , @xmath15 and @xmath16 , are expected .
/ 0d kek - th-596 + october 1998 msuhep-81010 * ehab malkawi@xmath17 and c.p . yuan@xmath18 * @xmath19department of physics , jordan university of science & technology + irbid 22110 , jordan @xmath20theory group , kek , tsukuba , ibaraki 305 - 0801 , japan @xmath18department of physics and astronomy , michigan state university + east lansing , mi 48824 , usa pacs numbers:12.15.ji , 12.60.-i , 12.60.cn , 13.20.- v , 13.35.-r |
the study of charmonium production in hadronic collisions is an interesting test of our understanding of the physics of strong interactions .
while the production of the @xmath3 pair can be addressed in a perturbative - qcd approach , the subsequent binding of the pair is an essentially non - perturbative process , involving soft partons and occurring on a rather long timescale ( @xmath4 fm / c ) .
various theoretical approaches have been proposed ( see @xcite for a recent review ) , but a satisfactory description of j/@xmath0 production in collisions is still missing @xcite . in interactions , the heavy - quark pair is created in the nuclear medium , and the study of its evolution towards a bound state can add significant constraints to the models .
for example , the strength of the interaction between the evolving @xmath3 pair and the target nucleons , that can lead to a break - up of the pair and consequently to a suppression of the j/@xmath0 yield , may depend on its quantum state at the production level ( color - octet or color - singlet ) , and on the kinematic variables of the pair @xcite .
in addition to final - state effects , also initial - state effects may influence the observed j/@xmath0 yield in . in particular , parton shadowing in the target nucleus @xcite may suppress ( or enhance , in the case of anti - shadowing ) the probability of producing a j/@xmath0 , while the energy loss of the incident parton @xcite in the nuclear medium , prior to @xmath3 production , may significantly alter the j/@xmath0 cross section and kinematic distributions .
furthermore , effects as final - state energy loss and the presence of an intrinsic charm component in the proton may also play a significant role @xcite . clearly , the correct understanding and the disentangling of the various nucleus - related effects on j/@xmath0 production is a non - trivial task , which poses significant challenges to theory , but at the same time offers important insights on the j/@xmath0 production and interaction mechanisms . finally , a suppression of the j/@xmath0 has been proposed a long time ago as a signature of the formation , in ultrarelativistic nucleus - nucleus collisions , of a state where quarks and gluons are deconfined ( quark - gluon plasma ) @xcite .
results from collisions , taken in the same kinematic conditions of , and properly extrapolated to nucleus - nucleus collisions , are therefore necessary to calibrate the contribution of the various cold nuclear matter effects to the overall observed suppression @xcite . having to deal with a rather complicated interplay of various physical processes , the availability of accurate sets of data , spanning large intervals in the incident proton energy , and covering large @xmath5 and @xmath6 regions , is essential for a thorough understanding of the involved mechanisms . at fixed target energies , high - statistics j/@xmath0
samples have been collected in recent years by the desy experiment hera - b @xcite , at 920 gev incident energy , by e866 @xcite at fnal at 800 gev , and by the cern - sps experiment na50 at 400 and 450 gev @xcite . in this letter , we present an extension of these measurements towards lower energies , carried out by the na60 experiment @xcite .
j/@xmath0 production has been studied in collisions at 158 gev , the same energy used in collisions at the cern sps .
data have also been taken with a 400 gev beam , in order to provide a result that can be compared with previous sets of data taken by the na50 experiment @xcite .
we have performed a systematic study of nuclear effects by analyzing the @xmath7dependence of the j/@xmath0 production cross section on seven different target nuclei ( be , al , cu , in , w , pb and u ) .
the results are relative to a region close to midrapidity , and are presented differentially in @xmath8 and @xmath2 and compared with previous results from the higher-@xmath9 measurements mentioned above .
the influence of parton shadowing in the nuclear targets on the results is also discussed , based on various recent parametrizations of this effect @xcite . a comparison between our results at the two energies as a function of @xmath2 ,
the fraction of the nucleon momentum carried by the parton in the nuclear target that produces the j/@xmath0 , is carried out . by doing so
, the shadowing effects can be factorized and the importance of the other nuclear effects can be investigated .
the na60 experiment has measured muon pair production in and collisions at the cern sps .
its experimental apparatus was based on a muon spectrometer ( ms ) , positioned downstream of a hadron absorber with a total thickness of 12 nuclear interaction lengths ( @xmath10 ) .
the ms is coupled to a vertex spectrometer ( vt ) based on si pixel detectors . the experiment triggered on muon pairs detected in the ms , which were then matched , during reconstruction , to the corresponding tracks in the vt . for details on the detector set - up and matching between the ms and
vt we refer to @xcite .
the j/@xmath0 mesons have been identified through their decay to a muon pair . in the data taking
, the sps proton beam , with an average intensity of 5@xmath1110@xmath12 s@xmath13 , was hitting a target system composed of nine sub - targets with thicknesses between 0.005 and 0.012 @xmath10 , with relative spacing between targets from 0.8 to 1 cm . to limit the possible influence of position - dependent systematic effects on the measurement of the nuclear dependence of the j/@xmath0 yield
, the subtargets were placed in a mixed - a order , namely al , u , w , cu , in , be , be , be and pb .
the analysis described in this letter has been carried out on a data sample consisting of 2.8@xmath1110@xmath14 events at 158 gev , and 1.5@xmath1110@xmath14 events at 400 gev , containing a dimuon reconstructed in the ms .
the mass region [email protected] gev / c@xmath16 , dominated by j/@xmath0 decays , contains 3.2@xmath1110@xmath17 and 2.1@xmath1110@xmath17 events at 158 and 400 gev , respectively . for about 50% of these events ,
each of the two ms tracks can be matched in direction and momentum with a track in the vt .
such a value of the dimuon matching efficiency is due the smaller solid angle coverage of the vt with respect to the ms and to the inefficiency of the pixel sensors as will be discussed below .
a monte - carlo simulation has shown that the contamination from events where at least one of the ms tracks is matched to a wrong vt track is negligible in the j/@xmath0 mass region . for events containing a pair of hard muons from the decay of a j/@xmath0 ( the average muon momentum is 27 gev / c ) , the point of closest approach of the two muons gives an accurate estimate of their production point , with a resolution of @xmath18 650 @xmath19 m . in fig .
[ fig:0 ] we show the distribution of the longitudinal coordinates of the dimuon production vertex , for [email protected] gev / c@xmath16 .
the peaks corresponding to the position of the various production targets are clearly visible and well separated , showing that the nuclear target where the j/@xmath0 has been produced is unambiguously determined .
the ms detects j/@xmath0 produced in the rapidity range 3@xmath204 .
however , the acceptance of the vt for matched tracks from the j/@xmath0 decay is target - dependent , the more downstream targets covering smaller rapidities .
we have therefore selected events produced in the range [email protected] , which is covered for all the targets .
the cut [email protected] , on the polar angle of the decay muons in the collins - soper reference frame has also been applied , in order to remove events at the edge of the ms s acceptance .
the transverse momentum coverage for the j/@xmath0 measurement extends down to zero @xmath22 .
the acceptance varies by about 30% in the @xmath22 range accessible with the collected statistics ( @xmath23 3 gev / c ) .
the j/@xmath0 signal has been extracted for each target nucleus at the two energies by fitting the opposite sign mass spectra , in the region [email protected] gev / c@xmath16 , with a superposition of the mass shapes of the expected dimuon sources .
these include the drell - yan process ( dy ) , the semi - leptonic decays of correlated d - meson pairs ( @xmath24 ) , the j/@xmath0 and the @xmath25 resonances . the expected mass distributions for the dy and @xmath24 contributions have been calculated with pythia @xcite , using the grv94lo @xcite parton distribution functions .
the j/@xmath0 events have been generated , for the 400 gev data , using the @xmath26 and @xmath22 distributions measured with good accuracy by na50 @xcite . at 158 gev , the differential distributions have been tuned directly on the data , using an iterative procedure .
the events have been tracked through the set - up and then reconstructed with the same algorithm used for real data .
the contribution from the combinatorial pair background due to @xmath27 and @xmath28 decays ( completely negligible in the j/@xmath0 mass region ) has been estimated through an event mixing technique @xcite . in fig .
[ fig:1 ] we show , as an example , the opposite - sign dimuon invariant mass distribution , relative to the kinematical domain specified above , for interactions at 158 gev .
the quality of all the fits to the invariant mass spectra is satisfactory ( with @xmath29 ranging from 0.7 to 1.5 ) .
the mass resolution at the j/@xmath0 peak is @xmath1870 mev / c@xmath16 , and the number of j/@xmath0 events ranges from @xmath18800 to @xmath182000 , depending on the target .
the number of continuum events under the j/@xmath0 peak is very small ( @xmath304% ) , and is dominated by the dy contribution .
nuclear effects have been parametrized by fitting the @xmath7dependence of the production cross section with the simple power law @xmath31 , and then studying the evolution of @xmath32 as a function of various kinematic variables .
alternatively , nuclear effects have been quantified by fitting the data in the framework of the glauber model , having as input parameters the density distributions for the various nuclei .
the model gives as output the so - called j/@xmath0 absorption cross section @xmath33 .
clearly , both @xmath32 and @xmath33 represent effective quantities , including the contribution of the various sources of nuclear effects detailed in the introduction . the nuclear effects on j/@xmath0 production
have been evaluated starting from the cross section measured for each target , normalized to the cross section for the lightest one ( be ) : @xmath34 where , for the target with mass number @xmath35 , @xmath36 is the number of j/@xmath0 events , @xmath37 is the number of incident protons , @xmath38 is the number of target nuclei per unit surface , @xmath39 is the j/@xmath0 acceptance and @xmath40 is the detection efficiency . when building these ratios , the results obtained with the three be targets have been averaged .
the use of relative cross sections brings several advantages .
in particular , the number of incident protons cancels out ( a small attenuation factor , due to the inelastic scattering of beam particles , which is 6% for the most downstream target , has been corrected for ) since all the targets were simultaneously exposed to the beam , which had a transverse dimension much smaller than that of the targets .
furthermore , the fraction of the detection efficiency related to the ms also cancels out .
in fact , this detector can not distinguish the target where the dimuon has been produced .
this is due to the presence of the thick hadron absorber , to the large distance between the target system and the ms tracking chambers ( 6 to 16 m ) , and to the closely spaced targets .
therefore , the muon detection efficiency is independent of the production target .
contrary to the ms case , the angular acceptance covered by the vt is slightly target dependent . therefore , non - uniformities in the efficiency of a certain pixel plane lead to a rapidity - dependent inefficiency which is different from target to target .
the efficiency of each of the 15 vt planes has been estimated using a modified track reconstruction algorithm that excludes the plane under study .
for each reconstructed track we then check the presence of a hit in that plane in a fiducial region around the intersection of the track with the plane .
efficiency values are calculated on a run - per - run basis ( @xmath412 hours ) , for sub - regions of the pixel plane down to a size of @xmath180.2 mm@xmath16 , if the track statistics in the sub - region under study is large enough ( more than 40 tracks ) .
otherwise , contiguous regions are grouped in order to reach a statistically significant track sample .
the distribution of the efficiency values for the various regions of the vt is peaked at @xmath1890% , with 82% of the detector area having an efficiency larger than 60% . finally , the j/@xmath0 acceptances , estimated by means of the monte - carlo simulation , range between 14.9% and 23.0% in the kinematic range [email protected] , @xmath430.5 .
the quoted rapidity range corresponds to the center - of - mass rapidity windows [email protected] at 158 gev and @xmath450.33 at 400 gev . in table 1
we summarize , target by target and for the two energies under study , the relative values of the j/@xmath0 acceptance times detection efficiencies .
for completeness , we also report the relative integrated luminosities @xmath46 .
.the relative values , target by target , of the integrated luminosities , and of the products @xmath47 .
we have chosen as reference value the most upstream be target .
the al target was not in place during the data taking at 400 gev . [ cols="^,^,^,^",options="header " , ] [ tab:2 ]
two main features emerge from this comparison .
first , when going from negative towards positive @xmath5 , @xmath32 steadily decreases .
this effect was already known from the data of hera - b and e866 , taken at rather similar incident proton energies ( 920 , 800 gev ) , and a similar effect might be present in our results at 158 gev , even if the size of the errors is not negligible .
second , at a constant @xmath5 , the @xmath32 values are lower when the incident proton energy is smaller , as can be seen when comparing the hera - b / e866 results with our results at 400 and 158 gev . on the other hand
, it is also worth mentioning that from na3 results on j/@xmath0 production at 200 gev @xcite one extracts @xmath32 values which are in partial contradiction with these observations , being similar to those obtained with the higher energy data samples ( hera - b / e866 ) .
we also note that in our calculation of the na3 @xmath32 values , performed starting from their measured j/@xmath0 cross section ratios between and collisions , the small bias @xcite induced by the use of a light target in the determination of @xmath32 has been corrected for . a satisfactory theoretical interpretation of the complex observed pattern is missing for the moment .
various works have underlined the importance of several effects , including final state break - up , parton shadowing , initial and final state energy loss , and the presence of a charm component in the nucleon wavefunction @xcite .
however , the relative weight of these effects is still under debate .
the new results from na60 at 400 and 158 gev have been obtained with the same experimental apparatus and in very similar running conditions .
their direct comparison may therefore be a clean testing ground for models .
interesting information can also be obtained by a simple comparison of nuclear effects as a function of various kinematic variables . as an example
, we now consider the @xmath2 dependence of the @xmath32 parameter , @xmath48 being the fraction of the momentum of the target nucleon carried by the parton which produces the j/@xmath02 kinematics of the main j/@xmath0 production process @xmath49 is , effectively , that of a 2@xmath501 process ( @xmath51 ) i.e. the final state gluon is very soft .
the alternative 2@xmath502 approach is discussed in @xcite .
+ we also note that the relation strictly holds when @xmath52 in the c.m . frame . due to the low @xmath9 ,
this is only approximately true and induces a smearing in the estimation of @xmath53 .
however , in the investigated kinematic domain , the effect is similar for the two energies .
the resulting @xmath54 shift in the @xmath2 values does not affect the discussion in the text . ] .
this kinematic variable is particularly interesting since @xmath2 is the quantity that determines the amount of shadowing in the target nucleus ( or anti - shadowing in our kinematic range ) .
furthermore , the center of mass energy of the j/@xmath0-n system , which is a relevant quantity for the final state break - up of the j/@xmath0 in the cold nuclear medium , is , to a very good approximation , a function of @xmath2 alone ( @xmath55 ) .
since our results at the two energies cover , with good approximation , the same @xmath2 domain , we have performed an analysis of nuclear effects in five @xmath2 bins , covering the region 0.08 @xmath56 0.14 . in fig .
[ fig : shad ] we plot @xmath32 as a function of @xmath2 for the two energies under study .
we also show in the same plot the expected @xmath32 value that would be obtained if shadowing were the only nuclear effect to be present .
various parameterizations of nuclear shadowing have been considered @xcite , including the recent eps09 result , where an estimate of the uncertainty of the calculation was also carried out .
looking at fig .
[ fig : shad ] we see that shadowing alone would lead to @xmath32 values larger than 1 in our @xmath2 acceptance .
this result implies that the other nuclear effects produce a significant suppression of the j/@xmath0 yield .
the other remarkable feature emerging from fig .
[ fig : shad ] is that @xmath32 is not the same at a fixed @xmath2 , for the two values of @xmath9 . in our analysis ,
the systematic uncertainties related to the measurement of @xmath32 are partly correlated , in particular those corresponding to the measurement of the j/@xmath0 detection efficiency .
therefore , in fig . [ fig:4 ]
we plot , as a function of @xmath2 , the quantity @xmath57 which is affected by a significantly smaller systematic uncertainty .
the results clearly indicate that @xmath58 ( in the covered @xmath2 range we have @xmath59 ) , and that @xmath60 does not vary appreciably in the @xmath2 region under study .
having factorized the effect of shadowing , which is the same at the two energies at fixed @xmath2 , the observation of @xmath61 could be attributed to a stronger final state interaction at 158 gev .
since , for a given @xmath2 , the kinematics of the collision of the j/@xmath0 with the target nucleons is the same at 158 and 400 gev , the observed effect may be related to a change in the j/@xmath0 break - up cross section at fixed @xmath62 .
a different weight of the color - octet and color - singlet precursor @xmath63 states in the production process between the two energies could be a natural explanation .
however , non - relativistic qcd ( nrqcd ) calculations carried out in the fixed target energy domain @xcite , only show a weak @xmath9 dependence of the color - octet and color - singlet relative contributions .
therefore , a strong variation of the j/@xmath0 break - up cross section with @xmath9 seems unlikely . in such a case one should observe , for the two incident
proton energies , the same @xmath32 at a given @xmath2 in contrast with our observation . in this scenario , processes different from final state break - up and parton shadowing must be advocated in order to explain the observed nuclear effects on j/@xmath0 production . in particular , effects such as the initial state energy loss of the incident parton might play an important role and deserve further investigation . as a function of @xmath2 , for the 158 and 400 gev data samples .
also shown in the plot is a calculation of @xmath32 as resulting from shadowing alone .
various shadowing parameterizations have been considered.,title="fig : " ] = @xmath64 , as a function of @xmath2 .
the empty boxes around the points represent the uncorrelated systematic uncertainty .
the systematic error common to all the points is shown as a filled box around the @xmath65 line.,title="fig : " ]
the na60 experiment has measured j/@xmath0 production in collisions at the cern sps , at 400 and 158 gev , the latter energy being the lowest one where a detailed systematic study has been performed .
the results show a suppression of the j/@xmath0 yield in cold nuclear matter , which is larger at lower incident energy .
a comparison with results from previous experiments indicates that nuclear effects on j/@xmath0 production , at constant @xmath5 , exhibit a strong @xmath9-dependence .
the observation of different @xmath32 values at the two energies , for a constant @xmath2 , may suggest a strong change in the j/@xmath0 break - up cross section and/or the presence of other effects like initial state energy loss .
these data can also provide an important baseline for heavy - ion collision studies .
lansberg , eur .
j. * c60 * , 693(2009 ) .
n. brambilla et al.(quarkonium working group ) , eur .
j. * c71 * , 1534 ( 2011 ) .
r. vogt , nucl .
a700 * , 539 ( 2002 ) .
kopeliovich and b.g .
zakharov , phys .
rev . * d44 * , 3466 ( 1991 ) .
eskola , h. paukkunen and c.a .
salgado , jhep * 0904 * , 065 ( 2009 ) .
s. gavin and j. milana , pys .
* 68 * , 1834 ( 1992 ) .
r. vogt , phys .
rev . * c61 * , 035203 ( 2000 ) .
t. matsui and h. satz , phys .
b178 * , 416 ( 1986 ) .
b. alessandro et al.(na50 ) , eur .
j. * c39 * , 335 ( 2005 ) .
r. arnaldi et al .
( na60 ) , phys .
lett . * 99 * , 132302 ( 2007 ) .
i. abt et al .
( hera - b ) , eur .
j. * c60 * , 525 ( 2009 ) .
leitch et al .
( e866 ) , phys .
lett . * 84 * , 3256 ( 2000 ) .
b. alessandro et al .
( na50 ) , eur .
j. * c48 * , 329 ( 2006 ) .
r. arnaldi et al .
( na60 ) , eur .
j. * c59 * , 607 ( 2009 ) .
m. hirai et al . , phys . rev .
* c70 * 044905 ( 2004 ) .
eskola et al . , eur .
j. * c9 * 61 ( 1999 ) .
eskola et al . , jhep * 0807 * 102 ( 2008 ) .
t. sjstrand et al . ,
135 * , 238 ( 2001 ) .
m. glck et al . , z. phys .
* c67 * , 433 ( 1995 ) . b. alessandro et al .
( na50 ) , eur .
j. * c33 * , 31 ( 2004 ) .
h. devries , c.w .
dejager and c. devries , atomic data and nucl .
data tables * 36 * , 495 ( 1987 ) .
j. badier et al .
( na3 ) , z. phys . *
c20 * , 101 ( 1983 ) .
r. shahoyan , phd thesis , universidade tcnica de lisboa , 2001 .
n . tram and f. arleo , eur .
j. * c61 * , 847 ( 2009 ) . c. lourenco et al .
, jhep * 0902 * , 014 ( 2009 ) .
e.g. ferreiro et al .
, phys.lett .
* b680 * 50 ( 2009 ) .
r. vogt , phys . rev . *
c61 * , 035203 ( 2000 ) . | the na60 experiment has studied j/@xmath0 production in collisions at 158 and 400 gev , at the cern sps .
nuclear effects on the j/@xmath0 yield have been estimated from the a - dependence of the production cross section ratios @xmath1 ( a = al , cu , in , w , pb , u ) .
we observe a significant nuclear suppression of the j/@xmath0 yield per collision , with a larger effect at lower incident energy , and we compare this result with previous observations by other fixed - target experiments . an attempt to disentangle the different contributions to the observed suppression has been carried out by studying the dependence of nuclear effects on @xmath2 , the fraction of the nucleon momentum carried by the interacting parton in the target nucleus . |
recently , the reaction @xmath11 has become accessible in photoproduction experiments on the deuteron or nuclei @xcite . at energies around @xmath1 gev ,
an excess of @xmath0 production on the neutron compared to the proton case has been reported in these experiments . on the theoretical side ,
this excess has been interpreted as a potential signal for a non - strange member of an anti - decuplet of pentaquarks @xcite although the most prominent of these states , the @xmath12 , may come from a peak created by the experimental cuts , helped by a statistical fluctuation due to the limited statistic of the experiment @xcite . for the peak in @xmath11 , there are also other explanations mostly in terms of different interfering partial waves @xcite . in the presented work , the phenomenon is addressed within the chiral unitary framework developed in ref .
details on the results presented in these proceedings can be found in ref .
@xcite , in particular a thorough discussion on the stability of the results .
the hadronic interaction in the present framework @xcite is mediated by the weinberg - tomozawa interaction in the coupled channels @xmath13 , and @xmath2 , unitarized in a bethe - salpeter equation .
the model also contains explicit resonance states which account for the @xmath14 and a phenomenological background .
the gauge invariant implementation of the photon interaction follows refs .
@xcite ( see also refs .
@xcite ) . for the results on the quasi - free @xmath15 and @xmath16 in the deuteron
, we use the impulse approximation . in the presented work @xcite , a global fit of @xmath17 multipoles ,
@xmath9-wave cross sections and partial waves has been performed for the reactions @xmath18 , @xmath19 , @xmath20 , @xmath21 , @xmath22 , @xmath23 , and @xmath24 where @xmath25 , but we concentrate here on the results for @xmath26 . .
solid ( dashed ) line : result , fermi motion included ( excluded ) . for the other curves ,
also , the @xmath27 , @xmath28 , and @xmath2 thresholds are indicated.,scaledwidth=47.0% ] in fig .
[ fig : deueta ] to the left , the present result is compared with the recent cross section data on the quasi - free @xmath16 and @xmath15 from ref .
the data are well reproduced . to the right ,
the ratio of these cross sections is shown ( solid line ) .
the dashed line indicates the ratio of cross sections on free nucleons , which becomes fermi smeared ( solid line ) .
the appearance of the sharp peak in @xmath29 is obviously due to the intermediate states @xmath28 and @xmath2 in the model . indeed , both in @xmath21 and @xmath20 , the photon can couple to charged pions and kaons in the intermediate @xmath30 and @xmath2 states ; however , in @xmath20 the photon coupling to the @xmath31 in the @xmath32 state is possible , while this is not possible in @xmath21 , because the corresponding intermediate state is given by @xmath33 @xcite . for the ratio @xmath29
, this difference manifests itself in the observed peak structure in fig .
[ fig : deueta ] to the right . indeed , removing the photon coupling to the @xmath32 state in the @xmath34 reaction , one obtains the ratio given by the dotted line in fig .
[ fig : deueta ] to the right ; the peak has disappeared . to check for the model dependence of the presented results
, we have replaced the hadronic final state interaction ( fsi ) with the weinberg - tomozawa ( wt ) term .
the photoproduction is then given by the triangle graph which contributes at next - to - leading order ( nlo ) in the chiral expansion of the amplitude @xcite .
the resulting ratio @xmath29 , shown as the dash - dotted line in fig .
[ fig : deueta ] to the right ( multiplied by an arbitrary factor of 20 ) , is , of course , very different in magnitude from the full result ( dashed line ) replacing the strong , non - perturbative fsi by the tree - level wt term is certainly an oversimplification .
however , the energy dependence shows the same pronounced @xmath35 cusps as the full result . in the future
, the present results could be tested within the hadron exchange framework of ref .
@xcite where @xmath35 states have been included recently . summarizing , the experimentally determined @xmath0 photoproduction cross sections on quasi - free neutron and proton can be explained quantitatively within the present model which accounts for the @xmath9 wave contribution only .
the chiral coupled channels su(3 ) dynamics and its interplay with the photon lead to the occurrence of the observed spike - like structure in @xmath29 .
the photoproduction of meson pairs is proving to be a rich field allowing us to widen our understanding of hadron dynamics and hadron structure . following much work devoted to the photoproduction of two pions in the last decade ,
@xmath37 photoproduction has attracted attention recently @xcite .
polarization observables are reported in refs .
the reaction was studied theoretically in refs .
@xcite and partial waves analyses of the reaction have been also performed @xcite , finding the @xmath38 partial wave to be important . in the chiral unitary approach of ref .
@xcite , the process turned out to be dominated at low energies by the excitation of the @xmath38 , which then decays into @xmath39 , with the @xmath40 subsequently decaying into @xmath30 . in that approach , predictions of the cross section are possible because the @xmath41 and @xmath42 couplings are known from the chiral unitary framework @xcite in which the @xmath38 appears dynamically generated . in ref .
@xcite the radiative decay width of the @xmath43 could be predicted , because the photon coupling to the mesons and baryons that constitute this resonance are all well known .
the result is in agreement with the phenomenologically known values @xcite from data analyses which are used in refs .
@xcite for the @xmath44 reaction .
the theoretical framework from refs .
@xcite is quite predictive since another coupling of the @xmath38 resonance is to the @xmath45 state and the evaluated ( differential ) cross sections for the reaction @xmath46
@xcite agree with the measurements published in ref .
@xcite . in ref .
@xcite , the chiral unitary amplitude from ref .
@xcite has been used to relate eleven different pion- and photon - induced reactions . from the list of the underlying processes included in ref .
@xcite , we show here only those in fig . [
fig : tree ] , which involve the @xmath47 and @xmath48 vertices predicted from the chiral unitary amplitudes @xcite .
these processes give the largest contributions to the @xmath49 and @xmath50 final states .
the latter appears from the unitarization of the meson - baryon amplitude in which the @xmath51 appears dynamically generated @xcite .
there are , however , also processes given by @xmath52-channel resonance exchange taken from the @xmath53 valencia model of ref .
@xcite . furthermore , there are contributions from the kroll - ruderman and meson pole terms , contributions from the normal and anomalous magnetic moments of the baryons , and combinations of those processes .
all free constants that appear in the model have been fixed from other processes , thus the results of ref .
@xcite can be regarded as predictions .
to @xmath10 .
this is the dominant process .
the complex @xmath54 coupling is a prediction within the chiral unitary framework .
right : other processes with @xmath55 and @xmath56 couplings and @xmath57 final states . for the full list of processes ,
see ref .
@xcite.,scaledwidth=30.0% ] the model resulting from fig .
[ fig : tree ] is gauge invariant .
this is so since the @xmath58 coupling is obtained from the experimental data through an expression which is manifestly gauge invariant @xcite . the recent work of ref .
@xcite presents another challenge since new observables are measured , i.e , the @xmath6 and @xmath7 polarizations as a function of the @xmath59 angle between the decay plane and the reaction plane ( for the precise definition of the reaction geometry , see e.g. ref .
@xcite ) . in the work presented here
@xcite , the chiral unitary amplitude developed in refs .
@xcite is used to straightforwardly evaluate @xmath6 and @xmath7 and compare to the data .
the predictions are shown in fig .
[ fig : is ] with the ( red ) solid lines , together with the data from ref .
@xcite .
, right : polarization observable @xmath60 .
shown are the three cases of @xmath15 , @xmath61 , and @xmath0 spectators and three different energies @xmath62 .
the data are from ref .
@xcite ( full circles ) .
the empty circles show @xmath63 and @xmath64 , respectively .
( red ) solid lines : present results , predicted from the model of refs .
( black ) dotted lines : without the processes from fig .
[ fig : tree ] .
( black ) dash - dotted lines : only contribution from the tree level diagram of fig .
[ fig : tree ] , left.,title="fig:",scaledwidth=50.0% ] , right : polarization observable @xmath60 . shown
are the three cases of @xmath15 , @xmath61 , and @xmath0 spectators and three different energies @xmath62 .
the data are from ref .
@xcite ( full circles ) .
the empty circles show @xmath63 and @xmath64 , respectively .
( red ) solid lines : present results , predicted from the model of refs .
( black ) dotted lines : without the processes from fig . [
fig : tree ] .
( black ) dash - dotted lines : only contribution from the tree level diagram of fig .
[ fig : tree ] , left.,title="fig:",scaledwidth=50.0% ] note the symmetries @xmath65 and @xmath66 @xcite .
the present results reproduce well the complex shapes of the angular distributions , while at the highest energies deviations start to become noticeable .
the dotted lines show the results without using the diagrams from fig .
[ fig : tree ] which demonstrates that the @xmath41 and @xmath42 couplings provide the essential dynamics . indeed , using only the tree level diagram to the left of fig .
[ fig : tree ] , the main features of the full results are already obtained ( dashed - dotted line in fig . [
fig : is ] ) . in ref .
@xcite , also the observables @xmath67 for the three spectator cases have been evaluated . while the agreement with data for the @xmath15 spectator case @xcite is good ,
there is no published data yet for the other two cases . in ref .
@xcite , also a discussion of the theoretical error can be found which turns out to be well under control .
summarizing , the theoretical predictions of @xmath6 , @xmath7 , and @xmath68 in the reaction @xmath69 agree well with the data recently measured at cbelsa / taps , providing support to a chiral unitary model , in which the @xmath38 appears dynamically generated .
this work is supported by the dfg ( deutsche forschungsgemeinschaft , gz : do 1302/1 - 2 and sfb / tr-16 ) and the cosy ffe grant no .
41445282 ( cosy-58 ) .
this work is partly supported by dgicyt and feder funds contract no .
fis 2006 - 03438 , the generalitat valenciana in the program prometeo and the eu integrated infrastructure initiative hadron physics project under contract rii3-ct-2004 - 506078 .
v. kuznetsov _ et al . _
[ graal collaboration ] , phys .
b * 647 * , 23 ( 2007 ) .
f. miyahara _ et al .
_ , prog .
phys . suppl .
* 168 * , 90 ( 2007 ) .
t. mertens _ et al . _
[ cbelsa and taps collaboration ] , eur .
j. a * 38 * , 195 ( 2008 ) . i. jaegle _ et al . _
[ cbelsa and taps collaboration ] , phys .
lett . * 100 * , 252002 ( 2008 ) .
v. kuznetsov _ et al .
_ , acta phys .
b * 39 * , 1949 ( 2008 ) .
d. diakonov , v. petrov and m. v. polyakov , z. phys .
a * 359 * , 305 ( 1997 ) .
a. fix , l. tiator and m. v. polyakov , eur .
j. a * 32 * , 311 ( 2007 ) .
a. martinez torres and e. oset , phys .
lett . * 105 * , 092001 ( 2010 ) .
r. a. arndt _ et al .
_ , phys .
c * 69 * , 035208 ( 2004 ) .
k. s. choi , s. i. nam , a. hosaka and h. c. kim , phys .
b * 636 * , 253 ( 2006 ) .
v. shklyar , h. lenske and u. mosel , phys .
b * 650 * , 172 ( 2007 ) .
r. shyam and o. scholten , phys .
c * 78 * , 065201 ( 2008 ) .
w. t. chiang , s. n. yang , l. tiator and d. drechsel , nucl .
phys . a * 700 * , 429 ( 2002 ) .
a. v. anisovich _ et al .
_ , eur .
j. a * 41 * , 13 ( 2009 ) . m. dring and k. nakayama , eur .
phys . j. a * 43 * , 83 ( 2010 ) .
m. dring and k. nakayama , phys .
b * 683 * , 145 ( 2010 ) .
h. haberzettl , k. nakayama and s. krewald , phys .
c * 74 * , 045202 ( 2006 ) .
k. nakayama , y. oh and h. haberzettl , arxiv:0803.3169 [ hep - ph ] .
b. borasoy , p. c. bruns , u .- g .
meiner and r. nissler , eur .
phys . j. a * 34 * , 161 ( 2007 ) .
v. bernard , n. kaiser and u .- g .
meiner , z. phys .
c * 70 * , 483 ( 1996 ) .
m. dring , c. hanhart , f. huang , s. krewald , u .- g .
meiner and d. rnchen , arxiv:1009.3781 .
t. nakabayashi _ et al .
_ , phys .
c * 74 * , 035202 ( 2006 ) .
j. ajaka _ et al .
_ , phys . rev
. lett . *
100 * , 052003 ( 2008 ) .
i. horn _ et al . _
[ cb - elsa collaboration ] , eur .
j. a * 38 * , 173 ( 2008 ) . v. l. kashevarov , eur .
j. a * 42 * , 141 ( 2009 ) .
e. gutz _ et al .
_ [ cbelsa / taps collaboration ] , eur .
j. a * 35 * , 291 ( 2008 ) .
e. gutz _ et al .
_ [ cbelsa / taps collaboration ] , phys .
b * 687 * , 11 ( 2010 ) . v. l. kashevarov _ et al . _
[ a2 collaboration ] , arxiv : 1009.4093 [ nucl - ex ] .
d. jido , m. oka and a. hosaka , prog .
phys . * 106 * , 873 ( 2001 ) .
m. dring , e. oset and d. strottman , phys . rev .
c * 73 * , 045209 ( 2006 ) .
a. anisovich , e. klempt , a. sarantsev and u. thoma , eur .
j. a * 24 * , 111 ( 2005 ) .
a. fix , v. l. kashevarov , a. lee and m. ostrick , phys .
c * 82 * , 035207 ( 2010 ) .
e. e. kolomeitsev and m. f. m. lutz , phys .
b * 585 * , 243 ( 2004 ) .
s. sarkar , e. oset and m. j. vicente vacas , nucl .
a * 750 * , 294 ( 2005 ) [ err .
: a * 780 * , 78 ( 2006 ) ] .
m. dring , nucl .
a * 786 * , 164 ( 2007 ) . c. amsler _ et al . _
[ particle data group ] , phys .
b * 667 * , 1 ( 2008 ) .
m. dring , e. oset and d. strottman , phys .
b * 639 * , 59 ( 2006 ) .
m. nanova _ et al .
_ [ cbelsa / taps collaboration ] , eur .
j. a * 35 * , 333 ( 2008 ) . t. inoue , e. oset and m. j. vicente vacas , phys .
c * 65 * , 035204 ( 2002 ) .
j. c. nacher , e. oset , m. j. vicente and l. roca , nucl . phys .
a * 695 * , 295 ( 2001 ) .
m. dring , e. oset and u .- g .
meiner , arxiv:1003.0097 [ nucl - th ] , + accepted for publication in eur .
. j. a. | the recently discovered enhancement of @xmath0 photoproduction on the quasi - free neutron at energies around @xmath1 gev is addressed within an su(3 ) coupled channels model .
the @xmath2 threshold plays a significant role and the quasi - free cross sections on proton and neutron , @xmath3 and @xmath4 , can be quantitatively explained . for the reaction @xmath5
, we evaluate the polarization observables @xmath6 and @xmath7 from a chiral unitary amplitude developed earlier .
the @xmath6 and @xmath7 observables have been recently measured for the first time by the cbelsa / taps collaboration .
we show the significance of the @xmath8 resonance and its @xmath9-wave decay into @xmath10 which further confirms the dynamical nature of this resonance .
address = institut fr kernphysik and jlich center for hadron physics , + forschungszentrum jlich , d-52425 jlich , germany |
since the top quark mass is the largest in the standard model , it enters theories as the largest quantum correction .
many models beyond the standard model hold a special role for top quarks . here
we present results for searches for physics beyond the standard model in models not involving supersymmetry .
all searches use proton - proton collisions at 8 tev , most of them use the full data set of lhc run i ( about 20 fb@xmath0 ) collected by the atlas and cms experiments .
the talk is divided in three sections according to the type of physics models and analysis techniques : searches for new resonances and heavy bosons , searches for dark matter produced in association with top quarks , and searches for new vector - like fermions .
an extended gauge sector featuring massive charged bosons , @xmath1 and @xmath2 , is predicted in many extensions of the standard model @xcite .
searches for a @xmath1 boson , a heavy partner of the standard model @xmath3 boson , are traditionally done in the lepton plus neutrino final state .
the current limit on this search is 3.74 gev .
this limit assumes the neutrino is lighter than the @xmath1 . instead , if the @xmath1 boson couples to a right - handed neutrino which is heavier than the @xmath1
, it would decay to top and bottom quarks .
searches are also performed for left - handed @xmath1 bosons where the interference between the @xmath1 and the s - channel single - top production is also taken into account .
table [ wprime - ljets ] shows the expected and observed limits as a function of @xmath1 boson mass for @xmath4 in the final state with a lepton , missing transverse energy , jets and at least one @xmath5-tag for the cms and atlas experiments .
results from the cms experiment were obtained by fitting the invariant mass of the reconstructed top and bottom quarks @xcite . in this
proceeding , atlas presents new preliminary results using a multivariate boosted decision tree approach ( fig . [ wprime - coupling ] ) @xcite .
a more detailed comparison of the cms and atlas analyses started at the top2014 conference .
currently studies indicate that cms and atlas are using a different theoretical assumptions and their limits might not be directly comparable
. @*7l| model&cms observed ( expected ) & atlas observed ( expected ) + @xmath6&1.84 ( 1.84 ) & 1.70 ( 1.54 ) + @xmath7&2.13 ( 2.12 ) & 1.92 ( 1.75 ) + ] ] searches for the @xmath1 boson were also done in the all - jets final state of @xmath4 .
atlas and cms use techniques for reconstructing highly boosted objects in order to reconstruct the invariant mass of the top and bottom quark system .
the multijet background estimate is derived from data .
table [ wprime - alljets ] shows the current limits for atlas and cms @xcite .
@*7l| model&cms observed ( expected ) & atlas observed ( expected ) + w@xmath8&1.94 ( na ) & 1.68 ( 1.63 ) [ no sm interefence considered ] + w@xmath9&2.02 ( 1.99 ) & 1.76 ( 1.85 ) + currently , the most stringent limits are produced by the cms experiment by combining the lepton+jets and all - hadronic channels resulting in a limit of 2.15 tev for @xmath7 ( fig.[wprime - combo ] ) @xcite .
many theories beyond the standard model predict the existence of new particles that decay to @xmath10 quarks .
searches for resonant production of @xmath10 quarks can test theories such as top - color , chiral color models , randall - sundrum ( rs ) models with warped extradimensions , etc . by looking for a bump in the @xmath10 invariant mass spectrum @xcite .
two benchmark models are tested : * a narrow resonance in comparison with detector resolution ( topcolor , leptophobic , @xmath11 ) with @xmath12=1.2% ( or 1% ) with a k - factor of 1.3 . * a broad resonance ( kaluza klein gluons from rs models with extradimensions , @xmath13 ) with @xmath12=15.3% ( or 10 - 15% ) with no k - factor .
the analyses presented use techniques for resolving boosted objects to reconstruct the @xmath14 pair and fit its invariant mass distribution to find a signal .
results from atlas analyzing the @xmath15 single lepton decay channel using resolved and boosted techniques are shown in table [ zprime ] @xcite .
note that this result uses a smaller dataset of 14fb@xmath0 in comparison to the other results presented in this proceeding .
cms results shown in table [ zprime ] are extracted after combining the results from the single - lepton and all - hadronic analyses @xcite .
@*7l| model&atlas observed ( expected ) & cms observed ( expected ) + & @xmath16 lepton+jets ( @xmath17=14 fb@xmath0 ) & z @xmath16 lepton+jets + all - jets ( @xmath17=20 fb@xmath0 ) + @xmath2 ( @xmath12=1.2% ) & 1.8 ( 1.9 ) & 2.1 ( 2.1 ) + @xmath13 ( @xmath12=15.3% ) & 2.0 ( 2.1 ) & 2.5 ( 2.4 ) +
searches for dark matter ( dm ) particles produced in a collider are most sensitive to low masses ( @xmath1810 gev ) and provide complementary information to to direct detection searches , which are most sensitve to larger dm masses . if particles that mediate the interaction between dark matter and standard model particles ( @xmath19 ) are too heavy to be produced , interactions can be described as contact operators in an effective field theory ( eft ) with operators expressed in terms of @xmath19 .
different operators are tested using models where the dark matter is produced with a top quark pair @xcite .
the cms experiment sets lower limits on the interaction of scale @xmath19 assuming a d1 coupling in the single - lepton and dilepton channel @xcite .
atlas presents a new result using single - lepton and all - jets @xmath10 decays and sets lower limits on three different operators , d1 , d11 , d9 @xcite . these new results are shown in fig .
[ dm - atlas ] . ] ] ] searches for missing energy produced with a single top quark ( mono - top ) have been done by cms and atlas .
cms extracts limits using the reconstructed top quark invariant mass in final states with three jets and large missing transverse energy and set limits within the eft interpretation where the dark matter particles are scalar and vector bosons , excluding scalar ( vector ) dm particle masses below 327 ( 655 ) gev @xcite .
newly presented at this conference were the limits on mono - top searches by the atlas experiment using the single - lepton channel @xcite .
the search was interpreted with two models : a @xmath20 production of a + 2/3 charged spin-0 boson , s , decaying into a right - handed top quark and a non - interacting neutral spin-@xmath21 fermion , @xmath22 , and as a @xmath23 production of a non - interacting neutral spin-1 boson , in association with a right - handed top quark , @xmath24 .
results are shown in fig .
[ monotop - atlas ] . ] ]
vector - like quarks are defined as quarks with left and right handed components transforming identically under ( su2)@xmath8 .
vector - like quarks ( vlq ) have been proposed in many theories beyond the standard model , mainly to address the naturalness problem in models like little higgs , composite higgs , etc .
they may exist as weak - isospin doublets , singlets , or triplets and can occur as down - type quark ( @xmath25 ) , up - type quark ( @xmath26 ) , or as a quark with even more exotic electric charges of 4/3 e or 5/3 e. predominantly they decay to third generation quarks , and therefore produce a variety of search final states .
searches are performed for pair or single production of @xmath27 and @xmath28 .
pair production searches are more common as the production of single @xmath25 and @xmath26 quarks via the weak interaction has lower cross sections . in general signatures
have large number of jets , leptons , and b - jets .
some of the final states have leptons with the same - sign charges .
selection of same - sign leptons is exploited to suppress sm backgrounds .
top quarks or bosons decaying from heavy quarks can have large momenta , showing a boosted signature .
many cms and atlas analyses use boosted tagging techniques to identify higgs , top quark , @xmath3 or @xmath29 bosons from heavy quarks .
the search strategy consists in discriminating signal from background using @xmath30 , the scalar sum of the transverse momenta of the reconstructed jets , leptons , and missing transverse energy , or @xmath31 , the reconstructed invariant mass of the final - state objects in each channel .
no new results were presented at this conference . in absence of an observed signal
the experiments produce limits on different couplings and vlq masses .
depending of the decay mode , @xmath25 and @xmath26 quarks are ruled out between 600 to 800 gev .
atlas and cms results can be found in references @xcite .
in addition searches for quarks with a charge of 5/3 e and 2/3 e have not seen any significant deviation from the sm and limits exclude top quark partners with masses below 800 gev for all possible values of the branching fractions @xcite .
the top quark plays an important role in beyond - standard model ( bsm ) theories . atlas and cms have used 8 tev proton - proton collisions from the lhc and the full dataset of about 20 fb@xmath0 to search for different bsm signatures .
here we have presented searches for @xmath1 and @xmath2 bosons as well as @xmath13 , production of dark matter in association with single top quark and top quark pairs , and lastly different searches for vector - like quarks produced singly or in pair .
no hint for new physics was found but stringent limits on these models have been set .
all the top quark decay channels were explored and many different reconstruction and analysis techniques were developed and improved for these searches .
9 g. burdman , b. a. dobrescu , and e. ponton , phys .
d74 ( 2006 ) 075008 , arxiv : hep - ph/0601186 ; h .- c .
cheng , c. t. hill , s. pokorski , and j. wang , phys .
d64 ( 2001 ) 065007 , arxiv : hep - th/0104179 ; t. appelquist , h .- c .
cheng , and b. a. dobrescu , phys .
d64 ( 2001 ) 035002 , arxiv : hep - ph/0012100 ; j. c. pati and a. salam , phys .
d10 ( 1974 ) 275289 ; r. n. mohapatra and j. c. pati , phys .
d11 ( 1975 ) 566571 ; g. senjanovic and r. n. mohapatra , phys .
d12 ( 1975 ) 1502 ; m. perelstein , prog .
58 ( 2007 ) 247291 , arxiv : hep - ph/0512128 ; d. j. muller and s. nandi , phys .
b383 ( 1996 ) 345350 , arxiv : hep - ph/9602390 ; e. malkawi , t. m. tait , and c. yuan , phys .
b385 ( 1996 ) 304310 , arxiv : hep - ph/9603349 . n. arkani - hamed , s. dimopoulos , and g. r. dvali , phys .
b 429 , 263 ( 1998 ) ; l. randall and r. sundrum , phys . rev .
83 , 3370 ( 1999 ) ; l. randall and r. sundrum , phys . rev . lett .
83 , 4690 ( 1999 ) ; k. agashe , a. belyaev , t. krupovnickas , g. perez , and j. virzi , phys . rev .
d 77 , 015003 ( 2008 ) ; n. arkani - hamed , a. g. cohen , and h. georgi , phys .
b 513 , 232 ( 2001 ) ; j. l. rosner , phys .
b 387 , 113 ( 1996 ) ; r. m. harris and s. jain , eur . phys .
j. c 72 , 2072 ( 2012 ) ; p. h. frampton and s. l. glashow , phys .
b 190 , 157 ( 1987 ) ; d. choudhury , r. m. godbole , r. k. singh , and k. wagh , phys . lett .
b 657 , 69 ( 2007 ) ; d. dicus , a. stange , and s. willenbrock , phys . lett .
b 333 , 126 ( 1994 ) ; r. frederix and f. maltoni , j. high energy phys . 01 ( 2009 ) 047 .
k. cheung et al . , jhep 2010 ( 2010 ) 081 ; m. beltran et al .
, jhep 2010 ( 2010 ) 117 ; j. goodman et al .
d 82 ( 2010 ) 116010 ; p. j. fox , r. harnik , j. kopp , and y. tsai , phys . rev .
d 85 ( 2012 ) 056011 ; p.
j. fox , r. harnik , r. primulando , and c .- t .
yu , phys .
d 86 ( 2012 ) 015010 ; a. rajaraman , w. shepherd , t. m. p. tait , and a. m. wijangco , phys . rev .
d 84 ( 2011 ) 095013 .
cms collaboration , cms - pas - b2g-14 - 004 .
cms collaboration , submitted to phys .
, arxiv:1410.1149 .
atlas collaboration , submitted to eur .
j. c , arxiv:1410.4031 .
atlas collaboration , submitted to eur .
j. c , arxiv:1410.5404 . | this proceeding describes the experimental results presented at the top2014 conference on searches for physics beyond the standard model in final states containing top quarks by the atlas and cms experiments at the lhc .
the searches presented were done in the context of a wide range of theoretical models except those related to supersymmetry .
results presented use data taken with about 20 fb@xmath0 of 8 tev proton - proton collisions from the lhc .
no significant excesses beyond the standard model are observed , therefore limits on potential new signals are set . |
rayleigh - bnard ( rb ) convection , which is a fluid layer heated from below and cooled from the top , is an idealized model to study turbulent flows involving heat transport and has attracted much attention during the past few decades @xcite .
the system is characterized by two control parameters : the rayleigh number @xmath0 , prandtl number @xmath23 , which are defined as @xmath24 and @xmath25 respectively . here
@xmath26 is the thermal expansion coefficient , @xmath27 the gravitational acceleration , @xmath28 the temperature difference between the bottom and the top plates , @xmath29 the height of the fluid layer between the plates , @xmath30 the kinematic viscosity , and @xmath31 the thermal diffusivity of the convecting fluid . in addition , the aspect ratio @xmath32 ( @xmath33 is the lateral dimension of the system ) also plays an important role in the structures and dynamics of the flow . in a fully developed rayleigh - bnard turbulent flow , most of the imposed temperature difference
is localized in the thermal boundary layers near the surface of the top and bottom plates , within which heat is transported via conduction @xcite .
the velocity field has the same character : velocity gradient is localized in a thin layer near the plates , which is called viscous boundary layer .
turbulent flow in the central region of the rb cell is approximately homogenous and isotropic@xcite .
as the top and bottom boundary layers contribute the main resistance to heat transfer through the cell and thus dominantly determine the nusselt number , they deserve special attention .
indeed , nearly all theories in rb convection are in essence boundary - layer ( bl ) theories .
for example , a turbulent bl was assumed in the early marginal stability theory @xcite and also in the models by @xcite and @xcite and by @xcite . on the other hand ,
a prandtl - blasius ( pb ) type laminar bl was assumption in the grossmann & lohse ( gl ) theory @xcite .
therefore , direct characterization of the bl properties is essential for testing and differentiating the various theoretical models , and will also provide insight into the physical nature of turbulent heat transfer .
one of the earlier measurements of temperature and also velocity profiles in turbulent rb convection was taken by @xcite in water ( @xmath34 ) at the fixed @xmath35 and at a fixed lateral position .
@xcite extended these measurements over the range @xmath36 in compressed gas ( air ) at room temperature ( pr=0.7 ) , but still at fixed lateral position .
@xcite measured the mean temperature profiles at various horizontal positions on the lower plate of a cylindrical convection cell , the result shows that the thermal layer thickness @xmath37 varied over the plate for the same @xmath0 , and the thinnest bl is closed to the center of the plate .
@xcite found similar results for a cubic cell .
@xcite measured high - resolution temperature profiles in rayleigh - bnard convection near the top plate of a cylindrical container with air ( @xmath38 ) as the working fluid .
their result shows that the thermal bl thickness @xmath39 in the cell with @xmath40 .
@xcite found that the thermal boundary layer thickness scales with @xmath41 in a rectangular cell , at @xmath42 and @xmath0 ranging from @xmath43 to @xmath44 .
for the velocity measurement , the methods for determining the velocity profiles near the solid walls of the cell are developed in recent years .
since strong temperature fluctuations exist in rayleigh - bnard convection , the well - established hot - wire anemometry could not be applied to this system . for the viscous boundary layer ,
the large temperature fluctuations make conventional laser doppler velocimetry ineffective because the temperature fluctuations cause fluctuations in the refractive index of the fluid that in turn make it difficult to steadily focus two laser beams to cross each other in the fluid @xcite .
@xcite introduced an electrochemical labeling method and measured the velocity profile and boundary layer thickness near the top plate of a cubic cell filled with water , but only at a single value of @xmath0 . in a later study , @xcite developed an indirect method the correspondence between the peak position of the cutoff frequency profile of the temperature power spectrum and the peak position of the velocity to infer the viscous boundary boundary layer thickness in gaseous convection .
this method has subsequently been used to infer the viscous layer in thermal convection in mercury @xcite .
it may just be that the method works in certain situations , but there is no theoretical basis for it .
@xcite , using a novel light - scattering technique developed by @xcite , conducted the first direct systematic measurement of velocity profiles in rb convection in a cylindrical cell as a function of @xmath0 .
they found @xmath45 from the velocity profile above the center of the lower plate .
@xcite extended these measurements to convection in cubic cells , finding the same scaling exponent @xmath46 at the bottom plate , but at the sidewall a different result @xmath47 . using various organic liquids , @xcite explored the @xmath23 dependence , finding @xmath48 . with the measured @xmath49 scaling relationship @xmath50 obtained in these studies ( in fact , it was the peclet number @xmath51 rather than @xmath10 in some of these studies ; please see @xcite for more detailed discussions ) , the above results imply a scaling relation @xmath52 . in recent years
, the technique of particle image velocimetry ( piv ) has been introduced to the experimental study of thermal convection @xcite .
@xcite further applied the piv technique to study the viscous bl in a rectangular cell .
their results show that @xmath53 and @xmath54 , which showed that the viscous bl in thermal turbulence has the same @xmath10-scaling as a prandtl - blasius laminar bl .
this result validates the laminar bl assumption made in the gl model in a scaling sense .
thus it appears there is a discrepancy in the measured scaling exponent of @xmath55 with respect to @xmath0 ( @xmath10 ) between those obtained in cylindrical and cubic cells and that obtained in rectangular cells . in @xcite
it was argued that because of the more complicated flow dynamics of the large - scale circulation ( lsc ) , such as the azimuthal motion in the cylindrical cell @xcite and the secondary flows in the cubic cell @xcite , the shear flow near the plates is less steady as compared to that in the rectangular cell which is more close to quasi - two - dimension ( quasi-2d ) . as the viscous boundary is created by the shear of the lsc , this may plausibly change the bl properties , resulting in a different exponent .
however , the above argument has not been substantiated experimentally .
part of the motivation of the present work is to determine how the three - dimensional lsc dynamics will affect the bl properties .
it is known that titling the cell by a small angle will `` lock '' the lsc in a fixed azimuthal plane in the sense that it will limit the range of the lsc s azimuthal meandering @xcite and reduce its azimuthal oscillation amplitude near the top and bottom plates of the cell @xcite . in the present work
, we present measurements of bl properties in a cylindrical cell with the cell titled over a range of angles . for small titling angles ,
we measure a boundary layer under a more steady shear comparing to the leveled " case when the lsc can freely meander in the azimuthal direction but presumably the bl is otherwise unperturbed under such a small titling angle ( @xmath56 ) .
we also examine how the bl scaling exponent and other bl properties behave when the titling angle becomes not so small ( @xmath57 ) , which would amount to a perturbation to the bls .
boundary layers play such an important role in turbulent thermal convection , it is therefore important to examine how bls respond to external perturbations .
understanding the stability or instability of bls is also relevant to the search for the so - called ultimate state of thermal convection , as the transition from the classical state " to the ultimate one is essentially an instability transition of the bl from being laminar to being turbulent .
in addition to the scaling of the bl thickness , the shape of the velocity profile near the top and bottom plates has attracted a lot of attention recently . although the bl has been found scaling wise to be of prandtl - blasius type ( at least in the quasi-2d case ) ,
the time - averaged velocity profiles are found to differ with the theoretically predicted one @xcite , especially for the region around the thermal bl .
recently , @xcite have proposed a dynamic scaling method that shows that the mean velocity profile measured in the laboratory frame can be brought into coincidence with the theoretical prandtl - blasius laminar bl profile , if it is resampled relative to the time - dependent frame that fluctuates with the instantaneous bl thickness .
this method was tested initially for the case of velocity profile in turbulent convection in a quasi-2d rectangular cell with water as working fluid ( @xmath58 ) . in a follow - up study using two - dimensional dns data
, @xcite found that the method is also valid for thermal boundary layers and for the case of @xmath59 as well .
more recently , these authors further shown , again using numerical data , that the method works also in other positions in the horizontal plate other than the central axis @xcite and in three - dimension ( 3d ) cylindrical cell for moderate values of @xmath0 @xcite .
however , @xcite and @xcite , both using numerical approaches , have found that dynamic scaling works less well in the 3d cylindrical geometry than in the quasi-2d case .
however , the method has not been tested experimentally so far in a 3d system . here
we would like to examine the dynamical scaling method using the experimentally obtained instantaneous velocity profiles in our three - dimensional cylindrical cell .
the remainder of this paper is organized as follows .
we give detailed descriptions of the experimental setup and measurement instrumentation in [ sec : apparatus ] and present and analyze experimental results in [ sec : result ] , which are divided into six subsections . in
[ subsec : thermalbl ] we present the measured temperature profiles and corresponding position - dependent fluid properties , which will be used to calculate the viscous and reynolds stresses . in
[ subsec : mean velocity profiles ] , the measured velocity profiles and their characterizations are presented . in
[ subsec : scaling for ra - scan ] , the scaling properties , with both @xmath0 and @xmath10 , of the thickness @xmath9 obtained from the mean velocity profiles and @xmath60 obtained form r.m.s .
velocity profiles , are presented and discussed .
we also discuss the influence of the cell tilting angle @xmath2 on the boundary layer scaling . in
[ subsec : statistical properties of the velocity field in the boundary layer ] statistical properties ( r.m.s . and skewness ) of the velocity field in the boundary layer region are discussed . in [ subsec : the quantities of turbulence ] we present results of the viscous and reynolds shear stresses distributions in the boundary layer , and discuss the scaling of the wall quantities . in [ subsec : instantaneous boundary layer properties ] we test the dynamic scaling method with respect to the measured instantaneous velocity profiles .
we summarize our findings and conclude in [ sec : summary ] .
the measurements were made in a cylindrical rayleigh - bnard convection cell , which has been described in detail previously @xcite .
here we give only its essential features .
the top and bottom conducting plates are made of pure copper with a thin layer of nickel to avoid oxidation .
the sidewall is made of plexiglas . to avoid distortions in the images viewed by the camera , a square - shaped jacket
is fitted around the sidewall of the convection cell .
as shown in figure [ fig : setupforpiv](b ) , the jacket is filled with water .
the diameter and height of the cell is @xmath61 cm and @xmath62 cm , respectively .
the aspect ratio @xmath63 is thus close to @xmath64 .
two ( three ) thermistors are embedded in the top ( bottom ) plate .
the top plate temperature is maintained constant by a refrigerated circulator ( polyscience model 9702 ) that has a temperature stability of @xmath65 .
a nichrome wire ( 26 gauge , aerocon systems ) surrounded by fiberglass sleeving and teflon tape is distributed inside the grooves carved under the bottom plate .
the wire is connected with five dc power supplies ( ge model gps-3030 ) in series to provide constant and uniform heating . during the measurement , the whole cell
is placed in a homemade thermostat box that is kept at the same temperature ( @xmath66 ) as that of the fluid at the centre of the cell . during the experiment
the cell was tilted by an angle @xmath2 such that the circulation plane of the lsc was parallel to the image plane of the camera ( the @xmath67-@xmath68 plane , see figure 1 ) .
the application of piv to thermal turbulence has been described in detail in several previous publications @xcite .
here we only provide details concerning the particular features of the present experiment .
the piv system consists of one ccd camera with @xmath69 pixels , a dual pulse nd - yag laser with @xmath70 mj per pulse , a synchronizer and software .
as the cell was titled , both the ccd and the laser light - sheet were titled accordingly with the same angle .
a 105 mm focal - length macro lens was attached to the ccd to achieve a measuring area with size varying from @xmath71 mm@xmath72 to @xmath73 mm@xmath72 .
each 2d velocity vector is calculated from a subwindow ( 32 pixels @xmath74 32 pixels ) that has 50@xmath75 overlap with its neighboring subwindows , so each vector corresponds to a region of 16 pixels @xmath74 16 pixels and each velocity map contains @xmath76 velocity vectors in the @xmath67-@xmath68 plane ( see figure 1 ) .
this corresponds to spatial resolutions of about @xmath77 mm@xmath72 to @xmath78 mm@xmath72 for velocities @xmath79 and @xmath80 measured in the horizontal @xmath67 and vertical @xmath68 directions , respectively . for the measurement at @xmath81 particles with diameter @xmath82 m were used , while particles with diameter of @xmath83 m were used for measurements with other three tilted angles . for each run , typically about 25200 image pairs were acquired with frame rate of @xmath84 hz .
sketch of the convection cell and the cartesian coordinates used in temperature and velocity measurements .
( a ) side view of the setup , and ( b ) the top view . ] measured at @xmath85 and @xmath81 .
( b ) a profile of dynamic viscosity @xmath86 obtained from the mean temperature profile in ( a ) . ]
piv measurements were made at four values of the titling angle @xmath2 @xmath87 , @xmath88 , @xmath89 , and @xmath6 .
for each @xmath2 , measurements over a range of @xmath0 were made .
table [ table : parameters ] lists the parameters ( @xmath2 , @xmath0 and @xmath23 ) of each measurement , which typically lasted for about 3.5 hours .
as already mentioned , titling the cell by a small angle has the effect of locking " the lsc s circulation plane at a fixed azimuthal angle ( in reality it restricts the angular range of the lsc s azimuthal meandering ) .
thus , measurements made with small @xmath2 are aimed at studying bl properties under more steady shear , but the bl itself is assumed to be unperturbed otherwise . for large values of @xmath2
we wish to examine how the bl responds to relatively large perturbations .
.control parameters of the experiment : cell tilt angle @xmath2 , the rayleigh number @xmath0 and the prandtl number @xmath23 ; and the measured maximum horizontal velocity @xmath90 and viscous boundary layer thickness @xmath9 .
the data are listed in chronological order . [
cols="^,^,^,^,^,^,^,^,^,^ " , ] scalings of the boundary layer scale @xmath60 determined from the measured rms velocity profiles : ( a ) versus @xmath0 and ( b ) versus @xmath13 .
the symbols represent : @xmath3 ( inverted triangles ) , @xmath4 ( squares ) , @xmath5 ( triangles ) , and @xmath6 ( circles ) .
the lines are power law fits @xmath91 and @xmath92 to the respective data sets , with the fitting results listed in table [ tab : table2 ] . ] in addition to the boundary layer thickness @xmath9 determined from the mean horizontal velocity profile , another length scale can also be defined based on the profile of the horizontal r.m.s .
velocity @xmath93 , which may be called the r.m.s . velocity boundary layer thickness , as defined in figure [ fig : tofindthickness ] . in figures [ fig : delta - rms - vs - re - rms](a ) and ( b ) we plot @xmath60 versus @xmath0 and @xmath13 respectively .
the @xmath0-scaling exponent varies from @xmath94 to @xmath95 , which appears to follow similar trend as that of @xmath9 , i.e. its absolute value increases with increasing @xmath2 .
but it and that of @xmath13-scaling exponent show significant difference with those obtained in previous studies . table [ tab : table2 ] shows the fitting results of @xmath96 and @xmath97 .
now we compare our result with previous experimental results obtained in the cells with different geometries .
as shown in table [ tab : table2 ] , the value of @xmath98 obtained in both cylindrical and cubic geometries and measured near the bottom plate of the cell is @xmath46 . in all these previous measurements ,
the reynolds number based on the maximum horizontal velocity near the plate was also obtained and they gave a scaling exponent @xmath99 via @xmath100 .
from this we obtain @xmath101 . in these studies ,
the convection cell was nominally leveled , i.e. not intensionally tilted . in the present study , for the small tilting angle cases , where we assume the bl is not strongly perturbed , the measured @xmath102 when combined with combined with @xmath103 give a @xmath104 ( note that the actual value of @xmath105 are obtained from fitting the @xmath9 vs @xmath10 data , not from the relationship between the exponents ) .
if we take these values to be close to the prandtl - blasius result , then scaling wise the viscous bl in a cylindrical geometry is also of a prandtl - blasius type , as was already found in a rectangular cell @xcite .
for the relatively large deviations found in the untilted case , it may be attributed to the random azimuthal motion of the lsc .
finally we remark that as far as the scaling of the viscous bl is concerned , there is no theoretical prediction for the dependence of @xmath9 on @xmath0 , only that on @xmath10 ( for example , @xmath106 for the prandtl - blasius bl ) . in the literature
, it is sometimes stated that @xmath9 should scale as @xmath107 for the prandtl - blasius bl .
this is based on the assumption that @xmath108 . from above
we have seen that the scaling exponent of @xmath10 with @xmath0 varies over a rather wide range .
it is therefore more meaningful to talk about the scaling of @xmath9 with @xmath10 , rather than with @xmath0 .
we further note that in @xcite it was found that @xmath109 and @xmath110 , which together give @xmath111 . in the present case
, we have @xmath112 and @xmath12 , which together give @xmath11 .
whether this is fortuitous or there is something deep here remains remains to be explored .
( left panels ) and vertical @xmath113 ( right panels ) velocity components measured at @xmath114 and @xmath115 , at @xmath116 and different distances @xmath68 from the bottom plate . ] and ( b ) vertical velocity @xmath113 measured at various distances from the bottom plate with @xmath114 and @xmath115 . ] ( @xmath117 ) and of ( b ) the skewness @xmath118 ( @xmath119 ) measured at @xmath114 and @xmath115 .
the vertical distance @xmath68 is normalized by the velocity boundary layer thickness @xmath9 . in both plots
the circles represent those for the horizontal velocity component @xmath79 and the crosses represent those for the vertical velocity component @xmath80 . ] in previous bl measurements in the cylindrical cell , owing to the nature of the dual - beam incoherent cross - correlation technique employed @xcite , only time - averaged velocity profiles are measured and no time - dependent quantities are obtained .
it is therefore interesting to examine these quantities and compare them with similar quantities obtained in other type of turbulent flows .
figure [ fig : different ] shows the time series of both the horizontal component @xmath120 ( left panel ) and the vertical component @xmath113 ( right panel ) of the velocity , measured at various positions from the plate .
the corresponding velocity histograms are shown in figure [ fig : different - histogram ] .
the measurements were made at @xmath114 and @xmath115 .
we show the velocity trace at several typical positions : ( i ) inside the thermal boundary layer , ( ii ) around the thermal boundary layer , ( iii ) around the viscous boundary layer ; ( iv ) at the position of the maximum velocity ; and ( v ) far away from the boundary layers .
the figures show that the absolute horizontal velocity is much higher than vertical velocity at each position .
one general feature we observed is that velocity time series and histograms look similar for different tilting angles .
for this reason , we show here results for only one tilting angle . at @xmath114 ,
the viscous bl thickness is @xmath121 mm .
it is seen from figures [ fig : different](a ) and [ fig : different](b ) that at positions inside the bl , the horizontal velocity @xmath120 skews toward the positive side , i.e. the velocity is skewed toward the mean flow direction .
this may be understood by the fact that close to the viscous sublayer the flow speed is very close to zero and a fluctuation smaller than the mean would mean a flow reversal , which is a rather rare event .
once outside of the bl , one observes more symmetric fluctuations around the mean velocity . for the vertical velocity @xmath113 , its mean velocity is very small at most positions .
but the fluctuation increases significantly when the position is outside of the bl , which are signatures of plume emissions at these positions .
these properties can also be seen from the velocity histograms shown in figure [ fig : different - histogram ] .
a notable difference of the present results from those observed in @xcite is that for positions outside of the bl the horizontal velocity fluctuates more or less symmetrically around the mean , rather than skewed toward the negative as seen in the rectangular cell .
the statistical properties of the velocity may be characterized more quantitatively by its root - mean - square ( r.m.s . )
value and its skewness , which are shown in figure [ fig : rmsnormalized ] .
figure [ fig : rmsnormalized](a ) plots the velocity r.m.s @xmath93 and @xmath117 normalized by maximum horizontal velocity @xmath90 versus the normalized distance @xmath122 .
figure [ fig : rmsnormalized](b ) shows the skewness profiles @xmath123 and @xmath124 for the horizontal and vertical velocities , respectively .
similar to @xcite , our result could not tell whether @xmath117 favors a power law or a logarithmic scaling with the distance @xmath68 , even though our measurement had a much higher spatial resolution .
this is partly due to the limited size of the measurement area . viscous stress @xmath125 ( solid squares ) , the reynolds stress @xmath126 ( open circles ) and the total stress @xmath15 ( open squares ) as functions of the normalized distance from the plate for ( a ) @xmath127 and ( b ) @xmath128 , both with @xmath81 . ]
the _ ra_-dependence of ( _ a _ ) the wall shear stress @xmath129 , ( _ b _ ) the friction velocity @xmath19 , ( _ c _ ) the wall thickness ( viscous sublayer ) @xmath17 , and ( _ d _ ) the friction coefficient @xmath21 , with power law fits shown as solid lines .
the symbols represent data for different tilting angles : @xmath3 ( inverted triangles ) , @xmath88 ( squares ) , @xmath89 ( triangles ) , and @xmath6 ( circles ) . ] the _ re_-dependence of ( _ a _ ) the wall shear stress @xmath129 , ( _ b _ ) the friction velocity @xmath19 , ( _ c _ ) the wall thickness ( viscous sublayer ) @xmath17 , and ( _ d _ ) the friction coefficient @xmath21 .
the symbols represent data for different tilting angles : @xmath3 ( inverted triangles ) , @xmath88 ( squares ) , @xmath89 ( triangles ) , and @xmath6 ( circles ) .
power law fits are indicated in the figure . ]
measured horizontal velocity profiles normalized by wall units for four different ra all taken at tilting angle @xmath81.,scaledwidth=75.0% ] one of the advantages of piv measurement is that it enables one to measure the horizontal and vertical velocities at the same time , so that one can calculate the reynolds shear stress @xmath130 . here
@xmath131 and @xmath132 are the fluctuations of the horizontal and vertical velocity components respectively , @xmath133 is the _ _ z-__dependent fluid density .
viscous shear stress is defined as @xmath134 , where @xmath135 is the dynamic viscosity dependent on position @xmath68 .
the reynolds stress represents the transport of momentum by turbulent fluctuations , whereas the viscous stress describes the momentum transfer by viscosity .
the total shear stress is then @xmath136
. figure [ fig : sheardistribution ] plots the profiles of the viscous shear stress , reynolds stress and total stress for ( a ) @xmath127 , and ( b ) @xmath137 .
it is seen that both ra have the same qualitative features . here
the examples are for @xmath138 , and results for other tilting angles are similar .
near the plate , it is seen that the reynolds stress @xmath126 is close to zero , while the viscous shear stress @xmath125 is maximum because of the large velocity gradient @xmath139 at the wall .
so the total stress at the wall @xmath140 comes almost entirely from the contribution of the viscous shear stress .
moving away from the plate , the velocity gradient becomes smaller and the viscous shear stress decreases to zero .
the reynolds stress @xmath126 increases and attains its maximum at @xmath141 .
it then decreases to around zero @xmath142 and becomes negative in the bulk flow .
it is also seen clearly from the figure that @xmath126 and @xmath125 cross at @xmath141 , where @xmath126 is close to its maximum value .
this suggests that the momentum transfer in the outer region is dominated by turbulent fluctuations .
but in the viscous boundary layer , the momentum transfer is still dominated by the viscous diffusion , which implies that the viscous boundary layer is still laminar in this range of @xmath0 . with the measured near - wall high - resolution velocity field , we are now in a position to check the dynamic wall properties in turbulent thermal convection .
we first consider the scaling of four basic wall quantities with both @xmath0 and @xmath10 .
these are the wall shear stress @xmath129 , the skin - friction velocity @xmath143 , the viscous sublayer length scale @xmath144 , and the skin - friction coefficient @xmath145 . here
@xmath146 and @xmath147 .
figure [ fig : wallra ] shows the scaling of these quantities with @xmath0 .
it is seen that within experimental uncertainties there is no difference between data with different @xmath2 .
this suggests that tilting the cell does not have any appreciable effect on bl properties near the wall . without differentiating the different data sets , power law fits to all data yield @xmath148 , @xmath149 , @xmath150 and @xmath151 . in a rectangular cell , sun _
et al _ ( 2008 ) found for the same quantities the fitted power law exponents @xmath152 , @xmath153 , @xmath154 , and @xmath155 respectively .
it is seen that the absolute values of these exponents are all larger than those obtained in the present experiment .
there is no theoretical prediction for the _ ra_-scaling of these quantities in turbulent thermal convection , so we do not know what the difference means
. it will be more useful perhaps to examine the scaling of these quantities with the reynolds number @xmath10 , since theoretical predictions exist for such scalings for wall - bounded shear flows @xcite .
figure [ fig : wallre ] plots these quantities as a function of _ re _ , the symbols are the same as in figure [ fig : wallra ] . for the quantities @xmath129 , @xmath19 , @xmath17 , and
@xmath21 our results give the exponents @xmath156 , @xmath157 , @xmath158 , and @xmath159 . for a laminar boundary layer over a flat plate ,
the theoretically predicted ` classical ' exponents for these quantities are @xmath160 , @xmath161 , @xmath162 , and @xmath163 respectively .
one sees that within the experimental uncertainties there is an excellent agreement between the present experiment and the theoretical predictions for all the wall quantities except for @xmath17 , which is a bit smaller . for reference ,
the previous measurement in rectangular cell gives @xmath164 , @xmath165 , @xmath166 , and @xmath167 for the corresponding quantities @xcite . to further compare the present system with classical boundary layers , we examine velocity profiles in terms of the wall units .
figure [ fig : wallnormalized ] shows the normalized mean horizontal velocity profiles for four different values of @xmath0 taken at @xmath81 in a semi - log plot , here @xmath168 and @xmath169 .
the linear scaling of @xmath170 over @xmath171 in the viscous sublayer below @xmath172 is reflected quite well by the measured profiles confirming that the boundary layer is not turbulent in the present range of @xmath0 and @xmath23 .
the velocity normalized by wall unit decrease after reaching the maximum value in @xmath173 . comparing to the same quantity measured in the rectangular cell @xcite , however , our result shows some deviation from the theoretical profile . this is a reflection of the fact that in the cylindrical cell it is more difficult to measure the profile accurately very close to the wall . , red circles ) and the laboratory frame ( @xmath174 , blue squares ) , measured at different tilt angles @xmath2 but with comparable values of @xmath0 . also shown for comparison
is the theoretical prandtl - blasius laminar velocity profile ( solid line ) .
( a ) @xmath3 , @xmath175 ; ( b ) @xmath176 , @xmath177 ; ( c ) @xmath178 , @xmath179 ; and ( d ) @xmath81 , @xmath180 . ] of profiles @xmath181 ( red circles ) obtained in the dynamical frame and of @xmath174 ( blue squares ) obtained in the laboratory frame as a function of @xmath0 and for different titling angles . ( a )
@xmath3 , ( b ) @xmath4 , ( c ) @xmath5 , and ( d ) @xmath6 .
the dashed line represents the value of @xmath182 for the theoretical prandtl - blasius laminar bl . ] and @xmath115 ) .
the corresponding instantaneous shape factor is also indicated on the plot .
the solid curves are the prandtl - blasius velocity profiles . ] between those of the rescaled instantaneous profiles and that of the prandtl - blasius profile , measured at ( a ) @xmath3 , ( b ) @xmath4 , ( c ) @xmath5 , and ( d ) @xmath6 . ]
the dynamic scaling method of @xcite has been found to work well when tested in quasi-2d experiment and 2d numerical simulations @xcite .
but it has not been examined in 3d experiments . here
we investigate how it works in our cylindrical geometry . as the method
has been well documented elsewhere @xcite , we will only give a brief description of it here .
from the measured instantaneous velocity profile @xmath183 one can obtain an instantaneous viscous boundary layer thickness @xmath184 using the same ` slope ' method as used for the mean velocity profiles . a local dynamical bl frame
can then be constructed by defining the time - dependent rescaled distance @xmath185 from the plate as @xmath186 the dynamically time averaged mean velocity profile @xmath187 in the dynamical bl frame is then obtained by averaging over all values of @xmath183 that were measured at different discrete times @xmath188 but at the same relative position @xmath189 , i.e. @xmath190 figure [ fig : profile-6e+8 ] shows the mean velocity profiles measured in the laboratory and the dynamical frames respectively at the four tilting angles and for comparable values of @xmath0 ( as indicated in the figure caption ) .
these results show that the dynamical scaling method appears to be more effective for larger values of @xmath2 .
this may be understood based on the fact that a larger tilt angle will place stronger restriction on the azimuthal meandering of the lsc so that it has less fluctuations in the horizontal direction perpendicular to the mean flow .
we note , however , regardless of the tilt angle , the method works less effectively than it is in quasi-2d experiment and 2d simulations .
a more quantitative approach to characterize the shape of the mean velocity profiles is to investigate their shape factor @xmath191 defined as the ratio between the displacement thickness @xmath9 and the momentum thickness @xmath192 , where @xmath193dz , ~\protect{and}~ \delta_{m}=\int_{0}^{\infty}[1-\frac{u(z)}{u_{max}}]\frac{u(z)}{u_{max}}dz.\ ] ] since @xmath174 decays after reaching its maximum value , the above integrations are evaluated only over the range from @xmath194 to where @xmath195 . for our profiles
the obtained shape factors range between @xmath196 to @xmath197 , which are smaller than @xmath198 , the value for a laminar prandtl - blasius boundary layer . a shape factor smaller than the theoretical value means the corresponding profile will approach its asymptotic value ( the maximum velocity ) slower than the theoretical profile does . in figure
[ fig : h - four - angles ] we show the shape factor @xmath29 for mean velocity profiles obtained in the laboratory and dynamical frames respectively for the four tilting angles and for all @xmath0 measured .
the dashed lines in the figure indicate the prandtl - blasius value of 2.59 .
it is seen that , despite the data scatter , there is a general trend that for both lab- and dynamical - frame profiles the deviation from the prandtl - blasius profile increases with @xmath0 .
this is no surprise , since , as the convective flow above the bl becomes more turbulent with increasing @xmath0 , the bl itself will experience stronger fluctuations and hence larger deviations from the laminar case .
this finding that the dynamical rescaling method works better for smaller ra than larger ones is consistent with those found in dns studies in the same geometry by @xcite for the temperature profile and by @xcite and by @xcite for the velocity profile .
the second feature is that for all @xmath2 and @xmath0 the profiles obtained in the dynamical frame in general show some degree of improvement towards that of prandtl - blasius value as compared to those obtained in the laboratory frame .
we also note that the degree of improvement does not seem to have an obvious dependence on @xmath0 , which is also consistent with the findings of @xcite .
some insight can be obtained by examining the rescaled instantaneous velocity profiles .
figure [ fig:1degree-6e+8-instantaneous ] show examples of rescaled instantaneous velocity profiles , where the distance from the plate has been normalized by the instantaneous bl thickness corresponding to that moment and the velocity has been normalized by the instantaneous maximum horizontal velocity .
it is seen that there are quite few cases where the rescaled instantaneous velocity profile is rather close to the theoretical prandtl - blasius profile ( up to the point of the maximum velocity ) and deviations of the instantaneous shape are likely caused by distubances such as plume emissions .
also shown in the figure are the shape factor @xmath199 of these instantaneous profiles . to quantify how the instantaneous profiles are distributed with respect to the prandtl - blasius profile , we examine the pdf of the shape factor difference @xmath200 where @xmath201 .
figure [ fig : deltah - pdf ] plots the pdfs of @xmath202 for the 4 tilting angles and for all measured @xmath0 respectively . despite the seemingly large variations among them ,
these pdfs show the general trend that the rescaled instantaneous profiles measured at lower values of @xmath0 ( @xmath203 ) are more of the time having a shape closer to that of the prandtl - blasius profile and that for higher values of @xmath0 the peak of the pdfs shift to smaller values of @xmath29 .
this indicates that with increasing @xmath0 the profiles around the bl thickness becomes more rounded , i.e. the approach to the maximum velocity becomes slower and slower .
we further note that these general trends are true across all tilt angles .
another feature observed in the present 3d case is that we did not find any strong correlation between the instantaneous bl thickness @xmath9 and the velocity @xmath120 just above the bl .
this is in contrast to the finding in the quasi-2d experiment where @xmath9 and @xmath120 are found to have a strong negative correlation , i.e. a large velocity above would exert a stronger shear and therefore thins the bl thickness @xcite .
this result suggest that in certain aspect the bls in the 3d and in the 2d / quais-2d cases are dynamically different .
we have conducted an experimental study of velocity boundary layer properties in turbulent thermal convection .
high - resolution two - dimensional velocity field was measured using the particle image velocimetry ( piv ) technique in a cylindrical cell of height @xmath62 cm and aspect ratio close to unity , with the rayleigh number @xmath0 varying from @xmath43 to @xmath204 and the prandtl number @xmath23 fixed at @xmath205 , with the convection cell tilted with respect to gravity at angles @xmath3 , @xmath88 , @xmath89 , and @xmath6 , respectively .
measurements made with small @xmath2 are aimed at studying bl properties under more steady shear , but the bl itself is assumed to be unperturbed otherwise . for large values of @xmath2
we wish to examine how the bl responds to relatively large perturbations .
we also examined effectiveness of the dynamical bl scaling method in a three - dimensional system .
it is found that the reynolds number @xmath10 ( @xmath206 ) based on the maximum mean horizontal velocity scales with @xmath0 as @xmath12 and the reynolds number @xmath13 ( @xmath207 ) based on the maximum rms velocity scales with @xmath0 as @xmath14 .
both exponents do not seem to have an apparent dependence on the tilt angle . on the other hand ,
the amplitude of @xmath10 seem to show a weak increasing trend with @xmath2 . with the measured horizontal velocity ,
we obtain two length scales , i.e. the viscous bl thickness @xmath9 based on the mean horizontal velocity profile and the length scale @xmath60 based on the rms horizontal velocity profile . it is found that as far as scaling with the reynolds number @xmath10 is concerned
, the behavior of @xmath9 can be divided into two regimes according to the tilting angle of the cell . for @xmath8 , it is found that @xmath11 , which within experimental uncertainty may be considered to be consistent with that of the prandtl - blasius bl .
it thus appears that the main effect of tilting the cell is to restrict the azimuthal meandering of the large - scale circulation but the bl is otherwise not strongly perturbed . for @xmath208 ,
the absolute value of the exponent is found to increase with @xmath2 and in this case the bl may be considered to be strongly perturbed . it is found that the scaling exponent of @xmath60 with respect to @xmath0 ( @xmath10 ) does not have a strong dependence on @xmath2 as @xmath9 does . but similar to @xmath9 , the absolute values of these exponents increase with increasing @xmath2 . it is also found that tilting the cell modifies the velocity profile in the bl region , i.e. for different tilt angles the shape of profiles is different .
but for the same tilting angle the velocity profiles measured at different @xmath0 can be brought to collapse on a single curve when the mean velocity is normalized by the maximum velocity @xmath90 and the distance from the plate by the viscous bl thickness @xmath9 . with simultaneously measured horizontal and vertical velocity components
, we also obtain the reynolds stress @xmath126 in the velocity boundary layer .
it is found that @xmath126 is stronger in the mixing zone comparing with the rectangular cell .
the wall quantities such as the wall shear stress@xmath129 , the viscous sublayer@xmath17 , the friction velocity @xmath19 are also measured .
their scaling exponents with the reynolds number are very close to those predicted for classical laminar boundary layers , which is also consistent with the measurement in rectangular cell .
regarding the dynamical scaling method , we found that the method in general works better when the cell is tilted at larger angle @xmath2 than it does at smaller angles , but the effect is somewhat marginal .
with respect to the influence of @xmath0 , it is found that in general the mean velocity profile sampled in both the laboratory and dynamical frames are more closer to the prandtl - blasius profile at smaller values of @xmath0 than they are at larger @xmath0 , which is consistent with findings from previous dns studies .
moreover , it is found that for smaller values of @xmath0 ( @xmath203 ) the pdf s of the shape factor @xmath29 for the rescaled instantaneous profiles exhibit a peak close to that for the prandtl - blasius profile , whereas for larger values of @xmath0 the peaks shift to smaller values of @xmath29 , indicating the profile s approach to the maximum velocity becomes slower and slower with increasing @xmath0 .
another finding is that the effectiveness of the dynamical scaling method , in terms of its ability of bringing the mean velocity profile closer to that of prandtl - blasius profile , does not have any apparent dependence on @xmath0 .
our general conclusion is that as far as the effectiveness of the dynamical scaling method is concerned the influence of titling angle is much smaller than that of the rayleigh number @xmath0 .
we note that the prandtl - blasius boundary layer theory is a 2d model , so it is perhaps no surprise that the dynamic method works less well in 3d than in 2d .
we would like to thank x .-
d . shang and s .- q .
zhou for kindly making their piv facility available to us and l. qu , y .- c . xie and s .- d .
huang for helping with the experiment .
this work was supported in part by the hong kong research grants council under project nos .
cuhk404409 and cuhk403811 . | we report an experimental study of the properties of the velocity boundary layer in turbulent rayleigh - bnard convection in a cylindrical cell .
the measurements were made at rayleigh numbers @xmath0 in the range @xmath1 and were conducted with the convection cell tilted with an angle @xmath2 relative to gravity , at @xmath3 , @xmath4 , @xmath5 , and @xmath6 , respectively . the fluid was water with prandtl number @xmath7 .
it is found that at small tilt angles ( @xmath8 ) , the measured viscous boundary layer thickness @xmath9 scales with the reynolds number @xmath10 with an exponent close to that for a prandtl - blasius laminar boundary layer , i.e. @xmath11 . for larger tilt angles , the scaling exponent of @xmath9 with @xmath10 decreases with @xmath2 .
the normalized mean horizontal velocity profiles measured at the same tilt angle but with different @xmath0 are found to have an invariant shape .
but for different tilt angles , the shape of the normalized profiles is different .
it is also found that the reynolds number @xmath10 based on the maximum mean horizontal velocity scales with @xmath0 as @xmath12 and the reynolds number @xmath13 based on the maximum rms velocity scales with @xmath0 as @xmath14 , with both exponents do not seem to depend on the tilt angle @xmath2 .
several wall quantities are also measured directly and their dependency on @xmath10 are found to agree well with those predicted for a classical laminar boundary layer .
these are the wall shear stress @xmath15 ( @xmath16 ) , the viscous sublayer @xmath17 ( @xmath18 ) , the friction velocity @xmath19 ( @xmath20 ) and the skin friction coefficient @xmath21 ( @xmath22 ) .
again , all these near - wall quantities do not seem to depend on the tilt angle .
we also examined the dynamical scaling method proposed bys zhou and xia [ _ phys .
rev .
lett . _ * 104 * , 104301 ( 2010 ) ] and found that in both the laboratory and the dynamical frames the mean velocity profiles show deviations from the theoretical prandtl - blasius profile , with the deviations increase with @xmath0 .
but profiles obtained from dynamical scaling in general have better agreement with the theoretical profile .
it is also found that the effectiveness of this method appears to be independent of @xmath0 . |
so far , many statistical theories of nematic liquid crystals have been proposed .
one of the most popular theories of the nematics is maier - saupe ( mst ) [ 1,2 ] which is a mean field study . although mst is successful to describe the nematic - isotropic transition for some nematics , it is not sufficent for some others . in this study , we used generalized maier - saupe theory ( gmst ) to study a nematic liquid crystal , n - p - methoxybenzylidene p - n - butylaniline(@xmath0 ) , where mst is known to fail .
most molecular theories of nematic liquid crystals assume that the constituent molecules are cylindrically symmetric and mst is one of these molecular theories .
this may be a useful approximation .
however , the molecules of real nematogens are lower symmetry
. therefore the dependence of the orientational properties of the uniaxial mesophase on the deviation from molecular cylindrical symmetry was calculated from the series expansion of the pseudopotential [ 3 ] .
this points out that the entropy of the system according to mst is of course not appropriate for the nematics .
because the molecules have lower symmetry , we should use another entropy definition , i.e. another thermostatistics . in this context ,
i use gmst within tt to investigate the nematics .
the aim of this study is to show that the tt may be an appropriate one of these thermostatistics .
we start describing the tsallis thermostatistics ( tt ) .
since the paper by tsallis [ 4 ] , there has been a growing tendency to nonextensive statistical formalism .
it has been shown that this formalism is useful , because it provides a suitable theoretical tool to explain some of the experimental situations , where standard thermostatistics seems to fail , due to the presence of long - range interactions , or long - range memory effects , or multi - fractal space - time constraints .
the axioms of tt are the following : tt has been applied to various concepts of thermostatistics and achieved in solving some physical systems , where boltzmann - gibbs statistics is known to fail [ 5 ] .
recently , mst has been generalized within tt and gmst has been applied to p - azoxyanisole ( @xmath1 ) in [ 6 ] in which kayacan et al .
used the second choice for the internal energy constraint which will be given below .
tt considers three possible choices for the form of a nonextensive expectation value .
these choices have been studied in [ 7 ] and applied to two systems ; the classical harmonic oscillator and the quantum harmonic oscillator . in that study ,
tsallis et al .
studied three different alternatives for the internal energy constraint .
the first choice is the conventional one and used in [ 3 ] by tsallis @xmath2 the second choice is given by @xmath3 and regarded as the canonical one .
both of these choices have been applied to many different systems in the last years [ 8 ] .
however both of them have undesirable difficulties .
the third choice for the internal energy constraint is @xmath4 this choice is commonly considered to study physical systems because it is the most appropriate one , and is denoted as the tsallis - mendes - plastino ( tmp ) choice .
@xmath5 index is called the entropic index and comes from the entropy definition [ 3 ] , @xmath6 where @xmath7 is a constant , @xmath8 is the probability of the system in the @xmath9 microstate , @xmath10 is the total number of configurations . in the limit @xmath11
, the entropy reduces to the well - known boltzmann - gibbs ( shannon ) entropy .
the optimization of @xmath12 leads to @xmath13 ^{\frac{1}{1-q}}}{% z_{q}^{(3)}}\ ] ] with @xmath14 ^{\frac{1}{1-q}}.\ ] ] this equation is an implicit one for the probabilities @xmath15 .
therefore the @xmath16 @xmath17 @xmath18 of an observable is defined as @xmath19 where @xmath20 denotes any observable quantity which commutes with the hamiltonian .
this expectation value recovers the conventional expectation one when @xmath21 .
as mentioned above , eq.(7 ) is an implicit one and in order to solve this equation , tsallis et al .
suggest two different approaches ; @xmath22 @xmath23 and @xmath24 transformation .
this study is the second one which use the gmst to study the nematics .
therefore another aim of this study is to take a step towards the enlargement of the use of tt in liquid crystals and application domain of gmst by using the third choice for the internal energy constraint .
it is worthwhile to note that in the previous paper [ 6 ] , kayacan et al .
used the second choice whereas the third choice is considered in the present study .
it is well - known from the literature that the first two choices have some disadvantages , so we use the third choice which is the most appropriate one , as mentioned above . also , there are some extensions of the mst , including molecular field terms into the potential energy , using bg statistics [ 9,10 ] . in this study ,
long - range interactions have been taken into account for @xmath25 , without including the terms into the interaction potential energy , and the experimental data have been successfully explained for second rank order parameter of @xmath0 , within tt , with a small departure from the standard theory .
mst assumes the intermolecular potential energy as @xmath26 which bases on dispersion forces . in this equation
, @xmath20 is a constant independent of pressure , volume and temperature , @xmath27 is molar volume , and @xmath28 is the angle between the long molecular axis and the preferred axis , @xmath29 is the second rank order parameter and given by @xmath30 in which @xmath29 denotes the legendre polynomial with @xmath31 .
the generalized second rank order parameter is calculated from @xmath32 within tt , where @xmath33 denotes the direction cosine of the long axis of an individual molecule with respect to the nematic axis .
now let us treat the helmholtz free energy of the system and the equilibrium condition within tt .
the helmholtz free energy is written as @xmath34 where @xmath35 thus the entropy of the system can be expressed as @xmath36 in which @xmath37 is given by eq.(6 ) . if substituting eqs.(12 ) and ( 13 ) into eq.(11 ) , the helmholtz free energy associated with the system is obtained .
the condition for the equilibrium in the nematic - isotropic liquid phase transition reads @xmath38 and this condition may be called `` the consistency relation '' .
g. sigaud et al . reported the measurements of the magnetic anisotropy and @xmath39 data of @xmath0 [ 11 ] .
fig.(1 ) shows the second rank order parameter as a function of temperature .
the curves are plotted in nematic range of @xmath0 .
it is seen that the fit to @xmath39 data is very good for @xmath40 , whereas the result of mst is very poor .
the variation of the free energy as a function of the order parameter at nematic - isotropic phase transition temperature is shown in fig.(2 ) .
it is seen that the free energy has two minima , one at @xmath41 representing the isotropic phase , and the other one at @xmath42 corresponding to the nematic phase .
the two states have equal free energies and at this temperature , a discontinuous transition takes place .
the minima corresponding to the nematic phase are at @xmath43 for @xmath21 ( mst ) and at @xmath44 for @xmath45 .
this result is in far better agreement with the experimental one , @xmath46 [ 12 ] than mst .
it is worthwhile to imply that the experimental values of @xmath47 are higher than those predicted by mst for most nematics[13 ] .
the most - known exceptions are @xmath1 and @xmath0 for which mst predictes higher values of @xmath39 than the experimental ones .
therefore , it seems that the @xmath48 curves of most nematics might be expected to be fitted appropriately with @xmath5 values greater than unity .
since the nematics are a class of the liquid crystals , this case is the expected one .
it is seen from this study that mst can be improved by using tt . in this manner , gmst explains the variation of the second rank order parameter and the value of the order parameter at the transition temperature for @xmath0 .
it is seen from the present and previous studies on some kinds of nematics that gmst is a useful tool to study the nematics .
some disadvantages of mst can be improved by using tt .
the author would like to thank u. tirnakli for fruitful comments and suggestions .
[ 1 ] w. maier , a. saupe , z. naturforsch . 13a , 564 ( 1959 ) .
orientational order parameter as a function of temperature in mbba for @xmath45 and @xmath21 .
the curves are plotted in nematic range of @xmath0 .
filled circles represent the experimental data[11 ] . | generalized maier - saupe theory ( gmst ) within tsallis thermostatistics ( tt ) has been used to investigate the second rank order parameter in a nematic liquid crystal , n - p - methoxybenzylidene p - n - butylaniline ( mbba ) .
also , the free energy has been investigated at the nematic - isotropic phase transition temperature and the effect of the nonextensivity has been demonstrated .
pacs number(s ) : 05.20.-y , 05.70.-a , 61.30.cz , 61.30 . gd .
keywords : tsallis thermostatistics , maier - saupe theory , nematic liquid crystals . |
ionization and electron transfer ( electron capture ) , which may occur in collisions between an atom and a bare nucleus , represent two of basic collision processes studied by atomic physics . in the process of ionization
the atom emits an electron , which after the colllision moves freely in space , while in the transfer process an electron initially bound in the atom is captured into a bound state of the moving ion .
both of these processes possess interesting physics , their study is of importance for many applications and various aspects of these processes have been attracting attention for decades .
quite an interesting situation is encountered when a combination of transfer and ionization occurs in a single collision event .
such a process , which becomes possible if the atomic target has at least two electrons , is called transfer - ionization . during the last decade transfer - ionization in collisions of protons with helium atoms
has attracted much attention @xcite-@xcite .
there are only a few known basic mechanisms governing transfer - ionization in fast colllisions . depending on whether the electron - electron interaction ( correlations ) plays in them a crucial role or not , these mechanisms can be subdivided into `` correlated '' and `` uncorrelated '' ones .
the group of uncorrelated mechanisms consists of the so called independent transfer - ionization ( iti ) and capture shake - off ( c - so ) . in the iti electron capture and emission occur due to the `` independent '' interactions of the electrons with the ionic projectile . in a theoretical description
this mechanism appears starting with second order perturbation theory in the ion - atom interaction and for its realization does not need any electron - electron interaction . according to the c
- so mechanism , a fast removal of the captured electron from the atom leads to a `` sudden '' change of the atomic potential in which the other electron is moving . as a result
, the electron tries to adjust its state to the new potential and has a nonzero probability to become unbound @xcite .
the correlated mechanisms are more interesting .
they include so called electron - electron thomas ( eet ) mechanism and a mechanism which will be termed here as electron - electron auger ( eea ) . according to the eet , transfer - ionization proceeds in two steps @xcite , @xcite . in the first step , as a result of a binary collision with the ion , one of the electrons acquires a velocity @xmath2 , where @xmath3 is the ion velocity , moving under the angle of @xmath4 with respect to the motion of the ion . in the second step
this electron scatters on the other electron acquiring , as simple kinematics shows , a velocity equal to the projectile velocity , both in absolute magnitude and direction , that makes the capture very probable .
the same kinematics also tells that the other electron in this process gets a velocity , which is perpendicular to the projectile velocity and whose absolute value is equal to @xmath3 .
thus , as a result of the eet one electron is captured and the other is emitted perpendicular to the projectile motion .
the electron - electron interaction is also the ( main ) driving force of the eea mechanism .
the physics of the latter becomes very transparent when it is viewed in the rest frame of the projectile nucleus .
the functioning of this mechanism is based on the fact that the presence of the second nucleus makes the initial configuration of atomic particles unstable with respect to a kind of auger decay .
indeed , in the presence of this nucleus one of the electrons , which initially belongs to a bound configuration of fast moving particles constituting the atom , undergoes a radiationless transition @xcite into a bound state of the ion by transferring ( most of ) energy excess to the another atomic electron which , as a result of this , is emitted from the atom @xcite , @xcite .
a clear signature of this mechanism is that in the rest frame of the atom the electron is emitted in the direction opposite to the projectile motion @xcite , @xcite , @xcite .
one has to emphasize that the mechanisms for transfer - ionization , discussed above , are in essence high - energy approximations , the validity of which improves with increasing impact energy .
therefore , the description of transfer - ionization in terms of these mechanisms becomes really meaningful only provided the collision velocity is high enough : @xmath5 , where @xmath6 and @xmath7 the typical velocities of the electron(s ) in the initial and final bound states , respectively , and @xmath8 ( @xmath9 ) is the charge of the nucleus of the target ( projectile ) .
this implies that in order to get an insight into the physics of transfer - ionization by considering this process as driven by these mechanisms , even in collisions with protons the impact velocity should lie in the range @xmath10 , where @xmath11 is the bohr velocity in atomic hydrogen .
although transfer - ionization was studied in a number of papers , most of them were concerned with the total cross section .
a better understanidng of the physics of this process can be obtained when differential cross sections are explored .
concerning such cross sections in the case of transfer - ionization in fast collisions only the cross sections singly differential in the momentum component of the emitted electron or the target recoil ion , parallel / antiparallel to the projectile velocity , have been considered ( see e.g. @xcite , @xcite-@xcite ) . however , the exploration of such singly differential cross sections even in principle can hardly allow one to clearly separate the contributions of the correlated and uncorrelated mechanisms ( and thus to study and understand them better ) .
compared to the singly differential cross sections the doubly differential cross sections , which are a function of both parallel and perpendicular to the projectile velocity components of the momentum of the emitted electron , can yield much more information about the process .
therefore , in the present paper we consider such cross sections for transfer - ionization in collisions of fast protons , alpha - particles and lithium nuclei with helium atoms .
it will , in particular , be shown that the study of such doubly differential cross sections may enable one to clearly separate and identify the different mechanisms contributing to transfer - ionization and to get a better insight into the physics of this process .
one should say that all the previous experimental studies devoted to the spectra of electrons emitted in transfer - ionization were dealing with relatively low impact velocities where , as was already mentioned , the discussion of this process in terms of the four mechanisms may not yet be very meaningful . therefore .
we hope that the present article could trigger the interest of experimental physicists to the exploration of this process at higher impact velocities . atomic units ( @xmath12 ) are used throughout the paper except where the otherwise stated .
in our description of transfer - ionization the correlated and uncorrelated mechanisms shall be treated separately ( and added in the cross section incoherently ) .
we begin with considering the correlated ones .
the ( approximate ) transition amplitude for transfer - ionization can be written @xmath13 here @xmath14 is the coulomb interaction between the electrons and @xmath15 and @xmath16 are the initial and final states of the electrons . in the nonrelativistic domain of atomic collisions the description of electron capture
is covariant under a galilean transformation and any galilean frame can be chosen to consider this process .
assuming that the target atom is ( initially ) at rest in the laboratory frame , we take for the moment the rest frame of the atom as our reference frame and choose its origin at the position of the atomic nucleus .
we denote the coordinates of the electrons by @xmath17 and @xmath18 . the projectile - nucleus with a charge @xmath9
is assumed to move along a straight - line trajectory @xmath19 , where @xmath20 is the impact parameter , @xmath21 the collision velocity and @xmath22 the time .
the coordinates of the first and second electrons with respect to the position of the projectile are denoted by @xmath23 and @xmath24 , respectively ( @xmath25 ; @xmath26 ) .
we choose the initial state as @xmath27 in eq.([e2 ] ) @xmath28 is the initial unperturbed two - electron atomic state with an energy @xmath29 and @xmath30 is a factor which accounts for the distortion of the initial atomic state by the field of the incident ion , its form shall be specified later .
we approximate the state @xmath28 according to @xmath31 where @xmath32 is the normalization factor , @xmath33 is the inter - electron distance and the parameters @xmath34 , @xmath35 and @xmath36 can be chosen from the following sets : ( i ) @xmath37 , @xmath38 ; ( ii ) @xmath39 , @xmath38 ; ( iii ) @xmath40 , @xmath41 ; ( iv ) @xmath42 , @xmath43 , @xmath38 ; and ( v ) @xmath44 , @xmath45 , @xmath46 . the final state is taken according to @xmath47 % \nonumber \\ & & \times \exp(-i ( \epsilon_k + \varepsilon_f ) t ) % \exp\left ( -i \frac{v^2}{2 } t \right ) .
\label{e4 } \end{aligned}\ ] ] here , @xmath48 is the final ( unperturbed ) bound state of the electron captured by the projectile , @xmath49 the energy of this state ( as viewed in the rest frame of the projectile ) and @xmath50 the so called translational factor .
further , @xmath51 denotes the state of the emitted electron which moves in the field of the residual atomic ion with ( asymptotic ) momentum @xmath52 and energy @xmath53 and @xmath54 describes the distortions of the states of captured and emitted electrons by the fields of the residual atomic ion and projectile , respectively .
now we turn to the discussion of the form of the distortion factors @xmath55 and @xmath54 .
let us remind the reader that in this paper we consider only collisions at high impact velocities in which one has @xmath56 .
besides , as will be seen below , in the transfer - ionization process the emitted electron has a high velocity ( @xmath57 ) with respect to the projectile . from the work on atomic ionization it is known that in such collisions the account of the distortion does not noticeably changes the result . at the same time
it is also known that for electron transfer reactions the effect of the distortion in general remains very important even at @xmath56 .
therefore , in our treatment we shall ignore the distortions for that electron , which is to be emitted , and account only for the distortions acting on that electron which is to be captured .
with such an assumption one can show that the transition amplitude in the momentum space , @xmath58 is given by @xmath59 here , @xmath60 where @xmath61 and @xmath62 is the momentum transfer in the collision .
note that @xmath63 is obtained from @xmath64 by interchanging @xmath65 and @xmath66 in eqs .
( [ e7])-([e8 ] ) .
the explicit form of the distortion factors is taken according to the continuum - distorted - state ( cdw ) model which has been proved quite successful in describing the total cross section for capture in a three - body collision system ( one active electron + two nuclei ) . in this model
the distortion factors read @xmath67 where @xmath68 , @xmath69 , @xmath70 , and @xmath71 and @xmath72 are the gamma and confluent hypergeometric functions , respectively ( see e.g. @xcite ) .
the inclusion of the distortion factor for the initial state in the form given by the first line of ( [ e10 ] ) means that in our treatment the electron , which is to be transferred , in its initial state moves not only in the field of the atom but also in the ( coulomb ) field of the projectile .
therefore , with such a factor the transition amplitude ( [ e1 ] ) describes both the eea and eet capture channels while when this factor is set to unity , @xmath73 , the calculated contribution of the eea mechanism becomes much larger but the eet mechanism simply `` vanishes '' . to conclude this subsection
let us note that the account of the distortion for the final state turned out to be not so crucial .
indeed , in cases tested the difference between results obtained with the distortion factor @xmath54 in the form given by the second line of ( [ e10 ] ) and by setting @xmath74 was not substantial .
therefore , taking into account that the neglect of this distortion greatly simplifies the calculation reducing the computation time , in what follows we shall report only results obtained when we suppose that @xmath75 .
let us now very briefly consider two uncorrelated mechanisms : the independent transfer - ionization and capture shake - off . according to the first of them transfer - ionization proceeds in two independent steps :
one electron is captured ( transfer ) and the other one is emitted ( ionization ) .
these transitions are driven by the interaction between the projectile and the electrons while the electrons do not need at all to interact with each other for the transitions to occur .
note that within this mechanism the projectile must interact with the target at least twice ( at least one interaction per electron ) . in the consideration of the present paper
the capture and ionization parts of the independent transfer - ionization are regarded as occurring in the collision between a projectile - nucleus and a hydrogen - like system .
the latter is described using an effective nuclear charge which was taken to be @xmath76 , both for capture and ionization . in the impact parameter space
the amplitude for this process is a product of the single - electron transition amplitudes for capture and ionization .
the latter ones are obtained using the three - body cdw ( capture ) and cdw - eis ( ionization ) @xcite models . in capture shake - off the `` instant '' removal of one of the electrons from the atom due to its capture by the fast projectile forces the other electron to react to a sudden change of the atomic potential . as a result , the second electron can be shaked off from the target and become unbound @xcite .
the amplitude for this channel is estimated as the product of the amplitude for single electron capture ( evaluated within the three - body cdw like in case of the iti ) and the amplitude for shake - off which is simply an overlap between the initial and final states of the `` second '' electron . in our calculations
we add the contributions of the correlated , the independent and capture shake - off channels incoherently . in the context of the present paper , which is focused on the correlated capture mechanisms ,
such an incoherent addition does not represent a big shortcoming since at the collision parameters considered here the correlated and uncorrelated have a small overlap in the momentum space of the emitted electrons . to conclude this section note that the validity of our approach to transfer - ionization in fast collisions has been already tested in @xcite-@xcite where the cross sections singly differential in the longitudinal momentum of the emitted electrons and target
recoil ions were calculated for proton on helium collisions at @xmath77 and @xmath78 a.u . and a good agreement with available experimental data has been found @xcite .
in this section we discuss the momentum spectra for electrons emitted in transfer - ionization in collisions of protons , alpha - particles and bare lithium nuclei with helium . as was mentioned in the previous section , in our evaluation of the iti and c - so mechanisms we use the effective charge of @xmath76 to describe the initial undistorted state of the electrons in helium .
therefore , for consistency , in our calculation of the contributions from the correlated mechanisms we use the set ii ) of the parameters for the state ( [ e3 ] ) ( except in figure [ corr - in - init - state ] , where the sets i ) and v ) are used ) .
the momentum spectra shown in figures [ 12 - 1]-[21 - 3 ] are given in the rest frame of the target ( laboratory frame ) and are represented by the doubly differential cross section @xmath79 where @xmath80 and @xmath81 are the longitudinal and transverse parts , respectively , of the momentum @xmath52 of the emitted electron .
the integration in ( [ e11 ] ) runs over the transverse part of the momentum transfer and the azimuthal angle @xmath82 of the emitted electron . in the range of collision parameters considered here an atomic electron
is mainly captured into the ground state of the projectile .
therefore , in what follows we consider transfer - ionization only for this channel .
the momentum spectra of electrons emitted in collisions with protons are displayed in figures [ 12 - 1 ] , [ 16 - 1 ] and [ 21 - 1 ] for impact energies of @xmath0 , @xmath83 and @xmath1 mev , respectively .
these energies correspond to @xmath84 , @xmath85 and @xmath86 a.u .
it is seen in the figures that there are three distinct maxima in the spectra .
the maximum , which is located at small values of @xmath87 , has its origin in the uncorrelated mechanisms : the independent transfer - ionization and capture shake - off . in high - velocity collisions ( @xmath88 ) the cross section for single electron
capture calculated within the cdw approximation scales approximately as @xmath89 . in our model
, this is obviously also the scaling for the contribution of the capture shake - off channel to the cross section .
since the ionization part of the independent transfer - ionization adds the factor @xmath90 , the cross section for this channel is proportional to @xmath91 and , compared to the capture shake - off , shows a steeper dependence both on the projectile charge and collision velocity . according to our model , in collisions with protons ( in the range of impact velocities considered )
the maximum at small @xmath87 is dominated by capture shake - off , that leads to the shape of the spectrum almost symmetric with respect to @xmath92 @xcite .
the situation becomes somewhat different in collisions with alpha - particles and lithium nuclei in which the independent transfer - ionization becomes relatively more important and , as a result , the emission spectrum acquires a slight forward - backward asymmetry with more emitted electrons moving in the forward semi - sphere ( see figures [ 16 - 2 ] , [ 21 - 2 ] and [ 21 - 3 ] ) .
the maximum at large ( negative ) @xmath93 appears due to the eea mechanism whereas the maximum at large @xmath94 is a signature of the eet channel .
\i ) let us consider the kinematics of these two correlated channels of transfer - ionization . to this end
it is convenient to go first to the rest frame of the projectile - nucleus . in this frame
the latter particle does not take part in the energy balance of the process ( because it is heavy and is initially at rest ) .
therefore , the energy balance can be written as @xmath95 . here
, @xmath96 is the velocity of the emitted electron , @xmath97 is the change in energy of the nucleus of the atom with @xmath98 being the change in its longitudinal momentum , @xmath99 is the initial energy of the two incident electrons and we have neglected the initial and final binding energies since @xmath100 and @xmath101 .
thus , the velocity @xmath96 of the emitted electron in the projectile frame is approximately given by @xmath102 \ii ) if the nucleus of the atom would be just a spectator in the collision ( and thus @xmath103 ) , one would obtain @xmath104 . taking into account that in the target frame the electron emitted via the eea mechanism moves in the direction , which is opposite to the projectile velocity , the momentum spectrum of electrons produced via the eea should then be centered in this frame around @xmath105 .
looking at the figures one sees , however , that only at the highest impact energy considered ( @xmath106 a.u . )
the electron spectrum is really having the maximum at the longitudinal momentum rather close to @xmath107 while at the lower velocities ( @xmath84 and @xmath85 a.u . )
this maximum is located at a noticeable distance from the point @xmath108 .
this means that only at sufficiently high impact energies the eea becomes ( almost ) purely electronic mechanism without the involvement of the nucleus of the atom . at lower impact velocities
the target nucleus does noticeably participate in this mechanism ( see also @xcite ) .
i ) \iii ) following the simple picture of the eet mechanism , which was mentioned in the introduction , one would expect that in the rest frame of the target atom the eet is characterized by electrons emitted with a velocity @xmath3 under the angle of @xmath109 with respect to the motion of the projectile .
correspondingly , the velocity of the electron with respect to the projectile should be equal to @xmath2 .
the latter value indeed agrees with the simple estimate for the electron velocity @xmath96 given above in this subsection ( if we assume that @xmath103 ) . however , according to the spectra shown in figures [ 12 - 1]-[21 - 3 ] , in the target frame the velocity @xmath110 of the emitted electron is on average slightly smaller than @xmath3 .
besides , the angle @xmath111 , which characterises the position of the `` center of mass '' of the eet maximum in the momentum spectrum , is somewhat larger than @xmath109 : @xmath112 , where @xmath113 . moreover , the differences @xmath114 and @xmath115 increase with increasing the charge of the projectile and/or decreasing the impact velocity .
the reason for this is that the simple picture does not take into account that the electron which moves together with the projectile is actually not free but bound @xcite . when the binding of the captured electron increases the differences between the result and what the simple picture suggests also growth .
only in the limit @xmath116 does the position of the eet maximum coincide with the prediction of this picture ( see also @xcite ) . assuming that at sufficiently high impact velocities the nucleus of the target atom is a spectator in the eet mechanism
, one can find a simple relation between the averaged velocity @xmath110 of the emitted electron and the angle @xmath115 . indeed , in the rest frame of the projectile the energy of this electron is approximately given by @xmath99 . taking into account that the same energy can also be expressed as @xmath117 , one obtains @xmath118 .
\iv ) two more observations can be drawn from figures [ 12 - 1]-[21 - 3 ] .
first , for a fixed projectile charge state the relative importance of the eet versus eea increases with @xmath3 .
second , for a fixed collision velocity @xmath3 the eea mechanism gains in relative importance when the charge of the projectile increases .
the first observation can be understood noting that the eea and eet are basically the first and second order processes , respectively @xcite , @xcite . as a result
, the eet weakens more slowly with increasing @xmath3 than the eea .
further , the dependence of the eea mechanism on @xmath9 is a bit steeper ( @xmath119 ) @xcite than that of the eet ( @xmath120 ) @xcite .
this enables one to understand the second observation .
both the eea and eet mechanisms crucially rely on the coulomb interaction between the electrons . on the other hand ,
the electron - electron correlations in the initial and final asymptotic states of the colliding system are also manifestation of this interaction @xcite .
the principal difference between them is that while the correlations in the asymptotic states are stationary in nature , the eea and eet are based on the electron - electron interaction in its dynamic variant .
an illustrative example of the correspondence between dynamic and stationary manifestations of basically the same force is represented by the interaction between an electron and its electromagnetic field ( the radiation field ) .
a stationary situation is realized , for instance , when one considers a free ( undistorted ) hydrogen atom in the ground state . in this case
the interaction with the radiation field has quite a weak impact on the system : it merely leads to a tiny shift of the energy of the ground state .
let us consider , however , a situation when the hydrogen atom collides with a fast ion .
now the same interaction may lead to electron transfer from the atom to the ion , which is called radiative electron capture . in this dynamic situation
the interaction with the radiation field leads to a drastic change in the state of the electron .
the difference between the stationary and dynamic manifestations of the electron - electron interaction is not that dramatic .
nevetheless , it is the dynamic electron correlations ( the eea and eet ) , which drive the process of transfer - ionization , whereas the stationary ones providing an `` environment '' also influence the process by determining , for instance , the mean electron - electron distance in the initial atomic state and , thus , the magnitude of the dynamic force acting between the electrons in the transfer process @xcite . these points can be seen in figure [ corr - in - init - state ] where we present the contributions to the momentum spectrum due to the eea and eet mechanisms calculated with two different approximations for the ground state of helium . in the left plot the parameters of the state ( [ e3 ] ) were taken as @xmath121 and @xmath122 which means that the electron - electron correlations in this state are completely ignored .
the right plot was obtained by choosing @xmath44 , @xmath45 and @xmath46 which includes ( in an approximate way ) both the radial and angular correlations in the ground state of helium .
it is seen that there is practically no difference in shape of these two spectra . however , their absolute intensities differ by about a factor of @xmath123 since the electron - electron interaction in the ground state of helium increases the mean electron - nucleus distances in the atom and , thus , decreases the effective strength of the eea and eet channels ( see also @xcite ) . to conclude our discussion in this section note that eq .
( [ velocity ] ) is of course also valid for the uncorrelated mechanisms .
in contrast to the correlated ones , however , in this case we have @xmath124 and , hence , @xmath125 . therefore , it is the nucleus of the atom which balances ( in the projectile frame ) the energy change of the captured electron , @xmath126 , both in the independent transfer - ionization and capture shake - off channels .
we have considered in some detail transfer - ionization in collisions of fast protons , alpha - particles and lithium nuclei with atomic helium . there are four basic mechanisms which are responsible for this process .
two of them ( the independent transfer - ionization and capture shake - off ) are so called uncorrelated mechanisms which means that they would not disappear if the electron - electron interaction would be `` switched off '' . in contrast , this interaction does play a crucial role in the other two ( the electron - electron and electron - electron - thomas ) mechanisms which both are governed by the dynamic electron - electron correlations . our consideration shows that at sufficiently high impact velocities the contributions of the correlated and uncorrelated mechanisms can be clearly separated in the cross section doubly differential in the longitudinal and transverse components of the momentum of the emitted electron .
the study of this cross section also enables one to separate the two correlated mechanisms from each other and get insight into subtle details of the dynamics of transfer - ionization . at high impact energies
@xmath127 the position of the center of the maximum in the momentum spectrum , caused by the eea mechanism , tends in the target frame to @xmath128 .
this means that the role of the nucleus of the atom in this mechanism weakens with increasing collision velocity and the eea eventually becomes a truly electronic one .
however , according to our model , even at impact velocities as high as @xmath129 and @xmath85 a.u . the helium nucleus
still noticeably participates in this process . according to the well known picture of the eet mechanism
the emitted electron should have a velocity equal to the collision velocity @xmath3 and fly under the angle @xmath109 with respect to the projectile motion .
our model predicts , however , that the velocity of the emitted electron is on average smaller than @xmath3 and that the electron is emitted under the angle which is larger than @xmath109 .
these two differences are interconnected and increase if the charge of the projectile increases and/or the impact velocity decreases .
an experimental exploration of the spectra of electrons emitted in transfer - ionization at high impact velocities is very desirable . at the highest velocity ( @xmath130 a.u . ) , considered in this article , the total cross section for transfer - ionization , according to our estimates , is of the order of @xmath131 , @xmath132 and @xmath133 mb in collisions with protons , alpha - particles and lithium nuclei .
these values are of course rather small .
note , however , that already several years ago it was possible ( see @xcite ) to measure the longitudinal momentum spectrum of the recoil target ions for transfer - ionization process with the total cross section of the order of @xmath132 mb .
a.b.v . acknowledges the support from the extreme matter institute emmi and the program for visiting international senior scientists of chinese academy of sciences . in general , using the rest frame of the ion this process can be viewed as a kind of `` auger decay '' in which one of the electrons initially belonging to a fast moving configuration of atomic particles undergoes a radiationless transfer into a bound state of the ion . when the angular correlations are included in the ground state of helium ,
the calculated contribution of the c - so channel to the emission spectrum becomes slighly asymmetric : the maximum of the spectrum is located at @xmath135 but more electrons are emitted in the direction opposite to the projectile motion . | we consider trasfer - ionization in collisions of fast ( @xmath0 @xmath1 mev / u ) protons , alpha - particles and lithium nuclei with helium atoms .
there are just a few basic mechanisms contributing to this process which can be grouped into correlated ones , which crucially depend on the electron - electron interaction , and uncorrelated , which do not need this interaction to proceed .
we show that by exploring momentum spectra of the emitted electrons the correlated and uncorrelated mechanisms can be cleary separated from each other .
this exploration also enables one to get insight into subtle details of the dynamics of transfer - ionization . |
since rubinstein s announcement of an algorithm to recognize the 3-sphere in 1993 @xcite , almost normal surfaces , the main technical innovation of that work , have received wide attention . in 1994
thompson @xcite reworked rubinstein s proof of the existence of an almost normal 2-sphere in @xmath7 using gabai s language of _ thin position _
@xcite . also in 1993 rubinstein announced that any strongly irreducible heegaard spliting could be made almost normal .
this result , combined with his later work with william jaco @xcite , implies that any non - haken 3-manifold admits at most finitely many heegaard splittings of a given genus . in 1996
stocking @xcite provided a new proof of the fact that strongly irreducible heegaard splittings can be made almost normal , by following thompson s use of thin position type arguments .
more recently the author @xcite and , independently , rieck and sedgwick @xcite used almost normal surfaces to prove that for any orientable 3-manifold with torus boundary , @xmath8 , in all but finitely many dehn fillings of @xmath8 the core of the attached solid torus can be isotoped onto every strongly irreducible heegaard splitting .
saul schleimer @xcite has used almost normal surfaces to show that the set of distances of the heegaard splittings of a given 3-manifold is finite .
the main result of this paper is that one may isotope a _
heegaard - scharlemann - thompson _ ( hst ) splitting into an appropriate normal form . roughly speaking ,
an hst splitting of a 3-manifold is a sequence of surfaces such that each consecutive pair cobounds a compression body , and if @xmath9 is such compression body then the subindex of @xmath10 is odd ( a more precise definition is given in section [ s : hst ] ) . for each odd @xmath2 the surface @xmath3
is referred to as a _ thick level _ , and for each even @xmath2 , @xmath3 is a _
thin level_. just as for classical heegaard splittings , the concept of _ strong irreducibility _ is a useful notion of non - triviality for an hst splitting .
the precise statement of our main result is the following : * theorem [ t : main ] * _ suppose @xmath0 ( @xmath6 or @xmath7 ) is an irreducible submanifold of a triangulated 3-manifold which is bounded by a normal or almost normal surface , and contains at most one maximal normal 2-sphere . if @xmath1 is a strongly irreducible hst splitting of @xmath0 then each thin level is isotopic to a normal surface , and each thick level is isotopic to an almost normal surface . _ consequences of theorem [ t : main ] include many of aforementioned results on the existence of almost normal surfaces .
we also show that several new results follow that may be of independent interest . * theorem [ nonsep ] * _ if a triangulated 3-manifold contains a non - separating surface then it contains a non - separating almost normal surface . _ * theorem [ compressible ] * _ suppose @xmath0 is a submanifold of a triangulated irreducible 3-manifold which is bounded by a connected normal surface , and contains at most one maximal normal 2-sphere . if @xmath11 is compressible in @xmath0 then there is an almost normal surface in @xmath0 which is topologically parallel to @xmath11 , and normalizes monotonically to @xmath11 .
_ a result similar to theorem [ compressible ] has been independently obtained by culler , dunfield , jaco and shalen in their program to solve the weak lopez conjecture .
an immediate corollary is the following : * corollary [ normalheegaard ] * _ if @xmath12 is a heegaard surface for a 3-manifold which is isotopic to a normal surface in the complement of a maximal normal 2-sphere then @xmath12 is isotopic to two , topologically parallel , almost normal surfaces .
_ we conclude the paper by proving a result similar to theorem [ t : main ] , which gives an exact characterization of handlebodies in terms of the existence of certain normal and almost normal surfaces .
we hope to use this result in subsequent work to give a new algorithm to determine when an almost normal surface is a heegaard surface .
the techniques used to prove theorem [ t : main ] may be of independent interest . while in this paper we make crucial use of ideas from such earlier papers as @xcite , our arguments differ from those of previous authors in several key respects .
we outline these differences here .
* for the existence of almost normal strongly irreducible heegaard splittings both rubinstein and stocking produce a sequence of surfaces which eventually converge to the desired one . for example , in rubinstein s iterated " sweepout approach he shows that almost normal surfaces occur at the maximum of a sweepout which is in some way minimal " .
but the first sweepout which he chooses may not have the desired surface as a maximum .
so a sequence of sweepouts of smaller and smaller submanifolds is chosen in which the maximum is eventually the desired surface . in the approach here
we show that if the right concept of minimal " is chosen for the initial sweepout then all _ local _ maxima are almost normal , and the desired surface must be one of these . *
several results are used here to streamline the argument which were not available to rubinstein or stocking .
for example , we appeal to scharlemann s _ local detection of strongly irreducible heegaard splittings _ @xcite to classify the intersection of certain surfaces with the interior of a 3-simplex . * rather than using sweepouts ( as does rubinstein ) or thin position ( as do thompson and stocking ) we use the general framework of _ relative _ hst splittings ( see section [ s : relhst ] ) .
we find that this general framework provides a much cleaner approach and further streamlines some of the arguments .
* we deal with 2-sphere components that may be pinched off " in the formation of an almost normal surface in an entirely different way than previous authors .
the author thanks saul schleimer for helpful conversations , especially regarding lemma [ l : untang ] , during the preparation of this paper . the author
also thanks peter shalen .
in this section , we give some of the standard definitions that will be used throughout the paper . the expert in 3-manifold theory may still want to skim this , as one or two new terms are introduced .
a 2-sphere in a 3-manifold which does not bound a 3-ball on either side is called _
essential_. if a manifold does not contain an essential 2-sphere , then it is referred to as _
irreducible_. a loop on a surface is called _ essential _ if it does not bound a disk in the surface .
given a surface , @xmath13 , in a 3-manifold , @xmath0 , a _ compressing disk _ for @xmath13 is a disk , @xmath14 , such that @xmath15 , and such that @xmath16 is essential on @xmath13 . if we let @xmath17 denote a thickening of @xmath18 in @xmath0 , then to _
compress @xmath13 along @xmath18 _ is to remove @xmath19 from @xmath13 , and replace it with @xmath20 . a _ compression body
_ is a 3-manifold which can be obtained by starting with some surface , @xmath13 ( not necessarily connected ) , forming the product , @xmath21 , attaching some number of 2-handles to @xmath22 , and capping off all remaining 2-sphere boundary components with 3-balls .
the boundary component , @xmath23 , is often referred to as @xmath24 .
the other boundary component is referred to as @xmath25 .
if @xmath26 , then we say the compression body is a _
a compression body is _ non - trivial _ if it is not a product . a surface , @xmath13 , in a 3-manifold
, @xmath0 , is a _ heegaard surface for m _ if @xmath13 separates @xmath0 into two compression bodies , @xmath9 , and @xmath27 , such that @xmath28 .
such a splitting is _ non - trivial _ if both @xmath9 and @xmath27 are non - trivial , and _ doubly - trivial _ if both @xmath9 and @xmath27 are trivial .
a separating surface , @xmath13 , in a 3-manifold , @xmath0 , is _ strongly irreducible _ if every compressing disk for @xmath13 on one side intersects every compressing disk for @xmath13 on the other .
[ l : haken ] if @xmath0 contains a non - trivial or doubly - trivial strongly irreducible heegaard surface , then @xmath11 is incompressible in @xmath0 .
lemma [ l : haken ] follows directly from a lemma of haken @xcite .
a _ heegaard - scharlemann - thompson ( hst ) splitting _ of a manifold , @xmath0 , is a sequence of closed , embedded , pairwise disjoint surfaces , @xmath29 , such that 1 . for each odd @xmath2
strictly less than @xmath30 , @xmath3 is a non - trivial or doubly - trivial heegaard splitting of the submanifold of @xmath0 cobounded by @xmath4 and @xmath5 .
if @xmath30 is odd , then there is a non - trivial compression body , @xmath9 , such that @xmath31 , and @xmath32 .
3 . @xmath33 . for each odd @xmath2 the surface @xmath3
is referred to as a _ thick level _ , and for each even @xmath2 , @xmath3 is a _
thin level_. if @xmath9 is a compression body , then @xmath34 is an hst splitting of @xmath9 .
the only thick level of this hst splitting is @xmath10 , and the only thin level is @xmath35 .
if @xmath0 is a closed 3-manifold , and @xmath13 is a heegaard surface for @xmath0 , then an hst splitting of @xmath0 is @xmath36 , with a unique thick level , @xmath13 .
if @xmath37 , then @xmath38 is an hst splitting of @xmath0 .
an hst splitting , @xmath29 , is _ strongly irreducible _ if for each odd @xmath2 strictly less than @xmath30 , the thick level , @xmath3 , is strongly irreducible in the submanifold of @xmath0 cobounded by @xmath4 and @xmath5 .
[ d : order ] for any surface , @xmath13 , let @xmath39 , where @xmath40 are the components of @xmath13 . if @xmath41 and @xmath42 denote compact , embedded surfaces in a 3-manifold , @xmath0 , then we say @xmath43 if @xmath44 .
note that this ordering is defined so that if @xmath41 is obtained from @xmath42 by a compression , then @xmath43 . for any hst
splitting , @xmath45 , let @xmath46 denote the ordered set @xmath47 is a thick level@xmath48 , where repeated integers are included , and the ordering is non - increasing . if @xmath49 and @xmath50 denote hst splittings of some 3-manifold , @xmath0 , then we say @xmath51 if @xmath52 , where the comparison is made lexicographically .
every 3-manifold admits a minimal hst splitting .
choose any hst splitting of @xmath0 .
now consider any decreasing sequence of hst splittings of @xmath0 that begins with the chosen one .
this sequence must be finite , since any decreasing sequence of sets of non - negative integers terminates under the lexicographical ordering .
the following two theorems of scharlemann and thompson originally appear in @xcite , with slightly different terminology .
one can find proofs of both that are more consistent with the language presented here in @xcite .
[ t : st ] [ scharlmann - thompson @xcite ] a minimal hst splitting is strongly irreducible .
[ t : stinceven ] [ scharlmann - thompson @xcite ] each thin level of a strongly irreducible hst splitting is incompressible .
to proceed , we must first understand arcs in a compression body . if @xmath9 is a compression body , recall that @xmath9 can be built by starting with a product , @xmath21 , and attaching 2- and 3-handles to @xmath22 . anything that remains of @xmath22 after
the attachment becomes part of @xmath35 .
we say a properly embedded arc , @xmath54 , is _ straight _ in @xmath9 if @xmath55 , where @xmath56 is a point such that @xmath57 .
we are now ready to generalize the definition of a compression body : [ d : kcompbod ] a _ @xmath58-compression body _ , @xmath59 , is 1 .
a 3-manifold , @xmath9 , which can be obtained by starting with some surface , @xmath13 ( not necessarily connected ) , forming the product , @xmath21 , attaching some number of 2-handles to @xmath22 , and capping some ( but not necessarily all ) remaining 2-sphere boundary components with 3-balls .
the boundary component , @xmath23 , is @xmath60 , and the other boundary component is @xmath35 .
2 . a 1-manifold , @xmath61 , such that 1 .
@xmath58 is a disjoint union of embedded arcs 2 .
each arc of @xmath58 has at least one endpoint on @xmath10 3 .
if @xmath54 is an arc of @xmath58 with @xmath62 , then there is a disk , @xmath63 , with @xmath64 , @xmath65 , and @xmath66 .
if @xmath54 is an arc of @xmath58 with one endpoint on @xmath10 , then @xmath54 is straight .
5 . for each 2-sphere component , @xmath67 , of @xmath68 , @xmath69 .
if @xmath67 is a 2-sphere component of @xmath10 then either @xmath67 does not bound a 3-ball component of @xmath9 or @xmath70 .
a @xmath58-compression body , @xmath59 , is _ non - trivial _ if either @xmath9 is not a product , or at least one arc of @xmath58 is not straight . if @xmath58 is a 1-manifold which is properly embedded in a 3-manifold , @xmath0 , then a _ heegaard splitting of the pair _
@xmath53 is an expression of @xmath0 as a union of @xmath71-compression bodies , @xmath72 and @xmath73 , such that @xmath74 , and @xmath75 .
such a splitting is _ non - trivial _ if both @xmath72 and @xmath73 are non - trivial .
when the context is clear , we will refer to the surface , @xmath60 , as a _
heegaard surface _ of @xmath53 .
if @xmath58 is a 1-manifold which is properly embedded in a 3-manifold , @xmath0 , then a _ heegaard - scharlemann - thompson ( hst ) splitting of the pair _ @xmath53 is a sequence of closed , embedded , pairwise disjoint surfaces , @xmath76 , such that 1 . for each odd @xmath77 between 1 and @xmath30
, there is a non - trivial @xmath78-compression body , @xmath79 , such that @xmath80 , and @xmath81 .
2 . for each
even @xmath77 between 2 and @xmath30 , there is a non - trivial @xmath82-compression body , @xmath83 , such that @xmath84 , and @xmath85 .
3 . @xmath86 4 .
@xmath87 . for each odd @xmath77 the surface @xmath88 is referred to as a _ thick level _ , and for each even @xmath77
, @xmath88 is a _
thin level_. [ e : thin ] suppose that @xmath89 is an arbitrary knot or link with no trivial components , and @xmath90 is some standard height function on @xmath7 ( so that for each @xmath91 , @xmath92 is a 2-sphere ) , which is a morse function when restricted to @xmath58 .
let @xmath93 denote the critical values of @xmath90 restricted to @xmath58 , and let @xmath94 be some point in the interval @xmath95 .
the following terminology is standard in _ thin position _ arguments ( see @xcite ) .
if @xmath77 is such that @xmath96 and @xmath97 , then we say the surface @xmath98 is a _ thick level _ of @xmath58 .
similarly , if @xmath99 and @xmath100 , then we say the surface @xmath98 is a _ thin level _ of @xmath58 .
suppose there are @xmath30 thick levels for @xmath58 .
let @xmath101 , @xmath102 denote the set of thick levels of @xmath58 , and @xmath103 denote the set of thin levels
. then @xmath104 is an hst splitting of @xmath105 , and the thick and thin levels of this hst splitting are precisely the thick and thin levels of @xmath58 .
if @xmath106 is some properly embedded 1-manifold then let @xmath107 denote @xmath0 with a regular neighborhood of @xmath58 removed . if @xmath8 is any subset of @xmath0 then let @xmath108 .
suppose @xmath13 is a surface in a 3-manifold @xmath0 , and @xmath58 is some properly embedded 1-manifold in @xmath0 , transverse to @xmath13 .
a _ relative compressing disk _ for @xmath13 is a disk , @xmath18 , such that @xmath109 , where @xmath110 , and @xmath111 .
if there are no relative compressions for some surface , then we say that it is _
relatively incompressible_. to _ perform a relative compression _ is to use a relative compressing disk to guide an isotopy , thereby reducing the number of intersections with @xmath58 by exactly two .
[ d : weakreduce ] suppose @xmath13 is a separating surface in a 3-manifold , @xmath0 , and @xmath58 is some properly embedded 1-manifold in @xmath0 , transverse to @xmath13 .
then @xmath13 is _ weakly reducible with respect to @xmath58 _ if there exist disks , @xmath18 and @xmath112 , on opposite sides of @xmath13 , such that either 1 .
@xmath18 and @xmath112 are disjoint compressing disks for @xmath113 in @xmath107 2 .
@xmath18 and @xmath112 are relative compressing disks for @xmath13 that are either disjoint , or meet in a point of @xmath58 3 .
@xmath18 is a relative compressing disk for @xmath13 , and @xmath112 is a disjoint compressing disk for @xmath113 in @xmath107 4 .
@xmath18 is a compressing disk for @xmath113 in @xmath107 , and @xmath112 is a disjoint relative compressing disk for @xmath13 .
we say @xmath13 is _ strongly irreducible with respect to @xmath58 _ if it is not weakly reducible with respect to @xmath58 .
[ e : thin3 ] let @xmath58 , @xmath90 , and @xmath114 be as in example [ e : thin ] .
the _ width _ of @xmath58 is defined to be the quantity @xmath115 .
a knot is said to be in _ thin position _ if @xmath90 is chosen so that the width of @xmath58 is minimal ( see @xcite ) .
recall from example [ e : thin ] that any choice of @xmath90 induces an hst splitting , @xmath116 , of @xmath105 .
if , when @xmath90 is chosen so as to minimize the width of @xmath58 , it turns out that @xmath117 ( _ i.e. _ @xmath58 is also in _ bridge position _
with respect to @xmath90 ) , then we claim that @xmath118 is a strongly irreducible heegaard splitting of @xmath105 .
suppose this were not the case .
then @xmath118 is weakly reducible with respect to @xmath58 .
let @xmath18 and @xmath112 be a pair of disks , as in the above definition of weak reducibility .
we now examine each of the possible cases in that definition : 1 .
_ @xmath18 and @xmath112 are relative compressing disks for @xmath118 that are either disjoint , or meet in a point of @xmath58 .
_ then there is an isotopy of @xmath58 which reduces its width , as depicted in figure [ f : thinmove ] .
+ = 2 in 2 . _
@xmath18 is a relative compressing disk for @xmath118 , and @xmath112 is a disjoint compressing disk for @xmath119 in @xmath107 .
_ then @xmath120 bounds disks , @xmath121 and @xmath122 , on @xmath118 ( since @xmath118 is a sphere ) .
one of these disks ( @xmath121 , say ) is disjoint from @xmath18 .
as @xmath112 is a compressing disk for @xmath119 , it must be that there is a relative compressing disk inside the sphere bounded by @xmath123 .
however , this relative compressing disk is disjoint from @xmath18 , so we are reduced to case 1 .
@xmath18 is a compressing disk for @xmath119 in @xmath107 , and @xmath112 is a disjoint relative compressing disk for @xmath13 .
_ this case and the previous one are completely symmetric .
@xmath18 and @xmath112 are disjoint compressing disks for @xmath119 in @xmath107 .
_ as in case 2 , @xmath120 bounds disks , @xmath121 and @xmath122 on @xmath118 . also , @xmath16 bounds disks , @xmath124 and @xmath125 on @xmath118 .
since @xmath126 , one of the disks , @xmath121 or @xmath122 is disjoint from one of the disks , @xmath124 or @xmath125 .
say @xmath127 . as in case 2
, there is a relative compression , @xmath128 , inside the ball bounded by @xmath123 , and a relative compression , @xmath12 , inside the ball bounded by @xmath129 .
but @xmath130 , so we are again reduced to case 1 above . in lemma [
l : untang ] we will present a generalization of this example for graphs in 3-manifolds , as opposed to 1-manifolds in @xmath7 .
proofs of the next two lemmas originally appeared in @xcite with different terminology .
one can find proofs of both that are more consistent with the language presented here in @xcite .
[ l : essential ] if @xmath13 is a strongly irreducible heegaard surface of @xmath53 and @xmath131 is an essential surface then @xmath13 can be isotoped rel @xmath58 so that no arc of @xmath132 cobounds a relative compressing disk for @xmath13 , and so that no loop of @xmath132 bounds a compressing disk for @xmath113 in @xmath107 .
[ l : relhaken ] if @xmath13 is a strongly irreducible heegaard surface of @xmath53 then @xmath133 is incompressible in @xmath107 , and @xmath11 is relatively incompressible in @xmath0 .
an hst splitting , @xmath76 of a pair , @xmath53 , is _ strongly irreducible _
if for each odd @xmath77 strictly less than @xmath30 , the thick level @xmath88 is strongly irreducible with respect to @xmath58 in the submanifold of @xmath0 cobounded by @xmath134 and @xmath135 .
we now present one further generalization of the concept of a hst splitting .
if @xmath136 is any graph then let @xmath137 and @xmath138 denote the sets of vertices and edges of @xmath136 .
if we are discussing a particular graph , @xmath136 , embedded in a 3-manifold , @xmath0 , then @xmath139 will denote @xmath0 with a regular neighborhood of @xmath137 removed .
let @xmath136 be a properly embedded graph in a 3-manifold , @xmath0 .
we define an _ hst splitting of @xmath140 _ to be an hst splitting of @xmath141 .
if @xmath136 is a graph properly embedded in an irreducible 3-manifold , @xmath0 , and we are given an hst splitting of @xmath140 , then we can define an hst splitting of @xmath0 by forgetting " @xmath136 .
[ d : underlyinghst ] if @xmath76 is an hst splitting of @xmath140 , where @xmath0 is an irreducible 3-manifold other than @xmath142 or @xmath7 , then we define its _ underlying hst splitting _ , @xmath143 $ ] , as follows : 1 . let @xmath144 be the onto , monotone function , such that @xmath145 iff the submanifold of @xmath0 co - bounded by the non-@xmath146 components of @xmath147 and @xmath88 is a product .
2 . for each @xmath2 between 0 and @xmath148 , choose some @xmath149 , and let @xmath3 denote the non-@xmath146 components of @xmath88 .
3 . let @xmath150 be the maximal subset of @xmath151 such that @xmath152 is an hst splitting of @xmath0 .
4 . define @xmath143=\{f_i\}_{i \in \sigma}$ ] .
we leave it to the reader to check that @xmath143 $ ] is well defined up to isotopy .
let @xmath136 be a properly embedded graph in a 3-manifold , @xmath0 .
for any surface , @xmath153 , let @xmath154 , where @xmath40 are the components of @xmath13 , and @xmath155 denotes @xmath156 with a neighborhood of @xmath136 removed .
if @xmath41 and @xmath42 denote compact , embedded surfaces in @xmath0 , then we say @xmath157 if @xmath158 .
in the next lemma we relate the thick and thin levels of an hst splitting of a pair to those of its underlying hst splitting
. this will be an important step in the eventual production of almost normal surfaces .
[ l : underlyingodd ] let @xmath136 be a properly embedded graph in an irreducible 3-manifold @xmath0 other than @xmath7 or @xmath142 .
let @xmath116 be an hst splitting of @xmath140 , and suppose @xmath159 $ ] .
then for each thick ( thin ) level , @xmath160 , of @xmath1 there is a thick ( thin ) level of @xmath116 which becomes parallel to @xmath160 when all @xmath146 components are removed .
let @xmath161 and @xmath150 be as given in definition [ d : underlyinghst ] .
let @xmath162 be the @xmath163 element of @xmath150 , for some odd number , @xmath54 .
@xmath164 is parallel to the non-@xmath146 components of @xmath88 , for all @xmath165 . since @xmath161 is monotone , @xmath166 either contains an odd number , or contains only one element
. assume the latter is true , and let @xmath167 .
the assumption that @xmath166 has only one element implies the non-@xmath146 components of @xmath168 , @xmath169 , and @xmath170 ( which we denote @xmath171 , @xmath172 , and @xmath173 ) are non - parallel .
now , let @xmath174 denote the predecessor of @xmath162 in @xmath150 . since @xmath54 is odd , @xmath175 . since @xmath176 is even , @xmath177 . as @xmath178 , we know that @xmath172 and @xmath171 are not parallel , so it must be that @xmath179 .
but @xmath178 also implies that @xmath180 is parallel to @xmath171 . since @xmath174 is the predecessor of @xmath162 in @xmath150 , @xmath181 .
since @xmath150 is maximal , @xmath175 implies @xmath182 . this is a contradiction , as @xmath183
, @xmath180 is parallel to @xmath171 , and @xmath179 . the proof in the case that @xmath54 is even is completely analogous .
in this section we examine the properties of strongly irreducible hst splittings of pairs and prove an important theorem about their existence .
we begin with a result which is completely analogous to theorem [ t : stinceven ] .
its proof is similar to that of theorem 7.4 . of @xcite .
[ t : minwrtk ] let @xmath58 be a properly embedded 1-manifold in a 3-manifold , @xmath0 .
each thin level of a strongly irreducible hst splitting of @xmath53 is incompressible in @xmath107 , and relatively incompressible in @xmath0 .
let @xmath184 be a strongly irreducible hst splitting of @xmath53 .
for each odd @xmath77 , let @xmath185 denote the submanifold of @xmath0 cobounded by @xmath134 and @xmath135 .
suppose @xmath186 is a loop that bounds a compressing disk , or an arc which cobounds a relative compressing disk , for some surface , @xmath187 , where @xmath188 is even .
if @xmath186 is a loop , then let @xmath189 denote a disk in @xmath0 , such that @xmath190 .
if @xmath186 is an arc , then let @xmath189 be a disk in @xmath0 such that @xmath191 , where @xmath192 . in either case
, choose @xmath189 so that @xmath193 is minimal .
it follows from lemma [ l : relhaken ] that for each odd @xmath77 , @xmath194 is incompressible and relatively incompressible in @xmath185 .
hence , @xmath189 can not lie entirely in @xmath195 or @xmath196 .
we conclude then that there is some loop or arc of intersection of the interior of @xmath189 with @xmath197 .
let @xmath198 denote an innermost such loop .
let @xmath199 denote the subdisk of @xmath189 bounded by @xmath198 .
@xmath199 lies in @xmath200 , for some odd number , @xmath201 .
as @xmath202 is incompressible in @xmath200 , @xmath198 must bound a disk , @xmath203 , on @xmath202 .
now , let @xmath204 be an innermost loop of @xmath205 , and let @xmath206 be the subdisk of @xmath203 bounded by @xmath204 .
then we can use @xmath206 to surger @xmath189 , and thereby obtain a new disk , with the same boundary as @xmath189 , contradicting our minimality assumption .
we conclude then that @xmath186 can not be a loop , and if @xmath186 is an arc , then @xmath189 contains no loops of intersection with @xmath197 .
let @xmath207 then denote an arc of intersection which is outermost on @xmath189 .
@xmath207 and a subarc of @xmath198 cobound a subdisk , @xmath208 , of @xmath189 .
@xmath208 is then a relative compressing disk for @xmath194 , for some @xmath77 , contradicting lemma [ l : relhaken ] . if @xmath136 is a properly embedded graph in an irreducible 3-manifold , @xmath0 , then each thin level of a strongly irreducible hst splitting of @xmath140 is incompressible in @xmath209 , and relatively incompressible in @xmath139 . for any hst
splitting , @xmath45 , of @xmath140 , let @xmath210 denote the ordered set @xmath211 is a thick level of @xmath212 , where repeated integers are included , and the ordering is non - increasing . if @xmath49 and @xmath50 denote hst splittings of @xmath140 , then we say @xmath213 if @xmath214 , where the comparison is made lexicographically .
[ l : untang ] let @xmath136 be a properly embedded graph in an irreducible 3-manifold , @xmath0 .
if @xmath1 is a strongly irreducible hst splitting of @xmath0 then there exists a strongly irreducible hst splitting , @xmath116 , of @xmath140 such that @xmath143=\{f_i\}$ ] . among all hst splittings of @xmath140
whose underlying hst splitting is @xmath1 , choose one , @xmath184 , that is smallest .
we would like to show that @xmath116 is strongly irreducible . for each odd @xmath77
strictly less than @xmath148 , let @xmath185 denote the submanifold of @xmath0 cobounded by @xmath134 and @xmath135 .
let @xmath215 be a thick level , so that @xmath201 is odd .
we need to show that @xmath215 is strongly irreducible with respect to @xmath216 in @xmath200 . if not
, then there are disks , @xmath18 and @xmath112 , in @xmath200 , as in definition [ d : weakreduce ] .
note that we may assume that @xmath16 does not bound a disk , @xmath124 , on @xmath215 such that @xmath217 bounds a ball containing a single unknotted arc of @xmath138 .
if this were the case then there is a relative compressing disk , @xmath218 , inside the ball bounded by @xmath219 . as @xmath220
, we may use @xmath218 instead of @xmath18 .
we make a similar choice for @xmath112 . our reason for such choices for @xmath18 and @xmath112 will be made clear shortly .
let @xmath221 , @xmath222 , and @xmath223 be the surfaces obtained from @xmath215 by ( relative ) compression along @xmath18 , ( relative ) compression along @xmath112 , and ( relative ) compression along both @xmath18 and @xmath112 . by our choice of @xmath18 from the previous paragraph
we know that if some component @xmath67 of @xmath221 bounds a ball in @xmath200 then @xmath224 .
this a necessary condition for @xmath221 and @xmath225 to cobound a @xmath138-compression body .
( see condition 2(f ) of definition [ d : kcompbod ] . )
we now produce a new hst splitting from @xmath116 .
there are four cases : 1 .
@xmath226 , @xmath227 . + remove @xmath215 from @xmath116 . in its place , insert @xmath228 and reindex .
2 . @xmath229 , @xmath227 .
+ replace @xmath230 with @xmath231 .
@xmath226 , @xmath232 .
+ replace @xmath233 with @xmath234 .
4 . @xmath229 , @xmath232 .
+ replace @xmath235 with @xmath223 and reindex . in all cases
it is a routine matter to check that we have defined a new hst splitting of @xmath140 which is smaller than @xmath116 .
what remains to be checked is that the underlying hst splitting is still @xmath1 .
we claim that the only way in which this can fail is if the non-@xmath146 components of both @xmath221 and @xmath222 are not parallel to the non-@xmath50 components of @xmath215 .
the proof of this assertion is similar in all four cases .
we present a proof for the first ( and most difficult ) case and leave the remaining cases as exercises .
let @xmath150 be as in definition [ d : underlyinghst ] .
if the non-@xmath146 components of two surfaces , @xmath203 and @xmath236 , are parallel then we write @xmath237 .
suppose that @xmath238 .
then @xmath18 is either a relative compressing disk , or @xmath16 bounds a disk on @xmath215 that meets @xmath136 ( and hence @xmath16 is essential on @xmath239 , but inessential on @xmath215 ) .
since @xmath223 is obtained from @xmath222 by ( relative ) compression along @xmath18 , we conclude that @xmath240 . in case 1 above , the new hst splitting that we construct looks like @xmath241 as @xmath242 , and @xmath243 , the underlying hst splitting of this is the same as the underlying hst splitting of the following sequence : @xmath244 since @xmath112 is a disk in the submanifold of @xmath0 cobounded by @xmath215 and @xmath245 we will not include the index of @xmath222 in @xmath150 when forming the underlying hst splitting of this sequence .
hence , the underlying hst splitting of this last sequence is the same as that of @xmath116 .
what we have established is that if the underlying hst splitting of the new hst splitting of @xmath246 constructed above is not equal to @xmath1 then both @xmath18 and @xmath112 were compressing disks for @xmath215 . as @xmath215 is parallel to some thick level of @xmath1 ( by lemma [ l : underlyingodd ] ) , and
@xmath18 and @xmath112 are disjoint , this contradicts our assumption that @xmath1 was strongly irreducible .
in this section , we discuss the necessary background material on normal surfaces .
a _ normal curve _ on the boundary of a tetrahedron is a simple loop which is transverse to the 1-skeleton , made up of arcs which connect distinct edges of the 1-skeleton .
the _ length _ of such a curve is simply the number of times it crosses the 1-skeleton .
a _ normal disk _ in a tetrahedron is any embedded disk , whose boundary is a normal curve of length three or four , and whose interior is contained in the interior of the tetrahedron , as in figure [ f : normal ] .
= 3.25 in a _ normal surface _ in @xmath0 is the image of an embedding , @xmath201 , of some closed surface , @xmath13 , into @xmath0 , such that @xmath247 is a union of normal disks .
in addition , we say @xmath247 is an _ almost normal surface _ if it consists of all normal disks , plus one additional piece in one tetrahedron .
this piece can be either a disk with normal boundary of length 8 ( depicted in figure [ f : octagon ] ) , or two normal disks connected by a single unknotted tube ( as in figure [ f:1tube ] ) .
almost normal surfaces were first explored by rubinstein in @xcite , and later used by thompson @xcite and stocking @xcite . in @xcite
, we produce new applications of almost normal surfaces by generalizing them to include surfaces with non - empty boundary .
one application of the results we have discussed thus far comes about when we let @xmath136 be the 1-skeleton of a triangulation of a 3-manifold , @xmath0 .
if @xmath249 is such a triangulation then we denote the @xmath30-skeleton of @xmath249 by @xmath250 .
we now focus on strongly irreducible hst splittings of @xmath248 .
[ bubble ] suppose @xmath13 is an embedded surface in @xmath0 , equipped with a triangulation , @xmath249 .
a _ bubble _ for @xmath13 is a ball , @xmath236 , such that @xmath251 , where @xmath252 and @xmath253 are disks , @xmath252 is contained in a single tetrahedron , @xmath254 , @xmath255 , and @xmath256 .
[ l : nobubbles ] suppose @xmath0 is a submanifold of a 3-manifold , @xmath257 , with triangulation , @xmath249 , which is bounded by a normal or almost normal surface .
let @xmath116 be a strongly irreducible hst splitting of @xmath248 .
then we may isotope @xmath116 so that for all @xmath77 , @xmath88 does not contain any bubbles .
suppose @xmath236 is a bubble for @xmath88 , for some @xmath77 , where @xmath258 , as in definition [ bubble ] .
we can use @xmath236 to guide an isotopy from @xmath253 to @xmath252 .
this may push other surfaces of @xmath116 which had non - empty intersection with @xmath259 , but it can only destroy bubbles for those surfaces , too . the isotopy is supported on a neighborhood of @xmath236 , which is disjoint from @xmath260 . hence ,
if @xmath116 was strongly irreducible before the isotopy then it will remain so .
since there are a finite number of surfaces in @xmath116 , and a finite number of bubbles for each , we eventually arrive at an hst splitting with the desired properties .
[ t : normal ] suppose @xmath0 is a submanifold of a 3-manifold , @xmath257 , with triangulation , @xmath249 , which is bounded by a normal or almost normal surface .
let @xmath116 be a strongly irreducible hst splitting of @xmath248 .
then we may isotope @xmath116 so that each thin level is a normal surface .
we begin by using lemma [ l : nobubbles ] to isotope @xmath116 , removing all bubbles .
now , let @xmath77 be some even number , let @xmath261 be some tetrahedron in @xmath249 , and let @xmath262 be a face of @xmath261 .
first , we examine the possibilities for @xmath263 .
let @xmath264 be an innermost simple closed curve , bounding a disk , @xmath252 in @xmath262 . by theorem [ t : minwrtk ] , @xmath264 must bound a disk , @xmath253 , in @xmath88 .
@xmath265 is irreducible ( it s a handlebody ) , so @xmath266 bounds a bubble for @xmath88 .
this is a contradiction , so we see no simple closed curves in any face .
if there are any curves which run from one edge of @xmath262 to itself , then there is an outermost such one .
let @xmath18 denote the sub - disk it cuts off in @xmath262 .
then @xmath18 is a relative compressing disk for @xmath88 , also contradicting theorem [ t : minwrtk ] .
we conclude that @xmath263 is a collection of normal arcs .
we now consider the possibilities for @xmath267 .
it is easy to show that the only possibilities for normal loops are curves of length 3 , or 4n ( see , for example , @xcite ) .
if there are any curves of length greater than 4 , then there must be a disk , @xmath18 , such that @xmath109 , where @xmath268 , and @xmath269 ( see @xcite ) .
this is a relative compressing disk for @xmath88 , which is again a contradiction .
we conclude that @xmath267 consists of normal loops of length 3 and 4 .
finally , it follows from theorem [ t : minwrtk ] that every loop of @xmath267 bounds a disk on @xmath88 .
since we have already ruled out simple closed curves in faces of @xmath261 , such disks must lie entirely inside @xmath261 .
we conclude @xmath88 is a normal surface .
our goal now is to show that once bubbles are removed from the thick levels of @xmath116 , they become almost normal in @xmath0 .
first , we shall need a few lemmas .
[ l : nlarcs ] suppose @xmath0 is a submanifold of a 3-manifold , @xmath257 , with triangulation , @xmath249 , which is bounded by a normal or almost normal surface .
let @xmath184 be a strongly irreducible hst splitting of @xmath248 , where @xmath11 normal implies @xmath148 is even , and @xmath11 almost normal implies @xmath148 is odd .
then we may isotope @xmath116 , so that each thin level is a normal surface , and so that for each thick level , @xmath88 , @xmath263 is a collection of normal arcs , for every 2-simplex , @xmath262 , of @xmath249 .
we begin by appealing to lemma [ l : nobubbles ] and theorem [ t : normal ] to isotope @xmath116 so that there are no bubbles for any of its surfaces , and so that each thin level is normal .
we now produce a further isotopy of @xmath116 , which fixes all thin levels , and which results in an hst splitting with all of the desired properties .
the isotopy will be defined individually for each odd value of @xmath77 ( _ i.e. _ , for each thick level ) .
let @xmath77 be some odd number , and let @xmath185 denote the submanifold of @xmath0 cobounded by @xmath134 and @xmath135 .
the strong irreducibility of @xmath116 implies that @xmath88 is a strongly irreducible heegaard surface for @xmath270 .
lemma [ l : essential ] now implies that a standard innermost disk / outermost arc argument can be used to isotope @xmath88 so that no arc of @xmath271 cobounds a relative compressing disk for @xmath88 in @xmath185 , and so that no loop of @xmath271 bounds a compressing disk for @xmath272 in @xmath273 . now
, suppose there are non - normal components of @xmath263 , for some 2-simplex , @xmath262 .
if there are loops of @xmath263 , then let @xmath264 be a loop which is innermost on @xmath262 .
then @xmath264 bounds a subdisk , @xmath18 , of @xmath262 .
but , as no loop of @xmath271 bounds a compressing disk for @xmath272 in @xmath273 , it must be that @xmath264 also bounds a disk , @xmath124 , on @xmath272 .
but then @xmath129 bounds a bubble for @xmath88 . as all such bubbles have been removed , this is a contradiction .
we conclude then that if there are non - normal components of @xmath263 , then they must be arcs which run from some edge of @xmath260 back to itself .
but then if @xmath198 is such an arc which is outermost on @xmath262 , then @xmath198 cobounds a subdisk of @xmath262 which is a relative compressing disk for @xmath88 , a contradiction .
[ 348 ] let @xmath274 and @xmath116 be as given by the conclusion of lemma [ l : nlarcs ] .
then for each odd @xmath77 , @xmath88 meets the boundary of every tetrahedron in normal curves of length 3 , 4 , and at most one curve on at most one tetrahedron of length 8 .
this lemma is taken straight from @xcite .
we refer the reader to this paper for its proof .
the necessary assumptions are that @xmath88 meets every tetrahedron in normal arcs , and that there is no pair of disks which form a weak reduction for @xmath88 .
a connected surface in a ball , @xmath236 , is _ unknotted _ if it is isotopic to a neighborhood of the graph obtained by coning @xmath30 points on @xmath275 to a point in the interior of @xmath236 . for the proof of theorem [ t : an ]
we will need the following result : [ t : sch ] [ scharlemann @xcite ] suppose @xmath13 is a strongly irreducible heegaard surface for a 3-manifold , @xmath0 , and @xmath236 is a ball in @xmath0 . if @xmath276 is incompressible in @xmath277 , then @xmath278 is connected and unknotted . to prove theorem [ t : an ]
, we will need to restate the previous result for heegaard surfaces of @xmath53 : [ t : relsch ] suppose @xmath13 is a strongly irreducible heegaard surface of @xmath279 and @xmath236 is a ball in @xmath107 . if @xmath276 is incompressible in @xmath280 , then @xmath278 is connected and unknotted .
the proof of this theorem is exactly the same as scharlemann s proof of theorem [ t : sch ] .
the point is that there is very little difference between a heegaard surface for @xmath0 and a heegaard surface for @xmath53 , as viewed from inside a ball that lies entirely in the complement of @xmath58 .
[ t : an ] suppose @xmath0 is a submanifold of a 3-manifold , @xmath257 , with triangulation , @xmath249 , which is bounded by a normal or almost normal surface .
let @xmath184 be a strongly irreducible hst splitting of @xmath248 , where @xmath11 normal implies @xmath148 is even , and @xmath11 almost normal implies @xmath148 is odd .
then we may isotope @xmath116 , so that each thin level is a normal surface and each thick level is almost normal .
isotope @xmath116 to satisfy the conclusion of lemma [ l : nlarcs ] .
let @xmath77 be some odd number , so that @xmath88 is a thick level .
let @xmath261 be some tetrahedron in @xmath249 .
let @xmath67 be a copy of @xmath281 , pushed slightly into @xmath261 .
we first claim that @xmath282 is incompressible in @xmath265 , to the outside of @xmath67 ( the side which is not contained in @xmath261 ) .
suppose not , and let @xmath18 denote a compressing disk which meets @xmath281 minimally .
as @xmath18 lies in @xmath265 , it must be that @xmath283 is a ( possibly empty ) collection of simple closed curves , for each 2-simplex , @xmath262 , of @xmath281 .
let @xmath264 denote an innermost such simple closed curve .
then @xmath264 bounds a subdisk of @xmath262 which is disjoint from @xmath88 ( since @xmath88 meets @xmath262 only in normal arcs ) .
hence , we can use this subdisk to surger @xmath18 , thereby obtaining a compressing disk for @xmath282 which meets @xmath281 in one fewer curve .
our conclusion , then , is that @xmath284 .
but then @xmath18 lies between @xmath67 and @xmath281 .
as these two surfaces are parallel , this is impossible .
now , choose a complete collection of compressing disks for @xmath282 inside @xmath285 , and surger @xmath67 along this collection .
we obtain in this way a collection of spheres , @xmath286 .
@xmath287 bounds a ball , @xmath288 , in @xmath261 , and by definition , @xmath289 is incompressible in the complement of @xmath290 in @xmath291 .
these are the conditions necessary to apply theorem [ t : relsch ] .
the conclusion is that inside each @xmath288 , @xmath88 is a connected surface , which looks like the neighborhood of a graph which is the cone on some collection of points in @xmath292 .
so , in particular , if @xmath293 is a single curve , then it bounds a disk in @xmath288 , and hence so does the corresponding curve in @xmath281 .
= 5 in suppose there is some @xmath2 such that @xmath293 consists of three or more curves , of length 3 or 4 .
the only ways this can happen are shown in figure [ f:2tubes ] . in all cases
we see a compressing disk on one side of @xmath88 which is disjoint from a relative compressing disk on the other side ( see figure [ f : hlcomp ] ) .
= 3.25 in now suppose that for some @xmath2 , @xmath293 consists of two normal curves , of length 3 or 4 .
theorem [ t : relsch ] tells us that the picture must be two normal disks , tubed together by a single unknotted tube , as in figure [ f:1tube ] .
note that in this situation , we see a relative compressing disk on one side , and a compressing disk on the other .
hence , there can not be more than one place where we see this picture .
otherwise , we d see either two disjoint compressing disks on opposite sides , or a compressing disk on one side disjoint from a relative compressing disk on the other .
neither of these situations can happen for a surface which is strongly irreducible with respect to @xmath260 .
= 3.75 in furthermore , suppose @xmath267 contains a curve of length 8 .
then we see relative compressing disks on both sides as in figure [ f : octagon ] , and hence , there can not be a tube anywhere else ( including attached to this disk ! ) .
=3 in we conclude that @xmath88 is made up of all normal disks , with the exception of either a single disk with a boundary curve of length 8 , or a single place where there are two normal disks tubed together by an unknotted tube .
this is the precise definition of an almost normal surface .
our proof is complete by noting that there must be an octagonal disk or a tube _ somewhere _ , because @xmath88 is a thick level , and hence there is at least one compressing or relative compressing disk on both sides .
if there were no tubes or octagons , then we would not have this .
the next theorem is an application of the results of this section .
this result is an important step in the proof of rubinstein and thompson s recognition algorithm for @xmath7 .
[ s3 ] [ rubinstein @xcite , thompson @xcite ] let @xmath249 be a triangulation of some 3-manifold , @xmath0 .
if @xmath236 is a ball in @xmath0 such that @xmath275 is the unique normal 2-sphere in @xmath236 , and such that @xmath236 does not contain any of the vertices of @xmath294 , then @xmath236 contains an almost normal 2-sphere with an octagon as its exceptional piece . inside @xmath236
the 1-skeleton is simply a 1-manifold .
choose an hst splitting , @xmath116 , of @xmath295 which is minimal .
we first claim that @xmath116 is strongly irreducible .
if not , then we can perform one of the four operations described in the proof of lemma [ l : untang ] to produce an hst splitting which is smaller than @xmath116 .
we now call upon theorem [ t : an ] to isotope @xmath116 so that all thick levels are almost normal .
suppose @xmath296 is such a thick level .
let @xmath67 be the sphere component of @xmath296 which contains the exceptional piece , which is either an octagon or a tube . in the latter case @xmath297
is obtained from @xmath296 by compressing the tube , yielding two normal 2-spheres . as the only such normal 2-spheres are normally parallel to @xmath275 , the sequence @xmath298 is also an hst splitting of @xmath140 , contradicting the minimality of @xmath116 .
we conclude then that the exceptional piece of @xmath67 was an octagon .
in this section we tie together the main results of this paper to show how one can prove that various topologically interesting surfaces can be made almost normal .
if @xmath0 is an irreducible submanfold of a triangulated 3-manifold then @xmath299 is a _
maximal normal 2-sphere _ in @xmath0 if there is no other normal 2-sphere which bounds a ball in @xmath0 which contains @xmath67 .
if @xmath0 is not homeomorphic to @xmath142 or @xmath7 , and has a unique maximal normal 2-sphere , @xmath67 , then we recall two basic facts about @xmath67 from @xcite or @xcite .
first , @xmath67 bounds a ball , @xmath236 , in @xmath0 which contains all of the vertices of @xmath300 .
second , there are no almost normal 2-spheres in @xmath0 that are disjoint from @xmath236 .
[ t : main ] suppose @xmath0 ( @xmath6 or @xmath7 ) is an irreducible submanifold of a triangulated 3-manifold which is bounded by a normal or almost normal surface , and contains at most one maximal normal 2-sphere .
if @xmath1 is a strongly irreducible hst splitting of @xmath0 then each thin level is isotopic to a normal surface , and each thick level is isotopic to an almost normal surface .
let @xmath67 denote a maximal normal 2-sphere in @xmath0 .
@xmath67 bounds a ball , @xmath236 which contains all of the vertices of @xmath300 . collapsing @xmath236 to a point
, @xmath301 , then turns @xmath260 into a properly embedded graph , @xmath136 , whose unique vertex is @xmath301 . note that any surface which is disjoint from @xmath301 after the collapse can be identified with a surface which is disjoint from @xmath236 before the collapse .
let @xmath1 be a strongly irreducible hst splitting of @xmath0 .
lemma [ l : untang ] implies that there is a strongly irreducible hst splitting , @xmath116 , of @xmath140 such that @xmath143=\{f_i\}$ ] . by lemma [ l : underlyingodd ] , for each thick ( thin ) level of @xmath1 there is a thick ( thin ) level of @xmath116 whose non-@xmath146 components are parallel to it .
note that as @xmath116 is a hst splitting of @xmath140 , which , by definition , is a hst splitting of @xmath302 , we can also think of @xmath116 as a strongly irreducible hst splitting of @xmath303 . by theorem [ t : an ]
we may isotope @xmath116 so that each thin level is normal and each thick level is almost normal . note that for each almost normal thick level , the exceptional piece is not contained in any 2-sphere component , because there are no almost normal 2-spheres in @xmath304 .
hence , every thick level of @xmath1 is isotopic to an almost normal surface .
[ hs ] [ rubinstein @xcite , stocking @xcite ] in any triangulation of a closed , irreducible 3-manifold other than @xmath7 , any strongly irreducible heegaard surface is isotopic to an almost normal surface .
let @xmath13 be a strongly irreducible heegaard surface for a closed , irreducible 3-manifold , @xmath0 .
then @xmath36 is a strongly irreducible hst splitting of @xmath0 with @xmath13 as a thick level . by the results of @xcite or @xcite
, @xmath0 contains a unique maximal normal 2-sphere .
hence , theorem [ t : main ] implies that @xmath13 can be made almost normal .
[ sch ] [ schleimer @xcite ] suppose @xmath0 is a submanifold of a 3-manifold with triangulation , @xmath249 , such that @xmath0 is bounded by two topologically , but not combinatorially , parallel normal surfaces .
if , in addition , @xmath0 does not contain a normal 2-sphere then @xmath0 contains a boundary parallel almost normal surface . by assumption
, @xmath0 is homeomorphic to @xmath21 , for some subsurface , @xmath13 , of @xmath11 .
the sequence , @xmath38 , is then a strongly irreducible hst splitting of @xmath0 , with @xmath305 as a thick level .
theorem [ t : main ] now implies that @xmath305 can be made almost normal .
let @xmath67 denote a maximal normal 2-sphere in an irreducible 3-manifold , @xmath306 .
as in the proof of theorem [ hs ] , @xmath67 bounds a ball which we collapse to a point , transforming @xmath260 to a 1-vertex graph , @xmath136 .
it is well known that if a 3-manifold contains a non - separating surface then it contains an essential one .
we may also assume that we have such a surface , @xmath13 , that is disjoint from @xmath67 , and that this surface can be normalized in @xmath307 .
let @xmath308 and @xmath309 denote two parallel copies of @xmath13 . finally , let @xmath0 denote the submanifold of @xmath306 cobounded by @xmath308 and @xmath309 which contains @xmath67 .
as @xmath0 is cobounded by surfaces which are essential in @xmath0 there is a strongly irreducible hst splitting , @xmath29 , of @xmath0 such that @xmath310 and @xmath311 .
suppose @xmath160 is a thick level of this splitting .
then @xmath160 separates @xmath308 from @xmath309 in @xmath0 , but is non - separating in @xmath306 .
theorem [ t : main ] may now be applied to @xmath0 to show that @xmath160 can be made almost normal .
suppose @xmath118 is a normal surface in a triangulated 3-manifold , and @xmath13 is any surface disjoint from @xmath118 .
then we say @xmath13 _ normalizes monotonically to _
@xmath118 if there is a @xmath58-compression body , @xmath59 , such that @xmath312 , @xmath313 , and @xmath314 .
the following theorem characterizes when a normal surface compresses in terms of the existence of certain almost normal surfaces . in @xcite
we use this to give a new algorithm to recognize when a given normal surface is compressible .
[ compressible ] suppose @xmath0 is a submanifold of a triangulated irreducible 3-manifold which is bounded by a connected normal surface , and contains at most one maximal normal 2-sphere .
if @xmath11 is compressible in @xmath0 then there is an almost normal surface in @xmath0 which is topologically parallel to @xmath11 , and normalizes monotonically to @xmath11 .
let @xmath315 denote a maximal collection of compressing disks for @xmath11 in @xmath0 .
let @xmath316 denote the surface obtained from @xmath11 by compressing along all disks in @xmath315 .
let @xmath317 denote side of @xmath316 opposite @xmath11 in @xmath0 .
alter @xmath316 by throwing away any 2-sphere component which bounds a ball in @xmath317 , and continue to denote this new surface as @xmath316 .
if @xmath318 , then @xmath0 is a handlebody , and @xmath319 is a strongly irreducible hst splitting of @xmath0 .
otherwise @xmath316 is essential in @xmath317 , and we can find a strongly irreducible hst splitting , @xmath320 , of @xmath317 , where @xmath321 and @xmath322 . in this case , let @xmath323 , so that @xmath324 is a strongly irreducible hst splitting of @xmath0 . in any case , @xmath323 is a thick level .
hence , we may apply theorem [ t : main ] to show that it is isotopic to an almost normal surface in @xmath0 .
[ normalheegaard ] if @xmath12 is a heegaard surface for a 3-manifold which is isotopic to a normal surface in the complement of a maximal normal 2-sphere then @xmath12 is isotopic to two , topologically parallel , almost normal surfaces .
let @xmath325 denote a normal representative of @xmath12 . then
@xmath325 bounds two compression bodies , @xmath9 and @xmath27 . both of these satisfy the hypotheses of theorem [ compressible ] , and hence , both contain almost normal surfaces which are topologically parallel to @xmath325 .
the following result is important because it gives an exact characterization of handlebodies in terms of the existence of normal and almost normal surfaces . in @xcite
we use this to give a new algorithm to recognize when a given almost normal surface is a heegaard surface .
[ t : handlebody ] suppose @xmath9 is a submanifold of a triangulated 3-manifold such that @xmath326 is almost normal , and such that @xmath9 does not contain a normal 2-sphere .
then @xmath9 is a handlebody if and only if there is a sequence of pairwise disjoint surfaces , @xmath327 , such that 1 .
@xmath328 , 2 .
@xmath329 , 3 . for each even @xmath77 , @xmath88 is normal , 4 . for each odd @xmath77 , @xmath88 is almost normal , and normalizes monotonically to @xmath134 and @xmath135 , and 5 .
@xmath330 for all @xmath331 ( with respect to the ordering given in definition [ d : order ] ) .
the proof that the existence of such a sequence implies that @xmath9 is a handlebody is left as an exercise to the reader .
the proof of the converse is similar to that of theorem [ t : main ] .
suppose that @xmath9 is a handlebody that satisfies the hypotheses .
then @xmath332 is a strongly irreducible hst splitting of @xmath9 .
lemma [ l : untang ] implies that that there is a strongly irreducible hst splitting , @xmath184 , of @xmath333 such that @xmath143=\{\emptyset , \partial w\}$ ] .
this immediately implies that @xmath329 , @xmath334 , and that for all @xmath331 , @xmath330 . by theorem [ t :
an ] we may isotope @xmath116 so that each surface of even index is normal , and each surface of odd index is almost normal . furthermore , by definition there is a @xmath260-compression body between consecutive thick and thin level , so each almost normal thick level normalizes monotonically to the next ( and previous ) normal thin level . as @xmath326 is assumed to be almost normal ,
we conclude that @xmath335 , for some @xmath30 . | we define a heegaard - scharlemann - thompson ( hst ) splitting of a 3-manifold @xmath0 to be a sequence of pairwise - disjoint , embedded surfaces , @xmath1 , such that for each odd value of @xmath2 , @xmath3 is a heegaard splitting of the submanifold of @xmath0 cobounded by @xmath4 and @xmath5 .
our main result is the following : suppose @xmath0 ( @xmath6 or @xmath7 ) is an irreducible submanifold of a triangulated 3-manifold , bounded by a normal or almost normal surface , and containing at most one maximal normal 2-sphere .
if @xmath1 is a strongly irreducible hst splitting of @xmath0 then we may isotope it so that for each even value of @xmath2 the surface @xmath3 is normal and for each odd value of @xmath2 the surface @xmath3 is almost normal .
we then show how various theorems of rubinstein , thompson , stocking and schleimer follow from this result .
we also show how our results imply the following : ( 1 ) a manifold that contains a non - separating surface contains an almost normal one , and ( 2 ) if a manifold contains a normal heegaard surface then it contains two almost normal ones that are topologically parallel to it .
keywords : heegaard splitting , normal surface , almost normal surface |
samples of ferroelectric single crystals often posses a surface layers .
its existence greatly influences properties of bulk samples@xcite as well as of thin films.@xcite equilibrium domain structure in the system , mentioned above in the abstract , and the role of the surface layers was first discussed by bjorkstam and oettel@xcite in a special case of shorted electrodes . in our recent paper@xcite we reconsidered this problem in a general case of nonzero voltage between electrodes , discussing the response of the domain structure to external electric field .
our calculations are valid also for thin films and present , in fact , continuation of our discussion of domain structures of thin films@xcite . in@xcite we used our theoretical results for prediction of extrinsic contribution to permittivity of the sample . in the next two sections we give a short review of notation , description of the model and basic results from@xcite . in last two sections we discuss as an example the extrinsic contribution to @xmath0 piezoelectric coefficient of @xmath2 .
we compare our predictions with the recent measurements of record values @xmath3 in temperature range 35k under critical temperature 146k@xcite ( see also@xcite ) .
we consider a plate - like electroded sample of infinite area with major surfaces perpendicular to the ferroelectric axis @xmath4 .
central ferroelectric part with antiparallel domains ( 2 . ) is separated from the electrodes ( 0 . ) , ( 4 . ) by nonferroelectric layers ( 1.),(3 . )
( see fig.[arta2:fig : geometry ] ) .
the spatial distribution of the electric field @xmath5 is determined by the applied potential difference @xmath6 and by the bound charge div@xmath7 on the boundary of ferroelectric material , where @xmath7 stands for spontaneous polarization .
geometrical , electrical and material parameters of the system are shown in fig.[arta2:fig : geometry ] .
we further introduce the symbols @xmath8 and several geometrical parameters : the slab factor @xmath9 the domain pattern factor @xmath10 and the asymmetry factor @xmath11 the ferroelectric material itself is approximated by the equation of state @xmath12 where @xmath13 is the spontaneous polarization along the ferroelectric axis .
this linear approximation limits the validity of our calculations to the temperature region not very close below the transition temperature @xmath14 .
domain walls are assumed to have surface energy density @xmath15 and zero thickness .
rather cumbersome calculations@xcite lead to the following formula for gibbs electric energy per unit area of the system ( in @xmath16 ) , which includes the domain wall energy , the electrostatic energy whose density is @xmath17 and the work performed by external electric sources @xmath18 , where @xmath19 is the charge on positive electrode .
@xmath20 the first term represents domain wall contribution while the last one is the depolarization energy . in the second term
we recognize the effect of layers ( 1 . ) and ( 3 . ) and of the applied voltage . in this model
we neglect the mechanical interactions between components of the system . for given slab factor
@xmath21 and voltage @xmath22 , the equilibrium domain structure , characterized by @xmath23 and @xmath24 , corresponds to local minimum of @xmath25 . in general a minimum
can be found by numerical methods , but for @xmath26 and @xmath27 , the @xmath28 and @xmath24 can be approximated by explicit formulae . for purpose of this paper
we use the following formula for @xmath24 @xmath29^{-1 } \label{arta2:eq : aeqv}\ ] ] where @xmath30 is considered as a small correction and @xmath31 is equilibrium value of @xmath32 for zero voltage @xmath22 . for the extrinsic contribution to permittivity of the sample we get from ( [ arta2:eq : aeqv ] ) ( see also@xcite ) @xmath33 \right\}^{-1 } \label{arta2:eq : eext}\ ] ]
as an example , in this section we work out the approximate prediction of extrinsic contribution to @xmath0 of @xmath34 ( rdp ) , based on our simple model .
rdp is a ferroelastic with spontaneous strain @xmath35 , opposite in opposite polarized domains . in the fig.[arta2:fig : symds]a there is @xmath36 cut through symmetric domain structure ( @xmath37 , @xmath38 ) .
the situation after application of @xmath39 is shown in fig.[arta2:fig : asymds]b ( @xmath40 ) .
a simple geometric consideration leads to the formula for average strain @xmath41 of the sample ( we neglect the mechanical coupling of the central part with the rest of the sample ) @xmath42 for the extrinsic coefficient @xmath0 we get from ( [ arta2:eq : aeqv ] ) and ( [ arta2:eq : avgstrain ] ) @xmath43^{-1 } \label{arta2:eq : d36}\ ] ] neglecting the small @xmath44 and for @xmath45 we get more simple formula in ( [ arta2:eq : d36simple ] ) is not correct , because the assumptions needed for validity of ( [ arta2:eq : aeqv ] ) are violated if @xmath21 is very small . ]
an extremely high @xmath0 under @xmath14 for rdp was first reported in@xcite . in a recent paper@xcite ,
sidorkin deduced the dispersion law of domain wall vibrations , however , in his treatment the existence of a surface layers is not explicitly considered .
we can fit our theoretical results to measured ones@xcite - @xmath47 in the 35k range plateau under @xmath14 . using values that roughly apply to rdp : , see@xcite . ] in ( [ arta2:eq : d36simple ] ) resp .
( [ arta2:eq : eext ] ) we come to an agreement for reasonable value of @xmath48 . naturally for lower temperatures ,
the motion of the walls is limited by freezing " and both @xmath49 and @xmath50 decrease to zero .
it is also interesting , for measurements in@xcite with alternating @xmath51 , that corresponding amplitude of alternating @xmath52 from ( [ arta2:eq : aeqv ] ) is only @xmath53 and displacement of the walls with @xmath54 is of the order @xmath55 . | this work presents a continuation of our last paper , concerning the theory of the response of an antiparallel domain structure in a plate - like electroded sample to external electric field .
the theory is based on the exact formula for free energy of the system , formed of a central ferroelectric part , isolated from electrodes ( with a defined potential difference ) by a surface layers .
our calculations are applicable also to thin films .
it is usual to use the term ` extrinsic ' for the contribution of domain walls displacement to macroscopic properties of a sample . in our last paper
we discussed the extrinsic contribution to permittivity . in this work we concentrate on extrinsic contribution to piezoelectric coefficients in ferroelectrics which are simultaneously ferroelastics .
as an example , we calculate the extrinsic contribution to @xmath0 piezoelectric coefficient of @xmath1 , that was recently measured in a wide range of temperature below curie point . |
kinetic theory of dilute gases has played crucial roles in the history of nonequilibrium statistical mechanics . the theoretical treatment of linear transport problems is well established including the derivation of navier - stokes equation and the relation between boltzmann equation and the linear - response theory through a number of studies of boltzmann equation@xcite .
the transport problems under the nonlinear - nonequilibrium circumstance , however , is still controversial .
in fact , we know burnett s classical work in which he obtained an approximate hydrodynamic equation at the second order nonequilibrium condition@xcite , but we realize that hydrodynamic equation he obtained is unstable for perturbations@xcite .
therefore , one may suggest that the validity of chpaman - enskog method is so limited that we should use grad expansion@xcite to discuss the higher order perturbations such as burnett order and super - burnett order@xcite .
on the other hand , when we obtain the steady solution of burnett equation , the stability of the steady inhomogeneous solution may be different from that of dynamical burnett equation for homogeneous perturbation .
therefore we can not judge whether the calculation at burnett order based on chapman - enskog method is useless even when burnett s hydrodynamic equation is unstable .
unfortunately , these difficulties might be regarded as unimportant , because the solution at burnett order is practically useless except for few cases such as the analysis of shock waves . in these days , however , the solution of boltzmann equation at burnett order has attracted much interest among researchers who try to extend the linear nonequilibrium statistical mechanics to nonlinear nonequilibrium statistical mechanics@xcite .
in fact , since all of attempts are formulated to match them with the linear response theory , their consistency with nonlinear theory like burnett solution should be important .
recently , kim and hayakawa@xcite have extended burnett s classical work on boltzmann equation and obtained the explicit steady solution of boltzmann equation under the steady heat conduction at second ( burnett ) order . the accuracy and numerical stability of this solution
is confirmed by fushiki@xcite from his molecular - dynamics simulation .
this stable behavior does not contradict with blown - up tendency of burnett equation for perturbation of short wavelength or high frequency , because the simulation does not involve any perturbation with short wavelength or high frequency .
their solution is useful@xcite to examine the validity of nonequilibrium theories such as information theory@xcite and the steady - state thermodynamics ( sst)@xcite based on thermodynamic arguments .
they also succeed to apply the solution to calculate the rate of chemical reaction under the nonequilibrium circumstance in which the linear nonequilibrium effects are canceled from the consideration of symmetries@xcite .
the purpose of this paper is to ( i ) summarize the achievement of kim and hayakawa , ( ii ) to discuss the stability of the steady solution at burnett order , and ( iii ) to introduce the result of the molecular - dynamics ( md ) simulation for nonequilibrium knudsen effect to examine sst .
the organization of this paper is as follows . in the next section
, we briefly summarize the outline of the solution obtained by kim and hayakawa@xcite . in section iii
, we will discuss the stability of the solution . in section iv
, we will discuss knudsen effect under a nonequilibrium condition based on md simulation .
in this section , we briefly summarize the solution of steady boltzmann equation under the heat conduction obtained by kim and hayakawa@xcite .
suppose a system of dilute gas in a steady state whose velocity distribution function is @xmath0 .
the steady state boltzmann equation is written as @xmath1 where the collision integral @xmath2 for hard core molecules is expressed as @xmath3 here the velocity distribution @xmath4 and @xmath5 change to @xmath6 and @xmath7 by a binary collision .
the relative velocity of the two molecules before and after the interaction has the same magnitude @xmath8 .
the relative position of the two molecules is represented by the impact parameter @xmath9 and @xmath10 which represents the angle of the plane to which @xmath11 and @xmath12 .
the impact parameter @xmath9 is identical to explicitly determined by specifying the kind of molecules .
the velocity distribution function @xmath4 can be expanded as : @xmath13 where the small parameter of the perturbation is knudsen number , i.e. mean free path of the molecules is much less than the characteristic length of the change of macroscopic variables . @xmath14 in eq.([be3 ] ) is the local maxwellian distribution function written as @xmath15,\label{be3.5}\ ] ] where @xmath16 is mass of the molecules . @xmath17 and @xmath18 denote the density and the temperature at position @xmath19 , respectively . substituting eq.([be3 ] ) into the steady state boltzmann equation ( [ be1 ] ) , we arrive at the following set of equations : @xmath20\phi^{(1)}_{\mathrm{1}}={\bf v}_{\mathrm{1}}\cdot\nabla f^{(0)}_{\mathrm{1 } } , \label{be4}\end{aligned}\ ] ] for @xmath21 and @xmath20\phi^{(2)}_{\mathrm{1}}={\bf v}_{\mathrm{1}}\cdot\nabla f^{(1)}_{\mathrm{1}}-j(f^{(1)}_{\mathrm{1}}f^{(1)}_{\mathrm{2}}),\label{be5}\end{aligned}\ ] ] for @xmath22 , where the linear integral operator @xmath23 $ ] is defined as@xmath20x_{\mathrm{1}}\equiv\int \int \int f^{(0)}_{\mathrm{1}}f_{\mathrm{2}}^{(0)}(x_{\mathrm{1}}^{\prime}-x_{\mathrm{1}}+x^{\prime}_{\mathrm{2}}-x_{\mathrm{2}})g b db d\epsilon d{\bf v}_{\mathrm{2}}. \label{be6}\end{aligned}\ ] ] it is important to consider the solubility conditions of these integral equations , for the steady state boltzmann equation , which are given by @xmath24 for @xmath21 , where @xmath25 is one of the collision invariants : @xmath26 substituting eq.([be3.5 ] ) into the solubility condition ( [ be7 ] ) , it can be shown that @xmath27 is uniform .
we use this result in our calculation . the solubility condition for @xmath22
is given by @xmath28 to make the solutions of the integral equations ( [ be5 ] ) and ( [ be6 ] ) definite , five further conditions must be specified ; we identify the density : @xmath29 the temperature : @xmath30 and the mean flow : @xmath31 here we assume that no mean flow @xmath32 exists in the system .
skipping the derivation of the solution at burnett order , let us write the explicit form of the steady solution for hard core molecules as under the temperature gradient : @xmath33 , \label{be46}\end{aligned}\ ] ] where @xmath34 , @xmath35 and @xmath36 for @xmath37 can be determined as listed in tables.[b0r ] .
note that we have confirmed the convergence of series @xmath38 by changing @xmath39 . here
@xmath40 represents @xmath41component of the heat - flux in the cell on the right hand side in the nonequilibrium steady state which corresponds to the @xmath42 component of the heat flux in eq.([be39 ] ) . in eq.([be46 ] ) , @xmath43 is the gamma function , @xmath44 , and @xmath45 is the sonine polynomial defined by @xmath46 this kind of analysis , of course , is possible to two - dimensional dilute hard - core gases .
kim@xcite has already obtained the result to compare his result with md simulation . in the derivation
the solubility conditions for @xmath22 is important .
thus , eq.([be9 ] ) leads to the condition : @xmath47 where @xmath48 , i.e. the heat flux for @xmath49 can be obtained as @xmath50 with @xmath51 listed in table.[b0r ] .
it must be emphasized that @xmath52 should satisfy @xmath53 from the symmetry consideration of @xmath54 .
thus , the solubility conditions for @xmath55 of the steady state boltzmann equation lead to the fact that the heat flux is constant to the second order . from eqs.([be38 ] ) and ( [ be39 ] ) , we also obtain an important relation between @xmath56 and @xmath57 as @xmath58 from eq.([be39.5 ] ) , the term of @xmath57 can be replaced by the terms of @xmath56 . amongst several applications of the solution , in this paper , we emphasize that the solution can examine the validity of sst proposed by sasa and tasaki@xcite .
they propose a non - trivial nonequilibrium effect in a simple setup as follows ( fig .
[ sst ] ) .
there are two cells whose size along @xmath42 axis is @xmath59 .
both of the cells are filled with dilute gases and connected to each other by a small hole at @xmath60 whose linear dimension along @xmath42 axis and its diameter are less than the mean free path .
the cell on the left hand side is at equilibrium of temperature @xmath61 , while the cell on the right hand side is in a nonequilibrium steady state under a temperature gradient caused by the right wall at temperature @xmath62 at @xmath63 and the thin mid - wall at temperature @xmath61 .
sasa and tasaki predicted the following results@xcite .
when the system shown in fig.[sst ] is in a steady state , the osmosis @xmath64 @xmath65 always positive , where @xmath66 is the pressure in the equilibrium cell and @xmath67 is the @xmath68-component of the pressure tensor @xmath69 in the nonequilibrium cell which is defined by @xmath70 in addition , there is a relation among @xmath71 , @xmath67 , @xmath72 and @xmath73 as @xmath74 where @xmath71 is the density of the cell in the nonequilibrium steady state around the wall and @xmath72 is the density of the cell at equilibrium .
a nonequilibrium steady state is achieved when the mean mass flux at the hole is zero . neglecting the knudsen layer effect , i.e. the slip effect around the wall@xcite
, we can calculate the balance of mass flux as @xmath75 where @xmath76 represents the components of the velocity which are orthogonal to @xmath77 , i.e. @xmath78 and @xmath79 . from eq.([be48 ] ) , we can obtain the relation between the density of the cell in the nonequilibrium steady state around the hole @xmath71 and that of the cell at equilibrium @xmath72 as @xmath80 , \label{be50}\end{aligned}\ ] ] to the second order , where @xmath81 in the approximation until 7th order expansion of @xmath38 .
similarly we can calculate pressure tensor @xmath69 as @xmath82,\label{be51}\end{aligned}\ ] ] with the unit tensor @xmath83 and the number tensor @xmath84 and @xmath85 . from substitution of eqs.([be50 ] ) and ( [ be51 ] ) into eq.([be53 ] ) leads to the relation @xmath86 in sst .
on the other hand , our results conflict with this relation : our results become @xmath87 for hard core molecules .
this is a negative result to sst , though the treatment of the wall in our calculation is too primitive .
we also indicate that our treatment in eq.([be50 ] ) assumes that we can use the local distribution function because of small knudsen number , though , for dilute gases with finite temperature gradient , the place with the mean - free path away from the hole has different temperature from that at the hole because knudsen number should be finite .
in this section , let us discuss the stability of the second order solution in eq.([be46 ] ) . honestly speaking ,
it is difficult to obtain the complete stability analysis of the problem .
therefore , here , we focus on the stability of five hydrodynamic modes which are zero - eigen modes of linearized boltzmann operator and most dangerous ones in the stability
. we should recall that the steady solution with the temperature field @xmath88 is obtained under the assumption of the zero mean ( hydrodynamic ) flow . with the aid of eq.([q_2 ] ) one of the hydrodynamic equations corresponding to the solution is the steady diffusion equation : @xmath89 where @xmath90 is the heat conductivity .
another hydrodynamic equation is nothing but equilibrium condition of hydrostatic pressure : @xmath91 where @xmath92 is the density determined from the solution .
when we keep the constraint of the static solubility conditions ( [ be38 ] ) and ( [ q_2 ] ) , there is no room to appear @xmath52 in the hydrodynamic equations . even when the perturbation violates the static solubility condition , the possible second order
heat current is only @xmath93 in the perturbation , where @xmath94 , and @xmath95 is the shear viscosity .
it should be noted that the above expression is from eq.(15.3.6 ) in chapman - cowling@xcite with neglecting the nonlinear terms of the velocity field .
we also note that the perturbed pressure tensor has the form of navier - stokes order , because the velocity field itself is a perturbed quantity .
thus , the linearized hydrodynamics has a similar form as that of navier - stokes equation except for a term of the heat current in eq.([q_2new ] ) . nevertheless , it is still not easy to discuss the complete stability of hydrodynamic modes around the steady burnett solution .
let us discuss the perturbation of density and temperature @xmath96 where @xmath97 is the temperature after we include the perturbation .
here we keep that the macroscopic velocity does not exist for the sake of simplificity . in this case
it is easy to confirm that the linearized hydrodynamic equations are reduced to @xmath98 and @xmath99 where the heat conductivity is assumed to be @xmath100 with @xmath101 .
this equation apparently shows that the solution of @xmath102 is relaxed to zero for all wavelength , as time goes on .
thus , the steady solution of boltzmann equation at burnett order is stable for the perturbation ( [ hydro - vari ] ) without macroscopic flow .
in this section , we show our preliminary result of two - dimensional molecular - dynamics ( md ) simulation to examine the knudsen effect under the heat conduction .
although we adopt two - dimensional systems , md can check the validity of their implicit assumptions of the calculation by kim and hayakawa@xcite where their calculation neglect the effects of the thermal wall at @xmath60 and use the distribution function at @xmath60 in eq .
( [ int ] ) .
we prepare two situations where @xmath103 and @xmath104 .
number of total particles is 10,000 for both cases . neglecting the non - gaussian correction in velocity distribution function , the mean - free paths @xmath105 for the center regions of four cells are respectively evaluated as @xmath106 and @xmath107 for the case @xmath104 and @xmath108 and @xmath109 for @xmath103 , where @xmath110 and @xmath111 represent the mean - free paths for an equilibrium cell and a nonequilbrium cell , respectively .
figure 2 shows the result of both situations . @xmath112 and
the height @xmath113 .
the linear dimension along x - axis of the hole is 0.40845 @xmath114 and the diameter of the hole is @xmath115 .
this system has the area fraction @xmath116 . the vertical axis in fig.2 plots the pressure divided by the density and temperature of equilibrium cell @xmath117 .
the horizontal axis shows the time of simulation measured by the number of collisions for each particle .
we adopt the diffusive boundary condition for the walls at @xmath60 and @xmath118 , and the periodic boundary condition at @xmath119 . in fig.2 ,
the suffices 1 and 2 in the figure correspond to the cases of @xmath103 and @xmath104 .
our result is contrast to those of both sst@xcite and the calculation by kim and hayakawa@xcite . here
, the osmosis @xmath64 depends on the gradient of temperature in which @xmath64 becomes positive for @xmath103 and @xmath64 becomes negative for @xmath104 .
this simulation suggests that both sst and the calculation by kim and hayakawa can not be used for knudsen effect in physical situations .
there are two possibilities to have linear dependence of the osmosis @xmath64 on the temperature gradient : ( i ) knudsen s layer effect is important on the thermal wall , and ( ii ) the assumption of small knudsen number is violated . as a result of finiteness of knudsen number which is around 1/50 - 1/100 ,
the temperature at the place deviated from the hole by the mean - free path is significantly different from the temperature at the hole .
we believe that the second possibility plays a dominant role , because the gap of the pressure is enhanced when the nonequilibrium cell has higher temperature , _
i.e. _ larger mean - free path @xmath120 than that for low temperature nonequilibrium cell @xmath106 .
the details of md simulation will be reported elsewhere .
in conclusion , we have explained our recent finding of a new explicit solution of boltzmann equation under the steady heat conduction .
we apply the solution to examine the validity of sst and information theory , and find that both of them are not appropriate as they are .
we also discuss the stability of the steady solution we have obtained , and confirmed stability at least for the case of no systematic flow . through md simulation of knudsen effect under the heat conduction , we have obtained the result that @xmath64 depends on the direction of the temperature gradient .
this indicates that our naive application of the steady solution of boltzmann equation to knudsen effect is not accurate .
this also suggests that the basic assumption of sst can not be valid if the influence of physical wall exists . 0.5 cm acknowledgment : the authors wish to thank m. fushiki for fruitful discussion and for his sending us the unpublished result of his simulation .
this work is partially supported by the grant - in - aid of ministry of education , culture , sports , science and technology(mext ) , japan ( grant no .
15540393 ) and the grant - in - aid for the 21st century coe center for diversity and universality in physics from mext , japan .
s. chapman and t. g. cowling , the mathematical theory of non - uniform gases , third edition ( cambridge university press , 1970 ) .
p. rsibois and m. de leener , classical kinetic theory of fluids ( john wiley @xmath121 sons , new york , 1977 ) .
c. cercignani , _ mathematical methods in kinetic theory_(plenum press , new york , 1990 ) .
d. burnett , _ the distribution of molecular velocities and the mean motion in a non - uniform gas _ , proc .
1935 , * 40 * ( nov.15 ) , 382 - 435 .
a. v. bobylev , _ the chapman - enskog and grad method for solving boltzmann equation _ , sov .
phys . dokl .
1982 , * 27 * , 29 . h. struchtrup and m. torrihon , _ regularization of grad s 13 moment equation : derivation and linear analysis _ , phys .
fluids , 2003 , * 15 * ( 9 ) , 2668 - 2680
. h. grad , _ on the kinetic theory of rarefied gas _ , commun .
pure appl . math .
1949 , * 2 * 331 - 407 . j. casas - vrquez and d. jou , _ temperature in non - equilibrium states : a review of open problems and current proporsal _ , rep .
prog . phys .
2003 , * 66 * , 1937 - 2023 .
s. sasa and h. tasaki , _ steady state thermodynamics _ , cond - mat/0411052 , 2004 .
see also their previous version in cond - mat/0108365 .
kim hyeon - deuk and h. hayakawa , _ kinetic theory of a dilute gas system under steady heat conduction
_ j. phys .
soc . jpn .
2003 , * 72 * ( 8) , 1904 - 1916 .
m. fushiki , private communications .
( and his unpublished result ) .
kim hyeon - deuk and h. hayakawa , _ test of information theory on the boltzmann equation _ j. phys
. soc . jpn .
2003 * 72 * , ( 10 ) , 2473 - 2476 .
kim hyeon - deuk and h. hayakawa , _ contributions of steady heat conduction to the rate of chemical reaction _ , chem . phys .
2003 , * 372 * , ( nos.3 - 4 ) 314 - 319 .
kim hyeon - deuk , in preparation . | the significance of the recent finding of the velocity distribution function of the steady - state boltzmann equation under a steady heat current obtained by kim and haykawa ( j. phys .
. jpn . * 72 * , 1904 ( 2003 ) ) is discussed . through the stability analysis
, it seems that the steady solution is stable .
one of possible applications to the nonequiliburium knudsen effect in which one cell at equilibrium is connected to another cell under the steady heat conduction is discussed .
this solution apparently shows that steady - state thermodynamics proposed by sasa and tasaki can not be used in a naive setup .
the preliminary result of our simulation based on molecular - dynamics for nonequilibrium knudsen effect is also presented to verify the theoretical argument . |
the standard model to describe the generation of energy in the nucleus of quasars and agn is the thin accretion disk ( shakura & sunyaev 1973 ; novikov & thorne 1973 ) .
this model not only explains the luminosities of quasars , but also predicts observables such as the disk size and its dependence on wavelength .
in particular , the scaling of the disk size with wavelength , or equivalently , the temperature profile of the disk , is a fundamental test for any disk theory .
well away from the inner edge , a temperature profile @xmath1 translates into a size profile @xmath8 , where the characteristic size can be thought of as the radius where the rest wavelength matches the disk temperature .
the simple shakura & sunyaev ( 1973 ) disk has @xmath9 and hence @xmath4 .
one of the most effective means of measuring accretion disk sizes is gravitational microlensing of lensed quasars ( wambsganss 2006 ) .
there are now many quasar size measurements based on microlensing either of individual objects ( kochanek 2004 ; anguita et al .
2008 ; eigenbrod et al .
2008 ; poindexter et al . 2008 ; morgan et al .
2008 ; bate et al . 2008 ; mosquera et al . 2009
; agol et al . 2009 ; floyd et al . 2009 ; poindexter & kochanek 2010 ; dai et al . 2010 ; mosquera et al . 2011 ; mediavilla et al .
2011a ; muoz et al .
2011 ; blackburne et al .
2011b , hainline et al . 2012 ; motta et al . 2012 ; mosquera et al . 2013 ; blackburne et al .
2013 ) or of samples of lenses ( pooley et al . 2007 ;
morgan et al .
2010 ; blackburne et al .
2011a , jimnez - vicente et al . 2012 ) .
a generic outcome of these studies is that the disks are larger than predicted .
one possible solution is to modify the disk temperature profiles , and this can be tested using microlensing by measuring the dependence of the disk size on wavelength .
there are fewer estimates of the wavelength dependence of the size ( see blackburne et al .
2011a and morgan et al .
2010 and references therein ) . most of the measurements ( anguita et al . 2008 , poindexter et al . 2008 , bate et al . 2008 , eigenbrod et al . 2008 , floyd et al . 2009 , blackburne et al .
2011a , blackburne et al .
2013 ) are based on broad band photometry that can be affected by line contamination .
these studies generally find that the shorter wavelength emission regions are more compact , but the estimates of the scaling exponent @xmath10 have both large individual uncertainties and a broad scatter between studies . only blackburne et al .
( 2011a ) have found little or no dependence of size on wavelength .
single epoch broad band photometry is affected by several potential systematic problems : i ) broad lines , which are known to be less affected by microlensing both observationally ( e.g. guerras et al .
2013a ) and theoretically ( e.g. abajas et al . 2002 ) , ii ) differential extinction of the images by the lens galaxy ( e.g. falco et al . 1999 , muoz et al .
2004 ) , and iii ) uncertainties in the macro model lens magnification .
the effects of extinction and modeling uncertainties can be eliminated by analysing the time variability produced by microlensing , but the effects of line contamination must still be modeled ( e.g. see the dai et al .
2010 analysis of rxj1131@xmath111231 ) . we can remove the effect of broad line contamination by using spectroscopy ( mediavilla et al . 2011a , motta et al .
2012 ) or narrow band photometry ( mosquera et al . 2009 , mosquera et al . 2011 ) to measure the continuum between the lines .
if the continuum and line flux ratios can be separately measured , then the line flux ratios can also be used to eliminate the effects of extinction and model uncertainties on the continuum flux ratios to the extent that the lines are little affected by microlensing .
even with spectra , there are residual issues coming from line contributions to the apparent continuum .
the most important of these are the emission from feii in the wavelength range from 1800 to 3500 and the balmer continuum in the range from 2700 to 3800 ( wills , netzer & wills 1985 ) .
the contribution of these pseudo - continua to the measured continuum flux varies from object to object , and its effect on the estimate of the continuum microlensing magnifications is difficult to assess .
if the uv emitting fe ii comes from a region similar in size to the continuum source as suggested by guerras et al .
( 2013b ) , it would be similarly affected by microlensing . on the other hand , the balmer continuum likely forms on larger scales and would be less affected by microlensing ( see maoz et al .
1993 ) . like including a broad line
, it would dilute the microlensing of the continuum .
since the balmer continuum peaks near the balmer limit at 3646 , it primarily affects the redder portion of the wavelength ranges we consider here and so could slightly bias our results towards steeper slopes ( higher @xmath10 ) .
the scale of the effect is difficult to estimate , but it should be modest .
previous results based on spectroscopic or narrow band studies ( sbs0909 + 532 in mediavilla et al .
2011a , sdssj1004 + 4112 in motta et al .
2012 , he1104@xmath111805 in motta et al .
2012 and in muoz et al .
2011 , he0435@xmath111223 in mosquera et al .
2011 ) have found a slightly shallower wavelength dependence than the predictions of the standard thin disk with @xmath12 rather than @xmath4 .
the main objective of the present work is to extend the studies of the exponent of the size - wavelength scaling , @xmath10 ( @xmath13 ) , using spectroscopic and narrow band data that are less affected by the presence of emission lines or extinction ( 2 ) . in
3 we will calculate an estimate for the average value of the exponent @xmath10 based on this less line contaminated data .
the main conclusions are presented in 4 .
in the present study we use estimates of microlensing magnifications measured either from optical / infrared spectra or narrowband photometry available in the literature .
we collected data for the 8 lens systems in table [ tab1 ] .
spectroscopic data are available for he0512@xmath113329 ( wucknitz et al . , 2003 ) , sbs0909 + 532 ( mediavilla et al .
2011 ) , qso0957 + 561 , sdssj1004 + 4112 and he1104@xmath111805 ( motta et al .
2012 ) . in the latter case , there were two epochs of data showing significant changes in the microlensing that were examined in motta et al .
we also selected objects with narrowband photometry for which ( most of ) the observed filters do not contain contamination from strong emission lines and where there is little extinction or where we can make extinction correction .
this adds two more objects : he0435@xmath111223 ( mosquera et al .
2011 ) and q2237 + 0305 ( muoz et al . 2014 ) . for spectroscopic data ,
the microlensing magnification between two images , 1 and 2 , at a given wavelength @xmath14 is calculated as @xmath15 , where @xmath16 is the flux ratio in an emission line and @xmath17 is the flux ratio of the continuum adjacent to that emission line . with this method ,
the microlensing magnification is isolated from extinction and from the mean lensing magnification , as these affect the line and the adjacent continuum equally ( mediavilla et al .
2009 ) . with narrow band photometry
we can also isolate the microlensing in the continuum from the contamination of strong emission lines . in this case , however , additional information from ir or spectroscopic observations may be needed to safely remove extinction ( see mosquera et al .
2011 , muoz et al . 2011 ) and to provide an unmicrolensed baseline . for each object
( except q2237 + 0305 for which muoz et al . 2014 carried out a detailed analysis of multi - epoch data in five different bands ) , we used the measured microlensing magnifications at three different wavelengths , trying to keep the wavelength baseline as large as possible in order to maximize our sensitivity to chromatic microlensing .
therefore , we used the bluest and reddest ends of the observed ranges , with a third point at an intermediate wavelength , using data from the literature as summarized in table [ tab1 ] . in some cases the microlensing magnification at the three wavelengths
is directly obtained from the original references . in other cases , the values where calculated from the published line and continuum magnitude differences . for he0435@xmath111223 ,
for which we have narrowband photometry , the continuum magnifications were taken from the observations and the h band magnifications were used as an unmicrolensed baseline ( mosquera et al .
similarly , mid - infrared flux ratios from minezaki et al .
( 2009 ) were used as an unmicrolensed baseline for q2237 + 0305 .
there may be some residual differential extinction in q2237 + 0305 , but it is very small compared to the measured microlensing chromaticity and can be ignored ( see muoz et al .
the values of the differential microlensing magnification at each wavelength and their estimated errors are shown in table [ tab1 ] along with the definition of the unmicrolensed baseline and the source of the data .
llrrrc he0435@xmath111223 & b@xmath11a & [email protected] @ 1305 & [email protected] @ 1737 & [email protected] @ 2815 & nir + he0512@xmath113329 & b@xmath11a & @xmath190.12 @ 1026 & @xmath200.11 @ 1474 & @xmath210.06 @ 1909 & lines + sbs0909 + 532 & b@xmath11a & @xmath220.05 @ 1459 & @xmath230.10 @ 4281 & @xmath240.07 @ 6559 & lines + qso0957 + 561 & b@xmath11a & @xmath250.09 @ 1216 & @xmath260.09 @ 1909 & @xmath270.09 @ 2796 & lines + sdssj1004 + 4112 & b@xmath11a & @xmath280.07 @ 1353 & @xmath290.07 @ 2318 & @xmath300.07 @ 4572 & lines + sdssj1029 + 2623 & b@xmath11a & @xmath310.10 @ 1216 & @xmath320.10 @ 1549 & @xmath330.10 @ 1909 & lines + he1104@xmath111805@xmath34 & b@xmath11a & [email protected] @ 1672 & @xmath350.07 @ 2452 & [email protected] @ 4669 & lines + he1104@xmath111805@xmath36 & b@xmath11a & @xmath370.07 @ 1306 & [email protected] @ 1728 & [email protected] @
2796 & lines + q2237 + 0305 & a@xmath11d , b@xmath11d , c@xmath11d & & & & mir + we used a monte carlo statistical analysis to estimate the average size and its wavelength dependence for this sample of objects .
the procedure is very similar to the one described in jimnez - vicente et al .
( 2012 ) , but in this case we apply it to the estimates of the microlensing magnification at three different wavelengths instead of two .
the structure of the accretion disk is described by a gaussian profile @xmath38 , where the characteristic size @xmath39 is a wavelength - dependent parameter ( related to the half - light radius by @xmath40 ) .
the wavelength dependence of the source size is parametrized by a power law such that @xmath41 , and as noted earlier , the standard shakura & sunyaev ( 1973 ) accretion disk model has @xmath4 .
magnification maps for each image are generated using the inverse polygon mapping ( ipm ) algorithm ( mediavilla et al .
2006 , 2011b ) .
the surface densities near the lensed images are generally dark matter dominated , so we put 5% of the surface mass density in the form of stars following the estimates of mediavilla et al .
the one exception was q2237 + 0305 , where 100% of the mass is in form of stars because the lensing is dominated by the bulge of a low redshift spiral galaxy .
we used a fixed microlens mass of @xmath42 .
all linear scales can be rescaled with the square root of the microlens mass ( @xmath43 ) .
the values of the convergence @xmath44 and shear @xmath45 associated with the macro lens models were taken from mediavilla et al .
( 2009 ) . for sbs0909 + 532 ,
the improved values in mediavilla et al .
( 2011b ) were used .
the maps are 2000@xmath462000 pixels in size , with a pixel size of 0.5 lt - days . for the objects in our sample
this size corresponds to between 43 and 80 einstein radii ( with the exception of q2237 + 0305 , for which the map size is 14 einstein radii ) .
we explored a grid in the parameter space of the logarithmic slope @xmath10 and the disk size at a reference wavelength @xmath47 .
the reference wavelength is taken to be the bluest wavelength for which we have a measured microlensing magnification , which is 1026 ( rest frame ) .
the grid runs over @xmath48 for @xmath49 .
we use a natural logarithmic grid in @xmath39 such that @xmath50 for @xmath51 . for every pair of values @xmath52
, the magnification maps are convolved with gaussians of the relevant sizes at the three wavelengths @xmath53 with @xmath54 from which we can compare the microlensing magnification statistics with the observed values at those wavelengths .
the likelihood of observing the three microlensing magnifications @xmath55 for lens @xmath56 at these three wavelengths @xmath57 ( @xmath54 ) given the parameters @xmath10 and @xmath58 is calculated as @xmath59 where @xmath60 is the number of trials with @xmath61 , @xmath62 and @xmath63 for the case with parameters @xmath64 and , @xmath65 , and @xmath66 the estimated errors @xmath67 at the different wavelengths are given in table [ tab1 ] .
the likelihood given by equation [ eq1 ] is calculated for each of the 8 lenses shown in table [ tab1 ] . for each value of the pair
@xmath68 , the likelihood is calculated using @xmath69 trials by sampling each of the two magnification maps at @xmath70 points distributed on a regular grid .
we construct a joint likelihood function for the parameters @xmath10 and @xmath58 by multiplying the individual likelihood functions for the 8 lenses @xmath71 these likelihoods are shown in figures 1 ( for the individual objects ) and 2 and 3 ( for the whole sample ) .
they can be directly used to estimate values for the parameters @xmath10 and @xmath58 . in this context
we have used the maximum likelihood estimator . we can also take a bayesian approach and convert these likelihoods into posterior probabilities by multiplying them by suitable priors .
we have used uniform priors for @xmath58 ( equivalent to a logarithmic prior for @xmath39 ) and the exponent @xmath10 . with this choice for the priors , figures 1 ,
2 and 3 also represent the bayesian posterior probabilities ( normalization constants aside ) . in this approach ,
the mean value of the posterior probabilities and its dispersion are used as estimators of the parameters .
lrrrrlll he0435@xmath111223 & @xmath72 & @xmath73 & @xmath74 & @xmath75 & 0.50 & civ & peng et al . ( 2006 ) + he0512@xmath113329 & @xmath76 & @xmath77 & @xmath78 & @xmath79 & - & - & - + sbs0909 + 532 & @xmath80 & @xmath81 & @xmath82 & @xmath83 & 1.95 & @xmath84 & assef et al . (
2011 ) + qso0957 + 561 & @xmath85 & @xmath86 & @xmath87 & @xmath88 & 0.72 & @xmath84 & assef et al .
( 2011 ) + sdssj1004 + 4112 & @xmath89 & @xmath90 & @xmath91 & @xmath92 & 2.02 & civ & peng et al .
( 2006 ) + sdssj1029 + 2623 & @xmath93 & @xmath94 & @xmath95 & @xmath96 & - & - & - + he1104@xmath111805 & @xmath97 & @xmath98 & @xmath99 & @xmath100 & 0.59 & @xmath84 & assef et al .
( 2011 ) + q2237 + 0305 & @xmath101 & @xmath102 & @xmath103 & @xmath104 & 1.20 & @xmath84 & assef et al .
( 2011 ) + the individual likelihood distributions for each of the 8 objects are shown in figure [ fig1 ] .
for the case of he1104@xmath111805 , where there is data at two different epochs , the distribution is the product of the probability distributions for the two epochs .
q2237 + 0305 combines the results for the three independent pairs at six different epochs ( see muoz et al .
2014 ) . in figure [ fig1 ]
we also show the results from previous studies . in order to make this comparison ,
these sizes from the literature have , as necessary , been converted to a gaussian @xmath39 at our nominal wavelength of 1026 using the reported values of @xmath10 , a microlens mass of @xmath105 , and assuming a fixed half - light radius , because estimates of the half - light radius are largely model independent ( mortonson et al . 2005 ) .
the general agreement is very good , with our results being compatible with previous estimates given the uncertainties .
the maximum likelihood and bayesian estimates for the size of the accretion disk @xmath39 at the reference wavelength ( 1026 ) and for the chromaticity exponent @xmath10 are given in table [ tab2 ] .
figure [ fig2 ] shows the resulting joint likelihood .
previous estimates are also shown , although only for estimates using logarithmic priors on size to avoid overcrowding .
again , the general agreement is very good .
the maximum likelihood estimates are @xmath2 and @xmath106 , where these are 1@xmath107 confidence intervals for 1 parameter .
this corresponds to a size of @xmath108 lt - days for a microlens mass of 1 @xmath109 .
we repeated the calculations using only the objects with spectroscopic data ( i.e. excluding he0435@xmath111223 and q2237 + 0305 ) to check whether including narrowband photometry introduces any bias , but the results are nearly identical .
the most noticeable effect is that , without these objects ( mostly due to the effect of q2237 + 0305 ) , the distribution is slightly more extended towards larger values of @xmath10 and smaller sizes . the upper limit on @xmath10 increases by a small amount , resulting in an estimate of @xmath110 . in order to compare the present result for the size with recent estimates , we convert it to a reference wavelength of 1736 ( using @xmath111 ) and for a mass of the microlenses of 0.3 @xmath109 .
the resulting ( average ) accretion disk size is @xmath112 lt - days , which is in good agreement with the value of @xmath113 lt - days estimated by jimnez - vicente et al .
( 2012 ) from microlensing of ( basically ) this lens sample without chromatic information .
the result for the size is , however , an average value for objects which , in principle , must have a spread in their sizes due to differences in their black hole masses .
in order to minimize this effect when combining the likelihoods of different objects , we repeated the calculations scaling the reference size of the disk as @xmath114 the 2/3 exponent is expected for shakura & sunyaev disks radiating close to the eddington limit , and morgan et al .
( 2010 ) used microlensing to estimate this scaling with mass and found results consistent with the theoretical prediction .
the black hole mass estimates are based on source luminosities and emission line widths .
we use estimates from the literature as reported in table [ tab2 ] , using the @xmath115 estimates where available . generally ,
these black hole mass estimates are viewed as logarithmically reliable with uncertainties of factors of 2 - 3 .
the resulting joint likelihood for the parameters @xmath116 and @xmath10 is shown in figure [ fig3 ] .
the maximum likelihood value is @xmath117 , corresponding to @xmath118 lt - days which is very similar to our previous result .
this was to be expected , as the average black hole mass for this sample is @xmath119 , which is very close to the nominal mass used in the scaling .
if we assume a typical value of @xmath120 for the mass of the microlenses , our estimate for the size of the accretion disk for a nominal black hole mass of @xmath121 is very similar to the estimate by morgan et al .
the estimate of the logarithmic slope @xmath10 did not change , remaining at @xmath2 . on scales of @xmath122
, we would not expect a strong object dependence of the value of @xmath10 for objects sharing the same emission mechanism . in spite of this
, there is no general agreement on the value of @xmath10 from previous studies .
most earlier determinations for individual objects seemed to point near the canonical value of @xmath4 ( e.g. eigenbrod et al .
, 2008 , poindexter et al .
2008 , floyd et al .
, 2009 , mosquera et al . ,
2011 ) , with the notable exception of the sample analyzed by blackburne et al .
( 2011a ) , which produced an unexpectedly low value of @xmath123 that is essentially compatible with a disk size that is independent of wavelength .
nevertheless , several recent estimates based on different techniques ( muoz et al . , 2011 , mediavilla et al . , 2011 ,
motta et al .
, 2012 , blackburne et al .
2013 ) have been finding values close to @xmath124 , although with uncertainties large enough to be consistent with the canonical value of @xmath4 .
figure [ fig4 ] shows the bayesian estimate of @xmath10 found for our sample after marginalizing over @xmath116 .
we find @xmath3 , with the value of the standard thin disk model ( @xmath4 ) being excluded at the 98% confidence level .
the significance drops to 90% when objects with only narrow band data are excluded .
we have used microlensing measurements for 10 image pairs from 8 lensed quasars to study the structure of their accretion disks . by using spectroscopic and narrow band photometric data we can eliminate the effects of differential extinction , which can be mistaken with chromatic microlensing and minimize any contamination from the weakly microlensed emission lines that can reduce the microlensing amplitude and potentially alter its dependence on wavelength .
we find a best estimate of the wavelength dependence of the accretion disk size of @xmath2 for a maximum likelihood analysis , and @xmath125 for a bayesian analysis ( @xmath0 ) .
this estimate is significantly smaller than the @xmath4 value predicted by the shakura & sunyaev ( 1973 ) thin disk model with a temperature profile of @xmath126 .
thus , our result seems to indicate a steeper temperature profile for the accretion disk .
agol & krolik ( 2000 ) proposed a model in which the magnetic connection between the black hole and the disk is able to spin the disk up leading to a steeper temperature profile with @xmath127 , which is more consistent with our results .
another alternative is that there is some other structure on top of the accretion disk . along this line
, abolmasov & shakura ( 2012 ) , proposed that super eddington accretion in the disks of quasars would generate an optically thick spherical envelope that could explain the lower values of the exponent @xmath10 found by blackburne et al .
( 2011a ) , although they need an ad - hoc mechanism to explain the small sizes of the x - ray - emitting regions of microlensed quasars .
note , however , that dai et al .
( 2010 ) find that the scattered light fraction has to be very high ( @xmath128 ) before microlensing size estimates are significantly shifted by scattering .
abolmasov & shakura ( 2012 ) proposed that this mechanism would only work for the quasars with lower black hole masses , although we do not find any trend of the parameter @xmath10 with the mass of the black hole . in this line ,
fukue & iino ( 2010 ) found that a fully thermalized relativistic spherical wind around the black hole would show a profile of the type @xmath129 which is very similar to the produced by our analysis .
note that such steeper temperature profiles make it more difficult to reconcile the larger disk sizes found by microlensing .
solutions to this problem would favor shallower profiles closer to a strongly irradiated disk with @xmath130 ( see the discussion in morgan et al . , 2010 ) .
our estimate for the size of the accretion disk of @xmath131 lt - days ( at 1026 in the rest frame ) is in agreement with other recent determinations and considerably larger than the expected value from a thin disk radiating as a black body .
the large slope of the temperature profile of quasar accretion disks estimated in this work makes the already serious problem posed by the large sizes determined by microlensing for these objects even worse . as scattering
does not seem to be the answer to this discrepancy , and other suitable theoretical models are lacking , evidence seems to indicate that some key ingredient in the emission mechanism of accretion disks in these objects may be missing in our present understanding . on the observational side , optical / infrared spectroscopy ( or even better spectroscopic monitoring ) for a larger sample of lensed quasars would be desirable to unambiguously confirm the large sizes and slopes of the temperature profile of quasar accretion disks .
in the meantime , on the theoretical front , a thorough revision of the standard model of accretion disks in quasars may also be necessary .
we gratefully acknowledge the anonymous referee for useful suggestions that improved the presentation of this work .
this research was supported by the spanish ministerio de educacin y ciencia with the grants aya2011 - 24728 , aya2007 - 67342-c03 - 01/03 , aya2010 - 21741-c03/02 .
is also supported by the junta de andaluca through the fqm-108 project .
is also supported by the generalitat valenciana with the grant prometeo/2009/64 .
is supported by nsf grant ast-1009756 .
v.m . gratefully acknowledge support from fondecyt 1120741 .
abajas , c. , mediavilla , e. , muoz , j. a. , popovi , l. c. & oscoz , a. 2002 , , 576 , 640 abolmasov , p. & shakura , n. i. 2012 , , 427 , 1867 agol , e. & krolik j. , 1999 , , 524 , 49 agol , e. , gogarten , s. m. , gorjian , v. , & kimball , a. 2009 , , 697 , 1010 anguita , t. , schmidt , r. w. , turner , e. l. et al .
2008 , , 480 , 327 assef , r. j. , denney , k. d. , kochanek , c. s. et al .
2011 , , 742:93 bate , n. f. , floyd , d. j. e. , webster , r. l. & wyithe , j. s. b. 2008 , , 391 , 1955 blackburne , j. a. , pooley , d , rappaport , s. & schechter , p. l. 2011a , , 729 , 34 blackburne , j. a. , kochanek , c. s. , chen , b. , dai , x. , chartas , g. 2011b , arxiv:1112.0027v1 blackburne , j. a. , kochanek , c. s. , chen , b. , dai , x. , chartas , g. 2013 , arxiv:1304.1620 dai , x. , kochanek , c. s. , chartas , g. et al .
2010 , , 709 , 278 eigenbrod , a. , courbin , f. , meylan , g. et al .
2008 , , 490 , 933 falco , e. e. , impey , c. d. , kochanek , c. s. et al .
1999 , , 523 , 617 floyd , d. j. e. , bate , n. f. , & webster , r. l. 2009 , , 398 , 233 fukue , j. & iino , e. 2010 , pasj , 62 , 1399 guerras , e. , mediavilla , e. , & jimnez - vicente , j. et al . 2013a , , 764 , 160 guerras , e. , mediavilla , e. , & jimnez - vicente , j. et al .
2013b , , 778 , 123 hainline l. j. , morgan c. w. , beach j. n. et al .
2012 , , 744 , 104 jimnez - vicente , j. , mediavilla , e. , muoz , j. a. , & kochanek , c. s. 2012 , , 751 , 106 kochanek , c. s. 2004 , , 605 , 58 maoz , d. et al .
1993 , , 404 , 576 mediavilla , e. , muoz , j. a. , lopez , p. , et al . 2006 , , 653 , 942 mediavilla , e. , muoz , j. a. , falco , e. , et al .
2009 , , 706 , 1451 mediavilla , e. , muoz , j. a. , kochanek , c. s. , et al .
2011a , , 730 , 16 mediavilla , e. , mediavilla , t. , muoz , j. a. et al .
2011b , , 741 , 42 minezaki , t. , chiba , m. , kashikawa , n. , inoue , k. t. , & kataza , h. 2009 , , 697 , 610 morgan , c. w. , kochanek , c. s. , dai , x. , morgan , n. d. , & falco , e. e. 2008 , , 689 , 755 morgan , c. w. , kochanek , c. w. , morgan n. d. & falco e. e. 2010 , , 712 , 1129 mortonson , m. j. , schechter , p. l. & wambsganss , j. 2005 , , 628 , 594 mosquera , a. m. , muoz , j. a. , & mediavilla , e. 2009 , , 691 , 1292 mosquera , a. m. , muoz , j. a. , mediavilla , e. , kochanek , c. s. 2011 , , 728 , 145 mosquera , a. m. , kochanek , c. s. , chen , b. et al .
2013 , , 769 , 53 motta , v. , mediavilla , e. , falco , e. , & muoz , j. a. 2012 , , 755 , 82 muoz , j. a. , falco , e. e. , kochanek , c. s. , mcleod , b. a. , & mediavilla , e. 2004 , , 605 , 614 muoz , j. a. , mediavilla , e. , kochanek , c. s. , falco , e. e. & mosquera , a. m. 2011 , , 742 , 67 muoz , j. a. , et al .
2014 , in preparation novikov , i. d. & thorne , k. s. 1973 , in black holes , ed . c. de witt & b. de witt ( new york : gordon & breach ) , 343 peng , c. y. , impey , c. d. , rix , h. w. et al .
2006 , , 649 , 616 poindexter , s. , morgan , n. & kochanek , c. s. 2008 , , 673 , 34 poindexter , s. & kochanek , c. s. 2010 , , 712 , 668 pooley , d. , blackburne , j. a. , rappaport , s. , & schechter , p. l. 2007 , , 661 , 19 shakura , n. i. & sunyaev , r. a. 1973 , , 24 , 337 wambsganss , j. 2006 , gravitational lensing : strong , weak , and micro , ed .
g. meylan , p. north , & p. jetzer ( berlin : springer ) , 453 wills , b. j. ; netzer , h. ; wills , d. 1985 , , 288 , 94 wucknitz , o. , wisotzki , l. , lopez , s. , & gregg , m. d. 2003 , , 405 , 445 | we use multi - wavelength microlensing measurements of a sample of 10 image pairs from 8 lensed quasars to study the structure of their accretion disks . by using spectroscopy or narrow band photometry
we have been able to remove contamination from the weakly microlensed broad emission lines , extinction and any uncertainties in the large - scale macro magnification of the lens model .
we determine a maximum likelihood estimate for the exponent of the size versus wavelength scaling ( @xmath0 corresponding to a disk temperature profile of @xmath1 ) of @xmath2 , and a bayesian estimate of @xmath3 , which are significantly smaller than the prediction of thin disk theory ( @xmath4 ) .
we have also obtained a maximum likelihood estimate for the average quasar accretion disk size of @xmath5 lt - day at a rest frame wavelength of @xmath6 for microlenses with a mean mass of @xmath7 , in agreement with previous results , and larger than expected from thin disk theory . |
the matrix models generate the ensemble of discretized two dimensional surfaces as a feynman diagrammatic expansion , thus providing a discrete formulation of the liouville gravity @xcite ( see also the reviews @xcite ) . in the case of the one ( hermitian ) matrix model ( omm ) ,
the matrix potential can be tuned in the continuum limit in order to achieve the @xmath0 minimal liouville gravity ( mlg@xmath0 ) , which consists of matter , liouville and ghosts fields .
deformations of the critical potential can also be introduced , which produce the kdv renormalization group flows between the critical points . in the liouville gravity context
, this picture corresponds to perturbing the action with vertex operators .
as suggested in @xcite , the exact relation between the matrix model ( kdv ) and the liouville gravity couplings involves a non - trivial resonance transformation due to the contact terms .
this transformation was obtained at the first order in @xcite , and the identification of the bulk one and two point correlation functions was performed . in @xcite , a. belavin and
a. zamolodchikov obtained the resonance transformation reproducing the three - point correlation function of the mlg@xmath0 from the matrix model .
their result leads to a conjecture for the explicit form of the resonance transformation to all orders @xcite .
this conjecture was checked up to the fifth order @xcite .
it was also shown in @xcite that this transformation also works for the bulk one - point correlation function on the disc . to complete our understanding of the discrete formulation of the liouville gravity
, we have to study the realization of boundary conditions ( bcs ) and boundary operators .
such approach has already been taken in various matrix models , including the rsos and @xmath3 models @xcite and the two matrix model @xcite .
recently , two of the authors constructed new boundaries for the omm using additional vector fields , providing the general mlg@xmath0 bcs in the continuum limit @xcite .
such bcs , referred as fzzt brane @xcite , depend on two parameters and will be labeled @xmath4 .
the parameter @xmath5 is related to the boundary cosmological constant @xmath6 which determines the bc for the liouville field , @xmath7 the matter bc is given by the cardy state @xmath8 where @xmath9 runs from one to @xmath10 @xcite and we use the kac notation .
the matrix model construction of @xcite was interpreted in @xcite as the realization of a linear relation among fzzt branes .
these results were also generalized to the two matrix model which provides the general mlg@xmath11 . between the two bc @xmath12 and @xmath13 , we have to introduce one of the following boundary operators of dimension one , @xmath14^{(t;m)},\quad k=|\ell - m|+1,\cdots,\ell+m-1;2 \label{bcc operator}\end{aligned}\ ] ] where @xmath15 denotes the liouville field and @xmath16 the @xmath17 primary matter field and @xmath18 has increment of two .
the liouville charge @xmath19 is related to the dimension of @xmath20 through the kpz relations @xcite and satisfies the seiberg s bound ( @xmath21 ) , @xmath22 the diffeomorphism invariance of the correlation functions imposes to integrate the boundary operators [ bcc operator ] over the boundary .
the three conformal killing vectors on the disc allow to fix the position of three boundary operators with the appropriate ghost dressing .
since there are no remaining coordinate dependence , these correlation functions are referred as ` correlation numbers ' .
so far from the matrix model , only correlators involving the @xmath23 operators were explicitly known .
the purpose of this article is to construct the matrix model operators providing in the continuum limit the general boundary operators [ bcc operator ] .
this will be achieved by the introduction of powers of the matrix inside the correlators .
this paper is organized as follows . in the section [ one matrix model with vectors ] ,
we briefly review the construction of the omm boundaries using the vector description .
then , we discuss in section [ boundary perturbation ] the perturbation of the boundary term , and the general form of the matrix boundary operators . these operators will be explicitly determined in the section [ boundary two - point correlation numbers ] using the requirement of vanishing boundary two - point functions for two different operators in the mlg@xmath0 . as a consistency check
, we recover the expression of the two - point functions for two identical operators .
finally , the section [ summary ] is devoted to summary and discussions .
we consider the omm coupled to a vector model , with partition function @xmath24 where @xmath25 is the @xmath26 hermitian matrix , and @xmath27 , @xmath28 are some polynomials of @xmath25 .
the vectors @xmath29 ( and their hermitian conjugates @xmath30 ) belong to the fundamental ( conjugate ) representation of @xmath31 , and `` @xmath32 '' represents the contraction of the @xmath33 dimensional indices .
the flavors indices @xmath34 and @xmath35 label the boundary condition , @xmath36 .
the coupling @xmath37 is related to the bulk cosmological constant @xmath38 which controls the area fluctuation of the worldsheet .
the vectors allow to emphasize the role played by the boundary , but can be integrated out , and the partition function [ starting point ] is reduced to the following expansion , @xmath39 here , @xmath40 represents the connected part of the expectation value with respect to the matrix integration with potential @xmath27 and @xmath41 stands for the trace over both flavor and matrix indices .
the correlators @xmath42 describe a worldsheet with @xmath43 disconnected boundaries , or holes .
they can be expanded in powers of @xmath33 , @xmath44 ; each @xmath45 describes a worldsheet with @xmath37 handles and @xmath43 holes , and @xmath46 is the corresponding euler characteristic . in this paper , we focus on the disk partition function ( @xmath47 ) in the planar limit ( @xmath48 ) , @xmath49 in the feynman diagrammatic expansion of the integral ( [ starting point ] ) , boundaries are formed by loops of the vectors , each flavor being associated to a different boundary @xcite .
the diagonal elements of the matrix @xmath28 create the corresponding boundaries , whereas the off - diagonal elements produce a flavor mixing leading to a boundary changing effect .
when the matrix @xmath28 is purely diagonal with respect to the flavor indices , the correlator [ disk partition function ] reduces to a sum over the disk correlators with different boundaries @xmath34 , @xmath50 , where @xmath51 denotes the trace over the matrix indices . the simplest , and most studied , case is to consider only a single boundary and define @xmath52 .
its derivative with respect to the boundary cosmological constant @xmath53 is the well - known resolvent , @xmath54 in order to achieve the @xmath10th critical point corresponding to the mlg@xmath0 , the potential @xmath27 is fine - tuned in the continuum limit .
the bulk cosmological constant gets renormalized , and we define @xmath55 where @xmath56 is the critical value for which the area blows up and the cut - off @xmath57 is the lattice spacing . in this limit ,
the worldsheets with large boundaries dominate , and @xmath53 is sent to its critical value @xmath58 .
this value will be taken to be zero by a suitable shift of the matrix @xmath25 , so that we define the renormalized boundary cosmological constant as @xmath59 , which is identified with the parameter @xmath6 of the mlg@xmath0 . in this limit
, the singular part of the resolvent @xmath60 is rescaled as @xmath61 , and has a branch over the support of the eigenvalue density @xmath62-\infty ,- u_0]$ ] .
this branch cut is resolved using the parameterization [ mub ] , @xmath63 the resolvent @xmath64 is identified with the disc boundary one point function of the dressed identity operator , with bc @xmath65 of the mlg@xmath0 .
the new matrix boundaries introduced in @xcite are defined as @xmath66 with @xmath67 where we denoted @xmath68 and @xmath69 in the continuum limit , these couplings get renormalized as @xmath59 , @xmath70 , and the critical resolvant @xmath71 takes the same value over the variables @xmath72 , @xmath73 . in this limit
, the matrix operator @xmath74 creates a boundary with @xmath4 bc , i.e. with a matter bc given by the cardy state @xmath75 and a renormalized boundary cosmological constant @xmath76 .
this interpretation follows from the study of the boundary one and two point correlators @xcite , and more generally from a linear relation among fzzt branes @xcite .
in this section , we study the perturbation away from the critical boundary constructed by @xmath77 .
the bulk perturbations are introduced as a modification of the potential @xmath27 , @xmath78 where the @xmath79 are the @xmath80th critical potentials leading to the mlg@xmath81 @xcite .
the couplings @xmath82 generate the kdv flows between the critical points of the matrix model ; they are related to the perturbation of the mlg@xmath0 with bulk vertex operators by the resonance transformation . in a similar way
, we would like to write the boundary term in [ starting point ] as a perturbation of a critical background @xmath83 , @xmath84 let us first consider the case of a single boundary @xmath85 ; the perturbed disc partition function corresponds to @xmath86 where the @xmath87 are the perturbative couplings related to the operators @xmath88 and @xmath89 .
this correlator can be expanded in the perturbation series , @xmath90 and we recover in the planar limit a disc correlator with insertion of zero , one , two , etc boundary operators .
these operators @xmath88 should satisfy two requirements : they have to be local and scaling operators .
the first requirement imposes that @xmath88 is a polynomial in @xmath25 .
indeed , the insertion of powers of the matrix corresponds to the insertion of a finite edge on the boundary , which is renormalized toward a point in the continuum limit . to understand the implication of the second requirement ,
let us recall that in the continuum limit , the critical matrix correlators obey a scaling property which allows to define their scaling dimension .
be a matrix correlator depending on the couplings @xmath91 .
this correlator can be decomposed into a regular @xmath92 and critical @xmath93 part . in the continuum limit , the regular part , which is polynomial in at least one of the boundary cosmological constant ,
is non - universal and shall be discarded .
if the couplings take the critical values @xmath94 , we define the renormalized couplings as @xmath95 and correlator @xmath96 where @xmath97 and @xmath98 are the corresponding scaling dimensions .
the scaling property writes @xmath99 . ]
a matrix operator has scaling dimension @xmath80 if its introduction within matrix correlators modifies their scaling dimension from @xmath97 to @xmath100 .
because of the shift of the matrix that imposed @xmath101 , the matrix powers @xmath102 are scaling operators of dimension @xmath80 ( see appendix [ scaling property ] ) . in the continuum limit , there exist only two scale parameters , given by the bulk and boundary cosmological constants .
thus , imposing that @xmath88 is a monic polynomial in @xmath25 with scaling dimension @xmath80 , it is restricted to be of the form @xmath103 } h_{nm}x^{n-2m}(\kappa
-\kappa^*)^{m}m^{j - n } , \label{pol}\end{aligned}\ ] ] where @xmath104 are @xmath105-numbers and we discarded terms that do not contribute to the continuum limit . a natural choice is to take @xmath106 since this generates a boundary flow similar to the kdv flow introduced by the perturbation [ bulk perturbation ] .
when one of the coupling @xmath87 goes from zero toward infinity , the continuum boundary flows from @xmath12 to @xmath107 .
we now consider the perturbation of the off - diagonal elements of @xmath108 between the two boundaries @xmath36 and @xmath109 . to simplify the notations
, we denote the polynomials by @xmath110 ; they should be local , monic and of degree @xmath80 . inserted between two different boundaries , the boundary operators can now depend on three scaling parameters in the continuum limit , given by the bulk and the two boundary cosmological constants .
however , in this case , there is no clear picture of boundary flows available at this stage . in the next section
we will follow a practical approach and determine the boundary operators directly in the mlg@xmath0 frame , i.e. after the resonance transformation . in this frame , the critical part of the matrix model correlators for a disc with two boundaries should satisfy the condition @xmath111 where @xmath112 is proportional to the liouville boundary two - point function in the continuum limit .
we will obtain the polynomials @xmath113 using the condition [ orthogonality ] only when @xmath114 . as a consistency check ,
we compute the correlators for @xmath115 , and recover the liouville boundary two - point function , the analysis of which shows that @xmath116 couple to the @xmath117 boundary operator .
we now turn to the determination of the polynomials @xmath113 . in the rest of this paper , we systematically drop out all the non - universal part of the matrix correlators since they are irrelevant in the continuum limit . in order to study the insertion of the matrix powers on the boundary created by @xmath77 and @xmath118
, we introduce the ratios @xmath119 , @xmath120 the correlators @xmath121 were studied in @xcite ; they have scaling dimension @xmath122 and can be expressed in the continuum limit in terms of the critical resolvent @xmath123!}{[\ell -1]![m-1 ] ! } \frac{\omega(\xi)-\omega(\z_{l+m } ) } { \prod_{j=-(\ell + m-2),2}^{\ell + m-2}(\xi_j-\z ) } , \label{olm}\end{aligned}\ ] ] where @xmath124 , @xmath59 , @xmath125 and we used the notations @xmath126!\equiv \prod_{k=1}^{n}[k],\quad [ k]\equiv\dfrac{q^k - q^{-k}}{q - q^{-1}},\quad q\equiv e^{i\pi b^2}.\end{aligned}\ ] ] the expression [ olm ] is known to reproduce in the continuum limit the liouville part of the @xmath127 boundary two - point functions .
the ratios @xmath119 can be computed recursively using the relation @xmath128 from its definition [ def_gjlm ] , @xmath119 is a scaling quantity of dimension @xmath80 and we can define the renormalized ratio @xmath129 in the continuum limit . substituting the expression ( [ olm ] ) for @xmath130 , the previous recursion relation can be written as @xmath131}{[\ell + m-2 ] } ( \xi_{\pm ( \ell + m-2)}-\z)g_{j-1}^{(\ell -1m)}(\xi_{\mp 1},\z ) .
\label{recursion}\end{aligned}\ ] ] we note that @xmath132 , and @xmath133 depends on the renormalized cosmological constants of the boundary @xmath134 and @xmath135 , and of the bulk through @xmath136 ; it is a polynomial of @xmath137 with degree @xmath80 .
are obviously polynomials in @xmath135 of degree @xmath80 for all @xmath138 .
the dependence on @xmath134 follows from the symmetry @xmath139 . ]
the first two ratios are explicitly given by @xmath140\xi+[\ell -1]\zeta } { [ \ell + m-2 ] } , \nonumber\\ g_2^{(\ell m)}(\xi , \zeta)&=\frac{[\ell -1][\ell -2]\xi^2 + [ m-1][m-2]\zeta^2}{[\ell + m-2][\ell + m-3 ] } , \nonumber\\ & \ ; + \frac{[2][\ell -1][m-1]\xi \zeta } { [ \ell + m-2][\ell + m-3 ] } + u_0 ^ 2\sin^2 \pi b^2\frac{[\ell -1][m-1]}{[\ell + m-3 ] } .\end{aligned}\ ] ] now that the correlators [ def_gjlm ] are determined , we investigate the consequence of the orthogonality condition [ orthogonality ] for @xmath114 .
we first write the polynomials @xmath141 as a sum of monomials , @xmath142 with @xmath143 .
the orthogonality condition can be recursively linearized , leading to @xmath144 where we used the symmetry @xmath145 , and introduced the ratios @xmath146 , @xmath147 the system of equations ( [ orthogonality 2 ] ) being linear , there is a unique solution for @xmath148 and @xmath149 in terms of @xmath150 , which is given by @xmath151 where @xmath152 is the following determinant , @xmath153 and @xmath154 its first minor with respect to the last row and the @xmath155th column .
the sum over the monomials [ expansion ] can be performed , leading to a compact expression for the polynomials @xmath156 it can be shown that the coefficients @xmath148 and @xmath149 are scaling variables of dimension @xmath157 and @xmath158 , respectively . in the continuum limit ,
@xmath159 and @xmath160 are polynomials in @xmath134 and @xmath135 of degree @xmath157 and @xmath158 .
is a polynomial in @xmath135 of degree @xmath161 , using the recursion relation [ recursion ] , we can show that it has zeros for @xmath162 , @xmath163 with multiplicity @xmath164 .
the same procedure also applies for the numerators of [ ajk ] which are polynomial of degree @xmath165 , leading to the same zeros . ]
the scaling operators @xmath166 get renormalized into @xmath167 , where @xmath168 , @xmath169 and @xmath170 , and the first polynomials are given by @xmath171\xi+[\ell -1]\zeta}{[\ell + m-2 ] } , \nonumber\\ p^{(\ell m)}_2(\xi,\zeta,\eta)&=\eta^2 -\frac{[2]([m-2]\xi+[\ell -2]\zeta)}{[\ell + m-4]}\eta -u_0 ^ 2 \sin^2 \pi b^2 \frac{[\ell -1][m-1]}{[\ell+m-3 ] } \nonumber\\ & + \frac{[m-1][m-2]\xi^2+[\ell -1][\ell -2]\zeta^2+[2][\ell -2][m-2]\xi\zeta } { [ \ell + m-3][\ell+m-4 ] } \ ] ] we now show that the solution we have found for @xmath172 with the polynomials @xmath173 previously determined indeed reproduce the liouville part of the boundary two - point function .
unfortunately , the expression given for @xmath149 is not very handy , so that we will take a different approach and work directly in the continuum limit where @xmath160 is polynomial .
we first notice that applying the recursion relation ( [ recursion ] ) to ( [ orthogonality 2 ] ) we obtain the identity @xmath174 where @xmath175 is a non - negative integer less than @xmath80 .
similarly , if we apply ( [ recursion ] ) to ( [ orthogonality 2 ] ) with @xmath176 and eliminate the term proportional to @xmath177 using the previous relation with @xmath178 , we get @xmath179}{[\ell + m-2 ] } ( \xi_{\pm ( \ell + m-2)}-\z ) \sum_{k=0}^j g_{k+j-1}^{(\ell -1m)}(\xi_{\mp 1},\z ) a_{jk}^{(\ell m)}(\xi,\z).\end{aligned}\ ] ] repeating this process we find that @xmath149 has zeros at @xmath180 for @xmath181 . since its degree is @xmath158 , its functional form in @xmath135 is completely determined as @xmath182 the remaining coefficient is independent of @xmath134 and @xmath135 ; it is obtained in the appendix [ ap asymptotic form ] from the asymptotic form of the recursion relation [ recursion ] when @xmath135 goes to infinity .
this determines @xmath149 , and we finally find the the expression of @xmath183 , @xmath184![\ell + m-2j-1]!}{[\ell + m - j-1 ] ! } o^0_{\ell -jm - j}(\xi,\z ) .
\label{final answer}\end{aligned}\ ] ] by substituting the expressions of [ correct solution ] , and of the resolvent ( [ resolvent ] ) in the continuum limit , we verify that this quantity is proportional to the mlg@xmath0 boundary two - point function @xmath185 ( see appendix a of the first reference in @xcite ) .. ]
in this paper , we derived the matrix realization of arbitrary boundary operators using a polynomial insertion inside the correlators .
these polynomials were determined directly in the cft frame by imposing the orthogonality condition for the two - point function [ orthogonality ] .
the insertion of the polynomial @xmath186 of degree @xmath80 given in [ expression for p ] leads to consider the boundary operator @xmath187 instead of @xmath188 inserted between @xmath12 and @xmath189 boundaries . as a consistency check
, we recovered the expression for the liouville boundary two - point function coupled to a minimal model .
note that the case of operators inserted between two identical boundaries is obtained in the ( non - singular ) limit of equal boundary labels , @xmath190 .
the recursion relation [ rec ] is one of the main identities we employed to derive the expression of the matrix operators . in the continuum limit
, this relation can be interpreted as the insertion of two boundary ground ring operators @xmath191 , as shown in appendix c. it would be interesting to study the realization of the ring structure with the one matrix model in more details .
finally , we emphasize that the expression we found for the boundary operators was obtained for the disc with two different boundaries .
it should be noted that because of the presence of resonant terms , we may have to introduce corrections to this expression when the operators are inserted into correlators describing a higher genus or a higher number of boundaries ( possibly disconnected ) .
this is the main open question we hope to address in a near future .
we are indebted to i. kostov for valuable discussions . this work is partially supported by the national research foundation of korea ( knrf ) grant funded by the korea government ( mest ) 2005 - 0049409 ( b , i & r ) and r01 - 2008 - 000 - 21026 - 0 ( r ) .
[ scaling property ] in this appendix , we show that the matrix powers @xmath102 are scaling operators of dimension @xmath80 .
let us first consider the correlators @xmath192 one can be shown that these correlators have the scaling dimension @xmath193 using recursively the identity @xmath194 and the dimension @xmath195 for the resolvant at the @xmath10-th critical point .
then , we consider an insertion of @xmath25 in the trace in ( [ n correlator ] ) , and use the relation @xmath196 to decompose this correlator as @xmath197 on the right - hand side , both terms have the dimension @xmath198 , so that the operator @xmath25 has scaling dimension one .
this conclusion can be easily generalized to the insertion of the higher power of @xmath25 , the operator @xmath102 leading to correlators of dimensions @xmath199 , i.e. it has dimension @xmath80 .
[ ap asymptotic form ] in this appendix , we calculate @xmath160 in the large @xmath135 limit . in this limit ,
the ratios @xmath201 are independent of @xmath134 and the relation ( [ recursion ] ) can be easily solved , @xmath202}{[\ell+m-1-k]}. \label{g in large y}\end{aligned}\ ] ] by substituting ( [ g in large y ] ) , @xmath152 defined in ( [ d = a / a ] ) is written as @xmath203}{[\ell + m-1-k]}\right)^{j+1-k}b_j^{(\ell m)},\end{aligned}\ ] ] where @xmath204 is a determinant of a @xmath205 matrix given by , @xmath206}{[\ell + m-2 ] } & \prod_{k=1}^2\frac{[\ell -k]}{[\ell + m-1-k ] } & \cdots & \prod_{k=1}^j\frac{[\ell -k]}{[\ell + m-1-k ] } \\ 1 & \frac{[\ell -2]}{[\ell + m-3 ] } & \prod_{k=1}^2\frac{[\ell -1-k]}{[\ell + m-2-k ] } & \cdots & \prod_{k=1}^j\frac{[\ell -1-k]}{[\ell + m-2-k ] } \\
\vdots & \vdots & \vdots & & \vdots \\ 1 & \frac{[\ell -1-j]}{[\ell + m-2-j ] } & \prod_{k=1}^2\frac{[\ell -j - k]}{[\ell + m-1-j - k ] } & \cdots & \prod_{k=1}^j\frac{[\ell -j - k]}{[\ell + m-1-j - k ] } \\
\end{array } \right ) .
\label{def of b}\end{aligned}\ ] ] @xmath204 satisfies the following recursion relation , @xmath207^j(-1)^j \prod_{k=1}^j
\frac{[k]}{[\ell + m-1-k][\ell + m-2-k ] } b_{j-1}^{(\ell -1m-1)}. \label{recursion for b}\end{aligned}\ ] ] this is obtained by subtracting @xmath208-th row from @xmath209-th row for @xmath210 in the determinant in ( [ def of b ] ) and using the identity , @xmath211[b]-[a][b+c]=[b - a][c].\end{aligned}\ ] ] solving ( [ recursion for b ] ) with @xmath212 , we find @xmath213[m - k]}{[\ell + m-1-k]}\right)^{j+1-k } \nonumber\\ & \times \prod_{q=1}^j \prod_{k=1}^{j+1-q } \left ( \frac{[k]}{[\ell + m+1 - 2q - k][\ell + m-2q - k ] } \right).\end{aligned}\ ] ] by taking the ratio as in ( [ d = a / a ] ) , we get the asymptotic form of @xmath149 : @xmath214[\ell -k][m - k]}{[\ell + m-1-k][\ell + m - j - k][\ell + m - j-1-k ] } \right).\end{aligned}\ ] ]
the ground ring structure discovered in @xcite was investigated within the minimal string theory in @xcite .
it was then generalized to the boundary by i. kostov in @xcite , and we employ here the notations of @xcite . on the boundary ,
the ring is generated by two operators denoted @xmath215 whose coordinate derivatives are brst - exact , allowing to move them freely on the boundary .
these operators are built over degenerate matter and liouville operators and have truncated ope with the vertex operators .
below , only the @xmath216 operator , build over the @xmath217 matter primary field , will play a role .
the operator @xmath218 can be inserted between two boundaries with parameters satisfying @xmath219 or @xmath220 .
two different dressed boundary vertex operators can be built , depending on the choice of the root of the kpz equation , @xmath221 where @xmath222 denotes the liouville momentum related to the liouville charge by @xmath223 .
physically realized boundary operators are known to be @xmath224 when @xmath225 and @xmath226 when @xmath227 .
we suppose here that the operators have a negative momentum , @xmath228 , so that we only need to consider the two following fusion relations between the boundary ground ring ( bgr ) and vertex operators , which simply writes within a suitable normalization @xmath229 note that the free field ( @xmath230 ) fusion relations are deformed by the liouville potential , leading to complicated fusion relations in the presence of integrated boundary operators .
similarly , the matter screening charge arising from the coulomb gas representation of the minimal model further modifies these relations .
however , such modifications only arise when considering the fusion of @xmath216 with vertex operators @xmath224 , and the relations [ rel_fusion ] are exact . in this appendix , we interpret the insertion of the monomial @xmath231 on the boundary constructed by @xmath232 as two insertions of the bgr operator @xmath216 on the fzzt brane @xmath233 .
we consider the correlator @xmath234 defined in ( [ n correlator ] ) . in the continuum limit
, the matrix operator @xmath232 creates the boundary @xmath233 , and between boundaries @xmath233 and @xmath235 , the boundary operator @xmath236 should be inserted , with momentum @xmath237 ( see figure [ fig1 ] ) .
this identification does not involve any resonance term since the mlg coupling has maximal scaling so it appears only once in the resonance transformation .
are commuting , leading to conjecture that the order of the boundaries does not matter in the liouville gravity side .
more precisely , @xmath238 with @xmath239 , @xmath240 , @xmath241 and @xmath242 . at the level of the 3pt function , this can be shown using the reflection property . ]
let us now examine the relation @xmath243 from the rhs , we see that the effect of @xmath231 is to modify the criticality of the @xmath209th boundary , @xmath244 , thus shifting the momentum of the neighboring boundary operators ( @xmath245 , @xmath246 ) , and to shift its boundary cosmological constant @xmath247 ( see figure [ fig2 ] ) .
these effects are exactly those obtained by the insertion of two bgr operators @xmath248 ( see figure [ fig3 ] ) . indeed , using the relations [ rel_fusion ] to fuse the bgr operator to the nearby vertex operators ,
we exactly recover the continuum limit of the identity [ mat_bgr ] .
99 a. polyakov , phys . lett .
* b103 * ( 1981 ) 207 .
v. knizhnik , a. polyakov and a. zamolodchikov , mod .
phys . lett .
* a3 * ( 1988 ) 819 .
f. david , mod .
lett . *
a3 * ( 1988 ) 1651 .
j. distler and h. kawai , nucl .
* b321 * ( 1989 ) 509 .
e. brezin and v. kazakov , phys .
* b236 * ( 1990 ) 144 .
m. douglas and s. shenker , nucl .
* b335 * ( 1990 ) 635 .
d. gross and a. migdal , phys .
* 64 * ( 1990 ) 127 ; nucl .
* b340 * ( 1990 ) 333 .
m. douglas , phys .
* b238 * ( 1990 ) 176 .
v. kazakov , mod .
lett * a4 * ( 1989 ) 2125 .
p. di francesco and d. kutasov , nucl .
* b375 * ( 1992 ) 119 .
g. moore , n. seiberg , m. staudacher , nucl .
* b362 * ( 1991 ) 665 .
i. kostov , b. ponsot , d. serban , nucl.phys . * b683 * ( 2004 ) 309 .
j. e. bourgine , k. hosomichi and i. kostov , nucl .
* b832 * , 462 ( 2010 ) ; j. e. bourgine , jhep * 0909 * ( 2009 ) 020 ; j .- e .
bourgine and k. hosomichi , jhep * 0901*:009 ( 2009 ) .
i. kostov , nucl.phys .
* b658 * ( 2003 ) 397 ; j. stat .
mech . * 0708*:p08023 ( 2007 ) .
e. j. martinec , g. w. moore and n. seiberg , phys .
lett . *
b263 * , 190 ( 1991 ) .
k. hosomichi , jhep * 0806*:029 ( 2008 ) .
g. ishiki and c. rim , phys .
* b694 * , 272 ( 2010 ) . | the one matrix model is known to reproduce in the continuum limit the @xmath0 minimal liouville gravity .
recently , two of the authors have shown how to construct arbitrary critical boundary conditions within this matrix model .
so far , between two such boundary conditions only one boundary operator was constructed . in this paper , we explain how to construct all the set of boundary operators that can be inserted . as a consistency check ,
we reproduce the corresponding liouville boundary 2pt function from the matrix model correlator .
in addition , we remark a connection between a matrix model relation and the boundary ground ring operator insertion in the continuum theory .
boundary operators in the one - matrix model jean - emile bourgine@xmath1 , goro ishiki@xmath1 chaiho rim@xmath2
+ + _ sogang university , seoul 121 - 742 , korea _ |
clusters of galaxies are sites of strong morphological evolution of disk galaxies .
while the fraction of ellipticals in clusters remains relatively unchanged with redshift , the fraction of spirals is 23 times larger in intermediate redshift clusters as compared to the present , with a corresponding decrease in the s0 population ( dressler 1980 ; dressler et al .
this dramatic change in the disk galaxy population since @xmath2 suggests that processes associated with cluster virialisation cause the transformation of spirals to s0s .
although a variety of mechanisms have been proposed to explain the morphological change of cluster disk galaxies , there is growing evidence to suggest that gravitational tidal effects are the predominant mechanism for these transformations .
intermediate redshift clusters have an unexpectedly high proportion of galaxies with unusual morphology , suggestive of merging and tidally interacting systems ( e.g. lavery & henry 1988 ; thompson 1988 ; lavery , pierce & mcclure 1992 ; dressler et al .
1994 ; couch , ellis & sharples 1994 ; oemler , dressler & butcher 1997 ) .
furthermore , the dynamically interacting galaxies seem to be responsible for most of the galaxies that show spectroscopic signs of starbursts ( oemler et al .
1997 ) , as may be expected from the consequences of tidal interactions ( e.g. barton et al .
these tidal effects and gravitational interactions are able to transform spirals to s0s by gas stripping ( e.g. valluri & jog 1991 ) ; by the loss of angular momentum of the disk gas , igniting a powerful central starburst which can assist in the formation of a bulge ( e.g. hernquist & mihos 1995 ; barnes & hernquist 1996 ) ; by tidal heating of the disk stabilising it against gravitational instability and suppressing subsequent star formation ( e.g. gnedin 2003b ) ; by truncation of the halo , halting any further infall of cold gas ( e.g. merritt 1984 ; gnedin 2003b ) ; and by minor merger of two spirals to form an s0 ( e.g. bekki 1998 ) . the correlation of star formation rate and local projected density which holds for galaxies up to several virial radii from the centres of clusters ( e.g. kodama et al . 2001 , 2003 ; lewis et al . 2002 ; gomez et al .
2003 ) suggests that pre - processing due to tidal forces in galaxy groups is a significant means to change galaxy morphology ( e.g. zabludoff & mulchaey 1998 ) .
however , recent numerical simulations demonstrate that the tidal forces associated with on - going cluster formation and sub - cluster merging also play a major role in the morphological transformation of cluster disk galaxies ( e.g. bekki 1999 ; gnedin 1999 , 2003a , 2003b ) .
changes in cluster disk galaxy morphology were dramatic in the past .
however it is important to consider whether such changes are continuing in clusters at the present epoch .
if morphological transformations of cluster disk galaxies are occuring at the present , albeit on a reduced scale , the processes involved may be studied in much greater detail than is easily possible at higher redshifts .
in fact , there is growing evidence that activity similar to that seen in distant clusters is still on - going at the present epoch . in previous work ( moss &
whittle 1993 ; moss , whittle & pesce 1998 ; moss & whittle 2000 ) , we have shown that the residual spiral population of low - redshift clusters shares characteristics of the more abundant spiral population in higher redshift clusters .
there is an enhancement of circumnuclear starburst emission in cluster spirals as compared to their counterparts in the field .
moreover , a high proportion of the spirals in low - redshift clusters ( e.g. @xmath3 40% in coma ) have a distorted morphology , strongly correlated with circumnuclear starburst emission , typical of tidally disturbed systems .
similarly caldwell & rose ( 1997 ) in a study of 5 nearby clusters , conclude that 15% of early - type galaxies show signs of on - going or recent star formation , characteristic of starburst or post - starburst galaxies , and that the frequency of such galaxies is enhanced as compared to the field .
further evidence that tides and interactions are not solely the provenance of intermediate redshift clusters is provided by conselice and gallagher ( 1999 ) who show that distorted , interacting and fine - scale substructures are common in low - redshift cluster galaxies .
our previous work was an h@xmath0 survey of spirals ( types sa and later ) in 8 low - redshift clusters , which we used to compare the incidence of star formation in clusters and the field .
in the present paper we use the same data set to complete the survey for early - type galaxies , and compare emission for these galaxies between clusters and the field .
such a comparison is of interest for several reasons .
firstly , previous such comparisons have generally concluded that emission is reduced in cluster early - type galaxies as compared to similar galaxies in the field .
however these earlier results have been cast in doubt due to the discovery of a previously unrecognised selection effect ( cf .
biviano et al .
our survey data is expected to be free of this selection effect , and we are able to provide an unbiased comparison of emission in cluster and field early - type galaxies ( see discussion in section [ scfcomp ] below ) . secondly , previous results for low - redshift clusters ( cf .
caldwell et al .
1993 ; caldwell & rose 1997 ) , lead to an expectation of an enhancement of starbursts in the cluster early - type population as compared to the field .
using the present survey data , we can attempt to confirm this prediction . and
finally , data for detected early - type and late - type emission - line galaxies ( elgs ) may be combined to gain further insight into processes affecting cluster galaxies particularly from the kinematical properties of the various elg subgroups .
this last investigation is the topic for a subsequent paper ( moss 2004 , in preparation ) .
identifications of elgs from combined h@xmath0 + [ nii ] emission for early - type galaxies from the h@xmath0 survey data have already been published for abell 1367 ( cf .
moss , whittle & pesce 1998 ) in section [ opsurvey ] we list the early - type galaxies surveyed , and present elg identifications for the remaining 7 clusters . a comparison of emission in field and cluster early - type galaxies is made in section [ scfcomp ] . in section [ emmech ]
we investigate possible enhancement of starburst emission in cluster early - type galaxies .
conclusions are given in section [ concl ] .
an objective prism survey for combined h@xmath0 + [ nii ] emission from cluster galaxies in 8 low - redshift clusters ( viz .
abell 262 , 347 , 400 , 426 , 569 , 779 , 1367 and 1656 ) has been undertaken using the 61/94-cm burrell schmidt telescope on kitt peak .
the survey technique and methods , and the plate material used have been described in detail in previous papers ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* papers i iv respectively ) . for convenience ,
a brief summary of these details is given here .
for each field , two plates were taken using hypersensitised iiiaf emulsion , with an emulsion / filter combination giving a @xmath3 350 bandpass centred on 6655 with a peak sensitivity of @xmath3 6717 .
all plates were taken in conditions of good seeing and good transparency , and an h@xmath0 detection was accepted only if the galaxy was independently detected on both plates .
previous work has shown that the approximate detection limit of h@xmath0 + [ nii ] emission is an equivalent width , @xmath4 @xmath5 20 .
the initial survey list comprised all 727 cgcg galaxies within a radial distance from the cluster centre , @xmath6 , where @xmath7 is the abell radius @xcite .
in addition , 79 cgcg galaxies ( all in abell 1367 , except one in abell 400 ) which lie in the region , @xmath8 , were included in the survey .
adopted values of the abell radius for each cluster , and plate boundaries of the survey plate material are given in papers iii and iv .
of the 806 cgcg galaxies in the initial list , 37 are double systems .
the components of these were surveyed separately , giving a total of 843 galaxies .
galaxy types according to the revised de vaucouleurs system ( de vaucouleurs 1959 , 1974 ) were obtained for all galaxies in the initial survey list .
types were either taken from the ugc @xcite or determined by inspection from a variety of schmidt plate material .
details of the type classification procedure are given in papers iii and iv .
types for galaxies in abell 1367 are given in paper iii , and for galaxies of types sa and later for the remaining clusters in paper iv .
papers iii and iv list types and h@xmath0 + [ nii ] emission detection for 460 galaxies , which are mainly of types sa and later .
the present paper completes the survey by listing types and h@xmath0 + [ nii ] emission detection for the remaining 383 galaxies of the survey , which are predominantly of types s0/a and earlier . combining the data from this paper with those for papers iii and iv
, there is a total of 95 galaxies omitted from the survey due to plate defects ( 68 ) , or a velocity @xmath9 12000 km @xmath10 ( 27 ) .
thus the final total of surveyed galaxies is 748 . in table
[ tsurvey ] we give galaxy types for the remaining 383 galaxies of the initial survey list , together with the heliocentric velocity taken from the nasa extragalactic database ( ned ) . with the exception of 8 galaxies in abell 262 typed as spirals and one galaxy in abell 347 typed as ` peculiar ' , accidentally overlooked in previous work ,
the remaining 374 galaxies are all of types s0/a and earlier , or are untyped ( type class ` @xmath11 ' ) .
of the galaxies listed in table 1 , 30 could not be surveyed due to plate defects ( these galaxies are listed in the notes to the table ) , or due to a velocity @xmath9 12000 km @xmath10 , since in the latter case any h@xmath0 emission is redshifted beyond the sensitivity limit of the plate .
thus there is a total of 353 surveyed galaxies . of these 353 predominantly early - type galaxies ,
28 galaxies were detected in emission . the emission - line galaxies ( elgs )
are listed in table [ gsurvey ] . for galaxies in table 2 which are untyped ( type class ` ... ' ) , we list additional type information from ned , where available . the visual classification of the detected emission according to visibility ( s strong ; ms medium - strong ; m medium ; mw medium - weak ; and w weak ) and concentration ( vc very concentrated ; c concentrated ; n normal ; d diffuse ; and vd very diffuse ) is given according to the scheme used in previous work ( cf .
papers i iv ) .
similarly , for the subsequent analysis , we choose binary ranks for the h@xmath0 appearance , yielding two parameters : _ compact _ emission ( concentration classes vc , c or n ) ; and _ diffuse _ emission ( concentration classes d or vd ) .
notes on individual objects are appended to the table . in table
[ etypes ] we summarise emission detection frequency with morphological type . in order to approximate a volume - limited sample ( cluster galaxies and galaxies proximate to the cluster in the supercluster field ) , we restricted the sample to galaxies with velocities within 3@xmath12 of the cluster mean .
for each morphological type class , the table lists the total sample number ( @xmath13 ) , the numbers of detected galaxies with compact and diffuse emission ( @xmath14 and @xmath15 respectively ) , and the overall percentage of galaxies detected in emission ( @xmath16 ) .
corresponding values are also given when ned types , where available , have been included for galaxies with undetermined types ( @xmath17 , @xmath18 , @xmath19 , @xmath20 ) .
similar results have been given and discussed previously ( cf .
paper iv ) .
the present work provides a greatly increased sample for early - type galaxies ( s0/a and earlier ) and confirms , as expected , a much lower detection frequency for early - type galaxies as compared to spirals and later types .
moreover the detected emission for early - type galaxies is seen to be predominantly compact emission .
the likely origin for this emission is discussed in section [ emmech ] below .
the earliest studies which compared the frequency of emission between field and cluster galaxies were in agreement in finding , _ for a given galaxy morphological type _ , a lower frequency of elgs in clusters ( e.g. osterbrock 1960 , gisler 1978 , dressler et al .
1985 , hill & oegerle 1993 ) . however biviano et al . ( 1997 ) identified a hitherto unsuspected systematic effect which causes an overestimate in the fraction of elgs at fainter magnitudes , due to the bias that operates against the successful determination of redshifts for faint galaxies without emission lines . when field galaxies are on average fainter than the cluster galaxy sample and redshift data are incomplete , this systematic effect works to overestimate the frequency of emission in field as compared to cluster galaxies
. based on the eso nearby abell cluster survey spectral data ( 5634 galaxies in the directions of 107 cluster candidates ) , biviano et al . concluded that the observed difference in frequency of elgs between field and cluster galaxies could be entirely accounted for by the variation in this frequency with galaxy morphological type , and the differing morphological mix between field and cluster galaxies . by inference ,
there is expected to be no difference in the elg frequency for galaxies of a given morphological type between cluster and field environments , in disagreement with all earlier studies . for the present prism survey , field and cluster galaxies
have been surveyed in an identical manner to the same magnitude limit .
redshift data are available for most ( @xmath3 96% ) of the galaxy sample , and field and cluster samples have both been limited to galaxies with velocities within 3@xmath12 of the cluster mean .
field galaxies thus comprise an approximately volume - limited sample , and both field and cluster galaxies are expected to have similar distributions in both apparent and absolute magnitude .
accordingly , we expect comparative frequencies of cluster and field elgs determined from this survey to be free of the bias identified by biviano et al . as well as systematic effects due to any dependence of the h@xmath0 emission on absolute magnitude . using data from the prism survey
, we have previously shown that for galaxies of types sa and later , there is indeed no difference in the elg frequency between field and cluster environments , in accord with the conclusions of biviano et al .
paper iv and references therein ) .
a similar result has been obtained by gavazzi et al .
( 1998 ) for galaxies of types sa and later .
these authors compared h@xmath0 e.w.s for volume - limited samples of cluster and supercluster field galaxies to the same magnitude limits in the coma supercluster , and concluded that there was no significant difference between the two samples .
[ cols="<,^,<,>,<,^,<,>,<,^ , < , > " , ] @l@l@cc & & & + 160 - 160 & pec & hii & 2 + 160 - 179 & s : pec & hii & 2 + 160 - 180 & pec & hii & 2 + ' '' '' 160 - 193 & s : ... & sbrst & 2 + explanation of the columns of the table .
cgcg number ( zwicky et al .
19601968 ) .
see notes to column 1 of table [ tsurvey ] above .
galaxy type taken from papers iii & iv , and this paper .
@xmath21 indicates that types were taken from ned .
columns 3 & 4 : spectral classification taken from : 1 .
miller & owen ( 2002 ) : sf star - forming galaxy ; ab+ predominantly absorption - line spectrum , although with slight emission of [ nii ] and sometimes [ sii ] ; sey seyfert ; lin liner ; mix nuclear spectrum that of a seyfert or liner , with off - nuclear spectrum showing star formation .
ned : either ned galaxy classification ( sbrst , sy , sy1 , sy2 ) or classification as hii or agn emission from the line ratios of emission lines in the uzc spectral archive ( faldo et al . 1999 ) .
the above expected percentages of compact elgs of different galaxy types with hii emission may be compared with actual percentages from spectroscopic data . in table
[ spt ] , we list spectral classifications for 54 compact elgs in the 8 clusters surveyed . these represent 69% of the total of 78 compact elgs which have a velocity within 3@xmath12 of the cluster mean .
spectral classifications were taken from miller & owen ( 2002 ) or derived from data given in ned .
for these latter , the galaxy classification was either taken from ned , or , when not available , classification of the emission as either hii or agn was made using emission line ratios in the uzc spectral archive ( faldo et al .
1999 ) , following the method of veilleux and osterbrock ( 1987 ) .
lrrrcrrr & & & + & & & & & + & & & & & + & & & & & & & + & 302 & 3.6 & 1.3 & & 28 & 7.1 & 0.0 + & 177 & 19.2 & 7.3 & & 45 & 8.9 & 0.0 + note to the table : @xmath22 @xmath23 and @xmath24 @xmath25 are the percentages of galaxies with compact hii emission and which both have compact hii emission and are disturbed , for cluster and field galaxies respectively .
ned types for galaxies with otherwise indeterminate type have been included where available .
for the three type groups , e , s0,s0/a ; sa , sab , sb ; and sbc , sc , sc
irr , restricted to galaxies within one abell radius of the cluster centres ( or within 0.5 abell radii of each of the two subcluster centres in abell 569 ) , the observed percentages of galaxies with hii emission are 68% ( @xmath26 ) , 80% ( @xmath27 ) , and 100% ( @xmath28 ) respectively .
for the two later type groups , these values are in good agreement with the expected percentages of galaxies with hii emission . for galaxies classified as peculiar ,
the observed percentage is 90% ( @xmath27 ) in accord with the suggestion above that these galaxies are predominantly on - going mergers .
however for early type galaxies , the observed percentage is higher than expected ( significance level @xmath3 2.8@xmath12 ) , and suggests an enhancement of hii emission for the early - type cluster galaxies with compact emission . if this enhancement of hii emission in early - type cluster galaxies is real , what is its likely cause ? obviously , the simplest explanation is that previously suggested for cluster later types , viz .
the effects of gravitational tidal interactions on the galaxies . for the present sample of cluster early - type galaxies ( types e , s0 and s0/a ) , restricted to galaxies within one abell radius of the cluster centres ,
there is a very strong correlation between the incidence of compact hii emission and a disturbed morphology for the galaxy ( significance level of 8.2@xmath12 ) .
four out of nine disturbed galaxies show compact hii emission , while only 6 out of 270 undisturbed galaxies show this emission .
this is a similar result to that found previously for cluster galaxies of types sa and later , and indeed suggests that any enhancement of hii emission in cluster early - type spirals is due to gravitational tidal effects as is the case for later type cluster galaxies . in table
[ percents ] we compare the fractions of galaxies in the field and in clusters which have compact hii emission ( @xmath24 and @xmath22 respectively ) and the fractions which have compact hii emission and in addition are tidally disturbed ( @xmath29 and @xmath30 respectively ) .
cluster galaxies are defined to be those within @xmath31 of the cluster centres ( or within @xmath32 of each of the two subcluster centres in abell 569 ) , while field galaxies are defined to be those outside @xmath31 for the clusters abell 262 , 347 , 400 , 569 and 779 , and outside @xmath33 for abell 426 , 1367 and 1656 ( cf .
paper iv ) .
we make the comparison for e , s0,s0/a and sa + later galaxies separately .
first consider the later types .
we find that the increase in the percentage of galaxies with compact hii emission in clusters is largely accounted for by an additional population of _ disturbed _ galaxies with compact hii emission
indeed there are _ no _ such galaxies in the field .
this restates our earlier findings in paper iv which focussed specifically on the later types . for the early - type galaxies , again there is a population of _ disturbed _ galaxies with compact hii emission which have no counterparts in the field .
the number of these galaxies ( @xmath34 ) is exactly that required to explain the observed excess of hii emission in cluster early type elgs as determined from a comparison with the pssn spectroscopic data for field galaxies ( cf . section [ emmech ] above ) .
thus for early - types as well as later types , an enhancement of compact hii emission in cluster galaxies is readily explained by gravitational tidal interactions associated with cluster virialisation .
, the percentage of early - type galaxies with compact hii emission in the field is numerically greater than in the cluster .
however , as discussed above in section [ scfcomp ] , this cluster
field difference is not statistically significant . ]
in a series of papers ( cf .
moss et al . 1988 ; moss & whittle 1993 ; moss et al .
1998 ; moss & whittle 2000 ) we have undertaken an h@xmath0 survey of an essentially complete sample of 748 cgcg galaxies in 8 low - redshift clusters ( viz .
abell 262 , 347 , 400 , 426 , 569 , 779 , 1367 and 1656 ) .
this paper has presented previously unpublished data for 383 mainly early - type galaxies and completes publication of data for the survey .
the combined survey data show that emission detection increases as expected from earlier to later galaxy types ( 3% for e , e s0
; 6% for s0 ; 9% for s0/a ; and 41% for types sa and later ) .
a comparison of cluster and field early - type galaxies shows a similar frequency of emission detection . together with the same result obtained for cluster and field galaxies of types sa and later , these data confirm the inference of biviano et al .
( 1997 ) that differences between the frequency of elgs between clusters and the field can entirely be accounted for by the differing mix of galaxy morphological types in the two environments , while _ for a given morphological type _ there is no difference in the frequency of elgs between clusters and the field . as noted by biviano et al .
, this result is in disagreement with all or most previous studies .
it is to be noted that this result has been obtained for galaxies with relatively strong emission ( @xmath1 ) .
work is in progress to extend these results to fainter limits in equivalent width ( cf .
sakai , kennicutt & moss 2001 ) .
although the incidence of emission does not vary between cluster and field environments , the survey has shown that the type of emission does vary .
detected emission is classified as ` compact ' or ` diffuse ' , identified as circumnuclear starburst or agn emission and disk emission respectively . in previous work , we have shown that for galaxies of types sa and later , there is an enhancement of compact hii emission in cluster galaxies as compared to the field .
this type of emission has been shown to be strongly correlated with a tidally disturbed morphology of the galaxy . in the present work , a comparison of spectroscopic data for the cluster early - type elgs with that for field galaxies from the pssn ( ho et al .
1997 ) gives a modest indication ( significance level @xmath3 2.8@xmath12 ) for enhanced compact hii emission in early - type cluster elgs as compared to the field .
moreover , the compact hii emission in the early - type cluster galaxies is strongly correlated with a disturbed galaxy morphology ( significance level of 8.2@xmath12 ) . for both early - type and later types
, there is a population of disturbed galaxies with compact hii emission which have no counterparts in the field .
this suggests that for cluster galaxies of all types , enhancement of compact hii emission can readily be explained as an enhancement of circumnuclear starburst emission due to gravitational tidal effects . as discussed previously ( cf .
moss & whittle 2000 ) these gravitational tidal effects are most likely to be associated with sub - cluster merging and other processes of on - going cluster virialisation .
observations were made with the burrell schmidt telescope of the warner and swasey observatory , case western reserve university .
this research has made use of the nasa / ipac extragalactic database ( ned ) , which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration . | we have extended our h@xmath0 objective prism survey of 8 low - redshift clusters ( viz .
abell 262 , 347 , 400 , 426 , 569 , 779 , 1367 and 1656 ) to include a complete sample of early - type galaxies within 1.5 abell radii of the cluster centres . of the 379 galaxies
surveyed , 3% of e , e s0 galaxies ,
6% of s0s and 9% of s0/a galaxies were detected in emission . from a comparison of cluster and supercluster field galaxies ,
we conclude that the frequency of emission - line galaxies ( elgs ; @xmath1 ) is similar for field and cluster early - type galaxies .
a similar result has previously been obtained for galaxies of types sa and later .
together , these results confirm the inference of biviano et al . that the relative frequency of elgs in clusters and the field can be entirely accounted for by the different mix of morphological types between the differing environments , and that , for galaxies of a given morphological type , the fraction of elgs is independent of environment .
detected emission is classified as ` compact ' or ` diffuse ' , identified as circumnuclear starburst or agn emission and disk emission respectively . by comparing spectroscopic data for cluster early - type elgs with data for field galaxies from the palomar spectroscopic survey of nearby galactic nuclei ,
we demonstrate there is modest evidence for an enhancement of compact hii emission relative to agn emission in the early - type cluster elgs as compared to the field . for the cluster early - type galaxies ,
compact hii emission correlates strongly with a disturbed morphology .
this suggests that , as for later type cluster galaxies , this enhanced compact hii emission can readily be explained as an enhancement of circumnuclear starburst emission due to gravitational tidal interactions , most likely caused by sub - cluster merging and other on - going processes of cluster virialisation .
surveys galaxies : clusters : general galaxies : elliptical and lenticular , cd galaxies : evolution galaxies : interactions galaxies : starburst |
grb 970228 was detected as a 5.5-s brief , high energy event on 28 february 1997 , by the satellite bepposax ( costa et al .
1997a ) . a decaying x - ray emission ( the x - ray afterglow ) was found 8 hr after the onset of the event ( costa et al .
optical observations started 15.4 hr after the burst ( pedichini et al .
1997 ) , which allowed the search of variable objects in the intersection between the @xmath2-ray error box , the x - ray error box and the interplanetary network ( ipn ) annulus ( hurley et al . 1997 ) .
they led to the discovery of the first optical counterpart of a grb ( van paradijs et al .
1997 ) .
the recent measurement of the redshift of the grb 970508 optical counterpart has established that most grbs - if not all - lie at cosmological distances ( metzger et al .
while galactic grb models allow multiple outbursts from each source ( li and dermer 1992 ) , the cosmological models are usually based on singular , cataclysmic highly energetic events .
thus , the detection of recurrent optical emission for a grb would impose severe constraints on the cosmological models ( narayan , paczyki and piran 1992 , mszros and rees 1993 ) .
archival searches can be crucial in order to clarify this problem .
archival searches in grb error boxes rely upon the assumption that grbs do repeat and show optical transient emission . with perhaps the exception of ot 1928 ( schaefer 1981 , hudec et al .
1994 ) , no firm candidates have yet been established ( hudec et al . 1993 , and references therein , also gorosabel and castro - tirado 1998 ) .
the detection of the two first optical counterparts for grb 970228 ( van paradijs et al .
1997 , guarnieri et al . 1997 ) and grb 970508 ( bond 1997,castro - tirado et al .
1998 ) has been an important breakthrough .
however , for some other grbs no optical emission was detected although very fast and deep follow up observations were carried out ( castro - tirado et al .
1997 , groot et al .
in order to search for optical recurrent activity at the position of the grb 970228 optical counterpart we have examined 8004 plates at the harvard college observatory plate collection ( hco ) . table 1 displays the list of plates examined at hco .
the plates span 64 yr , from 1889 to 1952 , and the total exposure time amounts to @xmath0 yr
. the accurate position of the optical counterpart made the archival search very easy in comparison to other grbs with larger error boxes .
many spots were found on the plates , but none of them was consistent with the position of the true optical counterpart , except for a @xmath3 9 mag star - like spot found on the plate rh 4888 ( see fig .
this plate was taken on 18 march 1933 and represents three exposures of 6 min each .
two of the three exposures included the grb 970228 field and are shifted @xmath4 from each other .
the third one includes the north equatorial pole .
as expected for a real optical transient , we found two spots corresponding -within the astrometrical errors- to both the real and `` shifted '' positions of the true grb 970228 optical counterpart .
however , an accurate measurement of the two positions revealed that they are not trailed in a similar way to other stellar images .
we then realized that the two spots correspond to a pair of stars near the north pole that were imaged during the third exposure .
.plates examined at hco for grb 970228 [ cols="^,^,^,^,^,^ " , ] in order to estimate the rate of plate faults , we investigated the number of spots found on the ac plates in a @xmath5 radius region around the ot position ( see fig . 2 ) .
the rate of plate faults with 0.20.6 mag above the background is @xmath6
mm@xmath7 plate@xmath8 , a value comparable to that reported by greiner et al .
the @xmath3 8000 plates span @xmath31.1 yr .
many star - like spots were found . with the exception of the above - mentioned case ,
none of them was consistent with the position of the true grb 970228 optical counterpart .
therefore we can settle a lower limit of @xmath3 1.1 yr for any recurrent optical transient emission activity brighter than 12.5 mag ( or 4.2 mag if a 1-s flash is assumed ) . according to extrapolation of the gamma - ray spectra into longer wavelengths for the strongest bursts observed by batse
, the magnitude of the optical flash that -eventually -would arise simultaneously to the gamma - ray event , could reach a magnitude of 2.5 in 1-s ( for the spectral index @xmath9 @xmath10 1 , ford and band , 1996 ) . if this would have been the case for a previous bursting activity at the grb 970228 location , such an optical flash would have been easily detected in any of the hco plates .
the authors wish to thank r. hudec for fruitful conversations .
we are also indebted to the hco plate curator m. l. hazen , and to a. doane for the facilities given at the hco plate stacks . | we have examined 8004 plates at the harvard college observatory plate collection searching for optical transient emission from the gamma - ray burst grb 970228 .
this is the first archival search carried out so far for a gamma - ray burst with known transient optical emission .
the total exposure time amounts to @xmath0 yr .
no convincing optical activity was found above @xmath1 mag at the expected position of the grb 970228 optical counterpart . |
a central goal in quantum chromodynamics is to describe the structure and dynamics of hadrons at the amplitude level .
the light - front fock expansion provides a physical description of hadrons as composites of quarks and gluons analogous to the @xmath6 momentum - space wavefunction description of nonrelativistic bound states of the schrdinger theory .
the light - front wavefunctions @xmath7 are functions of the constituent light - cone fractions @xmath8 , relative transverse momenta @xmath9 , and spin projections @xmath10 they are relativistic and frame - independent , describing all particle number excitations @xmath11 of the hadrons .
the light - front fock expansion follows from the quantization of qcd at fixed light - front time @xmath12 .
the bound - state hadronic solutions @xmath13 are eigenstates of the light - front heisenberg equation @xmath14 @xcite .
the spectrum of qcd is given by the eigenvalues @xmath15 .
the projection of each hadronic eigensolution on the free fock basis : @xmath16 then defines the lf fock expansion in terms of the quark and transversely polarized gluon constituents in @xmath17 light - cone gauge .
the expansion has only transversely polarized gluons . the freedom to choose the light - like quantization four - vector provides an explicitly covariant formulation of light - front quantization and can be used to determine the analytic structure of light - front wave functions and to define a kinematical definition of angular momentum @xcite .
the front form thus provides a consistent definition of relative orbital angular momentum and @xmath18 conservation : the total spin projection @xmath19 is conserved in each fock state .
the cluster decomposition theorem @xcite and the vanishing of the anomalous gravitomagnetic moment " @xmath20 @xcite are immediate properties of the lf fock wavefunctions @xcite .
given the light - front wavefunctions @xmath21 , one can compute a large range of hadron observables .
for example , the valence and sea quark and gluon distributions which are measured in deep inelastic lepton scattering are defined from the squares of the lfwfs summed over all fock states @xmath11 .
form factors , exclusive weak transition amplitudes @xcite such as @xmath22 . and
the generalized parton distributions @xcite measured in deeply virtual compton scattering are ( assuming the handbag " approximation ) overlaps of the initial and final lfwfs with @xmath23 and @xmath24 .
the gauge - invariant distribution amplitude @xmath25 defined from the integral over the transverse momenta @xmath26 of the valence ( smallest @xmath11 ) fock state provides a fundamental measure of the hadron at the amplitude level @xcite ; they are the nonperturbative input to the factorized form of hard exclusive amplitudes and exclusive heavy hadron decays in perturbative qcd . the resulting distributions
obey the dglap and erbl evolution equations as a function of the maximal invariant mass , thus providing a physical factorization scheme @xcite . in each case , the derived quantities satisfy the appropriate operator product expansions , sum rules , and evolution equations .
however , at large @xmath27 where the struck quark is far - off shell , dglap evolution is quenched @xcite , so that the fall - off of the dis cross sections in @xmath28 satisfies inclusive - exclusive duality at fixed @xmath29 the light - front fock - state wavefunctions encode the bound state properties of hadrons in terms of their quark and gluon degrees of freedom at the amplitude level .
they display novel features , such as intrinsic gluons , asymmetric sea - quark distributions @xmath30 , @xmath31 , and intrinsic heavy - quark fock states @xcite .
intrinsic charm and bottom quarks appear at large @xmath27 in the light - front wavefunctions since this minimizes the invariant mass and off - shellness of the higher fock state .
one can use the operator product expansion to show that the probability of such states scales as @xmath32 in contrast to @xmath33 fall - off of abelian theory @xcite .
the remarkable observations of the selex experiment of the double - charm baryon @xmath34 in @xmath35 and @xmath36 at large @xmath37 @xcite provides compelling evidence for double - charm intrinsic fock states in the proton .
the coherence of multi - particle correlations within the fock states leads to higher - twist bosonic processes such as @xmath38 ; although suppressed by inverse powers of @xmath28 , such subprocesses are important in the duality regime of fixed @xmath39 , particularly in @xmath40 @xcite . in the case of nuclei
, one must include non - nucleonic hidden color " @xcite degrees of freedom of the deuteron lfwf .
the e791 experiments at fermilab @xcite has shown how one can measure the valence lfwf directly from the diffractive di - jet dissociation of a high energy pion @xmath41 into two jets , nearly balancing in transverse momentum , leaving the nucleus intact .
the measured pion distribution in @xmath27 and @xmath42 is similar to the form of the asymptotic distribution amplitude and the ads / cft prediction discussed below .
the e791 experiment also find that the nuclear amplitude is additive in the number of nucleons when the quark jets are produced at high @xmath43 , thus giving a dramatic confirmation of color transparency " , a fundamental manifestation of the gauge nature of qcd @xcite .
the phase structure of hadron matrix elements is an essential feature of hadron dynamics .
although the lfwfs are real for a stable hadron , they acquire phases from initial state and final state interactions .
in addition , the violation of @xmath44 invariance leads to a specific phase structure of the lfwfs @xcite .
contrary to parton model expectations , the rescattering of the quarks in the final state in dis has important phenomenological consequences , such as leading - twist diffractive dis @xcite and the sivers single - spin asymmetry @xcite .
the sivers asymmetry depends on the same matrix elements which produce the anomalous magnetic moment of the target nucleon as well as the phase difference of the final - state interactions in different partial waves .
the rescattering of the struck parton generates dominantly imaginary diffractive amplitudes , giving rise to an effective hard pomeron " exchange and a rapidity gap between the target and diffractive system , while leaving the target intact .
this bjorken - scaling physics , which is associated with the wilson line connecting the currents in the virtual compton amplitude survives even in light - cone gauge .
thus there are contributions to the dis structure functions which are not included in the light - front wave functions computed in isolation and can not be interpreted as parton probabilities @xcite .
diffractive deep inelastic scattering in turn leads to nuclear shadowing at leading twist as a result of the destructive interference of multi - step processes within the nucleus .
in addition , multi - step processes involving reggeon exchange leads to antishadowing .
in fact , because reggeon couplings are flavor specific , antishadowing is predicted to be non - universal , depending on the type of current and even the polarization of the probes in nuclear dis @xcite .
another interesting consequence of qcd at the amplitude level is the @xmath28-independent `` @xmath45 fixed - pole '' contribution @xmath46 to the real part of the compton amplitude , reflecting the effective contact interaction of the transverse currents @xcite .
deeply virtual compton scattering can also be studied in the timelike domain from @xmath47 ; the lepton charge asymmetry and single - spin asymmetries allow measurements of the relative phase of timelike form factors and the @xmath48 amplitude . in principle
, one can solve for the lfwfs directly from the fundamental theory using methods such as discretized light - front quantization ( dlcq ) @xcite , the transverse lattice @xcite , lattice gauge theory moments @xcite , dyson - schwinger techniques @xcite , and bethe
salpeter techniques @xcite .
dlcq has been remarkably successful in determining the entire spectrum and corresponding lfwfs in one space - one time field theories @xcite , including qcd(1 + 1 ) @xcite and supersymmetric qcd(1 + 1 ) @xcite .
the dlcq boundary conditions allow a truncation of the fock space to finite dimensions while retaining the kinematic boost and lorentz invariance of light - front quantization .
there are also light - front solutions for yukawa theory in physical ( 3 + 1 ) space - time dimensions @xcite with a limited fock space . as emphasized by weinstein and
vary , new effective operator methods @xcite which have been developed for hamiltonian theories in condensed matter and nuclear physics , could also be applied advantageously to light - front hamiltonian .
a review of nonperturbative light - front methods may be found in reference @xcite .
as we discuss below , one can use the ads / cft correspondence to obtain a model for hadronic light - front wavefunctions which display confinement at large inter - quark separation and conformal symmetry at short distances . in particular , the scaling and conformal properties of the lfwfs at high relative momenta agree with perturbative qcd .
these ads / cft model wavefunctions could be used as an initial ansatz for a variational treatment of the light - front qcd hamiltonian .
the ads / cft correspondence @xcite , between strongly - coupled conformal gauge theory and weakly - coupled string theory in the 10-dimensional @xmath49 space is now providing a remarkable new insight into the hadron wavefunctions of qcd . the central mathematical principle underlying ads / cft duality is the fact that the group @xmath50 of poincar and conformal transformations of physical @xmath51 space - time has an elegant mathematical representation on @xmath52 space where the fifth dimension has the anti - de sitter warped metric .
the group of conformal transformations @xmath50 in 3 + 1 space is isomorphic to the group of isometries of ads space , @xmath53 , @xmath54 , where @xmath55 represents the coordinate in the fifth dimension .
the dynamics at @xmath56 in 3 + 1 space thus matches the behavior of the theory at the boundary @xmath57 this allows one to map the physics of quantum field theories with conformal symmetry to an equivalent description in which scale transformations have an explicit representation in ads space . even though quantum chromodynamics is a broken conformal theory , the ads / cft correspondence has led to important insights into the properties of qcd . for example , as shown by polchinski and strassler @xcite , the ads / cft duality , modified to give a mass scale , provides a nonperturbative derivation of the empirically successful dimensional counting rules @xcite for the leading power - law fall - off of the hard exclusive scattering amplitudes of the bound states of the gauge theory .
the modified theory generates the hard behavior expected from qcd instead of the soft behavior characteristic of strings .
other important applications include the description of spacelike hadron form factors at large transverse momentum @xcite and deep inelastic scattering structure functions at small @xmath27 @xcite .
the power falloff of hadronic light - front wave functions ( lfwf ) including states with nonzero orbital angular momentum is also predicted @xcite . in the original formulation by maldacena
@xcite , a correspondence was established between a supergravity string theory on a curved background and a conformally invariant @xmath58 super yang - mills theory in four - dimensional space - time .
the higher dimensional theory is @xmath49 where @xmath59 is the radius of ads and the radius of the five - sphere and @xmath60 is the string scale .
the extra dimensions of the five - dimensional sphere @xmath61 correspond to the @xmath62 global symmetry which rotates the particles present in the supersymmetric yang mills supermultiplet in the adjoint representation of @xmath63 . in our application to qcd , baryon number in qcd is represented as a casimir constant on @xmath64 the reason why ads / cft duality can have at least approximate applicability to physical qcd is based on the fact that the underlying classical qcd lagrangian with massless quarks is scale - invariant @xcite
. one can thus take conformal symmetry as an initial approximation to qcd , and then systematically correct for its nonzero @xmath65 function and quark masses @xcite .
conformal template " approach underlies the banks - zak method @xcite for expansions of qcd expressions near the conformal limit and the blm method @xcite for setting the renormalization scale in perturbative qcd applications . in the blm method
the corrections to a perturbative series from the @xmath65-function are systematically absorbed into the scale of the qcd running coupling .
an important example is the generalized crewther relation " @xcite which relates the bjorken and gross - llewellyn sum rules at the deep inelastic scale @xmath28 to the @xmath66 annihilation cross sections at specific commensurate scales @xmath67 .
the crewther relation @xcite was originally derived in conformal theory ; however , after blm scale setting , it becomes a fundamental test of physical qcd , with no uncertainties from the choice of renormalization scale or scheme .
qcd is nearly conformal at large momentum transfers where asymptotic freedom is applicable .
nevertheless , it is remarkable that dimensional scaling for exclusive processes is observed even at relatively low momentum transfer where gluon exchanges involve relatively soft momenta @xcite .
the observed scaling of hadron scattering at moderate momentum transfers can be understood if the qcd coupling has an infrared fixed point @xcite . in this sense , qcd resembles a strongly - coupled conformal theory .
the duality between a gravity theory on @xmath68 space and a conformal gauge theory at its @xmath69-dimensional boundary requires one to match the partition functions at the @xmath70 boundary , @xmath71 .
the physical string modes @xmath72 , are plane waves along the poincar coordinates with four - momentum @xmath73 and hadronic invariant mass states @xmath74 . for large-@xmath55 or small-@xmath75 , @xmath76 , where the dimension @xmath5 of the string mode must be the same dimension as that of the interpolating operator @xmath77 which creates a specific hadron out of the vacuum : @xmath78 .
the physics of color confinement in qcd can be described in the ads / cft approach by truncating the ads space to the domain @xmath79 , where @xmath80 .
the cutoff at @xmath81 is dual to the introduction of a mass gap @xmath82 ; it breaks conformal invariance and is responsible for the generation of a spectrum of color - singlet hadronic states .
the truncation of the ads space insures that the distance between the colored quarks and gluons as they stream into the fifth dimension is limited to @xmath83 .
the resulting @xmath51 theory has both color confinement at long distances and conformal behavior at short distances .
the latter property allows one to derive dimensional counting rules for form factors and other hard exclusive processes at high momentum transfer .
this approach , which can be described as a bottom - up " approach , has been successful in obtaining general properties of the low - lying hadron spectra , chiral symmetry breaking , and hadron couplings in ads / qcd @xcite in addition to the hard scattering predictions @xcite . in this classical holographic model " , the quarks and gluons propagate into the truncated ads interior according to the ads metric without interactions . in effect , their wilson lines , which are represented by open strings in the fifth dimension , are rigid .
the resulting equations for spin 0 , @xmath84 , 1 and @xmath85 hadrons on @xmath49 lead to color - singlet states with dimension @xmath86 and @xmath87 .
consequently , only the hadronic states ( dimension-@xmath88 ) @xmath89 @xmath90 pseudoscalar and vector mesons , the ( dimension-@xmath87 ) @xmath91 @xmath92 baryons , and the ( dimension-@xmath93 ) @xmath94 @xmath1 gluonium states , can be derived in the classical holographic limit @xcite .
this description corresponds to the valence fock state as represented by the light - front fock expansion .
hadrons also fluctuate in particle number , in their color representations ( such as the hidden - color states @xcite of the deuteron ) , as well as in internal orbital angular momentum .
the higher fock components of the hadrons are manifestations of the quantum fluctuations of qcd ; these correspond to the fluctuations of the bulk geometry about the fixed ads metric . for spinning strings orbital excitations of hadronic states
correspond to quantum fluctuations about the ads metric @xcite .
it is thus also natural to identify higher - spin hadrons with the fluctuations around the spin 0 , @xmath84 , 1 and @xmath85 classical string solutions of the @xmath95 sector @xcite . as a specific example , consider the twist - two ( dimension minus spin ) gluonium interpolating operator @xmath96 with total internal space - time orbital momentum @xmath97 and conformal dimension @xmath98 .
we match the large @xmath55 asymptotic behavior of each string mode to the corresponding conformal dimension of the boundary operators of each hadronic state while maintaining conformal invariance . in the conformal limit
, an @xmath99 quantum , which is identified with a quantum fluctuation about the ads geometry , corresponds to an effective five - dimensional mass @xmath100 in the bulk side .
the allowed values of @xmath100 are uniquely determined by requiring that asymptotically the dimensions become spaced by integers , according to the spectral relation @xmath101 @xcite .
the four - dimensional mass spectrum follows from the dirichlet boundary condition @xmath102 , @xmath103 , on the ads string amplitudes for each wave functions with spin @xmath104 2 .
the eigenspectrum is then determined from the zeros of bessel functions , @xmath105 .
the predicted spectra @xcite of mesons and baryons with zero mass quarks is shown in figs .
[ fig : mesonspec ] and [ fig : baryonspec ] .
the only parameter is @xmath106 gev , and @xmath107 gev for mesons and baryons , respectively .
gev : ( a ) vector mesons and ( b ) pseudoscalar mesons .
the dashed line is a linear regge trajectory with slope 1.16 @xmath108.,width=528 ] = 0.22 gev : ( a ) nucleons and ( b ) @xmath5 states.,width=542 ] current matrix elements in ads / qcd are computed from the overlap of the normalizable modes dual to the incoming and outgoing hadron @xmath109 and @xmath110 and the non - normalizable mode @xmath111 , dual to the external source @xmath112 where @xmath113 is the spin of the interpolating operator @xmath114 , which creates an @xmath11-fock state @xmath115 at the ads boundary .
@xmath111 has the value 1 at zero momentum transfer as the boundary limit of the external current ; thus @xmath116 .
the solution to the ads wave equation subject to boundary conditions at @xmath117 and @xmath118 is @xcite @xmath119 . at large enough @xmath120 ,
the important contribution to ( [ eq : ff ] ) is from the region near @xmath121 . at small @xmath75 ,
the @xmath11-mode @xmath122 scales as @xmath123 , and we recover the power law scaling @xcite , @xmath124^{\tau - 1}$ ] , where the twist @xmath125 , is equal to the number of partons , @xmath126 . a numerical computation for the proton magnetic form factor in the space and time - like regions , for the model described here , gives the predictions shown in fig .
[ fig : protonff ] .
the results correspond to a @xmath127 proton state .
it is interesting to compare the holographic predictions with a model - independent analysis of nucleon form factors using dispersion relations @xcite .
the data are from the compilation given in ref .
the prediction in the domain @xmath128 represents an analytic continuation into the unphysical region .
the results should be modified for values @xmath129 , where the simple form for the cavity current @xmath119 is not valid . ]
the ads / qcd correspondence provides a simple description of hadrons at the amplitude level by mapping string modes to the impact space representation of lfwfs .
it is useful to define the partonic variables @xmath130 , where @xmath131 are the physical position coordinates , @xmath132 are frame - independent internal coordinates , @xmath133 , and @xmath134 is the hadron transverse center of momentum @xmath135 , @xmath136 .
we find for a two - parton lfwf the lorentz - invariant form @xmath137 the @xmath138 are the zeroes of the bessel functions reflecting the dirichlet boundary condition .
the variable @xmath139 @xmath140 , represents the invariant separation between quarks . in the case of a two - parton state
, it gives a direct relation between the scale of the invariant separation between quarks , @xmath141 , and the holographic coordinate in ads space : @xmath142 .
the ground state and first orbital eigenmode are depicted in the figure below . as function of the constituents longitudinal momentum fraction @xmath27 and @xmath143 and the impact space relative
coordinate @xmath144 : ( a ) @xmath127 and ( b ) @xmath145.,title="fig:",width=278 ] as function of the constituents longitudinal momentum fraction @xmath27 and @xmath143 and the impact space relative coordinate @xmath144 : ( a ) @xmath127 and ( b ) @xmath145.,title="fig:",width=278 ] the distribution in @xmath27 and @xmath42 measured in the e791 experiment for diffractive dijet production @xmath146 is consistent with the ads / cft prediction @xcite .
the holographic model is quite successful in describing the known light hadron spectrum and hadronic form factors . since only one parameter , the qcd scale @xmath4 ,
is introduced , the agreement with the pattern of masses of the physical hadronic states and the space and time - like proton form factor data is remarkable . in particular ,
the ratio of @xmath5 to nucleon trajectories is determined by the ratio of zeros of bessel functions .
as we have described , non - zero orbital angular momentum and higher fock - states require the introduction of a casimir operator derived from quantum fluctuations .
it is interesting to note that the predicted mass spectrum @xmath2 at high orbital angular momentum , in contrast to the quadratic dependence @xmath147 found in traditional string theory .
the only mass scale is @xmath4 .
only dimension-@xmath148 and 4 states @xmath149 , @xmath150 , and @xmath151 appear in the duality at the classical level , thus explaining the suppression of @xmath152 odderon exchange .
we have also shown how one can use the extended ads / cft space - time theory to obtain a model for the form of hadron lfwfs .
the model wavefunctions display confinement at large inter - quark separation and conformal symmetry at short distances .
in particular , the scaling and conformal properties of the lfwfs at high relative momenta agree with perturbative qcd @xcite .
these ads / cft model wavefunctions could be used as an initial ansatz for a variational treatment of the light - front qcd hamiltonian .
the dominance of the quark - interchange mechanism in hard exclusive processes also emerges naturally from the classical duality of the holographic model . presented by sjb at the 11th international conference on hadron spectroscopy , hadron05 , 21 - 26 aug 2005 , rio de janeiro , brazil .
this work was supported by the department of energy contract de
ac0276sf00515 .
s. j. brodsky , h. c. pauli and s. s. pinsky , phys .
rept . * 301 * , 299 ( 1998 ) [ arxiv : hep - ph/9705477 ] .
s. j. brodsky , j. r. hiller , d. s. hwang and v. a. karmanov , phys .
d * 69 * , 076001 ( 2004 ) [ arxiv : hep - ph/0311218 ] .
s. j. brodsky and c. r. ji , phys .
d * 33 * , 2653 ( 1986 ) ; j. namyslowski , proceedings of the 9th international conference on few body problems , eugene , oregon , 1980 .
s. j. brodsky , d. s. hwang , b. q. ma and i. schmidt , nucl .
b * 593 * , 311 ( 2001 ) [ arxiv : hep - th/0003082 ] .
s. j. brodsky and d. s. hwang , nucl .
b * 543 * , 239 ( 1999 ) [ arxiv : hep - ph/9806358 ] .
s. j. brodsky , m. diehl and d. s. hwang , nucl .
b * 596 * , 99 ( 2001 ) [ arxiv : hep - ph/0009254 ] .
g. p. lepage and s. j. brodsky , phys .
b * 87 * , 359 ( 1979 ) .
a. v. efremov and a. v. radyushkin , phys .
b * 94 * , 245 ( 1980 ) .
g. p. lepage and s. j. brodsky , phys . rev .
d * 22 * , 2157 ( 1980 ) .
s. j. brodsky and g. p. lepage , slac - pub-2294 , proc . of the workshop on current topics in high energy physics , cal tech . ,
pasadena , calif . ,
( 1979 ) .
s. j. brodsky , p. hoyer , c. peterson and n. sakai , phys .
b * 93 * , 451 ( 1980 ) .
m. franz , m. v. polyakov and k. goeke , phys .
d * 62 * , 074024 ( 2000 ) [ arxiv : hep - ph/0002240 ] .
j. engelfried [ selex collaboration ] , nucl .
a * 752 * , 121 ( 2005 ) .
s. j. brodsky , e. l. berger and g. p. lepage , slac - pub-3027 , proceedings of the workshop on drell - yan processes , batavia , ill .
, ( 1982 ) .
g. bertsch , s. j. brodsky , a. s. goldhaber and j. f. gunion , phys .
lett . * 47 * , 297 ( 1981 ) .
s. j. brodsky and a. h. mueller , phys .
b * 206 * , 685 ( 1988 ) .
s. j. brodsky , s. v. gardner , and d. s. hwang ( in progress ) .
s. j. brodsky , p. hoyer , n. marchal , s. peigne and f. sannino , phys .
d * 65 * , 114025 ( 2002 ) [ arxiv : hep - ph/0104291 ] .
s. j. brodsky , d. s. hwang and i. schmidt , phys .
b * 530 * , 99 ( 2002 ) [ arxiv : hep - ph/0201296 ] .
s. j. brodsky , i. schmidt and j. j. yang , phys .
d * 70 * , 116003 ( 2004 ) [ arxiv : hep - ph/0409279 ] .
s. j. brodsky , f. e. close and j. f. gunion , phys .
d * 5 * , 1384 ( 1972 ) . h. c. pauli and s. j. brodsky , phys .
d * 32 * , 2001 ( 1985 ) .
w. a. bardeen , r. b. pearson and e. rabinovici , phys .
d * 21 * , 1037 ( 1980 ) .
s. dalley , few body syst .
* 36 * , 69 ( 2005 ) [ arxiv : hep - ph/0409139 ] .
m. burkardt and s. dalley , prog . part .
* 48 * , 317 ( 2002 ) [ arxiv : hep - ph/0112007 ] .
l. del debbio , m. di pierro , a. dougall and c. t. sachrajda [ ukqcd collaboration ] , nucl .
suppl . * 83 * , 235 ( 2000 ) [ arxiv : hep - lat/9909147 ] .
maris and c. d. roberts , int .
j. mod .
e * 12 * , 297 ( 2003 ) [ arxiv : nucl - th/0301049 ] .
d. j. gross , a. hashimoto and i. r. klebanov , phys .
d * 57 * , 6420 ( 1998 ) [ arxiv : hep - th/9710240 ] .
k. hornbostel , s. j. brodsky and h. c. pauli , phys .
d * 41 * , 3814 ( 1990 ) .
m. harada , j. r. hiller , s. pinsky and n. salwen , phys .
d * 70 * , 045015 ( 2004 ) [ arxiv : hep - th/0404123 ] .
s. j. brodsky , j. r. hiller and g. mccartor , annals phys .
* 305 * , 266 ( 2003 ) [ arxiv : hep - th/0209028 ] . s. j. brodsky , j. r. hiller and g. mccartor , arxiv : hep - ph/0508295 .
m. weinstein , arxiv : hep - th/0410113 .
h. zhan , a. nogga , b. r. barrett , j. p. vary and p. navratil , phys .
c * 69 * , 034302 ( 2004 ) [ arxiv : nucl - th/0401047 ] .
s. j. brodsky , few body syst .
* 36 * , 35 ( 2005 ) [ arxiv : hep - ph/0411056 ] .
j. m. maldacena , adv .
* 2 * , 231 ( 1998 ) [ int .
j. theor .
phys . *
38 * , 1113 ( 1999 ) ] [ arxiv : hep - th/9711200 ] .
j. polchinski and m. j. strassler , phys .
* 88 * , 031601 ( 2002 ) [ arxiv : hep - th/0109174 ] .
s. j. brodsky and g. r. farrar , phys .
lett . * 31 * , 1153 ( 1973 ) .
v. a. matveev , r. m. muradian and a. n. tavkhelidze , lett .
nuovo cim .
* 7 * , 719 ( 1973 ) .
j. polchinski and l. susskind , arxiv : hep - th/0112204 . j. polchinski and m. j. strassler , jhep * 0305 * , 012 ( 2003 ) [ arxiv : hep - th/0209211 ] .
s. j. brodsky and g. f. de teramond , phys .
b * 582 * , 211 ( 2004 ) [ arxiv : hep - th/0310227 ] .
g. parisi , phys .
b * 39 * , 643 ( 1972 ) .
s. j. brodsky , y. frishman and g. p. lepage , phys .
b * 167 * , 347 ( 1986 ) .
t. banks and a. zaks , nucl .
b * 196 * , 189 ( 1982 ) .
s. j. brodsky , g. p. lepage and p. b. mackenzie , phys .
d * 28 * , 228 ( 1983 ) .
s. j. brodsky , g. t. gabadadze , a. l. kataev and h. j. lu , phys .
b * 372 * , 133 ( 1996 ) [ arxiv : hep - ph/9512367 ] .
r. j. crewther , phys .
* 28 * , 1421 ( 1972 ) .
g. f. de teramond and s. j. brodsky , arxiv : hep - ph/0507273 .
s. j. brodsky , arxiv : hep - ph/0408069 .
g. f. de teramond and s. j. brodsky , arxiv : hep - th/0409074 ; j. erlich , e. katz , d. t. son and m. a. stephanov , arxiv : hep - ph/0501128 ; s. hong , s. yoon and m. j. strassler , arxiv : hep - ph/0501197 ; l. da rold and a. pomarol , nucl .
b * 721 * , 79 ( 2005 ) [ arxiv : hep - ph/0501218 ] ; h. boschi - filho , n. r. f. braga and h. l. carrion , arxiv : hep - th/0507063 .
g. f. de teramond and s. j. brodsky , phys .
lett . * 94 * , 201601 ( 2005 ) [ arxiv : hep - th/0501022 ] .
s. s. gubser , i. r. klebanov and a. m. polyakov , nucl .
b * 636 * , 99 ( 2002 ) [ arxiv : hep - th/0204051 ] .
r. baldini , s. dubnicka , p. gauzzi , s. pacetti , e. pasqualucci and y. srivastava , eur .
j. c * 11 * , 709 ( 1999 ) . | the ads / cft correspondence has led to important insights into the properties of quantum chromodynamics even though qcd is a broken conformal theory .
we have recently shown how a holographic model based on a truncated ads space can be used to obtain the hadronic spectrum of light @xmath0 and @xmath1 bound states .
specific hadrons are identified by the correspondence of string modes with the dimension of the interpolating operator of the hadron s valence fock state , including orbital angular momentum excitations .
the predicted mass spectrum is linear @xmath2 at high orbital angular momentum , in contrast to the quadratic dependence @xmath3 found in the description of spinning strings . since only one parameter , the qcd scale @xmath4 ,
is introduced , the agreement with the pattern of physical states is remarkable . in particular ,
the ratio of @xmath5 to nucleon trajectories is determined by the ratio of zeros of bessel functions .
the light - front quantization of gauge theories in light - cone gauge provides a frame - independent wavefunction representation of relativistic bound states , simple forms for current matrix elements , explicit unitarity , and a trivial vacuum .
the light - front fock - state wavefunctions encode the bound state properties of hadrons in terms of their quark and gluon degrees of freedom at the amplitude level .
one can also use the extended ads / cft space - time theory to obtain a model for hadronic light - front wavefunctions , thus providing a relativistic description of hadrons in qcd at the amplitude level . the model wavefunctions display confinement at large inter - quark separation and conformal symmetry at short distances . in particular , the scaling and conformal properties of the lfwfs at high relative momenta agree with perturbative qcd .
these ads / cft model wavefunctions could be used as an initial ansatz for a variational treatment of the light - front qcd hamiltonian .
we also show how hadron form factors in both the space - like and time - like regions can be predicted .
address = stanford linear accelerator center , stanford university , stanford , california 94309 , usa address = universidad de costa rica , san jos , costa rica |
the diffusion limited aggregation ( dla ) model was introduced in 1981 by t. witten and l. sander @xcite .
the model has been shown to underlie many pattern forming processes including dielectric breakdown @xcite , two - fluid flow @xcite , and electro - chemical deposition @xcite .
the model begins with fixing one particle at the center of coordinates in @xmath3-dimensions , and follows the creation of a cluster by releasing random walkers from infinity , allowing them to walk around until they hit any particle belonging to the cluster . upon hitting they are attached to the growing cluster .
the model was studied on- and off- lattice in several dimensions @xmath4 ; here we are only interested in the off - lattice versions in two dimensions . dla has attracted enormous interest over the years since it is a remarkable example of the spontaneous creation of fractal objects .
it is believed that asymptotically ( when the number of particles @xmath5 ) the dimension @xmath6 of the cluster is very close to 1.71 @xcite , although there exists to date no accepted proof for this fact in spite of several interesting attempts @xcite . in addition , the model has attracted interest since it was among the first @xcite to offer a true multifractal measure : the harmonic measure ( which determines the probability that a random walker from infinity will hit a point at the boundary ) exhibits singularities that are usefully described using the multifractal formalism @xcite .
nevertheless dla still poses more unsolved problems than answers .
it is obvious that a new language is needed in order to allow fresh attempts to explain the growth patterns , the fractal dimension , and the multifractal properties of the harmonic measure .
such a new language was proposed recently by hastings and levitov @xcite , @xcite .
these authors showed that dla in two dimensions can be grown by iterating stochastic conformal maps .
we adopt their basic strategy and will see that it provides a new formulation of the problem which may lead to new insights and results .
the basic idea is to follow the evolution of the conformal mapping @xmath7 which maps the exterior of the unit circle in the mathematical @xmath8plane onto the complement of the cluster of @xmath1 particles in the physical @xmath9plane .
@xmath0 is unique by the riemann mapping theorem , provided that it satisfies the boundary condition @xmath10 here @xmath11 is a real positive coefficient , fixing the argument of @xmath12'$ ] to be zero at infinity .
@xmath7 is related to the complex electric potential @xmath13 by @xmath14 where @xmath15^{-1 } ( z)$ ] is the inverse mapping .
letting @xmath16 in eq.([eq - boundary ] ) it is easy to verify that eq.([eq - potential ] ) implies @xmath17 as it should be at @xmath18 .
the equation of motion for @xmath7 is determined recursively .
the choice of the initial map @xmath19 is rather flexible , and in this paper we select ( arbitrarily ) an initial condition @xmath20 .
we expect the asymptotic cluster to be independent of this choice .
then suppose that @xmath21 is given .
the cluster of @xmath1 `` particles '' is created by adding a new `` particle '' of constant shape and linear scale @xmath22 t o the cluster of @xmath23 `` particles '' at a position which is chosen randomly according to the harmonic measure .
we denote points on the boundary of the cluster by @xmath24 where @xmath25 is an arc - length parametrization . the probability to add a particle on an infinitesimal arc @xmath26 centered at the point @xmath27 on the cluster boundary is @xmath28 the pre - image of @xmath27 and @xmath26 in the @xmath8-plane
are @xmath29 and @xmath30 respectively . clearly
, @xmath31 so the harmonic measure on the real cluster translates to a uniform measure on the unit circle in the mathematical plane .
the image of the cluster of @xmath1 particles under @xmath32 is , by definition , just the unit circle . on the other hand ,
the image of the cluster of @xmath1 particles under @xmath33 is the unit circle with a small bump whose linear scale is @xmath34 where @xmath35 is the image ( under @xmath36 ) of the point @xmath37 on the real cluster at which the growth occurred .
let us define now a new function @xmath38
. this function maps the unit circle to the unit circle with a bump of linear scale @xmath39 around the point @xmath40 . for @xmath41 , @xmath42 ( with positive real proportionality coefficient ) . using @xmath43 the recursion relation for @xmath7
is given by ( see fig . [ fig - mapping ] ) : @xmath44 according to the above discussion @xmath45 is given by @xmath46 so the rhs of eq.([eq - recursion ] ) is determined completely by @xmath21 ; eq.([eq - recursion ] ) induces the recursive dynamics of @xmath7 .
the recursive dynamics can be represented as iterations of the map @xmath43 , @xmath47 this composition appears as a standard iteration of stochastic maps .
this is not so .
the order of iterations is inverted the last point of the trajectory is the inner argument in this iteration . as a result
the transition from @xmath7 to @xmath48 is not achieved by one additional iteration , but by composing the @xmath1 former maps eq.([comp ] ) starting from a different seed which is no longer @xmath49 but @xmath50 .
we note that in the physical plane the `` particles '' are roughly of the same size . to achieve this the linear scales @xmath51 vary widely as a function of @xmath1 and @xmath52 .
we will see that the distribution of @xmath51 and their correlations for different values of @xmath1 determine many of the scaling properties of the resulting cluster .
in particular their moments are related to the generalized dimensions of the harmonic measure .
there are many functions @xmath53 which conformally map the unit circle to the unit circle with a bump .
a simple choice is a function which behaves linearly for large @xmath8 and has a simple pole inside the unit circle which will induce a bump in the image .
the pole has to be at @xmath54 in order to localize the bump near @xmath55 and make it of linear size of the order @xmath56 .
the residue has to be @xmath57 , in order for the bump s height to be also of the order @xmath56 .
consider then @xmath58 careful thinking leads to the conclusion that this function and other similar functions are inappropriate : they have long `` tails '' . in other words ,
the unit circle is slightly distorted everywhere .
this small global distortion may result in a loss of conformality or in the growth of non - constant size particles in the physical plane in numerical applications .
it was proposed in ref .
@xcite that a choice for @xmath43 that is free of global distortion is given by @xmath59 -1 \right \ } ^a \\ & & \phi_{\lambda,\theta } ( w ) = e^{i \theta } \phi_{\lambda,0}(e^{-i \theta } w ) \ , , \label{eq - f}\end{aligned}\ ] ]
the parameter @xmath60 is confined in the range @xmath61 .
as @xmath60 decreases the bump becomes flatter , with the identity map obtained for @xmath62 .
as @xmath60 increases towards unity the bump becomes elongated normally to the unit circle , with a limit of becoming a line ( `` strike '' in the language of @xcite ) when @xmath63 .
naively one might think that the shape of the individual particle is irrelevant for the large scale fractal statistics ; we will see that this is not the case .
the dependence on @xmath60 is important and needs to be taken into account .
notice that this map has two branch points on the unit circle .
the advantage of this is that the bump is strongly localized . on the other hand repeated iterations of this map
leads to rather complex analytic structure .
the aim of this paper is therefore to investigate the scaling and statistical properties of such iterated stochastic conformal maps with a view to discovering the scaling properties induced by the dynamics which any analytic theory must ultimately explain . in section
ii we present the numerical procedure used to generate the fractal clusters , and in section iii give the necessary mathematical background to describe such mappings .
in particular we discuss the laurent expansion of the conformal map from the unit circle to the @xmath1-particle cluster ; the coefficients of the laurent series have interesting scaling behavior with the size of the cluster which is intimately related to the fractal dimension of the cluster and to the generalized dimensions of the harmonic measure . in section
iv we present numerical results regarding the scaling properties of averages of the laurent coefficients and of the size parameter @xmath64 .
the results are accompanied by a theoretical analysis and interpretation . in section
v we conclude with some remarks on the road ahead .
the algorithm simulating the growth of the cluster is based on ref .
@xcite . the @xmath1 `` particle '' cluster
is encoded by the series of pairs @xmath65 .
having the first @xmath66 pairs , the @xmath67 pair is found as follows : choose @xmath68 from a uniform distribution in @xmath69 $ ] , independent of previous history .
then compute @xmath64 from eq.([eq - lambda ] ) , where the derivative of the iterated function @xmath70 involves @xmath71 , @xmath72 , @xmath73 etc , computed respectively at the points @xmath74 , etc .
notice that the evaluation of both @xmath75 and @xmath2 after the addition of one particle involves @xmath76 operations since the seed changes at every @xmath1 .
this translates into @xmath77 time complexity for the growth of an @xmath1-particle cluster .
this is inferior to the best algorithms to grow dla ( using hierarchical maps @xcite , with close to linear efficiency ) , but the present algorithm is not aimed at efficiency . rather , it is used since the laplacian field and the growth probability which is derived from it are readily available at every point of the cluster and away from it . the typical time to grow
a 10,000 particle cluster is 8 minutes on a 300 mhz pentium - ii .
naively one would expect that any choice of @xmath61 would yield dla clusters , since @xmath60 only determines the shape of the particles ( the aspect ratio is @xmath78 for small @xmath79 ) , and the microscopic details of the particles ( except their linear size ) should not affect the global properties .
three typical clusters with particles of various aspect ratios @xmath60 are shown in fig .
[ fig - clusters ] .
we mark in black the interior of the image of the unit circle under the conformal map @xmath7 .
the objects look very much like typical dla clusters grown by standard off - lattice techniques , and in the next section we demonstrate that they have fractal dimensions in close agreement with the latter . for @xmath60 significantly different from @xmath80 , disadvantages of the algorithm get amplified . since the functional form of @xmath2 is fixed ( only the size and position of the `` bump '' change ) , particles of constant shape and size are obtained only if the magnification factor @xmath81 ( the inverse of the field ) is approximately constant in the @xmath8plane around the `` bump '' of @xmath2 . if the particles are elongated along the cluster , then the variation of the field along the cluster affects the shape : large otherwise deeply invaginated regions , where @xmath82 is large , are filled up with a single particle , and the resulting cluster tends to be more compact .
this effect , slightly noticeable even at @xmath83 , is quite significant at the otherwise natural choice of @xmath84 , where the particles are half circles . in fig .
fig - clusters ] we show such a cluster and point out to the area filling dark regions which represent such unwanted events .
the other extreme , when the particles stick out of the cluster , leads to sensitivity to variations in the field going _ away _ from the cluster .
especially if a bump is grown on a tip of a branch , where the field decreases rapidly as one goes away from the tip ( such that @xmath82 increases significantly ) , then the map of the bump gets magnified , resulting in particles of very unequal sizes .
it is necessary to stress that even for @xmath83 , when this procedure appears to yield nice ramified structures , the problem of fill - ups does not go away : in a few rare cases the particle if it happens to land on a place where @xmath85 is large is significantly distorted .
the net effect is that large areas surrounded by the cluster ( where the growth probability is small ) are filled up entirely by one distorted particle . for the value of @xmath84 it appeared that the errors may be unbounded .
our numerics indicates that for @xmath83 the errors were bounded for the cluster sizes that we considered .
we do not have a mathematical proof of boundedness of the errors , and our disregard of this danger is only based on the sensible appearance of our clusters at this value of @xmath60 .
in this section we discuss the laurent expansion of our conformal maps , and introduce the statistical objects that are studied numerically in the next section .
since the functions @xmath7 and @xmath86 are required to be linear in @xmath8 at infinity , they can be expanded in a laurent series in which the highest power is @xmath8 : @xmath87 where @xmath88 the recursion equations for the laurent coefficients of @xmath7 can be obtained by substituting the series of @xmath89 and @xmath2 into the recursion formula ( [ eq - recursion ] ) .
we find @xmath90 we note that the @xmath1-dependence of @xmath91 follows from the dependence on the randomly chosen @xmath68 at the @xmath1th step , from which follows the dependence of @xmath64 on @xmath1 .
the latter is however a function of all the previous growth steps , making the iteration ( [ eq - f1-recursion ] ) -([eq - coeff - recursion ] ) rather difficult to analyze .
a general relation between the laurent coefficients is furnished by the so - called area theorem which applies to univalent mappings . since our maps
solve the laplace equations with boundary conditions only at infinity and on the cluster boundary where the potential is zero , they map the @xmath8 plane uniquely ( and with a unique inverse ) to the @xmath9 plane . in other words ,
the pressure lines and the stream lines are non - degenerate .
such mappings have the property @xcite that the area of the image of the unit disc in the @xmath1th step is given by : @xmath92 a second theorem that will be useful in our thinking is a consequence of the so - called one - fourth theorem , see appendix a. there a statement is proven that the interior of the curve @xmath93 is contained in the @xmath9-plane by a circle of radius @xmath94 .
now as the area @xmath95 is obtained simply from the superposition of @xmath1 bumps of roughly the same area @xmath96 , it has to scale like @xmath97 , for large @xmath1 .
on the other hand any typical radius of the cluster should scale like @xmath98 where @xmath6 is the dimension of the cluster .
we can thus expect a scaling of @xmath99 that goes like @xmath100 we note in passing that this scaling law offers us a very convenient way to measure the fractal dimension of the growing cluster .
indeed , we measured the dimension d for a range of @xmath60 in this way by averaging @xmath99 over 100 clusters .
we found that for a range of @xmath60 spanning the interval @xmath101 $ ] the dimension is constant , around 1.7 .
we can infer from eq.([eq - scalingrad ] ) that the sum in eq .
( [ area ] ) which subtracts positive contributions from @xmath102 contains terms that cancel the behavior of @xmath103 ( remember that @xmath104 ) , leaving a power of unity for the scaling of @xmath95 . indeed , we will show below both numerical and theoretical evidence for the scaling behavior of the @xmath105 for @xmath106 which is in agreement with @xmath103
. we can give a direct physical interpretation for the coefficients @xmath107 by comparing them to the coefficients of the series for @xmath108 , cf .
eq.([eq - potential ] ) : @xmath109 the coefficient of @xmath110 is unity so that the electric flux is unity .
this corresponds to the normalization of the probability .
the constant @xmath111 is the laplace radius which is the radius of a charged disk which would give the same field far away .
the rest of the @xmath112 s are conventional multipole moments .
the relations between the laurent coefficients of @xmath108 and @xmath0 are : @xmath113 the first line shows that @xmath114 , the laplace radius , in accordance with the one - fourth theorem .
the second line shows that the dipole moment @xmath115 is @xmath116 .
we can interpret this coefficient as a distance , the wandering of the center of charge due to the random addition of the particles .
we will take the point of view that this quantity is less `` intrinsic '' than the others to the dynamics of the dla growth .
in fact , if we set @xmath117 , ( we could imagine shifting the cluster as we grow it ) we can rewrite the rest of the equations : @xmath118 etc .
this leads to the interpretation of @xmath119 in terms of the multipole expansion of the electric field .
our growth process is stochastic .
accordingly , it is natural to introduce averages over the randomness . in our thinking
there are two important averages , one over histories of the whole random trajectory @xmath120 , and the other only over the random choice of @xmath68 at the @xmath1th step . to distinguish between the two we denote the first by angular brackets and refer to it as `` history - average '' , while the second is denoted by an overbar and referred to as a `` cluster - average '' .
there is a possibility that for very large clusters ( @xmath5 ) the two averages result in the same numbers .
we will refer to such a property as `` self - averaging '' . the cluster average of moments of @xmath64 offers a relationship to the generalized dimensions of the harmonic measure @xcite .
the latter are defined by dividing the plane into boxes of size @xmath121 , and estimating the probability for a random walker to hit the piece of the boundary of the cluster which is included in the @xmath122th box by @xmath123 where @xmath124 is the modulus of the electric field @xmath125 at some point in the @xmath122th box
. the generalized dimensions are defined by the relation @xmath126 where @xmath127 is the number of boxes of size @xmath121 that are needed to cover the boundary , and @xmath128 is the linear size of the largest possible box , which is of the order of the radius of the cluster . substituting ( [ pi ] ) we find @xmath129 taking @xmath121 very small , of the order of @xmath22 , and assuming that the field is smooth on this scale we have : @xmath130 where @xmath131 is the length of the boundary , @xmath26 is an arc - length differential , and we have used the scaling law @xmath132 .
the connection to our language is obtained by considering the cluster average of powers of @xmath64 .
we grow a cluster of @xmath66 particles , perform repeated random choices of growth sites ( without growing ) , and compute @xmath64 for each choice .
the cluster average can be represented as an integral over the unit circle , @xmath133 , and is given by @xmath134 recalling eq .
( [ eq - lambda ] ) we observe that @xmath135 .
the last relation , eq.([dteta ] ) , and eq.([gener - dim ] ) imply the scaling relation @xmath136
in this section we present results on three topics : + ( i ) the coefficients of the laurent expansion .
the scaling behavior of these quantities is described and discussed in the first subsection .
+ ( ii ) the microscopic fluctuations in the conformal map .
we show that the assumption of self - averaging is valid for eq .
( [ lamdq ] ) and that the multi - fractal exponents are in a good agreement with the known ones .
+ ( iii ) distribution functions of the laurent coefficients .
we analyze numerically the width of those functions and find that @xmath99 tends to a deterministic function of @xmath1 .
we attribute this effect to non - trivial temporal correlations in the field , and give some evidence of their existence .
all the coefficients of the laurent series of @xmath7 are complex numbers except @xmath137 which is real by the choice of zero phase at infinity , see eq.([eq - boundary ] ) .
most of our discussion below pertains to the amplitudes of the coefficients @xmath138 .
we need to stress , however , that the phases are not irrelevant .
if we attempted to use the correct amplitudes with random phases , the resulting series will in general not be conformal .
one of the main results of this paper is that in a addition to the expected scaling behavior of the linear coefficient @xmath99 ( given in eq .
( [ eq - scalingrad ] ) ) the rest of the amplitudes of the laurent coefficients @xmath139 exhibit also a scaling behavior .
we find numerically that in the mean the magnitudes of the laurent coefficients scale as powers of @xmath1 : @xmath140 the exponents @xmath141 are given in fig .
[ fig - coeffs ] .
we first discuss the consequences of the scaling behavior of @xmath99 .
the scaling behavior ( [ eq - scalingrad ] ) has immediate consequences for the scaling behavior of the bump areas @xmath64 .
the connection appears from the recursion eq.([eq - f1-recursion ] ) of @xmath142 which together with @xmath143 reads @xmath144^a \ , .
\label{f1a}\ ] ] taking history averages we find @xmath145^a\rangle \\
\ln\langle f^{(n)}_1 \rangle & \approx & a\sum_{k=1}^{n } \langle \lambda_k \rangle \\ d\ln\langle { f^{(n)}_1}\rangle / dn & \approx & a\langle{\lambda_n}\rangle .
\label{flam}\end{aligned}\ ] ] the last two equations are obtained by expanding the logarithm and keeping only divergent sums : both the mean of @xmath142 and the mean of the sum of @xmath146 increase as a function of @xmath1 .
all other sums of powers of @xmath146 converge as a function of @xmath1 , cf .
subsection b. thus , if we assume that @xmath147 , cf .
eq.([eq - scalingrad ] ) , fractal scaling of the radius ( see below ) implies that @xcite @xmath148 in the next subsection we show that this is indeed supported by the simulations .
note that @xmath149 is inversely proportional to @xmath1 for any value of the fractal dimension @xmath6 . on the other hand , if we assume the property of self - averaging , eq .
( [ meanlam ] ) implies a relationship between the generalized dimension @xmath150 and the fractal dimension @xmath6 . comparing eqs.([lamdq ] ) and
( [ meanlam ] ) leads immediately to the relation @xmath151 this scaling relation was derived by halsey @xcite using much more elaborate considerations .
we see that in the present formalism this scaling relation is obtained very naturally .
in fact the present formulation is more powerful since eq.([meanlam ] ) predicts not only the exponent of the third moment of the electric field , but also the prefactor .
it is also noteworthy that the scaling relation ( [ d3d ] ) results simply from the existence of a power law behavior for the radius @xmath142 .
we found the exponent of @xmath153 to be @xmath154 , see fig .
[ fig - coeff-0 ] . to estimate the scaling behavior of @xmath152
theoretically we note that @xmath155 accordingly we can write @xmath156 in writing the second line we assumed that the main contribution to the correlation function is short ranged , @xmath157 the justification for this is that the field is expected to exhibit wild variations as we trace the boundary @xmath27 .
in addition the main contribution to the integral is expected to come from the support of the harmonic measure where the radius is of the order of @xmath128 .
> from the estimate ( [ f0 ] ) and eq.([gener - dim ] ) we then find @xmath158 in agreement with our measurement of @xmath159 .
( we used here @xmath160 in correspondence with the numerical finding reported in section [ ssec - lambda ] .
any of the values of @xmath161 quoted in the literature would yield @xmath159 in the range @xmath162 . )
the exponents @xmath141 for @xmath163 are smaller than @xmath164 but approach it asymptotically , see fig .
[ fig - coeffs ] .
this behavior is expected from the area theorem , and also from a direct estimate of the integral representation of the coefficients for large @xmath165 @xmath166 in appendix b we show that this integral can be estimated using the multifractal formalism of the harmonic measure with the final result @xmath167 where @xmath168 and @xmath169 are the strength of singularities of the harmonic measure and the dimension of the sets of points that exhibit these singularities respectively @xcite . for our purposes the important consequence of eq.([fkalp ] ) is the scaling relation ( assuming self - averaging ) @xmath170 with @xmath171 .
one knows from the theory of multifractals that @xmath172 , and therefore we can bound @xmath173 from above and from below , @xmath174 .
this is in accord with our numerical simulations in the range @xmath175 , although the calculation in the appendix is only valid for large values of @xmath165 .
we found agreement with eq.([scalefn ] ) with @xmath176 and @xmath177 with @xmath178 .
note that this scaling behavior has important consequences for both the area theorem and for conformality .
absolute convergence of the sum @xmath179 in the area theorem requires @xmath180 which is not the case .
the situation is even more serious for the existence of conformality . to insure the latter
the sum @xmath181 must exist .
this would require @xmath182 .
the reason that the sums exist in the theory is only due to the ultraviolet cutoff at @xmath22 .
this cutoff introduces a highest @xmath165 in the laurent expansion which we estimate as @xmath183 where @xmath131 is the perimeter of the cluster .
here we test eq.([lamdq ] ) . in fig .
[ fig - multifrac ] we display double - logarithmic plots of @xmath184 vs.@xmath1 for @xmath185 and @xmath186 .
the values of the exponents derived from our simulations agree very well ( within the uncertainties ) with the generalized dimensions @xmath187 obtained in the past @xcite for @xmath188 using standard methods .
in addition we reproduce numbers in agreement with the theoretical prediction of @xmath189 and @xmath190 .
this agreement is a strong indication for self averaging at least for the purpose of computing moments of @xmath64 ( i.e. @xmath191 ) .
we previously discussed the scaling behavior of @xmath105 and showed that their history averages obey eq .
( [ scalefn ] ) .
however @xmath139 are random variables with broad scaling distributions . fig .
[ fig - sharpness ] describes the rescaled standard deviation @xmath192 of the laurent coefficients , @xmath193 for @xmath194 as a function of the cluster size @xmath1 .
as is seen clearly from the graphs the widths of the distributions for all @xmath195 tend asymptotically to a finite value .
this is the normal behavior for scaling distributions .
the exceptional case is @xmath196 . even though it exhibits a scaling law of the type ( [ scalefn ] ) ( see section iii ) , with @xmath197 the rescaled distribution width of @xmath198 tends to zero as @xmath1 goes to infinity .
this means that the rescaled distribution function of @xmath142 tends asymptotically to a delta function .
the importance of this result for the evaluation of the fractal dimension of the cluster warrants an immediate discussion of this sharpening phenomenon .
the conclusion of the numerics on @xmath137 is that there exists a universal constant @xmath199 such that @xmath200 where @xmath201 is cluster independent !
moreover , we found that @xmath202 , which is in accordance with the role played by @xmath22 as an ultraviolet inner lengthscale , which is the only lengthscale that appears in the mappings .
note that the constant @xmath203 in eq.([eq - constant ] ) depends on the parameter @xmath60 .
we measured @xmath203 values of 0.6 , 0.87 , 1.2 and 1.8 for @xmath60 values of 1/3 , 1/2 , 2/3 and 4/5 respectively .
the observed sharpening is not obvious since we knowthat @xmath99 is built from a product of random variables @xmath64 , whose moments change with @xmath1 in multi - fractal manner according to eq .
( [ lamdq ] ) . one could attempt to connect the sharpening of @xmath99 to the existence of other sharp functions of @xmath1 .
considering the full expansion of eq.([f1a ] ) we find @xmath204 we could understand eq .
( [ eq - constant ] ) easily if all the sums of all the powers of @xmath205 converged to constants , @xmath206 with @xmath207 cluster independent . in fact ,
this is not the case .
the sums of powers are not cluster independent .
a clear demonstration of this is a simulation that we performed in which the initial condition was very far from a circle .
the individual sums in eq .
( [ eq - expand ] ) were very different from the average values , but nevertheless @xmath208 converged to the right value .
it is our conclusion that each of the sums in ( [ eq - expand ] ) is not cluster independent , and yet somehow the resummed form is cluster independent .
this remarkable sharpening calls for further discussion ; it appears that its interpretation requires better understanding of the time correlations of the field : an independent choice of random realization of a series of @xmath205 according to their multi - fractal distribution can only generate @xmath99 with the proper scaling exponent but can not trivially yield a highly peaked distribution of @xmath99 .
therefore we consider now some evidence for the existence of temporal correlations .
the first outstanding evidence appears in the context of the scaling behavior of @xmath152 , which was discussed in the first subsection .
we show that if we assume that there exist no correlations between different growth stages , the exponent @xmath159 will be very different from the measured and calculated value .
> from the recursion relations of the laurent coefficients ( eq . [ eq - coeff - recursion ] ) we can estimate , in the limit of large @xmath1 when @xmath64 is very small on the average , @xmath209 the second line is obtained because @xmath210 is proportional to the radius of the whole cluster and should not be correlated with @xmath211 .
the crucial approximation comes next : if @xmath211 and @xmath212 can be treated as independent for @xmath213 , then ( since @xmath214 and @xmath215 are independent ) eq.([eq - f0 ] ) simplifies to @xmath216 the numerical simulation resulted in an exponent of the order of 0.7 , in serious disagreement with eq .
( [ eq - diag ] ) .
we think that the assumption of independence , eq .
( [ eq - onlydiag ] ) is the culprit .
another fact which illustrates the importance of the time - angle correlation eq .
( [ eq - onlydiag ] ) is the difference between the exponents of @xmath152 and @xmath217 ( @xmath218 whereas @xmath219 ) differ , for small @xmath64 , by two terms only .
the first one is the term @xmath220 in the rhs of the equation for @xmath217 which is absent in the equation for @xmath152 .
we checked numerically that neglecting this term leads to a very small change in the exponent .
the second difference between is that the term @xmath221 in eq.([eq - f0 ] ) is replaced by @xmath222 .
the change in the exponent can therefore be directly attributed to the existence of important time - angle correlations .
we tried to analyze numerically the time - angle correlations @xmath223 .
the results for some @xmath165 s are shown in fig .
[ fig - anglecorr ] .
it appears that as we increase the size of the ensemble , @xmath224 with the usual @xmath225 dependence on the ensemble size .
if we believe these numerical results ( doubts may exist due to the relative smallness of the ensembles analyzed ) , then the previous results must be related to more subtle correlation of higher order nature .
lastly we would like to discuss the importance of early stages of the growth .
@xmath226 might be written in the following way @xmath227 ( see eq .
( [ eq - f1-recursion ] ) ) . neglecting the correlations in time in the above product
one may approximate @xmath228 numerical evaluation of the two objects in eq.([approx ] ) shows that they differ by a few percent ( see fig .
[ fig - differ ] ) .
the numerics indicate the scaling laws @xmath229 where @xmath230 . to get further intuition we checked also the object @xmath231 for various values of @xmath165 .
the results are shown in fig .
[ fig - differ ] .
as it seems from this graph , time correlations between the initial and late stages of the growth are much more important than local correlations in the late stages .
we checked also two - point time correlations @xmath232 for some @xmath165 s .
the results are plotted in fig .
[ fig - corr ] .
as it turns out from this graph , @xmath233 up to statistical fluctuations .
the language proposed by hastings and levitov appears to offer many appealing features .
it generates dla clusters in such a way that the conformal map @xmath0 from the circle to the boundary of the cluster is known at every instant . in this paper we examined carefully the numerical procedure used to generate the conformal maps , and pointed out the advantages and the shortcoming of the algorithm .
the new results of this paper pertain to the scaling behavior of the laurent coefficients @xmath234 of the conformal map @xmath0 and of the moments of @xmath64 which are related to moments of the field .
we presented a theoretical discussion of the exponents characterizing moments of @xmath234 and @xmath64 .
we pointed out the relations to the multfifractal analysis of the harmonic measure , and derived scaling relations .
of particular interest is the scaling relation @xmath190 that was derived first by halsey and which appears here as a very natural consequence of the formalism .
one important result which is not adequately interpreted in this paper is the sharpness of the distribution of @xmath137 .
this coefficient is proportional to the radius of the cluster , and its sharpness is directly related to the existence of a universal fractal dimension independently of the details of the shape of the cluster . understanding the sharpness appears to be connected to understanding the existence of universal fractal dimension , and we believe that it poses a very worthwhile and focussed question for the immediate future .
this work has been supported in part by the israel science foundation founded by the israel academy of sciences and humanities .
lms and es are supported by doe grant defg-02 - 95er-45546 , and would like to thank the weizmann institute of science for hospitality .
in this appendix we prove that every univalent function of the type ( [ eq - laurent - f ] ) is bounded in a circle of radius @xmath235 .
this fact is based on two basic properties of univalent functions
@xcite : 1 .
there is one - to - one correspondence between univalent functions of the form @xmath236 ( @xmath237-class ) and univalent functions of the form @xmath238 ( @xmath239-class ) .
this correspondence is given by @xmath240 2 .
the koebe one - quarter theorem . _
the image of the unit disc under every function of class s contains the disc _ @xmath241 .
consider a function @xmath242 of the form ( [ eq - laurent - f ] ) .
this is a @xmath239-class function with linear coefficient @xmath137 .
let us denote its conjugate ( by eq .
( [ correspondence ] ) ) @xmath237-class function as @xmath243
. the linear coefficient of @xmath244 is @xmath245 .
consider now the smallest circle in the @xmath246plane which bounds the image of the unit circle under @xmath89 , @xmath247 . from eq .
[ correspondence ] it is clear that the circle @xmath248 is the largest circle which is contained in the image of the unit disc under @xmath244 .
thus the koebe one - quarter theorem ensures that @xmath249 , which implies @xmath250 .
to estimate the large @xmath165 and large @xmath1 dependence of @xmath252 , the components are first written as integrals over the boundary of the cluster @xmath253 where @xmath254 for the purposes of sect .
4 we are interested in @xmath255 : @xmath256 } |e(s)| ( or equivalently , of @xmath257 ) , an examination of eq ( [ estim ] ) shows that for large @xmath165 the fluctuations in the values of the integrands depend more crucially on the phase variations than on the field and radius variations .
the phase varies appreciably when @xmath52 changes an amount @xmath258 and therefore it is useful to split up the integral eq.([a1 ] ) into a sum of essentially independent contributions coming from the electric field singularities with exponents @xmath168 .
this exponent is determined by the scaling law relating the measure ( which is proportional to @xmath259 ) of a box to its size @xmath260 : @xmath261 @xcite .
the integral is split into contributions made of contour sections of different lengths @xmath262 dependent on the singularity but each giving rise to the same change @xmath259 .
if one can estimate both the magnitude of the contribution of a specific multifractal electric field singularity @xmath168 to the integral and the number of such contributions @xmath263 @xcite , then one can write @xmath264 where @xmath265 to estimate @xmath266 we recall that by definition @xmath267 > from eq.([a3 ] ) @xmath268 using eq.([a6 ] ) and the fact that the fractal dimension of singularities of size @xmath168 is @xmath269 , we can now also estimate the number of singularities of size @xmath168 which contribute to the integral as @xmath270 to estimate @xmath271 we note that the major contribution to eq .
( [ ialph ] ) comes from the support of the harmonic measure where @xmath272 .
accordingly @xmath273 where we have made use of eqs.([a2 ] ) , ( [ a3 ] ) . combining eq.([a4 ] ) with the estimates ( [ a7 ] ) and ( [ a9 ] ) then yields @xmath274 we note that the approximation adopted in this appendix differs from the delta - function assumption ( [ delfunc ] ) in asserting that for high values of @xmath165 the variation of the phase dominates the decay of the integrand compared to the rapid decorrelation of the field .
one would guess that for @xmath165 of the order of unity the field decorrelates faster due to the rapid variation over the arc length . for high values of @xmath165
the phase decorrelation is strongly amplified and we adopt the assumption used here . | the creation of fractal clusters by diffusion limited aggregation ( dla ) is studied by using iterated stochastic conformal maps following the method proposed recently by hastings and levitov .
the object of interest is the function @xmath0 which conformally maps the exterior of the unit circle to the exterior of an @xmath1-particle dla . the map @xmath0 is obtained from @xmath1 stochastic iterations of a function @xmath2 that maps the unit circle to the unit circle with a bump .
the scaling properties usually studied in the literature on dla appear in a new light using this language .
the dimension of the cluster is determined by the linear coefficient in the laurent expansion of @xmath0 , which asymptotically becomes a deterministic function of @xmath1 .
we find new relationships between the generalized dimensions of the harmonic measure and the scaling behavior of the laurent coefficients . 2 |
the hill problem arises as an approximation for nearly circular trajectories to newton s gravitational equations written in rotating coordinates for bodies moving around a central mass .
the original example is provided by the moon - earth - sun system @xcite .
later , hill s equations have also been applied to stellar dynamics @xcite , with a `` star cluster '' replacing moon and earth , and the `` galactic center '' playing the role of the sun . for the center - of - mass @xmath0 where @xmath1 the inter - particle interactions cancel , leaving us , in the planar case , with the equations @xcite , llll & -2 - 3 ^ 2x&=&0 , + & + 2&=&0 , [ hillsymmeqs ] where @xmath2 is the angular velocity @xmath3 for a circular keplerian trajectory with radius @xmath4 .
note here the `` only-@xmath5 '' anisotropic oscillator term , which is the remnant of the centrifugal and newtonian forces under linear approximation .
the general solution of eqn .
( [ hillsymmeqs ] ) is a combination of simply `` hall '' motions of a guiding center , combined with elliptic motions around the guiding center @xcite .
the decomposability into center - of - mass and relative coordinates relies on having an abelian symmetry made of ( generalized ) translations and boosts , characteristic for galilean symmetry @xcite .
these symmetries are also given by the solutions of eqn .
( [ hillsymmeqs ] ) ; the conserved quantities associated with them span two copies of heisenberg algebras with central parameters @xmath6 @xcite . the velocity - dependent terms in eqn .
( [ hillsymmeqs ] ) come from the coriolis force , and are analogous to a uniform magnetic field . whence the similarity to the landau problem and to the hall effect . on the other hand , some insight to the hall effect
could be gained by assuming that the charged particles admit an additional `` exotic '' structure , which makes the coordinates @xmath5 and @xmath7 non - commuting @xcite . in this note
we generalize hill s equations to exotic particles , providing us with a combination of the hill results in @xcite with the hall - type behavior found before in the non - commutative landau problem @xcite .
our main result is that , for the critical value @xmath8 of the rotation i.e. of the radius , the only allowed motions are those which satisfy the hall law .
our basic tool is the generalization to the exotic hill case of the chiral decomposition of ref .
the planar hill equations ( [ hillsymmeqs ] ) derive from the symplectic form and hamiltonian @xcite lll _ 0&=&dp^xdx+dp^ydy + 2 m dxdy , + h&=&+kx^2 , k=-3m^2 .
[ 0sympham ] for an exotic particle , the symplectic form is modified as @xmath9 while the hamiltonian is unchanged @xcite .
the constant @xmath10 is the _ non - commutative parameter _
, as justified by the associated poisson brackets @xmath11 where = 1 - 2m[delta ] is the [ square - root of the ] determinant of the `` exotic '' symplectic form ( [ exosymp ] ) .
hamilton s equations could be worked out and the solution found .
we find it , however , more convenient to deduce them in a smarter way , presented in the next section .
following @xcite , we introduce chiral coordinates @xmath12 , p^1=_+x_+^2+_-x_-^2 , p^2=-_+x_+^1-_-x_-^1 , where the coefficients @xmath13 and @xmath14 are determined from the requirement that both the symplectic form and the hamiltonian should split
. then the calculation analogous to the one in @xcite yields & = & - dx_+^1dx_+^2 + dx_-^1dx_-^2 , + h&= & ( x_+^1x_+^1 + x_+^2x_+^2 ) - x_-^1x_-^1,[decomposedhill ] where = 1-/2 m .
for @xmath15 we have @xmath16 , and the commutative hill case @xcite is recovered . for @xmath17 the oscillator term is switched off , and the system reduces to the purely magnetic non - commutative landau problem discussed in @xcite .
the decomposition ( [ decomposedhill ] ) implies the poisson brackets @xmath18 completed with @xmath19 , providing us with separated equations of motion , _
-^1&=&0 , & & _ -^2&=&-x_-^1 , + _ + ^1&= & x_+^2 , & & _
+ ^2&=&-2 x_+^1 .
[ x+-eq ] off the critical case , @xmath20 , the solution is therefore x_-^1(t)&=&x_0,x_-^2(t)=-x_0t+y_0 , [ x - hall ] + x_+^1(t)&=&^*t-^*t , ^*= , + x_+^2(t)&=&(2^*t+2^*t ) , @xmath21 performs , hence , for all @xmath20 and @xmath22 , a simple uniform translational motion , and can be identified with the _ guiding center coordinate_. @xmath23 , which moves along a flattened ellipses , describes instead motions about the guiding center . note that the @xmath24 dynamics depends on the non - commutative parameter @xmath10 through @xmath25 , while the guiding center dynamics is @xmath10-independent . in terms of the original coordinates ,
the motion is x(t)&=&^*t -^*t + x_0 , + y(t)&=&(2^*t+2^*t ) -x_0t+y_0.[gensol ] see fig .
[ k - figs ] .
is approached , `` non - hall '' trajectories speed up , while the guiding center ( in red ) moves following the hall law .
the latter are the only allowed motions when @xmath26 . after crossing the critical value ,
the sense of rotation is reversed .
_ , title="fig : " ] + is approached , `` non - hall '' trajectories speed up , while the guiding center ( in red ) moves following the hall law .
the latter are the only allowed motions when @xmath26 . after crossing the critical value ,
the sense of rotation is reversed .
_ , title="fig : " ] + is approached , `` non - hall '' trajectories speed up , while the guiding center ( in red ) moves following the hall law .
the latter are the only allowed motions when @xmath26 . after crossing the critical value ,
the sense of rotation is reversed .
_ , title="fig : " ] the particular case @xmath27 is rather harmless : it simply switches off the oscillations of the second coordinate of @xmath24 , whereas its first coordinate still oscillates , namely with frequency @xmath28 .
intuitively , we have sorts of `` half - hall motions '' , see fig .
[ k4-figs ] .
the oscillations of the @xmath7 coordinate disappear but @xmath29 still oscillates around the hall trajectory of the guiding center.__,title="fig : " ] + in the critical case @xmath30 i.e. for angular velocity and radius = _ c= , r_c^3=4gmm^2 ^ 2 , [ cr ] respectively , the system is singular , and ( [ x+-eq ] ) only allows @xmath31 thus the elliptic motions about the guiding center are eliminated , leaving us with the guiding center motion alone , @xmath32 in ( [ x - hall ] ) .
remarkably , the latter _ satisfies the hall law _ : the center - of - mass moves in the planar `` magnetic '' field @xmath33 perpendicularly to the `` electric '' field @xmath34 with appropriate `` hall '' velocity @xmath35 . as seen on fig .
[ k - figs ] , when the critical value is approached , the rotation speeds more and more up , and , _ with the exception of those whose initial conditions are consistent with the hall law _ , all motions become `` instantaneous '' @xcite .
the same equations ( [ hillsymmeqs ] ) also describe the symmetries .
having decomposed our system into chiral components , the symmetry plainly follows from those of our chiral solutions .
for the separated equations ( [ x+-eq ] ) the initial conditions @xmath36 and @xmath37 , p&= & x_0 + y_0= x_-^1(t ) + x_-^2(t)+tx_-^1(t ) , + k&= & -b/^ * + 2a/^ * = + & & x_+^1(t)^ * t- /2x_+^2(t)^ * t + 2x_+^1(t)^ * t + x_+^2(t)^ * t , are plainly constants of the motion .
they are interpreted as translations and generalized boosts , respectively . for their lie algebra structure
we find , off the critical case @xmath20 , the commutation relations \{p^1
, p^2}=- , [ x - halg ] \{k^1 , k^2}= , [ x+halg ] supplemented with @xmath38 .
we have therefore _ two uncoupled heisenberg algebras with different central charges @xmath39 and @xmath40 , respectively_. adding the obvious time - translation - symmetry would provide us with exotic newton - hooke symmetry without rotations cf .
@xcite .
in the critical case @xmath26 the @xmath24 dynamics becomes trivial .
@xmath41 drops out , and the symmetry algebra reduces to the single heisenberg algebra of @xmath42 alone [ plus time translations ] , with extension parameter determined by the non - commutative parameter , @xmath43 .
guided by the analogy with the non - commutative landau problem @xcite , we extended our previous study of hill s equation to exotic particles .
our most interesting result says that for a critical angular velocity i.e. for a critical radius determined by the non - commutative parameter @xmath10 , cf .
( [ cr ] ) , the only motions are those determined by the hall law .
the role of @xmath10 is to enhance the `` hall - type '' behavior , eliminating all the others in the critical case @xmath26 .
we note that hall motions in stellar dynamics have been considered before @xcite . except for the lack of rotational symmetry due to the anisotropic oscillator term in ( [ hillsymmeqs ] ) , our results are reminiscent to those for the non - commutative landau problem @xcite , and generalize those derived in @xcite for @xmath15 .
this paper grew out from joint work with gary gibbons @xcite , to whom we are indebted for his interest and advice .
p.a.h acknowledges hospitality at the _ institute of modern physics _ of the lanzhou branch of the chinese academy of sciences .
this work was partially supported by the national natural science foundation of china ( grant no .
11035006 and 11175215 ) and by the chinese academy of sciences visiting professorship for senior international scientists ( grant no .
2010tij06 ) .
b. bok `` the stability of moving clusters , '' harvard college observatory circular , * 384 * ( 1934 ) , 1 ; h. mineur , `` quilibre des nuages galactiques et des amas ouverts dans la voie lacte .
volution des amas , '' annales dastrophysique * 2 * ( 1939 ) 1 .
t. fukushige and d. c. heggie , `` the time scale of escape from star clusters , '' mon . not .
astron . soc . *
318 * ( 2000 ) 753 [ arxiv : astro - ph/9910468 ] ; d. heggie , `` escape in hill s problem , '' in : b.a .
steves , a.j .
maciejewski ( eds . ) , _ the restless universe _ , scottish universities summer school in physics and institute of physics publishing , bristol , 2001 , p. 109 .
[ arxiv : astro - ph/0011294 ] .
p. m. zhang , g. w. gibbons and p. a. horvathy , `` kohn s theorem and newton - hooke symmetry for hill s equations , '' phys . rev .
* d85 * ( 2012 ) 045031 , [ arxiv:1112.4793 [ hep - th ] ] .
souriau , _ structure des systmes dynamiques _ ,
dunod paris ( 1970 ) .
_ structure of dynamical systems . a symplectic view of physics _ , birkhuser , boston ( 1997 ) .
in the newton - hooke context , see g. w. gibbons and c. n. pope , `` kohn s theorem , larmor s equivalence principle and the newton - hooke group , '' ann .
* 326 * , 1760 ( 2011 ) .
[ arxiv:1010.2455 [ hep - th ] ] ; p. m. zhang and p. a. horvathy , `` kohn s theorem and galilean symmetry , '' phys . lett . *
b702 * , 177 ( 2011 ) .
p. m. zhang , p. a. horvathy , `` kohn condition and exotic newton - hooke symmetry in the non - commutative landau problem , '' phys .
b 706 * , 442 ( 2012 ) [ arxiv:1111.1595 [ hep - th ] ] ; k. andrzejewski , j. gonera and p. kosinski , `` nonlinear realizations , the orbit method and kohn s theorem , '' [ arxiv:1203.3311 [ hep - th ] ] .
c. duval and p. a. horvthy , `` the `` peierls substitution '' and the exotic galilei group . '' phys . lett .
* b 479 * ( 2000 ) 284 [ ` hep - th/0002233 ` ] ; `` exotic galilean symmetry in the non - commutative plane , and the hall effect . '' journ . phys .
* a34 * ( 2001 ) 10097 [ ` hep - th/0106089 ` ] ; p. a. horvthy , `` the non - commutative landau problem , '' ann . phys .
( n. y. ) * 299 * ( 2002 ) 128 [ ` hep - th/0201007 ` ] ; p. a. horvthy and m. s. plyushchay , `` nonrelativistic anyons in external electromagnetic field , '' nucl . phys .
* b 714 * 269 ( 2005 ) [ ` hep - th/0502040 ` ] , etc . p. d. alvarez , j. gomis , k. kamimura , m. s. plyushchay , `` anisotropic harmonic oscillator , non - commutative landau problem and exotic newton - hooke symmetry , '' phys .
* b 659 * , 906 ( 2008 ) [ arxiv:0711.2644 ] ; `` ( 2 + 1)d exotic newton - hooke symmetry , duality and projective phase , '' annals phys .
* 322 * ( 2007 ) 1556 [ hep - th/0702014 ] .
j. saavedra , r. troncoso and j. zanelli , `` degenerate dynamical systems , '' j. math . phys .
* 42 * ( 2001 ) 4383 [ hep - th/0011231 ] ; f. de micheli and j. zanelli , `` quantum degenerate systems , '' arxiv:1203.0022 [ hep - th ] . | our previous study of a system of bodies assumed to move along almost circular orbits around a central mass , approximately described by hill s equations , is extended to `` exotic '' [ alias non - commutative ] particles . for
a certain critical value of the angular velocity , the only allowed motions follow the hall law .
translations and generalized boosts span two independent heisenberg algebras with different central parameters . in the critical case
, the symmetry reduces to a single heisenberg algebra . |
systematic measurements of the centrality dependence of transverse energy production and charged particle multiplicity at midrapidity provide excellent characterization of the nuclear geometry of the reaction and are sensitive to the dynamics of the colliding system .
for example , measurements of and in collisions at = 200 gev and 130 gev as a function of centrality expressed as the number of participant nucleons , , exhibit a nonlinear increase with increasing .
this has been explained by a two - component model proportional to a linear combination of the number of collisions , , and @xcite . in a previous study by the phenix collaboration , measurements of and for collisions at 200 , 130 , and 62.4
gev are presented along with comparisons to the results of several models @xcite .
the models that were examined included hijing @xcite , a final state parton saturation model called ekrt @xcite , an initial state parton saturation model called kln @xcite , and a multiphase transport model called ampt @xcite .
the comparisons showed that most models could reproduce some of the features of the data , but most failed in describing all of the data with the hijing and ampt models best describing the overall trends , including the nonlinear increase of and as a function of .
it was also proposed that is linearly proportional to the number of constituent - quark participants without a significant contribution from a hard scattering component @xcite .
recently , the phenix collaboration at brookhaven national laboratory s relativistic heavy ion collider ( rhic ) , presented distributions at midrapidity for collisions at = 200 , 130 , and 62.4 gev , collisions at = 200 gev , and collisions at = 200 gev @xcite .
the data are better described by a model based upon the number of constituent quark participants than by the wounded nucleon model @xcite . here , this study will be extended to include both and measurements at midrapidity in collisions down to = 7.7 gev .
this study will also examine the centrality dependence of and for smaller systems including , , , and .
recent lattice quantum chromodynamics ( qcd ) calculations indicate that the transition from quark to hadronic matter is a crossover transition at high temperature and small baryochemical potential , @xmath13 @xcite . at high values of @xmath13 and low temperatures , model calculations
indicate the presence of a first - order phase transition and the possibility of a critical end point in the qcd phase diagram @xcite .
relativistic heavy ion collisions serve as excellent probes of the qcd phase diagram @xcite .
the region of the qcd phase diagram sampled by the collisions can be controlled by changing the beam energy .
lowering the beam energy corresponds to raising the value of @xmath13 . from 2010 to 2014 , rhic executed a beam energy scan program to explore the qcd phase diagram , look for evidence of the phase boundaries and search for evidence of the critical end point .
presented here are and measurements from the beam energy scan as a function of centrality expressed as the number of nucleon participants , , from collisions at = 200 , 130 , 62.4 , 39 , 27 , 19.6 , 14.5 , and 7.7 gev . over the past 15 years
, phenix has collected a comprehensive dataset covering a wide variety of colliding nuclei and collision energies , including the collision beam energy scan mentioned above . presented here
will be charged particle multiplicity and transverse energy measurements from the following systems : collisions at = 200 , 130 , 62.4 , 39 , 27 , 19.6 , 14.5 , and 7.7 gev ; collisions at = 200 and 62.4 gev ; collisions at = 200 gev ; collisions at = 193 gev ; collisions at = 200 gev ; collisions at = 200 gev ; and collisions at = 200 gev .
the results will be discussed in the context of scaling with the number of participant nucleons ( ) and the number of participant quarks ( ) .
phenix has previously published charged particle multiplicity distributions from collisions at = 200 gev @xcite , collisions at = 130 gev @xcite , and collisions at = 19.6 gev @xcite .
phenix has also previously published transverse energy distributions from collisions at = 200 gev @xcite , collisions at = 130 gev @xcite , collisions at = 62.4 gev @xcite , collisions at = 19.6 gev @xcite , and minimum - bias distributions for and collisions at = 200 gev @xcite . here
the previously published phenix results will be presented along with data from the many new collision systems in a consistent format to facilitate comparisons .
similar measurements have been published by the other rhic experiments .
charged particle multiplicity distributions have been published by brahms for collisions at = 200 and 130 gev @xcite , star for collisions at = 130 gev @xcite , and phobos for collisions at = 200 , 130 , 62.4 , 56 , and 19.6 gev along with collisions at = 200 and 62.4 gev , collisions at = 200 gev , and collisions at = 410 and 200 gev @xcite .
transverse energy distributions have been published by star for collisions at = 200 gev @xcite .
presented here are many collision systems and energies that have not been previously published by phenix or the other rhic experiments , especially for the transverse energy measurements .
the first complete results on charged particle multiplicity and transverse energy from the rhic beam energy scan program conducted from 2010 to 2014 are also included .
this paper is organized as follows .
the phenix detector and the methods used for centrality determination in each dataset will be described in section [ sec : detector ] .
the analysis of the data to measure and including a description of estimates of the systematic uncertainties is described in section [ sec : analysis ] .
the centrality dependent results at midrapidity from the beam energy scan in terms of are presented in section [ sec : besresults ] .
a description of the centrality dependent results at midrapidity for and collisions in terms of are found in section [ sec : curesults ] .
a description of the centrality dependent results at midrapidity for collisions in terms of are found in section [ sec : uresults ] .
section [ sec : smallresults ] contains a description of the centrality dependent results at midrapidity of and collisions in terms of .
a review all of the centrality dependent results in terms of is presented in section [ sec : quarks ] .
section [ sec : summary ] contains a summary of the results .
data tables for all data sets are tabulated in the appendix .
the phenix detector comprises two central spectrometer arms , two muon spectrometer arms , and a set of forward detectors .
all of the detector components and their performance is described elsewhere @xcite .
the analysis of charged particle multiplicity utilizes detectors in the central arm spectrometer @xcite including the drift chamber ( dc ) and pad chamber 1 ( pc1 ) detectors .
the drift chambers are cylindrically shaped and located radially from 2.0 to 2.4 meters .
the dc covers the pseudorapidity region @xmath140.35 and @xmath15 in azimuth for each arm .
the dc has a resolution better than 150 @xmath16 m in r-@xmath17 , better than 2 mm in the z direction , and a two track separation better than 1.5 mm .
the pc1 detector is a multiwire proportional chamber mounted on the outer radius of the drift chamber at 2.5 m from the beam axis .
pc1 covers the full central arm acceptance .
pc1 measures minimum ionizing particles with an efficiency greater than 99.5% with a position resolution of 1.7 mm by 3 mm and a two track separation of 4 cm .
reconstructed tracks from the drift chamber with an associated hit from pc1 are counted as charged particle tracks in the multiplicity measurement .
the analysis of transverse energy utilizes five of the lead - scintillator ( pbsc ) electromagnetic - calorimeter ( emcal ) sectors in the central arm spectrometers @xcite .
each calorimeter sector covers a pseudorapidity range of @xmath140.38 and subtends @xmath18 in azimuth for a total azimuthal coverage of @xmath19 .
the front face of each sector is located 5.1 m from the beam axis .
each sector contains 2592 pbsc towers arranged in a 36 @xmath20 72 array .
each tower has a 5.535 @xmath20 5.535 cm surface area and a thickness of 0.85 nuclear interaction lengths or 18 radiation lengths .
the pbsc emcal energy resolution has been measured using test beam electrons to be @xmath21 2.1% , with a measured response proportional to the incident electron energy to within @xmath22 over the range @xmath23 gev . for all data sets ,
a minimum - bias trigger is provided by a pair of beam - beam counters ( bbc ) @xcite .
each bbc comprises 64 individual erenkov counters .
each bbc covers 2@xmath24 azimuthally and a pseudorapidity range of 3.0 @xmath25 3.9 . for , , and collisions ,
an event is required to have at least one counter fire in each bbc . for all other collisions
, at least two counters must fire in each bbc .
the event vertex is reconstructed with a resolution of 2.0 cm in collisions and 0.5 mm in central collisions using the timing information from the bbcs .
all events are required to have an event vertex within 20 cm of the center of the detector .
centrality determination in the original = 200 gev and = 130 gev phenix analysis is based upon the total charge deposited in the bbcs and the total energy deposited in the zero - degree calorimeters ( zdc ) @xcite .
the zdcs are a pair of hadronic calorimeters that cover the pseudorapidity range @xmath266 . for subsequent data sets taken after 2002
, only the bbc information is used for the centrality determination , including the following data sets : at = 200 gev , at = 200 gev , at = 193 gev , at = 200 gev , and at = 200 gev . as the collision energy decreases ,
the width of the pseudorapidity distribution of produced particles becomes more narrow @xcite . as a result , for energies below = 130
gev , the acceptance of the zdc is reduced , therefore only the bbc information is used for collisions at = 62.4 and 39 gev , and for collisions at = 62.4 gev . below = 39 gev ,
the bbc acceptance becomes sensitive to the presence of beam fragments , which affects the linear response of the bbc to the centrality . to avoid this nonlinear response
, the reaction - plane detector ( rxnp ) @xcite is used for the centrality determination for collisions at = 7.7 gev , which was taken during the 2010 running period .
the rxnp comprises two sets of plastic scintillators positioned on either side of the collision vertex .
each rxnp detector is arranged in 12 azimuthal segments separated into an inner and outer ring .
the rxnp has an azimuthal coverage of 2@xmath24 .
the pseudorapidity coverage is [email protected] and [email protected] for the inner and outer ring , respectively .
a 2 cm thick lead converter is located directly in front of the rxnp scintillators with respect to the collision region , which allows the rxnp to also measure contributions from neutral particles through conversion electrons .
the rxnp is designed to measure the reaction plane angle , but it can also function well as a centrality detector , because the magnitude of the total charge measured by the rxnp is dependent on the centrality of the collision . in order to minimize contamination from beam fragments , only the outer ring of the rxnp
is used for centrality determination for collisions at = 7.7 gev . for the 2011 data taking period and
later when the data sets at = 27 , 19.6 , and 14.5 gev were collected , the rxnp was removed in order to install a silicon vertex detector , which was being commissioned during this time .
so , for these two data sets , the multiplicity of hits in the pc1 detector were used to determine the centrality .
a summary of the centrality detectors used for each dataset is included in table [ tab : datasetsummary ] .
table [ tab : datasetsummary ] provides a summary of the data sets used in this analysis . for collisions at = 62.4
gev , the analysis uses data taken in 2004 @xcite and the analysis uses data taken in 2010 . the number of events are those events that pass the minimum - bias trigger condition for the dataset and have an event vertex within 20 cm of the center of the detector .
51ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ] + 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1103/physrevlett.86.3496 [ * * , ( ) ] link:\doibase 10.1016/s0370 - 2693(01)00457 - 9 [ * * , ( ) ] link:\doibase 10.1103/physrevc.71.049901 , 10.1103/physrevc.71.034908 [ * * , ( ) ] , link:\doibase 10.1103/physrevd.44.3501 [ * * , ( ) ] link:\doibase 10.1016/s0550 - 3213(99)00720 - 8 [ * * , ( ) ] link:\doibase 10.1103/physrevc.64.011902 [ * * , ( ) ] link:\doibase 10.1103/physrevc.67.064905 [ * * , ( ) ] link:\doibase 10.1103/physrevc.89.044905 [ * * , ( ) ] link:\doibase 10.1016/0550 - 3213(76)90329 - 1 [ * * , ( ) ] link:\doibase 10.1038/nature05120 [ * * , ( ) ] link:\doibase 10.1103/physrevd.78.074507 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.81.4816 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.86.3500 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.87.052301 [ * * , ( ) ] link:\doibase 10.1016/j.nuclphysa.2005.02.130 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.87.112303 [ * * , ( ) ] link:\doibase 10.1103/physrevc.83.024913 [ * * , ( ) ] link:\doibase 10.1103/physrevc.70.054907 [ * * , ( ) ] link:\doibase 10.1016/s0168 - 9002(02)01950 - 2 [ * * , ( ) ] link:\doibase 10.1016/s0168 - 9002(02)01952 - 6 [ * * , ( ) ] link:\doibase 10.1016/s0168 - 9002(02)01954-x [ * * , ( ) ] link:\doibase 10.1016/s0168 - 9002(02)01956 - 3 [ * * , ( ) ] link:\doibase 10.1103/physrevc.70.021902 [ * * , ( ) ] link:\doibase 10.1016/j.nima.2011.01.034 [ * * , ( ) ] link:\doibase 10.1088/0954 - 3899/25/9/308 [ * * , ( ) ] link:\doibase 10.1146/annurev.nucl.57.090506.123020 [ * * , ( ) ] link:\doibase 10.1016/s0375 - 9474(96)00388 - 0 [ * * , ( ) ] link:\doibase 10.1103/physrevc.59.2173 [ * * , ( ) ] link:\doibase 10.1016/s0375 - 9474(02)01424 - 0 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.109.152303 [ * * , ( ) ] link:\doibase 10.1103/physrevc.58.3523 [ * * , ( ) ] link:\doibase 10.1016/s0370 - 2693(00)00916 - 3 [ * * , ( ) ] link:\doibase 10.1103/physrevc.81.024911 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.106.032301 [ * * , ( ) ] link:\doibase 10.1016/j.physletb.2012.02.045 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.34.836 [ * * , ( ) ] link:\doibase 10.1103/physrevd.27.140 [ * * , ( ) ] link:\doibase 10.1007/3 - 540 - 45792 - 5_6 [ * * , ( ) ] , @noop `` , '' link:\doibase 10.1140/epjc / s10052 - 006 - 0128-z [ * * , ( ) ] link:\doibase 10.1103/physrevlett.94.132301 [ * * , ( ) ] link:\doibase 10.1103/physrevc.83.021902 [ * * , ( ) ] link:\doibase 10.1103/physrevc.85.034905 [ * * , ( ) ] link:\doibase 10.1103/physrevc.89.064908 [ * * , ( ) ] link:\doibase 10.1103/physrevc.72.037901 [ * * , ( ) ] link:\doibase 10.1016/j.physletb.2009.08.025 [ * * , ( ) ] link:\doibase 10.1016/j.physletb.2015.07.078 [ * * , ( ) ] link:\doibase 10.1103/physrevc.90.034902 [ * * , ( ) ] @noop ( ) link:\doibase 10.1103/revmodphys.28.214 [ * * , ( ) ] link:\doibase 10.1016/s0370 - 2693(96)01362 - 7 [ * * , ( ) ] | measurements of midrapidity charged particle multiplicity distributions , @xmath0 , and midrapidity transverse - energy distributions , @xmath1 , are presented for a variety of collision systems and energies . included
are distributions for au@xmath2au collisions at @xmath3 , 130 , 62.4 , 39 , 27 , 19.6 , 14.5 , and 7.7 gev , cu@xmath2cu collisions at @xmath3 and 62.4 gev , cu@xmath2au collisions at @xmath3 gev , u@xmath2u collisions at @xmath4 gev , @xmath5@xmath2au collisions at @xmath3 gev , @xmath6he@xmath2au collisions at @xmath3 gev , and @xmath7@xmath2@xmath7 collisions at @xmath3 gev .
centrality - dependent distributions at midrapidity are presented in terms of the number of nucleon participants , @xmath8 , and the number of constituent quark participants , @xmath9 .
for all @xmath10@xmath2@xmath10 collisions down to @xmath11 gev , it is observed that the midrapidity data are better described by scaling with @xmath9 than scaling with @xmath8 .
also presented are estimates of the bjorken energy density , @xmath12 , and the ratio of @xmath1 to @xmath0 , the latter of which is seen to be constant as a function of centrality for all systems . |
jets of active galactic nuclei ( agn ) are among the most fascinating objects in the universe . from the time when the term jet " was first introduced by @xcite until today
it is still unclear how these jets are created and formed .
a prime source to gain insight into the physics of extragalactic jets is the broad - line radio galaxy 3c111 ( pks b0415 + 379 ) at @xmath9 = 0.049 and @xmath10 @xmath11 ; @xmath12 ) ] .
the object can be described with a classical fr ii morphology ( @xcite ) exhibiting two radio lobes with hot spots and a single - sided jet ( @xcite ) .
untypical for radio galaxies , a small inclination angle of only @xmath13 to our line of sight has been determined on parsec scales ( @xcite ) .
moreover , 3c111 has a blazar - like spectral energy distribution ( sed ) ( @xcite ) and shows one of the brightest radio cores in the mm - cm wavelength regime of all frii radio galaxies .
superluminal motion was detected in this radio galaxy by and @xcite making this source one of the first radio galaxies to exhibit this effect .
the egret source 3eg j0416 + 3650 has been associated with 3c111 ( , @xcite ) and @xmath14-ray emission from 3c111 has been confirmed by _
fermi / lat _ ( @xcite , @xcite , @xcite ) .
a major flux - density outburst in 1997 was investigated by @xcite with 10 years of radio monitoring data ( 19952005 ) .
in addition @xcite and @xcite report on a possible connection between the accretion disk and the jet of 3c111 .
+ in this paper a new major flux density outburst from 2007 and the associated jet kinematics will be discussed with data from the very large baseline array ( vlba ) , the university of michigan radio astronomy observatory ( umrao ) and the f - gamma program .
the broad - line radio galaxy 3c111 has been part of the vlba 2 cm survey program since its start in 1995 ( @xcite ) and its successor mojave ( monitoring of jets in active galactic nuclei with vlba experiments , @xcite ) in 2002 .
twenty - four epochs of data have been taken from 2006 to 2010 within the mojave program of this source .
phase and amplitude self - calibration as well as hybrid mapping by deconvolution techniques were performed as described by @xcite . utilizing the program difmap ( @xcite ) ,
two - dimensional gaussian components have been fitted in the ( _ u , v_)-plane to the fully calibrated data of each epoch .
we refer to the inner @xmath15mas as the `` core '' region , which can usually be modeled with two gaussian components .
all models have been aligned by assuming the westernmost components to be stationary and all component positions are measured with respect to it .
conservative errors of 15% are assumed for the flux densities of the model - fit components accounting for absolute calibration uncertainties and formal model - fitting uncertainties ( @xcite ) . within the umrao radio - flux - density monitoring program ( @xcite ) ,
more than three decades of single - dish flux - density data have been collected for 3c111 at 4.8ghz , 8.0ghz , and 14.5ghz . in addition , 3c111 is observed monthly by the f - gamma program ( @xcite , @xcite ) at multiple frequencies throughout the cm- and mm - bands since 2007 .
figure [ lc_umrao ] ( left ) shows the long - term radio lightcurve of 3c111 at 4.8ghz . since the start of the measurements
the source has been in a low activity state for almost two decades with only minor activity .
the source showed a major outburst in 1996/1997 ( @xcite ) . starting in 2004 , minor outbursts
are observed and the flux density level increases .
a major flux density outburst starts in early 2007 at high frequencies ( see figure [ lc_umrao ] , right ) , peaking @xmath16 and is subsequently seen at lower frequencies .
a secondary outburst starts mid 2008 at high frequencies .
the overall flux density level at 4.8ghz has been decreasing since the major outburst . in this work ,
we focus on the time period 2006 through 2010 , which contains 24 mojave epochs with an average image rms noise of 0.2mjy@xmath17 and a maximum of 0.4mjy@xmath17 .
the restoring beams of different epochs are very similar with an average of ( 0.84 @xmath18 0.58 ) mas at p.a .
a peak flux density of 6.54jy was measured in may 2008 .
figure [ evolution_3c111 ] shows the evolution of the parsec scale jet of 3c111 observed within the mojave vlba program at 15ghz since may 2008 ( excluding six epochs before the ejection of the primary component ) . a major bright feature appears in may 2008 and is travelling downstream .
the source brightness distribution was modelled with the clean algorithm by within difmap . in figure [ distvsmjd ] , the radial distances to the stationary component in the core region
are plotted for every component as a function of time .
the identification throughout the epochs is based on the comparison of the positions and flux densities of these model components .
this identification is preliminary and will be discussed in more detail in a forthcoming paper . in the context of this paper we focus on the components which can be associated with the major flux density outburst of 2007 ( see fig .
[ lc_umrao ] ) .
+ a linear regression fit has been performed to measure component speeds and ejection dates ( see figure [ distvsmjd ] ) .
the ejection dates of a primary component and a secondary component are quasi identical ( @xmath16 ) within the errors and were found to coincide with the peak of the outburst at high frequencies ( see figure [ lc_umrao ] , right ) .
the components flux density evolution ( see figure [ core_x10x11_fluxvsmjd ] ) shows that the core region was extremely bright during the time of the outburst but dropped significantly after ejecting the bright primary and secondary component .
the determined apparent speed for the primary component is @xmath20 and for the secondary component @xmath21 .
+ the primary component remains at a constant flux density of @xmath22 until mid 2009 .
after that , the component is splitting into multiple parts : a new leading component with trailing components in its wake .
+ the secondary component has a higher flux density than the primary in the beginning of its lifetime which rapidly decays .
this decay suggests that this component disappeared in mid 2009 though an identification with the first trailing component is possible based on position alone .
the calculated apparent speed of the new leading component after mid 2009 is @xmath23 with the flux density decreasing .
the first trailing component has an apparent speed of @xmath24 and shows a constant flux density evolution .
the second trailing component was first observed 2010.53 with a flux density of @xmath25 and could be modeled until 2010.91 with a flux density of less than 100mjy .
the flux density evolution of the second trailing component suggests that this component faded away and thus could not be modeled in epoch 2010.98 .
an association of this second trailing component with the 2008 component is possible based on the flux density evolution and position but needs further investigation of the 2008-outburst . + a similiar behaviour with a leading , secondary and trailing
components has been seen in the evolution of the components associated with the outburst from 1997 by @xcite .
the components were interpreted as a forward and backward shock with the backward shock fading very fast . in this model ,
the plasma of the forward shock entered a region of rapidly decreasing external pressure allowing it to expand into the jet ambient medium and accelerate . in the following ,
the plasma recollimated and trailing features were formed in the wake of the leading component ( ) . + with the addition of the `` core '' region , presented by stars .
the black arrow at the bottom indicates the calculated and almost identical ejection date of the primary and secondary component based on the derived jet kinematics.,scaledwidth=45.0% ]
in this paper , the ejection of new jet components on parsec scales were associated with a major flux density outburst of 3c111 in 2007 .
it was shown that the major flux density outburst can be associated with the ejection of a primary jet component and secondary component .
the evolution of the leading component suggests a split into multiple components .
the full multi - epoch kinematical analysis of the vlba jet of 3c111 between 2006 and 2011 will be presented elsewhere ( grossberger et al . , in prep ) . | we discuss the parsec - scale structural variability of the extragalactic jet 3c111 related to a major radio flux density outburst in 2007 .
the data analyzed were taken within the scope of the mojave , umrao , and f - gamma programs , which monitor a large sample of the radio brightest compact extragalactic jets with the vlba , the university of michigan 26 m , the effelsberg 100 m , and the iram 30 m radio telescopes .
the analysis of the vlba data is performed by fitting gaussian model components in the visibility domain .
we associate the ejection of bright features in the radio jet with a major flux - density outburst in 2007 .
the evolution of these features suggests the formation of a leading component and multiple trailing components .
* keywords * : galaxies : individual : 3c111 - galaxies : active - galaxies : jets - galaxies : nuclei * structural variability of 3c111 on parsec scales * * c. grossberger@xmath0 , m. kadler@xmath1 , j. wilms@xmath0 , c. mller@xmath0 , t. beuchert@xmath0 , e. ros@xmath2 , r. ojha@xmath3 , m. aller@xmath4 , h. aller@xmath4 , e. angelakis@xmath5 , l. fuhrmann@xmath5 , i. nestoras@xmath5 , r. schmidt@xmath5 , j.a .
zensus@xmath5 , t.p .
krichbaum@xmath5 , h. ungerechts@xmath6 , a. sievers@xmath6 , and d. riquelme@xmath6 * _ @xmath0dr .
remeis sternwarte & ecap , universitt erlangen - nrnberg , sternwartstr . 7 , 96049 bamberg , germany + @xmath7institut fr theoretische physik und astrophysik , universitt wrzburg , am hubland , 97074 wrzburg , germany + @xmath8departament dastronomia i astrofsica , universitat de valncia , 46100 burjassot , valncia , spain + @xmath5max - planck - institut fr radioastronomie , auf dem hgel 69 , 53121 bonn , germany + @xmath3goddard space flight center , nasa , greenbelt , 8800 greenbelt rd . , greenbelt , md , 20771 , usa + @xmath4department of astronomy , university of michigan , ann arbor , mi , 48109 - 1042 , usa + @xmath6instituto de radio astronoma milimtrica , avenida divina pastora 7 , local 20 , 18012 , granada , spain + _ |
little is known about galactic extinction in the far ultraviolet region of the spectrum ( fuv ; 912 1150 or 8.7 - 11.0 @xmath4m@xmath2 ; in this paper we include wavenumbers from 3.3 8.7 @xmath4m@xmath2 in what we refer to as the uv since _ iue _ covered this spectral region ) other than that it is relatively strong , it increases with increasing wavenumber , and is variable among sight lines .
the strength of fuv extinction requires that it be properly considered when recovering the intrinsic spectrum of an object .
this is true for most observations taken in the fuv as well as for ultraviolet ( uv ; 1150 3300 or 3.3 8.7 @xmath2 ) or optical observations of significantly red - shifted objects .
the light extinguished at fuv wavelengths plays a major role in the energetics of star - forming galaxies since a large fraction of energy in starlight is processed to long wavelengths through absorption and re - emission by dust . molecular cloud physics and chemistry
are also greatly affected by the amount of uv and fuv radiation that penetrates such regions ; it is these wavelengths that will most affect the formation and destruction of molecules .
finally , the strength and variation of fuv extinction make it a valuable diagnostic for better characterizing the mass of material in and the compositions of small grains ( @xmath5 ) .
extinction curves in the ultraviolet ( 3.3 8.7 @xmath4m@xmath2 ) have been well approximated by two empirical relationships .
for the first of these , fitzpatrick & massa ( 1986 ; 1988 ; 1990 , hereafter fm ) found an expression for extinction that relies on six parameters , two that account for an underlying linear extinction ( @xmath6 and @xmath7 ) , three that describe the strength , location , and width of the 2175 bump ( @xmath8 , @xmath9 and @xmath10 ) and one that describes the curvature of the rise toward the fuv ( @xmath11 ) .
these parameters are simply coefficients in mathematical functions that were fit to uv data , and have no known physical basis .
@xcite argues that the location of the 2175 bump is stable and that the underlying linear extinction can be varied with a single parameter , so that fm fits can be well determined with only four parameters
. the consistent position of the 2175 feature is confirmed by @xcite who find no shifted bumps in a sample of over 400 galactic sight lines .
however , in this paper we will use the traditional six - parameter fm fits . in a different empirical analysis ,
cardelli , clayton & mathis ( 1988 ; 1989 , hereafter ccm ) used data from the infrared to the uv to find an average relationship for extinction as a function of wavenumber that relies on a single parameter , the ratio of total - to - selective extinction , @xmath0 = a@xmath12/e(b - v ) .
@xmath0 values are interpreted as being related to the grain environment since they correlate with particle size in dust grain models .
the ccm relationship seems to be appropriate for the vast majority ( @xmath13 ) of galactic sight lines with known extinction curves @xcite indicating that extinction from the infrared ( ir ) to uv varies in a systematic way among a wide range of dust environments .
a potential value of the ccm formulation is its general predictive capabilities , providing an extinction curve from the ir through uv ( 0.3 8.7 @xmath4m@xmath2 ) based only on photometric observations at ir and optical wavelengths
. the first measurements of extinction beyond 8.7 @xmath4m@xmath2 ( below 1150 ) were made with _
copernicus_. from these initial few sight lines : @xmath14 oph , @xmath15 per , @xmath16 cam and @xmath17 sco @xcite it became evident that the fuv curves are generally a continuation of the uv extinction trend , and that the extinction continues to rise out to 10 @xmath4m@xmath2 .
ccm made a tentative extension of their extinction relation beyond 8 @xmath4m@xmath2 using the _ copernicus _ data .
although the extrapolation was based only on a few sight lines , the congruence of these extinction curves demonstrated that the average @xmath0-dependent law extends beyond 8 @xmath4m@xmath2 in the galaxy .
in addition , ccm showed that a simple extension of the fm fit to fuv wavelengths gives , on average , a fair representation of the actual extinction in this spectral region .
the @xcite `` average '' interstellar extinction curve beyond 9 @xmath4m@xmath2 is based entirely on _ copernicus _ observations of three sight lines . in the nearly thirty years since _ copernicus _ ,
a few additional reddened sight lines in the galaxy have been measured in the fuv using _ voyager _ , _ hut _ , _ orfeus _ , _ fuse _ and suborbital rockets @xcite .
the extinction curves produced in these papers are broadly consistent with extrapolations of the fm and ccm relationships into the fuv .
though the cited studies revealed some evidence that extinction curves continue to rise beyond 8.7 @xmath4m@xmath2 , the paucity of appropriate data did not allow for a systematic study of fuv extinction , including a rigorous determination of whether an extrapolation of fm or ccm to fuv wavelengths would be appropriate in a diverse set of interstellar environments .
instrumental issues of scattered light , time - variable sensitivity , and a limited sample of target and comparison stars have limited the utility of the _ copernicus _ data set @xcite .
the _ voyager _
uvs data also suffer from low resolution which prevents explicit characterization and removal of the effects of molecular hydrogen absorption .
_ hopkins ultraviolet telescope _ and
_ orfeus _ have produced insufficient data for a systematic survey , and rocket - borne instruments have not probed diverse regions @xcite .
_ fuse _ is the first instrument that has produced an appropriate data set for undertaking a thorough and careful analysis of fuv extinction over a wide range of sight line conditions .
this is the first in a series of three papers that will explore extinction in the fuv .
paper ii in the series @xcite will present our first results for fuv extinction in the magellanic clouds .
( in preparation ) will employ a much larger sample of sight lines ( @xmath18 70 ) in order to provide a greater statistical significance to trends in observed extinction properties .
the present study investigates fuv extinction characteristics in nine galactic sight lines that span a wide range of dust environments as measured by @xmath0 ( a proxy for dust grain properties ) .
we explore how the extinction curves from 8.7 9.5 @xmath4m@xmath2 ( 1150 - 1050 ) relate to extrapolations from the fm and ccm relationships .
we also apply a simple model to the extinction curves in order to investigate how the addition of fuv data might change the constraints on dust characteristics .
though our 3-component grain model does not account for the full complexity of dust in the interstellar medium , the model results do provide a rough guide for estimating mass requirements among the grain components . in
2 we present the observations , data reduction procedures , and modeling algorithm .
we discuss the nature of the sight lines , analyze the extinction curves in terms of fm and ccm fits , and consider model implications for the `` small grain '' population in 3 .
a summary is given in 4 .
the fuv data used for this study were obtained with _ fuse _ through the low - resolution ( lwrs ) aperture resulting in a resolution on the order of r = 20,000 and covering a wavelength range from approximately 950 to 1185 (or 10.5 8.4 @xmath4m@xmath2 ) .
the observations , originally obtained for a variety of programs ( including our own ) , were made between october 1999 and september 2001 .
the basic characteristics of the observations including the _ fuse _ target designations , the observing modes , and the exposure times are summarized in table 1 .
the data were retrieved from the _ fuse _ archive and recalibrated using calfuse version 2.2.1 .
we found that calibrations with versions of calfuse earlier than 2.0 did not possess sufficient photometric accuracy for determining reliable extinction curves , for instance see @xcite .
the eight calibrated spectra for each data set were mapped to a common wavelength array spanning the entire range of the data .
the data were merged by finding the mean flux at each wavelength point , weighted by three broad signal - to - noise bins determined for each individual spectrum .
the error arrays were merged in a manner consistent with that of the data arrays . for several observations ,
there were obvious misalignments of the star in the aperture , resulting in low or absent flux values for individual spectra .
such bad data segments were flagged before merging the spectra and were given zero weight in the co - addition .
the _ fuse _ data anomaly known as the `` worm '' @xcite was also present in a majority of the observations .
the flux values that were lowered by this anomaly were also flagged , and thus did not contribute to the final merged spectra .
merged spectra were co - added in the case of multiple observations of an individual sight line . for our sample of stars
, we note that the flux values in the wavelength overlap region between the _ iue _ and _ fuse _ data ( 1150 1185 ) match well .
this contrasts with the lower signal - to - noise magellanic cloud sight line data of @xcite who , in some cases , needed to scale their _ fuse _ data in order to provide a better match to the _ iue _ flux levels .
absorption features produced by interstellar molecular hydrogen can have a substantial effect on the observed fuv spectrum of a star .
that is to say , a considerable fraction of the light below @xmath201108 may be absorbed by this molecule .
although the strongest lines of each h@xmath21 band can reduce the observed flux to nearly zero , the broad overlapping damping wings are able to be modeled and removed ( i.e. , the spectrum is rectified to the true continuum level ) .
in addition , the numerous narrow lines due to rotationally excited states can be reliably removed , with the exception of strongest line cores .
even some of the lightly reddened stars ( some of our comparison stars ) can show enough h@xmath19 absorption that we chose to remove it .
our techniques for measuring the column densities of each rotational state are described by @xcite . in brief , we use profile fitting to measure the strongest lines , and a curve of growth analysis for the weaker lines .
the latter analysis also gives an `` effective '' @xmath22-value for the rotationally excited lines , which is necessary because we generally do not have velocity structure information for h@xmath19 .
we generate a model transmission spectrum based on the retrieved column densities and @xmath22-value . because slight wavelength mismatches will result in a poor removal of the h@xmath19 spectrum , we perform a running cross - correlation between the model and the observed spectrum .
this produces a map of the appropriate wavelength shift as a function of pixel number , and once those shifts are applied to the model , one can simply divide the observed spectrum by the model to remove h@xmath21 . we flag ( and exclude from further analysis ) all pixels for which the transmission model values fall below 30% of the continuum level .
the spectra displayed in figure 1 have the effects of h@xmath21 removed where the spectrum is 70% or more of the continuum level .
the spectra are not plotted at wavelengths where the h@xmath21 absorbs greater than 30% of the continuum .
the solid vertical lines in figure 1 indicate the locations of the h i lyman absorption features . using a variant of the method outlined by @xcite ( see gordon et al .
2003 for details ) , we have estimated the column densities of h i and have attempted to reconstruct the continuum near the lyman absorption . as shown by the spectra in figure 1 ,
the lyman features are not precisely accounted for , particularly in the line cores .
we take a conservative approach and avoid the regions near the lyman lines when determining the extinction curves .
the spectral regions between the vertical dashed lines in figure 1 show the regions around the lyman features that have been excluded from the extinction curve determinations below .
the crowding of the lyman lines and the profusion of h@xmath21 absorption at wavelengths shorter than ly@xmath23 ( 1026 or 9.75 @xmath4m@xmath2 ) means that there are no reliable continuum data above 9.5 @xmath4m@xmath2 .
we combine data from 2mass near - infrared photometry @xcite , optical photometry ( references are given in table 1 ) , archival low - dispersion @xmath24 uv spectra processed with the @xcite method , and @xmath25 fuv spectra for our target sight lines ; the spectra were binned to 5 .
we determined extinction curves from those data using the pair method @xcite .
the spectral matches between the reddened and comparison stars have been rigorously evaluated by examining their spectra in the ir through uv spectral regions .
the stellar pairs used for this study are shown in figure 1 where the spectrum of the reddened star in each pair lies below the comparison star s spectrum .
the _ fuse _ data , at wavenumbers greater than 8.4 @xmath4m@xmath2 , obviously have lower signal - to - noise as compared to the longer wavelength data .
our method for assembling the complete extinction curves from the ir to the fuv follows that of @xcite and @xcite .
the derived extinction curves for each pair are shown in figure 2 .
note that the curves have a smooth transition from the _ iue _ to _ fuse _ wavelengths around 8.4 @xmath4m@xmath2 .
we employ the maximum entropy method ( mem ) algorithm for fitting the extinction curves as developed by kim , martin & hendry ( 1994 ; hereafter kmh ) , with slight modifications to facilitate application to the _ fuse _ data .
additional information regarding the mem implementation may also be found in @xcite . as recommended by kmh
, we use the `` mass distribution ''
m(a ) da = mass of dust grains in the grain - radius interval a to a+da instead of the more traditional number of grains or `` size distribution . '' in this form , the classical mrn - type model @xcite becomes @xmath26 .
the grain sizes are divided into 50 logarithmically - spaced bins over the range 0.0025 - 2.7 . as discussed by kmh
, the shape of the mass distribution is strongly constrained only for data over the region @xmath200.02 - 1 . below a size of 0.02 ,
the extinction even in the ultraviolet becomes increasingly dominated by absorption as sizes approach the small - particle ( rayleigh ) limit and thus constrains only total mass .
above 1 , the `` gray '' nature of opacity ( i.e. , extinction efficiency approaching 2 ) provides a similar type of integral constraint ( on surface area ) .
thus there is danger of a large unconstrained mass if the default mass distribution used in the definition of entropy is not well chosen .
we therefore specify the defaults ( templates ) using the functional form of a power - law with an exponential decay above some characteristic large size @xmath27 ( ped , see kmh ) : @xmath28 .
the ped is essentially the traditional gamma distribution function .
the `` observations '' to which the model is fit are taken from the fm parameterization of the uv data and from the visible / infrared extinction data , resampled at 34 wavelengths . for model wavelengths longward of the observed data ,
the extinction is calculated using the ccm relation .
however , their inclusion in the modeling effort is an artifact of the current mem implementation , and we do not wish for them to contribute to the resulting mass distributions . as a result ,
their errors are set to a large number that effectively reduces their weight to zero .
the `` errors '' that are used to weight the data in the mem algorithm are calculated to represent only the scatter or random errors in the extinction curve . in the visible and infrared , these values
are taken directly from the random error estimate produced during the extinction curve generation .
unfortunately , the same approach for the uv points tends to produce `` errors '' which are much smaller than the point - to - point scatter in the data . as a result , for each model uv bin we have adopted the average deviation of the fm curve from the actual extinction data . in this work , we consider ( only ) three - component models of homogeneous , spherical grains : modified astronomical silicate " @xcite , amorphous carbon @xcite , and graphite @xcite , with mass densities ( @xmath29 ) of 3.3 g/@xmath30 , 1.8 g/@xmath30 , and 2.3 g/@xmath30 , respectively .
the graphite component is included , as in many models of interstellar extinction , to produce the extinction bump at 2175 .
only small graphite particles are appropriate for this purpose , and so for that component we chose @xmath31 , much smaller than for the other components ( @xmath32 ) . at each wavelength , the appropriately - weighted grain cross section is integrated for each bin ( and for each separate composition , see equation 1 in kmh ) . in order to incorporate abundance constraints ( see below ) ,
these are expressed as extinction per unit h column density ( accounting for h i and h@xmath19 ) .
the total mass of dust is limited by the observed dust - to - gas ratio and cosmic " abundances .
the mem algorithm includes explicit constraints that do not allow the model to use more carbon or silicon than is `` available '' .
the cosmic abundances adopted were 358 c and 35 si atoms per million h. the c abundance is an estimate from young f and g stars @xcite , and the si abundance is solar @xcite .
we note that the cosmic c abundance used here is generous ; recent results suggest that the solar c / h may be as low as @xmath33 c per million h @xcite . in order to allow generally for observed carbon gas - phase abundances , we require that no more than 70% of the adopted cosmic c may be used ( i.e. , 251 c per million h ) , whereas we allow up to 120% of the cosmic si
( i.e. , 42 si per million h ) .
these relaxed constraints allow for uncertainties in abundances and depletion , and in the dust - to - gas ratio ( here @xmath34 $ ] ) .
note that the model does not have to use this much c or si , but often does . to provide consistency between the individual analyses , we use the same abundance constraints , defaults , and initial guesses of the mass distribtion ( to which the model is not particularly sensitive ) .
in order to isolate the effect of including the _ fuse _ data in the analyses , each model is run with and without the _ fuse _ data , ( the latter case giving no weight to data beyond 8.7 @xmath2 ) .
the mem algorithm proceeds iteratively to find mass distributions which reproduce the data .
the goodness of fit is judged by a reduced-@xmath35 , the `` target '' value of which should be close to unity for an acceptable solution ( we used 1.0 as our target ) . in the mem algorithm used , this value is a constraint ( along with abundance limits , which must be satisfied for any solution to be considered valid . in other words , a successful mem solution will have a reduced-@xmath35 equal to the target value .
the assessment of goodness of fit is relative to the adopted errors , and so it is important that these be appropriately calculated ( systematic errors are not included ) . from this point of view , our assessment of random error ( see above ) appears reasonable , producing `` good '' fits for a reduced-@xmath36 .
the basic characteristics of the nine sight lines used in this study are shown in table 2 .
the reddened sight lines cover a wide range of galactic environments as indicated by the large spread in @xmath0 values , from 2.43 to 3.81 ( see table 2 ) .
these values of @xmath0 were determined using a @xmath35 minimization method ; see @xcite for the details of the method and associated uncertainties . as noted in the introduction , the extinction along these sight lines is unusual compared to a typical line of sight measured in the local interstellar medium . nevertheless , the sight lines toward hd 14250 , hd 73882 , hd 99872 , hd 167971 , hd 239729 , and hd 239683 , though having different values of @xmath0 , are still fit well by ccm @xcite .
hd 14250 and hd 73882 were part of the original ccm sample .
hd 239729 and 239683 are in trumpler 37 along with hd 204827 but do not share any of the anomalous properties of the dust in the latter sight line @xcite .
hd 197770 shows a broad , weak spectral feature in the ultraviolet interstellar linear polarization centered close to 2175 bump @xcite .
hd 197770 is an evolved , spectroscopic , eclipsing binary with two b2 stars @xcite . the star seems to lie on the edge of a large area of molecular clouds and star formation including lynds 1036 and 1049 in the cygnus region @xcite . on the _ iras
_ maps there is a bright 60 shell , 14 in radius centered on hd 197770 , surrounded by an apparent bubble cleared of dust with a radius of approximately 24 @xcite . only five galactic sight lines ( hd 29647 , hd 62542 , hd 204827 , hd 210121 , and hd 283809 )
are known to show systematic deviations from ccm , excluding the fuv @xcite ; these sight lines all sample dense , molecule - rich clouds .
two of these sight lines , toward hd 62542 and hd 210121 , both of which show strong uv extinction , are included in our data set .
the gas toward hd 62542 is rich in cn and ch molecules @xcite , and the sight line lies on the edge of material swept up by a stellar wind bubble .
the dust in the molecular cloud associated with hd 210121 @xcite seems likely to have been processed too as it was propelled into the halo during a galactic fountain or other event . as for hd 62542 ,
the non - ccm extinction curve toward hd 210121 has a weak bump , and an even steeper fuv rise , indicating size distributions that are skewed toward small grains @xcite , possibly from exposure to shocks or strong uv radiation that disrupted large grains .
the empirical six - parameter fm relationship well reproduces galactic uv extinction curves up to 8.7 @xmath4m@xmath2 .
ccm is able to reliably predict extinction curves up to 8.7 @xmath4m@xmath2 with the exception of @xmath37 of observed galactic sight lines @xcite .
an extension of these relationships into the fuv would be valuable for correcting extinction effects .
in addition , a valid extension of ccm would be especially important because of its ability to predict ( to some degree ) the extinction curves based on ground - based photometric parameters alone .
figures 3a and 3b show the extinction curves for our sample superimposed with three fits to each .
the fits include each : an fm fit based on _ iue _ data alone , an fm fit based on _ iue _ and _ fuse _ data , and a ccm fit based on the measured value of @xmath0 = a@xmath12/e(b - v ) .
each of the galactic extinction curves in our sample is well reproduced by fm curves fitted to data in the _ iue _ and _ fuse _ wavelength regions .
the six parameters used for these fm fits are shown in table 3 .
the good agreement between the extinction curves and fm fits in the fuv is not necessarily expected given that @xcite based their empirical fit to the extinction curves on data that went up to only 8.7 @xmath4m@xmath2 ( i.e. , only in the _ iue _ wavelength region ) .
@xcite found that the curvature of the extinction curve from 5.9 - 8.7 @xmath4m@xmath2 always has the same shape regardless of the dust size distribution or physical environment .
our small sample of sight lines suggest that this relationship extends to the fuv and that the extinction curve from 5.9 - 9.5 @xmath4m@xmath2 can be fit with a single continuous function ; the same function found by @xcite .
the continuity of the curvature shape at wavenumbers above 8.7 @xmath4m@xmath2 is verified by the fact that the fm fits based on the _ iue _ data alone are , in most cases , indistinguishable from the fm fits to the combined _ iue _ and _ fuse _ data ( the solid and dotted lines in figure 3 ) .
the largest deviation between the two fm fits for a given extinction curve occurs for hd 197770 where the fits differ by less than 0.5 magnitudes at 9.5 @xmath4m@xmath2 ; this is within the expected uncertainty for this noisier portion of the extinction curve .
analysis of our sample suggests that fm fits can be extrapolated safely from _ iue _ wavelengths up to 9.5 @xmath4m@xmath2 . @xcite
propose that the stability of the extinction curve shape at short wavelengths results only from a dust component optical property , as opposed to a function involving grain size .
while we certainly find a similar behavior in shape of the extinction curve from 1050 - 1700 , we can not exclude the possibility of systematic variations in particle sizes as viable mechanisms .
for these lines of sight , selected for their unusual extinction , the predictive capability of the ccm relationship for fuv extinction is more problematic .
note , however , that in contrast to the fm fits , no ultraviolet data , neither uv nor fuv , are used to determine the single ccm fitting paramater @xmath0 ; thus the ability to predict the uv extinction for most stars is a remarkable feature of ccm .
for the two sight lines in our sample where ccm does not match even the uv extinction , toward hd 62542 and hd 210121 , it also fails to reproduce the fuv .
this is not a surprising result since the longer wavelength data already suggested that the grains in these sight lines have likely experienced processing that is different from that in more `` average '' galactic sight lines .
the departure from the ccm relationship grows with increased wavenumber over the measured region up to a maximum of about 3 magnitudes per a@xmath12 at 9.5 @xmath4m@xmath2 ; for hd 62542 and hd 210121 , ccm underpredicts the fuv extinction as much as 40% and 35% , respectively .
hd 62542 is known to have an anomalous 2175 bump unusually weak and peaking at an atypically high wavenumber ; @xcite find @xmath38 and we find @xmath39 .
this suggests that the failure of ccm to reproduce the fuv extinction may be associated with anomalous properties of the particles producing the 2175 bump along this sight line .
the sight lines toward four stars , hd 73882 , hd 99872 , hd 167971 and hd 239729 , are , within uncertainties , well represented by ccm over the ir to uv wavelength region , but much less so in the fuv .
the ccm functions for these sight lines differ from the well - fit fm curves by 2 @xmath17 or more ( based on the @xmath0 uncertainties ) at 9.5 @xmath4m@xmath2 . as with the hd 62542 and hd 210121 extinction curves ,
the differences between the ccm predictions and the measurements increase with increasing wavenumber .
the fact that ccm fits the extinction curves in the ir through uv , but diverges in the fuv may indicate that the systematic processing of larger grains does not necessarily continue to the small grain population , at least in the same way .
ccm fuv extinction in two of these sight lines ( hd 73882 and hd 239729 ) is too high while it is too low in the other two ( hd 99872 and hd 167971 ) .
the @xmath0 values for this group of four sight lines ranges from 2.99 to 3.81 .
the final three sight lines in our sample , toward hd 14250 , hd 197770 and hd 239683 , have ccm curves that predict well the fuv extinction .
the @xmath0 values of these three sight lines range from 2.44 to 2.98 ; smaller than the group for which ccm did poorly .
while one might suggest that ccm can be extrapolated into the fuv more reliably for smaller @xmath40 values and the potential connection to an increased relative abundance of small grains , further generalization awaits a larger sample of fuv extinction ( as will be presented in paper iii ) .
the resulting model - data comparisons are presented in figure 4 , which shows the amount of extinction provided by each of the three materials and the total extinction of the model compared to the measured extinction curve .
the percentages of the carbon and silicon along the sight lines that are used to produce the displayed fits are given in table 4 .
our models use three grain types to reproduce the observed extinction , specifically astronomical silicates , amorphous carbon , and graphite .
a more realistic dust model would include other compositions such as oxides and grains composed of a combination of carbonaceous and silicaceous materials .
it would also account for non - spherical particles and the effects of coatings on and vacuums in the grains .
we have chosen to use a simple model here because of the computational complexity and arbitrary assumptions often associated with a more physically complete approach @xcite . by doing so , we acknowledge that the detailed results of the models , such as specific populations of grain types and sizes , and the abundance constraints that they imply , are to be used with caution . however , these simple models can provide some useful information about dust . since the absorption characteristics of a grain may be well constrained by its bulk behavior , the models can provide estimates of the mass distribution among grain - size populations that are appropriate to more complex systems .
the left and right columns in figure 4 show the mem fits to the extinction curves based on data up to 9.5 @xmath4m@xmath2 ( _ fuse _ wavelengths ) , and 8.7 @xmath4m@xmath2 ( _ iue _ wavelengths ) , respectively .
although some of the fuv extinction is produced by the grains that are also responsible for extinction at longer wavelengths , it is clear that models based on data only to 8.7 @xmath4m@xmath2 do not always fit the fuv extinction well .
other than toward hd 167971 and hd 197770 , the extinction curves require a larger fraction of grain material to be in small dust particles , mostly silicates in our model , to adequately account for the fuv .
more specifically , within the context of our simple three - component model , the inclusion of the fuv requires up to 10% more silicon to be in grains ( see table 4 ) .
since the fm parameter @xmath11 is a measure of the strength of the far uv rise in an extinction curve , it is not surprising that we find a correlation between this value and an increased need for small silicate grains .
a linear fit to this relationship ( a higher - order polynomial is not statistically justifiable ) indicates that in order to fit the fuv constraint , 12% more silicon is needed per increase of 1.0 in @xmath11 ; the @xmath41 scatter about the line is a 2% change in the required silicon .
the largest portion of the scatter is likely attributable to the errors assigned to the extinction curve data ( see 2.4 ) .
since the models tend to underfit the extinction when less constrained ( when the data have larger uncertainties ; see the left panels in figure 4 ) , we are likely underestimating the extra mass needed in small silicates for our simple model to fit the fuv .
the fraction of c allocated to amorphous carbon grains often decreases in the models that include the fuv extinction constraint ( see table 4 ) ; amorphous carbon serves as a source of visible opacity in these models . on the other hand ,
the abundance of c that the models allocate to graphite grains is often enhanced when the fuv is included .
this general increase in graphite and decrease in amorphous carbon represents the model s reallocation of c to fit better the extinction out to 9.5 @xmath2 ; the upward curvature in the graphite extinction at wavenumbers greater than 8 @xmath4m@xmath2 fits the upward turn often seen in the extinction curves at these wavelengths .
nevertheless , one should also consider the possibility that there is something wrong with the dielectric function of our silicate component in not being able to produce this upward curvature , or perhaps that other components missing entirely , e.g. , pahs , nanoparticles .
we note that the fits to x @xmath42 @xmath2 in figure 4 are often poorest in the fuv where the models tend to underestimate the extinction .
since a disproportionate share of the target @xmath35 is coming from the fuv , we have experimented with lowering the target reduced-@xmath35 in the models to better fit this region .
we show two examples of forcing a tighter fuv fit in figure 5 .
the figure displays three mem models for the sight lines toward hd 62542 and hd 210121 , one based on each : data up to 8.7 @xmath2 with a reduced-@xmath43 , data up to 9.5 @xmath2 with a reduced-@xmath43 , and data up to 9.5 @xmath2 with a reduced-@xmath44 .
the left panels show the fits to the data for each model .
the right panels show the resultant mass distribution of each component relative to the mass of hydrogen .
the reduced-@xmath36 , x @xmath45 @xmath2 model for hd 62542 does not use all of the si or c available to it , so the model can draw on both of these elements when it tries to fit better the fuv data ( see table 4 ) .
the reduced-@xmath44 model does , in fact , take the entire allocation of si ( 120% ) , however it uses less of the available c than the less - constrained fit ; specifically , the more - constrained model puts 30% of the total carbon abundance into amorphous c and 7% into graphite .
the top right panel of figure 5 shows that much of the `` extra '' silicon used to fit better the fuv was put into small grains .
the grain sizes below 200 are not well resolved , so the shape of the mass curve at these small sizes is guided by the model template and not the actual distribution in size . however , the changes in the overall mass of grains below 0.02 @xmath4 m , as shown in the figure , are significant .
figure 5 also shows that the model is reallocating the size distributions of the grains so that , for instance , the graphite is contributing more extinction at 2175 relative to the amorphous carbon ( note the decrease in amorphous carbon grains with sizes around 0.2 as the fuv is better fit ) .
recall however that at reduced-@xmath44 , the model is overconstrained if our treatment of the errors is appropriate .
we caution against overinterpreting these size distribution results ; we present them merely to show that this simplified model requires changes in the grain size distributions in order to better fit the fuv .
the reduced-@xmath36 model for hd 210121 uses the maximum available si , so it can only draw on c in order to put more mass into grains for a better fuv fit .
the reduced-@xmath44 model continues to incorporate the maximum allocated si , and increases the c that it uses for each of the amorphous carbon grains and graphite by 1% of the total c abundance .
this overconstrained model is able to fit the extinction primarily by reallocating the size - distributions of the grains in a way that is similar to the reduced-@xmath44 model for hd 62542 .
a rigorous analysis that relates grain size distributions to extinction will require a more complex model as well as extinction curves with smaller uncertainties in the fuv .
we present the extinction curves from the ir to the fuv ( up to 9.5 @xmath4m@xmath2 ) for nine galactic sight lines observed with _
fuse_. \4 .
ccm does fit the ir uv in seven of our sight lines . for three of these ,
all with @xmath0 values below the galactic average , an extrapolation of ccm does properly predict the fuv extinction .
the other four sight lines all have larger @xmath0 values than the previous group .
an extrapolation of ccm over - predicts the fuv extinction in two of these sight lines , and under - predicts it in the other two .
our simple mem grain models are generally able to fit the extinction curves that include the fuv constraints .
the more highly constrained models usually require more mass to be in small grains ( radii @xmath46 ) in order to reproduce an extinction curve that includes the ir fuv as compared to those using only the ir uv .
allende prieto , c. , lambert , d. l. , & asplund , m. 2002 , , 573 , l137 bohlin , r. c. 1975 , , 200 , 402 buss , r. h , allen , m. , mccandliss , s. , kruk , j. , liu , j. , & brown , t. 1994 , , 430 , 630 cardelli , j. a. , clayton , g. , c. , & mathis , j. s. 1988 , , 329 , l33 . 1989 , , 345 , 245 cardelli , j. a. , & savage , b. d. 1988 , , 325 , 864 cardelli , j. a. , suntzeff , n. b. , edgar , r. j. , & savage , b. d. 1990 , , 362 , 551 cartledge , s. i. b. , clayton , g. c. , gordon , k. d. , rachford , b. l. , draine , b. t. , martin , p. g. , mathis , j. s. , misselt , k. a. , sofia , u. j. , whittet , d. c. b. , & wolff , m. j. 2005 , , submitted clayton , g. c. , anderson , c. m. , magalhaes , a. m. , code , a. d. , nordsieck , k. h. , meade , m. r. , wolff , m. , babler , b. , bjorkman , k. , schulte - ladbeck , r. e. , taylor , m. , & whitney , b. a. 1992 , , 385 , l53 clayton , g. c. , gordon , k. d. , salama , f. , allamandola , l. j. , martin , p. g. , snow , t. p. , whittet , d. c. b. , witt , a. n. , & wolff , m. j. 2003 , , 592 , 947 clayton , g. c. , wolff , m. j. , sofia , u. j. , gordon , k. d. , & misselt , k. a. , 2003 , 588 , 871 cousins , a. w. j. , & stoy , r. h. 1962 , r. obs . bull . , 64 , 103 denoyelle , j. 1977 , , 27 , 343 dworetsky , m. m. , whitelock , p. a. , & carnochan , d. j. , , 201 , 901 feinstein , a. 1967 , , 149 , 107 . 1969 , , 143 , 273 fitzpatrick , e. l. 1999 , , 111 , 63 fitzpatrick , e. l. , & massa , d. 1986 , , 307 , 286 . 1988 , , 328 , 734 . 1990 , , 72 , 163 forbes , d. 1984 , iau inform . bull .
stars , 2605 , 1 garrison , r. f. , & kormendy , j. 1976 , , 88 , 865 gaustad , j. e. , & van buren , d. 1993 , , 105 , 1127 gordon , k. d. , & clayton , g. c. 1998 , , 500 , 816 gordon , k. d. , clayton , g. c. , misselt , k. a. , landolt , a. u. , & wolff , m. j. 2003 , , 594 , 279 gordon , k. d. , clayton , g. c. , smith , t. l. , aufdenberg , j. p. , drilling , j. s. , hanson , m. m. , anderson , c. m. , & mulliss , c. l. 1998 , , 115 , 2561 green , j. c. , snow , t. p. , cook , t. a. , cash , w. c. , & poplawski , o. 1992 , , 395 , 289 hardie , r. h. , heiser , a. m. , & tolbert , c. r. 1964 , , 140 , 1472 hendry , p. d. , & mochnacki , s. w. 200 , , 531 , 467 hill , p. w. , kilkenny , d. , & van breda , i. g. 1974 , , 168 , 451 hill , p. w. , & lynas - gray , a. e. 1977 , , 180 , 691 holweger , h. 2001 , in solar and galactic composition , ed .
r. f. wimmer - schweingruber ( berlin : springer - verlag ) , 23 hutchings , j. b. , & gaiasson , j. 2001 , , 113 , 1205 jenkins , e. b. , savage , b. d. , & spitzer , l. 1986 , , 301 , 355 kim , s .- h . , martin , p. g. & hendry , p. d. 1994 , , 422 , 164 laor , a. , & draine , b. t. 1993 , , 402 , 441 larson , k. a. , whittet , d. c. b. , & hough , j. h. 1996 , , 472 , 755 larson , k. a. , wolff , m. j. , roberge , w. g. , whittet , d. c. b. , & he , l. 2000 , , 532 , 1021 lewis , n. k. , cook , t. a. , & chakrabarti , s. 2005 , , 619 , 357 longo , r. , stalio , r. , polidan , r. s. , & rossi , l. 1989 , , 566 , 267 martin , p. , sommerville , w. , mcnally , d. , whittet , d. , allen , r. , walsh , j. , & wolff , m. 1995 , in the diffuse interstellar bands , ed . a. g. g. m. tielens , and t. p. snow ( dordrecht : klewer ) , 271 massa , d. , & fitzpatrick , e. l. 2000 , , 126 , 517 massa , d. , savage , b. d. , & fitzpatrick , e. l. 1983 , , 266 , 662 mathis , j. s. , rumpl , w. , & nordsieck , k. h. 1977 , , 217 , 425 mendoza , e. e. 1967 , bol .
tacub . , 4 , 149 rachford , b. l. , snow , t. p. , tumlinson , j. , shull , j. m. , blair , w. p. , ferlet , r. , friedman , s. d. , gry , c. , jenkins , e. b. , morton , d. c. , savage , b. d. , sonnentrucker , p. , vidal - madjar , a. , welty , d. e. , & york , d. g. 2002 , , 577 , 221 rachford , b. l. , snow , t. p. , tumlinson , j. , shull , j. m. , roueff , e. , andre , m. , desert j. m- . , ferlet , r. , vidal - madjar , a. , & york , d. g. 2001 , , 555 , 839 sahnow , d. j. 2003 , in future euv / uv and visible space astrophysics missions and instrumentation , ed .
j. c. blades , & o. h. w. siegmund , ( bellingham : international society for optical engineering ) , 610 sahnow , d. j. , friedman , s. d. , oegerle , w. r. , moos , h. w. , green , j. c. , & siegmund , o. h. 1996 , in space telescopes and instruments iv , ed .
p. y. bely , & j. b. breckinridge , ( bellingham : international society for optical engineering ) , 2 sasseen , t. p. , hurwitz , m. , dixon , w. v. , & airieau , s. 2002 , , 566 , 267 savage , b. d. , & mathis , j. s. 1979 , , 17 , 73 skrutskie , m. f. , et al .
, in the impact of large scale near - ir sky surveys , ( dordrecht : klewer ) , 25 snow , t. p. , allen , m. m. , & polidan , r. s. 1990 , , 359 , l23 snow , t. p. , & york , d. g. 1975 , , 34 , 19 sofia , u. j. , & meyer , d. m. 2001 , , 554 , l221 valencic , l.a . ,
clayton , g.c . ,
gordon , k.d . ,
& smith , t.l .
2003 , , 598 , 369 valencic , l.a . ,
clayton , g.c . , & gordon , k.d .
2004 , , 616 , 912 weingartner , j.c . , & draine , b.t .
2001 , , 548 , 296 welty , d. e. , & fowler , j. r. 1992 , , 393 , 193 wolff , m.j . , clayton , g.c . ,
kim , s.h . ,
martin , p.g . , & anderson , c.m .
1997 , , 478 , 395 york , d. g. , drake , j. f. , jenkins , e. b. , morton , d. c. , rogerson , j. b. , & spitzer , l. 1973 , , 182 , l1 zubko , v. g. , mennella , v. , colangeli , l. , & bussoletti , e.,1996 , , 282 , 1321 llllc [ table_obs ] bd+32 270&a063 - 07 & 4,900&hist & 1 + hd 14250 & a118 - 06 & 8,200&ttag & 2 + hd 37332 & b060 - 12 & 2,300&hist & 3 + hd 51013 & a063 - 09 & 4,100&hist & 4 + hd 62542 & p116 - 02&11,400&ttag & 5 + hd 73882 & x021 - 03&25,500&ttag & 6 + & p116 - 13&13,600&ttag & + hd 97471 & a118 - 04 & 4,200&hist & 7 + hd 99872 & a120 - 06 & 3,300&hist & 5 + hd 99890 & p102 - 46 & 4,600&hist & 7 + hd 114444&a118 - 05 & 5,500&ttag & 8 + hd 116852&p101 - 38 & 5,100&hist & 8 + hd 167971&p116 - 21 & 9,500&ttag & 9 + hd 197770&a118 - 13 & 6,000&ttag & 10 + hd 210121&p116 - 30&13,800&ttag & 11 + hd 239683&a118 - 10 & 9,200&ttag & 12 + hd 239729&a118 - 07 & 9,900&ttag & 12 + lllcclccccc
hd 14250 & bd+32 270&b1 iv & 5.5&&b2 v & & [email protected]&0.48@xmath48&[email protected] & 1 + hd 62542 & hd 37332 & b5 v & 6.5&&b5 v & & [email protected]&[email protected]&[email protected] & 2 + hd 73882 & hd 116852&o9 iii&12.9&&o9 iii&0.62 & [email protected]&[email protected]&[email protected] & 3 + hd 99872 & hd 51013 & b3 v & 3.5&&b3 v & & [email protected]&[email protected]&[email protected] & 1 + hd 167971&hd 99890 & b0 v & 7.1&&b0.5 v & & [email protected]&[email protected]&[email protected] & 2,4 + hd 197770&hd 114444&b2 iii&10.7&&b2 iii & & [email protected]&[email protected]&[email protected] & 1 + hd 210121&hd 51013 & b3 v & 5.6&&b3 v & & [email protected]&[email protected]&[email protected] & 1 + hd 239683&hd 51013 & b3 iv & 5.6&&b3 v & & [email protected]&[email protected]&[email protected] & 1 + hd 239729&hd 97471 & b0 v & 11.8&&b0 v & 0.80 & [email protected]&[email protected]&[email protected] & 2 lcccccc [ table_fm ] + x @xmath1 9.5 @xmath2 & -0.351 & 0.779 & 3.945 & 0.454 & 4.593 & 0.941 + error & 1.031 & 0.082 & 0.507 & 0.044 & 0.010 &
0.022 + x @xmath1 8.7 @xmath2 & -0.352 & 0.780 & 3.942 & 0.448 & 4.593 & 0.941 + error & 0.079 & 0.076 & 0.451 & 0.052 & 0.012 & 0.016 + + x @xmath1 9.5 @xmath2 & -1.240 & 1.253 & 0.446 & 1.047 & 4.780 & 0.798 + error & 0.395 & 0.141 & 0.124 & 0.128 & 0.080 &
0.014 + x @xmath1 8.7 @xmath2 & -1.203 & 1.242 & 0.486 & 1.131 & 4.787 & 0.826 + error & 0.401 & 0.146 & 0.137 & 0.158 & 0.080 & 0.014 + + x @xmath1 9.5 @xmath2 & 0.879 & 0.593 & 3.123 & 0.810 & 4.603 & 1.248 + error & 0.219 & 0.046 & 0.266 & 0.046 & 0.030 &
0.021 + x @xmath1 8.7 @xmath2 & 0.817 & 0.604 & 3.168 & 0.800 & 4.600 & 1.254 + error & 0.163 & 0.040 & 0.234 & 0.049 & 0.020 & 0.020 + + x @xmath1 9.5 @xmath2 & 0.101 & 0.488 & 5.830 & 0.360 & 4.590 & 1.180 + error & 0.038 & 0.080 & 1.310 & 0.097 & 0.076 & 0.020 + x @xmath1 8.7 @xmath2 & 0.097 & 0.487 & 5.862 & 0.370 & 4.590 & 1.182 + error & 0.033 & 0.081 & 1.312 & 0.115 & 0.076 & 0.020 + + x @xmath1 9.5 @xmath2 & 1.538 & 0.330 & 2.811 & 0.381 & 4.577 & 0.808 + error & 0.256 & 0.050 & 0.288 & 0.038 & 0.021 & 0.013 + x @xmath1 8.7 @xmath2 & 1.450 & 0.347 & 2.826 & 0.351 & 4.575 & 0.810 + error & 0.249 & 0.044 & 0.277 & 0.044 & 0.018 & 0.013 + + x @xmath1 9.5 @xmath2 & 1.217 & 0.517 & 5.359 & 0.733 & 4.616 & 1.233 + error & 0.326 & 0.056 & 0.586 & 0.058 & 0.032 & 0.021 + x @xmath1 8.7 @xmath2 & 1.112 & 0.546 & 5.115 & 0.637 & 4.609 & 1.207 + error & 0.298 & 0.053 & 0.496 & 0.083 & 0.030 & 0.020 + + x @xmath1 9.5 @xmath2 & -3.296 & 1.882 & 1.913 & 0.672 & 4.611 & 1.093 + error & 0.676 & 0.262 & 0.450 & 0.106 & 0.066 &
0.018 + x @xmath1 8.7 @xmath2 & -3.314 & 1.887 & 1.902 & 0.663 & 4.609 & 1.090 + error & 0.760 & 0.268 & 0.454 & 0.119 & 0.064 & 0.018 + + x @xmath1 9.5 @xmath2 & 0.116 & 0.703 & 4.726 & 0.515 & 4.614 & 1.235 + error & 0.031 & 0.057 & 0.601 & 0.047 & 0.044 &
0.022 + x @xmath1 8.7 @xmath2 & 0.137 & 0.699 & 4.738 & 0.523 & 4.616 & 1.237 + error & 0.077 & 0.050 & 0.560 & 0.068 & 0.046 & 0.025 + + x @xmath1 9.5 @xmath2 & 0.484 & 0.642 & 4.342 & 0.937 & 4.597 & 1.219 + error & 0.102 & 0.066 & 0.487 & 0.098 & 0.019 &
0.020 + x @xmath1 8.7 @xmath2 & 0.462 & 0.648 & 4.293 & 0.916 & 4.595 & 1.213 + error & 0.106 & 0.067 & 0.484 & 0.103 & 0.016 & 0.020 + x @xmath1 9.5 @xmath2 & 114% & & 51% & 24% +
x @xmath1 8.7 @xmath2 & 112% & & 51% & 24% + + x @xmath1 9.5 @xmath2 & 99% & & 36% & 8% + x @xmath1 8.7 @xmath2 & 90% & & 37% & 7% + + x @xmath1 9.5 @xmath2 & 90% & & 53% & 14% + x @xmath1 8.7 @xmath2 & 84% & & 55% & 12% + + x @xmath1 9.5 @xmath2 & 121% & & 76% & 24% + x @xmath1 8.7 @xmath2 & 120% & & 77% & 24% + + x @xmath1 9.5 @xmath2 & 73% & & 54% & 16% + x @xmath1 8.7 @xmath2 & 73% & & 54% & 16% + + x @xmath1 9.5 @xmath2 & 85% & & 32% & 18% + x @xmath1 8.7 @xmath2 & 85% & & 35% & 16% + + x @xmath1 9.5 @xmath2 & 120% & & 36% & 14% + x @xmath1 8.7 @xmath2 & 119% & & 36% & 12% + + x @xmath1 9.5 @xmath2 & 78% & & 41% & 16% + x @xmath1 8.7 @xmath2 & 77% & & 42% & 15% + + x @xmath1 9.5 @xmath2 & 60% & & 27% & 12% + x @xmath1 8.7 @xmath2 & 57% & & 28% & 10% + | we present extinction curves that include data down to far ultraviolet wavelengths ( fuv ; 1050 1200 ) for nine galactic sight lines .
the fuv extinction was measured using data from the _ far ultraviolet spectroscopic explorer_. the sight lines were chosen for their unusual extinction properties in the infrared through the ultraviolet ; that they probe a wide range of dust environments is evidenced by the large spread in their measured ratios of total - to - selective extinction , @xmath0 = 2.43 3.81 .
we find that extrapolation of the fitzpatrick & massa relationship from the ultraviolet appears to be a good predictor of the fuv extinction behavior .
we find that predictions of the fuv extinction based upon the cardelli , clayton & mathis ( ccm ) dependence on @xmath0 give mixed results .
for the seven extinction curves well represented by ccm in the infrared through ultraviolet ( x @xmath1 8 @xmath2 ) , the fuv extinction is well predicted in three sight lines , over - predicted in two sight lines , and under - predicted in 2 sight lines . a maximum entropy method analysis using a simple three component grain model shows that seven of the nine sight lines in the study require a larger fraction of grain materials to be in dust when fuv extinction is included in the models .
most of the added grain material is in the form of small ( radii @xmath3 200 ) grains . |
this is the third paper of a series whose purpose is to prepare for the projected final summary review of our 1991 - 2001 hubble space telescope experiment to determine the hubble constant by calibrating the absolute magnitude of type ia supernovae .
the method used in this observing program has been to measure cepheid distances to the host galaxies by ( 1 ) correcting the raw cepheid photometric data for reddening to determine the true distance moduli of the parent galaxies using an adopted cepheid period - luminosity relation and ( 2 ) correcting the sne data for their own intrinsic and extrinsic reddening and absorption values , plus all known second - parameter corrections for light - curve inhomogeneities .
the first two papers of this preparatory series concerned revisions in the cepheid p - l relation and its lack of universality from galaxy to galaxy , showing significant differences in the p - l slope and zero point between the galaxy @xcite and the lmc @xcite . following the emergence of sneia as powerful standard candles @xcite
, it became increasingly clear as the accuracy of the data became better that the dispersion in the sneia absolute magnitude at maximum became smaller with each advance in the accuracy of the data , and also with the application of corrections for light - curve inhomogeneities that had been found and subsequently systematically improved ( @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; also @xcite for a review ) .
the purpose of this third paper of the series is to update the discussion by @xcite on reddening , absorption , and second parameter corrections ( decline rate , color , galaxy type ) to the sneia light curves , and to determine a revised hubble diagram based on these corrections using the totality of the modern sneia photometry available to date .
the sample here consists of 124 supernovae compared with the sample of 35 sneia used by @xcite .
the organization of the paper is this : \(1 ) the photometric data for 124 sneia are compiled in [ sec : data ] from the extensive literature , together with the types and recession velocities of their host galaxies ( table [ tab : sn1 ] ) .
\(2 ) the _ intrinsic _ colors @xmath18 , @xmath19 ( i.e. 35 days after @xmath13 maximum ) , and @xmath20 are derived in [ sec : color ] from 34 normal sneia which have occurred in e / s0 galaxies or in the outer regions of spiral galaxies and which are assumed after correction for galactic reddening to suffer no reddening in their host galaxies .
\(3 ) after correction for galactic reddening , the reddening of the normal sneia due to the host galaxy is derived by three independent methods in
[ sec : hostgal ] , leading to the intrinsic ( unreddened ) colors of each of the sne .
these corrected data are listed in table [ tab : sn3]a .
\(4 ) new absorption - to - reddening ratios for that part of the reddening due to the host galaxy are derived in [ sec : absorption ] , leading to the absorption corrections to the observed sn apparent magnitudes ( also corrected for the k - term effect due to redshift ) .
the absorption - free magnitudes are also listed in table [ tab : sn3]a .
\(5 ) new dependencies of the fully corrected sn magnitudes on galaxy type , decline rate @xmath21 , and intrinsic color are derived in [ sec : other ] .
magnitudes @xmath22 , corrected for absorption and normalized to @xmath23 and @xmath24 complete table [ tab : sn3]a .
\(6 ) a discussion is made in
[ sec : snpec ] of the photometric properties of the small group ( 15% of the total sample ) of peculiar sneia , i.e. sne-91 t , sne-99aa , sne-91bg ( named after their eponymous prototypes ) , and two singular objects ( sn-2000cx , sn-1986 g ) , with the conclusion that sne-91 t , including sne-99aa , are brighter than normal sneia by @xmath25 and @xmath26 , respectively , even if normalized to @xmath23 , while sne-91bg are underluminous in @xmath13 by @xmath27 .
\(7 ) hubble diagrams using @xmath22 are derived for various subsamples of normal sneia in
[ sec : hubble ] , together with the rms magnitude deviations for each of the subsamples used .
the robustness of the solutions is emphasized .
\(8 ) twelve principal research points made in the paper are summarized in
[ sec : conclusions ] .
the conventions used here are these : colors @xmath28 are always @xmath29 .
@xmath30 stands for colors corrected for galactic reddening , @xmath31 for colors corrected for galactic reddening _ and _ reddening in the host galaxy .
absolute magnitudes @xmath32 and all listed distance moduli are calculated from the redshift using a flat universe with @xmath33 , @xmath34 with an arbitrarily assumed hubble constant of @xmath5 .
none of the conclusions as to magnitude corrections and trends of the second - parameter corrections depend on @xmath35 , hence there is no loss of generality using any assumed value .
a list of 19 `` local '' ( @xmath36 ) and 94 `` nearby '' ( @xmath37 ) supernovae of type ia ( sneia ) after 1985 with available maximum magnitudes in @xmath13 and @xmath14 , and , if possible in @xmath15 , is given in table [ tab : sn1 ] .
the list aims to be complete as to published template - fitted maximum magnitudes in @xmath13 and @xmath14 , however the few nearby sneia with not so well observed lightcurves by @xcite are not considered .
eleven earlier sneia ( 9 `` local '' and 2 `` nearby '' ) with useful photometry are added .
table [ tab : sn1 ] is subdivided into a ) 111 spectroscopically `` normal '' sneia in the sense of @xcite including objects which are not spectroscopically confirmed but share the photometric properties of normal sneia ( which is assumed here in first approximation also for the seven 1999aa - like objects ) ; b ) the four sne-91 t which are spectroscopically peculiar at early phases @xcite ; in addition included is the unique object sn2000cx whose spectrum resembles sn1991 t at least near maximum phase @xcite ; and c ) the seven sne-91bg objects which constitute a homogeneous sub - class with very fast decline rates ( @xmath38 ) , and which are quite red and underluminous at maximum .
also included here is sn1986 g ; it has , however , a faster decline rate and much of its apparent redness is probably due to absorption in the host galaxy .
table [ tab : sn1 ] is organized as follows : + col . 1 : the name of the sn + col . 2 : the name of the host galaxy + col . 3 : the coded type of the host galaxy : e@xmath39 , e / s0@xmath40 , s0@xmath41 , s0/a@xmath42 , the later types as in @xcite . the galaxy types are taken from the same sources as the photometry , in some cases from @xcite .
+ col . 4 : the mean heliocentric radial velocity as given in the nasa extragalactic database ( ned ) .
+ col . 5 : the radial velocity corrected for streaming motions .
the nearer galaxies are corrected for a self - consistent virgocentric infall model with a local infall vector of @xmath43 @xcite .
the more distant galaxies are corrected for the cmb motion , where it is assumed that the co - moving volume extends to @xmath44 , justified by the kinematics of the `` local bubble '' that merges into the background field kinematics near @xmath45 @xcite . even if merging takes place as far out as @xmath46 @xcite it has no noticeable effect on the present results .
if @xmath47 , the value of @xmath48 is listed ( denoted with v ) .
all larger velocities are in the cmb frame .
virgo and fornax cluster members are listed with the mean cluster velocity ( denoted with v and f ) .
also two galaxies in the w - cloud ( ngc4527 , 4536 ) outside the virgo cluster proper , are tentatively assigned the cluster velocity .
6 - 8 : observed @xmath49 , @xmath50 , and @xmath51 magnitudes as compiled from @xcite , @xcite , and @xcite .
the latter has published the @xmath14 and @xmath15 magnitude at the epoch of the @xmath13 maximum ; they were transformed to @xmath50 and @xmath51 by interpolating hamuy s et al .
( 1996c ) template light curves ( their table 10 ) with the appropriate decline rate @xmath21 .
@xcite have published only @xmath13 and @xmath14 maxima , their @xmath15 photometry was reduced by @xcite by the template method of @xcite which is also the procedure that provides the k - corrections for the effects of redshift on the photometry and which is the basis of all magnitudes in this paper . where available the @xmath49 were taken from @xcite who have uniformly reduced the available data including various other sources .
the corresponding @xmath50 are not published , but were kindly provided by dr .
g. altavilla .
several additional sources are quoted in column 14 .
+ col . 9 : the decline rate @xmath21 taken from the same sources as the apparent maximum magnitudes .
the slight dependence of @xmath52 on the reddening @xcite is neglected .
+ col . 10 : the galactic reddening following @xcite + col . 11 : the reddening @xmath53 in the host galaxy as judged from the @xmath54 color at maximum assuming intrinsic colors of @xmath18 as given in equation ( [ eq : color : mean : bv ] )
. + col .
12 : the reddening in the host galaxy at maximum as judged from the observed color @xmath19 35 days past @xmath13 maximum as explained in [ sec : hostgal : bvmax ] .
+ col . 13 : the reddening in the host galaxy as judged from the @xmath55 color at maximum assuming intrinsic colors @xmath20 as given in equation ( [ eq : color : mean : vi ] )
. + col . 14 : the source(s ) of the photometry
it is expected that sne in early - type galaxies ( s0 and earlier ) as well as far outlying sne in spiral galaxies suffer negligible reddening . in the case of sneia in spiral galaxies we require that their distance @xmath56 from the center of the host galaxy be larger than @xmath57 , where @xmath58 is the isophote corresponding to @xmath59 per arcsec@xmath60 ( cf . * ? ? ?
a complete list of 34 such sneia is given in table [ tab : color : minred ] ( omitting only the very red sn1974 g ) .
their colors @xmath54 , @xmath61 , i.e. the color 35 days past @xmath13 maximum from
[ sec : hostgal : bvmax ] below , and @xmath55 , all corrected for galactic reddening ( taken from column 10 of table [ tab : sn1 ] ) , are listed in table [ tab : color : minred ] . for the derivation of @xmath55
it was assumed that @xmath62 .
the distribution of the colors is shown in figure [ fig : color : tab2 ] , their dependence on the decline rate @xmath21 in figure [ fig : color : tab2dm15 ] .
lrrrrrrr + 1990af & @xmath63 & @xmath64 & @xmath65 & @xmath66 & @xmath67 & & @xmath68 + 1992j & @xmath69 & @xmath70 & @xmath71 & @xmath72 & @xmath73 & @xmath74 & @xmath75 + 1992ae & @xmath76 & @xmath77 & @xmath78 & @xmath79 & @xmath80 & & @xmath81 + 1992au & @xmath76 & @xmath82 & @xmath83 & @xmath84 & @xmath85 & @xmath86 & @xmath75 + 1992bk & @xmath76 & @xmath87 & @xmath88 & @xmath89 & @xmath90 & @xmath91 & @xmath92 + 1992bo & @xmath69 & @xmath93 & @xmath94 & @xmath95 & @xmath96 & @xmath97 & @xmath75 + 1992bp & @xmath69 & @xmath98 & @xmath99 & @xmath100 & @xmath101 & @xmath102 & @xmath103 + 1992br & @xmath76 & @xmath64 & @xmath104 & @xmath79 & @xmath105 & & @xmath75 + 1993o & @xmath69 & @xmath87 & @xmath106 & @xmath107 & @xmath108 & @xmath109 & @xmath110 + 1993ac & @xmath76 & @xmath111 & @xmath112 & @xmath113 & @xmath114 & @xmath115 & @xmath116 + 1993ae & @xmath76 & @xmath117 & @xmath118 & @xmath119 & @xmath120 & @xmath121 & @xmath122 + 1993ag & @xmath69 & @xmath123 & @xmath124 & @xmath125 & @xmath126 & @xmath127 & @xmath81 + 1993ah & @xmath63 & @xmath128 & @xmath129 & @xmath130 & @xmath131 & @xmath132 & @xmath133 + 1994d & @xmath63 & @xmath134 & @xmath135 & @xmath136 & @xmath137 & @xmath109 & @xmath138 + 1994 m & @xmath76 & @xmath139 & @xmath140 & @xmath141 & @xmath142 & @xmath143 & @xmath133 + 1994q & @xmath63 & @xmath77 & @xmath83 & @xmath72 & @xmath144 & @xmath145 & @xmath146 + 1996x & @xmath76 & @xmath139 & @xmath99 & @xmath147 & @xmath148 & @xmath149 & @xmath150 + 1997e & @xmath63 & @xmath151 & @xmath152 & @xmath153 & @xmath101 & @xmath154 & @xmath155 + 1998dx & @xmath76 & @xmath156 & @xmath157 & @xmath158 & & @xmath159 & @xmath160 + 1999gh & @xmath76 & @xmath161 & @xmath162 & @xmath163 & @xmath164 & @xmath165 & @xmath75 + 2000b & @xmath76 & @xmath111 & @xmath166 & @xmath167 & @xmath168 & @xmath169 & @xmath170 + 2000dk & @xmath76 & @xmath64 & @xmath171 & @xmath172 & @xmath173 & @xmath174 & @xmath175 + + 1972e & am & @xmath98 & @xmath176 & @xmath177 & @xmath67 & @xmath178 & @xmath179 + 1990n & @xmath180 & @xmath93 & @xmath104 & @xmath181 & @xmath182 & @xmath183 & @xmath179 + 1990o & @xmath184 & @xmath77 & @xmath185 & @xmath186 & @xmath168 & @xmath187 & @xmath188 + 1990 t & @xmath184 & @xmath134 & @xmath106 & @xmath189 & @xmath190 & @xmath191 & @xmath192 + 1991s & @xmath180 & @xmath193 & @xmath104 & @xmath194 & @xmath195 & @xmath196 & @xmath197 + 1991ag & @xmath180 & @xmath198 & @xmath199 & @xmath200 & @xmath201 & @xmath202 & @xmath203 + 1992a & @xmath204 & @xmath205 & @xmath206 & @xmath207 & @xmath105 & @xmath208 & @xmath122 + 1992bc & @xmath209 & @xmath210 & @xmath135 & @xmath211 & @xmath212 & @xmath213 & @xmath146 + 1992bl & @xmath204 & @xmath156 & @xmath214 & @xmath215 & @xmath216 & @xmath217 & @xmath218 + 1995d & @xmath204 & @xmath82 & @xmath162 & @xmath219 & @xmath220 & @xmath221 & @xmath222 + 1998v & @xmath180 & @xmath111 & @xmath223 & @xmath224 & @xmath225 & @xmath226 & @xmath227 + 1998eg & @xmath228 & @xmath70 & @xmath229 & @xmath230 & & @xmath231 & @xmath232 + & @xmath186 & @xmath114 & @xmath233 & + & @xmath234&@xmath235&@xmath236 & + & @xmath237 & @xmath238 & @xmath239 & + since the sne in table [ tab : color : minred ] are assumed to suffer no reddening in their host galaxies , their galactic absorption - corrected colors @xmath30 are equal to the colors @xmath31 corrected for galactic _ and _ host galaxy reddening .
the assumption of zero reddening is supported by several arguments .
( a ) if one assumes that the average error of @xmath49 and @xmath50 is @xmath240 , then the observed scatter of @xmath241 in table [ tab : color : minred ] is fully explained by these errors . in that case
the distribution of @xmath18 in figure [ fig : color : tab2]a must be gaussian , which is statistically acceptable in spite of the four very blue sneia with @xmath242 .
( b ) as figure [ fig : color : tab2 ] shows , the color distribution of the sneia in table [ tab : color : minred ] is skewed bluewards , if anything , while any noticeable reddening would skew the distribution towards red colors .
( c ) the colors @xmath243 and @xmath244 of the sneia in table [ tab : color : minred ] are _ uncorrelated _ ( in fact insignificantly anti - correlated ) ; in case of absorption in the host galaxies both values would be redder than average .
( d ) if the spread in color in figure [ fig : color : tab2]a was due to absorption , the absolute magnitudes @xmath245 would correlate with the color @xmath246 from column 5 with a slope of @xmath247 , whereas an insignificant slope of @xmath248 is observed .
the least squares solutions of the data in figure [ fig : color : tab2dm15 ] are @xmath249 @xmath250 @xmath251 the dependence of @xmath243 on @xmath21 in equation ( [ eq : color : mean : bv ] ) is weaker than suggested by @xcite , @xcite , and @xcite .
in fact the dependence is only marginally significant . yet
it is adopted here at face value .
the similarity of snia spectra at @xmath252 and the similar slopes of the color curves at this late phase have led @xcite , @xcite , @xcite , and @xcite to assume that @xmath19 is independent of @xmath21 .
however , equation ( [ eq : color : mean : bv35 ] ) shows this color to become _ bluer _ with increasing @xmath21 at a significance of @xmath253 . the @xmath244 colors exhibit in equation ( [ eq : color : mean : vi ] ) a clear dependence on @xmath52 , becoming _
redder _ with increasing decline rate .
the mean intrinsic colors of normal sneia , normalized to @xmath254 , become from equations ( [ eq : color : mean : bv]@xmath255[eq : color : mean : vi ] ) @xmath256 having argued that sneia in table [ tab : color : minred ] are essentially redding - free , it is nevertheless clear that their reddening is not exactly zero .
their true colors must therefore be slightly bluer than expressed in equation ( [ eq : color : mean ] ) .
we will return to this point in [ sec : conclusions ] where it is stressed that any rest reddening is inconsequential for the use of sneia as standard candles as long as they are reduced to some _ uniform _ color .
the reddening @xmath1 of sneia at maximum phase due to selective absorption in the host galaxy is determined in three different ways .
the reddening in the host galaxy is given by @xmath257 @xmath53 follows for all sneia in table [ tab : sn1]a and tentatively for the peculiar sneia in table [ tab : sn1]b by inserting the apparent magnitudes and the galactic reddening from columns 6 , 7 , and 10 in table [ tab : sn1]a , b and by taking the intrinsic color from equation ( [ eq : color : mean : bv ] ) .
the resulting @xmath53 are given in column 11 of table [ tab : sn1]a & b. color excesses @xmath258 at phase @xmath259days were kindly provided by s. jha for 59 sneia in table [ tab : sn1]a .
he has based them on an adopted intrinsic tail color of @xmath260 for all sneia .
additional tail color excesses were published by @xcite and @xcite .
they are in a slightly different system as those by jha . in a first step
the altavilla excesses were transformed into jha s system by the regression @xmath261 in a second step the phillips excesses were transformed by comparison with the joint set from jha and altavilla by means of @xmath262 the homogenized excesses @xmath258 of sneia with more than one determination were averaged .
if one adds to these values jha s adopted intrinsic color of @xmath260 , one recovers the `` observed '' tail colors @xmath61 , not listed , for 105 of the sneia .
new color excesses @xmath258 are now determined by making allowance for the @xmath21 dependence of the intrinsic color @xmath19 .
this is achieved by subtracting @xmath19 as given in equation ( [ eq : color : mean : bv35 ] ) from the `` observed '' colors @xmath61 .
the adopted tail excesses @xmath258 apply to the quite red phase at @xmath252 .
they can not directly be applied to the ( blue ) sn colors at maximum , because @xmath263 must depend on color due to bandwidth effects .
in fact , if one plots the colors @xmath264 ( corrected for galactic reddening and reduced to @xmath23 with equation ( [ eq : color : mean : bv ] ) ) against the adopted values of @xmath258 one obtains the following linear regression ( figure [ fig : color : e35 ] ) .
@xmath265 since the _ intrinsic _ color @xmath266 at maximum must not be a function of reddening it follows from equation ( [ eq : color : bvmax ] ) that @xmath267 the resulting values @xmath268 are listed in table [ tab : sn1]a , column 12 .
the excesses from equations ( [ eq : ebvmax ] & [ eq : color : ebvmax35 ] ) are compared in figure [ fig : color : comp1 ] .
the agreement is satisfactory with no systematic differences .
the excesses @xmath268 for four peculiar sneia in table [ tab : sn1]b are given in parentheses .
they are smaller on average than their excesses from @xmath269 , suggesting that they are redder at @xmath252 than normal sneia ( see also
[ sec : snpec : t ] ) .
the excesses @xmath268 of the sn1991bg - like sne in table [ tab : sn1]c have not been used , but they are consistent with the assumption that these sne in early - type galaxies are almost reddening - free . a third possibility to determine the reddening @xmath263 in the host galaxy is provided by the magnitudes @xmath50 and @xmath51 in table [ tab : sn1]a&b , columns 7 & 8 .
we note @xmath270 where @xmath271 is adopted for the galactic reddening .
if one inserts the intrinsic colors from equation ( [ eq : color : mean : vi ] ) , with the @xmath21-values appropriate for each snia , into equation ( [ eq : color : evihost ] ) one obtains @xmath272 .
since @xmath263 is required , not @xmath272 , the latter must be converted into the former .
this is done by means of figure [ fig : color : evihost ] , where @xmath273 ( as in figure [ fig : color : e35 ] ) is plotted against @xmath272 . from the equation at the bottom of figure [ fig : color : evihost ] follows @xmath274 the color excesses @xmath275 , listed in column 13 of table [ tab : sn1 ] , are compared with the mean of the excesses @xmath53 and @xmath268 in figure [ fig : color : comp2 ] . again the agreement is satisfactory . in view of the consistency of the color excesses @xmath53 , @xmath268 , and
@xmath275 they are averaged with equal weights .
the means are adopted as the best estimates of the reddenings in the host galaxy , @xmath1 .
the values of @xmath1 are listed in table [ tab : sn3 ] , column 2 .
the ensuing colors @xmath18 , fully corrected for galactic and host galaxy reddening , are in column 3 . table [ tab : sn3 ] is organized as follows : + col . 1 : the name of the sn . + col . 2 : the adopted reddening @xmath1 in the host galaxy being the mean of columns 11 - 13 of table [ tab : sn1 ] .
3&4 : the intrinsic colors @xmath18 and @xmath20 corrected for galactic reddening and reddening in the host galaxy .
+ col . 5 - 7 :
the apparent magnitudes @xmath276 at maximum , corrected for absorption in the galaxy and in the host galaxy .
+ col . 8 - 10 : the absorption - corrected apparent magnitudes @xmath277 reduced to a standard decline rate of @xmath23 and @xmath24 by the precepts developed in [ sec : other : corrected ] . + col . 11 : the photometric absorption - free distance modulus @xmath278 from @xmath279 in column 9 and the mean luminosity @xmath280 in equation ( [ eq : other : mabs ] )
. the latter assumes @xmath281 . in order to free the @xmath55 colors from the host galaxy reddening the relation between @xmath282 and @xmath1
is needed .
it is obtained from figure [ fig : color : evibvhost ] , where @xmath55 is plotted versus @xmath1 . the figure yields ( slightly rounded ) @xmath283 which we adopt .
the conversion factor is consistent with equation ( [ eq : color : ebvvihost ] ) , but here @xmath1 has high weight being the mean of three determinations .
the rms dispersion of @xmath1 , if based on the three determinations , is @xmath284mag .
the template - fitting of the sn light curves by the method of @xcite is applied in a way as to compensate the k - correction , i.e. the effect of the redshifts @xmath285 on the magnitudes and colors , in a single step . to verify the success of the method we plot the adopted colors @xmath286 and @xmath287 , normalized to @xmath23 , against @xmath288 in figure [ fig : color : kcorr ] .
any dependence is hardly significant and is neglected in the following .
the galactic reddenings of @xcite are converted into absorption values @xmath289 , @xmath290 , and @xmath291 by means of conventional absorption - reddening ratios of @xmath292 , @xmath293 , and @xmath294 ( hence @xmath295 as used in
[ sec : color ] ) .
the resulting absorption values are subtracted from the observed sn magnitudes to obtain @xmath296 , @xmath297 , and @xmath298 at maximum .
conventional values of @xmath299 have been adopted because the available range of @xmath300 is too small to derive meaningful galactic @xmath299 values from the sneia themselves .
previous attempts to correct sn magnitudes for absorption in the host galaxies have frequently assumed conventional ( galactic ) values of @xmath301 , although ( * ? ? ?
* and references therein ) have argued for significantly smaller values .
more recent determinations are ( quoted @xmath302 values are transformed here to @xmath303 by @xmath304 ) @xmath305 @xcite , @xmath306 @xcite , @xmath307 @xcite , @xmath308 @xcite , @xmath309 @xcite , @xmath310 @xcite .
the present large sample of sneia is well suited for an independent determination of the average value of @xmath301 of extragalactic sneia . for this purpose the absolute magnitudes @xmath311 , based on @xmath281 and corrected only for galactic absorption ,
are plotted against the values of the host galaxy reddening @xmath312 as listed in column 13 of table [ tab : sn1 ] ( figure [ fig : absorption : mabs ] ) .
the sn sample used here is called in the following the `` fiducial '' sample .
it contains the 62(58 ) sneia in table [ tab : sn1]a with @xmath13 , @xmath14 ( and @xmath15 ) magnitudes and with @xmath313 .
it excludes sn1996ai which suffers excessive absorption in its host galaxy ( see table [ tab : sn1]a ) .
it also excludes five sneia , which deviate by more than @xmath314 .
the apparently normal sn1999ej is too faint for unknown reasons by @xmath315 .
sn1992ag is very red at @xmath316 , causing its tail excess @xmath317 to be large , which in turn makes @xmath243 exceptionally blue and the absolute magnitude very bright ( cf .
table [ tab : sn1]a and [ tab : sn3 ] ) .
it is believed that either its color @xmath318 is in error or its light curve is peculiar .
if the color excess was based only on the color at maximum its absolute magnitude would become normal .
the overly bright sn1997cw is suspicious not so much because its photometry begins only 16 days after @xmath13 maximum , but because it was classified as sne-91 t @xcite , which is now revised to sne-91t/99aa @xcite .
also the very luminous sn1998ab and 1999dq are of type sne-99aa @xcite .
the properties of this new class of sneia and their overluminosity are taken up again in
[ sec : snpec : t ] .
four other less overluminous sne-99aa , designated with a dagger @xmath319 in table [ tab : sn1]a , are kept in the sample of normal sneia .
the reason is that they can be isolated only from pre - maximum spectroscopy , which is lacking for many objects in table [ tab : sn1]a .
thus a few of the `` normal '' sneia may actually belong to the elusive class of sne-99aa without having as extreme luminosities like sn1997cw , 1998ab , and 1999dq . as shown by the equations in figure [ fig : absorption : mabs ] the resulting values of @xmath320 are about 3.6 , 2.4 , and 1.2
. however , these values are still open to criticism . as will be shown below sneia in e / s0 have lower intrinsic luminosities _ and _ less extinction than average .
therefore there is some dependence of @xmath311 on @xmath263 which can not be attributed to absorption , but which is due to the intrinsic underluminosity of sneia in early - type galaxies .
this type - dependent effect can be compensated by the decline rate @xmath21 in [ sec : other : lightcurve ] .
the interplay of @xmath299 and @xmath21 requires therefore a simultaneous solution which is given in equation ( [ eq : other : dm15r ] ) below . from this
the improved values of @xmath301 become @xmath321 the @xmath299 values here are in addition adjusted to fulfill the condition @xmath322 .
they are quite close to the provisional values in figure [ fig : absorption : mabs ] .
it is also satisfactory that they agree quite well with those by other authors cited above . if one assumes that the highly reddened sn1996ai is a normal sneia with average luminosity and that its peculiar motion is @xmath323 , one finds @xmath324 from this single object . the solutions in table [ tab : other : dm15r ] for @xmath302 and @xmath325 imply a ratio @xmath326 , which may be compared with @xmath327 from equation ( [ eq : color : evibvhost ] ) .
the small error of the latter value is due to the fact that its determination is independent of any assumed value of @xmath299 , which is notoriously difficult to determine .
it is noted that the adopted value of @xmath81 is the same as the conventional galactic value of 1.3 .
the values of @xmath1 ( table [ tab : sn3 ] , column 2 ) were multiplied with the @xmath301 values in equation ( [ eq : absorption : r ] ) to be subtracted from the apparent magnitudes @xmath13 , @xmath14 , and @xmath15 ( table [ tab : sn1 ] , columns 6 - 8 ) , after they are corrected for the conventional galactic absorption .
the resulting fully corrected apparent magnitudes @xmath328 , @xmath329 , and @xmath330 are tabulated in columns 5 - 7 of table [ tab : sn3 ] .
if one plots the absolute magnitudes @xmath331 as introduced in [ sec : absorption ] , but now also corrected for absorption in the host galaxy and _ without _ the @xmath21 correction , against the hubble type of the host galaxy , a clear correlation emerges between sn luminosity and the hubble type as first shown by @xcite , the sneia in early - type galaxies being fainter by @xmath332 than in late spirals ( figure [ fig : other : type ] ) .
the obvious conclusion is that young populations produce brighter sneia .
yes , but why ?
the most secure conclusion from the data , which are beyond doubt ( * ? ? ? * ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
6 ) , is that different channels of sneia production exist between most sneia in e / s0 galaxies compared with most sne ia in spirals .
both channels are fed by old , highly evolved stars near one solar mass that are in close binaries . by one of two processes ( mass transfer from a normal star to a white dwarf , or mergers of two white dwarfs due to orbital energy loss caused by gravitational radiation nudging one or both stars over the chandrasekhar limit ) ,
catastrophic collapse eventually occurs , but with different gestation times .
hence , the delay times differ for progenitor formation to explosion , according to the process .
the consequence is that the star formation rates of the initial progenitors differ between e / s0 galaxies and spirals ( e / s0 with a single burst of star formation early on ; spirals with a more continuous star formation rate ) .
hence , the average delay times differ between e / s0 and spirals ( l. greggio , private communication ) , making a difference in the fe production rate , hence a chemical composition difference .
this expected chemical composition difference apparently affects the absolute magnitude at maximum in the sn explosion , as suggested by @xcite .
since also the decline rate is a function of the hubble type , one expects a correlation between the luminosity and the decline rate as well .
this pskovskii effect was first quantified by the decline rate parameter @xmath21 by @xcite .
the corresponding correlation of the form @xmath333 is in fact even tighter than the one in figure [ fig : other : type ] ; it is shown in figure [ fig : other : dm15 ] .
the normalization to @xmath334 corresponds roughly to the median value of normal sneia .
the magnitudes @xmath335 are corrected for galactic and internal absorption described in [ sec : absorption ] . since @xmath336 and @xmath299 and hence @xmath337 are not quite independent , as discussed in
[ sec : absorption : host ] , these quantities must be simultaneously solved for .
the corresponding regression with two independent variables takes the form of @xmath338 where the constant term , @xmath339 , is the mean absolute magnitude of sneia , corrected for galactic and host galaxy absorption and reduced to @xmath23 .
least - squares solutions for the same fiducial sn sample as used for figure [ fig : absorption : mabs ] give the values of @xmath340 , @xmath301 , and @xmath339 as shown in table [ tab : other : dm15r ] .
lccc @xmath336 & @xmath341 & @xmath342 & @xmath343 + @xmath299 & @xmath344 & @xmath345 & @xmath346 + @xmath347 & @xmath348 & @xmath349 & @xmath350 + @xmath351 & @xmath352 & @xmath352 & @xmath352 + @xmath353 & 62 & 62 & 58 + the values of @xmath340 in table [ tab : other : dm15r ] will be adopted in the following .
the dependence of @xmath354 on @xmath52 is illustrated in figure [ fig : other : dm15 ] , where @xmath340 appears as the slope of the correlation .
the values of @xmath301 in table [ tab : other : dm15r ] differ slightly from the values in equation ( [ eq : absorption : r ] ) , which we have adopted as the final ones , because the latter comply with the additional conditions that @xmath322 and @xmath355 from equation ( [ eq : color : evibvhost ] ) . in equation ( [ eq : color : mean : bv ] ) it was shown that the color @xmath18 is only marginally dependent on @xmath52 .
this is vindicated here by the @xmath336 s in table [ tab : other : dm15r ] .
they yield for the relation @xmath356 an insignificant value of @xmath357 .
the corresponding coefficient for the @xmath20 colors becomes @xmath358 , which is consistent with @xmath359 from equation ( [ eq : color : mean : vi ] ) .
it becomes clear from the above that any change of @xmath360 affects necessarily the coefficients @xmath336 , because an excess @xmath361 causes a change of the color excess of @xmath362 this affects the absorption - corrected magnitudes @xmath363 , taken as an example , by @xmath364 hence , if the _ intrinsic _ dependence of @xmath245 on @xmath52 is @xmath365 , one observes instead @xmath366 the analogue holds for the color excesses @xmath258 derived from tail colors @xmath318 .
if the true colors vary like @xmath367 , any deviation @xmath368 from @xmath369 leads to a variation of the color excess of @xmath370 such that @xmath371 , being the mean of @xmath372 and @xmath373 , varies like @xmath374 @xmath375\delta m_{15 } = ( \alpha_{b } + \delta\alpha)\delta m_{15}.\ ] ] we have adopted , besides @xmath247 , @xmath376 ( equation [ eq : color : mean : bv ] ) , @xmath377 ( equation [ eq : color : mean : bv35 ] ) as the intrinsic values . yet
some authors have adopted other values of @xmath299 , @xmath360 and @xmath369 which necessarily reflect on the observed value of @xmath336 .
for instance @xcite have chosen @xmath378 , @xmath379 , and @xmath380 .
if one subtracts the resulting @xmath381 from their proposed value @xmath382 one obtaines @xmath383 . @xcite
have taken @xmath384 , @xmath385 , and @xmath386 ; this leads to @xmath387 which , subtracted from their preferred value @xmath388 , yields @xmath389 .
these two determinations of @xmath390 embrace the adopted value of @xmath391 in table [ tab : coeff3fit ] within statistics .
if we had set the coefficient @xmath392 in equation ( [ eq : color : mean : bv ] ) equal to 0 , as @xcite did , the present fiducial sample would give @xmath393 which is the same as their value within the errors .
the correlation of type - ia sn luminosity and light curve shape has a considerably history .
first suggested by @xcite it was taken up by @xcite .
@xcite used first the decline rate @xmath21 to parametrize the correlation , however he had to depend on inaccurate tully - fisher and surface brightness fluctuation distances , which resulted in an overestimate of the slope @xmath390 by a factor of @xmath394 .
the steep slope was the reason why @xcite questioned the correlation altogether . if phillips had been correct in the steepness of the pskovskii effect , implying a luminosity variation of @xmath395mag in @xmath13 over the full range of @xmath21 , the use of sneia as standard candles would have been substantially compromised .
only when sufficient relative velocity distances became available the determination of the slope @xmath336 began to converge @xcite .
the remaining differences are now explained by the sensitivity of @xmath336 on the adopted relation between color and @xmath21 ( equations [ eq : color : mean : bv]@xmath255[eq : color : mean : vi ] )
. from the above it is clear that a positive / negative excess of the coefficient @xmath360 in equation ( [ eq : color : mean : bv ] ) propagates into @xmath1 and into the absorption values such that the magnitudes @xmath396 ( or @xmath397 ) are too faint / bright by @xmath398 .
yet it is to be noted that the magnitudes @xmath399 , normalized to @xmath254 , are not affected by this error , because they are corrected by an additional term @xmath400 of opposite sign .
the error compensation is only exact if @xmath1 is determined from the intrinsic colors given in equation ( [ eq : color : mean : bv ] ) .
in the present case also the intrinsic colors from equations ( [ eq : color : mean : bv35 ] & [ eq : color : mean : vi ] ) are used to determine @xmath1 , but the three determinations agree so well that it is sufficient here to show that any effect of an incorrect value of @xmath360 on @xmath399 is compensated by the adopted value of @xmath336 in table [ tab : other : dm15r ] .
@xcite have introduced the `` stretch factor '' , @xmath401 , instead of @xmath21 to characterize the light curve shape .
the two parameters @xmath401 and @xmath21 show a tight correlation @xcite ; they are interchangeable for all practical purposes .
an alternative method to allow for the correlation between luminosity and light curve shape and to determine at the same time the total absorption @xmath290 was introduced by ( * ? ? ?
* ; * ? ? ?
* see also @xcite ; @xcite ; @xcite ) .
their various forms of `` multi - color light curve shape '' ( mlcs ) fitting are difficult to apply and offer no obvious advantages . following a proposal by @xcite and
@xcite we test the question whether sn luminosity depends on the color @xmath18 and/or @xmath20 in figure [ fig : other : color ] .
dependencies at the @xmath402 level seem to exist of @xmath403 and @xmath404 on @xmath18 and of @xmath404 on @xmath20 ( see equations at the panel bottoms of figure [ fig : other : color ] ) .
the correlations should be considered , however , with some reservations because the range of about @xmath405mag in either color is only about two times the estimated observational error of a single color determination ( about @xmath406mag ) . in the following the correlation of sn luminosity on @xmath18
is taken at face value .
@xmath407 colors are preferred over @xmath408 colors because the latter are not available for all sneia of the fiducial sample .
the allowance for the color dependence does not shift the mean luminosity of sneia , but brings a marginal decrease of the luminosity dispersion , at least in @xmath15 . in any case the allowance for color can bring no harm . from the discussion in
[ sec : other : lightcurve ] and [ sec : other : color ] follows that minimum scatter of the absolute magnitude @xmath335 can be achieved by correcting for the decline rate @xmath21 _ and _ color @xmath243 .
the corresponding ansatz @xmath409 + c\ ] ] leads with the fiducial sample of 62 sneia to the coefficients @xmath410 , @xmath411 , and @xmath412 in table [ tab : coeff3fit ] .
lrrrrr + @xmath413 : & @xmath414 & @xmath415 & @xmath348 & @xmath352 & 62 + @xmath416 : & @xmath414 & @xmath417 & @xmath349 & @xmath352 & 62 + @xmath418 : & @xmath419 & @xmath420 & @xmath421 & @xmath422 & 58 + if one subtracts the first two terms of equation ( [ eq : other:3fitd ] ) from the fully absorption - corrected magnitudes @xmath276 in columns 5 - 7 table [ tab : sn3 ] one obtains the magnitudes @xmath277 ( columns 8 - 10 ) which are in addition reduced to @xmath23 and a corresponding color of @xmath24 .
the distribution functions of the resulting absolute magnitudes @xmath423 of the fiducial sample always on the basis of @xmath281 are shown in figure [ fig : mabshisto ] .
they are well fitted by gaussians after five sneia , discussed in
[ sec : absorption : host ] , are excluded .
sne-91 t , with sn1991 t as the prototype and long suspected to be overluminous , are spectroscopically well defined ( e.g. * ? ? ? * and references therein ) .
the observed and derived photometric properties of the known quartet are listed in table [ tab : sn1]b and [ tab : sn3]b , respectively .
they are characterized by quite slow decline rates @xmath21 .
their intrinsic color @xmath18 , being unknown , was _ assumed _ to be given by equation ( [ eq : color : mean : bv ] ) like that of normal sneia .
if one derives color excesses @xmath312 and absorption values on this basis , one finds the mean photometric parameters given in table [ tab : sneia - t ] .
the fact that the resulting mean value of @xmath424 is similar to normal sneia shows that the present assumption on @xmath243 leads at least to consistent results .
their absolute magnitudes @xmath425 , as given in table [ tab : sneia - t ] , turn out to be brighter than normal sneia ( equation [ eq : other : mabs ] below ) by @xmath426 on average with remarkably small scatter .
even if one reduces their luminosity to @xmath23 ( equation [ eq : absorption : dm15 ] ) they remain overluminous by @xmath427 . this conclusion
could only be avoided by postulating that sne-91 t were very blue at maximum , i.e.@xmath428 .
this is , however , not supported by interstellar absorption lines which give @xmath429 to @xmath430 for sn1991 t @xcite and @xmath431 for sn1997br @xcite .
lcrrrrrr sne-91 t & mean & @xmath432 & @xmath156 & @xmath433 & @xmath434 & @xmath435 & @xmath436 + & & @xmath437 & & @xmath438 & @xmath439 & @xmath439 & @xmath440 + & n & 4 & 4 & 3 & 4 & 4 & 3 + + sne-99aa & mean & 0.92 & @xmath441 & @xmath442 & @xmath443 & @xmath444 & @xmath445 + & & @xmath437 & @xmath446 & @xmath447 & @xmath239 & @xmath448 & @xmath448 + & n & 7 & 7 & 7 & 7 & 7 & 7 + + sn2000cx & & @xmath449 & @xmath156 & @xmath450 & @xmath451 & @xmath452 & @xmath453 + & & or & @xmath454 & @xmath455 & @xmath456 & @xmath457 & @xmath458 + sne-91 t are rare events .
four representatives out of a total of 124 corresponds to a relative frequency of about @xmath459 .
a few unrecognized sne-91 t may have to be added , but on the other hand their presumed overluminosity enhances their discovery chance , which is further enhanced by their slow decline rates . hence a rate of roughly @xmath459 seems realistic .
@xcite have noted in the pre - maximum spectra of some sneia features typical for normal sneia and others typical for sne-91 t ( see also * ? ? ?
* ) . at maximum and
later they have normal spectra .
the prototype is sn1999aa ; the class is here designated as sne-99aa .
seven sneia in table [ tab : sn1 ] were classified as sne-99aa by @xcite ; they are marked with a dagger .
if one determines their host galaxy extinction on the assumption that they have @xmath407 colors at maximum and at @xmath252 like normal sneia , their average color @xmath244 becomes roughly the same as for normal sneia as seen in table [ tab : sneia - t ] .
since sne-99aa are spectroscopically a milder form of sne-91 t one expects that their luminosity lies between the latter and normal sneia .
this expectation is born out in table [ tab : sneia - t ] .
sne-99aa are brighter by only @xmath460 than normal sneia , which is further reduced to @xmath461 if normalized to @xmath23 .
if overluminosity and strikingly slow decline rates ( @xmath462 ) are typical for sne-99aa , it is noted that table [ tab : sn1]a contains 13 additional possible candidates .
at least three of them are normal sneia ( sn1997bp , @xcite and references therein ; sn1998ef , @xcite ; and sn2000e , @xcite ) , leaving 10 candidates at most .
if they are added to the 7 known cases , the total of known or unrecognized sne-99aa in table [ tab : sn1 ] is @xmath463 , i.e. less than estimated by @xcite
. their true frequency in a volume - limited sample may be even lower considering that their suggested overluminosity and their slow decline rates favor their discovery .
seven sneia in table [ tab : sn1]b ( excluding sn1986 g ) are of type sne-91bg with sn1991bg as the prototype .
they are characterized by peculiar spectra ( * ? ? ?
* and references therein ) , very red color , low luminosity and unparallelledly fast decline rates @xmath21 .
their rate of occurrence seems high in early - type galaxies .
this offers a handle on their intrinsic properties , because the four known sne-91bg in early - type galaxies may be assumed to have minimum reddening . since the remaining three sne-91bg in spirals are not significantly redder [ @xmath464 versus @xmath465 we assume that the internal reddening of this class can be neglected .
minimal reddening is also independently confirmed for sn1999bg in the spiral ngc2841 from late - phase multi - color photometry @xcite .
the ensuing mean photometric parameters are compiled in table [ tab : sneia - bg ] .
ccrrrrrrr mean & & @xmath466 & @xmath467 & @xmath468 & & @xmath469 & @xmath470 & @xmath471 + & & @xmath472 & @xmath438 & @xmath437 & & @xmath473 & @xmath474 & @xmath475 + @xmath476 & & @xmath477 & @xmath352 & @xmath478 & & @xmath479 & @xmath480 & @xmath481 + n & & 7 & 7 & 7 & & 7 & 7 & 7 + + & & @xmath75 & @xmath482 & & & @xmath483 & @xmath483 & + [ 0cm][0cm]86 g & & or & @xmath484 & & & @xmath485 & @xmath486 & + the sne-91bg emerge besides for their spectral peculiarities as a separate class of sne with quite narrowly confined photometric properties .
their decline rate varies little about @xmath487 , while no normal snia is known with @xmath488 .
their extremely red intrinsic colors have only moderate scatter .
in fact their @xmath244 shows less scatter than normal sneia .
their absolute magnitudes show somewhat more variation than normal sneia , but they are all strikingly underluminous by @xmath489 , @xmath490 , and @xmath491 on average .
the underluminosity may be somewhat reduced if we have underestimated the host galaxy absorption , but we deem values as large as @xmath492 ( and correspondingly less in @xmath14 and @xmath15 ) as unlikely .
models of two sn-91bg ( sn1991bg and sn1999by ) were elaborated by @xcite and by @xcite .
the frequency per unit volume of sne-91bg can roughly be estimated by noting that the seven known cases all lie within @xmath493 , and that table [ tab : sn1]a contains 47 sneia within this distance limit . of these ,
45 have @xmath494 and the remaining two ( sn1981d and 1984a ) are normal sneia on the basis of their colors .
hence seven cases out of a total of 54 correspond to a frequency of @xmath495 , or somewhat more because their fast decline rates impair their discovery probability ( see also * ? ? ?
since they occur preferentially in early - type galaxies , it is suggested that they stem from an older population @xcite . only two sneia in table [ tab : sn1 ]
are neither normal nor belong to the sne-91 t ( including sne-99aa ) and sne-91bg classes .
the spectrum of sn2000cx resembles sne-91 t at early phases , but its light curve is highly unusual , in fact unique @xcite .
the observed photometric parameters are given in table [ tab : sn1]b .
the intrinsic photometric properties at maximum are calculated here for two different assumptions .
a ) the sn has @xmath496 like normal sneia , or b ) the sn suffers zero reddening in its s0 host galaxy .
the ensuing colors and luminosities are shown in table [ tab : sneia - t ] . in either case
sn2000cx may be with @xmath497 one of the brightest sn in the entire sample .
@xcite have argued that such high luminosities , although still depending on the exact value of @xmath35 , require one chandrasekhar mass of pure @xmath498ni which can not be provided by a single white dwarf .
they proposed therefore a model with two merging white dwarfs .
sn2000cx may be a prime candidate for this process .
the spectrum of sn1986 g is similar to sne-91bg , but less extreme ( * ? ? ?
* and references therein ) .
its photometric observables are in table [ tab : sn1]c .
they show that sn1986 g was observed to be even redder than sne-91bg , but to have a clearly slower decline rate .
much of its redness is certainly due to its position in the dust lane of ngc5128 ( cena ) .
the intrinsic color and luminosity are calculated here for two different assumptions .
a ) @xmath18 of sn1986 g is given by equation ( [ eq : color : mean : bv ] ) like normal sneia , and b ) it has the tail color @xmath499 which , from equation ( [ eq : color : mean : bv35 ] ) , is the tail color of a normal sn with @xmath500 .
the results for both cases are shown in table [ tab : sneia - bg ] . in the latter case , with the tail color as standard
, the extinction @xmath501 becomes @xmath502 , which seems on the low side given its position in a prominent dust lane .
however , even the small reddening makes sn1986 g bluer than average sne-91bg and its relatively slow decline rate ( @xmath500 ) remains untypical for this class , although with the distance modulus of ngc5128 of @xmath503 @xcite its absolute magnitudes @xmath504 become quite similar to sne-91bg . on
the other hand case a ) leads to @xmath505 and @xmath506 ( including galactic reddening ) in good agreement with the independent estimates of @xcite and @xcite .
hence , we prefer solution a ) and conclude that the luminosity of sn1986 g lies inbetween normal sneia and sne-91bg ( see table [ tab : sneia - bg ] ) .
an overview of the luminosity distribution in @xmath13 of the various types of sneia is given in figure [ fig : sne - pec ] .
the hubble diagram @xmath288 vs. @xmath22 is shown in figure [ fig : hubble ] for 108 normal sneia .
the recession velocities are corrected for streaming motions as described in the explanation of column 5 of table [ tab : sn1 ] ( [ sec : data ] ) .
the apparent magnitudes are corrected for galactic and host galaxy absorption as laid out in
[ sec : absorption ] and reduced to @xmath23 and @xmath507 ( equation [ eq : other:3fitd ] ) .
a slightly curved hubble line , corresponding to a flat universe with @xmath508 , is fitted to the points in figure [ fig : hubble ] .
the corresponding relation is given by @xcite @xmath509 where @xmath510^{-1/2}dz,\ ] ] and where the intercept is given by @xmath511 from this follows the decisive relation @xmath512 as equation ( [ eq : hubble:04 ] ) shows , the determination of @xmath35 requires an independent calibration of @xmath513 of normal sneia and an evaluation of the intercept @xmath514 .
while a revised calibration of @xmath513 will be provided in the forthcoming summary paper , the determination of @xmath514 is given here .
nine solutions for @xmath514 are presented in table [ tab : hubble ] for different subsets of the sneia in table [ tab : sn1]a .
the subsets are defined by velocity cuts and by different amounts of the total reddening @xmath515 .
it is very satisfactory that the solutions 1 - 3 and 6 - 8 for sneia beyond @xmath45 give the same values for @xmath514 in each waveband within the statistical errors .
@xmath516 varies by @xmath517 implying a variation in mean absolute magnitude @xmath518 of only @xmath519 , which influences @xmath35 by less than @xmath520 .
solution 3 is from the fiducial sample as defined in
[ sec : absorption : host ] .
it is least affected by photometric errors , peculiar motions , possibly remaining errors of the k - correction , and cosmological effects .
the corresponding values of @xmath521 with equation ( [ eq : hubble:03 ] ) and @xmath281 give @xmath522 it is also noted that @xmath523 corresponding to @xmath524 , and that @xmath525 corresponding to @xmath526 , both values being consistent with the observed colors in equation ( [ eq : color : mean ] ) .
lcccccccc 1 ) all : & @xmath527 & @xmath481 & @xmath528 & @xmath481 & @xmath529 & @xmath530 & @xmath531 & @xmath532 + 2 ) @xmath533 $ ] : & @xmath534 & @xmath535 & @xmath536 & @xmath535 & @xmath537 & @xmath538 & @xmath352 & @xmath539 + 3 ) @xmath540 : & @xmath536 & @xmath422 & @xmath541 & @xmath422 & @xmath542 & @xmath543 & @xmath422 & @xmath544 + 4 ) @xmath545 : & @xmath546 & @xmath547 & @xmath548 & @xmath547 & @xmath549 & @xmath550 & @xmath551 & @xmath552 + 5 ) @xmath553 : & @xmath554 & @xmath555 & @xmath556 & @xmath555 & @xmath557 & @xmath558 & @xmath559 & @xmath560 + 6 ) @xmath561 : & @xmath534 & @xmath535 & @xmath562 & @xmath535 & @xmath563 & @xmath564 & @xmath352 & @xmath544 + 7 ) @xmath565 : & @xmath566 & @xmath567 & @xmath568 & @xmath567 & @xmath569 & @xmath570 & @xmath352 & @xmath571 + 8) @xmath572 : & @xmath573 & @xmath535 & @xmath574 & @xmath535 & @xmath575 & @xmath576 & @xmath352 & @xmath577 + 9 ) @xmath578 , table2 : & @xmath579 & @xmath567 & @xmath573 & @xmath422 & @xmath577 & @xmath580 & @xmath239 & @xmath581 + the solutions 6 - 8 admit different amounts of the total extinction @xmath515 in the galaxy and in the host galaxy .
this has virtually no effect on @xmath582 , and even @xmath516 , which is most sensitive to absorption corrections , changes by not more than @xmath583 ( corresponding to @xmath584 ) .
this proves in favor of the adopted unconventional values of @xmath301 ( equation [ eq : absorption : r ] ) .
the sneia in table [ tab : sn1]a have @xmath585 , and an error of @xmath303 of more than @xmath586 would cause a larger discrepancy . the limitation of the maximum value of @xmath515 does not lead to a decrease of the scatter about the hubble line .
this suggest that the magnitude dispersion is mainly due to _ random _ errors of @xmath515 and of the corresponding absorption corrections .
if one averages the @xmath13 , @xmath14 , and @xmath15 magnitudes to define the hubble line in solutions 1 - 8 , the scatter is _ not _ reduced , which indicates that the magnitude errors are strongly correlated .
it is not meaningful to compare the values of @xmath521 derived here with those of other authors , because they depend on a number of choices taken here : the intrinsic color of sneia , the reddening and absorption law , and the reduction to the standard values of @xmath23 and @xmath507 .
it is therefore mandatory , if the present values of @xmath514 are to be used to obtain @xmath35 through equation ( [ eq : hubble:04 ] ) , that the calibrating sneia are treated according to the prescriptions given here .
the dispersion about the hubble line in solution 3 , i.e.@xmath587 , is quite satisfactory for a large sample of 62 ( 58 ) sneia .
@xcite has found from 29 sneia @xmath588 , but his reddening - to - absorption ratio of @xmath589 for galactic and host galaxy absorption seems unrealistic .
previous values of @xmath590 @xcite or even @xmath591 @xcite are based on smaller samples , and are in addition color - limited by @xmath592 .
@xcite have found @xmath593 for 18 of their low - extinction sneia . for comparison
the mlcs method of fitting the sn light curves has yielded @xmath594 @xcite and @xmath595 @xcite .
the `` bayesian adapted template match '' ( batm ) of @xcite yields a scatter of @xmath596 .
the very small scatter of @xmath597 achieved by @xcite with their `` color - magnitude intercept calibration '' may not be representative because it is derived from a severely color - restricted sample ; their color limit @xmath598 cuts into the realistic color distribution of even unreddened sneia ( cf.figure [ fig : color : tab2 ] ) .
a physically more relevant question concerns the _ intrinsic _ scatter of homogenized sn luminosities .
the best handle for this is offered by the 21 ( 20 ) sneia in table [ tab : color : minred ] with minimum absorption in their host galaxies and @xmath599 ( solution 9 of table [ tab : hubble ] ) .
the higher velocity limit is chosen here to guard against peculiar motions which , if @xmath600 , can contribute a magnitude dispersion of @xmath601mag still at @xmath602 .
the sample gives a puzzlingly large value of @xmath603 , but quite low values for @xmath590 and @xmath604 .
the small scatter in @xmath15 is particularly noteworthy because @xmath51 , occurring before @xmath49 , requires good photometry at an early phase .
allowing conservatively for an average photometric error of @xmath605 , we derive an intrinsic scatter of the homogenized @xmath51 magnitudes of @xmath606 , which in view of the many small and not so small spectroscopic and light curve differences between normal sneia is remarkably small .
we have refrained from weighting the sn data , as some authors have done , by the quoted magnitude errors or by the time interval between inferred @xmath13 maximum and the onset of the photometry , although these data are given for most , but not all sneia in table [ tab : sn1]a .
the reason is mainly that the quoted magnitude errors from many different sources carry an subjective element , and that they include sometimes systematic errors and sometimes not .
we have instead made alternative solutions excluding sneia with large quoted magnitude errors and/or with a late onset of their photometry . in no case
the result or even the statistical error was significantly changed . in view of the large sample considered here
this is no surprise .
a last comment to the solutions in table [ tab : hubble ] concerns the relative size of @xmath35 in the local universe .
the strong increase of @xmath607 in solution 4 and 5 is certainly due to peculiar motions whose influence becomes important at small redshifts .
the median velocity of the sneia in solution 4 is @xmath608 .
the observed magnitude scatter of @xmath609 corresponds therefore to typical peculiar motions of @xmath610 . within @xmath611
the median velocity is @xmath612 and the scatter corresponds here to @xmath600 .
these values are high relative to the small rms peculiar velocity of @xmath613 consistently derived over still smaller scales ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , but many of the sneia lie in the virgo cluster region which is known to be noisy @xcite , and peculiar velocities are generally expected to increase with increasing scale length up to some asymptotic value which , judging from the cmb dipole , is @xmath614 in the radial direction .
for @xmath615 @xmath616 becomes smaller by @xmath617 and @xmath582 by @xmath618 than in solution 3 .
this correspond from equation ( [ eq : hubble:04 ] ) to a local decrease of @xmath35 by @xmath619 or @xmath620 .
analogously the very local value of @xmath35 is , from solution 5 , @xmath621 or @xmath622 smaller than the cosmic value .
the local decrease of @xmath35 is in view of the small number of sneia and the statistical errors only suggestive .
the mean absolute magnitude @xmath280 in equation ( [ eq : other : mabs ] ) is combined with the individual apparent magnitudes @xmath279 in table [ tab : sn3]a , column 9 , to yield photometric absorption - free distance moduli for 111 normal sneia and hence of their parent galaxies ( table [ tab : sn3]a , column 11 ) .
the expected random error is only @xmath623 ; some moduli suspected for one reason or another to be less accurate are shown in parentheses .
the moduli have the advantage particularly for the nearby galaxies to be independent of peculiar motions . of course , all distances are still to be shifted by a constant amount , once the adopted value of @xmath281 can be replaced by the actual value of @xmath35 , which is the ultimate aim of this series .
the main purpose of this paper is to update an earlier summary by @xcite on the reddening , absorption and second parameter corrections for a color - restricted sample of `` branch normal '' type ia supernovae where the corrected data were used to construct the hubble diagram in @xmath13 , @xmath14 , and @xmath15 . here
the sample is increased to 124 type ia supernovae with no restrictions placed on the observed color , thereby avoiding the criticisms of the @xcite sample raised by others concerning observational selection biases .
all sn believed to be `` branch normal '' @xcite were used , together with 13 spectroscopically peculiar sneia ( types 91 t , 91bg , sn2000cx , 1986 g ) that deviate in one way or another from the main sample .
the peculiar supernovae are listed separately in tables [ tab : sn1]b , [ tab : sn1]c , [ tab : sn3]b , and [ tab : sn3]c and are discussed in [ sec : snpec ] .
because no a priori cut in observed color at maximum was imposed , many of the sne in our enlarged sample suffer non - negligible reddening in their host galaxies , and care has been taken at each step in the reductions to guard against systematic errors in the derived absorptions .
\(1 ) the 22 sneia in e or s0 galaxies and the 12 sneia in the outlying regions of spirals , set out in table [ tab : color : minred ] , are expected to have nearly negligible extinction due to the host galaxy .
their mean intrinsic colors , corrected for galactic reddening and nearly free of any host galaxy reddening , are @xmath624 , @xmath625 ( i.e. the tail color at phase @xmath252 ) , and @xmath626 .
the individual colors @xmath18 become marginally redder , the colors @xmath20 significantly redder with increasing decline rate @xmath21 , while the tail colors @xmath19 become clearly bluer with increasing @xmath52 ( equations [ eq : color : mean : bv]@xmath255[eq : color : mean : vi ] ) .
the latter result is in variance with @xcite , @xcite , @xcite , and @xcite , who assumed the tail colors to be independent of any second parameters .
\(2 ) the host galaxy reddenings for all other normal sneia in the sample are based on the intrinsic colors derived in item ( 1 ) above .
the difference between the observed , galactic - reddening - free colors @xmath246 , @xmath61 , and @xmath627 and the corresponding colors from equations ( [ eq : color : mean : bv]@xmath255[eq : color : mean : vi ] ) , i.e.adjusted to the appropriate value of @xmath21 , yield three different estimates of the reddening @xmath53 , @xmath258 , and @xmath628 .
\(3 ) due to the general color dependence of color excesses caused by the photometric bandwidth effect , the excesses @xmath258 , determined in item ( 2 ) from a late , quite red phase of the sn light curves , must be converted to the value applicable at the blue maximum phase .
this is achieved by requiring that @xmath18 as well as @xmath20 be independent of @xmath258 ( cf . figures [ fig : color : e35 ] & [ fig : color : evihost ] ) .
the requirement leads to equation ( [ eq : color : ebvmax35 ] ) which converts @xmath258 to @xmath317 , and to equation ( [ eq : color : ebvvihost ] ) which converts @xmath272 into @xmath275 .
the three determinations of @xmath53 are intercompared in figures [ fig : color : comp1 ] & [ fig : color : comp2 ] .
they agree well and are averaged to give the adopted reddening @xmath1 in the host galaxy .
the random error of @xmath1 is @xmath284mag .
\(4 ) the adopted color excesses @xmath1 depend on the assumption that the sample of sneia in e / s0 galaxies and the outliners in spirals in table [ tab : color : minred ] suffer no host - galaxy reddenings .
if , in fact , they do suffer a non - zero amount of reddening , the true intrinsic colors of the complete sample , determined by the method of items ( 1 ) through ( 3 ) , will be bluer , and the magnitudes brighter than adopted here , but this will have no effect on the calibration of @xmath35 to be made in the final summary paper of this series as long as the calibrating sneia in host galaxies are reduced to the _ same _ color . in that case , the local calibrators would suffer the same error in absorption as would the distant sneia defining the global hubble diagram , and the error cancels .
\(5 ) the magnitudes @xmath629 , corrected for conventional galactic absorption , were transformed into absolute magnitudes @xmath630 for a `` fiducial '' sample of 62(58 ) sneia which have @xmath631 ( excluding only the highly absorbed sn1996ai and five sneia deviating by more than @xmath253 and discussed in [ sec : absorption : host ] ) to guard against peculiar motions and other effects .
a flat universe with @xmath281 , @xmath632 was assumed .
we show in [ sec : absorption : host ] a strong correlation of @xmath630 with @xmath1 ( figure [ fig : absorption : mabs ] ) .
this correlation determines the ratio of absorption - to - reddening in the host galaxy .
the absorption - to - reddening values determined in this way are @xmath633 , @xmath634 , and @xmath635 .
these well determined values differ significantly from the canonical values in the galaxy of @xmath301 of @xmath9 , @xmath10 , and @xmath11 .
the significant difference with the galactic values has been pointed out before by @xcite , @xcite , @xcite , @xcite , @xcite , and @xcite .
an explanation is that the intense radiation of the supernovae in some manner modifies the size distribution of the dust grains in the host galaxy near the sn .
the canonical values of @xmath301 have been used to determine the galactic absorption and the new , smaller values of @xmath320 for the absorption in the host galaxy .
\(6 ) the absorption - free magnitudes @xmath636 of the fiducial sample are plotted against the hubble type of the host galaxies in figure [ fig : other : type ] .
it is confirmed that the supernovae in e / s0 galaxies average @xmath12 magnitude fainter than sne in late - type spirals , known already since 1995 ( @xcite ; @xcite ; @xcite ; @xcite , figure 6 ) .
it has been seen in a different form by the discovery by ( * ? ? ?
* their figure 4 ) , that the decline rate , @xmath21 , is correlated with galaxy type ( sneia in e / s0 galaxies have average decline rates near @xmath637 , while those in late - type spirals have a mean rate near @xmath638 ) , coupled with the fact that the sneia with large decline rates are fainter than those with smaller rates .
the reason for the @xmath396-galaxy type correlation is expected to be related to the different gestation times for type ia sne and the subsequent difference in the fe abundance between e / s0 galaxies and spirals of the same age .
the consequences may or may not be profound in using sneia for cosmological probes as is the current trend @xcite , depending on the precision with which the effect can be compensated for by the @xmath21 correlations with galaxy type .
\(7 ) the fully absorption - corrected magnitudes @xmath335 confirm the @xcite effect of a correlation with the @xmath21 decline rates ( figure [ fig : other : dm15 ] ) with correlation slopes of @xmath639 , @xmath640 , @xmath641 ( table [ tab : other : dm15r ] ) .
the present slopes @xmath340 , which correct for the magnitude differences of sneia in e / s0 galaxies and spirals , lie between the smaller values of @xcite and the larger values of @xcite and @xcite .
it is shown that the reason for this is the sensitivity of @xmath340 on the adopted slopes of the color-@xmath21 relation in equations ( [ eq : color : mean : bv]@xmath255[eq : color : mean : vi ] ) .
steeper slopes of these relations lead necessarily to larger values of @xmath340 .
incorrect slopes of the color-@xmath21 relation enter the color excesses and propagate strongly into the absorption and hence into the @xmath337 magnitudes .
they cancel , however , if the latter are normalized to a common value of @xmath21 ( @xmath642 ) .
\(8 ) the absolute magnitudes , @xmath643 show some dependence on the @xmath18 and @xmath20 colors ( figure [ fig : other : color ] ) , as suggested by @xcite and @xcite .
therefore the adopted magnitudes @xmath644 of normal sneia have been reduced to common values of @xmath23 _ and _ @xmath24 by equation ( [ eq : other:3fitd ] ) and table [ tab : coeff3fit ] .
this slightly reduces the dispersion about the hubble line without changing the _
mean _ magnitude of the sample .
\(9 ) the present sample of 124 sneia contains only four spectroscopically distinct sne-91 t , with sn1991 t as the prototype .
their apparent rarity must be real because of their pronounced overluminosity . the overluminosity amounts to @xmath426 on average if one assumes them to have the same color at maximum @xmath18 as normal sneia ( table [ tab : sneia - t ] ) .
the mean overluminosity is reduced to @xmath427 if one allows in addition for their below - average decline rates @xmath21 .
the new class of sne-99aa @xcite , named after sn1999aa , comprises seven objects in table [ tab : sn1]a ( marked with @xmath319 ) .
they are spectroscopically between normal sneia and sne-91 t , but are difficult to recognize because their spectroscopic peculiarity is confined to the pre - maximum phase .
their true frequency in a distance - limited sample may therefore be as high as @xmath645 .
they are less overluminous than sne-91 t , i.e.@xmath646 , or @xmath647 if reduced to @xmath23 , as compared to normal sneia .
sne-99aa have very slow decline rates ( @xmath648 ) and occur correspondingly in late - type galaxies .
the mean coded type of their host galaxies is @xmath649 ( @xmath650sbc ) as compared to @xmath651 for sne-91 t and @xmath652 ( @xmath650sab ) for normal sneia .
\(10 ) seven sne-91bg , whose eponymous star is sn1991bg , have sufficient photometry to be included in table [ tab : sn1]c .
they are very red and spectroscopically peculiar . in accordance with their exceptionally fast decline rates
( @xmath654 ) they lie preferentially in early - type galaxies ( @xmath655 ) .
this suggests little absorption in their host galaxies ; zero absorption is assumed here to derive their intrinsic photometric properties ( table [ tab : sneia - bg ] ) .
the sne-91bg emerge as a surprisingly homogeneous and separate class with @xmath13 magnitudes at maximum about @xmath27 below that of normal sneia .
the luminosity of sn1986 g seems to lie between sneia and sne-91bg ( see table [ tab : sneia - bg ] ) .
\(11 ) the hubble diagrams in @xmath13 , @xmath14 , and @xmath15 for all 111 normal sneia in table [ tab : sn1]a are shown in figure [ fig : hubble ] using fully corrected apparent magnitudes @xmath277 .
solutions for the intercepts @xmath514 of the hubble line are given for various subsets with different restrictions in table [ tab : hubble ] .
the fiducial sample of 62 ( 58 ) sneia with @xmath656 is taken as the most reliable sample ( solution 3 in table [ tab : hubble ] ) .
the resulting values of @xmath514 inserted in equation ( [ eq : hubble:04 ] ) and @xmath281 give @xmath657 , @xmath658 , and @xmath659 . the dispersion of the @xmath22 magnitudes ( corrected for galactic and internal absorption and normalized to @xmath254 ) about the hubble line is @xmath587 ( solution 3 in table [ tab : hubble ] ) .
it increases for distances shorter than @xmath45 due to peculiar motions ( solutions 5 and 6 ) .
much of the dispersion beyond @xmath45 is caused by random errors of the host galaxy absorption , as seen by the sneia with minimum absorption in table [ tab : color : minred ] which give @xmath604 , if restricted in addition to @xmath599 .
this is for normal sneia , after they are normalized to common values of @xmath21 and @xmath18 , an _
upper _ limit to their intrinsic luminosity dispersion in @xmath15 in view of remaining observational magnitude errors .
the intrinsic dispersion in @xmath13 and @xmath14 may be somewhat larger . in spite of the noticeable scatter introduced by the corrections for host galaxy absorption , the _
mean _ of the corrections must be nearly correct .
this is demonstrated by solutions 3 and 6 - 8 in table [ tab : hubble ] , where different cuts of the accepted maximum value of the total color excess @xmath515 lead to statistically insignificant variations of the intercept @xmath514 of only @xmath660 ( corresponding to @xmath661 ) . the combination of the absolute magnitude @xmath280 from above with the individual apparent magnitudes @xmath279 yields the absorption - free luminosity distances of 111 parent galaxies of normal sneia ( table [ tab : sn3]a , column 11 ) .
the distances are particularly valuable for nearby galaxies because they are insensitive to peculiar motions .
if the value @xmath281 is to be replaced by any other value , the listed distance moduli need only a bulk correction by an additive term .
had we adopted a flat universe with @xmath662 the value of @xmath514 in solution 3 of table [ tab : hubble ] would change by only @xmath663 .
this would decrease @xmath35 by merely @xmath664 .
the insensitivity of @xmath35 to the cosmological model is due to the velocity cut at @xmath665 of solution 3 .
\(12 ) the main goal of the present paper is to determine an up - to - date hubble diagram using a large and unbiased sample of branch - normal sneia , whose magnitudes are corrected for galactic and host galaxy absorption and normalized to standard values of the decline rate @xmath21 and of the color @xmath18 . regardless whether one considers the unrestricted sample of non - local or restricted samples with only a minimum of extinction one finds remarkably robust solutions for the intercept @xmath514 , which through equation ( [ eq : hubble:04 ] ) is decisive for the determination of the value of @xmath35 .
the remaining uncertainty of @xmath514 of @xmath666 will carry an error into the derived value of @xmath35 of only @xmath667 .
the consequence is that the accuracy of the derived large - scale value of @xmath35 will depend almost exclusively on the goodness of the luminosity calibration that we can achieve in a forthcoming paper by means of the sneia in host galaxies with known cepheids .
we are greatly indebted to dr .
g. altavilla for kindly providing his unpublished template - fitted @xmath50 magnitudes for a large sample of sneia .
we are equally indebted to dr .
s. jha for his generosity to make his unpublished tail colors @xmath668 available to us .
thanks an anonymous donator for financial support , while g.a.t .
thanks the swiss national science foundation for support over many decades .
is indebted to the carnegie institution for ongoing post - retirement facilities support .
altavilla , g. , et al .
2004 , mnras , 349 , 1344 anapuma , g.c .
1997 , aj , 114 , 2054 barbon , r. , ciatti , f. , & rosino , l. 1973 , a&a , 25 , 241 barris , b.j . , & tonry , j.l .
2004 , apj , 613 , 21 barris , b.j .
, et al . 2004 , apj , 602 , 571 benetti , s. , et al . 2004 ,
mnras , 348 , 261 berlind , p. , jha , s. , garnavich , p. , & kirshner , r. 1997 , iau circ . , 6700 , 2 bottari , c. , nakano , s. , kushida , y. , garnavich , p. , challis , p. , kirshner , r. , berlind , p. , & kleyna , j. 1996 , iau circ .
, 6422 branch , d. , & tammann , g.a .
1992 , ara&a , 30 , 359 branch , d. , fisher , a. , & nugent , p. 1993 , aj , 106 , 2383 candia , p. , et al .
2003 , pasp , 115 , 277 carroll , s.m . , press , w.h . , &
turner , e.l .
1992 , ara&a , 30 , 499 cristiani , s. , et al .
1992 , a&a , 259 , 63 dale , d.a . , & giovanelli , r. 2000 , in asp .
201 , cosmic flows workshop , eds .
s. courteau & j. willick ( san francisco : asp ) , 25 de vaucouleurs , g. , de vaucouleurs , a. , & corwin , h.g . 1976 , second reference catalogue of bright galaxies ( austin : univ .
of texas press ) federspiel , m. , sandage , a. , & tammann , g.a .
1994 , apj , 430 , 29 filippenko , a.v .
, et al . 1992 , apj , 384 , l15 fisher , a. , branch , d. , hatano , k. , & baron , e. 1999 , mnras , 304 , 67 garavini , g. , et al .
2004 , aj , 128 , 387 garnavich , p.m. , et al .
2004 , apj , 613 , 1120 germany , l.m . ,
reiss , d.j . ,
schmidt , b.p . , stubbs , c.w .
, & suntzeff , n.b .
2004 , a&a , 415 , 863 hamuy , m. , phillips , m.m . ,
maza , j. , suntzeff , n.b . , schommer , r.a . , & aviles , r. 1995 , aj , 109 , 1 hamuy , m. , phillips , m.m . ,
suntzeff , n.b . ,
schommer , r.a . ,
maza , j. , & aviles , r. 1996a , aj , 112 , 2391 hamuy , m. , phillips , m.m . ,
suntzeff , n.b . ,
schommer , r.a . ,
maza , j. , & aviles , r. 1996b , aj , 112 , 2398 hamuy , m. , phillips , m.m . ,
suntzeff , n.b . ,
schommer , r.a . ,
maza , j. , smith , r.c . , lira , p. , & aviles , r. 1996c , aj , 112 , 2438 hamuy , m. , trager , s.c . ,
pinto , p.a . ,
phillips , m.m . ,
schommer , r.a . ,
ivanov , v. , & suntzeff , n.b .
2000 , aj , 120 , 1479 hamuy , m. , et al .
1996d aj , 112 , 2408 hflich , p. , gerady , c.l . ,
fesen , r.a . , & sakai , s. 2002 , apj , 568 , 791 howell , d.a .
2001 , apj , 554 , l193 jha , s. 2002 , phd .
thesis , harvard univ .
jha , s. , et al . 1999 ,
apjs , 125 , 73 karachentsev , i.d . , & makarov , d.a .
1996 , aj , 111 , 794 knop , r.a .
, et al . 2003 , apj , 598 , 102 kowal , c.t . 1968 , aj , 73 , 1021 kraan - korteweg , r.c . 1986 , a&as , 66 , 255 krisciunas , k. , hastings , n.c . ,
loomis , k. , mcmillan , r. , rest , a. , riess , a.g . , & stubbs , c. 2000 , apj , 539 , 658 krisciunas , k. , et al .
2001 , aj , 122 , 1616 krisciunas , k. , et al .
2003 , aj , 125 , 166 li , w. , filippenko , a.v . , &
riess , a.g .
2001a , apj , 546 , 719 li , w. , filippenko , a.v . , treffers , r.r . ,
riess , a.g .
, hu , j. , & qiu , y. 2001b , apj , 546 , 734 li , w. , et al .
2001c , pasp , 113 , 1178 li , w.d .
, et al . 1999 , aj , 117 , 2709 lira , p. 1995
, master thesis , univ .
chile macri , l.m . ,
stetson , p.b . ,
bothun , g.d . , freedman , w.l .
, garnavich , p.m. , jha , s. , madore , b.f . , & richmond , m.w .
2001 , apj , 559 , 243 mandel , k. , jha , s. , matheson , t. , challis , p. , & kirshner , r.p .
2001 , baas , 33 , 1370 mazzali , p.a . ,
chugai , n. , turatto , m. , lucy , l.b . , danziger , i.j .
, cappellaro , e. , della valle , m. , & benetti , s. 1997 , mnras , 284 , 151 nobili , s. , goobar , a. , knop , r. , & nugent , p. 2003
, a&a , 404 , 901 parodi , b.r .
, saha , a. , sandage , a. , & tammann , g.a .
2000 , apj , 540 , 634 perlmutter , s. , et al .
1997 , apj , 483 , 565 perlmutter , s. , et al .
1999 , apj , 517 , 565 phillips , m.m .
1993 , apj , 413 , l105 phillips , m.m . , lira , p. , suntzeff , n.b . ,
schommer , r.a . ,
hamuy , m. , & maza , j. 1999 , aj , 118 , 1766 pignata , g. , et al .
2004 , mnras , in press ( astro - ph/0408234 ) pskovskii , y.p .
1967 , sov .
astron . , 11 , 63 pskovskii , y.p .
1971 , sov .
, 14 , 798 pskovskii , y.p .
1984 , sov .
astron . , 28 , 658 rich , r.m .
1987 , aj , 94 , 651 riess , a.g . , press , w.h . , & kirshner , r.p .
1996 , apj , 473 , 88 riess , a.g .
, et al . 1998 , aj , 116 , 1009 riess , a.g . , et al . 1999 ,
aj , 117 , 707 riess , a.g .
, et al . 2004 , apj , 607 , 665 saha , a. , sandage , a. , labhardt , l. , tammann , g.a . , macchetto , f.d . , & panagia , n. 1996 , apjs , 107 , 693 ( sn1960f ) saha , a. , sandage , a. , labhardt , l. , tammann , g.a . , macchetto , f.d . , & panagia , n. 1997 , apj , 486 , 1 ( sn1990n ) saha , a. , sandage , a. , tammann , g.a . , dolphin , a.e . ,
christensen , j. , panagia , n. , & macchetto , f.d .
2001b , apj , 562 , 314 ( sn1998aq ) saha , a. , sandage , a. , tammann , g.a . , labhardt , l. , macchetto , f.d . , & panagia , n. 1999 , apj , 522 , 802 ( sn1989b ) saha , a. , sandage , a. , thim , f. , labhardt , l. , tammann , g.a . , christensen , j. , panagia , n. , & macchetto , f. d. 2001a , apj , 551 , 973 ( sn1991 t ) sandage , a. 1986 , apj , 307 , 1 sandage , a. 1987 , apj , 317 , 557 sandage , a. , & tammann , g.a .
1982 , apj , 256 , 339 sandage , a. , & tammann , g.a .
1993 , apj , 415 , 1 sandage , a. , & tammann , g.a .
1995 , apj , 446 , 1 sandage , a. , & tammann , g.a .
1997 , in critical dialogues in cosmology , ed .
n. turok ( singapore : world scientific ) , 130 sandage , a. , tammann , g.a . , & reindl , b. 2004 , a&a , 424 , 43 sandage , a. , tammann , g.a . , & saha , a. 2001 , in supernovae and gamma - ray bursts : the greatest explosions since the big bang , eds . m. livio , n. panagia , & k. sahu ( cambridge : cambridge university press ) , 304 schaefer , b.e .
1998 , apj , 509 , 80 schlegel , d. , finkbeiner , d. , & davis , m. 1998 , apj , 500 , 525 stritzinger , m. , et al .
2002 , aj , 124 , 2100 strolger , l .- g .
, et al . 2002 , aj , 124 , 2905 tammann , g.a . ,
reindl , b. , thim , f. , saha , a. , & sandage , a. 2002 , in asp conf .
283 , a new era in cosmology , eds . t. shanks & n. metcalfe ( san francisco : asp ) , 258 tammann , g.a . , & sandage , a. 1985 , apj , 294 , 81 tammann , g.a . ,
sandage , a. , & reindl , b. 2003 , a&a , 404 , 423 thim , f. , tammann , g.a . ,
saha , a. , dolphin , a. , sandage , a. , tolstoy , e. , & labhardt , l. 2003 , apj , 590 , 256 tonry , j.l .
, et al . 2003 , apj , 594 , 1 tripp , r. 1998 , a&a , 331 , 815 tripp , r. , & branch , d. 1999 , apj , 525 , 209 valentini , g. , et al .
2003 , apj , 595 , 779 van den bergh , s. , li , w. , & filippenko , a.v .
2002 , pasp , 114 , 820 van den bergh , s. , li , w. , & filippenko , a.v .
2003 , pasp , 115 , 1280 vink , j. , et al .
2003 , a&a , 397 , 115 wang , l. , goldhaber , g. , aldering , g. , & perlmutter s. 2003 , apj , 590 , 944 wells , l.a .
, et al . 1994 , aj , 108 , 2233 llcrrrrrcrrrrl + 1937c & ic 4182@xmath669 & 5 & 321 & 330@xmath670 & 8.80 & 8.82 & & 0.85 & 0.014 & 0.001 & -0.045 & & a + 1960f & ngc 4496a@xmath669 & 5 & 1730 & 1179@xmath671 & 11.60 & 11.51 & & 0.87 & 0.025 & 0.099 & & & sa + 1972e & ngc 5253@xmath669 & 5 & 404 & 167@xmath670 & 8.49 & 8.49 & 8.80 & 1.05 & 0.056 & -0.030 & -0.052 & -0.069 & sa + 1972j & ngc 7634 & -1 & 3225 & 2855 & 14.80 & 14.56 & & 1.45 & 0.046 & 0.202 & -0.040 & & a + 1974 g & ngc 4414@xmath669 & 5 & 716 & 655@xmath670 & 12.48 & 12.30 & & 1.11 & 0.019 & 0.184 & 0.137 & & s + 1980n & ngc 1316 & 1 & 1760 & 1338@xmath672 & 12.49 & 12.42 & 12.70 & 1.30 & 0.021 & 0.064 & 0.039 & -0.045 & ha + 1981b & ngc 4536@xmath669 & 4 & 1808 & ( 1179@xmath671 ) & 12.04 & 11.98 & & 1.13 & 0.018 & 0.064 & 0.009 & & a + 1981d & ngc 1316 & 1 & 1760 & 1338@xmath672 & 12.59 & 12.40 & & & 0.021 & 0.242 & & & p + 1982b & ngc 2268 & 4 & 2222 & 2610@xmath670 & 13.65 & 13.40 & & 0.94 & 0.064 & 0.217 & 0.041 & & a + 1983 g & ngc 4753 & -1 & 1239 & 1497@xmath670 & 13.00 & 12.80 & & 1.30 & 0.034 & 0.180 & 0.281 & & a + 1984a & ngc 4419 & 1 & -261 & 1179@xmath671 & 12.50 & 12.30 & & & 0.033 & 0.240 & & & h + 1989b & ngc 3627@xmath669 & 3 & 727 & 549@xmath670 & 12.34 & 11.99 & 11.70 & 1.31 & 0.032 & 0.332 & 0.298 & 0.302 & w + 1990n & ngc 4639@xmath669 & 3 & 1010 & 1179@xmath671 & 12.75 & 12.73 & 12.95 & 1.05 & 0.026 & 0.020 & 0.070 & 0.012 & ha + 1990o & mcg 3 - 44 - 03 & 1 & 9193 & 9175 & 16.59 & 16.51 & 16.80 & 0.94 & 0.093 & 0.018 & -0.001 & -0.076 & ha + 1990 t & pgc 63925 & 1 & 11992 & 11893 & 17.27 & 17.35 & 17.36 & 1.13 & 0.053 & -0.111 & 0.055 & 0.114 & ha + 1990y & anon 0335 - 33 & -3 & 11702 & 11603 & 17.69 & 17.42 & 17.57 & 1.13 & 0.008 & 0.284 & 0.205 & 0.063 & ha + 1990af & anon 2135 - 62 & -1 & 15080 & 14966 & 17.91 & 17.84 & & 1.59 & 0.035 & 0.036 & -0.022 & & h + 1991s & ugc 5691 & 3 & 16489 & 16807 & 17.78 & 17.75 & 17.99 & 1.00 & 0.026 & 0.032 & 0.033 & 0.004 & ha + 1991u & ic 4232 & 4 & 9426 & 9724 & 16.67 & 16.61 & 16.62 & 1.11 & 0.062 & 0.021 & 0.056 & 0.108 & ha + 1991ag & ic 4919 & 3 & 4264 & 4161 & 14.67 & 14.56 & 14.83 & 0.87 & 0.062 & 0.082 & -0.012 & -0.032 & ha + 1992a & ngc 1380 & 0 & 1877 & 1338@xmath672 & 12.56 & 12.55 & 12.80 & 1.47 & 0.018 & -0.001 & -0.032 & -0.040 & ha + 1992j & anon 1009 - 26 & -2 & 13491 & 13828 & 17.88 & 17.76 & 17.88 & 1.69 & 0.057 & 0.060 & 0.032 & -0.011 & ha + 1992p & ic 3690 & 1 & 7615 & 7941 & 16.14 & 16.16 & 16.40 & 1.05 & 0.021 & -0.015 & 0.014 & 0.003 & ha + 1992ae & anon 2128 - 61 & -3 & 22484 & 22366 & 18.62 & 18.54 & & 1.30 & 0.036 & 0.059 & 0.053 & & a + 1992ag & eso 508-g67 & 5 & 7795 & 8095 & 16.64 & 16.47 & 16.50 & 1.20 & 0.097 & 0.092 & 0.359 & 0.059 & ha + 1992al & eso 234-g69 & 3 & 4197 & 4048 & 14.59 & 14.65 & 14.93 & 1.09 & 0.034 & -0.070 & -0.015 & -0.036 & ha + 1992aq & anon 2304 - 37 & 1 & 30279 & 30014 & 19.42 & 19.39 & 19.48 & 1.69 & 0.012 & 0.015 & 0.026 & 0.044 & ha + 1992au & anon 0010 - 49 & -3 & 18287 & 18092 & 18.17 & 18.13 & 18.51 & 1.69 & 0.017 & 0.020 & -0.014 & -0.141 & ha + 1992bc & eso 300-g9 & 2 & 5996 & 5876 & 15.13 & 15.25 & 15.56 & 0.90 & 0.022 & -0.109 & -0.108 & -0.027 & ha + 1992bg & anon 0741 - 62 & 1 & 10793 & 10936 & 17.39 & 17.30 & 17.32 & 1.15 & 0.185 & -0.074 & -0.015 & -0.002 & ha + 1992bh & anon 0459 - 58 & 4 & 13491 & 13519 & 17.68 & 17.63 & 17.75 & 1.13 & 0.022 & 0.050 & 0.061 & 0.070 & ha + 1992bk & eso 156-g8 & -3 & 17284 & 17237 & 18.07 & 18.14 & 18.21 & 1.67 & 0.015 & -0.087 & -0.018 & 0.056 & ha + 1992bl & eso 291-g11 & 0 & 12891 & 12661 & 17.34 & 17.37 & 17.59 & 1.56 & 0.011 & -0.038 & -0.033 & -0.024 & ha + 1992bo & eso 352-g57 & -2 & 5396 & 5164 & 15.86 & 15.84 & 15.95 & 1.69 & 0.027 & -0.010 & -0.030 & 0.020 & ha + 1992bp & anon 0336 - 18 & -2 & 23684 & 23557 & 18.53 & 18.53 & 18.68 & 1.52 & 0.069 & -0.064 & 0.020 & -0.024 & ha + 1992br & anon 0145 - 56 & -3 & 26382 & 26259 & 19.34 & 19.27 & & 1.69 & 0.026 & 0.041 & -0.020 & & h + 1992bs & anon 0329 - 37 & 3 & 18887 & 18787 & 18.33 & 18.33 & & 1.15 & 0.011 & 0.010 & 0.017 & & a + 1993b & anon 1034 - 34 & 3 & 20686 & 21011 & 18.71 & 18.65 & 18.71 & 1.30 & 0.079 & -0.004 & 0.122 & 0.045 & ha + 1993h & eso 445-g66 & 2 & 7257 & 7522 & 16.99 & 16.72 & 16.60 & 1.69 & 0.060 & 0.207 & 0.022 & 0.137 & ha + 1993l & ic 5270 & 4 & 1926 & 1854@xmath670 & 13.40 & 13.22 & & 1.47 & 0.014 & 0.173 & 0.212 & & a + 1993o & anon 1331 - 33 & -2 & 15589 & 15867 & 17.79 & 17.86 & 17.99 & 1.26 & 0.053 & -0.107 & -0.012 & 0.026 & ha + 1993ac & pgc 17787 & -3 & 14690 & 14674 & 18.42 & 18.25 & 18.12 & 1.25 & 0.163 & 0.024 & -0.004 & 0.101 & rpa + 1993ae & ic 126 & -3 & 5712 & 5410 & 15.42 & 15.52 & 15.65 & 1.47 & 0.038 & -0.131 & -0.058 & 0.019 & rpa + 1993ag & anon 1003 - 35 & -2 & 14700 & 15013 & 18.30 & 18.12 & 18.22 & 1.30 & 0.112 & 0.082 & 0.055 & -0.007 & ha + 1993ah & eso 471-g27 & -1 & 8803 & 8504 & 16.32 & 16.36 & 16.65 & 1.45 & 0.020 & -0.052 & 0.048 & -0.065 & ha + 1994d & ngc 4526 & -1 & 448 & 1179@xmath671 & 11.86 & 11.94 & 12.11 & 1.31 & 0.022 & -0.088 & -0.054 & 0.022 & ha + 1994 m & ngc 4493 & -3 & 6943 & 7289 & 16.35 & 16.30 & 16.40 & 1.45 & 0.023 & 0.035 & 0.083 & 0.052 & rpa + 1994q & pgc 59076 & -1 & 8863 & 8861 & 16.44 & 16.36 & 16.63 & 0.90 & 0.017 & 0.096 & 0.007 & 0.002 & rpa + 1994s & ngc 4495 & 3 & 4550 & 4831 & 14.79 & 14.83 & 15.14 & 1.02 & 0.021 & -0.034 & -0.055 & -0.038 & r + 1994 t & pgc 46640 & 1 & 10390 & 10703 & 17.35 & 17.23 & 17.38 & 1.45 & 0.029 & 0.099 & 0.028 & 0.016 & rpa + 1994ae & ngc 3370 & 5 & 1279 & 1575@xmath670 & 13.15 & 13.10 & 13.40 & 0.87 & 0.030 & 0.054 & 0.070 & -0.025 & rpa + 1995d & ngc 2962 & 0 & 1966 & 2129@xmath670 & 13.44 & 13.40 & 13.70 & 1.03 & 0.058 & 0.009 & 0.028 & -0.063 & rpa + 1995e & ngc 2441 & 3 & 3470 & 3496 & 16.82 & 16.10 & & 1.19 & 0.027 & 0.712 & 0.582 & & ra + 1995ak & ic 1844 & 1 & 6811 & 6589 & 16.09 & 16.10 & 16.21 & 1.45 & 0.038 & -0.040 & 0.172 & 0.033 & rpa + 1995al & ngc 3021 & 4 & 1541 & 1851@xmath670 & 13.36 & 13.25 & 13.50 & 0.89 & 0.014 & 0.129 & 0.066 & 0.018 & rpa + 1996c & mcg 8 - 25 - 47 & 1 & 8094 & 8245 & 16.59 & 16.56 & 16.77 & 0.94 & 0.013 & 0.048 & -0.007 & 0.039 & rpa + 1996x & ngc 5061 & -3 & 2039 & 2120@xmath670 & 13.26 & 13.21 & 13.39 & 1.32 & 0.069 & -0.005 & -0.016 & -0.024 & rpa + 1996z & ngc 2935 & 3 & 2277 & 2285@xmath670 & 14.61 & 14.25 & & 1.22 & 0.064 & 0.314 & 0.350 & & r + 1996ab & anon 1521 + 28@xmath673 & 5 & 37240 & 37370 & 19.54 & 19.42 & & 0.87 & 0.032 & 0.122 & 0.080 & & r + 1996ai@xmath674 & ngc 5005 & 4 & 946 & 1298@xmath670 & 16.96 & 15.26 & & 0.85 & 0.014 & 1.721 & 1.501 & & ra + 1996bk & ngc 5308 & -1 & 2041 & 2462@xmath670 & 14.84 & 14.41 & & 1.69 & 0.018 & 0.409 & 0.207 & & ra + 1996bl & anon 0036 + 11@xmath673 & 5 & 10793 & 10447 & 17.08 & 16.98 & 17.03 & 1.11 & 0.092 & 0.031 & 0.036 & 0.059 & rpa + 1996bo & ngc 673 & 5 & 5182 & 4893 & 16.15 & 15.78 & & 1.30 & 0.077 & 0.307 & 0.206 & & ra + 1996bv & ugc 3432 & 5 & 4998 & 5016 & 15.77 & 15.54 & 15.56 & 0.84 & 0.105 & 0.160 & 0.098 & 0.092 & rpa + 1997e & ngc 2258 & -1 & 4001 & 3998 & 15.59 & 15.45 & 15.46 & 1.39 & 0.124 & 0.026 & 0.013 & 0.031 & j + 1997y & ngc 4675 & 3 & 4806 & 4964 & 15.28 & 15.30 & 15.36 & 1.25 & 0.017 & -0.020 & 0.024 & 0.100 & j + 1997bp & ngc 4680 & 3 & 2492 & 2647@xmath670 & 14.12 & 13.92 & 14.09 & 0.97 & 0.044 & 0.185 & 0.286 & 0.036 & aj + 1997bq & ngc 3147 & 4 & 2820 & 2879 & 14.57 & 14.27 & 14.37 & 1.01 & 0.024 & 0.304 & 0.082 & 0.092 & j + 1997cw@xmath675 & ngc 105 & 2 & 5133 & 4782 & 16.00 & 15.52 & 15.33 & 1.02 & 0.073 & 0.434 & 0.335 & 0.233 & j + 1997dg & anonymous & 1 & 10193 & 9845 & 17.20 & 17.11 & 17.14 & 1.13 & 0.078 & 0.034 & & 0.081 & j + 1997do & ugc 3845 & 4 & 3034 & 3140 & 14.56 & 14.46 & 14.59 & 0.99 & 0.063 & 0.065 & & 0.044 & j + 1997dt & ngc 7448 & 4 & 2194 & 2356@xmath670 & 15.64 & 15.08 & 14.54 & 1.04 & 0.057 & 0.529 & & 0.463 & j + 1998v & ngc 6627 & 3 & 5268 & 5148 & 15.88 & 15.71 & 15.62 & 1.06 & 0.196 & -0.001 & -0.018 & 0.066 & j + 1998ab@xmath675 & ngc 4704 & 5 & 8134 & 8354 & 15.94 & 15.93 & 15.98 & 0.88 & 0.017 & 0.026 & 0.070 & 0.141 & j + 1998aq & ngc 3982@xmath669 & 3 & 1109 & 1514@xmath670 & 12.22 & 12.28 & & 1.12 & 0.014 & -0.051 & -0.053 & & sa + 1998bu & ngc 3368@xmath669 & 2 & 897 & 810@xmath670 & 12.22 & 11.86 & 11.65 & 1.15 & 0.025 & 0.356 & 0.209 & 0.272 & a + 1998dh & ngc 7541 & 5 & 2678 & 2766@xmath670 & 14.24 & 13.98 & 14.04 & 1.23 & 0.068 & 0.210 & 0.090 & 0.061 & j + 1998dk & ugc 139 & 5 & 3963 & 3609 & 14.93 & 14.74 & 14.81 & 1.05 & 0.044 & 0.172 & 0.127 & 0.091 & j + 1998 dm & mcg-01 - 4 - 44@xmath676 & 4 & 1968 & 1976@xmath670 & 14.70 & 14.48 & 14.28 & 1.07 & 0.044 & 0.201 & 0.259 & 0.258 & j + 1998dx & ugc 11149 & -3 & 14990 & 14895 & 17.71 & 17.74 & 17.86 & 1.55 & 0.041 & -0.068 & & 0.015 & j + 1998ec & ugc 3576 & 3 & 5966 & 6032 & 16.44 & 16.21 & 16.22 & 1.08 & 0.085 & 0.169 & & 0.092 & j + 1998ef & ugc 646 & 3 & 5319 & 5020 & 15.21 & 15.18 & 15.28 & 0.97 & 0.073 & -0.014 & & 0.056 & j + 1998eg & ugc 12133 & 5 & 7423 & 7056 & 16.62 & 16.50 & 16.49 & 1.15 & 0.123 & 0.018 & & 0.067 & j + 1998es@xmath675 & ngc 632 & 2 & 3168 & 2868 & 13.99 & 13.87 & 14.11 & 0.87 & 0.032 & 0.122 & 0.031 & 0.011 & j + 1999x & cgcg 180 - 22@xmath673 & & 7503 & 7720 & 16.45 & 16.29 & 16.31 & 1.11 & 0.032 & 0.151 & & 0.126 & j + 1999aa@xmath675 & ngc 2595 & 5 & 4330 & 4572 & 14.92 & 14.88 & 15.24 & 0.85 & 0.040 & 0.035 & -0.033 & -0.069 & aj + 1999ac@xmath675 & ngc 6063 & 4 & 2848 & 2944 & 14.34 & 14.28 & 14.32 & 1.00 & 0.046 & 0.042 & & 0.113 & j + 1999aw & anon
1101 - 06@xmath673 & & 11992 & 12363 & 16.86 & 16.84 & 17.24 : & 0.81 & 0.032 & 0.025 & -0.076 & -0.083 & sla + 1999cc & ngc 6038 & 5 & 9392 & 9452 & 16.85 & 16.82 & 17.03 & 1.46 & 0.023 & 0.014 & & -0.018 & j + 1999cl & ngc 4501 & 3 & 2281 & 1179@xmath671 & 15.11 & 13.89 & 13.08 & 1.19 & 0.038 & 1.201 & & 0.634 & j + 1999dk & ugc 1087 & 5 & 4485 &
4181 & 15.04 & 14.95 & 15.22 & 1.28 & 0.054 & 0.051 & 0.084 & -0.064 & ka + 1999dq@xmath675 & ngc 976 & 5 & 4295 & 4060 & 14.88 & 14.68 & 14.76 & 0.88 & 0.110 & 0.123 & 0.119 & 0.047 & j + 1999ee & ic 5179 & 4 & 3422 & 3163 & 14.93 & 14.61 & 14.66 & 0.92 & 0.020 & 0.332 & 0.237 & 0.135 & st + 1999ef & ugc 607 & 5 & 11733 & 11402 & 17.52 & 17.39 & 17.69 & 1.06 & 0.087 & 0.068 & & -0.089 & j + 1999ej & ngc 495 & 0 & 4114 & 3831 & 15.65 & 15.60 & 15.66 & 1.41 & 0.071 & -0.011 & & 0.041 & j + 1999ek & ugc 3329 & 4 & 5253 & 5278 & 17.97 & 17.19 & 16.54 & 1.13 & 0.561 & 0.241 & & 0.113 & j + 1999gd & ngc 2623 & & 5535 & 5775 & 17.02 & 16.61 & 16.27 & 1.14 & 0.041 & 0.391 & & 0.342 & j + 1999gh & ngc 2986 & -3 & 2302 & 2314@xmath670 & 14.46 & 14.27 & 14.25 & 1.69 & 0.058 & 0.129 & 0.046 & 0.076 & j + 1999gp@xmath675 & ugc 1993 & 3 & 8018 & 7806 & 16.25 & 16.10 & 16.43 & 0.94 & 0.056 & 0.125 & 0.027 & -0.071 & aj + 2000b & ngc 2320 & -2 & 5944 & 6020 & 15.94 & 15.77 & 15.94 & 1.46 & 0.068 & 0.109 & 0.028 & -0.030 & j + 2000e & ngc 6951 & 3 & 1424 & 1824@xmath670 & 14.31 & 13.83 & 13.49 & 0.94 & 0.366 & 0.145 & 0.167 & 0.096 & va + 2000bk & ngc 4520 & 0 & 7628 & 7974 & 16.98 & 16.95 & 16.81 & 1.69 & 0.025 & 0.002 & 0.116 & 0.178 & ka + 2000ce & ugc 4195 & 3 & 4888 & 4940 & 17.31 & 16.66 & 16.22 & 1.00 & 0.057 & 0.621 & 0.445 & 0.404 & aj + 2000cf & mcg 11 - 19 - 25@xmath673 & & 10793 & 10805 & 17.15 & 7.10 & 17.31 & 1.27 & 0.032 & 0.034 & & -0.008 & j + 2000cn & ugc 11064 & 5 & 7043 & 6969 & 16.82 & 16.63 & 16.71 & 1.58 & 0.057 & 0.135 & -0.061 & 0.024 & j + 2000dk & ngc 382 & -3 & 5229 & 4932 & 15.63 & 15.56 & 15.74 & 1.57 & 0.070 & 0.002 & -0.062 & -0.048 & j + 2000fa & ugc 3770 & 5 & 6378 & 6533 & 15.99 & 15.95 & 16.12 & 1.00 & 0.069 & -0.001 & 0.044 & 0.013 & j + 2001v & ngc 3987 & 3 & 4502 & 4810 & 14.64 & 14.60 & 14.84 & 0.99 & 0.020 & 0.048 & 0.036 & 0.010 & vim + 2001el & ngc 1448 & 5 & 1164 & 1030@xmath670 & 12.81 & 12.73 & 12.82 & 1.13 & 0.014 & 0.088 & 0.201 & 0.095 & k + 2002bo & ngc 3190 & 1 & 1271 & 1547@xmath670 & 14.04 & 13.58 & 13.52 & 1.13 & 0.025 & 0.457 & 0.311 & 0.180 & b + 2002er & ugc 10743 & 3 & 2569 & 2804@xmath670 & 14.89 & 14.59 & 14.49 & 1.33 & 0.157 & 0.156 & 0.116 & 0.079 & pi
+ + 1991 t & ngc 4527@xmath669 & 3 & 1736 & ( 1179@xmath671 ) & 11.69 & 11.50 & 11.62 & 0.94 & 0.022 & 0.199 & ( 0.064)@xmath677 & 0.088 & ha + 1995ac & anon 2245 - 08 & 1 & 14990 & 14635 & 17.21 & 17.21 & 17.36 & 0.95 & 0.042 & -0.012 & ( -0.011 ) & 0.052 & rpa + 1995bd & ugc 3151 & 5 & 4377 & 4326 & 17.27 & 16.50 & & 0.88 & 0.498 & 0.305 & ( 0.109 ) & & ra + 1997br & eso 576 - 40 & 5 & 2085 & 2193@xmath670 & 14.04 & 13.68 & 13.44 & 1.03 & 0.113 & 0.274 & ( 0.198 ) & 0.231 & aj + 2000cx & ngc 524 & -1 & 2379 & 2454@xmath670 & 13.44 & 13.27 & 13.62 & 0.93 & 0.083 & ( 0.118 ) & ( -0.228 ) & -0.105 & al + + 1991bg & ngc 4374 & -3 & 1060 & 1179@xmath671 & 14.75 & 13.96 & 13.51 & 1.94 & 0.041 & & ( 0.005)@xmath677 & & ha + 1992k & eso 269-g57@xmath678 & 3 & 3080 & 3334 & 16.31 & 15.42 & 15.16 & 1.94 & 0.101 & & ( -0.029 ) & & ha + 1997cn & ngc 5490 & -3 & 4855 & 5092 & 16.93 & 16.39 & 16.14 & 1.90 & 0.027 & & & & j + 1998bp & ngc 6495 & -3 & 3127 & 3048 & 15.73 & 15.25 & 15.01 & 1.83 & 0.076 & & ( 0.059 ) & & j + 1998de & ngc 252 & 2 & 4990 & 4671 & 17.56 & 16.80 & 16.48 & 1.94 & 0.058 & & ( 0.053 ) & & aj + 1999by & ngc 2841@xmath669 & 3 & 638 & 896@xmath670 & 13.66 & 13.15 & 12.91 & 1.90 & 0.016 & & ( 0.011 ) & & g + 1999da & ngc 6411 & -3 & 3806 & 3748 & 16.90 & 16.16 & 15.85 & 1.94 & 0.058 & & ( 0.027 ) & & ka + 1986 g & ngc 5128 & -1 & 547 & 317@xmath670 & 12.48 & 11.44 & & 1.69 & 0.115 & ( 0.922 ) & ( 0.489 ) & & a + lrrrrrrrrrc + 1937c & -0.022 & -0.012 & & 8.823 & 8.835 & & 8.967 & 8.991 & & 28.54 + 1960f & 0.099 & -0.034 & & 11.137 & 11.171 & & 11.284 & 11.308 & & 30.86 + 1972e & -0.050 & -0.006 & -0.316 & 8.444 & 8.450 & 8.767 & 8.462 & 8.486 & 8.774 & 28.04 + 1972j & 0.081 & 0.113 & & 14.316 & 14.203 & & 14.007 & 14.031 & & 33.58 + 1974
g & 0.161 & 0.000 & & 11.816 & 11.815 & & 11.793 & 11.817 & & 31.37 + 1980n & 0.019 & 0.030 & -0.332 & 12.334 & 12.304 & 12.637 & 12.175 & 12.199 & 12.504 & 31.75 + 1981b & 0.037 & 0.005 & & 11.832 & 11.827 & & 11.794 & 11.818 & & 31.37 + 1981d & 0.193 & -0.024 & & 11.801 & 11.825 & & 11.801 & 11.825 & & 31.38 + 1982b & 0.129 & 0.057 & & 12.917 & 12.860 & & 12.958 & 12.982 & & 32.53 + 1983 g & 0.231 & -0.065 & & 12.018 & 12.083 & & 11.924 & 11.948 & & 31.50 + 1984a & 0.190 & -0.024 & & 11.669 & 11.693 & & 11.669 & 11.693 & & 31.24 + 1989b & 0.311 & 0.007 & -0.162 & 11.075 & 11.068 & 11.223 & 10.925 & 10.949 & 11.105 & 30.50 + 1990n & 0.034 & -0.040 & -0.298 & 12.520 & 12.560 & 12.858 & 12.562 & 12.586 & 12.893 & 32.14 + 1990o & -0.020 & 0.007 & -0.385 & 16.281 & 16.274 & 16.659 & 16.357 & 16.381 & 16.704 & 35.93 + 1990 t & 0.019 & -0.152 & -0.104 & 16.983 & 17.135 & 17.239 & 17.053 & 17.077 & 17.332 & 36.63 + 1990y & 0.184 & 0.078 & -0.403 & 16.985 & 16.907 & 17.307 & 16.897 & 16.921 & 17.210 & 36.47 + 1990af & 0.007 & 0.028 & & 17.739 & 17.712 & & 17.404 & 17.428 & & 36.98 + 1991s & 0.023 & -0.019 & -0.304 & 17.590 & 17.609 & 17.912 & 17.647 & 17.671 & 17.952 & 37.22 + 1991u & 0.062 & -0.064 & -0.172 & 16.191 & 16.255 & 16.425 & 16.212 & 16.236 & 16.454 & 35.79 + 1991ag & 0.013 & 0.035 & -0.367 & 14.369 & 14.334 & 14.701 & 14.469 & 14.493 & 14.753 & 34.04 + 1992a & -0.025 & 0.017 & -0.241 & 12.576 & 12.559 & 12.801 & 12.321 & 12.345 & 12.605 & 31.90 + 1992j & 0.027 & 0.036 & -0.230 & 17.547 & 17.511 & 17.741 & 17.145 & 17.169 & 17.432 & 36.72 + 1992p & 0.001 & -0.042 & -0.268 & 16.052 & 16.093 & 16.361 & 16.094 & 16.118 & 16.398 & 35.67 + 1992ae & 0.056 & -0.012 & & 18.268 & 18.280 & & 18.138 & 18.162 & & 37.71 + 1992ag & 0.170 & -0.097 & -0.380 & 15.623 & 15.719 & 16.096 & 15.612 & 15.636 & 16.112 & ( 35.19 ) + 1992al & -0.040 & -0.054 & -0.271 & 14.598 & 14.651 & 14.923 & 14.624 & 14.648 & 14.952 & 34.20 + 1992aq & 0.028 & -0.010 & -0.143 & 19.267 & 19.277 & 19.420 & 18.896 & 18.920 & 19.150 & 38.47 + 1992au & -0.045 & 0.068 & -0.343 & 18.265 & 18.197 & 18.540 & 17.840 & 17.864 & 18.205 & 37.41 + 1992bc & -0.081 & -0.061 & -0.231 & 15.337 & 15.398 & 15.630 & 15.485 & 15.509 & 15.748 & 35.06 + 1992bg & -0.030 & -0.065 & -0.221 & 16.742 & 16.807 & 17.028 & 16.740 & 16.764 & 17.040 & 36.31 + 1992bh & 0.060 & -0.032 & -0.228 & 17.369 & 17.402 & 17.629 & 17.357 & 17.381 & 17.623 & 36.93 + 1992bk & -0.016 & -0.069 & -0.068 & 18.068 & 18.137 & 18.205 & 17.750 & 17.774 & 17.992 & 37.32 + 1992bl & -0.032 & -0.009 & -0.192 & 17.411 & 17.420 & 17.613 & 17.119 & 17.143 & 17.399 & 36.69 + 1992bo & -0.007 & 0.000 & -0.136 & 15.774 & 15.774 & 15.911 & 15.396 & 15.420 & 15.632 & 34.97 + 1992bp & -0.023 & -0.046 & -0.210 & 18.330 & 18.377 & 18.587 & 18.089 & 18.113 & 18.421 & 37.66 + 1992br & 0.010 & 0.034 & & 19.196 & 19.162 & & 18.795 & 18.819 & & 38.37 + 1992bs & 0.014 & -0.025 & & 18.234 & 18.259 & & 18.204 & 18.228 & & 37.78 + 1993b & 0.054 & -0.073 & -0.234 & 18.188 & 18.261 & 18.494 & 18.099 & 18.123 & 18.447 & 37.67 + 1993h & 0.122 & 0.088 & -0.119 & 16.299 & 16.211 & 16.327 & 15.860 & 15.884 & 15.976 & 35.43 + 1993l & 0.193 & -0.027 & & 12.640 & 12.666 & & 12.415 & 12.439 & & 31.99 + 1993o & -0.031 & -0.092 & -0.158 & 17.685 & 17.777 & 17.936 & 17.635 & 17.659 & 17.922 & 37.21 + 1993ac & 0.040 & -0.033 & -0.135 & 17.606 & 17.639 & 17.773 & 17.520 & 17.544 & 17.714 & 37.09 + 1993ae & -0.057 & -0.081 & -0.105 & 15.471 & 15.552 & 15.658 & 15.284 & 15.308 & 15.543 & 34.86 + 1993ag & 0.044 & 0.024 & -0.303 & 17.682 & 17.657 & 17.960 & 17.526 & 17.550 & 17.832 & 37.10 + 1993ah & -0.023 & -0.037 & -0.286 & 16.322 & 16.359 & 16.645 & 16.117 & 16.141 & 16.502 & 35.69 + 1994d & -0.040 & -0.062 & -0.146 & 11.916 & 11.978 & 12.124 & 11.814 & 11.838 & 12.064 & 31.39 + 1994 m & 0.056 & -0.029 & -0.204 & 16.050 & 16.079 & 16.282 & 15.839 & 15.863 & 16.133 & 35.41 + 1994q & 0.035 & 0.028 & -0.338 & 16.243 & 16.215 & 16.552 & 16.329 & 16.353 & 16.597 & 35.90 + 1994s & -0.042 & -0.019 & -0.281 & 14.858 & 14.877 & 15.159 & 14.904 & 14.928 & 15.190 & 34.48 + 1994 t & 0.047 & 0.044 & -0.250 & 17.058 & 17.014 & 17.264 & 16.797 & 16.821 & 17.054 & 36.37 + 1994ae & 0.033 & -0.013 & -0.382 & 12.907 & 12.920 & 13.302 & 13.040 & 13.064 & 13.393 & 32.61 + 1995d & -0.009 & -0.009 & -0.364 & 13.234 & 13.243 & 13.607 & 13.267 & 13.291 & 13.626 & 32.84 + 1995e & 0.647 & 0.046 & & 14.347 & 14.301 & & 14.244 & 14.268 & & 33.82 + 1995ak & 0.055 & -0.103 & -0.232 & 15.734 & 15.837 & 16.068 & 15.574 & 15.598 & 15.979 & 35.15 + 1995al & 0.071 & 0.025 & -0.362 & 13.043 & 13.018 & 13.379 & 13.138 & 13.162 & 13.431 & 32.71 + 1996c & 0.026 & -0.009 & -0.262 & 16.440 & 16.450 & 16.711 & 16.528 & 16.552 & 16.769 & 36.10 + 1996x & -0.015 & -0.004 & -0.250 & 13.032 & 13.036 & 13.286 & 12.883 & 12.907 & 13.173 & 32.46 + 1996z & 0.332 & -0.036 & & 13.135 & 13.172 & & 13.070 & 13.094 & & 32.64 + 1996ab & 0.101 & -0.013 & & 19.040 & 19.053 & & 19.173 & 19.197 & & 38.75 + 1996ai & 1.611 & 0.075 & & 11.023 & 10.948 & & 11.107 & 11.131 & & ( 30.68 ) + 1996bk & 0.308 & 0.104 & & 13.641 & 13.538 & & 13.192 & 13.216 & & 32.77 + 1996bl & 0.042 & -0.034 & -0.225 & 16.550 & 16.584 & 16.808 & 16.551 & 16.575 & 16.812 & 36.13 + 1996bo & 0.257 & 0.036 & & 14.898 & 14.862 & & 14.734 & 14.758 & & 34.31 + 1996bv & 0.117 & 0.008 & -0.311 & 14.913 & 14.905 & 15.213 & 15.050 & 15.074 & 15.301 & 34.62 + 1997e & 0.024 & -0.008 & -0.202 & 14.996 & 15.003 & 15.205 & 14.807 & 14.831 & 15.064 & 34.38 + 1997y & 0.035 & -0.072 & -0.128 & 15.084 & 15.156 & 15.283 & 15.025 & 15.049 & 15.256 & 34.60 + 1997bp & 0.169 & -0.013 & -0.451 & 13.322 & 13.335 & 13.782 & 13.394 & 13.418 & 13.831 & 32.97 + 1997bq & 0.159 & 0.117 & -0.341 & 13.890 & 13.774 & 14.112 & 13.848 & 13.872 & 14.035 & 33.42 + 1997cw & 0.334 & 0.073 & -0.346 & 14.481 & 14.408 & 14.748 & 14.463 & 14.487 & 14.703 & ( 34.04 ) + 1997dg & 0.057 & -0.045 & -0.207 & 16.670 & 16.716 & 16.922 & 16.667 & 16.691 & 16.927 & 36.24 + 1997do & 0.055 & -0.018 & -0.284 & 14.103 & 14.120 & 14.403 & 14.166 & 14.190 & 14.446 & 33.74 + 1997dt & 0.496 & 0.007 & -0.189 & 13.595 & 13.588 & 13.768 & 13.611 & 13.635 & 13.768 & 33.19 + 1998v & 0.016 & -0.042 & -0.186 & 15.019 & 15.060 & 15.246 & 15.055 & 15.079 & 15.278 & 34.63 + 1998ab & 0.079 & -0.086 & -0.177 & 15.581 & 15.667 & 15.842 & 15.759 & 15.783 & 15.990 & ( 35.33 ) + 1998aq & -0.052 & -0.022 & & 12.353 & 12.375 & & 12.339 & 12.363 & & 31.91 + 1998bu & 0.279 & 0.056 & -0.191 & 11.098 & 11.043 & 11.228 & 11.013 & 11.037 & 11.140 & 30.59 + 1998dh & 0.120 & 0.072 & -0.307 & 13.522 & 13.451 & 13.755 & 13.377 & 13.401 & 13.619 & 32.95 + 1998dk & 0.130 & 0.016 & -0.299 & 14.275 & 14.259 & 14.555 & 14.278 & 14.302 & 14.544 & 33.85 + 1998 dm & 0.239 & -0.063 & -0.173 & 13.646 & 13.710 & 13.878 & 13.692 & 13.716 & 13.923 & 33.27 + 1998dx & -0.026 & -0.045 & -0.139 & 17.638 & 17.683 & 17.822 & 17.377 & 17.401 & 17.641 & 36.95 + 1998ec & 0.131 & 0.014 & -0.293 & 15.614 & 15.600 & 15.890 & 15.600 & 15.624 & 15.868 & 35.17 + 1998ef & 0.021 & -0.064 & -0.223 & 14.833 & 14.897 & 15.120 & 14.941 & 14.965 & 15.210 & 34.52 + 1998eg & 0.043 & -0.046 & -0.206 & 15.959 & 16.005 & 16.211 & 15.944 & 15.968 & 16.207 & 35.52 + 1998es & 0.055 & 0.033 & -0.354 & 13.660 & 13.626 & 13.979 & 13.761 & 13.785 & 14.032 & ( 33.34 ) + 1999x & 0.139 & -0.011 & -0.245 & 15.812 & 15.823 & 16.065 & 15.797 & 15.821 & 16.050 & 35.37 + 1999aa & -0.022 & 0.022 & -0.382 & 14.838 & 14.815 & 15.198 & 14.959 & 14.983 & 15.270 & ( 34.53 ) + 1999ac & 0.077 & -0.063 & -0.202 & 13.869 & 13.932 & 14.133 & 13.957 & 13.981 & 14.209 & ( 33.53 ) + 1999aw & -0.045 & 0.033 & -0.382 & 16.893 & 16.860 & 17.243 & 17.031 & 17.055 & 17.323 & 36.61 + 1999cc & -0.002 & 0.009 & -0.237 & 16.763 & 16.754 & 16.991 & 16.520 & 16.544 & 16.806 & 36.09 + 1999cl & 0.918 & 0.264 & -0.451 & 11.605 & 11.341 & 11.773 & 11.351 & 11.375 & 11.495 & ( 30.93 ) + 1999dk & 0.024 & 0.012 & -0.372 & 14.732 & 14.720 & 15.091 & 14.597 & 14.621 & 14.982 & 34.17 + 1999dq & 0.096 & -0.006 & -0.350 & 14.078 & 14.084 & 14.432 & 14.200 & 14.224 & 14.514 & ( 33.77 ) + 1999ee & 0.234 & 0.066 & -0.385 & 13.992 & 13.927 & 14.307 & 14.040 & 14.064 & 14.312 & 33.61 + 1999ef & -0.010 & 0.053 & -0.399 & 17.201 & 17.148 & 17.547 & 17.172 & 17.196 & 17.501 & 36.75 + 1999ej & 0.015 & -0.036 & -0.172 & 15.304 & 15.340 & 15.512 & 15.123 & 15.147 & 15.386 & ( 34.70 ) + 1999ek & 0.177 & 0.042 & -0.313 & 15.023 & 14.981 & 15.291 & 14.959 & 14.983 & 15.223 & 34.53 + 1999gd & 0.366 & 0.003 & -0.197 & 15.515 & 15.513 & 15.702 & 15.472 & 15.496 & 15.662 & 35.05 + 1999gh & 0.084 & 0.048 & -0.166 & 13.917 & 13.869 & 14.033 & 13.506 & 13.530 & 13.714 & 33.08 + 1999gp & 0.027 & 0.067 & -0.438 & 15.922 & 15.855 & 16.293 & 15.957 & 15.981 & 16.288 & ( 35.53 ) + 2000b & 0.036 & 0.066 & -0.306 & 15.530 & 15.464 & 15.769 & 15.248 & 15.272 & 15.537 & 34.82 + 2000e & 0.136 & -0.022 & -0.315 & 12.313 & 12.335 & 12.648 & 12.410 & 12.434 & 12.716 & 31.98 + 2000bk & 0.099 & -0.094 & -0.023 & 16.518 & 16.611 & 16.632 & 16.205 & 16.229 & 16.430 & 35.78 + 2000ce & 0.490 & 0.103 & -0.281 & 15.287 & 15.184 & 15.456 & 15.261 & 15.285 & 15.395 & 34.84 + 2000cf & 0.013 & 0.005 & -0.269 & 16.971 & 16.966 & 17.235 & 16.847 & 16.871 & 17.136 & 36.42 + 2000cn & 0.033 & 0.100 & -0.197 & 16.467 & 16.366 & 16.563 & 16.087 & 16.111 & 16.250 & 35.66 + 2000dk & -0.036 & 0.036 & -0.224 & 15.473 & 15.438 & 15.662 & 15.144 & 15.168 & 15.406 & 34.72 + 2000fa & 0.018 & -0.047 & -0.284 & 15.640 & 15.687 & 15.971 & 15.717 & 15.741 & 16.034 & 35.29 + 2001v & 0.032 & -0.012 & -0.308 & 14.443 & 14.454 & 14.761 & 14.502 & 14.526 & 14.799 & 34.08 + 2001el & 0.128 & -0.062 & -0.277 & 12.285 & 12.347 & 12.622 & 12.293 & 12.317 & 12.640 & 31.87 + 2002bo & 0.316 & 0.119 & -0.390 & 12.783 & 12.665 & 13.048 & 12.666 & 12.690 & 12.917 & 32.24 + 2002er & 0.117 & 0.026 & -0.259 & 13.819 & 13.793 & 14.049 & 13.644 & 13.668 & 13.907 & 33.22
+ + 1991 t & 0.143 & 0.025 & -0.338 & 11.077 & 11.052 & 11.387 & 11.141 & 11.165 & 11.417 & + 1995ac & 0.020 & -0.062 & -0.231 & 16.965 & 17.027 & 17.257 & 17.083 & 17.107 & 17.355 & + 1995bd & 0.305 & -0.033 & & 14.113 & 14.147 & & 14.255 & 14.279 & & + 1997br & 0.252 & -0.005 & -0.240 & 12.656 & 12.661 & 12.896 & 12.686 & 12.710 & 12.911 & + 2000cx & 0.007 & + + 1991bg & 0.00 & 0.749 & 0.397 & 14.582 & 13.833 & 13.436 & & & & + 1992k & 0.00 & 0.789 & 0.129 & 15.896 & 15.107 & 14.978 & & & & + 1997cn & 0.00 & 0.513 & 0.215 & 16.819 & 16.306 & 16.091 & & & & + 1998bp & 0.00 & 0.404 & 0.141 & 15.418 & 15.014 & 14.873 & & & & + 1998de & 0.00 & 0.702 & 0.245 & 17.322 & 16.620 & 16.376 & & & & + 1999by & 0.00 & 0.494 & 0.219 & 13.594 & 13.100 & 12.881 & & & & + 1999da & 0.00 & 0.682 & 0.235 & 16.662 & 15.980 & 15.746 & & & & + 1986 g & + | photometric ( @xmath0 ) and redshift data corrected for streaming motions are compiled for 111 `` branch normal '' , four 1991t - like , seven 1991bg - like , and two unusual supernovae of type ia . color excesses @xmath1 of normal sneia , due to the absorption of the host galaxy , are derived by three independent methods , giving excellent agreement leading to the intrinsic colors at maximum of @xmath2 , and @xmath3 if normalized to a common decline rate of @xmath4 .
the strong correlation between redshift absolute magnitudes ( based on an arbitrary hubble constant of @xmath5 ) , corrected only for the extrinsic galactic absorption , and the derived @xmath1 color excesses leads to the well determined , yet abnormal absorption - to - reddening ratios of @xmath6 , @xmath7 , and @xmath8 .
comparison with the canonical galactic values of @xmath9 , @xmath10 , @xmath11 forces the conclusion that the law of interstellar absorption in the path length to the sn in the host galaxy is different from the local galactic law , a result consistent with earlier conclusions by others .
improved correlations of the fully corrected absolute magnitudes ( on the same arbitrary hubble constant zero point ) with host galaxy morphological type , decline rate , and intrinsic color are derived .
we recover the result that snia in e / s0 galaxies are @xmath12 magnitudes fainter than in spirals for possible reasons discussed in the text .
the new decline - rate corrections to absolute magnitudes are smaller than those by some authors for reasons explained in the text .
the four spectroscopically peculiar 1991t - type sne are significantly overluminous as compared to branch - normal sneia .
the overluminosity of the seven 1999aa - like sne is less pronounced .
the seven 1991bg - types in the sample constitute a separate class of sneia , averaging in @xmath13 two magnitudes fainter than the normal ia . new hubble diagrams in @xmath13 , @xmath14 , and
@xmath15 are derived out to @xmath16 using the fully corrected magnitudes and velocities , corrected for streaming motions .
nine solutions for the intercept magnitudes in these diagrams show extreme stability at the @xmath17 level using various subsamples of the data for both low and high extinctions in the sample , proving the validity of the corrections for host galaxy absorption . the same precepts for fully correcting sn magnitudes we shall use for the luminosity recalibration of sneia in the forthcoming final review of our hst cepheid - sn experiment for the hubble constant . |
in recent stellar evolution calculations herwig et al . ( 1997 ) have considered diffusive overshoot for all convective boundaries which leads to a considerable change in the models .
this method provides for asymptotic giant branch ( agb ) stars a sufficient amount of dredge - up to form low - mass carbon stars as required from observations .
additionally it leads to the formation of @xmath5c as a neutron source to drive the s - process in these stars .
more details are given in herwig et al .
( 1999 ) .
however the consequences for massive agb stars , which suffer from hot bottom burning , remain to be investigated in detail .
hot bottom burning , i.e. the penetration of the convective envelope into the hydrogen burning shell , provides lithium - rich agb stars and may delay or prevent the carbon star stage by turning @xmath6c into @xmath5c and @xmath7n .
we have extended our computations of agb stars with diffusive overshooting to hot bottom burning models including a self - consistent coupling of nucleosynthesis and mixing .
we used the evolutionary code of blcker ( 1995 ) .
nuclear burning was accounted for by a nuclear network covering 30 isotopes and 74 reactions up to carbon burning .
we considered the opacities of iglesias & rogers ( 1996 ) supplemented with those of alexander & ferguson ( 1994 ) for the low temperature regime .
convection was treated within the mixing length theory ( bhn - vitense 1958 ) with a mixing - length parameter @xmath8 .
mixing of chemical elements was treated by solving a diffusion equation .
overshoot was taken into account according to the prescription of herwig et al .
( 1997 ) which is based on the hydrodynamical calculations of freytag et al .
freytag et al .
( 1996 ) showed that mixing takes place well beyond the classical schwarzschild border due to overshooting convective elements with an exponentially declining velocity field . in the overshoot region
the corresponding diffusion coefficient is given by @xmath9 with @xmath10 : velocity of the convective elements immediately before the schwarzschild border ; @xmath11 : distance from the edge of the convective zone ; @xmath12 : the overshoot efficiency parameter .
we used an efficiency parameter of @xmath13 as appropriate to match the observed width of the main sequence .
diffusive overshooting was considered in all convective regions during the complete evolution .
abundance changes have been treated self - consistently by coupling time - dependent mixing and nuclear burning of all chemical elements , i.e. by solving @xmath14_{\rm mix}\ ] ] the simultaneous treatment of burning and mixing is essential to follow , e.g. the @xmath15-production in luminous agb stars via the cameron - fowler mechanism ( cameron & fowler 1971 , see also sackmann & boothroyd 1992 ) .
the self - consistent solution of time - dependent burning and mixing processes is , however , associated with a considerable increase of computing time due to the size of the problem ( 30 isotopes , @xmath16 mass shells ) .
we calculated sequences for initial masses of @xmath17 to @xmath18 and @xmath19 from the pre - main sequence stage up to the agb and through the thermal pulses . in order to disentangle the influence of mass loss from the one of overshoot we refer in this study only to sequences with small mass - loss rates ( reimers 1975 , @xmath20 ) .
complete sequences will also consider stronger ( and better suited ) rates ( cf .
blcker 1995 ) .
on the upper agb the helium burning shell becomes recurrently unstable raising the thermal pulses ( schwarzschild & hrm 1965 , weigert 1966 ) . during these instabilities the luminosity of the he shell increases rapidly for a short time of 100yr to @xmath21 to @xmath22l@xmath2 . the huge amount of energy produced forces the development of a pulse - driven convection zone which mixes products of he burning , carbon and oxygen , into the intershell region .
because the h shell is pushed concomitantly into cooler domains h burning ceases temporarily allowing the envelope convection to proceed downwards after the pulse , to penetrate those intershell regions formerly enriched with carbon and to mix this material to the surface ( 3@xmath23 dredge up , see , e.g. , blcker 1999 for a recent review ) .
overshoot leads to an enlargement of the pulse - driven convection zone and to enhanced mixing of core material from deep layers below the he shell to the intershell zone ( `` intershell dredge - up '' ) and to a deepening of the envelope convection .
if the determination of convective boundaries is solely based on the schwarzschild criterion as in our case , dredge up can easily be obtained if some envelope overshoot is present to overcome the h / he discontinuity .
the total amount of dredge up , however , depends mainly on the strength of the former intershell dredge - up ( cf .
herwig et al .
1999 ) . as in the case of low - mass stars
intershell dredge - up leads to considerable changes of the intershell abundances .
after intershell dredge - up the abundances ( mass fractions ) of he , c , o amount to ( 40,40,14 ) instead to ( 70,25,2 ) as in non - overshoot sequences .
= 0.8 = 0.9 fig .
[ fkiptbl6 m ] shows how the lower boundary of the envelope convection as well as the pulse - driven convection zone evolve during a pulse for a 6m@xmath2 model .
the introduction of overshoot and its application to all convective boundaries provides @xmath0 dredge up even for low - mass agb stars ( herwig et al .
1997 , 1999 ) and we found very efficient dredge up for higher masses as well . with the dredge - up parameter @xmath24 being the ratio of material mixed up during the dredge up and the material burnt during two consecutive thermal pulses we find @xmath25 already during the very first pulses . correspondingly , one observes a decrease of the mass of the hydrogen - exhausted core m@xmath26 ( see fig .
[ fmhydlumi ] ) . due to the overshoot we found high temperatures at the bottom of the pulse - driven convection zone . as
demonstrated in fig .
[ ftpulsmax ] one obtains for a 3m@xmath2 sequence temperatures already close to @xmath27k , the threshold temperature of the neutron source @xmath28ne(@xmath29,@xmath30)@xmath31 mg to operate .
more massive models show even temperatures of @xmath32k and more after a few pulses .
thus , the @xmath33-process nucleosynthesis in these stars will be governed by both the @xmath5c(@xmath29,@xmath30)@xmath34o and the @xmath28ne(@xmath29,@xmath30)@xmath31 mg neutron source .
in more massive agb stars ( @xmath35m@xmath2 ) the convective envelope becomes so extended downwards that it can cut into the hydrogen burning shell during the interpulse phase ( hot bottom burning ( hbb ) or envelope burning ; iben 1975 , scalo et al .
temperatures in excess of @xmath36k ( see fig . [ ftkonv456 ] ) are reached at the base of the convective envelope and material burnt there is immediately mixed to the surface .
hbb models do not obey paczynski s ( 1970 ) classical core - mass luminosity relation but , instead , evolve rapidly to very high luminosities ( blcker & schnberner 1991 ) . due to cno cycling of the envelope @xmath37 can be transformed into @xmath38 and @xmath39 .
consequently , a low @xmath6c/@xmath5c ratio is a typical signature for hbb which can prevent agb stars from becoming carbon stars ( iben 1975 ; renzini & voli 1981 ; boothroyd et al .
1993 , frost et al .
1998 ) .
the evolution of different cno isotopic ratios at the surface of our 5m@xmath2 sequence is shown in the lower panel of fig . [ fmhydlumi ] .
the @xmath40 ratio reaches its equilibrium value ( @xmath41 ) after the first pulses and @xmath42 starts to increase whereas @xmath43 decreases .
although dredge - up is very effective leading even to a core - mass decrease , hbb is so efficient that it prevents or delays the star to become a carbon star .
the evolution of the surface abundances depends on the competition between hbb and dredge up .
both effects depend on the envelope mass , i.e. on mass loss .
consequently , the c / o ratio reached at the end of the agb evolution depends crucially on the question which process will shut down first .
the critical minimum enevelope mass for efficient hbb appears to be larger than the one necessary for dredge up ( frost et al .
1998 ) , and it depends on the remaining dredge - up efficiency if finally carbon stars can be formed .
our calculations show that dredge up even operates during the post - agb stage if overshoot is considered with important conclusions for the formation of hydrogen - deficient stars .
another direct consequence of hbb is the formation of lithium - rich stars ( scalo et al .
1975 ) via the cameron & fowler ( 1971 ) mechanism : the timescale of the @xmath44-decay of @xmath45 which is produced at the bottom of the envelope convection is comparable to the convective timescale .
therefore , before @xmath45 is burnt in the lower part of the envelope it can be mixed up to cooler layers where it decays to produce @xmath15 which , in turn , is then convected to the surface .
this mechanism is effective for temperatures between 30 and @xmath46 at the bottom of the envelope .
the production of li - rich agb stars can only be followed with a simultaneous treatment of mixing and burning .
otherwise , @xmath45 would be burnt at the bottom of the envelope before it can be mixed up . since the luminosity is a unique function of the base temperature and li
is only produced in a certain temperature range , li - rich stars are only found between ( sackmann & boothroyd 1992 ) in good agreement with observations of the magellanic clouds ( smith & lambert 1990 , plez et al .
1993 ) .
on the other hand , the majority of li - rich galactic agb stars are observed at much lower luminosities , @xmath47 to @xmath48 ( abia & isern 1997 ) .
they are , however , carbon stars .
since @xmath15 enrichment requires @xmath49k whereas @xmath37 is efficiently destroyed via cno cycling for @xmath50 k one finds in standard evolution calculations only a narrow mass , age and brightness range where agb stars should be both lithium and carbon rich indicating that possibly additional mixing processes operate in these stars ( abia & isern 1997 ) .
recently , ventura et al .
( 1999 ) calculated agb models for the lmc with @xmath51 considering overshoot in a similar manner as herwig et al .
they found li - rich carbon stars only for @xmath52 and @xmath53 , i.e. for the upper part of the observed luminosity range .
our @xmath54 model ( @xmath55 ) becomes a carbon star after 10 thermal pulses due to very efficient dredge up and can be expected to become lithium rich during the further course of evolution due to the sufficient increase of envelope base - temperatures .
the @xmath56 model becomes a carbon star after a couple of thermal pulses as well but seems to miss the li - rich stage .
these findings are in line with those of ventura et al .
thus , overshoot alone appears not to be able to explain galactic li - rich carbon for the whole brightness range observed .
= 0.75 = 0.75 fig .
[ flilum ] illustrates the considerable @xmath15 enrichment during the first thermal pulses for the 6m@xmath2 sequence . within the first 3 pulses the li abundance increases by more than 5 orders of magnitude to reach almost 10 times the main sequence value .
the maximum lithium abundance is @xmath57 + 12\approx 4.4 $ ] in agreement with the results of sackmann & boothroyd ( 1992 ) .
mazzitelli et al .
( 1999 ) calculated lithium production in agb stars with application of overshoot to all convective regions as well .
the overshoot treatment is comparable to the method of herwig et al .
( 1997 ) but convection is dealt with the canuto & mazzitelli ( 1992 ) prescription .
they obtained a similar evolution of lithium enrichment and base temperatures but found for a solar metallicity @xmath58 model a complete penetration of the hydrogen burning shell by the convective envelope resulting in a lack of helium accumulation in the intershell region and corresponding prevention of thermal pulses .
a corresponding quenching of thermal pulses was not found for the present @xmath59 sequence ( see fig .
[ flilum ] ) .
we calculated models of massive agb stars ( solar metallicity ) with a self - consistent coupling of time - dependent mixing and nuclear burning for 30 isotopes and 74 reactions . treating exponential overshoot ( freytag et al .
1996 ) as in herwig et al .
( 1997 ) and applying it to all convective regions we find very efficient @xmath0 dredge - up with @xmath60 after a few thermal pulses .
overshoot leads to intershell dredge - up and provides considerably changes of the intershell abundances strongly enriched in carbon and oxygen , i.e. [ he , c , o]=[40,40,14 ] by mass .
these intershell abundances determine the strength of the third dredge up , whereas envelope overshoot is mainly required to penetrate the h / he discontinuity .
temperatures at the bottom of the flash - driven convection zone of the helium burning shell are in excess of @xmath27k after a few pulses activating the @xmath28ne(@xmath29,n)@xmath31 mg reaction for @xmath33-process nucleosynhesis . hot bottom burning occurs for @xmath1m@xmath2 within our sequences with overshoot .
only the @xmath61m@xmath2 model becomes a carbon star whereas carbon star formation in more massive agb stars is delayed or even prevented by hot bottom burning despite of the very efficient dredge - up . with the simultaneous treatment of mixing and burning the formation of li - rich agb stars due to the cameron - fowler mechanism
was studied .
already the @xmath61m@xmath2 model becomes lithium - rich leading to the formation of a lithium - rich carbon star . for @xmath62m@xmath2
the maximum li abundance was found to be @xmath4 .
f.h . has been supported by the _ deutsche forschungsgemeinschaft , dfg _ ( la587/16 ) . | we calculated models of massive agb stars with a self - consistent coupling of time - dependent mixing and nuclear burning for 30 isotopes and 74 reactions .
overshoot with an exponentially declining velocity field was considered and applied during all stages of evolution and in all convective regions .
very efficient @xmath0 dredge - up was found even overcompensating the growth of the hydrogen - exhausted core after a few thermal pulses .
hot bottom burning occurs for @xmath1m@xmath2 within the sequences with overshoot .
carbon star formation in these more massive agb stars is delayed or even prevented by hot bottom burning despite the very efficient dredge - up . with the simultaneous treatment of mixing and burning the formation of li - rich agb stars due to the cameron - fowler mechanism was followed . for a @xmath3m@xmath2 model
the maximum li abundance was found to be @xmath4 .
# 1 # 2 _ mem .
soc .
astron .
it . _ * # 1 * , # 2 # 1 # 2 _ the messenger _ * # 1 * , # 2 # 1 # 2 _ astron .
nach . _ * # 1 * , # 2 # 1 # 2 _ astron .
astrophys . _ * # 1 * , # 2 # 1 # 2 _ astron .
astrophys .
lett . _ * # 1 * , l#2 # 1 # 2 _ astron
. astrophys .
rev . _ * # 1 * , # 2 # 1 # 2 _ astron .
astrophys .
suppl .
ser . _ * # 1 * , # 2 # 1 # 2 _ astron .
j. _ * # 1 * , # 2 # 1 # 2 _ ann .
rev .
astron .
astrophys . _ * # 1 * , # 2 # 1 # 2 _ astrophys .
j. _ * # 1 * , # 2 # 1 # 2 _ astrophys .
j. lett .
_ * # 1 * , l#2 # 1 # 2 _ astrophys .
j. suppl .
_ * # 1 * , # 2 # 1 # 2 _ astrophys .
space sci .
_ * # 1 * , # 2 # 1 # 2 _ adv .
space res .
_ * # 1 * , # 2 # 1 # 2 _ bull .
astron .
inst .
czechosl . _ * # 1 * , # 2 # 1 # 2 _ j. quant . spectrosc .
. transfer _ * # 1 * , # 2 # 1 # 2 _ mon . not .
r. astr .
soc . _ * # 1 * , # 2 # 1 # 2 _ mem .
r. astr .
soc . _ * # 1 * , # 2 # 1 # 2 _ phys .
lett .
rev . _ * # 1 * , # 2 # 1 # 2 _ publ .
astron .
. japan _ * # 1 * , # 2 # 1 # 2 _ publ .
astr .
soc .
pacific _ * # 1 * , # 2 # 1 # 2 _ nature _ * # 1 * , # 2 epsf.sty |
supernovae driven winds are a key ingredient of current galaxy formation models , in order to suppress the formation of low mass galaxies and maybe to solve the so called `` overcooling '' problem in the current hierarchical scenario of structure formation in the universe .
the proper modelling of galactic winds is a difficult and unavoidable task , both in semi
analytical models ( hatton _ et al .
_ 2003 ) and in numerical simulation ( springel & hernquist 2003 , rasera & teyssier 2006 ) of galaxy formation .
observational evidence for galactic outflows have already been pointed out by several authors ( heckman et al 2000 , adelberger _
they are usually associated to massive starbursts , for which very strong outflows are reported : for one solar mass of star formed in the galaxy , between 1 and 5 solar masses of gas are ejected in those winds ( martin 1999 ) .
this translates into a wind efficiency , usually noted @xmath0 , ranging from 100% to 500% .
the effect of galactic outflows can also be measured in the enrichment of the intergalactic medium ( igm ) as observed in absorption lines of quasars spectra ( bouch _ et al .
_ 2006 ) . despite the difficulties of modelling supernovae explosions within a turbulent , multiphase and magnetized ism , understanding the physics of the resulting large scale outflows is also a challenge .
many questions arise : what are the conditions for a galactic wind to develop and escape from the galaxy potential well ? what is the mass ejection rate of such a wind ?
what is the metallicity of the wind and other associated observational signatures ? as explored by fujita _
( 2004 ) in the context of an isolated , pre - formed galactic disc , the ram - pressure of infalling material might be the main limiting factor for galactic winds to exist .
our goal is here to present new simulations of galactic winds performed with the ramses code , using adaptive mesh refinement with a state - of - the - art shock - capturing scheme ( teyssier 2002 ) . in this paper , we follow the approach of springel _ et al . _
( 2003 ) , considering an isolated navarro , frenk & white ( 1996 ) halo in hydrostatic equilibrium , that self - consistently cools down and forms a centrifugally supported disc in its center .
star formation and the associated supernovae explosions proceed according to rather standard recipies .
we have considered that initially gas and dark matter follow the same nfw mass density profile , with 15% of the total mass in baryons in hydrostatic equilibrium .
the concentration parameter was set to @xmath1 , independant of the halo mass .
the initial angular momentum is given by the fitting formula @xmath2 of bullock et al .
( 2001 ) , normalized to the halo spin parameter @xmath3 , for which we have considered two cases : @xmath4 and @xmath5 .
the gas is radiatively cooled using a metallicity - dependant cooling function ( courty & teyssier , in prep . ) , since metal enrichment by supernovae feedback is self - consistently simulated in this study . in order to take into account the thermal feedback of supernovae explosions , we have implemented a simplified form of the multiphase model of springel & hernquist ( 2003 ) , using a polytropic equation of state in the regions of star formation ( @xmath6 @xmath7 ) and where the temperature is fixed to the polytropic equilibrium temperature . in these regions a fraction of gas
is converted into star particles using a schmidt law , with a time scale parametrized by @xmath8 .
we chose @xmath9 or @xmath10 in our simulations , resulting in different overall star formation efficiencies ( see rasera & teyssier for details ) .
kinetic feedback due to supernovae explosions that we have implemented in ramses is the key process of this study . for each star particle
formed , we create `` debris particles '' , containing both supernovae ejecta and surrounding gas entrained by the blast wave .
the velocity of the `` debris particles '' is computed according to a local sedov solution .
debris propagate freely as collisionless particles over a distance corresponding to a blast wave radius of 2 cells . at this moment , debris release their mass , metal content , momentum and energy to the gas cell they have reached .
we have performed isolated galaxy simulations with two different virial masses @xmath11 and @xmath12 in a box of 6 @xmath13 and over 6 gyr .
the coarse grid of those simulations has @xmath14 cells and the maximum resolution is about 150 pc .
for three different halo masses : @xmath15 , @xmath16 , @xmath17 ( from left to right ) .
the shaded area corresponds to the allowed range , as demonstrated by our numerical simulations .
_ right _ : comparison of the temperature map obtained at @xmath18 6 gyr for 2 different simulations , with a galactic wind breaking out of the low mass halo ( left ) , and a galactic fountain remaining close to the disc in the high mass case ( right).,title="fig:",width=226 ] for three different halo masses : @xmath15 , @xmath16 , @xmath17 ( from left to right ) .
the shaded area corresponds to the allowed range , as demonstrated by our numerical simulations .
_ right _ : comparison of the temperature map obtained at @xmath18 6 gyr for 2 different simulations , with a galactic wind breaking out of the low mass halo ( left ) , and a galactic fountain remaining close to the disc in the high mass case ( right).,title="fig:",width=245 ] following fujita _
( 2004 ) , we anticipate that the infall of gas from the halo will be the key process in the success or the failure of galactic outflows .
gas infall fuels the star forming disc with fresh gas , and therefore controls the supernovae explosion rate in the galaxy but it is also the source of ram pressure that can confine any outflowing material from the disc .
we propose here a very simple model in which the total gas mass evolves according to the following `` open box solution '' @xmath19 the accretion rate @xmath20 is computed assuming that each shell of our nfw halo free - falls down to the center , where the disc has formed .
we then compare the supernovae luminosity in the disc to the accretion luminosity of infalling gas @xmath21 where @xmath22 is the terminal velocity of free - falling gas shells , computed using the nfw mass profile .
the key parameter of the model , @xmath23 , is called the `` hydrodynamical efficiency '' and refers to the conversion efficiency of supernovae energy into galactic wind energy .
@xmath24 10% is the mass fraction of stars that explode into type ii supernovae , according to standard stellar initial mass function , @xmath25 erg is the typical energy produced by one single supernova , and @xmath26 m@xmath27 is its typical progenitor mass .
a part of the energy of supernovae explosions creates heating and turbulence in the multiphase ism .
since this energy dissipated in the ism does not contribute in the global outflow , the hydrodynamical efficiency is likely to be small .
the fraction of energy that manage to escape from the dense gaseous disc will depend on the disc characteristic ( thickness , size , gas content ) . in our simulations
, these properties will be specified by the the spin parameter of the halo and by the star formation time scale of the schmidt law . in figure
[ figure_1 ] , the wind break - out epoch ( i.e. when the wind luminosity exceeds the accretion luminosity ) is shown for various halo masses , as a function of the unknown parameter @xmath23 .
one clearly sees that the smaller the hydrodynamical efficiency , the later the wind will blow out of the disc .
more importantely , for each halo mass , a minimum efficiency is required in order for a wind to appear : 0.7% ( resp .
3% and 15% ) for a @xmath28 m@xmath29 halo ( resp . @xmath30 and @xmath31 )
our various simulations draw a similar qualitative picture : a cold centrifugally supported disc form at the halo center from the inside out , whose size depends mainly on the halo spin parameter . the star formation rate in the galaxy rises sharply ( see figure [ figure_2 ] ) and hot supernovae bubbles start to break out of the disc . after roughly 1 to 3 gyr ,
either a galactic wind develops ( for the low mass halo ) or a galactic fountain sets in ( for the high mass halo ) .
figure [ figure_1 ] shows a map of the gas temperature obtained after 6 gyr in both cases .
one clearly sees a noozle - like structure escaping from a small , thick disc in the @xmath28 m@xmath27 case , while only hot plumes and cold clumps are seen in the @xmath30 m@xmath27 case , oscillating close to a large thin disc , as they are confined by a hot , ram pressure driven atmosphere . in figure [ figure_2 ]
we have plotted the mass flux flowing out of the disc , computed in a spherical shell of size @xmath32 $ ] , in the @xmath33 case only , for various values of @xmath34 and @xmath35 .
the wind break - out epoch can be determined with great accuracy : it appears as a sharp rise in the mass outflow rate curve . comparing the apparition time of the wind to the sfr
, we can asses that a strong sfr at early epoch of galaxy formation creates a young wind that carries out a big quantity of the hot gas of the halo during the first blow . using our toy model
, we can determine the corresponding hydrodynamical efficiency , which , depending on the simulation parameters , varies between 0.8 to 2% . injecting these values in the toy model for a @xmath36 halo
, we predict that no wind can break - out in this case .
our numerical simulations do confirm this , as they show only a galactic fountain with no gas leaving the halo potential well .
when the galactic wind is fully developped , it shows a typical nozzle shape , and the mass outflow rate reaches its asymptotic value , around 0.01 m@xmath29/yr .
one can then compute the wind efficiency , defined as @xmath37 , which ranges from 10 to 20% , an order of magnitude below what is expected from lyman break galaxies observations ( martin 1999 ) .
halos calculated between @xmath32 $ ] and sfr ( right ) for the @xmath38 and @xmath39 ( solid line ) , @xmath38 and @xmath40 ( dotted line ) , @xmath41 and @xmath42 ( dashed line ) and @xmath41 and @xmath43 ( dash - dotted line).,title="fig:",width=226 ] halos calculated between @xmath32 $ ] and sfr ( right ) for the @xmath38 and @xmath39 ( solid line ) , @xmath38 and @xmath40 ( dotted line ) , @xmath41 and @xmath42 ( dashed line ) and @xmath41 and @xmath43 ( dash - dotted line).,title="fig:",width=226 ] figure [ figure_3 ] shows the mean metallicity projected along the rotation axis of the disc for different times of the galaxy evolution ( 3 gyr , 4.5 gyr , 6 gyr ) .
the @xmath44 halo begins with a moderately enriched wind ( @xmath45 ) .
the wind propagates in the igm with a metallicity growing up to about @xmath46 at 6 gyr , while the disc barely reaches @xmath47 .
the behavior of the @xmath48 halo is quite different because the gas flowing out of the disc falls back .
there are two consequences : the gas in the galactic fountain remains at a constant metallicity around @xmath49 and the disc progressively accumulates metals up to @xmath46 .
the galactic disc can be seen as a closed system where no gas is lost in the galactic wind .
halo ( left ) and the @xmath48 halo ( right ) at different times : @xmath50 ( solid line ) , @xmath51 ( dotted line ) and @xmath52 ( dashed line ) for @xmath53 and @xmath54.,title="fig:",width=226 ] halo ( left ) and the @xmath48 halo ( right ) at different times : @xmath50 ( solid line ) , @xmath51 ( dotted line ) and @xmath52 ( dashed line ) for @xmath53 and @xmath54.,title="fig:",width=226 ]
using a quiescent model of star formation in isolated galaxies , self - consistently simulated from a cooling nfw halo , we have studied the conditions for a galactic wind to break - out of the dark matter halo potential well .
our simulations have shown that no wind can form in halo of mass greater than @xmath48 , even for our most favorable couple of halo parameters ( @xmath38 , @xmath55 ) . using a simple toy model
, we understand this failure as due to the ram pressure of infalling material confining the outflowing wind .
using a more realistic cosmological setting may result into non
spherical accretion flows , and therefore to a less stringent criterion for a wind to break - out .
a proper modelling of starburst ( yet to be invented ) might also provide an easier route for increasing the feedback efficiency of supernovae - driven outflows .
99 adelberger , k. l. , steidel , c. c. , shapley , a. e. , & pettini , m. 2003 , apj , 584 , 45 bouch , n. , lehnert , m. d. , & proux , c. 2006 , mnras , 367 , l16 bullock , j. s. , dekel , a. , kollat , t. s. , et al .
2001 , apj , 555 , 240 fujita , a. , mac low , m .-
ferrara , a. , & meiksin , a. 2004 , apj , 613,159 hatton , s. , devriendt , j. e. g. , ninin , s. , et al .
2003a , mnras , 343 , 75 heckman , t. m. , lehnert , m. d. , strickland , d. k. , & armus , l. 2000 , apjs , 129 , 493 c.l .
, 1999 , apj , 513 , 156 navarro , j. f. , frenk , c. s. , & white , s. d. m. 1996 , apj , 462 , 563 rasera , y. & teyssier , r. 2006 , a&a , 445 , 1 springel , v. & hernquist , l. 2003 , mnras , 339 , 289 teyssier , r. 2002 , a&a , 385 , 337 | observations give evidences of the presence of metals in the intergalactic medium ( igm ) .
the stars responsible for transforming hydrogen and helium into more complex atoms do not form outside the galaxies in the standard scenario of galaxy formation .
supernovae driven winds and their associated feedback was proposed as a possible solution to explain such enrichment of the igm .
it turned out that a proper modelling of supernovae explosions within a turbulent interstellar medium ( ism ) is a difficult task .
recent advances have been obtained using a multiphase approach to solve for the thermal state of the ism , plus some additional recipes to account for the kinetic effect of supernovae on the galactic gas .
we briefly describe here our implementation of supernovae feedback within the ramses code , and apply it to the formation and evolution of isolated galaxies of various masses and angular momenta .
we have explored under what conditions a galactic wind can develop , if one considers only a quiescent mode of star formation . we have also characterized the distribution and evolution of metallicity in the gas outflow spreading in the igm . |
fu ori is the prototype of a small group of pre - main - sequence stars , the so - called fuors .
as explained in the introduction to @xcite , the first members were optically detectable and so could be studied in some detail .
they were distinguished by flareups of several magnitudes followed either by near - constancy or slow decline ; by broad - lined absorption spectra resembling f- to k - type supergiants in the optical region ; by powerful outflowing winds shown by p cyg structure at h@xmath5 and the d lines ; and by high abundances ( an indication of youth ) .
all the known fuors are located in star - forming regions .
these characteristics defined the fu ori class ; their observed properties have been reviewed in detail by @xcite and by @xcite . in recent years , some authors have chosen to take a broader and more inclusive view of the fuor phenomenon .
for example , the list of @xcite contains some stars that are certainly not fuors in the original sense .
we see no advantage in lumping such incompletely characterized objects in with those about which we know a great deal , such as fu ori and v1057 cyg .
so we prefer to call them fu ori - like " ( following @xcite ) until their status can be clarified .
this paper deals with fu ori .
one of best - documented fuors is v1057 cyg , which before its flare - up by about 6 mag in 1971 was a faint variable star with an emission - line spectrum like that of a t tauri star . after the flare - up
, it appeared as a high - luminosity early - type star with only h@xmath5 in emission .
it has since been in slow decline , and is now ( 2008 ) only about 1 mag ( in b ) above its pre - outburst brightness .
fu ori itself increased in brightness from @xmath6 to 10 in 19371939 , and since that time has remained near maximum , except for a slow decrease in optical brightness of about 0.15 mag per 10 years @xcite . before the flare - up , fu ori was an irregular variable of unknown spectral type . optically , it is now classified as g0 , about luminosity class ib .
the spectral energy distribution shows an ir excess , typical for a pre main - sequence star with circumstellar disk and/or dust shell .
there is also excess flux in the uv region ( @xmath7 @xmath8 2600 ) indicating the presence of a hot continuum of @xmath9 k @xcite .
fu ori has a faint red companion , probably a pre main - sequence star , at a projected separation of 05 = 250 au @xcite .
it is unknown whether the presence of this star has a connection to the outbursts .
a widely known concept is that fuors are solar mass pre
main - sequence stars that have increased dramatically in brightness due to a sudden episode of enhanced accretion triggered by some kind of disk instability .
the rate of accretion has been estimated as @xmath10 yr@xmath2 by ( * ? ? ? * hereafter pkhn96 ) by fitting an accretion disk model to the observed spectral energy distribution .
that is three orders of magnitude larger than in a classical t tauri star .
the optical spectra of fuors contain broad metallic lines that , according to ( * ? ? ?
* hereafter khh88 ) , originate in the atmosphere of the inner accretion disk at radii of 880 @xmath11 .
there is general agreement that the ir excess seen in the sed of fu ori originates in an extended envelope presumably a disk whose outline is seen in the @xmath12-band interferometric image of @xcite . in the model of hartmann & kenyon , this structure extends inward to the central star and overwhelms it : during outburst , the disk outshines the central star by factors of 1001000 " @xcite .
thus the optical spectrum of fu ori is thought to be formed in an atmosphere atop the surface of this optically thick disk , and the contribution of the central star is small or negligible .
because of the plausibility of this hypothesis , and the many subsequent papers that built upon it , it largely escaped the critical scrutiny that usually follows upon such an interesting new proposal .
a simple yet critical test of the self - luminous accretion disk model can be provided by the higher - resolution optical spectra that have now become available .
these make it possible to approach the key question : whether the _ optical _ line profiles are in fact those of a keplerian disk or not .
one of the most explicit predictions of that fuor disk model is that line width ( or @xmath0 sin @xmath1 ) should depend on wavelength and excitation potential , because the warm central regions of the disk rotate faster than the cooler regions farther out .
observational evidence for such relationships in the optical spectra of fu ori and v1057 cyg were reported by @xcite , but not confirmed later by spectroscopic data of higher quality and resolution ( @xcite , hereafter hpd03 ) .
this paper deals with that issue in more detail .
another prediction of the model was that fuor absorption lines should appear double because of keplerian splitting .
no such splitting is observed in the high - resolution spectra of fu ori described in this paper . in v1057 cyg ,
some ( but not all ) absorption lines appeared to be double , but as the star faded , it became clear that the duplicity was caused by the emergence of emission cores in the line centers that are produced in a hot region ( which was called a chromosphere " by hpd03 for lack of a better word ) .
this paper reports an analysis of absorption line profiles of fu ori on new spectra obtained beween 2003 and 2006 with the hires spectrograph on the keck i telescope on mauna kea .
the dates of those observations are given in table 1 .
the 2003 spectrograms covered the region 43506750 , while the remaining four covered 47508690 , both with interorder gaps .
the nominal resolution was 48,000 .
the slit width projected on the sky was @xmath13 .
the spectra were reduced by conventional iraf procedures .
the velocity zero - point of each exposure was checked by measures of atmospheric lines .
the new spectra of fu ori appear similar to those of 19952000 , described in detail by hpd03 , and to the single observation of 1987 discussed by @xcite .
h@xmath5 has a p cyg profile with a strong blueshifted absorption extending to about 400 km s@xmath2 that is produced by an outflowing wind .
a weaker longward emission component is also present .
both absorption and emission components vary from night to night .
the wind absorption is always strong , while the emission component may disappear completely on some occasions ( see @xcite and figs.26 and 32 in hpd03 ) . on account of the high resolution and quality of the hires
spectra , the distinctive profiles of photospheric lines are quite apparent .
a sample of both strong and weak metallic lines is shown in figure 1 .
the weaker lines have symmetric wings , while the stronger have either wider shortward wings ( e.g. , 6439 ) or a deep blueshifted absorption component ( e.g. , 5183 , 5018 ) .
the stronger lines are formed in outer layers of the atmosphere ; such profiles were described by @xcite in terms of a disk wind model .
the weaker lines have a boxy " shape , with steep wings and flat bottoms ( fig .
1 , right panel ) .
fine structure can be seen at their bottoms , and sometimes narrow minima appear at one or both edges .
such profiles persist on a timescale of months and years .
figure 2 shows a sequence of six spectra of the 75307590 region , which contains a few outstanding metal lines .
a spectrum of @xmath14 aqr ( g0 ib ) taken with the same instrument is shown for comparison . in fu ori ,
a slight profile asymmetry that changes from one spectrum to another is present : either the shortward or longward side of the profile becomes deeper .
this variable asymmetry of photospheric lines was noticed earlier in the sofin spectra of fu ori ( hpd03 ) , where a period of 3.5 days was suspected .
the time spacing of the present hires observations is so large that although variations can be detected , the periodicity can not be verified .
here we concentrate on the profiles of weak photospheric lines , which are not affected by wind / shell absorptions and thus may carry information about the structure of photospheric layers . the equivalent widths ( ew ) of and other photospheric lines in fu ori are about the same as in a g - type supergiant , but due to their large widths , those lines are very shallow . the weakest lines that we could measure on the hires spectra have ews of approximately 0.05 and central depths of about 0.015 of the continuum . in most previous spectroscopic studies of fuors ,
cross - correlation was used to improve the signal - to - noise ratio and to derive an
average " line profile over a certain spectral interval , while in these hires spectra we are able to measure the profile of each line individually .
because of the large line widths , blending is a severe problem especially in the blue , so minimally blended lines were selected .
furthermore , some lines are accompanied shortward by wind ( or shell ) components , so only the longward half of those profiles could safely be considered to be purely atmospheric .
table 2 is a list of 59 lines selected as adequately free of blends or wind structure .
the symbol @xmath15 in the last column means that only the longward half of that profile is blend - free .
the line parameters in table 2 are taken from the vienna atomic line database @xcite . as a line width parameter
, we selected the half - width at half - depth ( hwhd ) of the longward side of the profile , as illustrated in figure 3 .
the standard deviation of one line measurement is 1.52.0 kms@xmath2 .
the value of hwhd , averaged over all the lines of table 2 , is given for each observation in table 1 , where @xmath16 in the fourth column is the standard deviation of the average in km s@xmath2 .
the average hwhd is not the same on all six dates .
the overall average is 61.88 @xmath3 0.19 kms@xmath2 , so the variations of hwhd , if real , do not exceed the @xmath16 in the fourth column of table 1 .
since hwhd is determined from the red wing only , any change in the radial velocity of fu ori would cause a change in hwhd as well .
therefore , the constancy of hwhd within @xmath30.19 kms@xmath2 is compatible with constancy of the radial velocity of fu ori at that level .
radial velocity shifts can also be obtained by cross - correlating each of the six spectra of fu ori with their average .
this procedure was done for the spectral intervals : 59956030 , 65706620 , and 77057765 ; the results are listed as @xmath17rv in table 1 .
these radial velocity shifts between spectra do not exceed @xmath30.6 kms@xmath2 ; the standard deviation is 0.33 km s@xmath2 . the accuracy of this method is lower than that described above because slight changes in profile asymmetry may partly contribute to the apparent velocity shifts ( hpd03 ) .
it is safe to conclude that the radial velocity of fu ori derived from these six hires spectra was constant within @xmath30.3 km s@xmath2 over this time interval .
these new hires spectra of fu ori make it possible to compare observations with the spectra predicted by accretion disk theory . in this section
we first test whether line width depends on wavelength and excitation potential ( ep ) .
then we compare the observed line _
profiles _ to those calculated for two different models : ( 1 ) a differentially rotating keplerian disk , and ( 2 ) a rigidly rotating spherical star with a dark polar spot . .
the first models for fu ori and v1057 cyg , by khh88 , were thin accretion disks with temperature @xmath18 decreasing with distance @xmath15 from the disk center as @xmath19 and keplerian rotational velocity @xmath20 .
in pkhn96 , the model was improved by including a self - consistent treatment of the boundary layer region . during the 20 years between the middle of 1980s , when the accretion disk model ( khh88 ) was put forward , and the beginning of our present hires observations in 2003 , the brightness of fu ori has dropped by only about 0.3 mag in @xmath21 , @xmath22 , and r @xcite , so we assume that the distribution of energy in the optical region did not change significantly , and so we begin by comparing the khh88 model with our observations .
that model predicts a specific dependence of line width on wavelength and excitation potential ( ep ) , a prediction that has in many subsequent papers on fuors been cited as observational support for the theory . according to the khh88 prescription
, the disk can be considered as a sum of annuli , each of radius @xmath15 , temperature @xmath23 , and projected rotational velocity @xmath24 , also a function of @xmath15 ; @xmath1 is inclination of the disk rotational axis to the line of sight .
we use the same parameters of the model as in khh88 : radius of the central star 5.5 @xmath11 , inner radius of the disk @xmath25 , peak temperature @xmath26 = 7200 k at @xmath27 , @xmath24 = 93 kms@xmath2 at @xmath15 = 1 @xmath28 ; @xmath24 is the main parameter that determines spectral line width .
the intensity spectrum of each annulus can be represented by a synthetic spectrum emerging from the stellar atmosphere at a certain angle @xmath29 between the line of sight and normal to the stellar surface .
these synthetic spectra were calculated using the code by berdyugina ( 1991 ) and kurucz s stellar atmosphere models @xcite .
we chose @xmath30 to make the spectra more appropriate for modeling an inclined disk .
however , this parameter is not critical for line profiles in the integrated spectrum of the disk , as long as @xmath31 .
the synthetic spectra were convolved with the rotational function ( defined in khh88 ) of the corresponding annulus in the disk , and convolved with a gaussian profile corresponding to the spectral resolution of hires .
in the inner 18 annuli , within @xmath15 = 6 @xmath28 , the temperature decreases gradually from 7200 to 3500 k. within this range , the following discrete set of @xmath18 was used for calculation of the synthetic spectra : 7200 , 6500 , 5750 , 5000 , 4500 , 3750 , 3500 k , with log @xmath32 , and @xmath33 = 2 kms@xmath2 .
this is about the same as the set of spectral templates from f2 to m1 types used in khh88 .
rotational velocity was calculated explicitly for each annulus as a function of its radius . for the outer 58 annuli with @xmath34 k ( @xmath15 = 613 @xmath35 )
, high - resolution spectra of late - type stars were obtained from the eso / uves spectral library : hd 95950 ( m2 ib ) and hd 34055 ( m6 v ) . the contribution from each annulus to the integrated spectrum ( as defined by its temperature and surface area ) was taken from the khh88 model .
spectra from each annulus with its contribution factor were added up to produce the integrated spectrum of the disk .
the outer annuli with t@xmath36 @xmath8 3500
k contribute only a few percent to the optical spectrum of the disk .
moreover , the atomic lines discussed in this paper are formed within the inner regions of the disk , which are represented by the kurucz model spectra .
nevertheless , we attempt to exclude from the analysis those atomic lines that fall near the strongest breaks in the continuum caused by molecular bands in the m - type spectra , which might distort the local continuum level in the integrated spectrum .
the only difference between this procedure and that described in khh88 is that we used synthetic spectra of stellar atmospheres , while in khh88 observed spectra of template stars were used .
this may be important when fine structure of photospheric lines is considered .
first , we examine whether line width is a function of ep and wavelength in both observed and calculated spectra .
table 3 is a list of lines measured in the disk model spectrum . it is shorter than that in table 2 because of more severe blending in the predicted spectrum of the disk .
the error in the measurement of hwhd for one line is about the same ( @xmath32.0 km s@xmath2 ) as in hires spectra .
although model spectra are noiseless , there is always some uncertainty in locating the local continuum level .
figure 4 shows line width vs. ep and vs. wavelength for the hires spectrum of highest quality ( 2004 november 21 ) as filled circles , and for the model spectrum of the disk as open circles .
hwhd vs. ep ( filled circles in the left panel ) show a correlation coefficient ( cc ) of 0.18 , which with 59 data points corresponds to a false alarm probability ( fap ) of 20% ; i.e. , there is no correlation of hwhd with ep . in any case , the line slope is opposite to that expected from the disk model . on the right panel , hwhd vs. wavelength ,
the filled circles show a weak correlation ( cc = 0.34 ) in the sense of line width increasing toward longer wavelength , but again , that is in the opposite sense to that predicted by the disk model . as expected , the disk model ( open circles ) predicts a clear dependence of line width on wavelength .
( the larger scatter of points in the model plot is caused by the dependence of line width on both of these parameters . )
so , in contrast to the model prediction , the observed hwhds in fu ori show no dependence on either ep or wavelength .
this result was checked for each of the six spectra of fu ori ( see fig .
it confirms the conclusion of our previous paper ( hpd03 ) , but the higher quality of the new keck spectra makes the result more firm .
the disk model also predicts a doubling of photospheric absorption lines , which is not evident in these spectra of fu ori , where the line profiles have rather flat bottoms .
nevertheless , when the hires spectra of fu ori are cross - correlated with the spectrum of a g0 ib template ( @xmath14 aqr ) , the cross - correlation function is slightly doubled , similar to that seen in the earlier observations of 19952000 ( hpd03 ) , and in the single spectrogram of 1987 @xcite .
this is due to the narrow absorption dips flanking the flat bottom of the line profile ; this effect is discussed in an alternative model below .
line doubling may be also caused by the presence of emission cores in the line centers , an effect that actually was observed in v1057 cyg after 1995 , when that star dropped in brightness and some emission cores rose above the continuum level @xcite .
such an emission core in @xmath76707 may be present in some of our spectra of fu ori .
in our discussion @xcite of a spectrogram of fu ori obtained in 1987 january ( at a resolution comparable to hires ) , we noted that although not striking , a doubling of metallic lines was detectable at that time by cross - correlation against a standard supergiant . in the case of the hires spectra discussed here , although variable structure is apparent , no duplicity is evident on cross - correlation .
the earlier sofin spectra showed that the line profiles change from night to night ( fig . 29 in hpd03 ) , so there is no evidence of a secular change in the line structure .
the new hires spectra were also compared to those predicted by the boundary layer model of pkhn96 . in that model for fu ori
, the boundary layer is an annulus extending from the stellar surface to about 2.2 @xmath35 .
the expected rotational profile differs from the keplerian in such a way that maximum rotational velocity of 47 kms@xmath2 is at 2.2 @xmath35 , while closer to the stellar surface the rotation decreases to 33 kms@xmath2 . within the boundary layer ,
the temperature ranges from 8200 to 5400 k. outside the boundary layer the disk parameters approach those of the classical keplerian thin disk .
thus , the optical spectrum is formed mostly within the boundary layer , and all spectral lines with ep from 2 to 8 ev have about the same width .
only lines of 01 ev , formed in the distant regions of the disk , are a few km s@xmath2 narrower .
the dependence of line width on ep for the pkhn96 model is shown by the dashed line in figure 4 ( left ) .
it is much lower than the observed curve because of the low rotational velocity in the boundary layer , as specified by pkhn96 for fu ori .
the line at 6170 used by pkhn96 to fit their model to observations is not a good indicator of line width because it is a blend of three lines ( 6169.04 , 6169.56 and 6170.51 ) that are of comparable strength at @xmath18 = 50006000 k. the blend can not be resolved because the width of each component is @xmath372.5 , due to rotational broadening .
we conclude that the observed line widths of fu ori do not follow the relations predicted by the keplerian disk model .
we remark also that the narrower widths of low excitation lines of the pkhn96 model would have been obvious on these hires spectra , if present .
still more information can be retrieved from the fu ori line profiles . in figure 6 observed line profiles are compared with those of the keplerian accretion disk model .
the main differences between observed and modeled profiles dependence of the model line width on ep , and the line doubling can be seen even in this small spectral region . when the pkhn96 model is used , the match is worse , because those modeled lines are narrower and deeper than observed and have more prominent doubled structure .
the wings of lines in fu ori are steeper than in the classical case of a rotating spherical star having reasonable limb darkening .
the keplerian disk also produces a rather smooth profile . a ring or torus ( like the boundary layer )
might produce lines with steep wings , but it would also produce a prominent hump in the line center .
the flat bottom and steep wings that are actually observed are typical of another kind of object : a fast - rotating star with dark polar spot .
therefore we now examine a simple star - spot " model ( as was first applied to fu ori by @xcite ) .
a dark spot on the surface of a rotating star produces a bump in the otherwise smooth rotationally broadened profile of an absorption line .
as the star rotates , the spot passes across the visible hemisphere , and the bump travels across the line profile .
this effect is utilized in the doppler imaging technique .
if a large dark spot is located at the pole of rotation , line profiles remain disturbed throughout the period .
figure 7 shows a star with a polar spot and illustrates the consequence : a flat bottom in the line profile .
figure 8 shows how the line profile ( of 7555.59 , as an example ) responds to different model parameters .
( there are similar spots on both poles in this model , but only one is visible . ) only two free parameters are required to fit the model to the observed profiles : the inclination of the rotational axis ( @xmath1 ) and the spot radius ( @xmath38 ) .
the value of @xmath0 sin @xmath1 = 70 km s@xmath2 is fixed by the observed width of lines at continuum level .
an upper limit to the spot temperature can be set by the absence of tio bands and other m - type features in the optical region ( see fig .
if the spot were hotter , the tio band would be apparent in the stellar spectrum . with @xmath39
k , the ratio of spot to photosphere brightnesses @xmath40 would be less than 0.17 in the red . with the spot area occupying 30% of the visible hemisphere in projection , the contribution of the spot to the total spectrum would be less than 5% .
the model spectrum of the spotted star was calculated using the synthetic intensity spectrum with @xmath41 k , log @xmath42 = 2 , @xmath33 = 2 kms@xmath2 for the photosphere , and @xmath43 k for the spot .
those spectra were convolved with the instrumental profile and integrated over the corresponding areas of the visible stellar hemisphere , with the limb darkening coefficient 0.7 , inclination @xmath1 , and @xmath0 = 70/sin @xmath1 km s@xmath2 . in figure 10 the observed line profiles
are compared to those of the star - spot " model that results from this calculation .
the best match was achieved with @xmath44 and @xmath45 ( there is freedom of a few degrees in the choice of those parameters ) .
such a spot occupies 12% of the global stellar surface ( about 30% in projection on the visible stellar hemisphere ) .
note the narrow absorption dips at both sides of the flat bottom , which appear at a certain inclination and spot radius .
these features make the profile appear slightly doubled : cross - correlation of the star - spot " model spectra with the template g0 spectrum also shows that double structure .
clearly , the star - spot " model fits the observations of fu ori much better than either disk model .
note that in making this fit , we consider relatively weak photospheric lines of metals having ew @xmath37 0.1 .
the structure of stronger lines often includes shortward - shifted absorption features formed in the expanding shell and , in some cases , emission cores .
figure 11 shows the observed spectrum of fu ori around the @xmath76707 line on two occasions : the li line is strong due to the enhanced li abundance in fu ori ; the same section calculated with the star - spot " model is shown for comparison .
the two spectra of fu ori differ in the shortward wing : the deep absorption at 40 kms@xmath2 is prominent on 2005 november 23 ( it is also present at lesser strength in other spectra of fu ori ) and , when present , causes the line to appear double .
comparison to the star - spot " model shows that any emission core , if present , is very weak .
such shortward - shifted shell lines tend to be stronger in the blue region .
there is no doubt that fu ori possesses a disk : radiation from the disk and envelope dominates the ir , the outline of the disk was resolved at au scale by ir interferometry @xcite , and the distribution of energy in the infrared is compatible with a disk model .
so our question was : what is it that is seen in the optical region , a central star or an inner accretion disk ?
we found that ( 1 ) all weak photospheric lines have the same line width and profile , as expected for a rigidly rotating body , but in definite conflict with prediction for the self - luminous keplerian disk model ; ( 2 ) those profiles can be explained if the central object is a rapidly rotating high - luminosity star with a dark polar spot ; and ( 3 ) there is no sign of the line doubling , or the dependence of line width on wavelength ( in the optical region ) that is expected for the disk model . in any disk , the temperature and rotational velocity change with radius , and this defines the width of spectral lines in integrated light
. it would be strange if the disk temperature and velocity profiles were designed by nature in such a way that all atomic lines formed in different parts of the disk were of equal width .
it is more natural to assume that the optical spectrum of fu ori is produced at the central object , while the ir spectrum is formed in the circumstellar disk ( as in classical t tauri stars ) , and the 2.3 @xmath46 m co lines a characteristic feature of fuors
are formed in a cool disk or a distant shell .
certainly there is no evidence of tio bands or other m - type features in the optical region : see figure 9 .
do these results undermine the self - luminous accretion disk hypothesis in a significant way ?
no , but they do demonstrate that some modification is in order .
it is remarkable that the hypothesis still stands even though two of the strongest observational arguments that were originally urged in its favor are now seen , in the case of the prototype , to be invalid .
p. p. acknowledges the support of grant intas 03 - 516311 during this investigation , and g. h. is indebted to the national science foundation , grant ast 02 - 04021 for partial support .
we thank ann boesgaard , who obtained the two 2002 hires spectrograms of fu ori at our request , and suzan edwards , bo reipurth , and steve stahler for helpful comments .
2003 jan 11 & 52650.85 & 61.77 & 0.29 & 0.56 + 2003 feb 11 & 52681.81 & 61.64 & 0.27 & 0.10 + 2004 sep 24 & 53273.09 & 61.96 & 0.29 & 0.31 + 2004 nov 21 & 53331.05 & 61.77 & 0.24 & -0.31 + 2005 nov 23 & 53698.12 & 62.04 & 0.28 & -0.53 + 2006 dec 10 & 54079.97 & 62.16 & 0.23 & 0.29 + [ table 1 ] cccc|cccc 5383.37 & & 4.31 & & 6726.66 & & 4.61 & r + 5717.83 & & 4.28 & & 6767.79 & & 1.83 & - + 5772.15 & & 5.08 & & 6810.26 & & 4.61 & - + 5775.08 & & 4.22 & & 6814.94 & & 1.96 & r + 5862.35 & & 4.55 & & 6828.59 & & 4.64 & r + 5899.29 & & 1.05 & & 7090.38 & & 4.23 & r + 5922.11 & & 1.05 & & 7122.19 & & 3.54 & - + 5934.66 & & 3.93 & & 7344.70 & & 1.46 & r + 5965.83 & & 1.88 & & 7393.60 & & 3.61 & - + 5987.07 & & 4.80 & r & 7445.75 & & 4.26 & r + 6016.66 & & 3.55 & & 7511.02 & & 4.18 & - + 6024.06 & & 4.55 & r & 7525.11 & & 3.63 & r + 6027.05 & & 4.09 & & 7555.60 & & 3.85 & - + 6056.01 & & 4.73 & r & 7568.89 & & 4.28 & - + 6108.11 & & 1.68 & & 7574.04 & & 3.83 & - + 6180.20 & & 2.73 & & 7586.01 & & 4.31 & r + 6270.23 & & 2.86 & & 7727.61 & & 3.68 & r + 6355.03 & & 2.85 & r & 7780.55 & & 4.47 & - + 6358.70 & & 0.86 & & 7788.94 & & 1.95 & - + 6380.74 & & 4.19 & r & 7855.44 & & 5.06 & - + 6411.65 & & 3.65 & & 7912.87 & & 0.86 & - + 6439.08 & & 2.53 & & 7937.13 & & 4.33 & - + 6471.66 & & 2.53 & r & 8075.15 & & 0.92 & r + 6475.62 & & 2.56 & r & 8080.55 & & 3.30 & - + 6569.22 & & 4.73 & & 8085.18 & & 4.45 & r + 6581.21 & & 1.49 & & 8426.51 & & 0.83 & r + 6586.31 & & 1.95 & & 8611.80 & & 2.85 & - + 6707.89 & & 0.00 & & 8621.60 & & 2.95 & - + 6717.68 & & 2.71 & r & 8648.47 & & 6.21 & - + 6721.85 & & 5.86 & & & & & + 5965.83 & & 1.88 & 49.9 & 7525.11 & & 3.63 & 54.2 + 6016.66 & & 3.55 & 66.8 & 7555.59 & & 3.85 & 57.6 + 6024.06 & & 4.55 & 59.3 & 7568.89 & & 4.28 & 53.7 + 6027.05 & & 4.08 & 53.7 & 7574.04 & & 3.83 & 55.6 + 6039.72 & & 1.06 & 47.1 & 7586.01 & & 4.31 & 56.8 + 6358.69 & & 0.86 & 52.5 & 7714.31 & & 1.93 & 52.3 + 6371.37 & & 8.12 & 74.8 & 7727.61 & & 3.68 & 58.5 + 6569.22 & & 4.73 & 60.0 & 7738.96 & & 1.46 & 41.9 + 6581.21 & & 1.49 & 47.4 & 7767.44 & & 0.90 & 40.0 + 6586.31 & & 1.95 & 54.7 & 7788.94 & & 1.95 & 49.8 + 6707.89 & & 0.00 & 50.5 & 7912.87 & & 0.86 & 45.8 + 6717.68 & & 2.71 & 59.3 & 7937.13 & & 4.33 & 57.0 + 6721.85 & & 5.86 & 69.3 & 8075.15 & & 0.92 & 47.9 + 6726.66 & & 4.61 & 57.5 & 8080.54 & & 3.30 & 50.3 + 6767.77 & & 1.83 & 57.2 & 8085.18 & & 4.45 & 57.3 + 6772.31 & & 3.66 & 61.4 & 8611.80 & & 2.85 & 48.6 + 6806.85 & & 2.73 & 52.2 & 8621.60 & & 2.95 & 51.8 + 6814.94 & & 1.96 & 48.8 & 8648.46 & & 6.21 & 59.5 + 7511.02 & & 4.18 & 57.1 & & & & + [ table 3 ] | new high - resolution spectra of fu ori , obtained with the hires spectrograph at the keck i telescope in 20032006 , make it possible to compare the optical line profiles with those predicted by the self - luminous accretion disk model .
a dependence of line width on excitation potential and on wavelength , expected for a keplerian disk , is definitely not present in the optical region , nor is the line duplicity due to velocity splitting .
the absorption lines observed in the optical region of fu ori must originate in or near the central object , and here their profiles are shown to be those expected of a rigidly rotating object .
they can be fitted by a rapidly rotating ( @xmath0 sin @xmath1 = 70 km s@xmath2 ) high - luminosity g - type star having a large dark polar spot , with axis inclined toward the line of sight . over these years
, the radial velocity of fu ori has remained constant to within @xmath30.3 km s@xmath2 , so there is no indication that the star is a spectroscopic binary .
these results apply to the optical region ( @xmath4 ) ; more distant , cooler regions of the disk contribute in the infrared . |
in @xmath14 seyferts 1 and quasars there is an interesting phenomenon , the ionized absorber or warm absorber ( wa ) ( reynolds 1997 ; crenshaw et al . 2003a ; piconcelli et al . 2005 ) .
this phenomenon was observed in the x - ray band for the first time in the spectrum of the quasar mr2251 - 178 by halpern ( 1984 ) with the einstein observatory .
more recently , high resolution observations by _ chandra _ and _ xmm - newton _ have revealed the presence of blueshifted absorption lines ( v@xmath15 to @xmath16 km s@xmath4 with respect to the agn host rest frame ; collinge et al .
2001 ; kaspi et al . 2002 ) , produced by warm ionized gas ( t@xmath17 k ; krongold et al .
these reveal the nature of the wa as an outflowing wind .
the absorbing gas usually shows several components with different states of ionization , as it is demonstrated , for instance , by the different transitions of oxygen ( vii ovii , viii oviii ) , or the different transitions of iron ( vii - xii fevii - xii and xvii - xxii fexvii - xxii ) .
transitions by iv civ , v nv , and vi ovi are observed in both the x - ray and uv spectra of these sources ( with similar outflow velocities , suggesting a connection between the narrow absorption line systems in the uv and the wa ( crenshaw et al .
1999 , kaastra et al .
2000 ; kaspi et al .
2002 ; krongold et al .
2003 ; steenbrugge et al .
2005 ) .
the possibility that the absorbing components with different ionization states are in pressure equilibrium with each other , has been suggested by the analysis of several seyfert 1 galaxies : ncg 3783 ( krongold et al .
2003 , netzer et al .
2003 ) , ncg 985 ( krongold et al . 2005 ; 2009 ) , ngc 4051 ( krongold et al .
2007 ) , mrk 279 ( fields et al .
these results strongly suggests that the absorbers consist of a discrete multi - phase medium .
is important point out that these studies were carried out by fitting observations using photoionization models .
a detailed model for the wa consisting of a thermal multiphase wind , with a density - size spectrum has been presented by chelouche & netzer ( 2005 ) ; chelouche ( 2008 ) .
another representation of the absorber ( which includes calculations of line radiative transfer , but does not fit directly the data ) may be a constant pressure medium with a radially stratified density distribution ( raska et al .
2006 ) .
ngc 5548 ( classified as a seyfert 1 with broad emission lines , full width half maximum ( fwhm@xmath18 ) = 5610 km s@xmath4 ; turner et al .
1999 ) is at a redshift @xmath19 ( de vaucouleurs , 1991 ; the determination of z was performed with the detection of an hi emission line with negligible error in the velocity , heckman et al .
1978 ) , and has a luminosity of l@xmath20 erg s@xmath4 ( branduardi - raymont , 1986 ; nicastro et al .
2000 ; present work ) .
this galaxy has been studied many times in the optical , uv , and x - ray bands . the optical variability of the source allows detailed reverberation mapping studies of the broad emission line region ( belr ; e.g. peterson , 1993 ) , yielding a central black hole mass of @xmath21 m@xmath22 ( peterson 2004 ) .
this is @xmath23 times larger than , for instance , the ngc 4051 black hole mass ( 2@xmath24 m@xmath22 , peterson et al .
2004)5 , when radiation pressure is taken into account in the virial theorem . ] , for which the wa has been extensively studied and has been described as an accretion disk wind ( krongold et al .
the belr radius calculated using the h@xmath25 emission line for ngc 5548 is 21.2@xmath26 l - d ( kaspi et al .
2000 ) and the he ii belr is 16.2 l - d ( bottorff 2002 ) .
this is @xmath27 and @xmath28 times larger , respectively , than the ngc 4051 belr radius .
the x - ray flux is variable ( branduardi - raymont , 1989 ; nandra et al .
1991 ; chiang et al .
2000 ; markowitz et al .
the long - term x - ray and optical continuum light curves are highly variable and are correlated ( uttley et al .
it has been suggested that these variations are due to thermal instabilities in the inner disk ( treves et al .
the short - term x - ray variability may be more concentrated toward the center of the accretion flow ( uttley et al . ) .
low resolution x - ray studies ( e.g. nicastro et al . 2000 , using _ bepposax _
satellite observations ) showed the presence of an ionized absorber , as the data presented transitions consistent with bound - free absorption by h - like and he - like ions of o. the first high resolution spectrum of ngc 5548 , obtained with the _ chandra _ low energy transmission grating spectrometer ( letgs ) , confirmed the presence of the ionized absorber with many absorption lines from transitions such as vi cvi , vii nvii , vii - viii ovii - viii , and ix - x neix - x ( kaastra et al . 2000 ) . in the uv band , spectra obtained with the goddard high - resolution spectrograph ( ghrs ) on the hubble space telescope ( hst )
presents multiple velocity components , with intrinsic absorption lines such as ly@xmath29 , iv civ@xmath30 , and v nv@xmath31 ( mathur et al .
1999 , crenshaw et al .
five different kinematic blueshift components are observed with velocities corresponding to -166 , -336 , -530 , -667 , and -1041 km s@xmath4 ( components 5 to 1 in crenshaw et al .
far ultraviolet spectroscopic ( fuse ) spectra also show intrinsic vi ovi and ly@xmath32 absorption ( brotherton et al .
2002 ) .
ngc 5548 is the first seyfert galaxy for which a possible connection between the x - ray and uv absorbers was suggested ( mathur et al .
1995 ) given the ionization state of the medium ( though a connection has already been suggested for the quasar 3c 351 by mathur et al .
however , the exact nature of this connection in ngc 5548 has been controversial .
brotherton et al .
( 2002 ) disputed this conclusion by the discrepancies in the vi ovi ionic column densities inferred in the two bands .
arav et al .
( 2003 ) then found consistency between these columns , once the effects of velocity - dependent covering fraction in the uv band are considered .
however using simultaneous _ chandra_-letgs and hst - stis observations , crenshaw et al .
( 2003)b report that the ratio @xmath33 , in the uv component 1 , is 2.5 , which suggests a gas with low ionization state , implying low vii ovii and viii oviii column densities .
crenshaw et al . , then suggested no relation between the uv and x - ray absorbers . using _
_ high energy transmission grating spectrometer ( hetgs ) and letgs data , steenbrugge et al .
( 2005 ) proposed again a link between that x - ray and uv absorbers , given the similar kinematics seen in the absorption lines in the two bands . a new study based on 2002 uv spectra suggested that at least a part of uv component 3 should produce absorption in x - rays range ( crenshaw et al .
2009 ) . while it is clear that some connection exists , the exact relation between the uv and x - ray absorbers remains uncertain .
ngc 5548 is perhaps an exception to the idea that the wa is a multi - phase medium possibly in pressure balance ( [ intro ] ) .
a continuous radial range of ionization states was suggested for the structure of the wa in this source ( steenbrugge et al .
it should be noted that ngc 5548 has an unusually low galactic column density ( 1.6@xmath34 @xmath35 , murphy et al .
1996 ) , so the x - ray spectrum remains less absorbed at soft x - ray wavelengths than in most agns .
this broader wavelength range exposes extra atomic transitions that can provide additional constraints on wa conditions ( for instance ngc 3783 , has a very well studied absorber , but has a substantially higher galactic column density than ngc 5548 , 9.5@xmath34 @xmath35 , murphy et al .
thus , it is worthwhile to test the multi - phase scenario for the absorber in this object . for these reasons , an analysis of all the seven _
chandra _
datasets seems in order .
we re - analyzed archival _
chandra _
letgs and hetgs high resolution spectra of ngc 5548 .
despite previous results that suggest that only a continuous radial flow could model the absorber ( steenbrugge et al .
2005 ) , we show that its physical structure is also consistent with a multi - phase medium .
we contrast our results with previous studies of this source . in section 2
we describe the observations and data reduction . in section 3
we describe our data modeling . in 4 we discuss our results , and in 5 their implications .
ngc 5548 was observed seven times with the _ chandra _ x - ray observatory ( weisskopf et al . 2000 ) between 1999 and 2005 ( see table [ tab : obs ] ) .
two observations were carried out with the hetgs ( canizares et al . 2000 ) with the light dispersed on to the advanced ccd imaging spectrometer ( acis ; garmire et al .
the remaining five observations were obtained with the letgs ( brinkman et al .
2000 ) and the high resolution camara ( hrc ; murray et al .
we retrieved the primary and secondary data products of the observations from the _ chandra _ data archive .
the data reduction was performed with the _ chandra _ interactive analysis of observations software ( ciao v.3.4 ; fruscione et al .
we followed the data analysis threads provided by the _ chandra _ x - ray center .
negative and positive first order spectra were extracted and response matrices created for each observation . in this paper , we explore the time - averaged properties of the spectra , focusing on the best possible determination of the different ionization and velocity components of the well - known ionized absorber in this agn .
thus , we co - added the spectra obtained with each instrument to get the maximum signal to noise ratio ( s / n ) .
this gave a total net exposure time of 236 ks for the meg and heg spectra , and 564 ks for the leg spectrum .
the source shows flux differences among observations .
the change in flux between the 2 hetgs observations is @xmath361.7 .
the change in flux among the letgs data is a factor of 6 during the six years of observations ( see fig.[fig : lc ] ) .
the change in flux in short time scales between consecutive ( letgs ) observations is @xmath361.2 - 1.6 ( see table [ tab : obs ] ) , ( for observations separated by 3 days ) .
a time - evolving analysis is postponed to a forthcoming paper ( andrade - velazquez et al . in preparation ) .
however , we discuss the possible effects that these variations in our analysis in section [ effects ] . we have analyzed the data in the range between ( 1.57 to 10.81 ) ( [ 1.15 to 7.90 ] kev ) for the heg , ( 2.46 to 24.58 ) ( [ 0.50 to 5.04 ] kev ) for meg , and ( 8.85 to 49.15 ) ( [ 0.25 to 1.40 ] kev ) for leg .
all the data were binned by a factor of 2 , corresponding to 0.005 , 0.01 , and 0.025 per bin for the heg , meg and leg , respectively .
we have used the ciao software sherpa ( freeman et al . 2001 ) to perform spectral fits to the data . in all the models described below
, the attenuation due to the galactic absorption was accounted for with a neutral absorber ( see 1 ) .
we first focus on the hetgs ( heg and meg ) data .
then , we compare the final hetgs model with the letgs data , and find a good fit at both short and long wavelengths ( see below ) . initially , we fit the data with a power law , following previous analysis ( reynolds et al .
1997 : advanced satellite for cosmology and astrophysics -_asca _ ; nicastro et al .
2000 _ bepposax _ ) and a black - body component ( kaastra @xmath37 barr 1989 , steenbrugge et al .
2005),see table [ tab : models ] , model a. we further added four gaussians to model the evident emission lines ( see table [ tab : el ] ) .
three emission lines are resolved , with fwhm between @xmath38500 - 700 km s@xmath4 ( meg has an fwhm resolution of @xmath39 155 km s@xmath4 at 19 ) .
these lines are in the soft band ( consistent with ix [email protected] , vii ovii@xmath40 22.10 , and vi nvi@xmath40 23.27 ) .
the short wavelength ( @xmath41 ) region of the spectra shows the presence of the fourth emission line .
this corresponds to the commonly found low - ionization ( @xmath42 xvii fexvii ) fe k@xmath29 line ( 6.4 kev , nandra et al .
2007 ) , and is located at 1.937 ( without any shift from the rest frame of the host galaxy ) .
yaqoob et al .
( 2001 ) report an ew of 133 ev and a width of 4515@xmath43 and suggest a possible origin in the outer broad line region ( blr ) .
we report a full width half maximum ( fwhm ) of 2644@xmath442469 km s@xmath4 consistent with the value reported by yaqoob et al .
we find a slightly lower ew than this authors , ew@xmath45 92.2@xmath46 ev , however , a detailed analysis of this line is beyond the scope of this paper .
though model a was statistically acceptable ( @xmath47/d.o.f.= 3813/4112 ) , strong negative residuals between ( 5 - 24.5 ) are present , consistent with absorption features by ionized atoms ( xvii - xxii fexvii - xxii , vii - viii ovii - viii , ix - x neix - x , xiii - xiv sixiii - xiv , vi - vii nvi - vii , and xi - xii mgxi - xii ) due to the well - known wa in this source ( see fig.[fig : moda ] ) .
to model the features produced by the ionized absorber , we used the code phase ( krongold et al .
the code has 4 free parameters per absorbing component , namely 1 ) the ionization parameter at the illuminated face of the absorber @xmath48 , where @xmath49 is the rate of h ionizing photons , @xmath50 is the distance to the source , @xmath51 is the hydrogen number density and @xmath52 is the speed of light ; 2 ) the hydrogen equivalent column density @xmath53 , 3 ) the outflow velocity of the absorbing material v@xmath54 and 4 ) the micro - turbulent velocity v@xmath55 of the material .
usually , the micro - turbulent velocity is not left free to vary , because the different transitions due to ionized gas are heavily blended , or because the different velocity components are also blended and can not be resolved ( e.g. krongold et al .
2003 , 2005 for ngc 3783 ) . in the case of ngc 5548 , with the hetgs data resolution , it is possible to distinguish two velocity components ( see below ) .
so , despite the fact that the individual absorption lines can not be resolved , we have left the micro - turbulent velocity free to vary ( at least for the strongest absorbing components , see below ) . phase has the advantage of producing a self - consistent model for each absorbing component , because the code starts without any prior constraint on the column density or population fraction of any ion .
we have assumed solar elemental abundances ( grevesse & noels 1993 ) .
so , given the intrinsic spectral energy distribution ( sed , the approximation of this input parameter is explained in the next subsection ) of the source , the column density , and the ionization parameter of the absorbing media , the ionization balance is fixed . in all the models described below
, we have assumed that the continuum source ( including the soft excess ) is fully covered by the ionized absorbing gas while the emission lines are produced outside the region where the absorbing material is located .
the x - ray to radio sed of ngc 5548 , used to calculate the ionization balance of the gas , was defined in the following way : between the radio and the near uv ranges , we used ( non simultaneous ) data from ned ( nasa / ipac extragalactic database , references are attached in appendix [ app ] ) . in the range between 100 @xmath56 m and 912 , we performing a linear fit to the over 100 references of flux measurements found in ned . for the region between ( 0.25 - 7.90 ) kev ( [ 1.57 - 49.15 ]
) , we used the power law ( @xmath0 ) as well as the black - body contribution , as inferred from the fits to the continuum presented in table [ tab : contmh ] . for energies
@xmath57 kev , we extrapolated the derived @xmath0 in our fits up to 130 kev , where a cut - off is observed ( nicastro et al .
given that the uv sed shape has only a second order effect on the ionization balance of the x - ray species , as these charge states have ionization potentials larger than 0.1 kev ( see netzer 1996 , steenbrugge et al .
2003 ) , we have connected the unobservable region of the sed between 912 and 0.25 kev with a simple power law ( haro - corzo et al .
the radio region ( at 21 cm ) was connected with the infrared ( 100 @xmath56 m ) also with a simple power law .
figure [ fig : sed ] shows the final sed used in this work .
an important caveat of averaging the different datasets is that the the physical properties of the absorbers might have changed among observations , either because of intrinsec changes or because of changes induced by the variations in the flux level .
we note , that kaastra et al .
2002 and steenbrugge et al .
2005 report similar values for the hydrogen equivalent column density ( within the uncertainties ) for the 1999 and 2002 observations ( see table [ tab : obs ] ) .
then , we do not expect a significative change in the equivalent hydrogen column density among the seven observations .
the same is true for the outflow velocity of the absorbing gas .
however , the flux level changes between the observations used in this work . assuming the gas is in photoionization equilibrium
, this would drive changes by the same factors observed in the ionization parameter ( u ) of the gas ( producing changes in the gas opacity ; nicastro et al . 1999 ) .
there is a flux variation by a factor of 1.7 between the hetg observations , thus , the expected variation between these two observations will be delta ( logu ) = 0.23 . for the letg observations , there is a large change in flux ( by a factor of 6 during 6 years ; see fig .
[ fig : lc ] ) between the first and last two observations . in photoionization equilibrium
this would produce a change delta(logu ) = 0.77 in the ionization parameter .
thus , our determination of this parameter in the co - added data will represent a weighted average over the different observations .
we note that for the letg , the 1999 and 2002 observations have the higher s / n and similar fluxes ( different by a factor @xmath58 , and also similar to the hetg observations ) .
then , the best fit value of the absorbing gas ionization parameter on the letg co - added spectrum will be highly dominated by these observations ( the inclusion of the 2005 fainter observations contribute to increase the s / n by @xmath59 ) .
we made a new model , b , by including a single ionized absorbing component to model a. table [ tab : comp ] shows the best fit parameters values for this model .
the outflow velocity is @xmath45 -740@xmath60 km s@xmath4 and could correspond to the uv component 2 , which has a value @xmath45 -667 km s@xmath4 .
given the ionization parameter ( logu@xmath61 ) and column density ( logn@xmath62 ) of this component ( see table [ tab : models ] ) , this model fits the absorption lines produced by ions such as xvii - xxii fexvii - xxii , ix - x neix - x , xiii sixiii , xi - xii mgxi - xii , and viii oviii ( see fig . [
fig : modb ] ) .
the fit gives a significant improvement over model a ( @xmath63 for 4108 d.o.f . ;
@xmath64 ) . an f - test gives a higher than @xmath65 confidence level for the presence of the absorber ( see table [ tab : models ] ) .
the absorbing component can reproduce @xmath45 20 absorption features ( including blends ) in the wavelength region ( 5 - 20 ) .
however , significant residuals remain at 6.18 , which are probably due to a xiv sixiv transition ( @xmath406.182 ) , and at 10.62 and 11.01 , probably corresponding to transitions by xxiv fexxiv(@xmath4010.619 ) and xxiii fexxiii(@xmath4011.019 ) , see figure [ fig : modbc ] ; indicating the possible presence of another absorbing component of gas with higher ionization .
strong residuals found close to 16.5 , could correspond to inner transitions of ix feix , further suggesting the possible presence of another absorber with lower ionization state .
in addition , a detailed inspection of the spectrum near the absorption lines fit by this model ( xi mgxi , ix neix , xvii fexvii , and xiv sixiv ) reveals a complex structure in the line profile , with additional absorption with an outflow velocity of @xmath45 -1040 km s@xmath4 ( see fig .
[ fig : modbb ] ) .
these results , implying additional components and complex velocity structure , are consistent with the analysis of steenbrugge et al .
we will model these additional components in the following sections .
model c includes two absorbing components , whose parameters are again free to vary .
this model gives a @xmath66 for 4104 d.o.f .
, significantly improving over model b ( @xmath67 , an f - test , gives a confidence of larger than @xmath65 for the addition of these parameters ) .
the ionization parameters ( logu@xmath68 and logu@xmath69 ) are different by almost a factor of 3 ( see table [ tab : comp ] ) . the component with lower ionization ( with logu@xmath69 slightly lower than logu@xmath70 ) models the transitions already fitted in model b : xiii - xiv sixiii - xiv , ix - x neix - x , xi - xii mgxi - xii , xvii - xxii fexvii - xxii , and viii oviii ( see fig .
[ fig : modc ] ) .
the higher ionization component ( with logu@xmath71logu@xmath70 ) contributes weakly to fit these ions , but fits the transitions due to xxiii - xxiv fexxiii - xxiv ( not fitted in model b ) .
the two components have different outflow velocities ( v@xmath72 = -1120 km s@xmath4 and v@xmath73 = -450 km s@xmath4 ) .
we note that while model b partially fits the xiv sixiv line , the higher ionization component in model c ( logu@xmath68 ) , which has also higher outflow velocity , improves the fit to the xiv sixiv line , i. e. this component fits the residuals that model b left around 6.18 ( see fig .
[ fig : modc ] ) . while model c does reproduce , then , the lines produced by xiv sixiv , xxi - xxiv fexxi - xxiv , it can not model the complex absorption profile observed also for other lines .
motivated by the line profile residuals , we built a new model , d , by adding a third absorbing component , whose parameters are again free to vary , as are those of the other two components .
model d gives a @xmath74 for 4100 d.o.f .
, improving significantly model c ( @xmath75 , f - test confidence level of @xmath65 ) .
all the three absorbing components have different ionization level .
two of the three absorbing components have similar values of logu to those found in model c. the third component has a much lower ionization level ( by at least a factor @xmath76 , see table [ tab : comp ] ) , and fits the vi - vii nvi - vii , vii ovii , ix - xi mgix - xi lines , as well as the vii - xii fevii - xii m - shell unresolved transitions array [ uta ] ( see fig . [
fig : modd ] ) .
the three absorbing components group in two different outflow velocity systems .
one system , with outflow velocity in the range between -450@xmath60 and -590 @xmath60 km s@xmath4 , contains the systems with intermediate ( logu@xmath77 ) and low ( logu@xmath78 ) ionization level .
the other velocity system ( v@xmath54 = -1116 @xmath79 150 km s@xmath4 ) is formed by the high ionization component .
we note that model d still leaves some residuals around 16.5 , where an ix feix transition lies .
the reason for this discrepancy may be due to our use of the abbreviated data to model the uta ( see krongold et al . 2003 and behar et .
al 2001 for further details ) .
model d does not solve the problem of the line profile for several absorption lines with intermediate ionization state , in particular the xvii fexvii and x nex transitions .
therefore , model e was built by adding a fourth absorbing component to model d. this case gives a @xmath80 for 4096 d.o.f .
( @xmath81 ) , which is statistically better than that of model d ( an f - test shows that the fourth absorber is required at the @xmath65 confidence level ) .
the best - fit parameters value for each absorbing component are listed in table [ tab:2abhv ] , and the global fit to the hetgs data is shown in figures [ fig : meg ] ( for meg data ) and [ fig : heg ] ( for heg data ) .
model e again consists of two outflow velocity systems ( hereafter high velocity or hv for the system with v@xmath54 = -1110 @xmath79 150 km s@xmath4 , and low velocity or lv with v@xmath54 = -490 @xmath79 150 km s@xmath4 ) , each having two different ionization components ) ] .
the ionization parameter of the high ionization phase in the lv system ( hereafter lv - hip ) is logu=0.67 with a column density of log@xmath53= 21.25 .
the low ionization phase ( lv - lip ) has log@xmath53= 20.74 and logu=-0.49 .
the features fitted by this system are the same ones fitted in model d by the components similar to that model .
the contribution of this velocity component to the global fit is shown in figure [ fig : lv ] .
the hv system has two absorbing components .
one of the absorbing components of this system has a high ionization level ( logu=0.671 , hereafter called hv - hip ) , and a column density of logn@xmath82=21.03 , and the other absorber has an even higher ionization level ( hereafter called hv super - high ionization phase or hv - ship ) with logu=1.23 and logn@xmath82=21.73 .
figure [ fig : hv ] shows the contribution of the hv system to the global fit of the spectrum .
the hv - ship fits the same features model d does .
the hv - hip of this system fits the same absorption lines as the lv - hip ( given that the ionization parameters of these two components are quite similar ) .
the difference between these two components lies mainly in their outflow velocities , finally improving the fit to the line profile ( as seen in fig .
[ fig : lines2c ] ) .
we note that this result is obtained without any constraint in the main parameters of the absorbers , which further confirms the presence of the two velocity systems . separating the two components
was possible given the excellent resolution of the hetgs , @xmath36155 km s@xmath4 ( 19 ) for meg .
in the previous models we fitted four strong emission lines ( the fe k@xmath83 line , and the ix neix , vii ovii , and vi nvi forbidden transitions ) .
however , in the meg data , there are four additional emission lines that can be detected with a significance level @xmath84 once the absorbers are properly modelled .
these lines correspond to the resonant and intercombination transitions of vii ovii , the resonant transition of vi nvi and xiii sixiii line .
we produced a new model including these lines ( model f ) .
the best fit parameters for the eight absorption lines included to fit the hetg data are given in table [ tab : el ] .
figure [ fig : meg ] already includes these features
. our analysis does not require the inclusion of broad emission features ( as was reported by steenbrugge et al .
the forbidden ( z ) , resonant ( w ) , and intercombination ( x+y ) lines of he - like ions ( gabriel & jordan 1969 ) can be used to derive the physical conditions of photoionized plasma ( porquet @xmath37 dubau 2000 ) . in particular , the ratio between the resonant and intercombination lines serve as a diagnostic of the electronic density n@xmath85 ( r(n@xmath85 ) = @xmath86 ) . using the r value of the vii ovii lines ( r(n@xmath85)= 2.23@xmath87 ) , and following porquet @xmath37 dubau ( their figure 8)
, we find that the gas must be photoionized and must have a density n@xmath88 .
kaastra et al .
( 2000 ) report an upper limit for n@xmath89 7.0@xmath90 , with a r(n@xmath85 ) = 2.22@xmath91 , which is in agreement with our result .
the final model ( f ) for the ionized absorber in the meg and heg spectra of ngc 5548 can be further tested by extrapolating to the letgs data ( the letgs resolution is a factor @xmath45 2.5 worse than the meg resolution at 19 ) . in doing this , we implicitly assume no significant opacity variations between the co - added hetgs and letgs spectra that may arise from flux variability ( nicastro et al . 1999 , krongold et al . 2005b , 2007 ) .
given that the average flux level in these two spectra differs by only 30% ( steenbrugge et al .
2005 ) , this is a reasonable assumption .
to fit the leg spectrum , we used model f , with the 4 absorbing components parameters fixed to the values obtained for the hetgs data ( hereafter model g ) .
we also used the same continuum components ( a power law and a black - body ) , but we allow these parameters to vary freely in the fit , as the continuum level and spectral shape might be different ( in addition , the letgs is more sensitive to low energy photons and so these data allows for a better determination of the black - body temperature ) . in model g we also let free to vary the parameters for the emission lines detected in the hetgs spectra ( except for the fe k@xmath92 and xiii sixiii lines that lie outside the spectral range of the letg ) .
we further added three additional emission lines , ( located in wavelengths larger than 25 , the lower bound of the meg data ) , corresponding to v nv@xmath93 , vi nvi@xmath94 , and v cv@xmath95 ) .
an acceptable fit is obtained to the letgs spectrum with this simple method ( @xmath96 ) , implying that the hetgs spectra alone can constrain in a reliable way the properties of the ionized absorber .
the best - fit continuum parameters for the letgs spectrum are collated in table [ tab : contmh ] .
table [ tab : el ] presents the parameters for the emission lines . the best - fit model over the data
is presented in figure [ fig : leg ] ( see also table [ tab : models ] ) .
this figure show that most absorption lines observed in the hetg data are well reproduced by the model in the letgs spectrum .
it also shows that model g fits well more than 12 low ionization absorption lines in the region ( 26 - 49 ) .
these lines correspond to transitions of ions such as v cv , v - vi nv - vi , and xii sxii . in the uta region
there are discrepancies between data and model , but as for the hetgs data , they are produced likely due to the use of the abbreviated data to model this feature .
we note that some low ionization lines show residuals consistent with an outflow velocity of -1040 km s@xmath4 as shown in the figure [ fig : lin5to ] .
this may indicate the presence of a hv - lip ( see [ d ] ) .
however , attempting to model these residuals with an additional absorbing component does not improve the fit in a statistical way .
the best fit model of the wa for the seyfert 1 galaxy ngc 5548 requires two outflow velocity systems , each with two different ionization phases ( [ model ] ) .
figure [ fig : lines2c ] shows that the profiles of the absorption lines ( produced by ions such as x nex , xviii fexviii , viii oviii , and xi mgxi ) are fit better with two narrow line profiles than with a single broad component .
each of these two velocity components requires absorption by gas with very different ionization level .
the lv component consists of gas with two different ionization states ( the lv - lip and the lv - hip ) .
the hv component requires also two absorbers with different ionization ( the hv - ship and the hv - hip ) .
these four absorbers , divided into two velocity systems , can reproduce all the spectral features observed in the spectra ( vii - xxiv fevii - xxiv , vi - viii ovi - viii , ix - x neix - x , ix - xii mgix - xii , xii - xiv sixii - xiv , v - vi cv - vi , ix - xii six - xii , and xi - xii arxi - xii ) .
we note that the hv system might also have an additional , low ionization component ( hv - lip ) .
figure [ fig : lin5to ] shows a double line profile for several low ionization level lines , suggesting a lip with higher outflow velocity ( -1041 km s@xmath4 ) . while the inclusion of this component is not statistically required by the data , its physical properties appear to be well constrained ( as can be observed in in figure [ fig:5aregp ] , where the 1 , 2 , and 3@xmath97 confidence regions for the ionization parameter vs. the column density of this component are presented ) .
further evidence for the presence of this component arises from the fact that uv system 1 can be associated with the hv absorber ( see [ c ] ) , and the hv - ship and hv - hip have a very hgh ionization state to produce such absorption ( ( see tables [ tab : ionden ] and [ tab : ionden2 ] ) . we conclude that the data clearly shows a shift to higher ionization material at higher velocity in ngc 5548 wa , with the lip becoming weak and the ship dominating in the hv system . on the other hand , the lv system does not show the presence of the ship .
the separation between components is a new and important feature in x - ray spectra of ionized absorbers , and could be achieved because of the large difference in velocity between the two systems ( @xmath98 km s@xmath4 ) , as well as the different dominant ionization levels between them ( see also steenbrugge et al .
2005 ) . in the study of the narrow vi ovi , v nv , and iv civ absorption lines in the uv region , crenshaw et al .
( 2003 ) finds five separate kinematic components in @xmath99 -stis and @xmath100 spectra .
our analysis reports two kinematic components ( hv and lv systems ) .
the hv system shows similar outflow velocity to the uv absorbing component 1 ( @xmath101 km s@xmath4 ) . according to crenshaw et al .
( 2003)b , the gas producing absorbing component 1 should produce only a weak x - ray wa , which is consistent with the marginal detection of the hv - lip in the _ chandra _ data ( [ d ] ) .
there is also a correspondence between the x - ray lv system and uv absorbing components 2 and 3 ( with outflow velocities @xmath102 and @xmath103 km s@xmath4 , respectively ) . crenshaw et al .
( 2003)b and crenshaw et al .
2009 suggest that at least component 3 ( and possibly component 2 ) should produce absorption in the x - ray range .
this is consistent with the detection of the lv - lip .
there is no x - ray detection of uv absorbing systems 4 and 5 ( with velocities @xmath104 and @xmath105 km s@xmath4 respectively ) , although we can not rule out a possible contribution from component 4 to the lv - lip , given the uncertainty in the outflow velocity of the x - ray systems . according to the analysis of the uv data by crenshaw et al . , system 5 has the lowest column density ( logn@xmath106=19.27 ) .
krongold et al .
( 2003 ) discuss how the difference in sensitivity of the current x - ray and uv detectors may affect the detection or non - detection of a given uv component in the x - rays .
their figure 11 shows the observational lower limits for the equivalent widths of the iv civ@xmath107 , vi ovi@xmath108 , vii ovii@xmath109 , viii oviii@xmath110 lines detectable in the uv and x - ray bands , as a function of the ionization parameter and h equivalent column density .
thus , given their ionization state and column density , systems 4 and 5 are consistent with not being detected in the x - rays , given the limited sensitivity of the x - ray observations . while a kinematic connection between the components is evident , an important question is how the predicted / measured column densities for the different analyses compare between them .
the ionic column densities predicted by our model are now listed in table [ tab : ionden ] and [ tab : ionden2 ] for the lv and hv systems , respectively . in these tables
, we also list the column densities predicted ( or measured ) by crenshaw et al .
( 2009 ) , based on uv data , and by steenbrugge et al .
( 2005 ) , based on the letg 2002 data ( their model b , see [ cps ] ) .
the column densities for ions civ , nv y ovi between our predicted values and those measured by crenshaw et al .
( 2009 ) in the uv differ by factors of 16 , 4 and 1.2 , respectively .
we note that ovi ( where no diference is found ) is detected in the x - ray analysis as well .
the differences between the uv and x - ray columns could be produced because of 3 different factors : 1 ) the blending of uv components 2 and 3 in the x - rays ( and maybe an additional contribution from uv component in our lv - lip model .
2 ) the different seds used in the 2 analysis : the sed used by crenshaw et al . includes much more photons in the far ( unobservable ) uv , and 3 ) a detailed analysis on the dependence of the covering factor on velocity could not be performed in the analysis of the uv line profile .
this could underestimate the columns as shown by arav et al .
2003 . taking into account these effects
, we consider that the predictions of our model for the columns measured in the uv are very reasonable , and conclude that the lv - lip component can produce the absorption features produced by uv component 3 ( and maybe 2 and 4 ) .
however , we note that the values predicted by crenshaw et al . for the high ionization charge states differ from those measured in this analysis . the reason for this discrepancy , is that these ions are not observable in the uv , and only three ions ( civ , nv y ovi ) were used in the uv model . with these considerations , and given the strong coincidence in kinematics between the uv and x - ray outflows ,
we conclude that a common origin is not only possible , but quite likely .
such connection was originally suggested by mathur et al .
( 1994 ) for ngc 5548 , and has also been suggested for several other objects ( ngc 985- krongold et al .
2005 ; ngc 3783-gabel et al .
2003 , krongold et al . 2003 ,
netzer et al .
2003 ; ngc 7469- blustin et al .
2003 , kriss et al . 2003 ) .
thus , these studies on the wa may represent a general feature of the structure for quasars , as proposed by elvis ( 2000 ) .
the presence of two different absorbing components with different temperature but similar outflow velocity suggests that the absorber may be formed by a multi - phase medium ( e.g. elvis , 2000 ; krongold et al .
given the two velocity systems in ngc 5548 , we could be seeing the presence of two different outflows ( with different outflow velocity ) , each being a multi - phase medium .
this is further supported by the pressure equilibrium of the two absorption components in each velocity system . in figure
[ fig : sc1 ] we show the thermal stability curve of the gas ( also known as the `` s - curve '' , krolik et al .
1981 ) for the sed used in our analysis ( 3.2.1 ; fig .
[ fig : sed ] ) .
the s - curve marks the points of thermal equilibrium in the t vs. u / t plane , where t is the photoionization equilibrium temperature of the gas , and u / t is inversely proportional to the gas pressure ( u / t @xmath111 ) . in figure
[ fig : sc1]a , the position on the s - curve of the two lv absorbing components is shown .
the lv - hip and the lv - lip have quite different temperatures but , within the errors , are consistent with a single value of the u / t , and so can be in pressure equilibrium .
figure [ fig : sc1]b shows the hv - ship and hv - hip on the s - curve .
again , the pressure equilibrium between the two components is allowed .
note that , given the possible range of ionization parameters of the tentative hv - lip ( figure [ fig:5aregp ] ) , this component , if indeed present ( as suggested by our results and by those by crenshaw et al .
2009 ) , could also be in pressure equilibrium with the higher ionization ones that form this system .
thus , both the hv and the lv systems are formed by a two - phase medium ( or maybe three - phase for the hv system ) , with one component pressure confining the others ( see krongold et al .
2005a for a more detailed explanation ) .
ngc 5548 thus joins the group of agns for which pressure equilibrium applies : ngc 3783 ( krongold et al .
2003 , netzer et al .
2003 ) , ngc 985 ( krongold et al . 2005 , 2008 ) , ngc 4051 ( krongold et al .
2007 ) , mrk 279 ( fields et al .
2007 ) , ugc 11763 ( cardaci et al .
no other well studied wa in a seyfert galaxy is inconsistent with the description of a multi - phase medium .
we note however , that we will further test this idea for the wa over a large sample of objects ( andrade - velazquez et al . in preparation ) .
ngc 5548 is the first agn in which a two - phase wa medium is likely for two velocity systems .
the presence of a multi - phase medium obviously requires that the different ionization components that form it lie at the same distance from the ionizing continuum .
we do not have the means to test this requirement ( though see krongold et al . 2010 ) .
however , for ngc 4051 , krongold et al .
( 2007 ) measure the distance of the two absorbing components and found that they were indeed consistent with being co - located , giving further plausibility to multi - phase absorbing winds . the ship of the hv system and the hip of both the lv and hv systems lie on unstable regions of the s - curve , implying a possible inconsistency with the multi - phase scenario .
however , the shape of this curve depends on the exact shape of the sed of the source , which has major uncertainties , particularly in the extreme uv region ( haro - corzo et al .
the chemical composition of the absorbing gas has an important effect on this curve also ( see komossa & mathur , 2001 , fields et al .
in addition , recent studies of the s - curve with new and more reliable dielectronic recombination rate coefficients ( badnell et al .
2006 ) show that there is a larger probability of having a thermally stable wa at 10@xmath112 k , ( crenshaw et al .
( 2009 ) obtain a t @xmath45 1@xmath113 for the uv component 3 ) , and then a greater plausibility for its multiphase nature ( chakravorty et al .
2008 ) .
steenbrugge et al .
( 2005 ) analyzed three _
_ grating observations of ngc 5548 ( those carried out on 2002 , table 1 ) with a total exposure time of 494 ks .
they presented a model showing that a continuous distribution of column densities and ionization parameters could fit the data .
they further concluded that the data was not consistent with discrete phases of material in pressure equilibrium .
this conclusion is in contradiction to the results presented here , making a detailed comparison of both analysis necessary . in their model c , steenbrugge et al .
( 2005 ) explored a model similar to the one presented here .
they used a self - consistent model that takes advantage of photoionization balance calculations to make a global fit to the data .
they concluded that following this approach , and fitting fe separately , a good fit could be obtained , but constraining only two absorbing components .
they further tested the presence of discrete components by adjusting equivalent h column densities and ionization parameters to the absorbing components of their model b , where ionic column densities were measured on an ion by ion fit . in this case
they find that at least five different components with different ionization state were required .
this approach led them to conclude that discrete components were not a good representation of the data .
the most fundamental difference between the analysis carried out by steenbrugge et al .
( 2005 ) and the present analysis is that their models b and c used only a single velocity outflow with a large turbulent broadening to fit the data
. however , as they report in their paper , and as discussed here , the data are better fitted with two different velocity components , the hv system and the lv system .
furthermore , as shown here ( and as also noted by steenbrugge et al . )
the degree of ionization is different for the different velocity components , as the hv system becomes dominant for high ionization states .
this led them to conclude that more and more ionization components were required to fit the data and thus a structure of discrete clumps of gas was unlikely .
however , here we have shown that , when each velocity component is modelled independently , both can be well described by two ionization components , making the presence of discrete components plausible and likely .
furthermore , the addition of extra components does not improve the fit .
so , given the difference between number of free parameters from the different models , we can claim by occam s razor that our global approach is preferable .
table [ tab : ionden ] and [ tab : ionden2 ] show the column densities for our model and the steenbrugge et al .
model b. the differences found between our analysis and that by steenbrugge et al .
( that analysis shows sistematically larger columns than ours ) are likely produced because only one velocity component was used by steenbrugge et al . to model the line profile produced by the two velocity systems .
a detailed fitting including the line profile of each velocity system produces more reliable column densities .
steenbrugge et al .
( 2005 ) further tested whether their five discrete components could be in pressure balance , and they concluded that they could not .
in fact , their s - curve ( their fig.10 ) has no region where the slope is negative , making it impossible for the gas to be in pressure balance .
we stress again that the shape of the s - curve depends strongly on the metallicity of the absorbing gas , and the shape of the sed used to produce it ( in this work we use the best multiwavelength data available extracted from ned ) . given this , and the new results by chakravorty et al .
( 2008 ) , we conclude that the possibility that the different phase are in pressure equilibrium can not be discarded . here
we have shown that when the velocity systems are modelled independently , each one is consistent with pressure balance .
we further note that chelouche ( 2008 ) also find that the wa in ngc 5548 can be described by a multiphase medium .
the author models the ionized flow assuming a thermal wind . while his model is more physically driven than ours , it does not take into account the two different velocity systems revealed by the data .
however , given the similar results , the idea of a multi - phase medium absorber is reinforced .
the presence of two different velocity systems , each formed by gas with very different ionization state , can give a clue about the structure of the flow .
in a continuous * radial * range of ionization structures , the wind must extend in our line of sight for parsecs ( or more ) , forming a large scale outflow along our line of sight .
we find difficult to reconcile such a distribution of material with the two velocity systems found in ngc 5548 , even if inhomogeneities in the flow ( like those suggested by ramrez et al . 2005 ) are invoked .
the presence of the two velocity components , then , suggests that we are not seeing the wind in a radial direction . on the other hand ,
the presence of the different kinematic components can be naturally explained if we are looking at an outflow in a transverse direction , which is consistent with the detection of transverse motion in the uv absorbers ( marthur et al .
1995 , arav et al .
2002 , crenshaw et al .
2003 ) . in this case , an important component of the velocity ( and acceleration ) of the wind could take place in the direction perpendicular to our line of sight .
such a configuration for the wind is further supported by the finding that the different phases observed at each velocity form a single multi - phase outflow . if the wind arises from the inner accretion disk ( konigl & kartje , 1994 , elvis et al .
2000 , and krongold et al 2007 ) , and flows in a transverse direction with respect to our line of sight , the two velocity components could be simply explained as forming in different regions of the disk ( elvis , 2000 ) .
the two velocity components in ngc 5548 are consistent with originate in the accretion disk , as discussed by krongold et al .
furthermore , crenshaw et al . ( 2009 ) suggest that uv component 3 ( clearly associated to the lv - lip ) , should be accelarated by magnetocentrifugal forces , and thus should form at an accretion - disk scale ( bottorff et al .
we note that an accretion - disk wind does not imply a small scale outflow .
the wind could still extend for large scales , but our line of sight would cross it only at a given distance from the central ionizing source .
we find that the wa of ngc 5548 can be modelled in a simple way with two different velocity systems ( v@xmath54 = -1110 km s@xmath4 and v@xmath114 km s@xmath4 ) each with two phases of gas in pressure equilibrium .
this results strongly suggests that the structure of the absorber in ngc 5548 could consist of two different outflows ( with different outflow velocity ) , each formed by a multi - phase medium in pressure balance .
the ionized absorber in ngc 5548 now fits a pattern established for the other well studied systems .
there are no remaining cases in which a multi - phase medium is inconsistent with the observations .
the presence of two velocity systems , each formed by a multi - phase medium suggests that we are looking at the flows in a transverse direction with respect to our line of sight ( as also suggested by the uv data ) .
the velocity of the two x - ray absorbing systems is in agreement with components 1 , 2 , and 3 found in the uv band ( crenshaw et al .
the x - ray data are further consistent with the non detection of uv components 4 and 5 , given the limited sensitivity and resolution of the current x - ray ( compared to those in the uv ) .
there is also a reasonable agreement in the column densities of civ , nv , and ovi predicted by our x - ray analysis , and those measured in the uv .
thus , our models elegantly solve the long - standing problem of apparent discrepancies between the uv and x - ray absorbers , and give further support to the idea that uv and x - ray absorbers are part of the same outflow .
this work was supported by the unam papiit grant in118905 and the conacyt grant j-49594 .
a - vm acknowledges support from conacyt scholarship .
nsb acknowledges support from nasa to the chandra x - ray center through nas8 - 03060 .
this work was supported by nasa grant nnx08ab81 g .
1 arav , n. , kaastra , j. , steenbrugge , k. , brinkman , b. , edelson , r. , korista , k. t. , & de kool , m. 2003 , , 590 , 174 badnell , n. r. 2006 , , 651 , 73b .
behar , e. , sako , m. , & kahn , s. m. 2001 , , 563 , 497 blustin , a. j. , et al .
2003 , , 403 , 481 bottorff , m. c. , korista , k. t. , & shlosman , i. 2000 , , 537 , 134 bottorff , m. c. , baldwin , j. a. , ferland , g. j. , ferguson , j. w. , & korista , k. t. 2002 , , 581 , 932 branduardi - raymont , g. 1986 , the physics of accretion onto compact objects , 266 , 407 branduardi - raymont , g. 1989 , active galactic nuclei , 134 , 177 brinkman , bert c. et al . 2000 , , 4012 , 81 brotherton , m. s. , green , r. f. , kriss , g. a. , oegerle , w. , kaiser , m. e. , zheng , w. , & hutchings , j. b. 2002 , , 565 , 800 canizares , c. r. , et al .
2000 , , 539 , l41 cardaci , m. v. , santos - lle , m. , krongold , y. , hgele , g. f. , daz , a. i. , & rodrguez - pascual , p. 2009
, , 505 , 541 chakravorty , s. , kembhavi , a. k. , elvis , m. , ferland , g. & badnell , n. r. 2008 , , 384 , 24 chelouche , doron ; netzer , hagai 2005 , , 625,95c chelouche , doron 2008 , arxiv0812.3621c chiang , j. ; reynolds , c. s. ; blaes , o. m. ; nowak , m. a. ; murray , n. ; madejski , g. ; marshall , h. l. ; magdziarz , p.2000 , , 528 , 292c collinge , m. j. ; brandt , w. n. ; kaspi , shai ; crenshaw , d. michael ; elvis , martin ; kraemer , steven b. ; reynolds , christopher s. ; sambruna , rita m. ; wills , beverley j. 2001 , , 5557 , 2 crenshaw , d. m. , kraemer , s. b. , boggess , a. , maran , s. p. , mushotzky , r. f. , & wu , c .- c .
1999 , , 516 , 750 crenshaw , d. m. , kraemer , s. b. , & george , i. m. 2003a , , 41 , 117 crenshaw , d. m. ; kraemer , s. b. ; gabel , j. r. ; kaastra , j. s. ; steenbrugge , k. c. ; brinkman , a. c. ; dunn , j. p. ; george , i. m. ; liedahl , d. a. ; paerels , f. b. s. ; turner , t. j. ; yaqoob , t. 2003b , , 594 , 116 crenshaw , d. m. , kraemer , s. b. , schmitt , h. r. , kaastra , j. s. , arav , n. , gabel , j. r. , & korista , k. t. 2009 , , 698 , 281 de vaucouleurs , g.1991 , science , 254 , 1667 elvis , m. 2000 , , 545 , 63 fields , d. l. , mathur , s. , krongold , y. , williams , r. , & nicastro , f. 2007 , , 666 , 828 freeman , p. , doe , s. , & siemiginowska , a. 2001 , , 4477 , 76 fruscione , a. 2002 , chandra news , 9 , 20 gabel , j. r. , et al.2003 , , 595 , 120 gabriel , a. h. , & jordan , c. 1969 , , 145 , 241 garmire , g. p. , bautz , m. w. , ford , p. g. , nousek , j. a. , & ricker , g. r. , jr .
2003 , , 4851 , 28 grevesse , n. , & noels , a. 1993 , origin and evolution of the elements , 15 halpern , j. p. 1984
, , 281 , 90h haro - corzo , s. a. r. , binette , l. , krongold , y. , benitez , e. , humphrey , a. , nicastro , f. , & rodrguez - martnez , m. 2007 , , 662 , 145 heckman , t. m. ; balick , b. ; sullivan , w. t. , iii kaastra , j. s. ; barr , p. 1989
, , 226 , 59k .
kaastra , j. s. , mewe , r. , liedahl , d. a. , komossa , s. , brinkman , a. c. 2000 , , 354 , 83 kaastra , j. s. , steenbrugge , k. c. , raassen , a. j. j. , van der meer , r. l. j. , brinkman , a. c. , liedahl , d. a. , behar , e. , & de rosa , a. 2002 , , 386 , 427 kaspi , shai ; smith , paul s. ; netzer , hagai ; maoz , dan ; jannuzi , buell t. ; giveon , uriel 2000 , , 533 , 631 kaspi , s. , et al .
2002 , , 574 , 643 komossa , s. ; mathur , s. 2001 , , 374 , 914 konigl , a. , & kartje , j. f. 1994 , , 434 , 446 kriss , g. a. ; blustin , a. ; branduardi - raymont , g. ; green , r. f. ; hutchings , j. ; kaiser , m. e. 2003 , , 403 , 473 krolik , j. h. , mckee , c. f. , & tarter , c. b. 1981 , , 249 , 422 krongold , y. , nicastro , f. , brickhouse , n.s . , elvis , m. , liedahl d.a . &
mathur , s. 2003 , , 597 , 832 ( k03 ) krongold , y. , nicastro , f. , elvis , m. , brickhouse , n. s. , mathur , s. , & zezas , a. 2005a , , 620 , 165 krongold , y. , nicastro , f. , brickhouse , n. s. , elvis , m. , & mathur , s. 2005b , , 622 , 842 krongold , y. , nicastro , f. , elvis , m. , brickhouse , n. , binette , l. , mathur , s. , & jimnez - bailn , e. 2007 , , 659 , 1022 krongold , y. , et al.2009 , , 690 , 773 krongold , y. , et al .
2010 , mathur , smita ; wilkes , belinda ; elvis , martin ; fiore , fabrizio .
1994 , , 434 , 493 .
mathur , s. , elvis , m. , & wilkes , b. 1995 , , 452 , 230 mathur , s. , elvis , m. , & wilkes , b. 1999 , , 519 , 605 markowitz , a. ; edelson , r. ; vaughan , s. 2003 , , 598 , 935 m .
marconi , alessandro ; axon , david j. ; maiolino , roberto ; nagao , tohru ; pastorini , guia ; pietrini , paola ; robinson , andrew ; torricelli , guidetta .
2008 , , 678 , 693 m murphy , e. m. , lockman , f. j. , laor , a. , & elvis , m. 1996 , , 105 , 369 murray , stephen s. , chappell , jon h. , kenter , almus t.,juda , michael , kraft , ralph p. , zombeck , martin v. , meehan , gary r. , austin , gerald k. , gomes , joaquim j. 2000 , , 4140 , 144 nandra , k. , pounds , k. a. , stewart , g. c. , george , i. m. , hayashida , k. , makino , f. , & ohashi , t. 1991 , , 248 , 760 nandra , k. , oneill , p. m. , george , i. m. , & reeves , j. n. 2007 , , 382 , 194 netzer , hagai 1996 , , 473 , 781 netzer , h. , et al.2003 , , 599 , 933 nicastro , f. , fiore , f. , perola , g. c. , & elvis , m. 1999 , , 512 , 184 nicastro , f. , et al.2000 , , 536 , 718 peterson , b. m. 1993 , , 105 , 247 peterson , b. m. 2004 , the interplay among black holes , stars and ism in galactic nuclei , 222 , 15 piconcelli , e. ; jimenez - bailn , e. ; guainazzi , m. ; schartel , n. ; rodrguez - pascual , p. m. ; santos - lle , m. 2005 , , 432 , 15 porquet , d. , & dubau , j. 2000 , , 143 , 495 ramrez , j. m. , bautista , m. , & kallman , t. 2005 , , 627 , 166 reynolds , c. s. 1997 , , 286 , 513 raska , a. , goosmann , r. , dumont , a .-
, & czerny , b. 2006 , , 452 , 1 steenbrugge , k. c. ; kaastra , j. s. ; de vries , c. p. ; edelson , r. 2003 , , 402 , 477s steenbrugge , k. c. , et al .
2005 , , 434 , 569 treves , a. , belloni , t. , chiappetti , l. , maraschi , l. , stella , l. , tanzi , e. g. , & van der klis , m. 1988 , , 325 , 119 turner , t. j. ; george , i. m. ; nandra , k. ; turcan , d. 1999 , , 524 , 667 t uttley , philip ; edelson , rick ; mchardy , ian m. ; peterson , bradley m. ; markowitz , alex 2003 , , 584l , 53u .
weisskopf , m. c. , tananbaum , h. d. , van speybroeck , l. p. , & odell , s. l. 2000 , , 4012 , 2 yaqoob , t. ; george , i. m. ; nandra , k. ; turner , t. j. ; serlemitsos , p. j. ; mushotzky , r. f. 2001 , , 546 , 759
in this appendix we have enumerated ( from high to low energy ) the 31 references for the 185 photometric data points from ned used to build the spectral energy distribution .
1 . kaspi , shai ; maoz , dan ; netzer , hagai ; peterson , bradley m. ; vestergaard , marianne ; jannuzi , buell t. 2005 , , 629 , 61k 2 .
mckernan , b. ; yaqoob , t. ; reynolds , c. s. 2007 , , 379 , 1359 m 3 .
jack w. sulentic , rumen bachev , paola marziani , c. alenka negrete , and deborah dultzin 2007 , , 666 , 757s 4 .
muoz marn , vctor m. ; gonzlez delgado , rosa m. ; schmitt , henrique r. ; cid fernandes , roberto ; prez , enrique ; storchi - bergmann , thaisa ; heckman , tim ; leitherer , claus 2007 , , 134 , 648 m 5 .
anderson , kurt s. 1970 , , 162 , 743a 6 .
mcalary , c. w. ; mclaren , r. a. ; mcgonegal , r. j. ; maza , j. 1983 , , 52 , 341 m 7 .
de vaucouleurs , g. ; de vaucouleurs , a. ; corwin , h. g. , jr . ; buta , r. j. ; paturel , g. ; fouque , p. 1991 , rc3.9 , c. 8 .
koulouridis , e. ; plionis , m. ; chavushyan , v. ; dultzin - hacyan , d. ; krongold , y. ; goudis , c. 2006 , , 639 , 37k 9 .
suganuma , masahiro ; yoshii , yuzuru ; kobayashi , yukiyasu ; minezaki , takeo ; enya , keigo ; tomita , hiroyuki ; aoki , tsutomu ; koshida , shintaro ; peterson , bruce a. 2006 , , 639 , 46s 10 .
zwicky , fritz ; herzog , e. 1963 , cgcg2.c ... 0000z 11 .
takamiya , m. ; kron , r. g. ; kron , g. e. 1995 , , 110 , 1083 t 12 .
petrosian , artashes ; mclean , brian ; allen , ronald j. ; mackenty , john w. 2007 , , 170 , 33p 13 .
bentz , misty c. ; denney , kelly d. ; cackett , edward m. ; dietrich , matthias ; fogel , jeffrey k. j. ; ghosh , himel ; horne , keith d. ; kuehn , charles ; minezaki , takeo ; onken , christopher a. ; peterson , bradley m. ; pogge , richard w. ; pronik , vladimir i. ; richstone , douglas o. ; sergeev , sergey g. ; vestergaard , marianne ; walker , matthew g. ; yoshii , yuzuru 2007 , , 662 , 205b 14 .
bentz , misty c. ; peterson , bradley m. ; pogge , richard w. ; vestergaard , marianne ; onken , christopher a. 2006 , , 644 , 133b 15 .
lebofsky , m. j. ; rieke , g. h. 1980 , natur , 284 , 410l 16 .
wisniewski , w. z. ; kleinmann , d. e. 1968 , , 73 , 866w 17 .
spinoglio , luigi ; malkan , matthew a. ; rush , brian ; carrasco , luis ; recillas - cruz , elsa 1995 , , 453 , 616s 18 .
balzano , v. a. ; weedman , d. w. 1981 , 243 , 756b 19 .
mcalary , c. w. ; mclaren , r. a. ; crabtree , d. r. 1979 , , 234 , 471 m 20 .
rieke , g. h. 1978 , , 226 , 550r 21 .
pacholczyk , a. g. ; weymann , r. j. 1968 , , 73 , 870p 22 .
penston , m. v. ; penston , m. j. ; selmes , r. a. ; becklin , e. e. ; neugebauer , g. 1974 , mnras , 169 , 357p 23 .
rodrguez - ardila , a. ; riffel , r. ; pastoriza , m. g. 2005 , , 364 , 1041r 24 . de vaucouleurs , antoinette ; longo , giuseppe 1988 , virph .
stein , w. a. ; weedman , d. w. 1976 , , 205 , 44s 26 .
maiolino , r. ; shemmer , o. ; imanishi , m. ; netzer , h. ; oliva , e. ; lutz , d. ; sturm , e 2007 , , 468 , 979 m 27 .
kleinmann , d. e. ; low , f. j. 1970 , , 161l , 203k 28 .
rieke , g. h. ; low , f. j. 1972 , , 176l , 95r 29 .
gorjian , v. ; werner , m. w. ; jarrett , t. h. ; cole , d. m. ; ressler , m. e. 2004 , , 605 , 156 g 30 .
moshir , m. ; et al . 1990 , irasf , c , 0 m 31 .
gorjian , v. ; cleary , k. ; werner , m. w. ; lawrence , c. r. 2007 , , 655l , 73 g observation & ut start date & exposure@xmath115 & flux@xmath116 & s / n(22 ) + ( 2)837 & 2000 - 02 - 05 15:37 & 82.3 & 1.14&6.63 + ( 3)3046 & 2002
- 01 - 16 06:12 & 153.9 & 1.96&4.69 + total&&236.2 & & 8.12 + ( 1)330 & 1999 - 12 - 11 22:51 & 85.98 & 2.99&34.27 + ( 4)3045 & 2002 - 01 - 18 15:57 & 169.68 & 1.38&35.71 + ( 5)3383 & 2002 - 01 - 21 07:33 & 171.02 & 2.19&33.32 + ( 6)5598 & 2005 - 04 - 15 05:18 & 116.43 & 0.56&14.45 + ( 7)6268 & 2005 - 04 - 18 00:31 & 25.19 & 0.46&9.38 + total&&568.3&&62.09 + llllllccccc a & 1.52@xmath117 & 4.87@xmath118 & 0.012@xmath79 & 1.00@xmath119 & 4 & & & - & - - & 3813 , 4112 , 0.93 + b & 1.560@xmath120 & 5.20@xmath121 & 0.085@xmath122 & 6.98@xmath123 & 4 & &@xmath124 & - - & - & 3053 , 4108 , 0.74 + c & [email protected] & 5.00@xmath125 & [email protected] & 7.28@xmath126 & 4 & @xmath124 & @xmath124 & - - & - & 2987 , 4104 , 0.73 + d & 1.595@xmath127 & 5.00@xmath128 & [email protected] & 5.65@xmath129 & 4 & @xmath124 & @xmath124 & @xmath124 & - & 2946 , 4100 , 0.72 + e & 1.597@xmath130 & 5.40@xmath128 & 0.110@xmath131 & 5.17@xmath132 & 4 & @xmath124 & @xmath124 & @xmath124&@xmath124 & 2904 , 4096 , 0.71 + f & 1.598@xmath133 & 5.46@xmath128 & 0.116@xmath134 & 4.51@xmath135 & 8 & @xmath124 & @xmath124 & @xmath124&@xmath124 & 2885 , 4102 , 0.70 + g & 1.778@xmath136 & 7.84@xmath128 & 0.083@xmath137 & 5.93@xmath138 & 7 & @xmath124 & @xmath124 & @xmath124&@xmath124 & 2013 , 1618 , 1.24 + ccccc fe k@xmath83 & 1.937 & 1.9368@xmath139 & 2644.4@xmath140 & 27.9 @xmath141 + xiii sixiii&6.648 & 6.739@xmath142 & 754@xmath143 & 3.7 @xmath144 + ix neix & 13.699 & 13.678 @xmath145 & 485@xmath146 & 22.6@xmath147 + vii ovii&21.607 & 21.597 @xmath148 & 936@xmath149 & 35.9@xmath150 + vii ovii & 21.807 & 21.785 @xmath151 & 928@xmath152 & 73.2@xmath153 + vii ovii & 22.101 & 22.075 @xmath154 & 915@xmath155 & 244.2@xmath156 + vi nvi & 23.024 & 23.035 @xmath157 & 327@xmath158 & 26.2@xmath159 + vi nvi&23.277 & 23.306@xmath160 & 673@xmath161 & 80.9@xmath162 + ix neix & 13.699 & 13.677 @xmath163 & 60.7@xmath164 & 5.9@xmath165 + vii ovii&21.607 & 21.623 @xmath166 & 944@xmath167 & 34.9@xmath168 + vii ovii & 21.807 & 21.760 @xmath169 & 935@xmath170 & 42.8@xmath171 + vii ovii & 22.101 & 22.075 @xmath172 & 923@xmath173 & 119.0@xmath174 + vi nvi&29.534 & 29.502@xmath175 & 462@xmath176 & 23.9@xmath177 + vi nvi&29.081 & 29.029&10 & 11.5@xmath178 + v cv & 41.472 & 41.429 &
20@xmath179 & 121.1@xmath180 + ccccc b1 & [email protected] & [email protected] & -740@xmath79150 & 536@xmath7935 + c1 & [email protected] & [email protected]&-1120@xmath79150 & 117@xmath7926 + c2&[email protected] & 21.28@xmath79 0.04&-450@xmath79150&241@xmath7937 + d1&[email protected]&[email protected]&-1120@xmath79150&113@xmath7942 + d2&[email protected]&[email protected]&-450@xmath79150 & 275@xmath79117 + d3&[email protected]&[email protected]&-590@xmath79150&105@xmath79 47 + log u@xmath115 & 1.23@xmath190 & 0.67@xmath191 & 0.67@xmath191 & -0.49@xmath192 + log n@xmath82 ( @xmath35)@xmath115 & 21.73@xmath193 & 21.03@xmath194 & 21.26@xmath195 & 20.75@xmath196 + v@xmath197 ( km s@xmath4 ) & 175 @xmath198 & 100 @xmath199 & 177 @xmath200 & 105 @xmath201 + v@xmath202 ( km s@xmath4)@xmath115 & -1040@xmath60 & -1180@xmath60 & -400@xmath60 & -590@xmath60 + t ( k)@xmath203 & @xmath204 & @xmath205 & @xmath206 & @xmath207 + @xmath208log t ( k)@xmath209 $ ] & 6.47@xmath193 & 5.89@xmath190 & 5.72@xmath193 & 4.54@xmath136 + log
t / u ( @xmath210p@xmath211 ) & 5.21@xmath212 & 5.15@xmath213 & 5.04@xmath214 & 5.06@xmath215 + lccccc iv civ & 15.20 & 10.13 & 15.20 & 13.99 & + v cv & * 16.75 * & 13.76 & 16.75 & 16.90 & 17.20 + vi cvi & * 17.03 * & 16.03 & 17.07 & 18.10 & 17.60 + v nv & * 15.53 * & 11.07 & 15.53 & 14.93 & - + vi nvi & * 16.46 * & 14.06 & 16.46 & 17.20 & 17.20 + vii nvii & * 16.26 * & 15.86 & 16.41 & 18.0 & 17.50 + v ov & * 16.36 * & 11.11 & 16.36 & 15.70 & 16.80 + vi ovi & * 16.89 * & 13.15 & 16.89 & 16.80 & 16.60 + vii ovii & * 17.40 * & * 15.68 * & 17.41 & 18.50 & 18.18 + viii oviii & * 16.78 * & * 17.06 * & 17.24 & 18.70 & 18.53 + ix neix & * 15.92 * & * 16.10 * & 16.32 & 18.00 & 17.20 + x nex & 14.70 & * 16.88 * & 16.88 & 17.60 & 17.90 + xi mgxi & 15.04 & * 16.20 * & 16.23 & 17.30 & 17.00 + xii mgxii & 13.31 & * 16.51 * & 16.51 & 16.50 & 17.30 + viii siviii & 15.46 & 12.06 & 15.46 & 16.70 & 16.20 + ix siix & 15.83 & 13.40 & 15.82 & 17.20 & 16.00 + xi sixi & 15.44 & 14.34 & 15.47 & 16.80 & 16.00 + xiii sixiii & 13.81 & * 16.45 * & 16.45 & 16.40 & 17.00 + xiv sixiv & 11.74 & * 16.39 * & 16.39 & 15.20 & 17.00 + xi sxi & 15.35 & 14.99 & 15.51 & 16.90 & 16.00 + xii sxii & 14.84 & 15.63 & 15.70 & 16.70 & 16.20 + xvii fexvii & 13.82 & 15.73 & 15.74 & 15.10 & 16.30 + xviii fexviii & 13.54 & * 16.19 * & 16.19 & & - + ccccc iv civ & 10.36 & 9.04 & 10.38 & + v cv & 12.98 & 12.74 & 13.18 & 17.20 + vi cvi & * 16.26 * & 15.63 & 16.35 & 17.60 + v nv & 11.30 & 9.57 & 11.31 & + vi nvi & 14.30 & 12.96 & 14.32 & 17.20 + vii nvii & 16.09 & 15.50 & 16.19 & 17.50 + v ov & 10.86 & 9.30 & 10.87 & 16.80 + vi ovi & 13.37 & 11.30 & 13.37 & 16.60 + vii ovii & 15.90 & 14.51 & 15.92 & 18.18 + viii oviii & * 17.29 * & * 16.76 * & 17.40 & 18.53 + ix neix & * 15.86 * & 14.68 & 15.89 & 17.20 + x nex & * 16.65 * & * 16.55 * & 16.90 & 17.90 + xi mgxi & * 15.96 * & 15.10 & 16.02 & 17.00 + xii mgxii & * 16.28 * & * 16.44 * & 16.67 & 17.30 + viii siviii & 11.83 & 7.98 & 11.83 & 16.20
+ ix siix & 13.16 & 7.98 & 13.16 & 16.00 + xi sixi & 14.10 & 11.85 & 14.10 & 16.00 + xiii sixiii & * 16.22 * & 15.75 & 16.35 & 17.00 + xiv sixiv & * 16.16 * & * 16.71 * & 16.82 & 17.00 + xi sxi & 14.17 & 9.71 & 14.17 & 16.00 + xii sxii & 14.75 & 12.37 & 14.75 & 16.20 + xvii fexvii & 15.49 & 13.24 & 15.49 & 16.30 + xviii fexviii & * 15.96 * & 14.12 & 15.97 & - + | we present an analysis of x - ray high quality grating spectra of the seyfert 1 galaxy ngc 5548 using archival _ chandra_-hetgs and letgs observations for a total exposure time of 800 ks . the continuum emission ( between 0.2 - 8 kev ) is well represented by a power law ( @xmath0 = @xmath1 ) plus a black - body component ( kt = @xmath2 kev ) .
we find that the well known x - ray warm absorber in this source consists of two different outflow velocity systems .
one absorbing system has a velocity of -@xmath3 km s@xmath4 and the other of -@xmath5 km s@xmath4 . recognizing the presence of these kinematically distinct components allows each system to be fitted independently , each with two absorption components with different ionization levels .
the high velocity system consists of a component with temperature of @xmath6 k , @xmath7 and another component with temperature of @xmath8 k , @xmath9 . the high - velocity , high - ionization component produces absorption by charge states xxi - xxiv fexxi - xxiv , while the high - velocity , low - ionization component produces absorption by ix - x neix - x , xvii - xx fexvii - xx , vii - viii ovii - viii .
the low - velocity system required also two absorbing components , one with temperature of @xmath10 k , @xmath11 , producing absorption by ix - x neix - x , xvii - xx fexvii - xx , vii - viii ovii - viii . the other with lower temperature ( @xmath12 k ) , and lower ionization ( @xmath13 ) ; producing absorption by vi - vii ovi - vii and the vii - xii fevii - xii m - shell uta .
once these components are considered , the data do not require any further absorbers .
in particular , a model consisting of a continuous radial range of ionization structures ( as suggested by a previous analysis ) is not required .
the two absorbing components in each velocity system are in pressure equilibrium with each other .
this suggests that each velocity system consists of a multi - phase medium .
this is the first time that different outflow velocity systems have been modelled independently in the x - ray band for this source .
the kinematic components and column densities found from the x - rays are in agreement with the main kinematic components found in the uv absorber .
this supports the idea that the uv and x - ray absorbing gas is part of the same phenomenon .
ngc 5548 can now be seen to fit in a pattern established for other warm absorbers : 2 or 3 discrete phases in pressure equilibrium .
there are no remaining cases of a well studied warm absorber in which a model consisting of a multi - phase medium is not viable . |
recent cosmological observations favoring the scenario of spatially flat and accelerated universe .
observations come from hubble diagram of sne ia supernovae are best fitted by spatially flat accelerated cosmological models @xcite-@xcite . besides , other observations from cosmic microwave background radiation ( cmbr ) and baryon acoustic oscillation ( bao ) are also strong evidences for this scenario @xcite-@xcite .
this currently observed speeding up expansion of the universe leads to the fact that there is a repulsive force acting in the universe space - time . according to our current understanding of the universe history ,
the radiation era was followed by a matter dominated era during which most of the universe structures like stars , galaxies and galaxy clusters we now observe was formed by gravitational instabilities . through this phase of the universe evolution , a particular form of gravitating but non - baryonic form of matter
is formed .
this sort of matter which is non - relativistic does not interact with radiation , and so is called dark matter ( dm ) .
late in this matter dominated phase , a new but very peculiar form of invisible and non - gravitating `` matter '' started to dominate .
this sort of `` matter '' does not interact with baryons , radiations , or any other sort of visible matter .
it is now known as dark energy ( de ) .
observational evidences show that , see for ex . @xcite and @xcite-@xcite , our universe at present contains approximately @xmath3 only of radiation and baryonic matter that is well known by the standard model of particle physics and can be detected in laboratory and so is considered visible . while about @xmath4 is the non - baryonic dark matter . the rest of our universe content , which is about @xmath5 is the exotic component known as dark energy . as a result , this dominated dark energy component antigravitates leading to the present observed universe expansion .
although dark energy is an acceptable and effective description for such large scale homogeneous and isotropic accelerated universe , identifying its origin and nature is one of the most debated questions in physics and cosmology .
large variety of models are proposed to describe the nature of such dark energy @xcite,@xcite . as the energy density due to de remains unchanged throughout the universe space - time ,
the simplest choice was to revive the idea of einstein about the cosmological constant @xmath6 @xcite,@xcite .
this together with cold dark matter ( cdm ) provides the standard model for cosmology @xmath6cdm @xcite-@xcite , which appears to be in a very good agreement with observational data . in einstein s field equation
, this term has the concept of intrinsic energy density of vacuum @xcite .
however , this arises an important question , why the observed value of vacuum energy density is very far below that is predicted from particle physics ? .
this is the famous fine tuning problem .
another important question is why the energy densities of the vacuum and the matter are of the same order of magnitude today ? , the famous coincidence problem .
dynamical models are other alternative to the cosmological constant .
some of these are the minimal coupled scalar fields or quintessence @xcite-@xcite , k - essence @xcite-@xcite , chaplygin gas @xcite-@xcite , all have an equation of state parameter @xmath7 .
these models adopt the idea of the possibility of unification of the two dark sectors , dark energy and dark matter .
models with @xmath8 knowing as phantom energy models @xcite , @xcite consider a growing vacuum energy with expansion .
a more general class of models considers a parameter varies with the scale factor , or the red shift @xcite , @xcite , @xcite , @xcite , @xcite .
there are also some observational analysis which consider the variation of @xmath9 with the red shift and with time @xcite , @xcite .
modification of the governing equation , or the so called modified gravity theory @xcite , @xcite , is also one way of thinking .
the idea of unification of de and dm is a promising one and chaplygin gas model is one of the most acclaimed models representing it .
this model was proposed by kamenshchik et .
al . @xcite as early as 2001 where they assumed that the universe , within the framework of standard cosmology , is filled with an exotic fluid obeys the chaplygin gas equation of state ( eos ) @xmath10 where @xmath11 is a positive constant and @xmath12 .
this model has the advantage of smooth transition between different phases of the universe .
many authors then extended this model using the idea of generalizing or modifying the eos of the background fluid in order to improve the behavior of the fluid throughout the universe evolution . in 2002 , bento et .
@xcite extended the model to a more general form known as the generalized chaplygin gas ( gcg ) model where the @xmath13 constant is considered in the range @xmath14 .
other various models based on chaplygin gas were also proposed such as the modified chaplygin gas model introduced in 2002 by benaoum @xcite where he could interpolate between standard fluids at high energy densities and chaplygin gas fluids at low energy densities using the eos @xmath15 where @xmath16 and @xmath11 and @xmath17 are positive constants , and the hybrid chaplygin gas model introduced by bili@xmath2 et .
al . in 2005
@xcite and leads to transient acceleration . in this work ,
we have adopted a class of dynamical models in which the eos has the advantage of interpolation between two different chaplygin gas models .
our fluid is a barotropic one with an eos can be considered as a correction to the vacuum eos by a power law .
this power law has the advantage of interpolation between the two equations of state of the de and dm which might describe some sort of smooth phase transition .
another advantage of this model is that it is not only interpolating between dust and de in early and late times , but also it has a more general eos for de that enables the cosmological constant as a special case .
this has the more advantage of a general asymptotic solution than chaplygin gas which goes to the cosmological constant .
we present an exact solution of the equations which enables more accurate results for the calculations of the luminosity distance , while observations of hubble diagram of sne ia supernovae are used to further constrain the model parameters .
we study the energy density and the equation of state parameter evolutions .
the analysis of the eos shows that the dynamical properties of both dark sectors can be described through this model which lets our choice seems natural because of the smooth transition .
we also calculate the deceleration parameter to examine the expansion of the universe due to the model .
the rest of the article is organized as follows ; in the following section we have formulated the model . in section [ sectr ] the mathematical treatment of the model is made .
section [ case1 ] treats the first case of study of our model where the causality constraints are examined and the dynamics of the universe is studied where the equation of the luminosity distance as a function of red shift is solved and parameters are constrained through hubble diagram observations , the energy density and the deceleration parameter evolutions due to the model are examined .
section [ case2 ] treats the second case of study of our model .
section [ seconc ] concludes our work .
we are working through the standard frw cosmology where the metric in the spatially flat geometry is @xmath18 where we consider units with @xmath19 . using the energy momentum tensor given by @xmath20
where in comoving coordinates @xmath21 , the einstein s equation will lead to the friedman equations @xmath22 while the conservation equation @xmath23 gives @xmath24 now we have to consider an eos to close the system .
it is a fact that dark energy dominates the universe today and an adequate eos for this component is @xmath25 .
observations indicate that the eos parameter @xmath9 is close to @xmath26 and people are now targeting to confirm whether @xmath9 is constant or varies as the universe expands . owing to this fact , nojiri and odintsov @xcite considered a general eos of the form @xmath27 where @xmath28 is an arbitrary function may be considered as a correction to the standard de equation of state .
@xmath0tefan@xmath1i@xmath2 @xcite , @xcite studied in details the case of @xmath29 with a constant @xmath30 , and nojiri et .
@xcite studied the future singularity associated with this model . in this work we adopted a barotropic fluid with an eos which is also a correction to the standard dark energy eos .
the fluid pressure has a general power law form for the density dependence @xmath31 where @xmath32 , @xmath33 , @xmath34 , and @xmath35 are constants which are considered as free parameters .
this enables interpolation between different powers for the density , so that the phase transitions during the universe evolution are described smoothly .
the eos parameter is given by @xmath36
let s now study the evolution of the physical quantities of the universe due to our model .
considering units with @xmath37 , eqns ( [ dot])-([cons ] ) reduce to @xmath38 where @xmath39 is the hubble parameter . combining eq([van1 ] ) to eq([rodot ] ) we get @xmath40 where again @xmath13 , @xmath41 , @xmath42 , and @xmath43 are constants related to those of ( [ van1 ] ) through the relations @xmath44 the solution of equation([dhdt ] ) indicates the evolution of the hubble parameter with time .
this equation is integrated to give @xmath45 however , our model is not describing a pure dark energy , in which case the values of the parameters are totally free and are not constrained .
instead , our model able to follow the smooth phase transition which takes place through the universe evolution .
this restricts the parameter values to obey the limits satisfied by observations so that for early time we must have the equation of state of the perfect fluid . accordingly
, we first have to study the asymptotic behavior of @xmath46 . from one side , this indicates to what extent our model agrees the real evolution of the universe , and from the other side may also fixes the limits of some of our model parameters .
table [ tab : ass ] shows this behavior for the different ranges of of the parameters @xmath42 and @xmath43 .
@cccc@ case & ranges of @xmath42 and @xmath43 & large @xmath47 & small @xmath47 + + @xmath48 & @xmath49 @xmath50 @xmath51 & @xmath52 & @xmath53 + @xmath54 & @xmath49 @xmath50 @xmath55 & @xmath56 & @xmath52 + @xmath57 & @xmath58 @xmath50 @xmath51 & @xmath52 & @xmath53 + @xmath59 & @xmath58 @xmath50 @xmath55 & @xmath56 & @xmath52 + [ tab : ass ] one can see from the table that the values of negative @xmath42 must be precluded in our calculations as a result of the asymptotic behavior of the @xmath46 function .
this is because for asymptotically perfect fluid behavior of our universe , the @xmath46 function at large @xmath47 forces @xmath42 to be positive if @xmath43 is negative , and forces @xmath60 to be positive if @xmath43 is positive , which does not satisfy constraints for the pre - assumptions . on the other hand , the asymptotic behavior of @xmath46 for the two cases of positive @xmath42 can attain the perfect fluid behavior at large @xmath47 if @xmath61 and @xmath62 for the case of @xmath63 and if @xmath64 and @xmath65 for the case of @xmath66 .
these two cases will be studied in details .
finally let s study the different classes of solutions due the fixed points in the model .
this general class of models can be studied using the phase space method developed in @xcite .
the phase diagram in fig.[fig : phase ] shows that there are no fixed points in the solution so that there are no different classes , here @xmath67 and @xmath68 .
[ fig : phase ]
this is the first case that match the dark fluid dynamics .
the asymptotic behavior of the eos at large @xmath47 , early times , has the form @xmath69 , so that for @xmath61 and @xmath62 we have the perfect fluid eos at early times .
let s now examine the constraints that can be obtained by applying the causality condition to this model .
spherical sound waves are lunched as a result of the overpressure due to the initial overdensity in the dm and gas .
these oscillations have a speed which is an indication to the velocity by which such perturbations are transmitted .
accordingly , it must not exceed the speed of light , which is a condition of causality .
in barotropic cosmic fluid , the speed of these sound waves is defined by the relation @xmath70 this must satisfy the condition @xmath71 that must be obeyed at all values of @xmath72 .
accordingly , applying causality constraints , one may get an indication for the available ranges for the values of the free parameters in the eos . now using ( [ van1 ] )
, we get @xmath73\gamma \rho^{n-1}}{\left(1+\delta \rho^m\right)^2 } \leq 1 \label{cas1}\ ] ] now for our case where @xmath34 and @xmath35 are @xmath74 , relation ( [ cas1 ] ) for large @xmath72 gives the condition @xmath75 this is fully satisfied for @xmath76 while for small @xmath72 , the condition will be @xmath77 this is satisfied if @xmath78 of course , this last condition is automatically satisfied if the the condition in ( [ cas7 ] ) is satisfied . now combining our conditions , ( [ cas7 ] ) and ( [ cas9 ] ) , we get to the conclusion that for this case we must consider @xmath79 which means @xmath80 in agreement with the conditions required by asymptotic behavior of the eos at early time for this case .
in fact , for the general values of @xmath72 we also have a large dependence on @xmath32 and @xmath33 .
accordingly , we have to make a check using the obtained optimized values for the parameters . in this section
we study the ability of our model to produce the physical quantities that can be really measured .
this , of course , depends on the compatibility of the results of our model with the astrophysical data that are indeed observed .
one of the most important observed data is that of the hubble diagram of type sne ia supernovae , which is a plot of the luminosity distance @xmath81 , that determines the flux of the source , as a function of the red shift .
we will use these data to constrain the parameters in our model . due to the asymptotic and causality constraints
, we have now only two out of four parameters that are free and needed to be optimized due to observations . to satisfy the causality and asymptotic behavior constraints , we choose @xmath60 as @xmath82 while @xmath83 has to be @xmath84 , so that we can attain the perfect fluid behavior of @xmath46 at large @xmath47 .
let s now start with our model by using the relation ( [ dhdt ] ) for @xmath85 in ( [ rodot ] ) , so that we have @xmath86 which in turn gives @xmath87 this has the solution @xmath88 \label{hofap}\ ] ] the function @xmath89 is the lambert w - function . to fix the constant c we consider that at some time where @xmath90 we have @xmath91 and @xmath92 , this gives @xmath93 so that @xmath94 \label{hofap1}\ ] ] accordingly @xmath95 \label{hofzp}\ ] ] the luminosity distance @xmath81 is then given by @xmath96 where @xmath97 is the proper distance given by @xmath98 accordingly @xmath99 d\zeta \label{dpofzp1}\ ] ] the integration of the above equation can not be solved analytically .
accordingly , we will use numerical methods for solving the integral .
this enables us to get an exact solution for the relation controlling the dependence of the luminosity distance on the red shift . in fig.[fig : lmnstyp ] we plot the apparent magnitude , @xmath100 , of the data set from conley et al .
( 2011)@xcite versus their red shifts @xmath101 together with the fitted results of our model .
data set are available at supernova cosmology project@xcite . using @xmath102 due to plank@xmath103 wmap@xmath103 bao measurements@xcite , our @xmath104 minimization fit results in the parameters @xmath105 , and @xmath106 .
these parameters are due a reduced @xmath104 value of @xmath107 .
having the full set of parameters , let s test the causality condition using these parameters .
a plot of causality relation ( [ cas1 ] ) using our parameters for this case is shown in fig([fig : pstvs ] ) .
we can see from the figure that the parameters satisfy the causality condition for the model throughout the whole stages of the evolution of the universe .
let us see how the energy will evolve with time due to the model in a homogenous and isotropic universe .
relations ( [ cons ] ) and ( [ van1 ] ) give @xmath108 to simplify our way of solution to this equation let s use our parameters . in this case
we have @xmath109 , so that the solution of relation ( [ roh1 ] ) will take the form @xmath110 where @xmath111 to study the properties of relation ( [ roh8 ] ) at early times we make use of the property @xmath112 then @xmath113 now , in the past , where @xmath114 , so that @xmath115 due to ( [ roh9 ] ) as @xmath35 is @xmath74 , taylor expansion of @xmath116-function can be approximated to @xmath117 , so that the exponential term of the @xmath116-function converges to @xmath118 , then the energy density for that time will take the asymptotic form @xmath119 as we have @xmath120 for this case , where @xmath121 is a constant .
accordingly , the fluid of the model behaves as a matter in the earlier times . on the other hand , for @xmath122 , @xmath123 , the energy density function due to relation ( [ roh333 ] ) will tend to @xmath124 accordingly @xmath125 using the value of @xmath35 we can see that this function is a very slowly decreasing function of @xmath126 for @xmath127 , and tends to be nearly constant at @xmath123 , i.e. at @xmath122 , a signature for a cosmological constant towards the future . the equation of state parameter @xmath9 for this case , where @xmath128 , will be given , from ( [ par1 ] ) , by @xmath129 using our parameters for this case where @xmath130 , we finally get @xmath131 indicating that for earlier times , where we have large values of @xmath72 , @xmath132 , a signature for matter era , while changes to @xmath26 gradually as the matter content of the universe diluted .
the acceleration of the universe is related to a dimensionless parameter known as the deceleration parameter @xmath133 defined as @xmath134 since you have to have @xmath135 to get an accelerating universe , the deceleration parameter will then measure if the expansion of the universe is accelerating or decelerating .
if @xmath136 , the expansion is accelerating , while if @xmath137 it is decelerating .
now as @xmath138 then @xmath139 now using ( [ dhdt ] ) , one gets the deceleration parameter for our model to be @xmath140 using @xmath141 , relation ( [ dec4 ] ) gives @xmath142 using the values of our parameters with @xmath143 , the present day deceleration parameter due to this model case has the value @xmath144 relation ( [ dec6 ] ) also shows that deceleration parameter evolves towards more negative values as the hubble parameter decreases with time .
for this case @xmath145 asymptotically at large @xmath47 , which means that we can attain the perfect fluid eos at early time if @xmath64 and @xmath65 . however , a first step again is the examination of the causality constraints for this case of the model . in this case
@xmath35 is @xmath146 , so that relation ( [ cas1 ] ) at large values of @xmath72 tends to @xmath147 which is satisfied if @xmath148 while for small @xmath72 we have @xmath149 as @xmath35 is @xmath146 , this is always satisfied under the condition @xmath150 combining ( [ cas3 ] ) and ( [ cas4d ] ) we get to the conclusion that in this case we must consider @xmath151 which means @xmath152 in full agreement with the asymptotic behavior of the eos at early time for this case . we ll now study the physical quantities representing the dynamics of our universe due to this case of the model and comparing our results with observations .
we could now constrain two parameters out of four through asymptotic behavior of the model and the causality condition , where @xmath42 has to be @xmath82 while @xmath13 has to be @xmath84 .
let s then calculate the luminosity distance function @xmath153 and use the observed data of this function to constrain the remaining two parameters of the model .
we start by using relation ( [ dhdt ] ) for @xmath64 in ( [ rodot ] ) , so that we have @xmath154 which in turn gives @xmath155 this has the solution @xmath156 \label{hofa}\ ] ] and let s assume that at some time @xmath90 we have @xmath91 and @xmath92 .
this gives @xmath157 so that @xmath158 \label{hofaa}\ ] ] in turn , we can write @xmath159 as @xmath160 \label{hofz}\ ] ] the luminosity distance @xmath81 , eqs ( [ dl ] ) , is then calculated using the proper distance which due to ( [ dp ] ) will be given by @xmath161 d\zeta \label{dpofz}\ ] ] again , the integration of relation ( [ dpofz ] ) can not be solved analytically so that we use numerical methods .
the apparent magnitude of conley et al .
( 2011)@xcite data set is plotted in fig.[fig : lmnsty ] together with our results .
our @xmath104 minimization fit results in the parameters @xmath162 and @xmath163 .
these parameters are due a @xmath104 value of @xmath164 .
a last step for this case now is to test the causality condition using these parameters . a plot of causality relation ( [ cas1 ] ) using our parameters
is shown in fig([fig : ngtvs ] ) .
the figure shows that the causality is satisfied for all values of @xmath72 , i.e. , it is ensured by the model in this case through the universe evolution .
let s solve relations ( [ roh1 ] ) for @xmath165 , the value of the parameter @xmath34 in this case .
this gives @xmath166 where @xmath167 studying the properties of relation ( [ roh2 ] ) we see that due to relation ( [ roh322 ] ) we have @xmath168 so that in the past where @xmath169 , so that @xmath115 due to ( [ roh3 ] ) since @xmath35 is @xmath146 , we can consider the @xmath116-function as approximated to @xmath117 and its exponential term converges to 1 , the energy density for that time will then has the asymptotic form @xmath170 where we have @xmath171 for this case and @xmath172 .
this shows that , the fluid of our model behaves as a matter in the earlier times . on the other hand , to study the fluid for @xmath122 , @xmath123
, we make use of relation ( [ roh3 ] ) together with the property of relation ( [ roh33 ] ) in the energy density function , relation ( [ roh2 ] ) , so that the energy density function takes the form @xmath173 accordingly @xmath174 using the value of the parameter @xmath35 we can see that this function is again a very slowly decreasing function of @xmath126 for @xmath127 and tends to be nearly constant at @xmath123 , i.e. , at @xmath122 , which is again a signature for the cosmological constant towards the future . finally , the equation of state parameter @xmath9 for this case will be given from ( [ par1 ] ) by @xmath175 indicating that for large values of @xmath72 , where we are deep in the matter era , @xmath176 as @xmath35 is @xmath146 , while changes to @xmath26 gradually as the matter content of the universe diluted . using @xmath64 and @xmath177 in relation ( [ dec4 ] )
we get @xmath178 and using the values of @xmath43 and @xmath41 together with @xmath143 , this relation gives a present day deceleration parameter of @xmath179 relation ( [ dec5 ] ) also shows that as @xmath43 is negative , the deceleration parameter evolves towards more negative values as the hubble parameter decreases to the future .
a barotropic fluid model is considered which initially can be described as a correction to the dark energy regime by a power law .
the power law enables interpolation between the dm and de equations of state , and so guarantees a smooth phase transition from matter era to de era .
accordingly , our model is a dark fluid model provides a unification for the two dark sectors in one eos that has the advantage of the general asymptotic solution .
this means that it can describe dark matter at early times while describing dark energy at late times , while adopts a general eos for de which enables the cosmological constant as a special case .
studying the asymptotic behavior of the equation of state constrains the parameters of the equation and restricts the model .
it also clarifies the cases that can be physically acceptable as dark fluid which are found to be two cases . on the other hand , studying the phase diagram of the model for each of these two cases shows that there are no fixed points for both , with the result that there is only one class of solution for each of them .
causality condition further constrains the parameters of the model , so that at the end , only two out of four parameters are free .
these two parameters are constrained using the hubble diagram of sne ia supernovae , where we present an exact solution for the red shift dependence of the luminosity distance in each model , or model case , avoiding the approximate formulas that are accurate only at small red shifts .
the @xmath180 fitting of the two models shows that they both fit the data well . for the model with @xmath181 we got a @xmath182 , and for the model with @xmath183 we got a @xmath184 .
however , it is clear that observational data prefer the @xmath181 model than the @xmath183 one .
more analysis such as density perturbation analysis may differentiate them and may favor one of the models than the other .
we also studied the energy density evolution which ensures the smooth transition between the two dark sectors .
furthermore , we examined the evolution of the eos parameter @xmath9 that again ensures its two extremes of @xmath185 and @xmath26 for early and late times .
we also calculated the deceleration parameter @xmath133 and examined its evolution .
calculations resulted in a negative @xmath133 for both model cases with the present day values of @xmath186 and @xmath187 due to the two models , while assures evolution towards more negativity to the future as the hubble parameter decreases , a clear evidence for accelerated expansion in agreement with observations . due to the above mentioned results ,
our model can be considered as a dark fluid model unifying the two dark sectors of the universe and succeeded in describing the evolution of the universe .
the author would like to thank dr .
adel awad for valuable discussions and guidance . | we consider a cosmological model based on a generalization of the equation of state proposed by nojiri and odintsov @xcite and @xmath0tefan@xmath1i@xmath2 @xcite , @xcite .
we argue that this model works as a dark fluid model which can interpolate between dust equation of state and the dark energy equation of state .
we show how the asymptotic behavior of the equation of state constrained the parameters of the model .
the causality condition for the model is also studied to constrain the parameters and the fixed points are tested to determine different solution classes .
observations of hubble diagram of sne ia supernovae are used to further constrain the model .
we present an exact solution of the model and calculate the luminosity distance and the energy density evolution .
we also calculate the deceleration parameter to test the state of the universe expansion . |
x - ray emission in early - type galaxies partly originate from the hot interstellar medium ( ism ) which is distributed throughout the galaxy , and partly from the population of point - like sources , known as low mass x - ray binaries ( lmxbs ) ( trienchieri & fabbiano 1985 ) .
the spectra of the x - ray emission in these galaxies are typically fitted with a two - component model : a soft 1kev thermal component for hot , diffuse ism and a hard 1.6 photon index power law component for the contribution from the point sources ( irwin et al .
though the diffuse x - ray emission is a powerful tool for tracing the past star formation and central massive black hole activity through their effects on the ism ( strickland et al .
2000 ) , the later component is more important in the case of fainter galaxies ( matsushita 2001 ) .
the superb angular resolution of the _ chandra _ x - ray telescope for the first time has made possible to spatially resolve the point sources in extragalactic environment . in this paper
we present morphology and spectral properties of hot gas and point - like sources in some early - type galaxies . +
x - ray data on target galaxies ngc 3245 , ngc 3607 , ngc 3923 and ngc 4365 were acquired from public archive of the _ chandra _ x - ray observatory .
level-1 processed event files provided by the pipeline processing of _ chandra _ x - ray centre were used for this study .
the acquired x - ray data was homogeneously processed using ciao ( chandra interactive analysis of observations ) science threads . a more detailed discussion on the x - ray data analysis is given in strickland et al .
( 2002 ) .
the point sources present in the field of view centred on the target galaxies were identified and adaptively removed using the task _ wavdetect _ available within ciao .
the holes " due to this removal were filled by extrapolating the background counts using _
dmfilth_. finally , the filled - in " smooth , diffuse x - ray images were produced in different energy bands i.e. , soft ( 0.3 - 1 kev ) , medium ( 1 - 2 kev ) and hard ( 2 - 10 kev ) using the task _
csmooth_. figure 1(a ) shows one of such image delineating the detected x - ray point sources as well as the point sources within ngc 4365 .
region , ( b ) x - ray spectrum of hot gas alone in ngc 3923.,title="fig:",width=245 ] region , ( b ) x - ray spectrum of hot gas alone in ngc 3923.,title="fig:",width=245 ] the spectra of the combined emission from the detected point sources with counts more than 100 net counts were fitted by a power - law spectrum with photon index about 1.6 .
these results are consistent with the spectral analysis of the combined emission from lmxbs in a sample of 15 early - type galaxies analyzed by irwin et al .
( 2003 ) . while , the spectra of the diffuse emission component in the target galaxies were fitted by a hot plasma code ( vmekal ) along with a power law function . figure 1(b ) represents x - ray spectrum of ngc 3923 for the diffuse gas alone .
xspec v11.2 package was used for the spectral fitting .
we determined x - ray hardness ratios of the detected point sources following the technique discussed by sarazin et al .
( 2000 ) , which are useful for characterizing the spectral properties of the sources . hardness ratios of all the detected sources in target galaxies are plotted as h31 vs. h21 and are shown in figure 2(a ) , while figure 2(b ) represents the x - ray luminosity function ( xlf ) of the discrete point sources detected in the target galaxies .
the spectrum derived for ngc 3923 ( figure 1(b ) ) shows lines from s , si , ne , mg and fe . we tried to determine the si / fe and s / fe ratios by fitting a modified vmekal model to the diffuse gas component .
all the other elemental abundances were allowed to vary freely , except he which was set to solar value .
we find z@xmath5/z@xmath6 ratio in ngc 3923 to be about 4.2 , which is an indicator that the ism is enriched by the supernova ( sn ) explosions .
the resulting best - fit power - law indices fitted for diffuse gas and detected point - sources are consistent with those derived from fitting the combined emission from the detected point sources .
this shows that the spectral index of lmxbs does not vary significantly with the luminosity .
emission from lmxbs is significant in these galaxies and have a contribution lying between 30 to 65% to the total x - ray emission of the target galaxies .
10 irwin j.a .
, sarazin c.l . , bregman j.n . ,
2000 , apj , 544 , 293 irwin j.a . , athey a. & bregman j. , 2003 , apj , 393 , 134 matsushita k. , 2001 , apj , 547 , 693 sarazin c.l .
, irwin j.a . , bregman j.n . ,
2000 , apj , 544 , l101 strickland d.k . ,
heckman t.m . , weaver k.a . , & dahlem m. , 2000 , aj , 120 , 2965 stricklind d.k .
, heckman t.m . , weaver k.a . ,
hoopes c.g . & dahlem m. , 2002 , apj , 568 , 689 trienchieri g. , & fabbiano g. , 1985 , apj , 296 , 447 | we present results of morphological and spectroscopic study of hot gas in some early - type galaxies based on the analysis of high resolution x - ray images acquired from the archive of chandra space mission .
distribution of the hot gas in target galaxies after eliminating contribution from the discrete sources ( lmxbs ) displays varied morphologies , ranging from very compact nuclear emission to very extensive emission , larger than even optical images of the host galaxies .
the surface brightness profile of the hot gas in program galaxies is well described by a single beta model , while spectrum of the diffuse emission is best fitted by a combined soft mekal model and a hard power law model .
we use these results to derive temperature and abundance profiles of the hot gas in host galaxies .
the deprojection of the diffuse emission shows a temperature gradient in some of the galaxies .
we also report on the 2-d distribution of the discrete sources ( lmxbs ) in host galaxies and compare it with their optical morphologies .
the x - ray spectrum of the resolved sources is well - fit by a hard power law model with x - ray luminosities ( 0.3 to 10 kev ) in the range from 5@xmath0 10@xmath1 to 2.5@xmath0 10@xmath2 erg s@xmath3 .
x - ray luminosity function ( xlf ) of the lmxbs shows a break near the luminosity comparable to the eddington luminosity for a 1.4 m@xmath4 neutron star . galaxies : elliptical and lenticulars , early - type ; ism : x - ray emission |
the interplay between charge quantization and superconductivity offers the possibility to study macroscopic quantum phenomena in small - capacitance josephson junction circuits .
one of the most extensively studied devices in this context is the cooper pair transistor ( cpt ) , which consists of two josephson junctions in series , with a gate capacitively coupled to the small island formed between the two junctions .
such single island circuits are fundamental building blocks for quantum electronic circuits based on the number - phase uncertainty of the cooper pair condensate , and have been the object of growing interest due to their potential to realize new types of quantum limited measurements and solid state quantum bits .
essential for all these applications is the ability to inhibit the excitation of quasiparticles which give rise to dissipation ( noise ) , and thereby decoherence . for the cpt
this means that the charge parity of the island must be well controlled , so that quasiparticles , or single unpaired electrons , are not transferred to or from the cpt island .
here we study the cpt in a tunable high frequency electrodynamic environment , where the dc current - voltage characteristic ( ivc ) of the cpt exhibits a coulomb blockade of cooper pair tunneling . by tuning the impedance of the environment ,
we demonstrate how the parity of the island is controlled by the environment .
cpts have been most extensively studied in a high frequency environment with low impedance , @xmath3 , where the gate voltage modulates the critical current flowing through the device . at low temperature and in the absence of quasiparticle excitations ,
this modulation should be periodic in gate voltage with period @xmath4 , where @xmath5 is the gate capacitance .
@xmath6-periodic modulation is a clear sign that charge transport through the device is due to cooper pairs , which do not change the parity of the island .
however , in many experiments this parity effect is spoiled by the transfer of quasiparticles to or from the island , resulting in an @xmath2-periodic gate modulation , or modified @xmath6-periodic modulation , as the unpaired electron occupies the island .
for applications like the coherent control of a quantum bit , these quasiparticles have to be avoided , and several experimental studies have been carried out in order to understand and control this parity effect .
normal metal leads , or small normal metal traps " , contacting the superconducting source and drain close to the tunnel junctions are thought to help bring long - lifetime out - of - equilibrium quasiparticles into thermal equilibrium , and thereby enhance the parity effect @xcite .
however , in many experiments where no traps were used , a parity effect was demonstrated @xcite .
et al . _
@xcite showed that the lifetime of the unpaired quasiparticle on the island could be dramatically reduced by fabricating the island of the cpt with a larger superconducting energy gap than that of the source - drain . here
we present an experimental study of the cpt in a high impedance environment , @xmath7 , where far less experimental work has been done due to the difficulty in constructing such an environment . in a high impedance environment ,
the ivc of the cpt shows not a supercurrent , but rather a coulomb blockade of cooper pair tunneling .
the coulomb blockade is modulated with the gate voltage and , in the absence of quasiparticles , this modulation is @xmath6-periodic .
previously , the high impedance environment has been achieved with small on - chip resistors ( @xmath8 ) located close to the cpt , which have enabled the observation of a coulomb blockade of cooper pair tunneling with @xmath6-periodic dependence on the gate charge @xcite .
more recently , one - dimensional squid arrays have been used to bias a cpt @xcite which is the technique used in the present study .
the great advantage of the squid array bias is that the impedance of the environment can be tuned _ in situ _
, without affecting the cpt .
in this way the role of fluctuations due to the environment can be easily identified . in this article
we study the ivc of the cpt as the environment impedance is changed over many orders of magnitude . starting from a superconducting - like ivc at low impedance , a well defined coulomb blockade of cooper pair tunneling with a back - bending ivc develops as the impedance of the environment
is increased .
this back - bending ivc is due to the coherent tunneling of single cooper pairs and is a clear indication that @xmath7 .
we show that the gate voltage dependence of the ivc changes from @xmath2-periodic to @xmath6-periodic as the environment impedance is increased .
the high impedance environment suppresses quasiparticle tunneling rates , thereby restoring the even parity of the island .
these observations indicate new approaches for controlling out - of - equilibrium quasiparticles in superconducting quantum circuits .
sem micrograph of a cpt ( in the center ) biased by four squid arrays .
two squid arrays are employed to apply a symmetric bias , while the other pair is used to measure the voltage across the cpt . ]
figure [ sem ] shows a sem micrograph of a cpt sample biased by four squid arrays .
the al / alox / al tunnel junctions are fabricated on a sio@xmath9/si substrate using e - beam lithography and the standard shadow evaporation technique @xcite .
four arrays , each consisting of 70 squids in series , connect the cpt to the bias and measurement circuitry . the squids are designed to have two identical junctions in order to suppress the josephson coupling in the array as low as possible .
the nominal area of each junction in a squid is 0.03 @xmath10m@xmath11 , and the loop area is 0.18 @xmath10m@xmath11 .
each cpt junction has a nominal area of 0.01 - 0.02 @xmath10m@xmath11 . to provide an effective high impedance environment
, the arrays have to be located close to the cpt , so as to avoid the shunting capacitance of the leads .
the squid arrays have high impedance due to the josephson inductance of the squid @xmath12 , which becomes infinite as the critical current of the squid , @xmath13 , is suppressed to zero with an external magnetic flux .
this inductance is also non - linear , as it depends on the phase across the squid , and therefore a linear circuit model can not give a complete description of the electrodynamics of these arrays .
however , for current much less than the critical current , a linear approximation can be made , which shows that the infinite array behaves like a transmission line with a gap , or region of frequency where no propagating modes exist .
a finite array will have a set of discrete modes , and for arrays such as the ones used in this study , the lowest frequency mode has @xmath14 ghz @xcite .
the impedance of arrays such as these has not yet been systematically characterized due to the very high frequencies involved .
however , we have shown that these arrays do provide an environment which induces a coulomb blockade of cooper pair tunneling and bloch oscillations in a single josephson junction @xcite .
as the magnetic flux through the squid loops is increased , the arrays undergo a quantum phase transition @xcite where the measured zero bias resistance of the array increases several orders of magnitude ( 50 k@xmath15g@xmath1 ) .
a key aspect of this type of experiments is that the arrays can be tuned in this transition region to a point where they provide a high enough impedance at high frequencies , while not having too large @xmath16 .
if @xmath16 becomes too large , the input offset current of the voltage amplifier will inhibit the measurement of the differential voltage across the cpt .
the chip with the arrays and cpt is bonded to a printed circuit board with connectors , which is mounted in a rf - tight box on a dilution refrigerator with a base temperature of 15 mk .
the cryostat leads are twisted constatan pairs , which act as rather good microwave filters attenuating 35 db / ghz up to 2 ghz , above which a background level of -80 db is measured with an open cryostat .
additional room temperature rc filters were used to suppress lower frequency noise when necessary .
no microwave filters between the sample and twisted pairs were implemented in these measurements . however , in the high impedance case , the arrays themselves are good filters , protecting the cpt from electromagnetic fluctuations coming from the bias side of the arrays .
the ivc of the cpt is measured in a four point configuration with the cpt symmetrically biased through one pair of squid arrays , while the other pair is used to measure the voltage across the sample .
the current is measured with a trans - impedance amplifier ( modified stanford research system 570 ) on one of the biasing leads .
the voltage across the cpt is measured with a high input impedance differential voltage amplifier ( burr - brown ina110 ) .
the bias voltages @xmath17 are applied symmetrically with respect to ground in order to avoid a common mode drain - source voltage . in this way
, a polarization charge can be induced on the island only by the voltage applied to the gate and not from a bias - induced common mode voltage acting across the stray capacitance of the cpt to ground @xcite . [ cols="^,^,^,^,^,^,^,^",options="header " , ] the parameters of the measured samples are listed in table 1 . the normal state resistance per junction @xmath18 in the cpt
is calculated from the slope of the ivc at large bias voltage , assuming that both junctions are identical .
the josephson energy is @xmath19 , where @xmath20ev is the superconducting energy gap of al .
the capacitance of one junction in the cpt is calculated from the junction area as measured from sem pictures , using the specific capacitance 45 ff/@xmath10m@xmath11 . from the capacitance @xmath21 ,
the charging energy of one junction in the cpt is calculated as @xmath22 .
for all the measured samples , @xmath23 , so that the total island capacitance is given by @xmath24 . in table 1
, @xmath25 indicates the measured maximum threshold voltage for cooper pairs , as determined from the ivc of the cpt when @xmath26 .
@xmath27 refers to the maximum value of the measured coulomb gap for single electrons , estimated from the offset of the normal state ivc , taken in high magnetic fields where superconductivity is completely suppressed .
schematic of a cooper pair transistor .
the environment is represented by the impedance @xmath28 . when @xmath29 ( @xmath30 ) the cpt is embedded in a low ( high ) impedance environment .
@xmath31 indicates the charge and @xmath32 the phase difference across each junction .
@xmath33 represents the current flowing through junction 1(2 ) and @xmath34 the displacement current from the island to the gate electrode.,scaledwidth=40.0% ] most experimental and theoretical studies to date have considered the case of a cpt embedded in a low impedance environment , @xmath29 @xcite .
a few studies have been made for the case of a high impedance environment , @xmath30 @xcite .
below we review the theoretical description of the cpt for the low impedance case , showing how it can be transformed to a description appropriate for the high impedance case .
we then give a qualitative description of the ivc and bloch oscillations for the later case , comparing this with experimental results .
we consider the circuit depicted in fig .
[ cpt_scheme ] , consisting of a symmetric cpt with @xmath35 and @xmath36 .
the charge induced by the gate on the island is @xmath37 , where @xmath38 is the voltage applied to the gate electrode .
@xmath31 indicates the charge and @xmath32 the phase difference across each junction .
let @xmath39 be the bias phase across the cpt and @xmath40 the phase of the island .
the phases @xmath41 and @xmath42 are conjugate respectively to the charge transferred across the device @xmath43 and the total island charge @xmath44 and obey the commutation relations @xmath45=[q,\phi_{g}]=-2ie$ ] . for the moment
we neglect the contribution of quasiparticles by assuming that the superconducting energy gap @xmath46 is large compared to all relevant energy scales .
the hamiltonian of a cpt with @xmath47 biased by a current @xmath48 can be written as follows @xmath49 the first term of the hamiltonian corresponds to the charging energy associated with the total charge across the series combination of the two junctions of the cpt .
the second term is the electrostatic energy associated with the island which sees the two junctions in parallel .
the third term is the effective josephson energy of the cpt , which , due to its dependence on @xmath42 , can be modulated by the gate voltage .
the last term represents the coupling between the bias current @xmath50 and the phase @xmath41 . in a low impedance environment , @xmath29 , quantum fluctuations of the phase
are suppressed and the bias phase @xmath41 behaves as a classical variable . due to the low impedance , cooper pair tunneling at zero bias voltage dominates the ivc of the cpt which shows the typical superconducting behavior characterized by a supercurrent branch at vanishing voltages .
when @xmath51 , charging effects are observable and result in a @xmath6-periodic modulation of the critical current with the gate charge . thus , when @xmath29 , the cpt behaves as a single josephson junction with a gate charge dependent critical current @xcite .
let us turn to the case of a high impedance environment , @xmath30 , where quantum fluctuations of the phase are strong and the conjugate charge now becomes a classical variable . for this case , a quantitative theory of the single josephson junction in terms of bloch energy bands has been developed @xcite .
the dynamics of the junction can be described by the quasicharge @xmath52 , where @xmath53 is the current flowing through the junction .
the quasicharge evolves in time according to the langevin equation @xmath54 where @xmath55 is the voltage across the junction and @xmath56 indicates a random noise component of the bias current , induced by the real impedance of the environment @xmath57\gg r_{q}$ ] at finite temperature .
the voltage @xmath55 is given by the derivative of the lowest bloch energy band , @xmath58 .
the ivc of the junction can be calculated analytically from the langevin equation @xcite and shows a voltage peak near zero current corresponding to the coulomb blockade of cooper pair tunneling , and a back - bending region at finite currents .
this region of negative differential resistance is the indication of the bloch oscillations , which occur with frequency @xmath59 . lowest energy bloch band for a cpt with @xmath60 . for fixed gate charges , the quasicharges @xmath61 and @xmath62 follow a diagonal trajectory
the solid and dotted lines indicate respectively the quasicharges trajectories for @xmath63 and @xmath64 . ]
the natural extension of the bloch band picture from a single josephson junction to a cpt leads to a description of the quasicharge dynamics in two - dimensional bloch energy bands . in fig .
[ bands2d ] the lowest energy band of a cpt with @xmath60 is plotted in a space spanned by the two quasicharges @xmath61 and @xmath62 , which describe the dynamics of the two junctions of the cpt .
this description becomes clear if we rewrite the hamiltonian ( [ hcpt1 ] ) in terms of the variables @xmath31 and @xmath32 . for @xmath65
the hamiltonian ( [ hcpt1 ] ) reads @xmath66
this hamiltonian is that of two single josephson junctions , @xmath67 , coupled by the term containing the gate charge @xmath68 . for small coupling , @xmath47 , the behavior of a cpt in a high impedance environment
is described by the quasicharges @xmath69 , where @xmath33 indicates the current flowing through junction 1(2 ) ( fig .
[ cpt_scheme ] ) . in this case
the lowest energy band of the cpt is approximately given by the sum of the lowest energy band of the two single junctions , that is @xmath70 .
the conservation of current results in the constitutive relation @xmath71 , where @xmath34 indicates the displacement current flowing from the island of the cpt to the gate . from this relation follows the condition @xmath72 , which , for a fixed gate charge , constrains the quasicharges to a diagonal trajectory in the two - dimensional bloch band ( fig .
[ bands2d ] ) .
a change in the gate voltage induces a shift in the trajectory of the two quasicharges , as shown in fig .
[ bands2d ] . in analogy with the one dimensional case
, the voltage across the cpt can be expressed via the derivative of the lowest energy band along the diagonal direction as @xmath73 , where @xmath74 defines the quasicharge of the cpt .
the maximum coulomb blockade voltage of the cpt , or critical voltage @xmath75 $ ] , is a @xmath6-periodic function of the gate charge .
the dependence of the critical voltage on the gate charge is determined by the slope of the energy band according to the bloch band theory @xcite .
the critical voltage is maximum when @xmath63 , where the slope of the lowest energy band is maximized ( solid line in fig .
[ bands2d ] ) , while the minimum critical voltage is obtained for @xmath64 ( dotted line in fig .
[ bands2d ] ) .
thus , when @xmath30 the cpt behaves as a single josephson junction with a gate charge dependent critical voltage . in the case
@xmath76 , the energy - charge relation of the cpt is identical to that of a single josephson junction and the voltage @xmath55 across the cpt has a @xmath6-periodic dependence on the quasicharge . from the band diagram
one can see that as the gate charge is shifted , the dependence of the voltage on the quasicharge is modified .
when @xmath64 , the bloch oscillations in each junction are out of phase , resulting in an oscillation of the voltage across the cpt with the quasicharge having frequency @xmath77 @xcite . for @xmath78 , the lowest bloch band becomes more sinusoidal along the quasicharge axes , and the diagonal trajectory at @xmath79 becomes flat for a symmetric cpt . in this case the coulomb blockade for cooper pairs can be modulated to zero and the out of phase bloch oscillations cancel one another . a pictorial description of the bloch oscillations and back - bending ivc of a cpt , or single josephson junction with tunable critical voltage @xmath80 , current biased by a high impedance resistor .
@xmath81 indicates the zener current .
the time dependence of the voltage across the junction is shown for 4 different operating points in the ivc .
the shape of the bloch oscillations refers to the case of a junction with @xmath82.,scaledwidth=40.0% ] the ivc of a cpt in the high impedance environment shows the same features as that for a single josephson junction and can be calculated from the langevin equation for the quasicharge ( [ langeq ] ) .
we can picture how the bloch oscillations occur in the cpt , or single josephson junction with tunable critical voltage , by simply considering a non - linear capacitor biased by a voltage source @xmath83 through a resistor @xmath84 . in fig .
[ bo_cpt ] the generic shape of the ivc is shown , with the time evolution of the junction voltage sketched for four different operating points of the circuit .
we consider the case @xmath85 , where the non - linear capacitor has a periodic , saw - tooth form of the voltage - charge relation , given by the derivative of the lowest energy band along a diagonal trajectory with a maximum critical voltage .
( a ) : ivc of the two biasing squid arrays of sample cpt 9a for different magnetic fields .
the measured values for @xmath86 are : light - gray curve @xmath87 50 k@xmath1 , dark - gray curve @xmath88 500 k@xmath1 , black curve @xmath88 20 m@xmath1 .
( b ) corresponding ivcs of cpt 9a . ] when the bias voltage is below the critical voltage , @xmath89 ( operating point a ) , the langevin equation ( [ langeq ] ) will have a stationary solution with @xmath90 , and the non - liner capacitor will simply charge up to the value @xmath83 with no voltage oscillation . as the bias
is increased so that @xmath91 , time dependent solutions to the langevin equation are possible , which describe overdamped , forced bloch oscillations in the junction .
these oscillations result in a periodic charging and discharging of the non - linear capacitor , which lead to a time average voltage @xmath92 ( operating point b ) . as the bias voltage is further increased , the bloch oscillations increase in frequency as higher current is forced through the junction . with faster bloch oscillations ,
the finite charging time due to the bias resistor becomes negligible , and the voltage oscillations become more symmetric around zero , with @xmath93 ( operating point c ) . increasing the bias voltage further would cause back - bending of the ivc to dissipation - less transport , or dc current with zero average voltage .
however , with higher frequency bloch oscillations , zener tunneling to higher energy bands takes place , resulting in a temporary build - up of the junction voltage to a level close to the superconducting energy gap @xcite . in this case , quasiparticle tunneling can no longer be neglected , and josephson - quasiparticle tunneling cycles , which are dissipative processes , can occur . as the zener tunneling probability increases with the bias current , the time average voltage across the junction increases ( operating point d ) .
the magnitude of the zener current depends on the energy gap between the lowest and first bloch energy band , which in turn depends on the ratio @xmath94 @xcite .
the ivc for three different cpts having approximately the same value of the zero bias resistance of the squid arrays ( @xmath95m@xmath1 ) .
black curve : cpt 9a , @xmath96 , light - gray curve : cpt 6 , @xmath97 dark - gray curve : cpt 8 , @xmath98 . ]
we find qualitative agreement with this theoretical picture in our experimental ivcs of cpts biased by squid arrays .
figure [ ivenv]a shows the ivc of two biasing squid of sample cpt 9a measured in series at three values of the magnetic field , between zero and @xmath99 . here
we see that the arrays change their zero bias resistance @xmath16 over several orders of magnitude as the josephson coupling of the squids is suppressed . in fig .
[ ivenv]b the corresponding ivcs of the cpt are plotted for the same magnetic fields . at low @xmath16 of the arrays ,
the ivc of the cpt shows a superconducting - like feature that disappears as the coulomb blockade region becomes progressively more defined for increasing values of @xmath16 .
when @xmath100
m@xmath1 , the coulomb blockade for cooper pairs is fully developed and is followed by a region of negative differential resistance .
this back - bending ivc is the hall mark feature of a coulomb blockade of cooper pair tunneling , and clear evidence that a high impedance of the environment has been achieved .
figure [ ivcpt ] shows the ivc of three different cpt samples with different values of @xmath94 . in each case
the squid arrays are tuned to approximately the same value of @xmath16 where the back - bending ivc is well defined .
figure [ ivcpt ] clearly demonstrates qualitative agreement with the theory described above . with increasing @xmath94 ratio , the critical voltage decreases , as the lowest bloch band develops more shallow minima .
we also see that the zener current , becomes larger with increasing @xmath94 ratio , as the gap to higher energy bands becomes larger .
as discussed in the previous section , the two - dimensional bloch band picture describing transport in the cpt in terms of the quasicharge leads to a 2@xmath2-periodic modulation of the ivc with gate voltage .
this picture only considers cooper pair tunneling , completely neglecting quasiparticle or single electron tunneling .
if the single electron tunneling rates are large enough , they can disrupt this picture , causing @xmath2-periodic modulation of the ivc . in our experiments
we find that the impedance of the environment can be used to suppress these quasiparticle tunneling rates low enough for stable 2@xmath2-periodic behavior to be observed .
figure [ e2e_env ] shows how the periodic modulation of the threshold voltage with gate at a constant current @xmath101pa evolves from 2@xmath2-periodic to @xmath2-periodic as the impedance of the environment is tuned . at high impedance , where the coulomb blockade is well developed
, we find 2@xmath2-periodic behavior , as expected from the two - dimensional bloch band picture ( fig .
[ e2e_env]c ) . as the impedance of the environment
is reduced , the effect of quasiparticle poisoning " is seen as a small peak in the threshold voltage at gate charges @xmath102 ( fig .
[ e2e_env]b ) , which grows until an @xmath2-periodic behavior is observed ( fig .
[ e2e_env]a ) .
a similar crossover from 2@xmath2 to @xmath2-periodic behavior as the environment impedance is tuned has been observed by watanabe @xcite and kuo _ et .
_ @xcite .
( a , b , c ) : gate - induced voltage modulation of sample cpt 9a for increasing values of the zero bias resistance of the biasing squid arrays .
the bias current is fixed at @xmath103pa .
( a , b , c ) : ivc of the cpt for the same values of @xmath86 .
( a , a ) : @xmath104 m@xmath1 , ( b , b ) : @xmath105 m@xmath1 , ( c , c ) : @xmath106 m@xmath1 . ] to understand how quasiparticles affect the critical voltage of a symmetric cpt in a tunable environment it is instructive to determine the regions in the bias voltage gate voltage plane where it is energetically favorable for a single quasiparticle to tunnel on or off the island , thereby changing the island parity .
as we will see , the shape of the corresponding stability diagrams depends on the impedance of the environment as well as on the ratio of the two characteristic energies for adding or removing a single quasiparticle to or from the island .
these energies are the charging energy @xmath107 of the island and the free energy difference @xmath108 that lifts the odd parity from the even parity states @xcite . here , the entropy term involves @xmath109 , the number of quasiparticle states available within temperature @xmath110 just above the gap ( @xmath111 is the normal - state density of states at the fermi level and @xmath112 the volume of the island ) . at zero temperature ,
@xmath113 and this energy difference is the superconducting energy gap @xmath46 , which is paid for having one unpaired electron on the island .
the free energy difference @xmath114 vanishes at a temperature @xmath115 , generally much lower than the superconducting gap .
let us first consider the situation of a free energy difference smaller than the island charging energy , @xmath116 . in fig . 8a , we plot the corresponding stability diagram for the case of a low impedance environment , where tunneling is elastic and we can apply the global rule @xcite
. the stability diagram contains lines which separate regions in the bias voltage gate voltage plane with stable parity from regions where it is energetically favorable for a single quasiparticle to tunnel on or off the island , thereby changing the parity . within the dotted rhombi
the transition rate @xmath117 from odd to even parity is zero , while within the solid rhombi the rate from even to odd parity @xmath118 vanishes .
outside the rhombi , a parity change corresponds to a decrease in energy and the transition rates @xmath117 and @xmath119 are non - zero at @xmath120 .
we conclude that the odd parity states are stable near the gate charges @xmath121 ( fig .
8a ) . at these gate charges
the system will be poisoned " by the excess quasiparticle which sits on the island in the lower energy state .
when in the odd parity state , the effective island charge is shifted by @xmath2 and we thus expect the gate voltage dependence of the critical voltage of the cpt to be shifted correspondingly ( from the solid to the dashed line in fig .
in particular , as gate voltage is changed , the critical voltage should have two maxima per @xmath6-period ( near gate charges @xmath121 and @xmath122 ) .
this is in qualitative agreement with the low impedance data of fig .
7a and 7b . in a high impedance environment ,
quasiparticle tunneling is inelastic as a tunnel event is accompanied by energy exchange with the environment and the local rule @xcite should be applied to establish the stability diagrams . as a result ,
the size of the rhombi doubles , and various stability regions start to overlap .
overlap means that there are bistable regions where either parity is stable , depending on the history of system .
however as can be seen in fig . 8b , for @xmath116 , odd stability regions
( dashed rhombi ) still exist near gate charges @xmath121 , and we expect the critical voltage to show two maxima per @xmath6 period .
this observation is in contradiction with the experimental high impedance data shown in fig .
7c , where only one critical voltage maximum per @xmath6 period is observed , near gate charges @xmath123 .
we therefore consider the stability diagram for the case @xmath124 , first in a low impedance environment , as shown in fig .
compared to the case @xmath125 ( fig .
8a ) , the odd stability regions now disappear , giving rise to regions near the gate charges @xmath121 where both rates @xmath126 and parity will fluctuate at small bias voltage . however , in a high impedance environment for @xmath127 , as shown in fig . 8e , we observe that the odd stability regions completely disappear , giving way to overlapping regions of even stability only , as long as the applied bias voltage satisfies the inequality @xmath128 .
this is markedly different from the low - impedance case , where regions of fluctuating parity exists at any finite bias .
we thus conclude that the critical voltage is @xmath6-periodic as a function of gate voltage ( fig .
8f ) for high impedance and @xmath129 , in agreement with the experimental findings of fig .
this stability analysis shows that quasiparticle poisoning can occur near the gate charges @xmath121 if conditions are not favorable .
the measured , long - time average cpt voltage near these gate charges will depend on the relative weight of the odd and even states , which are fluctuating on a much more rapid time scale determined by the rates @xmath130 .
the value of these rates can not be inferred from the stability analysis alone @xcite . for @xmath131
the effective island charge will most often be even , hence the measured voltage of the cpt will be minimal near @xmath121 ( see fig .
the effects of poisoning will not be noticed on the long time scale of the experiment and one observes a stable 2@xmath2-periodic modulation of transport . in the opposite case @xmath132 near @xmath121 ,
the decay of the odd state is much slower than that of the even state .
the effective island charge will be predominantly odd and the measured average voltage , instead of taking a minimum value , will reach a local maximum , that eventually may become equal to the maxima found at @xmath133 ( see fig .
8c ) . the measurements for low - impedance presented in fig .
7a show @xmath2-periodicity of the critical voltage with gate voltage , we conclude that the ratio @xmath134 .
this is in agreement with several other experiments on cpts with @xmath124 in low impedance environments @xcite .
we also note that the observed @xmath2-periodicity excludes that @xmath114 is large compared @xmath135 .
if @xmath136 , we would have found overlapping regions of stable even parity near @xmath121 , rather than regions of fluctuating parity and the resulting dependence of critical voltage on gate voltage would have been @xmath6-periodic in the low impedance case .
we finally comment on the effect of finite temperature .
the energy difference @xmath114 depends on temperature @xmath110 . for the islands used in this experiment , the characteristic crossover temperature @xmath137 where @xmath138 can be estimated to be approximately 250 mk .
the temperature @xmath110 appearing in the expression for @xmath114 does not necessarily correspond to the cryostat temperature , and could be an effective temperature , describing the fluctuations in the environment .
previous analysis of single junctions ivcs biased by similar arrays @xcite showed an effective noise temperature approximately 150 mk , which would result in @xmath139 in agreement with our conclusions above .
a second effect of thermal fluctuations is that they mix states of different parity whose energy difference is @xmath140 . on the level of the stability diagrams
this means that the lines separating stability regions of different parity smear into strips whose width is @xmath141 .
for the low impedance case this has drastic effects , even at low bias , because it leads to broadening of the parity fluctuation regions near @xmath121 thereby enhancing the @xmath2-periodic gate modulation .
for high impedance the even parity stability region has a gap with respect to the bias voltage , therefore thermal fluctuations do not effect the @xmath6-periodicity as long as @xmath142 .
the fact that parity stability is robust against fluctuations for small bias voltage in a high impedance environment is directly connected with the inelastic nature of tunneling in this limit .
this inelasticity is at the origin of the bias voltage gap and leads to an exponential suppression of quasiparticle tunneling at low bias voltage .
this effect is absent for low impedance as tunneling is elastic .
a , b , c , d : stability diagrams for the cases @xmath125 with @xmath29 and @xmath30 ( a and b ) , @xmath143 with @xmath29 and @xmath30 ( d and e ) .
c ) @xmath2-periodic gate voltage dependence of the critical voltage at finite bias voltage ( @xmath144 ) corresponding to the stability diagrams of a , b and d. f ) @xmath6-periodic gate voltage dependence of the critical voltage at finite bias voltage ( @xmath144 ) corresponding to the case e).,scaledwidth=48.0% ] the condition for stable @xmath6-periodic behavior is not preserved at all bias currents in the high impedance environment .
figure [ e2ebias ] shows the gate voltage modulation for high impedance ( @xmath145 m@xmath1 ) of the cpt voltage for three different values of the bias current .
> from low current levels up to the back - bending region of the ivc , the voltage modulation is stable 2@xmath2-periodic ( fig .
[ e2ebias]a ) . as the current
is increased toward the zener current , which for this sample is measured at @xmath146 pa , the rate of single electron transitions increases , as explained in the previous section .
zener tunneling causes an increase in the rate @xmath118 near @xmath121 , and the effect of poisoning can be observed ( fig .
[ e2ebias]b ) . at even higher currents ,
the quasiparticle tunneling rate is so large that only @xmath2-periodic modulation can be seen ( fig .
[ e2ebias]c ) .
gate voltage modulation for sample cpt 9a at various bias current .
the zero bias resistance of the biasing squid arrays is @xmath100 m@xmath1 .
the ivc of the cpt for this magnetic field is shown in fig .
[ ivenv ] . ]
the effect of the electrodynamic environment on single charge tunneling in the cooper pair transistor ( cpt ) has been studied experimentally by biasing and measuring the cpt through one - dimensional squid arrays .
the impedance of these arrays can be tuned with an external magnetic field without changing the parameters of the cpt , which allows one to clearly discern the effect of the environment on tunneling . as the impedance of the arrays
is increased , the cpt develops a coulomb blockade of cooper pair tunneling with a well defined back - bending current - voltage characteristic ( ivc ) , indicative of the coherent transfer of single cooper pairs .
the general shape of the measured ivc is that predicted by a theory based on quasicharge dynamics .
we find that the observed threshold voltage depends on the ratio between the josephson energy and the charging energy in a way that is in qualitative agreement with the bloch band theory for the cpt . as the impedance of the environment
is increased , we observe a transition from @xmath2 periodic to @xmath6-periodic modulation of the ivc , demonstrating that the high impedance environment restores the parity effect in the cpt .
enhanced parity effect , or reduced quasiparticle poisoning , is explained by the inelastic nature of single - charge tunneling in a high impedance environment , where a charging energy gap makes odd parity states unstable near gate charges @xmath64 .
this work was partially supported by the swedish vr , ssf nanodev program and the eu project squbit .
samples were made in the kth nanofabrication lab , provided by the k. a. wallenberg foundation .
fwjh acknowledges support from the institut universitaire de france. | we have measured the cooper pair transistor ( cpt ) in a tunable electromagnetic environment consisting of four one - dimensional squid arrays .
the transport properties of the cpt in the high impedance limit , @xmath0k@xmath1 , are studied for different ratios of the josephson coupling energy to the charging energy . as the impedance of the environment
is increased , the current - voltage characteristic ( ivc ) of the cpt develops a coulomb blockade of cooper pair tunneling and the measured ivcs agree qualitatively with a theory based on quasicharge dynamics for a cpt .
increasing the impedance of the environment induces a transition in the modulation of the ivc with the gate charge from @xmath2-periodic to 2@xmath2-periodic . |
quantization of electronic motion by a strong magnetic field leads to many qualitatively new effects .
they are most pronounced in two - dimensional systems , where the motion along the field is frozen by quantum confinement , so the continuous electronic spectrum is transformed into a series of discrete landau levels ( lls ) .
probably the most famous examples of such new effects are the integer and the fractional quantum hall effects @xcite , which arise because the degenerate electronic states on a ll are profoundly modified by disorder or lateral confinement potential , and by electron - electron interaction , respectively .
electron - phonon coupling can also produce new effects , especially when the energy of an inter - ll excitation matches that of an optical phonon .
such resonances may lead to strong - coupling effects even if the electron - phonon coupling in the absence of the magnetic field was weak .
magneto - phonon resonances were first discussed theoretically in relation to the dc electron transport in doped bulk semiconductors subject to a strong magnetic field where the resonance leads to an enhancement of the electron - phonon scattering rate @xcite .
they were later observed experimentally in insb @xcite .
a doublet structure was observed in the infrared absorption spectrum of bulk insb in a magnetic field , which was attributed to the resonant coupling of an electronic excitation , the cyclotron resonance mode , to optical phonons , when tuned in resonance @xcite .
this is a strong - coupling effect , whose detailed theory was presented in ref .
later , it was suggested that such magneto - phonon resonance need not be restricted to doublets : three branches of excitations involving zero , one , and two phonons may also cross at the same value of the magnetic field , resulting in a triplet structure of the absorption peak @xcite .
in fact , equidistancy of lls for electrons with parabolic bands leads to a rich variety of possible resonant combinations @xcite .
still , we are not aware of any observation of a resonant splitting of more than two excitation branches in conventional semiconductors .
graphene , with its conical electron dispersion and exceptional crystal quality , opens new possibilities for observation of fundamental physical effects . in a magnetic field
@xmath1 , perpendicular to the crystal plane , the ll energies are @xmath2 , where @xmath3 is the landau level index , @xmath4 is the dirac velocity , and @xmath5 is the magnetic length . due to the @xmath6 dependence ,
the spacing between lls with small @xmath7 is larger than for electrons with parabolic spectrum , which enabled the observation of the quantum hall effect at room temperature @xcite , unprecedented in conventional semiconductor structures .
magnetophonon resonances have been observed in raman scattering on optical phonons @xcite , where the raman g peak acquired a doublet structure .
here , we report an observation of magnetophonon resonances where the triple peak structure is clearly seen . to the best of our knowledge
, it is the first observation of such resonances in solid - state physics . despite the generality of the observed magnetophonon resonance effects
, graphene introduces some qualitatively new aspects .
first , electronic bands in graphene are conical rather than parabolic , so the lls are not equally spaced in energy .
this leads to a classification of possible resonant combinations which is quite different from that in conventional semiconductors .
such classification for graphene is developed below .
second , while in an undoped semiconductor all lls in the conduction band are empty , and those in the valence band are completely filled , in graphene there is always a partially filled ll ( unless the magnetic field is tuned to some special values ) ; if this partially filled level is involved in the resonance , the physics of the electron - phonon coupling is changed significantly .
namely , it turns out that the multiple - peak structure can not be viewed as a resonant splitting of a few discrete levels ; one has to face a true many - body problem , so the peak splitting is necessarily accompanied by a broadening of a similar magnitude , and complicated spectral shapes may be produced .
this physics was overlooked in the previous theoretical studies of magneto - phonon resonances in graphene @xcite .
our experiments are performed on extremely pure flakes of graphene which can be found on the surface of bulk graphite @xcite . to probe the magnetophonon resonances , we use raman scattering , which is a powerful and popular tool for characterization of graphene @xcite .
inter - ll electronic excitations can be observed in the raman spectrum of graphene @xcite , and the strongest features are due to transitions @xmath8 @xcite .
in addition , transitions @xmath9 are also observed @xcite even though theory predicts them to be weaker @xcite . in this work
, we trace the evolution of inter - ll excitations in graphene on graphite in magnetic fields up to @xmath10 t. such fields are needed to tune the inter - ll excitation energies across the energies of optical phonons and their combinations .
we observe different series of magnetophonon resonances , involving both phonons from the vicinity of the @xmath11 and @xmath0 points of the first brillouin zone .
the paper is organized as follows . in sec .
[ sec : qualitative ] we discuss qualitatively different types of magnetophonon resonances which can occur in graphene , and summarize our main results . in sec .
[ sec : experiment ] we give a detailed presentation of our experimental results . in sec .
[ sec : theory ] the theory is developed for the various cases discussed in sec . [ sec : qualitative ] .
( color online ) different types of magneto - phonon resonance in excitation decay . ( a )
an optical phonon decaying into an electronic excitation .
( b ) an electronic excitation decaying into an optical phonon and another electronic excitation .
the horizontal lines represent the landau levels , the black circles represent their filling .
the filled red circle represents the excited electron , the empty red circle represents the hole , and the wavy line represents the phonon .
, width=302 ] in the following , we use a shorthand notation @xmath12 for an excitation of an electron from a filled level @xmath13 to an empty level @xmath14 ( assuming that the partially filled level is @xmath15 ) . to gain a qualitative insight into various resonant processes , it is instructive to analyze the weak coupling limit first .
namely , consider an initial excitation which was created in the sample in the course of raman scattering , and analyze different channels of decay for this excitation ( see fig .
[ fig : decay ] ) .
one case is when the initial excitation is a phonon ; for example , the doubly degenerate @xmath16 phonon with the wave vector at the @xmath0 point of the first brillouin zone , giving rise to the @xmath17 peak in the raman spectrum .
it may decay into an electronic inter - ll excitation , as schematically shown in fig .
[ fig : decay](a ) .
the selection rule for the matrix element is @xmath18 , so the phonon decay rate is strongly enhanced when the phonon is resonant with one of the corresponding transitions , that is , the phonon energy @xmath19 for some @xmath7 .
this case corresponds to the magneto - phonon resonance in the raman scattering on phonons , which was observed experimentally @xcite , following the theoretical prediction in refs .
@xcite . in the present work ,
we focus on a different case , that of the initial excitation being an inter - ll excitation which decays into a phonon and another inter - ll excitation .
again , the corresponding decay rate is enhanced at resonance , but the resonance condition is different from the previous case .
for example , if the initial excitation is @xmath20 , it can decay if the electron on the level @xmath7 or the hole on the level @xmath13 emits a phonon and moves to another level @xmath21 or @xmath22 , respectively ( @xmath23 ) .
the decay is resonant if the phonon energy @xmath24 .
it is essentially the same process that is responsible for the magneto - phonon resonance in electronic transport in semiconductors @xcite , as well as in magneto - optical absorption @xcite .
the role of momentum conservation is quite different in the two processes .
as we focus here on optical excitation , the momentum of the incident and scattered photons can be neglected , so in both cases , the total momentum of the initial excitation is zero
. if the initial excitation is a single phonon , it can be only the @xmath16 optical phonon from the @xmath0 point , and the total momentum of the inter - ll excitation into which it decays should also be zero , which leads to the selection rule @xmath18 , and thus restricts this case to the magneto - phonon resonance studied in refs .
@xcite . for a multi - phonon initial excitation
, one can also think about the resonant enhancement of the phonon decay rate ; however , the width of the multi - phonon raman peaks is dominated by other factors than the phonon decay @xcite , so the modification of latter by the magnetic field is not observed in the raman spectra . for an initial electronic inter - ll excitation of zero total momentum ,
the momentum conservation allows the decay into a phonon with an arbitrary momentum and another inter - ll excitation with the opposite momentum .
still , the matrix element is strongly suppressed unless the phonon momentum is within @xmath25 from the @xmath0 or @xmath26 points of the first brillouin zone .
thus , one has to consider two kinds of resonances : ( i ) those involving the @xmath16 optical phonons around the @xmath0 point and intravalley electronic inter - ll excitations , and ( ii ) those involving intervalley inter - ll excitations and phonons around the @xmath26 points . among the latter phonons , those most strongly coupled to the electrons are the @xmath27 transverse optical phonons .
the corresponding phonon energies are @xmath28 and @xmath29 ( see sec .
[ sec : experiment ] ) . another important consequence of finite phonon momenta ( @xmath25 ) involved in the process , is the absence of any selection rule on the ll indices . in the above example of the decay of the @xmath20 excitation into a phonon and the @xmath30 , any @xmath21 is allowed ( energy conservation requires @xmath23 ) .
the above discussion of decay of the initial excitation is valid in the limit when the electron - phonon coupling is weak compared to the typical broadening of the landau levels or to the dispersion of optical phonons or inter - ll excitations on the momentum scale @xmath25 . in the opposite limiting case , the problem corresponds to the coupling between discrete energy levels , when the states become strongly mixed .
the energies of these hybrid excitations ( magnetopolarons ) exhibit an anticrossing as a function of the magnetic field as the latter is swept through the resonance region . for an initial electronic inter - ll excitation ,
different types of resonances are possible .
a resonance occurs when the phonon energy @xmath31 matches the energy difference between some empty landau levels : @xmath32 for some @xmath33 ( we assume to be at zero temperature and zero doping , so @xmath15 is the only partially filled level ) . due to electron - hole symmetry , @xmath34 , eq .
( [ resonance= ] ) automatically yields a similar condition for two filled levels , @xmath35 .
also , because of the @xmath6 dependence of @xmath36 , the lls are not equidistant , but still , some additional degeneracies may arise .
for example , if @xmath37 , this automatically yields @xmath38 , while if @xmath39 , there are no additional degeneracies .
( color online ) examples of different consequences of a resonance of the type of eq .
( [ resonance= ] ) .
( i ) irrational @xmath40 for the resonance @xmath41 .
no further resonant steps are possible .
( ii ) rational @xmath42 for the resonance @xmath43 , which automatically yields the next resonant step , @xmath44 .
the @xmath15 level is not involved .
( iii ) integer @xmath45 for the resonance @xmath46 , which yields the next resonant step , @xmath47 , involving the partially filled zero level .
the horizontal lines represent the landau levels , the black circles represent their filling .
the filled red circles represent the excited electrons , the vertical arrows show their transitions which are resonant with a phonon .
, width=188 ] all resonances of the kind ( [ resonance= ] ) can be classified into three groups , depending on whether the ratio @xmath48 is integer , rational , or irrational ( see fig .
[ fig : resonances ] ) .
let us see what happens in each of these cases , restricting our attention to the initial excitations of the type @xmath20 , which dominate the electronic raman scattering in the magnetic field @xcite , and to @xmath49 ( as well as its electron - hole - symmetric counterpart @xmath50 ) , which are also observed in the raman spectra , as discussed in sec .
[ sec : experiment ] .
( i ) if @xmath51 is irrational , the resonances do not proliferate .
then the simplest case is that of the initial @xmath49 excitation which is resonant with @xmath52 plus one phonon . in this case , there are no hole resonances ; indeed , @xmath53 at the same time would imply @xmath54 for some @xmath55 , which is impossible due to the concavity of the square root .
the symmetric excitation @xmath20 is then resonant with @xmath30 ( or @xmath56 ) plus one phonon , but also with @xmath57 plus two phonons , i. e. the electron and the hole resonances necessarily occur simultaneously .
as the excitations with zero , one , and two phonons disperse differently with the magnetic field , the symmetric excitation gives a triple anticrossing of the levels as a function of the magnetic field .
such triple anticrossings were discussed long ago in the context of optical magnetoabsorption in bulk semiconductors @xcite .
( ii ) let @xmath51 be rational but not an integer , and let @xmath14 be the integer part of @xmath51 . in this case , condition ( [ resonance= ] ) yields not just one , but @xmath14 resonances .
indeed , for any @xmath58 the energy @xmath59 matches the energy @xmath60 of the level @xmath61 ^ 2 $ ] .
the last level in this sequence has the energy @xmath62 , which does not allow for emission of any more phonons .
however , if we construct @xmath63 ^ 2 $ ] , then it turns out that @xmath64 , or , equivalently , @xmath65 .
this means that the phonon itself is resonantly hybridized with the electronic excitation from the filled level @xmath66 to the empty level @xmath67 ( as the phonon involved has a finite wave vector , there are no selection rules on @xmath68 ) .
thus , the asymmetric excitation @xmath69 is resonant with @xmath70 plus @xmath71 phonons for any @xmath72 , so we have an anticrossing of @xmath73 energy levels with different dispersions ( the hybridization of the phonon with @xmath74 or @xmath75 does not add anything , as these electronic excitations disperse with the magnetic field in the same way as @xmath76 ) .
the symmetric excitation @xmath20 can resonantly emit phonons both on the electron and hole sides , up to @xmath77 phonons in total , corresponding to the anticrossing of @xmath78 levels .
it is important that for non - integer @xmath51 none of the resonances involve the partially filled @xmath15 landau level .
( iii ) if @xmath79 , an integer , then the last term of the sequence @xmath80 from the previous case is @xmath81 .
the resonance counting is analogous to case ( ii ) .
namely , we have @xmath73 anticrossings for the asymmetric initial excitation @xmath69 and @xmath78 anticrossings for @xmath20 , so that the latter can thus be entirely transformed into @xmath77 phonons by a sequence of resonant transitions . however , involvement of the partially filled @xmath15 level brings about a qualitatively new feature : the possibility of creation of an arbitrary number of zero - energy electronic excitations on the partially filled level . as will be discussed in sec .
[ sec : theoryiii ] , this feature makes a consistent theoretical treatment of case ( iii ) quite problematic . our experimental results , presented in sec .
[ sec : experiment ] , include a clear observation of several resonances of type ( iii ) with double and , for the first time , triple avoided crossings .
near each avoided crossing , we describe our data phenomenologically by an effective two- or three - level model with some coupling between the resonant excitations .
we also observe some signatures of resonances of type ( i ) , which are , however , close to the limits of our experimental resolution , and they do not allow for a quantitative anlysis .
resonances of type ( ii ) are quite difficult to observe , as the smallest landau level indices @xmath82 which produce non - integer rational @xmath51 are @xmath83 , @xmath84 , leading to @xmath85 .
the corresponding magnetic field is too low , and the landau levels involved in the electronic excitation are too high , to be resolved in our experiment . the theory for all three cases is presented in sec .
[ sec : theory ] .
its main task is to justify ( or disprove ) the phenomenological description of each avoided crossing by an effective few - level model .
clearly , such description may be correct only if one can neglect the energy dispersion of the electronic excitations and of the phonons .
the former can be neglected if the coulomb interaction is sufficiently weak ( e.g. , due to screening by the conducting graphite substrate ) . as for the latter
, one should recall that the most important contribution to the dispersion of phonons near the @xmath0 or @xmath11 points comes from their coupling to the dirac electrons which produces the kohn anomaly @xcite . in a strong magnetic field
the spectrum of dirac electrons is quantized , so the main effect of the electron - phonon coupling is to induce magneto - phonon resonances .
the residual part of the phonon dispersion which is due to interaction with electrons in the @xmath86 bands is quite weak and can indeed be neglected , as the relevant scale of the phonon wave vectors is quite small , @xmath87 ( @xmath88 is the magnetic length ) .
still , it turns out , that even if one starts from dispersionless excitations , energy dispersion can be eventually generated due to resonant coupling .
this effect is especially dramatic for case ( iii ) , where plenty of zero - energy excitations are available on the partially filled @xmath15 landau level . as a result ,
the splitting between the energies of the coupled excitations is accompanied by a broadening , whose magnitude is of the same order as the splitting .
the broadened peaks have complicated spectral shapes , not described by any simple functional form , such as lorentzian , gaussian , or any other ( see fig .
[ fig : spectrint ] in sec .
[ sec : theoryiii ] for an example ) .
thus , the phenomenological description of our experimental data by effective two- or three - level models is very approximate .
the splitting of the @xmath89 electronic excitation , observed when it is in resonance with the @xmath0 point phonon near @xmath90 t @xcite , can be ascribed to either of the two types of resonances mentioned in the beginning of this section .
indeed , on the one hand , it can be viewed as the decay of the initial phonon into an electronic excitation . on the other hand
, it can also be viewed as the decay of the initial @xmath89 electronic excitation into a phonon and a zero - energy @xmath91 excitation , falling into case ( iii ) above . due to such `` double - faced '' nature of this particular resonance ,
the split peaks are strongly broadened .
this intrinsic broadening is missing in the simple description of refs .
for the resonance between the @xmath11 point phonon and the @xmath89 excitation , coupling to an external continuum of states was introduced in ref .
@xcite to describe the strong broadening of the peaks ; here we see that the broadening , in fact , originates from the electron - phonon coupling itself , and there is no need to invoke any extrinsic effects .
the broadening effects are less dramatic in the cases ( i ) and ( ii ) .
as discussed above , case ( ii ) necessarily involves landau levels with quite large @xmath92 .
this strongly suppresses the broadening effect : the main peaks are narrow and they contain the most of the spectral weight ( sec . [ sec : theoryii ] ) . in case ( i ) , discussed in sec .
[ ssec : irrational ] , for asymmetric transitions @xmath69 no broadening is generated at all , so their resonance with @xmath93 and a phonon can be faithfully described by an effective two - level model .
no broadening arises also for symmetric transitions @xmath20 resonating with @xmath30 and the @xmath11 point phonon , so the effective three - level description is justified . however ,
if @xmath0 point phonons are involved in the resonance , broadening is generated due to a peculiar phonon exchange effect : the phonons emitted by @xmath94 and @xmath95 excitations which constitute the initial @xmath20 excitation , can be reabsorbed in the opposite order ; this is impossible for the @xmath11 point phonons as the two emitted phonons must necessarily belong to the opposite valleys .
the low - temperature magneto - raman scattering response , at the m scale , has been measured using a home - made miniaturized optical probe based on optical fibers ( a 5 m core mono - mode fiber for the excitation and a 50 m core for the collection ) , lenses and optical band pass filters to clean the laser and to reject the laser line .
the excitation laser at @xmath96 nm is focused on the sample with a high numerical aperture aspherical lens down to a spot of the diameter @xmath97 .
the sample is placed on piezo stages to allow for the spatial mapping of the raman response of the sample .
this optical set - up is inserted in a closed jacket , with some helium exchange gas , which is then immersed in liquid helium and placed in the center of resistive solenoid producing magnetic fields up to 30 t. the unpolarized raman scattering response of our samples has been measured in the quasi - backscattering geometry with the magnetic field @xmath1 applied perpendicularly to the graphene crystal plane .
this experimental set - up has been described in more details in refs .
@xcite . to find a graphene flake on the surface of bulk graphite , we use the methodology based on the spatial mapping of the raman scattering response of the surface of bulk graphite with an applied magnetic field described in ref .
@xcite .
a ) characteristic raman scattering spectra at selected values of the magnetic field , from which the zero - field response has been subtracted .
the vertical gray bars mask the residual , after normalization , phonon contributions , due to the faraday effect in the experimental set - up .
b ) false color map of the raman scattering response of the graphene - like location , differentiated with respect to @xmath1 , as a function of the magnetic field ( with the step in @xmath1 of 100 mt ) .
main @xmath12 electronic excitations are indicated .
the horizontal solid line represents @xmath98 , the dashed line @xmath99 , the dashed - dotted line is at the energy position of the 2d band feature while the dotted line represents @xmath100 .
the three white ( black ) boxes identify spectral and magnetic field regions around @xmath101 and @xmath102 ( @xmath103 and @xmath104 ) where the electron-@xmath0 point ( k point ) phonon resonance affects the spectra . ]
the typical raman spectra , from which the @xmath105 response has been subtracted , are shown in fig .
[ fig : spectra](a ) .
the two gray bars mask the two strong phonon features , the g band around @xmath106 @xmath107 and the 2d band observed at @xmath108 @xmath107 for our @xmath109 nm excitation .
typically , the observed line widths for electronic features are @xmath110 @xmath107 . due to the pronounced difference in the scattered intensity of phonon excitations and of the different types of electronic excitations , we present our main results in fig .
[ fig : spectra](b ) in the form of a false color map of the scattered intensity as a function of the magnetic field which has been differentiated with respect to the magnetic field , with a step of @xmath111 mt .
the 2d band feature is weakly affected by the magnetic field , but because of a faraday effect affecting the intensity of the excitation laser when sweeping the magnetic field , it can not be cleanly removed by normalizing spectra acquired at a finite field by the @xmath105 response . thus , differentiation as a function of @xmath1 provides the cleanest color map .
the rich series of avoided crossings around @xmath106 @xmath107 is due to the well - studied magneto - phonon resonance in the raman scattering on the @xmath0-point optical phonon @xcite . in the following , instead of tracing the evolution of the phonon feature with @xmath1 , we study the evolution of the inter - ll electronic excitations .
this approach is today only possible for graphene on graphite which shows an exceptionally high quality @xcite , surpassing suspended graphene and graphene on bn @xcite in terms of line widths and of the variety of the observed excitations .
the numerous features in fig .
[ fig : spectra](b ) dispersing with the magnetic field can be attributed to inter - landau - level electronic excitations @xcite , and in the present experiment they can be observed up to energies as high as @xmath112 @xmath107 .
we observe @xmath20 and @xmath49 ( together with its electron - hole - symmetric counterpart @xmath50 ) with @xmath7 up to 6 .
their evolution with magnetic field can be reproduced using the value @xmath113 for the electronic velocity in graphene .
.the values of the magnetic fields , evaluated theoretically , and the corresponding resonances which can be observed in fig .
[ fig : spectra](b ) . [ cols="^,^,^",options="header " , ]
to conclude , we have presented a complete theoretical and experimental picture of magneto - phonon resonances in graphene , as observed in the magneto - raman scattering spectra . by considering the case of an electronic inter - landau - level excitation decaying into a phonon and another electronic excitation ,
we have extended the commonly accepted picture of magneto - phonon resonance in graphene ( i ) to both @xmath0 and @xmath11 point phonons and ( ii ) to multiphonon processes .
such multi - excitation final states have numerous possibilities to fulfill momentum conservation , including intervalley electronic excitations .
we have derived a simple classification of such magneto - phonon resonances and we have described the properties of each different type of resonances . in particular ,
resonant splitting of three crossing excitation branches is observed experimentally for the first time .
such multi - phonon processes are particularly important for the understanding of hot carrier relaxation in graphene .
we analyze the experimental data phenomenologically in terms of effective two- or three - level models , and then give a detailed discussion of validity of such effective description starting from a microscopic theory in the resonant approximation .
we also highlight the richness of the physics associated with the partially filled zero landau level and its low - energy electronic excitations .
when this level is involved in a resonance , the possibility to excite an arbitrary number of intra - level excitations with zero energy ( i ) leads to a genuine many - body problem which is quite hard to analyze theoretically , and ( ii ) results in complicated spectral shapes of the split peaks .
we thank m. o. goerbig and j .- n .
fuchs for fruitful discussions .
this work was supported by
let us first consider the case of coupling to @xmath26 point phonons , which results in eq .
( [ pinovertex= ] ) .
let us look for the eigenstates in the form @xmath114 is the ground state .
note that , generally speaking , @xmath115 as the phonons belong to different valleys and thus are effectively distinguishable in the two - phonon state @xmath116 .
eigenstates of the form ( [ psik= ] ) belong to the sector where the initial electron - hole pair @xmath117 is in the @xmath11 valley .
there are also eigenstates in the complementary sector , obtained by exchanging @xmath118 in the above expressions , with exactly the same energies , as the hamiltonian has the overall valley symmetry .
the optical excitation corresponds to the symmetric combination of states from the two sectors .
the schrdinger equation @xmath119 results in the following system of linear equations for the coefficients in eq .
( [ psik= ] ) : @xmath120^*\psi^{eh}+ \sum_{{\mathbf{q}}'}\tilde{j}^z_{nn'}({\mathbf{q}}')\ , e^{-i[{\mathbf{q}}'\times{\mathbf{q}}]_zl_b^2/2}\psi_{{\mathbf{q}},{\mathbf{q}}'},\\ & \sqrt{\frac{l_xl_y}{\lambda_k{v}^2}}\ , ( \omega - e_n - e_{n'}-\omega_k)\,\psi^h_{\mathbf{q}}=\nonumber\\ { } & { } = [ \tilde{j}^z_{nn'}({\mathbf{q}})]^*\psi^{eh } -\sum_{{\mathbf{q}}'}\tilde{j}^z_{-n',-n}({\mathbf{q}}')\ , e^{i[{\mathbf{q}}'\times{\mathbf{q}}]_zl_b^2/2}\psi_{{\mathbf{q}}',{\mathbf{q}}},\\ & \sqrt{\frac{l_xl_y}{\lambda_k{v}^2}}\ , ( \omega-2e_{n'}-2\omega_k)\psi_{{\mathbf{q}}_1,{\mathbf{q}}_2}=\nonumber\\ { } & { } = { } [ \tilde{j}^z_{nn'}({\mathbf{q}}_2)]^*e^{i[{\mathbf{q}}_2\times{\mathbf{q}}_1]_zl_b^2/2 } \psi^e_{{\mathbf{q}}_1}-{}\nonumber\\ & \quad{}-[\tilde{j}^z_{-n',-n}({\mathbf{q}}_1)]^ * e^{i[{\mathbf{q}}_2\times{\mathbf{q}}_1]_zl_b^2/2}\psi^h_{{\mathbf{q}}_2}.\end{aligned}\ ] ] let us look for the solution in the separable form @xmath121^*}{v_k}\,\psi^e,\\ & \psi^h_{\mathbf{q}}=\sqrt{\frac{\lambda_k{v}^2}{l_xl_y}}\ , \frac{[\tilde{j}^z_{nn'}({\mathbf{q}})]^*}{v_k}\,\psi^h,\\ & \psi_{{\mathbf{q}}_1,{\mathbf{q}}_2}=-\frac{\lambda_k{v}^2}{l_xl_y}\ , \frac{[\tilde{j}^z_{-n',-n}({\mathbf{q}}_1)]^ * [ \tilde{j}^z_{nn'}({\mathbf{q}}_2)]^*}{v_k^2}\times{}\nonumber\\ & \qquad\qquad{}\times { } e^{i[{\mathbf{q}}_2\times{\mathbf{q}}_1]_zl_b^2/2}\psi^{(2)}. \label{schrk4=}\end{aligned}\ ] ] then the normalization condition is @xmath122 where the effective hamiltonian matrix is given by @xmath123 the upper left matrix element of @xmath124 coincides with eq .
( [ pinovertex= ] ) , which proves their equivalence . for the coupling to @xmath0 point phonons
, we can look for the eigenstate in the form @xmath125 .
the basis states are defined as @xmath126 again , in the last line @xmath127 , and we focuse on the sector where the initial electron - hole pair @xmath117 is in the @xmath11 valley . note that @xmath128 and @xmath129 represent the same state , so the wave function in the two - phonon sector is assumed to be symmetric , @xmath130 , and the factor 1/2 is necessary in eq .
( [ psig= ] ) .
the schrdinger equation @xmath119 results in the following system of linear equations for the coefficients in eq .
( [ psig= ] ) : @xmath131^*\psi^{eh}+{}\nonumber\\ & \quad{}+ \sum_{{\mathbf{q}}',\sigma'}\tilde{j}^{\sigma'}_{nn'}({\mathbf{q}}')\ , e^{-i[{\mathbf{q}}'\times{\mathbf{q}}]_zl_b^2/2 } \psi_{{\mathbf{q}}',\sigma',{\mathbf{q}},\sigma},\label{schrge=}\\ & \sqrt{\frac{l_xl_y}{\lambda_\gamma{v}^2}}\ , ( \omega - e_n - e_{n'}-\omega_\gamma)\psi^h_{{\mathbf{q}},\sigma } = \nonumber\\{}&{}= [ \tilde{j}^\sigma_{nn'}({\mathbf{q}})]^*\psi^{eh}-{}\nonumber\\ & \quad{}-\sum_{{\mathbf{q}}',\sigma ' } \tilde{j}^{\sigma'}_{-n',-n}({\mathbf{q}}')\ , e^{i[{\mathbf{q}}'\times{\mathbf{q}}]_zl_b^2/2 } \psi_{{\mathbf{q}}',\sigma',{\mathbf{q}},\sigma},\label{schrgh=}\\ & \sqrt{\frac{l_xl_y}{\lambda_\gamma{v}^2}}\ , ( \omega-2e_{n'}-2\omega_\gamma ) \psi_{{\mathbf{q}}_1,\sigma_1,{\mathbf{q}}_2,\sigma_2 } = \nonumber\\{}&{}= [ \tilde{j}^{\sigma_1}_{nn'}({\mathbf{q}}_1)]^*\ , e^{i[{\mathbf{q}}_1\times{\mathbf{q}}_2]l_b^2/2}\psi^e_{{\mathbf{q}}_2,\sigma_2 } + \nonumber\\ & \quad{}+[\tilde{j}^{\sigma_2}_{nn'}({\mathbf{q}}_2)]^*\ , e^{i[{\mathbf{q}}_2\times{\mathbf{q}}_1]l_b^2/2}\psi^e_{{\mathbf{q}}_1,\sigma_1 } -\nonumber\\ & \quad{}-[\tilde{j}^{\sigma_1}_{-n',-n}({\mathbf{q}}_1)]^*\ , e^{i[{\mathbf{q}}_2\times{\mathbf{q}}_1]l_b^2/2}\psi^h_{{\mathbf{q}}_2,\sigma_2 } -\nonumber\\ & \quad{}-[\tilde{j}^{\sigma_2}_{-n',-n}({\mathbf{q}}_2)]^*\ , e^{i[{\mathbf{q}}_1\times{\mathbf{q}}_2]l_b^2/2}\psi^h_{{\mathbf{q}}_1,\sigma_1}. \label{schrg2=}\end{aligned}\ ] ] this system differs from eqs .
( [ schrk1=])([schrk4= ] ) by the presence of two extra terms in the last equation , due to the symmetry of the wave function in the two - phonon sector .
these terms correspond to the exchange of phonons between the electron and the hole , and because of them the whole system becomes non - separable .
in the dimensionless variables @xmath132 , @xmath133 , @xmath134 , eq . ( [ gselfcons= ] ) acquires the following form : @xmath135 the function @xmath136 is real unless @xmath137 .
it behaves as @xmath138 at @xmath139 . at @xmath140
it diverges as @xmath141 , the larger of the solutions of the equation @xmath142 . at @xmath143
, it has a square - root singularity , provided that @xmath144 . for @xmath15,it can be evaluated explicitly : @xmath145 $ ] . treating @xmath146 as a new unknown variable , we rewrite eq .
( [ 1g= ] ) as @xmath147 so that at any fixed @xmath7 , it contains only one parameter @xmath148 . numerical solution of eq .
( [ 1fnu= ] ) is facilitated by noting the following properties .
* at large positive @xmath149 , corresponding to large @xmath150 or large negative detuning @xmath151 , eq .
( [ 1fnu= ] ) has a real solution @xmath152 , corresponding to the unperturbed green s function @xmath153 .
another real solution is at @xmath154 . * as @xmath149 is decreased below some critical value @xmath155 , the two real solutions merge , and two complex solutions appear [ the sign of @xmath156 should be chosen to give the correct analytical properties of @xmath157 . * for @xmath149 large and negative ( i. e. , moderate @xmath150 and large positive @xmath151 ) , there is also a real solution @xmath158 , as well as another one close to @xmath159 .
the two solutions merge when @xmath160 .
it is easy to see that for @xmath161 real solution exist , so @xmath162 .
* in the region @xmath163 the solution is complex , which means that there are two disconnected intervals of @xmath164 where the spectral density is non - zero .
specifically , for @xmath165 we have @xmath166 , @xmath167 , the corresponding values of @xmath168 are @xmath169 and @xmath170 . for @xmath15 , @xmath171 and @xmath172
are the solutions of the equation @xmath173 , @xmath174 , @xmath175 ; the corresponding values of @xmath168 are @xmath176 which gives @xmath177 and @xmath178 .
the numerical smallness of @xmath179 gives rise to sharp but finite features in the spectral density .
for example , the shape of the spike in fig .
[ fig : spectrint](b ) is given by @xmath180 the typical width of the spike is @xmath181 , which appears very narrow on the scale of the figure .
37ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop _
( , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevlett.16.655 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1126/science.1137201 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.103.186803 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.105.227401 [ * * , ( ) ] link:\doibase 10.1143/jpsj.76.024712 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevb.85.125422 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.102.176804 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.103.136403 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.107.036807 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.195406 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1021/nl501578 m [ * * , ( ) ] link:\doibase 10.1103/physrevb.80.241404 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevb.81.155436 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.035433 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.93.185503 [ * * , ( ) ] link:\doibase 10.1103/physrevb.80.085423 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.108.247401 [ * * , ( ) ] link:\doibase 10.1103/physrevb.86.205431 [ * * , ( ) ] @noop ( ) link:\doibase 10.1103/physrevlett.114.126804 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevlett.98.166802 [ * * , ( ) ] link:\doibase 10.1103/physrevb.77.041409 [ * * , ( ) ] link:\doibase 10.1103/physrevb.78.081406 [ * * , ( ) ] @noop _ _ ( , ) @noop * * , ( ) | our low - temperature magneto - raman scattering measurements performed on graphene - like locations on the surface of bulk graphite reveal a new series of magneto - phonon resonances involving both k - point and @xmath0 point phonons . in particular , we observe for the first time the resonant splitting of three crossing excitation branches .
we give a detailed theoretical analysis of these new resonances .
our results highlight the role of combined excitations and the importance of multi - phonon processes ( from both k and @xmath0 points ) for the relaxation of hot carriers in graphene . |
several works ( e.g. , ratinov and roth , 2009 ; cohen and sarawagi , 2004 ) have shown that injecting dictionary matches as features in a sequence tagger results in significant gains in ner performance . however , building these dictionaries requires a huge amount of human effort and it is often difficult to get good coverage for many named entity types .
the problem is more severe when we consider named entity types such as gene , virus and disease , because of the large ( and growing ) number of names in use , the fact that the names are heavily abbreviated and multiple names are used to refer to the same entity @xcite . also , these dictionaries can only be built by domain experts , making the process very expensive .
this paper describes an approach for automatic construction of dictionaries for ner using large amounts of unlabeled data and a small number of seed examples .
our approach consists of two steps .
first , we collect a high recall , low precision list of _ candidate phrases _ from the large unlabeled data collection for every named entity type using simple rules . in the second step , we construct an accurate dictionary of named entities by removing the noisy candidates from the list obtained in the first step .
this is done by learning a classifier using the lower dimensional , real - valued cca @xcite embeddings of the candidate phrases as features and training it using a small number of labeled examples .
the classifier we use is a binary svm which predicts whether a candidate phrase is a named entity or not .
we compare our method to a widely used semi - supervised algorithm based on co - training @xcite .
the dictionaries are first evaluated on virus @xcite and disease @xcite ner by using them directly in dictionary based taggers .
we also give results comparing the dictionaries produced by the two semi - supervised approaches with dictionaries that are compiled manually .
the effectiveness of the dictionaries are also measured by injecting dictionary matches as features in a conditional random field ( crf ) based tagger .
the results indicate that our approach with minimal supervision produces dictionaries that are comparable to dictionaries compiled manually .
finally , we also compare the quality of the candidate phrase embeddings with word embeddings @xcite by adding them as features in a crf based sequence tagger .
we first give background on canonical correlation analysis ( cca ) , and then give background on crfs for the ner problem . the input to cca consists of @xmath0 paired observations @xmath1 where @xmath2 are the feature representations for the two views of a data point .
cca simultaneously learns projection matrices @xmath3 ( @xmath4 is a small number ) which are used to obtain the lower dimensional representations @xmath5 where @xmath6 .
@xmath7 are chosen to maximize the correlation between @xmath8 and @xmath9 .
consider the setting where we have a label for the data point along with it s two views and either view is sufficient to make accurate predictions . and give strong theoretical guarantees when the lower dimensional embeddings from cca are used for predicting the label of the data point .
this setting is similar to the one considered in co - training @xcite but there is no assumption of independence between the two views of the data point .
also , it is an exact algorithm unlike the algorithm given in . since we are using lower dimensional embeddings of the data point for prediction , we can learn a predictor with fewer labeled examples .
crf based sequence taggers have been used for a number of ner tasks ( e.g. , mccallum and li , 2003 ) and in particular for biomedical ner ( e.g. , mcdonald and pereira , 2005 ; burr settles , 2004 ) because they allow a great deal of flexibility in the features which can be included .
the input to a crf tagger is a sentence ( @xmath10 ) where @xmath11 are words in the sentence .
the output is a sequence of tags @xmath12 where @xmath13 .
@xmath14 is the tag given to the first word in a named entity , @xmath15 is the tag given to all words except the first word in a named entity and @xmath16 is the tag given to all other words .
we used the standard ner baseline features ( e.g. , dhillon et al .
, 2011 ; ratinov and roth , 2009 ) which include : * current word @xmath17 and its lexical features which include whether the word is capitalized and whether all the characters are capitalized .
prefix and suffixes of the word @xmath17 were also added . * word tokens in window of size two around the current word which include @xmath18 and also the capitalization pattern in the window . *
previous two predictions @xmath19 and @xmath20 .
the effectiveness of the dictionaries are evaluated by adding dictionary matches as features along with the baseline features @xcite in the crf tagger .
we also compared the quality of the candidate phrase embeddings with the word - level embeddings by adding them as features @xcite along with the baseline features in the crf tagger .
this section describes the two steps in our approach : obtaining candidate phrases and classifying them .
we used the full text of 110,369 biomedical publications in the biomed central corpus to get the high recall , low precision list of candidate phrases .
the advantages of using this huge collection of publications are obvious : almost all ( including rare ) named entities related to the biomedical domain will be mentioned and contains more recent developments than a structured resource like wikipedia .
the challenge however is that these publications are unstructured and hence it is a difficult task to construct accurate dictionaries using them with minimal supervision .
the list of virus candidate phrases were obtained by extracting phrases that occur between `` the '' and `` virus '' in the simple pattern `` the ... virus '' during a single pass over the unlabeled document collection .
this noisy list had a lot of virus names such as _ influenza _ , _ human immunodeficiency _ and _ epstein - barr _ along with phrases that are not virus names , like _ mutant _ , _ same _ , _ new _ , and so on .
a similar rule like `` the ... disease '' did not give a good coverage of disease names since it is not the common way of how diseases are mentioned in publications .
so we took a different approach to obtain the noisy list of disease names .
we collected every sentence in the unlabeled data collection that has the word `` disease '' in it and extracted noun phrases following the patterns `` diseases like .... '' , `` diseases such as .... '' , `` diseases including .... '' , `` diagnosed with .... '' , `` patients with .... '' and `` suffering from .... '' .
having found the list of candidate phrases , we now describe how noisy words are filtered out from them .
we gather ( _ spelling _ , _ context _ ) pairs for every instance of a candidate phrase in the unlabeled data collection . _
spelling _ refers to the candidate phrase itself while _ context _ includes three words each to the left and the right of the candidate phrase in the sentence .
the _ spelling _ and the _ context _ of the candidate phrase provide a natural split into two views which multi - view algorithms like co - training and cca can exploit .
the only supervision in our method is to provide a few _ spelling _ seed examples ( 10 in the case of virus , 18 in the case of disease ) , for example , _ human immunodeficiency _ is a virus and _ mutant _ is not a virus .
we use cca described in the previous section to obtain lower dimensional embeddings for the candidate phrases using the ( _ spelling _ , _ context _ ) views . unlike previous works such as and
, we use cca to learn embeddings for candidate phrases instead of all words in the vocabulary so that we do nt miss named entities which have two or more words .
let the number of ( _ spelling _ , _ context _ ) pairs be @xmath0 ( sum of total number of instances of every candidate phrase in the unlabeled data collection ) .
first , we map the _ spelling _ and _ context _ to high - dimensional feature vectors . for the _
spelling _ view , we define a feature for every candidate phrase and also a boolean feature which indicates whether the phrase is capitalized or not . for the _ context _ view
, we use features similar to where a feature for every word in the _ context _ in conjunction with its position is defined .
each of the @xmath0 ( _ spelling _ , _ context _ ) pairs are mapped to a pair of high - dimensional feature vectors to get @xmath0 paired observations @xmath1 with @xmath21 ( @xmath22 are the feature space dimensions of the _ spelling _ and _ context _ view respectively ) . using cca
, we learn the projection matrices @xmath3 ( @xmath23 and @xmath24 ) and obtain _ spelling _ view projections @xmath25 .
the k - dimensional _ spelling _ view projection of any instance of a candidate phrase is used as it s embedding .
the k - dimensional candidate phrase embeddings are used as features to learn a binary svm with the seed _ spelling _ examples given in figure 1 as training data .
the binary svm predicts whether a candidate phrase is a named entity or not . since the value of k is small , a small number of labeled examples are sufficient to train an accurate classifier
the learned svm is used to filter out the noisy phrases from the list of candidate phrases obtained in the previous step .
to summarize , our approach for classifying candidate phrases has the following steps : * input : @xmath0 ( _ spelling _ , _ context _ ) pairs , _ spelling _ seed examples .
* each of the @xmath0 ( _ spelling _ , _ context _ ) pairs are mapped to a pair of high - dimensional feature vectors to get @xmath0 paired observations @xmath1 with @xmath21 . * using cca , we learn the projection matrices @xmath3 and obtain _ spelling _ view projections @xmath25 . *
the embedding of a candidate phrase is given by the k - dimensional _ spelling _ view projection of any instance of the candidate phrase .
* we learn a binary svm with the candidate phrase embeddings as features and the _ spelling _ seed examples given in figure 1 as training data . using this svm
, we predict whether a candidate phrase is a named entity or not .
we discuss here briefly the dl - cotrain algorithm @xcite which is based on co - training @xcite , to classify candidate phrases .
we compare our approach using cca embeddings with this approach . here , two decision list of rules are learned simultaneously one using the _ spelling _ view and the other using the _ context _ view .
the rules using the _ spelling _ view are of the form : full - string = human immunodeficiency @xmath26 virus , full - string = mutant @xmath26 not_a_virus and so on . in the _ context _ view , we used bigram rules where we considered all possible bigrams using the _ context_. the rules are of two types : one which gives a positive label , for example , full - string = human immunodeficiency @xmath26 virus and the other which gives a negative label , for example , full - string = mutant @xmath26 not_a_virus .
the dl - cotrain algorithm is as follows : input : ( _ spelling _ , _ context _ ) pairs for every instance of a candidate phrase in the corpus , @xmath27 specifying the number of rules to be added in every iteration , precision threshold @xmath28 , _ spelling _ seed examples .
algorithm : 1 .
initialize the _ spelling _ decision list using the _ spelling _ seed examples given in figure 1 and set @xmath29 . 2 .
label the entire input collection using the learned decision list of _ spelling _ rules .
add @xmath30 new _ context _ rules of each type to the decision list of _ context _ rules using the current labeled data .
the rules are added using the same criterion as given in , i.e. , among the rules whose strength is greater than the precision threshold @xmath28 , the ones which are seen more often with the corresponding label in the input data collection are added .
label the entire input collection using the learned decision list of _ context _ rules .
add @xmath30 new spelling rules of each type to the decision list of _ spelling _ rules using the current labeled data .
the rules are added using the same criterion as in step 3 .
set @xmath31 .
if rules were added in the previous iteration , return to step 2 .
the algorithm is run until no new rules are left to be added .
the _ spelling _ decision list along with its strength @xcite is used to construct the dictionaries .
the phrases present in the _ spelling _ rules which give a positive label and whose strength is greater than the precision threshold , were added to the dictionary of named entities .
we found the parameters m and @xmath28 difficult to tune and they could significantly affect the performance of the algorithm .
we give more details regarding this in the experiments section .
previously , introduced a multi - view , semi - supervised algorithm based on co - training @xcite for collecting names of people , organizations and locations .
this algorithm makes a strong independence assumption about the data and employs many heuristics to greedily optimize an objective function .
this greedy approach also introduces new parameters that are often difficult to tune . in other works such as and external structured resources like wikipedia
have been used to construct dictionaries . even though these methods are fairly successful they suffer from a number of drawbacks especially in the biomedical domain .
the main drawback of these approaches is that it is very difficult to accurately disambiguate ambiguous entities especially when the entities are abbreviations @xcite .
for example , _ dm _ is the abbreviation for the disease _ diabetes mellitus _ and the disambiguation page for _ dm _ in wikipedia associates it to more than 50 categories since _ dm _ can be expanded to _ doctor of management _ , _ dichroic mirror _ , and so on , each of it belonging to a different category . due to
the rapid growth of wikipedia , the number of entities that have disambiguation pages is growing fast and it is increasingly difficult to retrieve the article we want . also , it is tough to understand these approaches from a theoretical standpoint . used cca to learn word embeddings and added them as features in a sequence tagger .
they show that cca learns better word embeddings than cw embeddings @xcite , hierarchical log - linear ( hlbl ) embeddings @xcite and embeddings learned from many other techniques for ner and chunking .
unlike pca , a widely used dimensionality reduction technique , cca is invariant to linear transformations of the data .
our approach is motivated by the theoretical result in which is developed in the co - training setting .
we directly use the cca embeddings to predict the label of a data point instead of using them as features in a sequence tagger . also , we learn cca embeddings for candidate phrases instead of all words in the vocabulary since named entities often contain more than one word .
learn a multi - class svm using the cca word embeddings to predict the pos tag of a word type .
we extend this technique to ner by learning a binary svm using the cca embeddings of a high recall , low precision list of candidate phrases to predict whether a candidate phrase is a named entity or not .
in this section , we give experimental results on virus and disease ner .
the noisy lists of both virus and disease names were obtained from the biomed central corpus .
this corpus was also used to get the collection of ( _ spelling _ , _ context _ ) pairs which are the input to the cca procedure and the dl - cotrain algorithm described in the previous section .
we obtained cca embeddings for the @xmath32 most frequently occurring word types in this collection along with every word type present in the training and development data of the virus and the disease ner dataset .
these word embeddings are similar to the ones described in and .
we used the virus annotations in the genia corpus @xcite for our experiments .
the dataset contains 18,546 annotated sentences .
we randomly selected 8,546 sentences for training and the remaining sentences were randomly split equally into development and testing sentences .
the training sentences are used only for experiments with the sequence taggers . previously , tested their hmm - based named entity recognizer on this data . for disease ner
, we used the recent disease corpus @xcite and used the same training , development and test data split given by them .
we used a sentence segmenter to get sentence segmented data and stanford tokenizer to tokenize the data .
similar to , all the different disease categories were flattened into one single category of disease mentions .
the development data was used to tune the hyperparameters and the methods were evaluated on the test data .
[ t ] [ cols="^,^,^,^,^,^,^ " , ] first , we compare the dictionaries compiled using different methods by using them directly in a dictionary - based tagger .
this is a simple and informative way to understand the quality of the dictionaries before using them in a crf - tagger .
since these taggers can be trained using a handful of training examples , we can use them to build ner systems even when there are no labeled sentences to train .
the input to a dictionary tagger is a list of named entities and a sentence .
if there is an exact match between a phrase in the input list to the words in the given sentence then it is tagged as a named entity .
all other words are labeled as non - entities .
we evaluated the performance of the following methods for building dictionaries : * * candidate list * : this dictionary contains all the candidate phrases that were obtained using the method described in section 3.1 .
the noisy list of virus candidates and disease candidates had 3,100 and 60,080 entries respectively . *
* manual * : manually constructed dictionaries , which requires a large amount of human effort , are employed for the task .
we used the list of virus names given in wikipedia .
unfortunately , abbreviations of virus names are not present in this list and we could not find any other more complete list of virus names .
hence , we constructed abbreviations by concatenating the first letters of all the strings in a virus name , for every virus name given in the wikipedia list . + for diseases , we used the list of disease names given in the unified medical language system ( umls ) metathesaurus .
this dictionary has been widely used in disease ner ( e.g. , dogan and lu , 2012 ; leaman et al . , 2010 ) + . * * co - training * : the dictionaries are constructed using the dl - cotrain algorithm described previously .
the parameters used were @xmath33 and @xmath34 as given in .
the phrases present in the _ spelling _ rules which give a positive label and whose strength is greater than the precision threshold , were added to the dictionary of named entities .
+ + in our experiment to construct a dictionary of virus names , the algorithm stopped after just 12 iterations and hence the dictionary had only 390 virus names .
this was because there were no _ spelling _ rules with strength greater than @xmath35 to be added .
we tried varying both the parameters but in all cases , the algorithm did not progress after a few iterations .
we adopted a simple heuristic to increase the coverage of virus names by using the strength of the _ spelling _ rules obtained after the @xmath36 iteration .
all _ spelling _ rules that give a positive label and which has a strength greater than @xmath37 were added to the decision list of _ spelling _ rules .
the phrases present in these rules are added to the dictionary .
we picked the @xmath37 parameter from the set [ 0.1 , 0.2 , 0.3 , 0.4 , 0.5 , 0.6 , 0.7 , 0.8 , 0.9 ] using the development data .
+ the co - training algorithm for constructing the dictionary of disease names ran for close to 50 iterations and hence we obtained better coverage for disease names .
we still used the same heuristic of adding more named entities using the strength of the rule since it performed better . * * cca * : using the cca embeddings of the candidate phrases as features we learned a binary svm to predict whether a candidate phrase is a named entity or not .
we considered using 10 to 30 dimensions of candidate phrase embeddings and the regularizer was picked from the set [ 0.0001 , 0.001 , 0.01 , 0.1 , 1 , 10 , 100 ] .
both the regularizer and the number of dimensions to be used were tuned using the development data .
table 1 gives the results of the dictionary based taggers using the different methods described above .
as expected , when the noisy list of candidate phrases are used as dictionaries the recall of the system is quite high but the precision is very low .
the low precision of the wikipedia virus lists was due to the heuristic used to obtain abbreviations which produced a few noisy abbreviations but this heuristic was crucial to get a high recall .
the list of disease names from umls gives a low recall because the list does not contain many disease abbreviations and composite disease mentions such as _ breast and ovarian cancer_. the presence of ambiguous abbreviations affected the accuracy of this dictionary .
the virus dictionary constructed using the cca embeddings was very accurate and the false positives were mainly due to ambiguous phrases , for example , in the phrase _ hiv replication _ , _ hiv _ which usually refers to the name of a virus is tagged as a rna molecule .
the accuracy of the disease dictionary produced using cca embeddings was mainly affected by noisy abbreviations .
we can see that the dictionaries obtained using cca embeddings perform better than the dictionaries obtained from co - training on both disease and virus ner even after improving the co - training algorithm s coverage using the heuristic described in this section .
it is important to note that the dictionaries constructed using the cca embeddings and a small number of labeled examples performs competitively with dictionaries that are entirely built by domain experts .
these results show that by using the cca based approach we can build ner systems that give reasonable performance even for difficult named entity types with almost no supervision .
we did two sets of experiments using a crf tagger .
in the first experiment , we add dictionary features to the crf tagger while in the second experiment we add the embeddings as features to the crf tagger .
the same baseline model is used in both the experiments whose features are described in section 2.2 .
for both the crf experiments the regularizers from the set [ 0.0001 , 0.001 , 0.01 , 0.1 , 1.0 , 10.0 ] were considered and it was tuned on the development set . here , we inject dictionary matches as features ( e.g. , ratinov and roth , 2009 ; cohen and sarawagi , 2004 ) in the crf tagger . given a dictionary of named entities , every word in the input sentence
has a dictionary feature associated with it .
when there is an exact match between a phrase in the dictionary with the words in the input sentence , the dictionary feature of the first word in the named entity is set to @xmath14 and the dictionary feature of the remaining words in the named entity is set to @xmath15 .
the dictionary feature of all the other words in the input sentence which are not part of any named entity in the dictionary is set to @xmath16 .
the effectiveness of the dictionaries constructed from various methods are compared by adding dictionary match features to the crf tagger .
these dictionary match features were added along with the baseline features .
figure 2 indicates that the dictionary features in general are helpful to the crf model .
we can see that the dictionaries produced from our approach using cca are much more helpful than the dictionaries produced from co - training especially when there are fewer labeled sentences to train .
similar to the dictionary tagger experiments discussed previously , the dictionaries produced from our approach performs competitively with dictionaries that are entirely built by domain experts .
the quality of the candidate phrase embeddings are compared with word embeddings by adding the embeddings as features in the crf tagger . along with the baseline features , * cca - word * model adds word embeddings as features while the * cca - phrase * model adds candidate phrase embeddings as features . *
cca - word * model is similar to the one used in .
we considered adding 10 , 20 , 30 , 40 and 50 dimensional word embeddings as features for every training data size and the best performing model on the development data was picked for the experiments on the test data . for candidate phrase embeddings we used the same number of dimensions that was used for training the svms to construct the best dictionary .
when candidate phrase embeddings are obtained using cca , we do not have embeddings for words which are not in the list of candidate phrases .
also , a candidate phrase having more than one word has a joint representation , i.e. , the phrase `` human immunodeficiency '' has a lower dimensional representation while the words `` human '' and `` immunodeficiency '' do not have their own lower dimensional representations ( assuming they are not part of the candidate list ) .
to overcome this issue , we used a simple technique to differentiate between candidate phrases and the rest of the words .
let @xmath38 be the highest real valued candidate phrase embedding and the candidate phrase embedding be a @xmath39 dimensional real valued vector .
if a candidate phrase occurs in a sentence , the embeddings of that candidate phrase are added as features to the first word of that candidate phrase .
if the candidate phrase has more than one word , the other words in the candidate phrase are given an embedding of dimension @xmath39 with each dimension having the value @xmath40 .
all the other words are given an embedding of dimension @xmath39 with each dimension having the value @xmath41 .
figure 3 shows that almost always the candidate phrase embeddings help the crf model .
it is also interesting to note that sometimes the word - level embeddings have an adverse affect on the performance of the crf model .
the * cca - phrase * model performs significantly better than the other two models when there are fewer labeled sentences to train and the separation of the candidate phrases from the other words seems to have helped the crf model .
we described an approach for automatic construction of dictionaries for ner using minimal supervision . compared to the previous approaches ,
our method is free from overly - stringent assumptions about the data , uses svd that can be solved exactly and achieves better empirical performance .
our approach which uses a small number of seed examples performs competitively with dictionaries that are compiled manually .
we are grateful to alexander rush , alexandre passos and the anonymous reviewers for their useful feedback .
this work was supported by the intelligence advanced research projects activity ( iarpa ) via department of interior national business center ( doi / nbc ) contract number d11pc20153 . the u.s .
government is authorized to reproduce and distribute reprints for governmental purposes notwithstanding any copyright annotation thereon .
the views and conclusions contained herein are those of the authors and should not be interpreted as necessarily representing the official policies or endorsements , either expressed or implied , of iarpa , doi / nbc , or the u.s . government . | this paper describes an approach for automatic construction of dictionaries for named entity recognition ( ner ) using large amounts of unlabeled data and a few seed examples .
we use canonical correlation analysis ( cca ) to obtain lower dimensional embeddings ( representations ) for _ candidate phrases _ and classify these phrases using a small number of labeled examples .
our method achieves 16.5% and 11.3% f-1 score improvement over co - training on disease and virus ner respectively .
we also show that by adding candidate phrase embeddings as features in a sequence tagger gives better performance compared to using word embeddings . |
relativistic heavy ion collisions aim to create the quark - gluon plasma ( qgp ) and study its properties at the extreme conditions of high temperature and energy density @xcite . the collective flow as a soft probe is often used to study the qgp properties in experimental and theoretical investigations .
both hydrodynamics @xciteand transport theory @xcitecan describe the bulk data in heavy ion collisions .
for example , the string melting version of a multi - phase transport ( ampt ) model @xcite reasonably reproduces particle yields , @xmath1 spectra , and @xmath2 of low-@xmath1 pions and kaons in central and mid - central au+au collisions at 200a gev and pb+pb collisions at 2760a gev @xcite .
the small system data can be also satisfactorily described by ampt @xcite .
recent studies have shown that light parton @xmath0 is mainly generated by the anisotropic parton escape from the collision zone and hydrodynamics may play only a minor role @xcite .
it suggests that the mass ordering of @xmath0 , commonly believed as a signature of hydrodynamics , may arise from other mechanisms . in these proceedings ,
we first investigate the physic mechanisms of mass ordering in ampt @xcite .
we then study the production mechanism of the charm @xmath0 in light of the escape mechanism for the light quark @xmath0 .
we employ the string melting version of ampt @xcite in our study .
the model consists of four components : fluctuating initial conditions , parton elastic scatterings , quark coalescence for hadronization , and hadronic interactions .
the partons interaction are modeled by zhang s parton cascade ( zpc ) @xcite .
we use debye screened differential cross - section @xmath3 @xcite , with strong coupling constant @xmath4 and debye screening mass @xmath5/fm ( the total cross section is then @xmath6 mb ) for all ampt simulation in our work .
after partons stop interaction , a simple quark coalescence model is applied to combine two nearest partons into a meson and three nearest partons into a baryon(or antibaryon ) .
the hadronic interactions are described by the art model @xcite .
we terminate the hadronic interactions at a cutoff time .
we simulate three collision systems : au+au collisions with @xmath7 - 8.1 fm at the nucleon - nucleon center - of - mass energy @xmath8 gev , @xmath9+au collisions with @xmath10 fm at @xmath8 gev , and @xmath11+pb collisions with @xmath10 fm at @xmath12 tev by ampt .
we analyze the momentum - space azimuthal anisotropy of partons in the final state before hadronization , of hadrons right after hadronization before hadronic rescatterings take place , and of freeze - out hadrons in the final state .
the momentum anisotropy is characterized by fourier coefficients @xcite @xmath13 where @xmath14 is the azimuthal angle of the particle ( parton or hadron ) momentum .
the @xmath15 is the harmonic plane of each event from its initial configuration of all partons .
all results shown are for particles within pseudo - rapidity window @xmath16 .
) @xmath2 of @xmath17 , k , p(@xmath18 ) and charged hadrons ( @xmath19 ) at @xmath20 gev/@xmath21 at different stages of system evolution in @xmath7 - 8.1 fm au+au ( left ) and @xmath10 fm @xmath9+au ( right ) collisions at @xmath22=200 gev by ampt : right after coalescence hadronization and before hadronic rescattering ( initial @xmath2 of primordial hadrons ) , hadron initial @xmath2 including decays , and after hadronic rescattering at freezeout ( hadron final @xmath2 ) .
[ fig : fig3_s ] ] considering that observed pions and protons are made of only light constituent quarks , the difference between their @xmath2 should come from the hadronization process and/or hadronic rescattering .
we first study the primordial hadrons right after hadronization , before any decay and hadronic rescatterings take place .
figure [ fig : fig1](a ) shows the @xmath2 of primordial charged pion , kaon , and ( anti)proton as a function of @xmath1 ( solid curves ) .
it demonstrates that the mass ordering of @xmath0 at low @xmath1 in ampt comes from the dynamic in coalescence .
the dynamical `` selections '' of constituent quarks into pions , kaons , and protons are manifest in the constituent quark @xmath2 distributions shown by the dashed curves in fig .
[ fig : fig1](a ) , plotted at the respective _ hadron _ @xmath1 . figure [ fig : fig1](b ) shows the @xmath2 results by @xmath14-randomized ampt @xcite for primordial hadrons right after coalescence hadronization at the corresponding constituent quark @xmath2 s .
no hydrodynamic anisotropic flow is present in the @xmath14-randomized case @xcite , however , mass splitting is still present .
it implies that the mass splitting is mainly due to kinematics in the coalescence process .
it is therefore not a unique signature of collective anisotropic flow or hydrodynamics .
after hadronization , particles undergo decay and rescattering .
we evaluate the @xmath2 of hadrons before hadronic rescattering but including effects of resonance decays by setting the cut - off time to be 0.6 fm / c .
the results are shown in fig .
[ fig : fig2](a ) by the dashed curves .
the decay product @xmath2 is usually smaller than their parent @xmath2 @xcite . by including decay products ,
the hadron @xmath2 is reduced from that of primordial hadrons ( solid curves in fig .
[ fig : fig1](a ) ) .
the @xmath2 values before hadronic rescattering ( including resonance decay effects ) can be considered as the initial @xmath2 for the hadronic evolution stage .
the final - stage freezeout hadron @xmath2 s ( also including decay daughters ) are shown in fig .
[ fig : fig2](a ) by the solid curves .
pion @xmath2 increases while proton @xmath2 decreases from before to after hadronic rescattering .
this may be due to interactions between pions and protons , after which they tend to flow together at the same velocity .
thus , the same - velocity pions and protons ( i.e. small @xmath1 pions and large @xmath1 protons ) will tend to have the same anisotropy .
it will yield lower @xmath2 for the protons and higher @xmath2 for the pions at the same @xmath1 value .
this should happen whether or not there is a net gain in the overall charged hadron @xmath2 , which depends on the initial configuration geometry from which the hadronic evolution begins @xcite .
the similar results in @xmath9+au collisions as shown in fig .
[ fig : fig2](b ) @xcite . to summarize the origin of @xmath2 mass splitting , we plot the @xmath2 of pions , kaons , and protons within @xmath20 gev/@xmath21 , a typical region for different stages of the collision system evolution as shown in fig . [
fig : fig3_s ] .
the evolution stages contain : ( i ) right after coalescence hadronization including only primordial hadrons ( labeled `` prim . '' ) ; ( ii ) right after coalescence hadronization but including decay products ( labeled `` w/ decay '' ) ; and ( iii ) at final freezeout ( labeled w/ rescatt .
w/ decay ) .
we can see the mass ordering actually comes from interplay of several physics effects ; it comes from coalescence , and more importantly , from hadronic rescattering process .
in this section we discuss the production mechanism of charm quark flow in ppb .
we compare the light and charm quark freeze out @xmath2 , integrated over all @xmath1 at the same proper time .
this is shown in the left panel of fig .
[ fig : fig4_c ] .
the charm @xmath0 is systematically lower than light quark @xmath0 .
this may suggest that the charm quark is not as thermalized as light quarks .
the right panel of fig .
[ fig : fig4_c ] shows @xmath14-randomized results .
the charm and light quark @xmath2 are similar suggesting a common escape mechanism .
in these proceedings , we investigate the origin of mass splitting of identified hadron @xmath0 .
the mass splitting is due to coalescence and , more importantly , hadronic rescattering .
it is therefore not a unique signature of hydrodynamics .
we also study the development of charm @xmath0 and found the escape mechanism to be the major contribution in ppb similar to light quarks .
this work is supported in part by us department of energy grant no .
de - fg02 - 88er40412 ( lh , fw , wx ) and no . de - fg02 - 13er16413 ( dm ) .
hl acknowledges financial support from the china scholarship council . | a multi - phase transport ( ampt ) model has been shown to describe experimental data well , such as the bulk properties of particle spectra and elliptic anisotropy ( @xmath0 ) in heavy ion collisions .
recent studies have shown that ampt describes the @xmath0 data in small system collisions as well . in these proceedings , we first investigate the origin of the mass ordering of identified hadrons @xmath0 in heavy ion as well as small system collisions .
we then study the production mechanism of the charm @xmath0 in light of the escape mechanism for the light quark @xmath0 . |
for the first time in the history of cosmology , the variety and quality of current cosmological data provide the possibility of testing some fundamental hypotheses of the standard cosmological model .
one of these hypotheses is the validity of the so - called cosmic distance - duality relation ( cddr ) , which is derived from etherington reciprocity theorem ( etherington , 1933 ) .
the cddr , expressed as @xmath12 where @xmath13 is the angular diameter distance and @xmath14 is the luminosity distance to a given source at redshift @xmath3 , holds if photons follow null ( unique ) geodesic and the geodesic deviation equation is valid , along with the assumption that the number of photons is conserved over the cosmic evolution ( bassett & kunz , 2004 ) .
the cddr plays an essential role in observational cosmology and any departure from it could point to exciting new fundamental physics or unaccounted systematic errors ( for a general discussion , see ellis , 1971 ) . in recent years , several authors have proposed methods to test the cddr . generally speaking , we can roughly divide them in two classes , cosmological model - dependent tests , usually performed within the cosmic concordance @xmath15cdm model , and model - independent ones .
for instance , uzan et al .
( 2005 ) showed that the combination of the sunyaev - zeldovich effect ( sze ) plus x - ray observations , used to measure angular diameter distance to galaxy clusters , is dependent on the cddr validity . by using 18 measurements of @xmath13 to galaxy clusters described by a spherical model and the @xmath15cdm concordance model these authors showed
that the cddr validity was only marginally verified .
( 2010 ) and meng _ et al . _
( 2012 ) , also using sze / x - ray observations of galaxy clusters , showed that triaxial ellipsoidal model is a better geometrical hypothesis describing the structure of the galaxy cluster compared with the spherical model if the cddr is valid in cosmological observations .
( 2010 ; 2012 ) adopted a modified expression of the cddr , @xmath16 , to constrain , in the context of a flat @xmath15cdm model , the cosmic opacity from type ia supernova ( sne ia ) and @xmath17 data .
they found @xmath18 ( @xmath19 ) . in this line ,
holanda & busti ( 2014 ) probed the cosmic opacity at high redshifts with gamma - ray bursts and @xmath17 data and showed that this combination of data is compatible with a transparent universe ( @xmath20 ) at @xmath21 level , regardless of the dark energy equation - of - state parameter @xmath22 assumed in the analysis .
on the other hand , several cosmological model - independent tests for the cddr have also been proposed .
de bernardis , giusarma & melchiorri ( 2006 ) , holanda _ et al . _
( 2011 ) and li , wu & yu 2011 confronted @xmath13 measurements to galaxy clusters with sne ia data obtaining that the cddr is verified at 1@xmath10 when the galaxy clusters are described by ellipsoidal model . by showing that the x - ray gas mass fraction ( @xmath23 ) of galaxy clusters also depends on the cddr validity , gonalves
( 2012 ) proposed a test involving samples of gas mass fractions and sne ia observations .
another test using exclusively gas mass fractions was proposed by holanda _
these authors showed that the relation between @xmath24 and @xmath25 observations is given by @xmath26 , where @xmath27 quantifies departures of the cddr validity .
no violation was found .
it was shown no violation of the cddr .
santos - da - costa _ et al . _
( 2015 ) , applying gaussian process , proposed a test based on galaxy clusters observations and @xmath17 measurements and no evidence of deviation of the cddr validity was found . by using measurements of the cosmic microwave background black - body spectrum , ellis _
( 2013 ) showed that the reciprocity relation can not be violated by more than @xmath28 between decoupling and today .
recently , liao _ et al . _
( 2015 ) introduced a new method to test the cddr based on strong gravitational lensing systems ( cao et al .
2015 ) and the most recent sne ia compilation ( betoule et al . 2014 ) .
although most of the results so far have been consistent with the validity of the cddr , new methods with different astronomical observations and redshift range offer a path to validate the whole cosmological framework as well as to detect unexpected behaviour or systematic errors .
therefore , in this paper , we propose a new test which uses the strong gravitational lensing effect and luminosity distances from type ia supernovae .
the paper is organised as follows . in section 2
we describe our new method to test the cddr , while in section 3 the observational quantities used in this work are discussed .
the corresponding constraints on the departures of the cddr are investigated and discussed in section 4 .
we end this paper by summarizing our main results in section 5 .
strong gravitational lensing ( sgl ) is a powerful astrophysical effect to probing both gravity theories and cosmology .
it occurs whenever the source ( @xmath29 ) , the lens ( @xmath30 ) and the observer ( @xmath31 ) are so well aligned that the observer - source direction lies inside the so - called einstein radius of the lens ( for a complete discussion on this effect , see schneider , ehlers & falco , 1992 ) . in a cosmological context
the source is usually a galaxy or quasar with a galaxy acting as the lens .
recent papers have explored strong gravitational lensing systems for measuring cosmological parameters ( biesiada , malec , & piorkowska 2011 ; yuan & wang 2015 , cao et al .
2015 ) .
the einstein radius @xmath32 , under assumption of the singular isothermal sphere ( sis ) model for the lens , is given by @xmath33 where @xmath34 is the speed of light , @xmath35 and @xmath36 are the angular diameter distances between lens - source and source - observer , respectively , and @xmath37 is the velocity dispersion due to lens mass distribution.@xmath37 is not to be exactly equal to the observed stellar velocity dispersion @xmath38 , there is a strong indication that dark matter halos are dynamically hotter than the luminous stars based on x - ray observations . in order to account for this difference
it is introduced a phenomenological free parameter @xmath39 , defined by the relation : @xmath40 , where @xmath41 ( ofek et al .
2003 ) .
in our cddr test we are interested in the ratio between angular distances : @xmath42 if one assumes the cddr validity it is possible to transform the above expression into a luminosity distance ratio . for the angular diameter distance of the source @xmath29
we have @xmath43 whereas the luminosity and angular distances between lens and source are linked by ( bartelmann & schneider , 2001 ) @xmath44 from the above expressions , one may show that @xmath45 where @xmath46 therefore , obtaining @xmath47 from observations , the expression ( [ razao ] ) can be modified to investigate the cddr validity . in order to obtain @xmath47
it is necessary to know @xmath48 at the same redshift of source in the lensing system and @xmath49 , the luminosity distance between lens and source . in our work @xmath50
is obtained as follows : we first estimate the luminosity distance @xmath51 for each sne ia of the sample by using their distance modulus measurements .
the relation between the distance modulus @xmath52 and the luminosity distance is : @xmath53 where @xmath54 is the absolute magnitude of the source and @xmath55 is the apparent magnitude ( @xmath56 is for the @xmath56-band ) .
it is important to stress that the distance moduli were obtained fitting some sne ia light - curve parameters in a flat @xmath15cdm model , which makes them model dependent .
however , this dependence is much smaller than the errors coming from the gravitational lensing modelling , so it does not affect our results strongly .
applying the transformation from distance modulus to luminosity distance , we perform a polynomial fit to @xmath51 of the sne i data ( see fig .
( 1a ) and the discussion in the next section ) .
thus , we can calculate @xmath50 at each point of interest ( source in the lensing system ) .
previous papers used pairs of sne ia with some other astronomical object whose the redshifts difference between them were smaller than @xmath57 .
the advantage of using polynomial fit on the sne ia data is to avoid any redshift difference as well as to minimize the bias from the objects in the same redshift but in different direction on the sky . the distance @xmath49 , the luminosity distance between lens and source , is obtained by using two cosmological models : flat @xmath15cdm concordance model from the planck collaboration ( ade et al .
2015 ) and flat @xmath58cdm from wmap9 satellite results ( hinshaw et al .
the @xmath49 expressions for these models are @xmath59 where @xmath60 for @xmath15cdm model is @xmath61 and for @xmath58cdm is @xmath62 where the dark energy equation - of - state parameter is assumed to be given by @xmath63 . combining temperature and lensing data
, the planck collaboration found , for a flat @xmath15cdm universe , @xmath64 and @xmath65 @xmath66 .
for a flat @xmath58cdm model , we use the following values to parameters at 1@xmath10 obtained from wmap 9 years ( hinshaw et al .
2013 ) : @xmath67 , @xmath68 , @xmath69 and @xmath70 . in fig .
( 1b ) we plot @xmath71 and @xmath47 using the gravitational lensing and sne ia data . for simplicity
, we only show these quantities for the @xmath15cdm scenario .
recently , the supernova cosmology project reported the discovery of the most distant sne ia ( rubinet al . 2013 ) : the sne ia scp-0401 at @xmath72 and distance modulus of @xmath73 ( statistical errors ) . in our analysis we added this sne ia to union 2.1 sample ( suzuki et al . 2012 ) , comprising 581 data points .
the curves shown in fig .
1a stand for the polynomial fit of @xmath74 points from sne ia data and the corresponding 1@xmath10 error .
the open circle corresponds to the most distant ( @xmath75 ) spectroscopically confirmed sne ia .
we also use 95 data points from 118 sgl systems from sloan lens acs survey ( slacs ) , boss emission - line lens survey ( bells ) , lenses structure and dynamics survey ( lsd ) and strong legacy survey sl2s surveys ( cao et al .
we discarded systems with sources redshifts higher than @xmath76 .
it is important to stress that cao et al .
( 2015 ) assumed a spherically symmetric mass distribution in lensing galaxies , but relaxed the rigid assumption of sis model in favor of more general power - law index @xmath77 ( plaw ) , @xmath78 , where the distribution becomes a sis for @xmath79 . under this assumption
the einstein radius is @xmath80 where @xmath81 is the stellar velocity dispersion inside the aperture of size @xmath82 and @xmath83 ^ 2.\end{aligned}\ ] ] therefore , @xmath84 the sample used here is compiled and summarized in table 1 of cao et al .
( 2015 ) , in which all relevant information necessary to obtain @xmath71 from eqs .
( 3 ) and ( 13 ) to perform our fit can be found . also following these authors , we replace @xmath81 by @xmath38 in eq .
( [ newobservable ] ) .
this procedure makes @xmath71 more homogeneous for the sample of lenses located at different redshifts . [ cols="^,^,^,^,^,^",options="header " , ]
in this work we assume the following general expressions : @xmath85 ( linear parametrization ) and @xmath86 ( non - linear parametrization ) , which avoids divergences at high-@xmath3 .
the constraints to the @xmath87 parameter are derived by evaluating the likelihood distribution function , @xmath88 , with @xmath89 ^ 2}{\sigma^2 } \end{aligned}\ ] ] where @xmath90 is the observed @xmath91 ratio and @xmath92 stands for the statistical errors associated to sne ia , gravitational lensing obtained using standard propagation errors techniques . for the gravitational lensing error
one may show that @xmath93 in order to explore the influence of the model used to describe the lens we perform our analysis by using the eqs .
( 3 ) and ( 13 ) where the following flat priors on @xmath39 and @xmath94 were used : @xmath95 and @xmath96 .
the results of our statistical analysis are shown in figs .
( 2 ) and ( 3 ) . in figs . ( 2a ) and ( 2b ) we plot the results for the linear parametrization .
the parameter @xmath87 describing possible departures from the cddr is constrained in the regions : @xmath97 and @xmath98 for @xmath15cdm model using sis and plaw sgl models , respectively . for @xmath58cdm model
we obtain : @xmath99 and @xmath100 using the sis and plaw models , respectively . in figs .
( 3a ) and ( 3b ) we plot the results for the non - linear parametrization .
the parameter @xmath87 is constrained in the regions : @xmath101 and @xmath102 for @xmath15cdm model using sis and plaw sgl models , respectively . for @xmath58cdm model
we obtain : @xmath103 and @xmath104 using the sis and plaw models , respectively . as on
may see , the validity of the cddr is fully compatible with the data regardless the @xmath105 parametrization or cosmological model adopted in the analysis .
finally , in table i we show an extensive list of current bounds on a possible departure from the cddr obtained from different techniques .
it is worth observing that the validity of the cddr is verified at 2@xmath10 in all cases .
although the uncertainties in our analysis are larger than previous works , given the large scatter of distance measurements in the lensing observations , the advantage of using these systems are the few astrophysical assumptions and systematic uncertainties .
the cosmic distance - duality relation ( cddr ) , which relates in a simple way the luminosity and the angular diameter distances at a given redshift @xmath3 , plays an essential role in observational cosmology .
its validity has a direct impact on several cosmological probes , from gravitational lensing studies to the spectrum of the cosmic microwave background and galaxy clusters observations .
any source of attenuation , such as gray intergalactic dust or exotic photon interaction , contributes to violate the cddr since its proof is based on the conservation of the average number of photons . on the other hand , deviations from the cddr
can also indicate that light does not propagate on null geodesics or that a metric theory for gravitation is inadequate .
therefore , given the importance of the underlying hypotheses which hold the cddr , tests of the cddr are an important task for cosmology . in this paper we have discussed a new test of the cddr in terms of the ratio of angular diameter ( @xmath106 ) and luminosity ( @xmath5 ) distances , where the subscripts 1 and 2 correspond to redshift @xmath6 and @xmath7 . by allowing departures of the cddr we have proposed @xmath107 , where @xmath108 , and placed constraints on @xmath87 by using strong gravitational lensing and type ia supernovae data .
more specifically , we used 95 @xmath71 measurements from various gravitational lens and @xmath47 from 581 sne ia ( union 2.1 sample plus the most distant sne ia to date , i.e. , the scp-040 at @xmath72 ) .
we have explored the influence of the model used to describe the lens by performing the fits under the assumption of the singular isothermal sphere and singular isothermal sphere with the general power - law index @xmath94 .
we have obtained no violation of the cddr .
the results derived here show an excellent agreement with those originated from very different astrophysical environments ( see table i ) and reinforce the theoretical pillars of the cddr .
rflh acknowledges financial support from inct - a and cnpq ( no . 478524/2013 - 7 ) .
jsa is supported by cnpq ( no . 310790/2014 - 0 ; 400471/2014 - 0 ) , faperj and inespao .
vcb is supported by so paulo research foundation ( fapesp)/capes agreement , under grant number 2014/21098 - 1 .
99 ade , p. a. r. , et .
al . , 2015 , planck collaboration , arxiv:1502.01589 avgoustidis , a. , burrage , c. , redondo , j. , verde , l. & jimenez , r. , 2010 , jcap , 10 , 024 avgoustidis , a. , luzzi , g. , martins , c. j. a. p. & monteiro , a. m. r. v. l. , 2012 , jcap , 2 , 013 bartelmann , m. & schneider , p. , 2001
, 340 , 291 b. a. bassett and m. kunz , phys .
d * 69 * , 101305 ( 2004 ) cao , s. , biesiada , m. , gavazzi , r. , piorkowska , a. & zhu , z. u. , 2015 , apj , 806 , 185 betoule , m. et .
al . , 2014 ,
a&a , 568 , a22 biesiada , m. , piorkowska , a. , & malec , b. 2010 , mnras , 406 , 1055 biesiada , m. , malec , b. , & piorkowska , a. 2011 , raa , 11 , 641 cao , s. , pan , y. , biesiada , m. , godlowski , w. , zhu , z. , 2012 , jcap , 03 , 016 chen , z. , zhou , b. & fu , x. , 2015 , ijtp , 54 , 1 de bernardis , f. , giusarma , e. , & melchiorri , a. 2006 , ijmpd , 15 , 759 ellis , g.f.r . , poltis , r. , uzan , j .-
weltman , a. , 2013 , phys . rev .
d , 87 , 103530 gonalves , r. s. , holanda , r. f. l. , & alcaniz , j. s. , 2012 , mnras , 420 , l43 hinshaw , g. et al .
, 2013 , apjs , 208 , 19h holanda , r. f. l. , lima , j. a. s. & ribeiro m. b. , 2012b , a&a , 538 , 131 holanda , r. f. l. , lima , j. a. s. & ribeiro , m. b. , 2010 , apj , 722 , l233 holanda , r. f. l. & busti , v. c. , 2014 , prd , 89 , 103517 holanda , r. f. l. , gonalves , r. s. & alcaniz , j. s. , 2012a , jcap , 06 , 022 holanda , r. f. l. , carvalho , j. c. & alcaniz , j. s. , 2013 , 04 , 027 jhingan , s. , jain , d. , nair , r. , 2014 , journal of physics : conference series , 484 , 1 lazkoz , r. , nesseris , s. & perivolaropoulos , l. , 2008 , jcap , 0807 , 012 liao , et al . , 2015 , arxiv:1511.01318 li , z. , wu , p. & yu , h. , 2011 , apjl , 729 , l14 meng , x .-
l . , zhang , t .- j .
, & zhan , h. , 2012 , apj , 745 , 98 nair , r. jhingan , s. & jain , d. , 2011 , jcap , 05 , 023 newton , e. , et al . 2011 , accepted by apj , arxiv:1104.2608 ofek , e. o. , rix , h .- w . , & maoz , d. 2003 , mnras , 343 , 639 rubin , d. et al . , 2013 , apj , 35 , 763 santos - da - costa , s. , busti , v. c. & holanda , rodrigo f. l. , 2015 , jcap , 10 , 061 suzuki , n. et al . , 2012 , apj , 85 , 746 uzan , j.p . ,
aghanim , n. and mellier , y. , 2004 , phys .
d , 70 , 083533 wu , p. , li , z. , liu , x. & yu , h. , 2015 , prd , 92 , 023520 yang , x. et al .
, 2013 , jcap , 06 , 007 yuan , c. c. ; wang , f. y. , 2015 , mnras , 452 , 2423 uzan , j.p . ,
aghanim , n. & mellier , y. 2004 , prd , 70 , 083533 | we propose and perform a new test of the cosmic distance - duality relation ( cddr ) , @xmath0 , where @xmath1 is the angular diameter distance and @xmath2 is the luminosity distance to a given source at redshift @xmath3 , using strong gravitational lensing ( sgl ) and type ia supernovae ( sne ia ) data .
we show that the ratio @xmath4 and @xmath5 , where the subscripts 1 and 2 correspond , respectively , to redshifts @xmath6 and @xmath7 , are linked by @xmath8 if the cddr is valid .
we allow departures from the cddr by defining two functions for @xmath9 , which equals unity when the cddr is valid .
we find that combination of sgl and sne ia data favours no violation of the cddr at 1@xmath10 confidence level ( @xmath11 ) , in complete agreement with other tests and reinforcing the theoretical pillars of the cddr . |
conventionally , wireless communication nodes operate in half duplex ( hd ) mode under which they transmit and receive signals over orthogonal frequency or time resources
. recent advances , nevertheless , suggest that full duplex ( fd ) communications that allows simultaneous transmission and reception of signal over the same radio channel be possible @xcite .
in addition to the immediate benefit of essentially doubling the bandwidth , full duplex communications also find applications in simultaneous wireless information and power transfer ( swipt ) .
much interest has turned to full - duplex relaying in which information is sent from a source node to a destination node through an intermediate relaying node which is powered by means of wireless energy harvesting . in the literature , the studies on relay aided swipt largely considered hd relaying and adopted a time - switched relaying ( tsr ) approach @xcite@xcite .
authors in @xcite considered swipt in miso multicasting systems , in @xcite considered swipt in miso broadcasting systems , and in @xcite miso secrecy systems , where the joint transmit beamforming and receive power splitting problem for minimising the transmit power of the base station ( bs ) subject to signal - to - noise ratio ( snr ) and energy harvesting constraints at the receiver was investigated .
in contrast to the existing results , this paper studies the joint optimization of the two - way beamforming matrix for swipt in a multiple - input multiple - output ( mimo ) amplify - and - forward ( af ) full - duplex relay system employing a power splitter ( ps ) , where the sum rate is maximized subject to the energy harvesting and total power constraints . _
notations_we use @xmath0 to represent a complex matrix with dimension of @xmath1 . also , we use @xmath2 to denote the conjugate transpose , while @xmath3 is the trace operation , and @xmath4 denotes the frobenius norm .
in addition , @xmath5 returns the absolute value of a scalar , and @xmath6 denotes that the hermitian matrix @xmath7 is positive semidefinite .
the expectation operator is denoted by @xmath8 we define @xmath9 as the orthogonal projection onto the column space of @xmath10 ; and @xmath11 as the orthogonal projection onto the orthogonal complement of the column space of @xmath12
let us consider swipt in a three - node mimo relay network consisting of two sources @xmath13 and @xmath14 wanting to exchange information with the aid of an af relay @xmath15 , as shown in fig .
[ alex ] . in our model
, all the nodes are assumed to operate in fd mode , and we also assume that there is no direct link between @xmath13 and @xmath14 so communication between them must be done via @xmath15 . both @xmath13 and @xmath14 transmit their messages simultaneously to @xmath15 with transmit power @xmath16 and @xmath17 , respectively . in the broadcast phase ,
the relay @xmath15 employs linear processing with an amplification matrix @xmath18 to process the received signal and broadcasts the processed signal to the nodes with the harvested power @xmath19 .
we assume that each source node is equipped with a pair of transmitter - receiver antennas for signal transmission and reception respectively .
we use m@xmath20 and m@xmath21 to denote the number of transmit and receive antennas at @xmath15 , respectively .
we use @xmath22 and @xmath23 to , respectively , denote the directional channel vectors between the source node @xmath24 s @xmath25 transmit antenna to @xmath15 s receive antennas , and that between the relay s transmit antenna(s ) to source node @xmath24 s receive antenna .
the concurrent transmission and reception of signals at the nodes produces self - interference ( si ) which inhibits the performance of a full duplex system .
we consider using existing si cancellation mechanisms in the literature to mitigate the si ( e.g. , antenna isolation , analog and digital cancellation , and etc . )
@xcite . due to imperfect channel estimation ,
however , the si can not be cancelled completely @xcite . we therefore denote @xmath26 @xmath27 and @xmath28 as the si channels at the corresponding nodes . for simplicity , we model the residual si ( rsi ) channel as a gaussian distribution random variable with zero mean and variance @xmath29 , for @xmath30 @xcite .
we further assume that the relay is equipped with a ps device which splits the received signal power at the relay for energy harvesting , amplification and forwarding of the received signal .
in particular , the received signal at the relay is split such that a @xmath31 portion of the received signal power at the relay is fed to the information receiver ( ir ) and the remaining @xmath32 portion of the power to the energy receiver ( er ) at the relay .
when the source nodes transmit their signals to the relay , the af relay employs a short delay to perform linear processing .
it is assumed that the processing delay at the relay is given by a @xmath33symbol duration , which denotes the processing time required to implement the full duplex operation @xcite .
@xmath33 typically takes integer values .
we assume that the delay is short enough compared to a time slot which has a large number of data symbols , and thus its effect on the achievable rate is negligible . at time
instant @xmath34 the received signal @xmath35 $ ] and the transmit signal @xmath36 $ ] at the relay can be , respectively , written as @xmath37 & = \mathbf{h}_{ar}{s_a}[n ] + \mathbf{h}_{br}{s_b}[n]+ \mathbf{h}_{rr } \mathbf{x}_r[n ] + \mathbf{n}_r[n],\label{y1}\\ \mathbf{x}_r[n ] & = \mathbf{w y}^{\rm ir}_r(n-\tau),\label{x_r}\end{aligned}\ ] ] where @xmath38 $ ] is the signal split to the ir at @xmath15 given by @xmath39= \rho(\mathbf{h}_{ar}{s_a}[n ] + \mathbf{h}_{br}{s_b}[n]+\mathbf{h}_{rr } \mathbf{x}_r[n]\\ + \mathbf{n}_r[n ] ) + n_p[n],\end{gathered}\ ] ] where @xmath40 @xmath41 is the additional processing noise at the ir . using ( [ x_r ] ) and ( [ y21 ] ) recursively ,
the overall relay output can be rewritten as @xmath42=\mathbf{w } ( \rho(\mathbf{h}_{ar}{s_a}[n-\tau ] + \mathbf{h}_{br}{s_b}[n-\tau]\\ + \mathbf{h}_{rr } \mathbf{x}_r[n-\tau ] + \mathbf{n}_r[n -\tau ] ) + n_p[n-\tau]).\end{gathered}\ ] ] the capacity of a relay network with delay depends only on the relative path delays from the sender to the receiver and not on absolute delays @xcite .
thus , the relay output is given as @xmath43 \!\!\ ! & = & \!\!\ ! \mathbf{w } \sum_{j=0}^{\infty } ( \mathbf{h}_{rr } \mathbf{w})^j [ \rho(\mathbf{h}_{ar}{s_a}[n - j\tau-\tau ] \nonumber\\ \!\!\ ! & + & \!\!\ !
\mathbf{h}_{br}{s_b}[n - j\tau-\tau ] + \mathbf{n}_r[n - j\tau-\tau ] ) \nonumber\\ \!\!\ ! & + & \!\!\ !
n_p[n - j\tau-\tau ] ] , \label{xr_new}\end{aligned}\ ] ] where @xmath44 denotes the index of the delayed symbols . to simplify the signal model and to keep the optimization problem more tractable
, we add the zero forcing ( zf ) solution constraints such that the optimization of @xmath18 chooses to null out the rsi from the relay output to the relay input @xcite .
to realize this , it is easy to check from ( [ xr_new ] ) that the following condition is sufficient @xcite : @xmath45 consequently , ( [ xr_new ] ) becomes @xmath43\!\!\ ! & = & \!\!\!\mathbf{w } ( \rho ( \mathbf{h}_{ar}{s_a}[n-\tau ] + \mathbf{h}_{br}{s_b}[n-\tau ] \nonumber\\ \!\!\ ! & + & \!\!\ ! \mathbf{n}_r[n-\tau ] ) + n_p[n-\tau ] ) , \label{xr_new1}\end{aligned}\ ] ] with the covariance matrix @xmath46 thus the relay output power can be written as @xmath47 + \mathrm{trace } ( \mathbf{w}\mathbf{w}^\dagger).\end{gathered}\ ] ] the received signal at @xmath13 is given by @xmath48 \!\!\ !
& = & \!\!\!\mathbf{h^{\dagger}}_{ra } \mathbf{x}_r[n ] + h_{aa}s_a[n ] + n_a[n]\nonumber\\ \!\!\ ! & = & \!\!\ ! \rho ( \mathbf{h^{\dagger}}_{ra } \mathbf{w } \mathbf{h}_{ar } s_a [ n-\tau ] \nonumber\\ \!\!\ ! & + & \!\!\ !
\mathbf{h^{\dagger}}_{ra } \mathbf{w } \mathbf{h}_{br } s_b
[ n-\tau ] + \mathbf{h^{\dagger}}_{ra}\mathbf{wn}_r[n])\nonumber\\ \!\!\ ! & + & \!\!\ ! \mathbf{h^{\dagger}}_{ra } \mathbf{w}n_p[n ] + h_{aa}s_a[n]+ n_a[n].\label{y^{sa}}\end{aligned}\ ] ] after cancelling its own signal @xmath49,$ ] it becomes @xmath48 \!\!\ ! & = & \!\!\ ! \rho ( \mathbf{h^{\dagger}}_{ra } \mathbf{w } \mathbf{h}_{br } s_b [ n-\tau ] + \mathbf{h^{\dagger}}_{ra}\mathbf{wn}_r[n])\nonumber\\ \!\!\ ! & + & \!\!\ ! \mathbf{h^{\dagger}}_{ra } \mathbf{w}n_p[n ] + h_{aa}s_a[n]+ n_a[n].\label{y^{sa1}}\end{aligned}\ ] ] the received signal - to - interference - plus - noise ratio ( sinr ) at node @xmath50 , denoted as @xmath51 can be expressed as @xmath52 similarly , the received sinr @xmath53 at node @xmath54 can be written as @xmath55 the achievable rates are then given by @xmath56 and @xmath57 at nodes @xmath50 and @xmath54 , respectively .
the signal split to the er at @xmath15 is given as @xmath58 + \mathbf{h}_{br}{s_b}[n]+ \mathbf{h}_{rr } \mathbf{x}_r[n ] + \mathbf{n}_r[n]),\ ] ] where @xmath59 denotes the energy conversion efficiency of the er at the relay which accounts for the loss in energy transducer for converting the harvested energy to electrical energy to be stored . in this paper , for simplicity , we assume @xmath60 .
thus , the harvested energy at the relay is given by @xmath61 + \delta_r),\label{q1a}\ ] ] where @xmath62 and @xmath63 is the additive white gaussian noise ( awgn ) with zero mean and unit variance at the relay .
note that the conventional hd relay communication system requires two phases for @xmath13 and @xmath14 to exchange information @xcite .
fd relay systems on the other hand reduce the whole operation to only one phase , hence increasing the spectrum efficiency . for simplicity , we assume that the transmit power at the source nodes are intelligently selected by the sources . therefore , in this work , we do not consider optimization at the source nodes . to ensure a continuous information transfer between the two sources ,
the harvested energy at the relay should be above a given threshold so that a useful level of harvested energy is reached . as a result , we formulate the joint relay beamforming and receive ps ratio ( @xmath64 ) optimization problem as a maximization problem of the sum rate .
mathematically , this problem is formulated as @xmath65 where @xmath66 is the maximum transmit power at the relay and @xmath67 is the minimum amount of harvested energy required to maintain the relay s operation .
in this section , our aim is to maximize the sum - rate of the proposed fd mimo two - way af - relaying channel .
considering the fact that each source only transmits a single data stream and the network coding principle encourages mixing rather than separating the data streams from the two sources , we decompose @xmath18 as @xmath68 , where @xmath69 is the transmit beamforming vector and @xmath70 denotes the receive beamforming vector at the relay . then the zf condition is simplified to @xmath71 or equivalently @xmath72 because in general @xmath73 @xcite .
we further assume without loss of optimality that @xmath74 .
therefore , the optimization problem in ( [ problem ] ) can be rewritten as ( [ y6 ] ) ( see top of next page ) [ y6 ] @xmath75 where @xmath76 and @xmath77 observe in ( [ y6 ] ) that @xmath70 is mainly involved in @xmath78 and @xmath79 , so it has to balance the signals received from the sources .
according to the result obtained in @xcite , @xmath70 can be parameterized by @xmath80 as @xmath81 where @xmath82 is a non - negative real - valued scaler .
it should be made clear that ( [ alex1 ] ) is not a complete characterization of @xmath70 because it is also involved in the zf constraint @xmath83 but this parameterization makes the problem more tractable .
thus , given @xmath84 we can optimize @xmath69 for fixed ps ratio @xmath64 . then perform a 1-d search to find the optimal @xmath85 . for given @xmath70 and @xmath69 ,
the optimal receive ps ratio @xmath64 can be determined .
firstly , using the monotonicity between sinr and the rate , ( [ y6 ] ) can be rewritten as @xmath86 problem ( [ rho1 ] ) is a linear - fractional programming problem , and can be converted into a linear programming problem @xcite .
the receive ps ratio is determined by the equation set below : @xmath87 using the procedure in @xcite , the optimal @xmath64 can be found by @xmath88 we check whether the above solution satisfies the constraint ( [ rho1 ] ) . if it does , then it is the optimal solution .
otherwise the energy harvesting constraint should be met with equality thus giving the optimal receive ps ratio @xmath89 given by @xmath90 here , we first study how to optimize @xmath69 for given @xmath82 and @xmath91 then we perform a 1-d search on @xmath82 to find the optimal @xmath85 which guarantees an optimal @xmath92 as defined in ( [ alex1 ] ) for the given @xmath91 for convenience , we define a semidefinite matrix @xmath93
. then problem ( [ y6 ] ) becomes @xmath94 where @xmath95 is given in ( [ fw ] ) ( see next page ) .
@xmath96 clearly , @xmath95 is not a concave function , making the problem challenging . to solve ( [ fw ] )
, we propose to use the difference of convex programming ( dc ) to find a local optimum point . to this end
, we express @xmath95 as a difference of two concave functions @xmath97 and @xmath98 @xcite , i.e. , @xmath99 where @xmath100@xmath101@xmath102@xmath103@xmath104 and @xmath105@xmath106@xmath107 @xmath108 @xmath109 .
note that @xmath97 is a concave function while @xmath98 is a convex function .
the main idea is to approximate @xmath98 by a linear function .
the linearization ( first - order approximation ) of @xmath98 around the point @xmath110 is given in ( [ main^_prob ] ) .
@xmath111 then the dc programming is applied to sequentially solve the following convex problem : @xmath112 we solve ( [ main_prob2 ] ) by : * choosing an initial point @xmath113 . * for @xmath114 , solve ( [ main_prob2 ] ) until convergence .
notice that in ( [ main_prob2 ] ) , we have ignored the rank1 constraint on @xmath115 this constraint is guaranteed to be satisfied by the results in theorem 2 in @xcite and also in @xcite when m@xmath116 .
thus , the decomposition of @xmath18 leads to the optimal solution @xmath117 .
given @xmath69 , the optimal receive beamforming @xmath70 can be obtained by performing a 1-d search on @xmath82 to find the maximum @xmath85 which maximizes @xmath118 for a fixed value of @xmath31 .
see algorithm 1 .
the bounds of the rate search interval are obtained as follows .
the lower bound @xmath119 is obviously zero while the upper bound @xmath120 is defined as the achievable sum - rate at zero rsi . with optimal @xmath85
, the optimal @xmath121 can be obtained from ( [ alex1 ] ) .
now , the original beamforming and receive ps optimization in problem ( [ y6 ] ) can be solved by an iterative technique shown in algorithm 2 .
algorithm 2 continually updates the objective function in ( [ y6 ] ) until convergence .
set @xmath122 @xmath123 set @xmath124 as numerals and @xmath125 ( any value @xmath126 ) .
set @xmath127 obtain @xmath89 by considering ( [ lola ] ) .
set @xmath82 = non - negative scaler and obtain @xmath70 in ( [ alex1 ] ) . at step @xmath128 ,
set @xmath129 until @xmath130 where @xmath131 is the searching step size .
set @xmath132 , @xmath133 * repeat * \i ) set sum - rate @xmath134 \ii ) obtain the optimal @xmath113 by solving ( [ main_prob2 ] ) .
\iii ) update the value of @xmath135 with the bisection search method : if ( ii ) is feasible , set @xmath136 @xmath137 ; otherwise , @xmath138 .
* until * @xmath139 where @xmath140 is a small positive number .
thus we get the optimal @xmath82 which maximizes ( [ alex1 ] ) to give @xmath141 @xmath142 obtain @xmath143 by comparing @xmath144 .
initialise @xmath82 * repeat * \1 ) solve ( [ rho1 ] ) to obtain optimal @xmath91 \2 ) solve ( [ max1 ] ) using algorithm [ algorithm1 ] to obtain @xmath145 @xmath146 @xmath147
in this section , we evaluate the performance of the proposed algorithm through computer simulations assuming flat rayleigh fading environments . in fig .
[ sumr_eupisco ] , we show the sum - rate results against versus the transmit power budget @xmath148 ( db ) for various harvested energy constraint . the proposed scheme ( ` joint opt ' in the figure ) is compared with those of the fixed receive beamforming vector ( @xmath70 ) ( ` frbv'= 0.583 ) at optimal ps coefficient ( @xmath89 ) . remarkably , the proposed scheme yields higher sum - rate compared to the sum - rate of the frbv schemes which essentially necessitates joint optimization .
the impact of the rsi on the sum - rate is studied in fig .
[ si_eupisco ] .
results show that an increase in the rsi results in a corresponding decrease in the achievable sum - rate .
in this paper , we investigated the joint beamforming optimization for swipt in fd mimo two - way relay channel and proposed an algorithm which maximizes the sum - rate subject to the relay transmit power and harvested energy constraints . using dc and a 1-d search , we jointly optimized the receive beamforming vector , the transmit beamforming vector , and receive ps ratio to maximize the sum - rate .
simulation results confirm the importance of joint optimization .
a. a. nasir , x. zhou , s. durrani , and r. a. kennedy , `` relaying protocols for wireless energy harvesting and information processing , '' _ ieee trans .
wireless commun .
12 , no . 7 , pp .
36223636 . jul .
2013 .
x. ji , b. zheng , y. cai , and l. zou , `` on the study of half duplex asymmetric two way relay transmission using an amplify and forward relay , '' _ ieee trans . on vehicular technology _ ,
16491664 , may . 2012 . | in this paper , we investigate the problem of two - relay beamforming optimization to maximize the achievable sum - rate of a simultaneous wireless information and power transfer ( swipt ) system with a full - duplex ( fd ) multiple - input multiple - output ( mimo ) amplify - and - forward ( af ) relay . in particular , we address the optimal joint design of the receiver power splitting ( ps ) ratio and the beamforming matrix at the relay given the channel state information ( csi ) .
our contribution is an iterative algorithm and one - dimensional ( 1-d ) search to achieve the joint optimization .
simulation results are provided to demonstrate the effectiveness of the proposed algorithm . |
while isolated nonlinear systems are fairly well understood , much less is known about emergent dynamics of coupled systems @xcite .
complete synchronization , also known as zero - lag synchronization , where all nodes of a network behave identically , is just the simplest scenario .
nevertheless , it has important applications , for instance , as chaos synchronization for secure communications @xcite . other types of synchronization have been found as well , including phase - lag , cluster , or generalized synchronization @xcite .
recently a new form of dynamics called chimera state triggered a lot of research @xcite .
these chimera states arise in regular networks of identical phase oscillators and consist of parts with high spatial coherence and regions where spatial coherence is lost .
they can be realized with carefully prepared initial conditions and a nonlocal network topology , which both are crucial ingredients for these states .
similar phenomena , localized patterns surrounded by a chaotic background , were found in populations of threshold elements with nonlocal interaction , related to a continuity - discontinuity transition @xcite . in this paper we investigate in detail the transition from spatial coherence to incoherence in both time - discrete and time - continuous chaotic systems , using the well - known models of the logistic map and the rssler system , respectively .
see also ref .
@xcite for a first report on this transition .
we consider a ring topology and use the coupling strength and coupling range as bifurcation parameters . at the transition we observe partially incoherent chimera - like states that we identify as spatial chaos @xcite .
this paper is structured as follows : section [ sec : models ] introduces the model equations .
sections [ sec : diagram ] and [ sec : transition ] describe the general coherence - incoherence scenario and numerical results on the coherence - incoherence transition , respectively .
analytical results are presented in sec .
[ sec : analytical ] . in sec .
[ sec : lorenz ] we demonstrate that the observed scenario can also be found in the lorenz model .
finally we conclude in sec .
[ sec : conclusion ] .
in order to investigate the universality of the coherence - incoherence transition for @xmath0 coupled oscillators we consider different nonlinear models : as chaotic systems we employ the time - discrete , logistic map as well as the rssler system .
the latter is paradigmatic for chaos described by ordinary differential equations . as a control parameter
we use the coupling strength @xmath1 and the coupling radius @xmath2 , where @xmath3 denotes the number of nearest neighbors in each direction on a ring .
varying the value of @xmath3 from @xmath4 to @xmath5 we cover the range of regular network topologies from local ( nearest - neighbor ) to global ( all - to - all ) coupling . as an example for time - discrete systems , we study coupled maps : @xmath6,\ ] ] where @xmath7 ( @xmath8 ) are real dynamic variables and @xmath9 denotes the discrete time . in the coupling term
the index @xmath10 is periodic mod @xmath0 .
the function @xmath11 is a local one - dimensional map that we choose as the logistic map @xmath12 .
we fix the bifurcation parameter @xmath13 at the value @xmath14 unless stated otherwise .
this is a value that yields chaotic behavior with a positive lyapunov exponent @xmath15 . in order to show that the coherence - incoherence transition scenario is universal
, we will also discuss nonlocally coupled networks of time - continuous systems . as a prominent example for chaotic systems
, we consider the rssler model . in analogy to eq .
( [ eq : map ] ) the compound system is given by @xmath16 with @xmath17 .
again the index @xmath10 is periodic mod @xmath0 to realize the underlying ring topology .
the system s parameters are chosen as @xmath18 , @xmath19 , and @xmath20 showing chaotic behavior in the uncoupled case .
the @xmath21 coupled eqs .
( [ eq : roessler ] ) can also be written in compact form as follows @xmath22 with @xmath23 and a @xmath24 coupling matrix @xmath25 that determines , which components are transmitted .
the @xmath26 matrix @xmath27 describes the network topology . in the case of eqs .
( [ eq : roessler ] ) , we have the identity matrix for @xmath25 , and a circulant matrix with cyclically permutated rows @xmath28 is chosen for @xmath29 such that @xmath30 . this notation has been employed in the framework of the master stability function in order to investigate complete , potentially chaotic synchronisation @xcite .
based on a variational equations this formalism allows for calculation of the lyapunov exponent transverse to the synchronisation manifold @xcite .
the master stability function technique has been used in various contexts ranging from generic models @xcite to coupled lasers @xcite and neuroscience @xcite .
it can also be extended to cluster synchronization @xcite , but does not give conclusive results for the purpose of the investigation of spatially incoherent states .
-parameter plane for @xmath31 coupled logistic maps .
the system parameter is chosen as @xmath32 , @xmath33 , and @xmath34 for the solid , dashed ( red ) , and dotted ( blue ) curve , respectively . ]
let us discuss first the coherence - incoherence transition for coupled logistic maps described by eq .
( [ eq : map ] ) . for large values of coupling radius @xmath35 and coupling strength @xmath1
we observe complete chaotic synchronization .
there the oscillators behave identically ( @xmath36 ) , but chaotically in time following the dynamics of @xmath37 . using the fact that the coupling matrix of system ( [ eq : map ] )
is a circulant matrix , with cyclically permutated rows @xmath38 we obtain analytically the values for @xmath39 transversal lyapunov exponents of the completely synchronized solution @xmath40 : @xmath41 @xmath42 , where @xmath43 denotes the longitudinal lyapunov exponent corresponding to the one - dimensional local map , and the second term on the right - hand side is derived from the eigenvalues of the jacobian matrix of the system , using the fact that coupling matrix is circulant . with a decrease of the coupling radius @xmath35 or the coupling strength
@xmath1 the completely synchronized solution loses its stability in a blowout bifurcation , where the largest transverse lyapunov exponent becomes positive .
the position of the blowout bifurcation line in the @xmath44-plane depends on the local dynamics of the system elements .
figure [ fig : bb ] shows examples for the blowout bifurcation curves for @xmath45 , @xmath33 , and @xmath46 . for any of these values
the local logistic maps operate in the chaotic regime .
the region of complete chaotic synchronization is the region above the blowout bifurcation curves .
the completely synchronized state represents the simplest example of spatial coherence . to be precise , we will call a network state @xmath47 , @xmath8 , coherent on the ring @xmath48 as @xmath49 , if for any point @xmath50 the following limit holds : @xmath51 where @xmath52 denotes a network neighborhood of the point @xmath53 .
intuitively it means that the profile becomes smooth in the thermodynamic limit @xmath49 .
if the limit ( [ eq : coherencedef ] ) does not vanish for @xmath54 , at least for one point @xmath53 , the network state is considered incoherent .
parameter plane for @xmath31 logistic maps ( a ) and rssler systems ( b ) , labeled by the wave number @xmath55 .
gray scale ( color code ) inside the coherence regions distinguishes different periods and the coherence - incoherence bifurcation ( cib ) curve separates regimes with coherent and incoherent dynamics .
the light hatched region bounded by the blowout bifurcation curve bb refers to completely synchronized chaotic states .
the insets in panel ( a ) display snapshots of typical coherent states .
system parameters : panel ( a ) @xmath14 ; panel ( b ) @xmath18 , @xmath19 , and @xmath56 .
the vertical dashed lines refer to values of @xmath1 in fig .
[ fig : sweep ] .
triangles denote parameter values used to describe the bifurcation scenario in figures [ fig : maps_snap ] and [ fig : roesslerbifscenario].,title="fig : " ] parameter plane for @xmath31 logistic maps ( a ) and rssler systems ( b ) , labeled by the wave number @xmath55 .
gray scale ( color code ) inside the coherence regions distinguishes different periods and the coherence - incoherence bifurcation ( cib ) curve separates regimes with coherent and incoherent dynamics .
the light hatched region bounded by the blowout bifurcation curve bb refers to completely synchronized chaotic states .
the insets in panel ( a ) display snapshots of typical coherent states .
system parameters : panel ( a ) @xmath14 ; panel ( b ) @xmath18 , @xmath19 , and @xmath56 .
the vertical dashed lines refer to values of @xmath1 in fig .
[ fig : sweep ] .
triangles denote parameter values used to describe the bifurcation scenario in figures [ fig : maps_snap ] and [ fig : roesslerbifscenario].,title="fig : " ] below the blowout bifurcation curve , several further regions of coherence appear as the control parameters @xmath1 or @xmath35 decrease .
figure [ fig : diagram ] depicts these regions as shaded ( color ) tongues for the logistic map and the rssler system in panels ( a ) and ( b ) , respectively .
the blowout bifurcation line is added as dashed curve ( bb ) , and the regime of complete , chaotic synchronisation is indicated by the light hatching .
a coherent state has a smooth profile characterized by the number of maxima , i.e. , the wave number @xmath55 . only regions for wave numbers @xmath57 , @xmath58 , and @xmath59 are shown .
the corresponding coherent states @xmath47 are shown as snapshots in the insets of fig .
[ fig : diagram](a ) for the logistic map . further decrease of @xmath35 yields additional thin higher - order regions following a period - adding cascade @xmath60 ( not shown ) . inside the regions ,
the states are coherent in space and periodic in time and undergo a period - doubling cascade in time as @xmath35 or @xmath1 decrease . in the parameter space between the coherence regions
the network dynamics remains coherent but not periodic anymore .
the states alternate chaotically between the adjacent @xmath55-states and thus exhibit chaotic itineracy @xcite .
the combination of period - adding in space and period - doubling in time represents a remarkable feature of networks of coupled chaotic oscillators with nonlocal coupling .
figure [ fig : diagram](b ) shows stability diagram for the coupled rssler systems again in the @xmath44-plane .
it is easy to see that the overall structure of the diagram is similar to the one obtained for nonlocally coupled logistic maps ( fig .
[ fig : diagram](a ) ) . in analogy to the maps , there exist a blowout bifurcation boundary ( bb curve ) and a sequence of coherence tongues with wave numbers @xmath55 indicating the universality of the general dynamical scenario . in the shaded ( color ) regions shown in fig .
[ fig : diagram](b ) the coherent states with regular ( periodic or quasiperiodic ) dynamics are the only stable solution .
the bright ( yellow ) regions inside the coherence tongues below the stable period-2 tongues refer to multistability and more complex behavior including coexisting period-4 and torus solutions in between the period-2 and period-4 regions .
the dark ( blue ) region is the region of stability for the solutions with periods @xmath61 and larger .
since these regions appear to be very thin , we combine them for better visualization
. coherence tongue of fig .
[ fig : diagram](b ) at ( a ) @xmath62 and ( b ) @xmath63 .
the poincar section ( @xmath64 ) of @xmath65 is shown for the rssler systems in dependence on the coupling strength @xmath1 .
the black and gray ( red ) points correspond to a slow down and up sweep of @xmath1 , respectively .
long transients up to 3000 oscillation periods have been discarded .
other parameters as in fig .
[ fig : diagram ] . ] for further insight of the transitions within the coherence tongues figs .
[ fig : sweep](a ) and ( b ) depict cuts through the @xmath44-parameter space for a fixed coupling radius @xmath62 and @xmath66 , respectively ( see dashed lines in fig . [
fig : diagram](b ) ) .
the black and gray ( red ) values refer to the variable @xmath65 at a poincar section ( @xmath64 ) for a very slow decrease and increase of the coupling strength @xmath1 with steps @xmath67 , respectively . for each value of @xmath1
300 iterations are shown , where each initial condition is chosen as the final configuration of the previous value of @xmath1 .
panel ( a ) shows a sequence of chaotic , period-4 , torus , period-2 , coexisting period-2 and period-4 , period-4 , and again chaotic dynamics . in panel ( b )
there appears an additional regime of chaos in between the period-2 and period-4 region .
the coupling radius appears to be the crucial parameter in the system with nonlocal couplings . with increasing the system size @xmath0 and a proportional increase of the number of coupled neighbors ,
we obtain similar coherent solutions .
if with increasing @xmath0 the number of coupled neighbors @xmath3 does not change , then the coupling radius decreases , and the point of operation @xmath44 moves in the stability diagram ( fig . [
fig : diagram ] ) to the left , obtaining coherent solutions with larger wavme numbers or itineracy regions between them .
the bifurcation diagrams presented in fig . [ fig : diagram ] have been obtained by means of numerical simulations .
note , however , that in the thermodynamic limit @xmath68 it is also possible to obtain some analytic results .
details on such derivations are given in sec .
[ sec : analytical ] , e.g. , by identifying a scaling relation of the coherent profiles .
and @xmath69 . the gray regions labeled @xmath70 and @xmath71 refer to multistable incoherent parts of the profile . the coupling radius is fixed at @xmath72 . other parameters as in fig .
[ fig : diagram ] . ]
the described combination of period - adding in space and period - doubling in time for coherent solutions is valid for sufficiently large coupling strength ( @xmath73 ) . in ref .
@xcite we have shown that with decreasing coupling strength @xmath1 a coherence - incoherence bifurcation is observed in the time - discrete system ( [ eq : map ] ) .
the typical scenario of this transition is shown in fig .
[ fig : maps_snap ] for the logistic map : the smooth wave - like solution profile @xmath47 splits up into upper and lower branches .
furthermore , two narrow boundary layers of incoherence appear around the points @xmath74 and @xmath75 in fig .
[ fig : maps_snap](b ) .
they are indicated by shaded stripes @xmath70 and @xmath71 in figs .
[ fig : maps_snap](c ) and ( d ) .
the width of the incoherence stripes increases for smaller coupling strength @xmath1 ( see fig .
[ fig : maps_snap](d ) ) until the dynamics becomes completely incoherent as depicted in figs .
[ fig : maps_snap](e ) and ( f ) .
below the bifurcation parameter value @xmath76 numerous hybrid , that is , partially coherent states arise .
they are characterized by coherence on some intervals of the ring @xmath48 and incoherence on the complementary intervals .
[ fig : maps_snap](c ) and ( d ) display exemplary realizations of these partially coherent states . within the incoherence intervals ,
i.e. , @xmath70 and @xmath71 , any combinations of the upper and lower states are admissible for appropriately prepared initial conditions .
thus the coherence - incoherence bifurcation leads to spatial chaos @xcite that appears first at narrow incoherence intervals and , with decreasing coupling strength , encompasses the whole ring .
therefore we identify a chimera - like state of coexisting coherent and incoherent regions as a transitional state in the coherence - incoherence bifurcation scenario . opposed to previously reported chimera states in time - continuous systems @xcite , however , the temporal behavior in the present case is periodic rather than chaotic .
the complexity arises due to the huge variety of multistable incoherent states and corresponds to permutations of the sequence of upper and lower local states . with further decrease of the coupling strength ,
the chimera states cease to exist and we find completely incoherent behavior .
the findings above are based on numerical simulations of coupled chaotic logistic maps .
even if the local logistic maps operate in the periodic regime , one can observe a sequence similar to fig .
[ fig : maps_snap ] as reported in ref .
note that , when the local dynamics of the system is periodic , the region of stability of the completely synchronized solution covers almost all of the @xmath44 parameter plane , and the system is characterized by multistability .
( upper panels , @xmath77 as black dot ) and space - time plots ( lower panels ) for the rssler system .
the coupling strength is chosen as @xmath78 , and @xmath79 in ( a ) to ( d ) , respectively .
the coupling radius is fixed at @xmath80 .
other parameters as in fig .
[ fig : diagram ] . ] figures [ fig : roesslerbifscenario ] and [ fig : roessler_xovert ] display the typical scenario of coherence - incoherence transition for nonlocally coupled rssler systems , similar to the one described above for the logistic map . in the upper panels of fig .
[ fig : roesslerbifscenario ] , snapshots of the variables @xmath81 are shown . here
all nodes are shown in gray , and an exemplary node ( @xmath82 ) is highlighted as a black point .
similarly to the coupled maps , the smooth profile of the coherent ( @xmath83 ) solution of fig .
[ fig : roesslerbifscenario](a ) breaks up into two clusters in fig .
[ fig : roesslerbifscenario](b ) . with further decrease of the coupling strength
a partially coherent solution appears in the system as shown in fig .
[ fig : roesslerbifscenario](c ) .
finally we observe a chimera - like state around the upper cluster state in fig .
[ fig : roesslerbifscenario](d ) , which combines spatially coherent and spatially chaotic parts .
the lower panels of fig .
[ fig : roesslerbifscenario ] show space - time plots . here
the spatial structure and the breaking up of the profile is clearly visible .
furthermore one can already infer the temporal behavior that will be discussed in the following .
figure [ fig : roessler_xovert ] shows three - dimensional phase portraits of the @xmath81 variables and the superimposed time series of all @xmath84 , @xmath8 of the rssler systems for the same parameter values as in fig .
[ fig : roesslerbifscenario ] . again
the time series of node @xmath82 is highlighted as black ( red ) .
note that in figs .
[ fig : roessler_xovert](b ) and ( c ) the time series of the points from the incoherent regions are grouped into two clusters that are formed by the two coherent parts of the solution . in fig .
[ fig : roessler_xovert](d ) one can visualize the chaotic dynamics already in the short time interval shown . in the time domain
one can observe a period - doubling bifurcation in figs .
[ fig : roessler_xovert](a ) to [ fig : roessler_xovert](c ) eventually leading to chaos in fig .
[ fig : roessler_xovert](d ) ( lower panels ) . at the same time
the spatial irregularity becomes more pronounced as described above .
after the numerical exploration of the coherence - incoherence transition , we provide some analytical derivations in this section .
we will discuss an approximation for the value of the critical coupling strength at the coherence - incoherence transition and present a scaling of the profiles in different coherence tongues , which holds in the limit @xmath68 .
in the following , we derive an approximation of the critical coupling strength at the transition from coherent to incoherent states , see the line cib in fig .
[ fig : diagram](a ) .
we will restrict the investigation to the logistic map introduced in sec .
[ sub : map ] , but a similar argument holds for the rssler system as well . from a geometrical point of view and in the thermodynamic limit @xmath85 , coherent solutions @xmath47 approach a smooth profile @xmath86 of the spatially continuous version of eq .
( [ eq : map ] ) given by @xmath87 dy.\ ] ] a transition from coherence to incoherence occurs when the respective solution profile @xmath86 becomes discontinuous in some points @xmath53 of the ring @xmath88 . ) for coupled logistic maps for @xmath89 0.42 ( dashed ) , 0.40919 ( solid curve ) , @xmath90 ( dotted ) .
( b ) snapshots for even ( open dots ) and odd ( filled dots ) solutions for @xmath91 and @xmath72 .
other parameters as in fig .
[ fig : diagram ] . ]
let us now discuss in more detail this coherence - incoherence bifurcation .
consider a solution of system ( [ eq : continuous ] ) with wave number @xmath83 and period-2 dynamics in time .
hence we can reduce the dynamics by even and odd time steps @xmath92 and @xmath93 , respectively .
this leads to @xmath94 with @xmath95 . taking the spatial derivative yields @xmath96.\end{aligned}\ ] ] at the point @xmath53 where the smooth profile breaks up
, the spatial derivative becomes infinite .
considering that @xmath97 diverge to infinity , we can neglect the coupling term on the right - hand side of eq .
( [ eq : deriv ] ) .
the main contribution comes from the first term .
multiplying the equations for even and odd time steps we obtain : @xmath98 z_{0}'(x)z_{1}'(x),\ ] ] which yields the following condition @xmath99 keeping in mind that the local dynamics is governed by the logistic map @xmath12 and that its derivative is equal to @xmath100 , we introduce the function @xmath101 as a deviation from condition ( [ eq : cond ] ) . for fixed system parameters @xmath35 and @xmath1 that correspond to the region of coherent solutions with periodic behavior of period @xmath58 in time , and thus fixing @xmath92 and @xmath93 , we can calculate @xmath102 and determine values @xmath53 , where @xmath103 .
the obtained values @xmath53 are the locations at which the smooth profile break up into two parts .
figure [ fig : analytic ] illustrates @xmath102 for a fixed coupling radius @xmath72 for the finite case of 100 elements .
numerical evaluation shows that for a coupling strength @xmath104 the function @xmath105 approaches zero in two points .
figure [ fig : analytic](b ) displays the corresponding snapshot of the solution .
one can see that the profile breaks up at positions at which @xmath105 vanishes as indicated by the vertical dashed lines . to derive an approximation for this coupling strength let @xmath106 be the fixed point of the local logistic map : @xmath107 , hence @xmath108 . here
, for @xmath14 we have @xmath109 . under the assumption @xmath110
if @xmath111 , we obtain an approximation for @xmath1 : @xmath112 ^ 2 -1 = 0 \\ & \rightarrow \sigma \approx 1 - \frac{1}{a-2}.\end{aligned}\ ] ] for @xmath14 we get @xmath113 . from the numerical experiments we obtain that the transition from coherence to incoherence in system ( [ eq : map ] ) occurs for a coupling strength @xmath1 close to @xmath114 , which roughly agrees with our approximation
the deviation is due to the finite number of nodes in the simulations .
the derivation above was presented for wavenumber @xmath83 .
note that coherent solutions with higher wave numbers undergo the same bifurcation scenario , with the smooth profile breaking up into @xmath115 parts . as a second analytic result we will derive in the following a scaling relation that allows a mapping of coherent profiles in different regimes of coherence .
these regimes were depicted in fig .
[ fig : diagram ] by shaded ( color coded ) tongues labeled @xmath116 .
our considerations apply to rings of oscillators in the thermodynamic limit @xmath117 , @xmath0 being the number of coupled oscillators .
a comparison with numerical simulation shows that for sufficiently large finite @xmath0 the result holds in good approximation . for notational convenience
, we consider the case of time - continuous systems , but our findings are also valid for maps as we will show later ( fig . [
fig : kscale ] ) . the network model of a ring of @xmath0
linearly coupled oscillators , each coupled to its @xmath3 nearest neighbors , @xmath118 becomes , in the limit @xmath119 , @xmath120 \mbox{d}v \label{modeltherm}\ ] ] with coupling range @xmath35 and coupling strength @xmath1 as used in the previous sections .
$ ] denotes the spatial position on the ring .
compare eq .
( [ eq : continuous ] ) for the spatially continuous version of time - discrete maps .
the coupling scheme is determined by the matrix @xmath25 . in the case of the rssler system introduced by eqs .
( [ eq : roessler ] ) , @xmath25 is the identity matrix and the vector describing each node is given by @xmath122 . considering the logistic maps of eq .
( [ eq : map ] ) , the coupling term @xmath123 changes to @xmath124 , but the following argument still holds ( fig .
[ fig : kscale](a),(b ) ) . in short
, we will show that for fixed @xmath1 the spatially @xmath125-periodic dynamics of an oscillator at the position @xmath126 with coupling range @xmath127 is identical to that at a rescaled position @xmath128 of the corresponding @xmath129-periodic state in a network with coupling range @xmath130 .
from this we can conjecture to find a @xmath129-periodic state at @xmath131 , if a stable @xmath125-periodic state is found at @xmath127 , such that @xmath132 @xmath133 @xmath134 and @xmath135 , respectively , and ( c),(d ) for rssler systems with @xmath31 , @xmath136 , @xmath62 and @xmath137 , respectively .
gray ( red ) points in panels ( b),(d ) show rescaled @xmath83 solutions from panels ( a ) and ( c ) .
other parameters as in fig .
[ fig : diagram ] . ]
assume that a state @xmath138 with spatial period @xmath125 exists .
for simplicity and notational convenience consider @xmath139 .
this state evolves according to @xmath140 \mbox{d}v .
\label{unscaled}\ ] ] in order to compare another @xmath55-periodic state @xmath141 with the state @xmath142 , in the way described above , we assume @xmath143 with @xmath144 .
vice versa , scaling the spatial coordinate of @xmath145 with @xmath146 $ ] , as well as @xmath35 , by a factor @xmath147 leads to @xmath148 \mbox{d}v.\end{aligned}\ ] ] substituting @xmath149 gives @xmath150\ ; \mbox{d}w \label{scaled}\end{aligned}\ ] ] and , using eq .
, results in @xmath151 \mbox{d}w . \label{scaleeq}\ ] ] comparing the right - hand side of eq ( [ scaleeq ] ) to eq .
( [ unscaled ] ) , we find @xmath152 , if the coupling radius @xmath35 is appropriately rescaled by @xmath147 .
this derivation shows that one can recover the coupling parameters leading to a state with wave number @xmath129 from the parameters of a @xmath125 state .
this finding is illustrated in fig .
[ fig : kscale ] for @xmath139 and @xmath153 by corresponding snapshots of 100 logistic maps and 100 rssler system in panels ( a),(b ) and ( c),(d ) , respectively . already for @xmath31
there is strong agreement between the numerically obtained and scaled profiles . in the case of logistic maps ( rssler systems )
, we choose the coupling strength as @xmath154 ( @xmath155 ) and the coupling radius as @xmath134 ( @xmath156 ) in fig .
[ fig : kscale](a ) ( fig .
[ fig : kscale](c ) ) and , according to the scaling relation , as @xmath135 ( @xmath79 ) for the @xmath153 profile in fig .
[ fig : kscale](b ) ( fig .
[ fig : kscale](d ) ) .
the gray ( red ) dots in figs .
[ fig : kscale](b ) and ( d ) depict the scaled profile of fig . [
fig : kscale](a ) and ( b ) , respectively . ) for different values of coupling strength @xmath157 , and @xmath4 in panels ( a ) to ( d ) , respectively .
the coupling radius is fixed at @xmath158 and the number of nodes is @xmath31 .
parameters of the lorenz system : @xmath159 @xmath160 @xmath161 . ]
to investigate how universal the results obtained so far are , we provide one more example of a chaotic time - continuous system that exhibits the coherence - incoherence bifurcation scenario described above . figure [ fig : lorenzsnapshots ] displays the coherence - incoherence transition for nonlocally coupled lorenz systems : @xmath162 where the local parameters are fixed at the values @xmath159 @xmath160 @xmath161 . fig . [
fig : lorenzsnapshots ] reveals a striking similarity with the sequences of snapshots for nonlocally coupled chaotic logistic maps ( fig .
[ fig : maps_snap ] ) and rssler systems ( fig .
[ fig : roesslerbifscenario ] ) . for fixed coupling radius and with decreasing coupling strength
the smooth profile breaks up into two parts ( fig .
[ fig : lorenzsnapshots](b ) ) . further decrease of the coupling strength leads to the appearance and growth of incoherent parts , repeating again the typical scenario of the coherence - incoherence bifurcation described above .
additional numerical evidence shows that system ( [ eq : lorenz ] ) is characterized by high multistability involving different types of solutions .
there are no clearly separated tongues in the @xmath44-parameter plane present as in logistic maps and rssler systems , but there is strong overlap of the regimes of different wave numbers @xmath163 and the region of spatially coherent synchronized chaotic dynamics ( @xmath164 ) .
this is due to the completely different local dynamics , which is characterized by a double scroll and two symmetric saddle - foci .
the coupling induces stable spatially coherent steady states in the form of wavelike profiles with wave numbers @xmath165 ( fig .
[ fig : lorenzsnapshots](a ) ) .
by both numerical and analytical means we have investigated the transition from coherence to incoherence in coupled chaotic systems . in dependence upon the coupling strength and coupling range
we have observed a combination of period - adding in space and period - doubling in time for nonlocally coupled chaotic maps and rssler systems .
an inspection of the spatial profiles reveals a coherence - incoherence bifurcation , where the smooth coherent profile of the solution breaks up into parts and gives rise to partially coherent solutions and chimera states . upon further decrease of the coupling strength
completely incoherent states are found as a sign of fully developed spatial chaos . a similar transition from coherence to incoherence
is found for nonlocally coupled lorenz systems .
our findings hold for both time - discrete and time - continuous systems indicating the universality of the transition . in conclusion ,
we have analyzed in detail the transition between spatial coherence and incoherence that appears in different systems with nonlocal couplings , and we find similar bifurcation scenario for nonlocally coupled logistic maps , rssler and lorenz systems . beyond numerical investigations ,
we have determined analytically the critical coupling strength for this bifurcation , and established a scaling relation for profiles with different wave numbers .
io and ph acknowledge support by the bmbf under the grant no . 01gq1001b
( _ frderkennzeichen _ ) .
es and ym acknowledge support by deutsche forschungsgemeinschaft in the framework of sfb910 .
we thank edward ott for stimulating discussions . | we investigate the spatio - temporal dynamics of coupled chaotic systems with nonlocal interactions , where each element is coupled to its nearest neighbors within a finite range . depending upon the coupling strength and coupling radius
, we find characteristic spatial patterns such as wave - like profiles and study the transition from coherence to incoherence leading to spatial chaos .
we analyze the origin of this transition based on numerical simulations and support the results by theoretical derivations identifying a critical coupling strength and a scaling relation of the coherent profiles . to demonstrate the universality of our findings we consider time - discrete as well as time - continuous chaotic models realized as logistic map and rssler or lorenz system , respectively .
thereby we establish the coherence - incoherence transition in networks of coupled identical oscillators . |
the recent observations of an accelerating expansion @xcite , together with new cosmic microwave , large - scale structure and lensing data , give a strong indication that the universe fluid is composed of at least four different components : relativistic matter ( @xmath00.01% ) , baryons ( few per cent ) , dark matter ( @xmath130% ) and dark energy ( @xmath170% ) .
the present amount of these components raises some deep questions : why are the dark matter and dark energy , which supposedly have a different scaling with time , almost equal right now @xcite ?
why are the baryons strongly suppressed with respect to dark matter ?
why is the coincidence between the dark components relatively close to the equivalence between matter and radiation ( as pointed out by @xcite ) ?
the problem of the present coincidence between dark energy and dark matter could be simply solved if the two fluids have the same scaling with time ( let us call a system of fluids @xmath2 with identical scaling a `` stationary '' model , since @xmath3 ) .
it is interesting to remark that such a scaling can soon be observationally tested @xcite .
it is well known that , assuming an exponential potential for the scalar field representing the dark energy , the field scales as the dominant component @xcite , but this does not produce acceleration .
the question therefore arises of which is the most general system of dark matter ( modeled as a perfect fluid ) and dark energy ( a scalar field ) that allows an _ accelerated stationary global attractor_. in this paper , following the arguments of ref .
@xcite , we show that all models which contain an accelerated stationary global attractor reduce asymptotically to a system characterized by an exponential potential _ and _ a linear coupling between the two components .
systems of such a kind have been already analyzed first in refs .
@xcite and successively by many authors @xcite , while their perturbations have been studied in ref . @xcite .
in order to pass local gravity experiments @xcite , we argue that it is necessary to break the universality of the coupling , leaving the baryons uncoupled or weakly coupled ( see e.g. @xcite ) .
these assumptions completely define our cosmology .
the consequences of a linear coupling between the dark components are manifold .
first , as already remarked , this explains the cosmic coincidence and the acceleration .
second , the accelerated regime explains the decay of the baryons with respect to the dark components .
third , the near coincidence of the equivalence between the luminous components ( baryons and radiation ) and the beginning of the dark era ( dark equivalence from now ) is automatically enforced .
fourth , the baryons are the dominant component between the two equivalence times , producing a decelerated epoch in which gravitational instability is effective .
fifth , the present status of the universe is independent of the initial conditions , being on a global attractor .
in addition , it is to be noticed that the model requires only constants of order unity in planck units , and that they are all fixed by the observations of the present energy densities and the acceleration .
nothwithstanding these intriguing features , we show that the model we present here fails in satisfying at the same time the nucleosynthesis requirements together with producing an acceptable cosmic microwave background ( cmb ) angular power spectrum .
in fact if the conditions for a standard nucleosynthesis are adopted , during the recent accelerated regime a fast growth of the perturbations is induced , which in turn causes an excessive integrated sachs - wolfe ( isw ) effect , in contrast with the observations . nevertheless , we think that the dynamics we discuss is interesting on its own .
the conclusion we draw is that only models with a time - dependent coupling or a potential more complicated that an exponential may contain an accelerated global attractor and at the same time be compatible with nucleosynthesis .
a solution along these lines has been proposed in @xcite where a non - linear modulation of the coupling allows nucleosynthesis to happen and structure to form in a regime of weak coupling , while the acceleration is produced in the subsequent ( present ) regime of strong coupling . in ref .
@xcite it has been shown that a similar mechanism may be realized in superstring theories .
let us show first why all two - fluid systems with a stationary accelerated attractor reduce asymptotically to the one investigated below .
let us write a generic coupled two - fluid systems with equations of state @xmath4 in a flat friedmann metric as @xmath5 where the subscript @xmath6 stands for cold dark matter ( cdm ) and the subscript @xmath7 for a scalar field .
the friedmann equation is @xmath8 where @xmath9 and @xmath10 . as shown in ref .
@xcite , the stationary condition @xmath11 , that is @xmath12 , can be satisfied only if @xmath13 putting @xmath14 and observing that @xmath15 , we have @xmath16 where @xmath17 , the ratio of the potential to kinetic scalar field energy ( notice that @xmath18 ) . the stationary solution is accelerated if @xmath19 which , for @xmath20 , can be realized only if @xmath21 ( i.e. @xmath22 ) and @xmath23 so far we repeated the steps of ref .
now , let us consider the asymptotic behavior of @xmath24 .
if @xmath25 , the kinetic energy dominates over the potential energy , and the asymptotic solution is not accelerated . if , on the other hand , @xmath26 , the potential energy dominates , and the solution will be identical to that of a cosmological constant .
in fact , in this limit @xmath27 and @xmath28 and eq .
( [ cons ] ) gives @xmath29 where @xmath30 which implies @xmath31 .
therefore , barring oscillatory solutions , only if @xmath32 the scalar field behaves as stationary accelerated dark energy ( clearly this case includes also that of dark energy as a perfect fluid ) . when @xmath24 is constant , the coupling reduces to @xmath33 with @xmath34 which is the form we study below .
moreover , it is easy to show that @xmath35const .
implies an exponential potential @xmath36 where @xmath37 .\ ] ] the conclusion is that a linear coupling and an exponential potential represents the only non - trivial asymptotic case of stationary accelerated dark energy .
it is not difficult to see that this theorem extends also to a brans - dicke theory with an explicit coupling between matter components .
as will be shown below , such a solution is also a global attractor in a certain region of the parameter space .
the condition for acceleration , @xmath38 , together with @xmath39 imply in eq .
( [ etabeta ] ) that @xmath40 this limit is much larger than allowed by local gravity experiments on baryons , which give at most @xmath41 ( see e.g. @xcite ) .
therefore , the theory must break the universality of the coupling and let the baryons be decoupled from dark energy .
an immediate consequence of the species - dependent coupling , so far unnoticed , can be seen by observing that the energy density of the dark components scales as @xmath42 for any @xmath43 the energy density decays slower than in the standard matter - dominated friedmann universe
. therefore , any uncoupled ( or weakly coupled ) component , as the baryons , decays faster than the coupled ones ( see also @xcite ) .
the cosmology we study below is a more realistic version of the one above : we include in fact radiation and baryons , both of which are coupled to the dark components only through gravitation .
once baryons and radiation decay away , we recover the stationary accelerated attractor .
the einstein equations for our model have been already described in @xcite , in which a similar model ( but on a different attractor , i.e. , for different parameters ) was studied ( see also ref .
@xcite ) . here
we summarize their properties .
the conservation equations for the field @xmath7 , cold dark matter , baryons ( @xmath44 ) , and radiation ( @xmath45 ) , plus the friedmann equation , are @xmath46 where @xmath47 and @xmath48 ( we put @xmath49 instead of @xmath50 for generality ) .
the coupling @xmath51 can be seen as the relative strength of the dark matter - dark energy interaction with respect to the gravitational force .
the only parameters of our model are @xmath51 and @xmath52 ( the constant @xmath53 can always be rescaled away by a redefinition of @xmath7 ) .
for @xmath54 we reduce to the standard cosmological constant case , while for @xmath55 we recover the ferreira & joyce model of @xcite . as shown in ref .
@xcite , the coupling we assume here can be derived by a conformal transformation of a brans - dicke model , which automatically leaves the radiation uncoupled . to decouple the baryons one needs to consider a two - metric brans - dicke lagrangian as proposed in @xcite .
additional theoretical motivations for this kind of coupling have been put forward in ref .
@xcite and for coupled dark energy in general in ref .
@xcite .
the system ( [ sys ] ) is best studied in the new variables @xcite @xmath56 and @xmath57 and the time variable @xmath58
. then we obtain @xmath59 where the prime denotes derivation with respect to @xmath60 the cdm energy density parameter is obviously @xmath61 while we also have @xmath62 @xmath63 and @xmath64 .
the system is subject to the condition @xmath65 .
the critical points of system ( [ sys2 ] ) are listed in tab .
i , where @xmath66 is the scale factor exponent , @xmath67 , where @xmath68 , and where we used the subscripts @xmath69 to denote the existence of baryons , matter or radiation , respectively , beside dark energy . in tab .
ii we report the conditions of existence and stability of the critical points , denoting @xmath70 and @xmath71 . [ cols="^,^,^,^,^,^,^,^,^",options="header " , ] the radiation equivalence occurs at a redshift @xmath72 given by @xmath73 for realistic values .
the dark equivalence redshift @xmath74 can be found equating the baryon density and the dark energy density . from the conservation laws
@xmath75 putting @xmath76 and approximating @xmath77 ( valid for @xmath78 ) it turns out that @xmath79 ^{\frac{1}{3}\left ( 1+\frac{1}{r}\right ) } , \ ] ] for @xmath80 , we obtain @xmath81 .
the three main observations we compare our model to , nucleosynthesis , structure formation and present acceleration , are produced in turn during the three eras .
the background trajectory discussed above passes the nucleosynthesis constraint , yields the observed acceleration and explains the cosmic coincidence .
however , as we show next , it gives an exceedingly large integrated sachs - wolfe effect .
close to the critical points @xmath82 and @xmath83 the perturbations in the cold dark matter component ( @xmath84 ) and in the baryonic one ( @xmath85 ) grow as a power of the scale factor , that is @xmath86 ( see ref .
@xcite)@xmath87 in this expression the baryon bias @xmath44 and the growth exponent @xmath88 depend only on the parameters @xmath52 and @xmath89 considering only the dominating growing modes , during the baryonic phase the perturbations evolve asymptotically with the law ( see for instance ref .
@xcite)@xmath90 while in the last plateau the common growth exponent is @xmath91\ ] ] where@xmath92 in ref .
@xcite it is also reported how restrictions on the baryon bias would result in further constraints on the model but these will not be used here since experimental bias determinations still remain rather uncertain . the constraint provided by nucleosynthesis
can surely be considered on firmer grounds .
the nucleosynthesis constraint @xmath93 ( so that @xmath941 mev@xmath95 0.1 ) together with the limitation @xmath96 implies a value of @xmath51 comprised between @xmath97 and @xmath98 . for this range of values the growth exponent in the last era
is found to lie between @xmath99 and @xmath100 respectively . in fig .
4 numerical evolutions of @xmath84 and @xmath85 using the full set of equations are shown in the case @xmath101 and @xmath102 for a fluctuation of wavelength @xmath103 mpc @xmath104 : the fast growth during the final stage shows up clearly .
with such a conspicuous growth one expects a very large ( late ) isw effect on the cmb , which in fact appears in the numerical integrations of the model of fig . 5 ,
produced using a version of cmbfast modified for the dark energy .
the angular power spectrum is forced , by the normalization procedure , to be highly suppressed at the lowest angular scales respect to the observed values .
this determines the failure of the model investigated here . even neglecting the nucleosynthesis constraint ,
the existence of a radiation era requires @xmath105 , a value that induces again an unacceptably large isw effect .
our coupled dark energy model provides a cosmological scenario with quite unusual features .
the standard sequence of a dark matter era followed by a cosmological constant era , which forces to put the present universe on a unlikely transient , is here replaced by a baryonic era and a stationary dark era in which dark energy and dark matter share a constant fraction of the total density .
contrary to almost all models published so far , the present universe can be seen as already on the final global attractor ( except that , luckily , some baryons are still around ) .
the two new dimensionless constants introduced in our scenario , @xmath51 and @xmath52 , are determined by the present dark matter energy density and by the present acceleration , and can be of order unity . all cosmologies with an accelerated stationary global attractor reduce asymptotically to the model discussed in this paper . in this model the coincidence problem
is immediately solved by setting @xmath51 and @xmath52 to the same order of magnitude .
regardless of the initial condition , the universe evolves to a stationary state with @xmath106const . _ and
_ of order unity
. moreover , this model explains also why the accelerated epoch occurs _
just before _ the present or , equivalently , why there are far less baryons than cdm . the reason is provided by eq .
( [ zdark ] ) : fixing @xmath76 of order unity and @xmath107 of the order of a few per cent , we have @xmath74 near unity , regardless of the initial conditions .
that is , the fact that we observe a relatively small quantity of baryons around implies that the accelerated epoch is recent . much more or much less
baryons would push the beginning of the accelerated epoch far in the future or in the past .
the problem of the near coincidence between the radiation equivalence and the dark equivalence can be rephrased as why @xmath72 and @xmath74 are relatively close to each other .
the answer is that is the end of the radiation era that triggers the onset of the baryon era , which in turn lasts for a relatively short time because the system is heading toward the global attractor represented by the dark era . despite these positive features , the model
as it stands can not explain our universe , since baryons and cdm fluctuations grow excessively during the last accelerated phase if it is considered that a standard nucleosynthesis has taken place earlier .
this causes a cmb angular power spectrum suppressed at the lowest angular scales respect to the observational results .
thus we are forced to conclude that , for the universe to fall on the stationary attractor , a non - linear modulation in the coupling ( as in ref .
@xcite ) , and/or a potential that reduces only asymptotically to a pure exponential , is needed .
riess et al .
ap.j . , 116 , 1009 ( 1998 ) ; s. perlmutter et al .
ap.j . , 517 , 565 ( 1999 ) i. zlatev , l. wang & p.j
steinhardt , phys .
82 896 ( 1999 ) .
n. arkani - hamed , l. hall , c. kolda , h. murayama , phys .
85 , 4434 ( 2000 ) , astro - ph/0005111 n. dalal , k. abazajian , e. jenkins , a.v .
manohar , phys .
87 , 141302 ( 2001 ) b. ratra and p.j.e .
peebles , phys .
d37 , 3406 ( 1988 ) p.g .
ferreira and m. joyce , phys .
d58 , 2350 ( 1998 ) a.r .
liddle and r.j .
scherrer , phys.rev . d59 ( 1999 )
023509 w. zimdahl , d. pavon & l. p. chimento , phys.lett .
b521 ( 2001 ) 133 , astro - ph/0105479 c. wetterich , a & a , 301 , 321 ( 1995 ) d. wands , e.s .
copeland and a. liddle , ann .
, 688 , 647 ( 1993 ) l. amendola , phys .
d62 , 043511 ( 2000 ) , astro - ph/9908440 d.j .
holden and d. wands , gr - qc/9908026 , phys .
d61 , 043506 ( 2000 ) a.p billyard and a.a .
coley , phys .
d61 , 083503 ( 2000 ) gasperini m. , gr - qc/0105082 ; a. albrecht , c.p .
burgess , f. ravndal , and c. skordis , astro - ph/0107573 ; n. bartolo and m. pietroni , phys.rev .
d61 ( 2000 ) 023518 l. p.
chimento , a. s. jakubi & d. pavon , phys .
d62 , 063508 ( 2000 ) , astro - ph/0005070;w .
zimdahl , d. schwarz , a. balakin & d. pavon , astro - ph/0009353 ; a. a. sen & s. sen , mpla , 16 , 1303 ( 2001 ) , gr - qc/0103098 l. amendola , mnras , 312 , 521 , ( 2000 ) , astro - ph/9906073 t. damour , g. w. gibbons and c. gundlach , phys .
lett . , 64 , 123 , ( 1990 ) ; t. damour and c. gundlach , phys .
d , 43 , 3873 , ( 1991 ) j. a. casas , j. garcia - bellido and m. quiros , mod .
a7 , 447 ( 1992 ) l. amendola , phys .
lett . , 86 , 196 ( 2001 ) , astro - ph/0006300 r. bean and j. magueijo , phys.lett .
b517 ( 2001 ) 177 , astro - ph/0007199 l. amendola and d. tocchini - valentini , phys.rev .
d64 ( 2001 ) 043509 , astro - ph/0011243 m. gasperini , f. piazza and g. veneziano , gr - qc/0108016 a. nunes , j. mimoso , and t. charters , phys.rev .
d63 ( 2001 ) 083506 , gr - qc/0011073 s. carroll , phys .
81 , 3067 ( 1998 ) l. amendola , phys .
d60 , 043501 , ( 1999 ) , astro - ph/9904120 e.j .
copeland , a.r .
liddle , and d. wands , phys .
d57 , 4686 ( 1997 ) k. freese et al .
b , 287 , 797 ( 1987 ) r. bean , s. h. hansen and a. melchiorri , phys.rev .
d64 ( 2001 ) 103508 , astro - ph/0104162 u. seljak and m. zaldarriaga , astrophys .
j. 469 , 437 ( 1996 ) c. b. netterfield _ et al .
_ , astro - ph/0104460 .
bond , a.h .
jaffe and l. knox , ap .
j. , 533 , 19 ( 2000 ) f. occhionero , c. baccigalupi , l. amendola , s. munastra , 1997 , phys .
d56 , 7588 t. barreiro , e.j .
copeland , n.j .
nunes , astro - ph/9910214 , phys .
d61 ( 2000 ) 127301 l. amendola and d. tocchini - valentini , astro - ph/0111535 | we show that all cosmological models with an accelerated stationary global attractor reduce asymptotically to a dark energy field with an exponential potential coupled linearly to a perfect fluid dark matter . in such models
the abundance of the dark components reaches a stationary value and therefore the problem of their present coincidence is solved .
the requirement of a vanishing coupling of the baryons in order to pass local gravity experiments induces the existence of an intermediate baryon - dominated era .
we discuss in detail the properties of these models and show that to accomodate standard nucleosynthesis they can not produce a microwave background consistent with observations . we conclude that , among stationary models , only a time - dependent coupling or equation of state might provide a realistic cosmology . |
protoplanetary discs are ubiquitous around young stars , an inevitable consequence of angular momentum conservation during star formation . young stars accrete their mass through discs which then provide the reservoir of material from which planets may form .
any understanding of the processes involved in star and planet formation requires knowledge of how discs evolve and are finally dispersed .
observations of disc evolution have typically focused on thermal emission from the dust grains in the disc through the analysis of their continuum spectral energy distribution ( sed ) at several different wavelengths .
while empirical analysis and numerical modelling of the sed of discs around young stars have proved extremely useful to study this important phase of the star and planet formation process , a number of well known degeneracies conspire to render the interpretation of the data non - unique . to remedy these shortcomings recent
observational efforts are making use of various techniques to obtain spatially resolved information of many sources . herbig
ae / be stars are more luminous than their lower mass t - tauri counterparts and provide an opportunity to observe large scale morphology in resolved images .
discs observed at nearly edge - on inclinations are particularly useful in studying the large scale structure since the star and inner disc , which are much brighter than any extended emission , are obscured along the line of sight .
several edge - on discs around young stars exhibit extended emission above and below their mid - plane at nir wavelengths ( e.g. padgett et al .
1999 & perrin et al . 2006 ) and some have been further imaged through ao systems providing excellent spatial resolution , as in the case of pds 144n ( perrin et al .
2006 ) .
photoevaporation ( hollenbach et al . 1994 ) is the mechanism by which high energy photons ( uv - x - rays ) heat the surface layers of the discs to temperatures of order the escape temperature , forming a thermally driven wind from the disc out to large radius .
photoevaporation has been proposed as the dominant disc dispersal mechanism around low - mass stars ( e.g. clarke et al . 2001 ) and has been successful in explaining many of the observational statistics ( owen et al .
however there is only tentative evidence that the photoevaporative wind itself has been detected , through neii and oi emission ( hartigan et al .
1995 ; alexander et al . 2008 ; pascucci et al 2009 ; schisano et al 2010 ; ercolano & owen 2010 ) which probes the wind on a very local scale . in order to test photoevaporation , along with the various different heating mechanisms we need global observational diagnostics to validate the models .
edge - on discs provide the perfect opportunity to detect a disc wind on a global scale through light scattered by the dust grains that will inevitably be entrained by the wind .
the origin of crystalline dust grains detected in protoplanetary discs via infra - red ( ir ) spectroscopy ( e.g. bouwman et al . 2001 ) and directly within our own solar system ( e.g. wooden et al .
1999 ) is still a matter of discussion .
the source of all dust in star formation - the ism - is inferred to be entirely amorphous ( e.g. kemper et al . 2005 ) .
spectroscopy of discs ( van boekel et al . 2005 ; apai et al . 2005 ) indicates a non - negligible level of crystalline dust outside the crystalline radius , where grains are hot enough ( @xmath1k ) to be thermally annealed and converted into crystalline grains .
this discovery has lead to the development of disc models with radial mixing to allow crystalline grains that formed in the hot inner disc to be transported to larger radii ( e.g. morfill & voelk 1984 ; gail 2001 , 2002 ; wehrstedt & gail 2002 ; bockelee - morvan et al .
2002 ; dullemond et al .
2006 ; hughes & armitage 2010 ) .
another suggestion put forward by shu et al .
( 1996 ) , is that crystalline grains could be carried outwards by a x - wind . given that the crystallization radius of @xmath21 - 2au is smaller than the minimum launch radius for a photoevaporative wind @xmath3 around herbig ae / be stars and lower mass stars , the photoevaporative wind can not be the direct source of the crystalline grains at larger radii , but could work in combination with radial mixing to produce the observed enhancement .
furthermore crystallinity around edge - on discs can be used to probe the source of the dust in the extended emission , since a disc origin will give rise to crystalline dust grains in the wind while an in - falling envelope will contain only amorphous grains . in this paper
we present a simple model to investigate the imprint of photoevaporation on the extended emission observed around edge - on herbig ae / be stars .
we will compare our model to current observations of edge - on discs along with following the fate of crystalline grains in the wind as they are transported to large radius .
in section [ model ] we describe the model , including the hydrodynamic and radiative transfer methods . in section
[ sec : images ] we present synthetic images obtained from the models . in section
[ sec : cryst ] we describe the results of crystallinity calculations for the wind .
we compare our results to observations of edge - on discs in section [ sec : compare ] and we summarise our main findings in section [ sec : conclusions ] .
considering the force balance on a dust grain , it is easy to show that small dust particles are entrained in a photoevaporative wind , which carries them out to large distances .
takeuchi et al .
( 2005 ) showed that the drag force on a grain is approximately : @xmath4 where @xmath5 , @xmath6 and @xmath7 are , respectively ; the mass , the density and the radius of a dust grain ( assumed to be spherical ) and @xmath8 and @xmath9 are the density and velocity of the wind . at large radius to refer to cylindrical co - ordinates and @xmath10 to refer to spherical polar co - ordinates](@xmath11 ) the flow is approximately spherical implying @xmath12 falls off as @xmath13 .
since gravity also falls off as @xmath13 and @xmath9 increases monotonically with radius , then if a grain is still entrained several scale heights above the launching surface , it will be entrained permanently , allowing dust to be carried to very large distance from the star , since the drag force will dominate over gravity .
we build a simplified model to test the possibility of extended emission from dust grains due to a photoevaporative wind , and we choose a set of assumptions that allows us to place an upper limit on this expected level of emission .
namely we assume that the entire dust population is able to reach the launching surface of the wind via some turbulent mechanism and ignore the effects of settling and grain growth .
our simple model ignores the details of the bound regions of the disc ] and thus ignores any emission produced by this region , including _ only _ the wind itself and its contributions to emission .
this simplification means our model will not be accurate near the mid - plane , where emission from dust in the disc s upper atmosphere dominates over the wind emission .
therefore , we are unable to reproduce the observed intensities and optical depths seen at several scale height above and below the mid - plane in an edge - on disc ; however , here we are interested in the extended emission , which , in the absence of an infalling envelope is dominated by the photoevaporative wind .
our model naturally produces a ` wingnut ' morphology similar to that seen in several scattered light images ( e.g. perrin et al .
2006 ) , and produces a spatially variable dust distribution , due to the different maximum grain size that can be entrained along each streamline .
the radial morphology of the streamlines then yields that , at a given cylindrical radius , the maximum grain size increases with height , resulting in a spatial variation of colour in the scattered light images .
furthermore crystalline grains entrained in the wind are transported outwards from regions of the disc with higher crystallinity fractions resulting in an enhancement of crystalline grains in the wind over the disc s underlying distrubtion .
we can split the method for constructing scattered light images into three separate parts : ( i ) hydrodynamic calculations of the photoevaporative wind ; ( ii ) calculation of the dust profile distribution and crystallinity based on the hydrodynamic solution ; ( iii ) radiative transfer modeling of the dust distribution . as the mass - loss rates are not well known for herbig ae / be stars we have left the ionizing luminosity ( which sets the mass - loss rates ) for an euv driven wind ( hollenbach et al .
1994 ) as a free parameter and consider the effect of changing mass - loss rates on the morphology , colour of the emission and crystallinity distribution in the winds .
we consider an euv wind from a primordial disc around a 2.5m@xmath14 star with a range in ionizing luminosity from @xmath15 to @xmath16 s@xmath17 in steps of 1 dex .
these luminosities correspond to mass - loss rates in the range @xmath18m@xmath14yr@xmath17 and are similar to values calculated by alexander et al .
( 2004 ) . in order to calculate images we compute scattered light images at 1.6 , 2.1 & 3.8@xmath19 m with band passes corresponding to the h , k & l bands . in order to consider the relative contributions of crystalline to amorphous dust we take the calculated crystallinity fractions in the disc of dullemond et al .
( 2006 ) for an evolved herbig ae / be star .
we assume that this disc crystallinity fraction is the same at the base of our photoevaporative wind , such that the crystallinity fraction along the streamline can be calculated by following the thermal evolution of the dust entrained on that streamline .
a photoevaporative disc wind is a thermally driven hydrodynamic wind occuring when the disc s surface is heated to temperatures of order the escape temperature , allowing it to launch a freely expanding wind . while the wind driving source for lower mass ( t - tauri ) stars is likely to be x - rays ( owen et al .
2010a , ercolano & owen 2010 , ercolano & clarke 2010 , owen et al .
2010b ) the wind driving source around intermediate mass stars has not yet been thouroughly investigated .
x - ray photoevaporation may still occur to some degree ; however , the lower @xmath20 ratio of herbig ae / be compared to t - tauri stars and their higher euv fluxes , may reduce the role of x - rays in driving the wind .
we adopt here the euv driven wind of hollenbach et al ( 1994 ) and hydrodynamic solution of font et al .
( 2004 ) , which is a _ simple _ and _ scalable _ hydrodynamic solution , allowing us to consider a wide range of parameter space , something not possible with more complicated fuv models ( gorti & hollenbach 2009 ) or x - ray models ( owen et al .
the simplified euv treatment is suitable for the purpose of this work which aims at being the first approach in studying the qualitative aspects of scattered light emission from a disc wind .
we have repeated the calculation of font et al .
( 2004 ) and alexander ( 2008 ) and we refer the reader to these papers for a detailed description of the model setup , while the basics are summarised below . in order to determine an accurate kinematic structure of the wind we must compute a numerical solution to the problem .
we use the zeus2d code ( stone & norman 1992 ) ; we employ a spherical grid with @xmath21 $ ] , the radial grid cells are logarithmically spaced , such that we have adequate resolution at small radius to resolve the onset of the flow .
the calculation is an isothermal wind calculation with the sound speed set to @xmath22 km s@xmath17 and the radius scaled to the length scale @xmath23 i.e. the radius at which the internal energy of the gas is enough to unbind the gas from the star .
we use a radial range of @xmath24r_g$ ] with @xmath25 and @xmath26 . the number density at the base ( i.e. the density along the @xmath27 axis ) of the wind was calculated semi - analytically by hollenbach et al . ( 1994 ) and scales as @xmath28 for @xmath29 and as @xmath30 outside @xmath31 , as in font et al .
( 2004 ) and alexander et al .
( 2008 ) we adopt the smooth * base * density profile suggested by @xcite that varies between the two power laws : @xmath32^{1/5}\ ] ] where @xmath33 is the density at @xmath31 which was determined through the numerical calculations of @xcite to be : @xmath34 where @xmath35 is the ionizing luminosity .
we note that this hydrodynamic calculation does not include a cold bound ` disc ' component , since as discussed above the base density structure of the wind is known a priori .
this base density , along with keplerian rotation is reset at every time - step and the model is allowed to evolve to a steady state launching a wind from the grid s mid - plane ( representing the disc s surface ) .
as expected , we find excellent agreement with the results of the @xcite and @xcite calculations . in figure
[ fig : structure ] we show a plot of the converged wind structure showing that the wind is approximately spherical once it has reached several scale heights , in agreement with our earlier discussion in section [ model ] . in order to calculate the dust distribution in the wind , we must make some assumptions about the underlying dust distribution in the disc .
we adopt a dust to gas mass ratio of 0.01 , and a power law grain size distribution of index -3.5 ( mathis , rumpl & nordsiek , 1977 , mrn ) , with grain sizes ranging from @xmath36=5@xmath37 m to @xmath38=1 mm , we assume spherical grains with a density of 1 g @xmath39 .
we also assume that the dust is fully mixed within the disc up to the transition between the bound cold disc and the hot euv heated flow .
we then calculate streamlines from the base of the flow to the edge of the grid , then along each streamline , compute the force balance between the drag force ( calculated from equation 1 ) , gravity and the centrifugal force .
we take the dust grain as entrained if the net force along the streamline is @xmath40 .
we then obtain the maximum grain size entrained along the entire streamline ( making sure to check that each grain can be entrained the entire length of the streamline ) . in figure
[ fig : amax ] we show the obtained maximum grain size as a function of position in the flow for an ionizing luminosity of @xmath41s@xmath17 .
s@xmath17 , we note that most dust comes from within 1 - 5@xmath42 . ]
we can then compute the dust density from this maximum grain size under the assumption that the dust all comes from the same underlying population described above . in figure
[ fig : ddensity ] we show the obtained dust density , noting that the combination of a photoevaporative wind and a selective dust population naturally reproduces a ` wingnut ' structure . s@xmath17 calculation . ]
we calculate the radiative transfer as a two step process : first we calculate the temperature structure of the dust using a 3d grid calculated from the 2d dust density calculations .
we have modified the mocassin code ( ercolano et al .
2003 , 2005 , 2008 ) to allow cell - to - cell variations of the grain size distribution .
we interpolate the results from section [ sec : dust ] onto at 3d cartesian grid with resolution @xmath43 assuming azimuthal symmetry , using the range in the x and y axis of [ -32,32]@xmath42 and in the z axis of [ 0,32]@xmath42 .
the input sed used to irradiate the wind is taken from the model set of robitaille et al .
( 2006 ) , model number:3011099 , corresponding to a 2.51m@xmath14 star with an effective temperature of 9480k and radius of 2.18r@xmath14 , surrounded by an optically thick disc extending into the dust destruction radius .
the irradiating sed is placed at the origin of the cartesian grid ; this approximation is valid in this case where we are mainly interested in scattering of near - ir ( nir ) photons on the much larger scale of the wind i.e. on scale much greater than the scale the disc emission arises from . in order to calculate images of the dust distribution ,
we have written a simple ray tracing driver for mocassin , which we have called mocassinthinimage .
the standard monte - carlo radiative transfer techniques used to calculate scattered light images in mocassin ( i.e. by capturing scattered photons along the line of sight of interest ) is inappropriate for use in this case .
the wind structure is too optically thin to obtain a converged image with a computationally feasible number of photon packets , since most packets leave the grid without interacting with the distribution .
mocassinthinimage performs a ray tracing calculation throughout the grid in both scattered and thermal emission , under the assumption that the medium is optically thin ( true of the wind always ) .
we note again that since we have not explicitly included the dust in the ` bound ' region of the disc we will not calculate scattered light emission from this region .
we also assume the disc mid - plane is optically thick to all radiation and hence only consider one hemisphere of the wind . in order to consider the structure of the dust grains in the wind ,
we have built a very simple dust crystallinity model .
we assume that the grains are launched from a disc that is infinitesimally thin and has the crystalline structure calculated by dullemond et al .
( 2006 ) , for a herbig star with a mass of 2.5@xmath44 after 3.3myr of evolution .
any dust that was amorphous and is heated to a temperature @xmath45k while entrained in the wind is then assumed to immediately become totally crystalline .
any dust that starts off crystalline or becomes crystalline at any point in the flow is assumed to be crystalline for the remainder of the flow regardless of the temperature it reaches at any other stage ( we find that no dust in the flow reaches a temperature exceeding the evaporation temperature , which would destroy crystalline dust grains ) .
in order to calculate a radial profile of the fraction of dust in the wind that is crystalline / amorphous we must apply some cut off on the flow since the total mass of the flow is a function of time since it switched on , rather than a converged quantity ( it diverges logarithmically ) .
since much of the information on dust structure comes from the 10@xmath46 m silicate feature a temperature cut - off is most appropriate , thus we calculate the mass - fraction of crystalline grains out to a temperature of 100k ; as such this should provide a reasonable estimate of the observable crystallinity - fraction ( anything colder is extremely unlikely to make much impact on the 10@xmath46 m silicate feature ) . furthermore , since each cylindrical radial region will be cut by many streamlines with different crystallinity fractions , in order to calculate an average crystallinity fraction we perform a number density weighted average out to the temperature cut - off .
in figure [ fig : images ] we show calculated images of the 5 different models in the h , k & l bands , they all show a ` wingnut ' morphology .
the emission is strongest in the bluer bands with the h band being strongest and the l being the weakest .
as the mass - loss rate increases the emission becomes more extended and reaches larger radii , since the density along a streamline falls to a given value at larger radius .
we also find that all of the emission is scattered light from the central star and inner disc , many orders of magnitude above any thermal emission from the wind itself , in agreement with our assertion that the wind is optically thin .
the fact that the emission is dominated by scattered light allows us to understand the evolution of the color variation and emission morphology as the mass - loss rates change . for high mass - loss rates the emission morphology follows the gas density distribution of the wind ( i.e. essentially spherical with a gap on axis in a region where the wind is negligible ) , while more pronounced ` wingnut ' profiles are clearly visible for lower values of mass - loss rate .
this arises from the fact that scattering at a given wavelength is dominated by dust grains of comparable sizes .
given the bands we are interested in are in the range 1 - 5@xmath46 m then when there is large spatial variation in grains of this size then a profile that differs markedly from the gas density would be expected ( e.g. the wingnut ) . however , for large enough mass - loss rates dust grains with sizes in the 1 - 5@xmath46 m range can be entrained in the flow everywhere and the emission then simply follows the gas density distribution .
we find that all of the synthetic images are dominated by blue light ; this is somewhat expected since the smallest grains can always be entrained in the flow and outnumber the larger grains , thus dominating the opacity for an mrn size distribution as the one considered here .
furthermore , the negative slope with wavelength of the irradiating field , which in the 1 - 5@xmath46 m range is dominated by emission from the disc , means that short wavelength photons are more abundant , contributing to the blue appearance of the synthetic images .
.model luminosities [ cols="^,^,^,^ " , ] [ table : table ]
we have used the wind profiles and dust temperature information from the radiative transfer calculation ( section [ sec : rad ] ) to calculate crystallinity fractions in the wind .
we have assessed the impact of the wind on the observable crystallinity fraction for discs , where the inner disc is both obscured ( edge - on ) and unobscured . in the unobscured case , to observe the disc s crystallinity fraction the wind will also make a contribution along the line of sight .
we compare the disc s emission to the wind s emission at 10@xmath46 m and find that at all radii the disc dominates over the wind , meaning that the observed crystallinity fraction for non - edge on discs is unaffected by the photoevaporative wind .
however , when the disc is observed close to edge - on the disc s contribution will be undetectable and the crystallinity fraction in the wind maybe studied .
it is important to note at this point , that the radial distribution of the crystallinity fraction in a wind is likely to differ from the disc s ` native ' distribution . in figure
[ fig : cryst ] we show radial profiles of the crystallinity fraction obtained from a photoevaporative wind ( solid black line ) and compare it to the input crystallinity distribution of the disc ( dashed blue line ) .
the figure shows that at small radii @xmath47au there is no crystallinity in the wind simply because there is no wind inside this radius to carry the dust grains to large heights . between 2.5 - 200au
there is a falling crystallinity fraction ; however , it is always in excess of the disc s underlying fraction , whereas outside @xmath48au it continues to fall even though the crystallinity fraction in the disc has now risen above the level in the wind .
the crystallinity fraction in the wind remains the same for all ionizing luminosities and is insensitive to the temperature cut - off described in section [ sec : mcryst ] in the range 150 - 50k .
the observed wind crystallinity profile is a direct consequence of the streamline topology , which implies that the grains in the wind at any given cylindrical radius must have originated in the disc at a smaller radius and hence , in general , with a higher crystallinity fraction .
the rise in the disc crystallinity at a radii @xmath49au is not translated into the wind population , since most of the dust in the wind is entrained inside @xmath50au .
the low wind rates at @xmath49au , coupled with the geometric dilution of the wind , mean that the crystallinity fraction in the wind continues to fall with radius .
the fact that the crystallinity fraction is insensitive to the mass - loss rate is simply a consequence that the number of dust particles at a given point scales approximately as @xmath51 everywhere , the same as the gas , since it is largely unaffected by the change in maximum grain size for the mrn dust distribution chosen .
herbig star ( dashed line ) . ]
there are several objects that show extended emission above and below the dark dust lane indicative of an optically thick disc .
all the objects in the sample presented by padgett et al .
( 1999 ) , which includes the well known ` butterfly star ' , show extended scattered light emission above and below a dark dust lane .
the objects in the padgett et al .
( 1999 ) sample ; however , all are relatively low - mass young stars and the high fluxes detected in scattered light indicate that there is much more mass present than would be expected in a photoevaporative wind .
furthermore the observed structure is much more filamentary than our models predict .
these observations clearly indicate that a photoevaporative wind is too weak to have produced the observed extended emission .
padgett et al .
( 1999 ) suggested that these objects are young ( class i ) objects , with the extended emission arising from an infalling envelope .
wolf et al .
( 2003 ) and stark et al .
( 2006 ) have successfully reproduced these observations using detailed radiative transfer modelling that include the combination of a disc component , an in - falling envelope and an outflow cavity .
pds 144n ( perrin et al . 2006 ) ; however , is a more promising candidate for a photoevaporative wind , with images showing a morphology very similar to the ` wingnut ' shapes predicted for the lower mass - loss rates @xmath52erg s@xmath17 . furthermore adopting the @xmath41erg s@xmath17 , which predicts a surface brightness of 0.01 jy arcsec@xmath53 in h at a radius and height of @xmath54au
, we can derive a distance to pds 144n . by comparing the observed image to our predicted image
we find a distance of 200 - 400pc , which is in the range of previous estimates . as discussed in perrin et al .
( 2006 ) , it is highly unlikely that the morphology arises from foreground extinction as suggested explanation to the morphology of the padgett et al . ( 1999 ) sample .
perrin et al .
( 2006 ) indeed discuss photoevaporation as a possible source for the extended emission .
however , they oversimplify the photoevaporation model by just considering the radial scale @xmath42 , ignoring the fact that , as discussed in section 2 , dust grains can be carried to very large distances in the wind , and that at several flow scale heights from the disc the streamlines are spherical .
our models show that the use of @xmath42 in discussing the characteristic scales of the flow in dust is a poor approximation , since , as shown in figure [ fig : amax ] , dust entrained at @xmath55 can be easily carried to radii and heights of @xmath56 producing a dust density that varies on scales different to @xmath42 as shown in figure 3 .
perrin et al . (
2006 ) present models of the object in which they construct an in - falling envelope with a jet cavity surrounding an extended passive accretion disc . while their model ( figure 6 of perrin et al .
2006 ) can also reproduce extended emission , it can not reproduce the distinctive ` wingnut ' morphology seen in the observations and in the model presented in this work ( see the top panels of figure [ fig : images ] ) . however both the model presented by perrin et al .
( 2006 ) and the ones presented in this work fail to reproduce one important aspect of the observations .
the observed colour of pds 144n is such that the extended regions of the emission become dominated by the redder scattered light .
this is opposite to what is predicted by the perrin et al .
( 2006 ) model which predicts a colour variation of red to blue with height and the blue band dominating the emission at large height .
this is due to a dust population whose maximum grains size decreases with height , as you move from the disc population to the envelope ( ism like ) population . on the contrary , our model _
can _ reproduce the sign of the observed colour variation i.e that the relative strength of the red light increases with height .
this is demonstrated in figure [ fig : colour ] where we plot the change in h - k and h - l for the @xmath41 s@xmath17 model as a function of height above the mid - plane for the @xmath41 s@xmath17 , where there is variations in the grain population at sizes of a few microns , which dominate the scattering opacity at nir wavelengths .
while our models predict that the colours become redder with height above the disc ( due to the fact that larger grains are entrained at greater height , see figure [ fig : amax ] , and these scatter red light more efficiently ) , our models predict scattered blue light to remain dominant at large heights ( see figure 4 ) .
as discussed in section [ sec : images ] , in our model blue light dominates due to the spectral slope of the irradiating sed and the fact that small grains are present in the wind at all heights .
we estimate using figure 3 of perrin et al .
( 2006 ) a change in h - l ( @xmath57(h - l ) ) of @xmath581.5 , at least three times larger than our simple model can predict .
h - x ) as a function of height above the midplane , at a cylindrical radius of 100au .
green line : x = k band colour ; red line : x = l band colour .
the model shown is the @xmath41 s@xmath17 calculation . ] however , in order to allow the redder bands to dominate over the bluer bands one needs to ( i ) increase the available disc emission at the redder wavelengths so there is more original flux to scatter and/or ( ii ) remove the smaller grains which are currently dominating the scattered light ( iii ) increase the optical depth to the star / disc so the extended region becomes optically thick , resulting in the reddening of photons as they pass through .
grain growth in the disc provides a natural solution to both ( i ) and ( ii ) .
dullemond & dominik ( 2005 ) have shown that the dust grains in a disc easily grow to @xmath59 m reducing the opacity of the disc at shorter wavelengths and resulting in a disc spectrum that falls less steeply or even rises to longer wavelengths in the 1 - 5@xmath46 m range .
furthermore , the removal of small grains from the disc will also result in the removal of small grains from the wind , reducing the scattering efficiency at the shorter bluer radiation . future observations of the @xmath60 m silicate feature may help disentangle the origin of the dust particles in the extended region around pds 144n since an in - falling envelope would be entirely composed of amorphous ism type grains , while , as shown by our models , a photoevaporative wind origin implies the presence of crystalline silicate grains in the extended regions .
we have considered the fate of dust grains that can be entrained in a photoevaporative wind from the surface of a disc surrounding a herbig ae / be star . for a median mass - loss rate of @xmath61m@xmath14yr@xmath17
, we find that grains up to radii of several microns can be entrained in the wind .
we also show that , once entrained in the wind , the dust grains will remain entrained and be carried out to very large radius .
we have considered the observational imprint of this wind - entrained dust on edge - on discs , showing that the combination of a photoevaporative wind structure and a variable dust grain population resulting from the variable drag force at the base of the wind , can naturally reproduce a ` wingnut ' morphology of the dust density distribution in the wind . using a combination of monte - carlo and ray tracing radiative transfer techniques ,
we calculate scattered light images from the density distribution at near infared wavelengths .
we find our model is not applicable to the well known edge - on discs in the padgett et al .
( 1999 ) sample , which are too young and optically thick to be explained with a photoevaporative origin , but are likely to arise from an infalling envelope as shown in wolf et al .
( 2003 ) and stark et al .
these synthetic images show a wingnut morphology inferred from the dust density distributions , similar to observations of the edge - on disc around pds 144n ( perrin et al .
2006).the synthetic images , however , are dominated by emission from the smallest grains entrained in the flow , hence failing to reproduce the colour gradient of the observations , which show redder emission at larger heights above the disc .
grain growth in the disc , is a natural solution to the colour problem , and we estimate severely depleted abundance of grains with radii smaller than @xmath62 m . finally we consider the crystallinity of dust grains entrained in the flow : by following the thermal evolution of the grains , we find that crystallinity fraction will remain unchanged when the disc s photosphere can be observed since the observable mass in the wind in much less than the observable mass in the disc .
however when the disc photosphere is obscured ( i.e. for edge - on discs ) the crystallinity fraction in the wind is significantly enhanced above the discs underlying crystallinity fraction . finally , we suggest that detection of crystalline grains in extended emission around an edge on disc is indicative of a photoevaporative wind and argues against an envelope origin for the extended emission .
we wish to thank the referee , barbara whitney for her help with improving this paper .
jeo acknowledges support of a stfc phd studentship an is indebted to the university of exeter astrophysics department for hospitality during the completion of this work .
be is supported by a science and technology facility council advanced fellowship .
some of this work was performed using the darwin supercomputer of the university of cambridge high performance computing service ( http://www.hpc.cam.ac.uk/ ) , provided by dell inc . using strategic research infrastructure funding from the higher education funding council for england .
clarke c. j. , gendrin a. , sotomayor m. , 2001 , mnras , 328 , 485 dullemond c. p. , dominik c. , 2005 , a&a , 434 , 971 dullemond c. p. , apai d. , walch s. , 2006 , apj , 640 , l67 ercolano b. , barlow m. j. , storey p. j. , liu x .- w .
, 2003 , mnras , 340 , 1136 ercolano b. , barlow m. j. , storey p. j. , 2005 , mnras , 362 , 1038 ercolano b. , young p. r. , drake j. j. , raymond j. c. , 2008 , apjs , 175 , 534 ercolano b. , clarke c. j. , 2010 , mnras , 402 , 2735 ercolano , b. , & owen , j. e. 2010 , , 406 , 1553 kemper f. , vriend w. j. , tielens a. g. g. m. , 2005 , apj , 633 , 534 lucy l. b. , 1999 , a&a , 344 , 282 mathis j. s. , rumpl w. , nordsieck k. h. , 1977 , apj , 217 , 425 morfill g. e. , voelk h. j. , 1984 , apj , 287 , 371 perrin m. d. , duchne g. , kalas p. , graham j. r. , 2006 , apj , 645 , 1272 robitaille t. p. , whitney b. a. , indebetouw r. , wood k. , denzmore p. , 2006 ,
apjs , 167 , 256 schisano e. , ercolano b. , gdel m. , 2010 , mnras , 401 , 1636 | we have performed hydrodynamic and radiative transfer calculations of a photoevaporating disc around a herbig ae / be star to determine the evolution and observational impact of dust entrained in the wind .
we find that the wind selectively entrains grains of different sizes at different radii resulting in a dust population that varies spatially and increases with height above the disc at radii @xmath0 au .
this variable grain population results in a ` wingnut ' morphology to the dust density distribution .
we calculate images of this dust distribution at nir wavelengths that also show a wingnut morphology at all wavelengths considered . we have also considered the contribution that crystalline dust grains will have in the wind and show that a photoevaporative wind can result in a significant crystallinity fraction at all radii , when the disc is edge - on .
however , when the disc s photosphere is unobscured , a photoevaporative wind makes no contribution to the observable crystallinity fraction in the disc .
finally , we conclude that the analysis of extended emission around edge - on discs could provide a new and independent method of testing photoevaporation models .
[ firstpage ] accretion , accretion discs - circumstellar matter - planetary systems : protoplanetary discs - stars : pre - main - sequence . |
details of the experimental setup and data analysis , as well as the definitions of notations , can be referred to ref .
there are two categories of events relevant to the present discussion : physics signals ( phy ) and electronic noise ( ele ) .
the phy events are due to actual energy depositions at the ulege by gammas , neutrons , neutrinos , wimps and other radiations , while ele events are induced by the various stages in the readout electronics .
the majority of the events above the noise edge of 300 ev are phy - events .
a particularly good method to extract a clean sample of phy - events below the noise edge for further studies is to require that the ulege signals are in coincidence with the anti - compton veto ( acv ) tag .
however , phy - events due to neutrinos or dark matter interactions would have to be in anti - coincidence with acv .
a major goal of the experiment is to lower the threshold by suppressing ele while keeping phy - events in some substantially large and known fraction . in order for phy - events to be included in the final spectra where physics is extracted , they have to survive three procedures , the efficiency of each of which must be known .
these efficiency factors were summarized in table 1 of ref .
@xcite : 1 .
trigger efficiency ( @xmath38 ) @xmath1 that phy would produce a trigger signal to the data acquisition ( daq ) system .
this efficiency depends on the energy of phy .
daq efficiency ( @xmath39 ) @xmath1 that the trigger signal would actually activate the daq system resulting in a complete event recorded in the computer .
this efficiency is independent of the energy of phy .
analysis efficiency ( @xmath40 ) @xmath1 that offline software procedures would retain phy and suppress ele . some phy would be rejected in the process leading to an efficiency factor which is energy - dependent . among these three efficiency factors ,
the derivations of @xmath39 and @xmath40 were discussed in ref .
the @xmath39 was evaluated accurately by `` random trigger '' ( rt ) events to be 89% fraction of the rt events actually recorded in the computer .
( cited in table 1 as `` daq dead time '' of 11.0% . )
same procedures were used in our earlier work on neutrino magnetic moments @xcite .
the evaluation of @xmath40 by two different approaches was discussed in the text and the results were shown in figure 3 .
further elaborations are made in ref .
the rest of this reply would focus on @xmath38 .
ref . @xcite challenged our values of @xmath41 , stating that the `` discriminator threshold of 20 ev '' and `` rms resolution of pedestal ( that is , rt events ) of 55 ev '' would imply a daq rate ( @xmath42 ) of @xmath620k hz , based on a known relation @xcite @xmath43 ~~,\ ] ] where @xmath44 is the shaping time ( = 6 @xmath15s for the trigger signal @xmath16 ) , @xmath26 is the rms of the pedestal noise fluctuations and @xmath45 is the threshold level above the pedestal .
the calculated rate is much higher than the actual daq rate of @xmath65 hz for the ulege in the actual measurement .
this discrepancy originates from the fact that it is incorrect to use these two energy values together in eq .
[ eq::daqrate ] . in the experiment ,
the `` energy '' of an event is defined by integrating ( the `` q - mode '' ) the @xmath17 signal within time intervals ( -20,51.2 ) @xmath15s where t=0 denotes the trigger instant .
that is , the energy is measured through the summation of 71.2x20=1424 fadc numbers . under this definition , the pedestal ( rt )
events have an rms resolution of 55 ev , after appropriate calibration
. however , the triggering was done by comparing the @xmath16 signal with a pre - set discriminator level .
that is , the relevant quantity in this operation is the amplitude of the pulse ( the `` a - mode '' ) @xmath1 and in fact of a different pulse ( @xmath16 instead of @xmath17 , but this is a minor point ) . _
the rms - resolution of rt events in q - mode is not related to the trigger configuration in a - mode and therefore should not be taken as the @xmath26 of eq .
[ eq::daqrate ] which describes the trigger rate .
_ it is an experimental optimization frequently used ( also discussed in ref .
@xcite ) @xmath1 that the energy definitions are based on averaging over long periods , while the triggering schemes make use of the instantaneous response of the signal .
pulses due to ( a ) top : a random trigger event and ( b ) bottom : a physics event at 139 ev in coincidence with anti - compton signal .
the pedestal mean and the discriminator threshold were denoted by dotted and dashed lines , respectively .
, title="fig:",width=321 ] + pulses due to ( a ) top : a random trigger event and ( b ) bottom : a physics event at 139 ev in coincidence with anti - compton signal . the pedestal mean and the discriminator threshold were denoted by dotted and dashed lines , respectively .
, title="fig:",width=321 ] having identified the source of discrepancy which led to the incorrect conclusions in ref .
@xcite , we present the evaluation of @xmath38 in what follows . displayed in figures [ singleevent]a&b are typical @xmath16 signals for rt and phy events , respectively .
the amplitude is presented in fadc unit ( 2 v is equivalent to 256 fadc unit ) .
the phy - event was measured to be 139 ev in the q - mode , and in coincidence with acv .
superimposed is the discriminator level ( 38@xmath02 mv in hardware unit , uncertainties mostly from instabilities over time ) which is 4.8 fadc unit above the mean value of the pedestal level .
it can be seen that the rt event is below threshold while the phy event is above threshold by @xmath6 1.5 fadc unit , thereby provided a trigger to the daq system with good margin .
histograms of large samples of events like those of figures [ singleevent]a&b is presented in figure [ amode ] , where the horizontal axis is in fadc unit .
the `` random noise '' histogram corresponds to the distribution of the amplitudes of every time - bin of the rt - events .
the rms is 1.1 fadc unit and represents the noise fluctuation of the @xmath16 signal .
the discriminator threshold level is also shown .
accordingly , one can equate @xmath44=6 @xmath46 , @[email protected] and @xmath26=1.1 in eq .
[ eq::daqrate ] , giving @xmath47 .
the measured rate averaged over the entire daq periods of about 17 days and for four channels together is @xmath65 hz .
this is in good agreement .
the various distributions show the maximum amplitude of phy - events at various energy ranges measured by the q - mode .
all the events above 100 ev ( the relevant range for subsequent analysis ) exhibit at least 1 fadc unit of margin above threshold .
the trigger efficiencies were then derived using the maximum amplitude distributions of the rt events at e=0 and the phy - events between 300 ev to 1000 ev . the mean and rms for the e=0@xmath1300 ev regions were evaluated by interpolation ( rather than from actual data ) to avoid biased sampling .
the results are displayed in figure [ trigeff ] .
we note that the energy range that provides the most severe constraints to the dark matter analysis is that of 200 - 250 ev , where the trigger efficiency is close to unity . using phy - events as well as data taken with precision test pulser
, it can be shown that energy measurements with q - mode ( integration ) and a - mode ( maximum amplitude ) are both valid and equivalent for pulses which are large compared to the noise fluctuations .
moreover , the q - mode measurements are well - behaved and linear all the way towards the pedestal zero - energy level . on the contrary , the a - mode measurements become inaccurate and the calibration is non - linear as the energy decreases and approaches zero . _
although the threshold is well - defined in amplitude ( [email protected] fadc unit ) , the statement `` discriminator threshold at 20 ev '' in ref .
@xcite has been misleading and too simplistic .
it is incorrect to equate the threshold to a single number in q - mode as its energy scale .
_ that is , the trigger efficiency versus energy curve of figure [ trigeff ] is not a step - function .
for instance , @xmath48% corresponds to an energy range of about 80@xmath050 ev .
this is the root of the misunderstanding and is corrected in the revised text .
another point noted in ref .
@xcite is the comparison of spectra in ref .
@xcite with those taken at the y2l underground laboratory in south korea @xcite .
such comparisons would not be appropriate .
different detectors were used even though the specifications are similar .
the electronic modules and noise sources ( which are affected by many ambient conditions ) are not identical .
the most important difference is that we did not make an attempt to use pulse shape differentiation ( psd ) techniques to suppress ele - events with y2l data .
the spectra shown with y2l data are always without psd cuts , such that the raw `` background '' @xmath1 dominated by ele - events @xmath1 below the noise edge are extremely high . daq rates in both cases are similar , at the range 1 - 10 hz . displayed in figure [ ksy2l ] are the raw spectra for both ks and y2l , as well as that after psd suppression for ks .
it can be seen that the raw spectra are comparable .
the residual differences are due to ambient electronic noise and radioactive background conditions .
we are indebted to profs .
f. avignone , p. barbeau and j. collar for careful reading and critical comments on ref .
their input not only expose the parts of the original text where the experimental procedures were not adequately explained , but also stimulate us to more thorough thoughts on the issue .
avignone iii , p.s .
barbeau and j.i .
collar , arxiv : 0806.1314v2 ( 2008 ) .
lin et al . , texono collaboration , arxiv:0712.1645 ( 2007 ) .
lett . * 90 * , 131802 ( 2003 ) ; h.t .
wong et al . , phys . rev . *
d 75 * , 012001 ( 2007 ) .
li et al . , reply to arxiv:0806.1341v2 ( 2008 ) .
statham , x - ray spect .
* 6 * , 94 ( 1977 ) .
texono internal report ( 2006 ) .
li , s.t .
lin , s.k .
lin , a.k .
soma , h.t .
wong@xmath36 + ( texono collaboration ) + @xmath36 corresponding author : [email protected] additional questions were raised in a recent paper @xcite on the experimental concepts and systematic issues of our recent results on wimp dark matter searches in the low - mass ( @xmath37 ) domain with a ultra - low energy germanium ( ulege ) detector at a threshold of 220 ev @xcite .
we provide clarifications and justifications on these issues , and conclude that there are no flaws in our procedures .
in a previous version(v2 ) @xcite , the authors raised questions on our daq dead time , trigger efficiencies and apparent inconsistencies between the kuo - sheng reactor laboratory ( ks ) and the yang - yang underground laboratory ( y2l ) data . in an earlier reply @xcite ,
these questions were addressed in details .
no further comments were made along these lines .
a new version(v3 ) of the comments @xcite was subsequently posted , where questions on the experimental concepts and various systematic issues were raised .
we responded on these comments in this article .
section ii deals with the basic concepts and performance of the psd cuts and efficiencies , while section iii elaborates on the reasons behind the various choices made in the experiment .
details of the experimental setup and data analysis , as well as the definitions of notations , can be referred to ref .
the objectives of applying selection criteria ( `` cuts '' ) on experimental data is to reject undesirable `` background '' events and increase the fraction of `` signal '' events in the data sample .
the events surviving the cuts need not be all signals , only that usually , the signal - to - background ratios are enhanced by the selection
. there are much freedom in the choice of these cuts
. there may be effective or ineffective cuts , but in general , all cuts are valid . the variables ( `` figure of merits '' fom ) on which the cuts are applied are defined to optimize the performance
. the foms can be mathematical constructions and need not correspond to , or be linear to , certain physical variables .
once the events are selected by the cuts , the physical parameters ( like energy ) can be derived for these events with different algorithms which are themselves constructed to give the best resolutions for these quantities . when a set of cuts are applied to experimental data ,
the corresponding `` selection efficiencies '' must be evaluated .
these are the probabilities that potential signal events that will survive these cuts .
the goals of applying cuts are to suppress background events as much as possible while keeping the selection efficiencies as large as possible . in this particular analysis
@xcite , the acv and crv cuts are straightforward .
they are signals from detector components other than the ulege target , and their efficiencies were evaluated by random trigger ( rt ) events .
their performance are summarized in table 1 of ref .
@xcite .
it is the psd - cut which was questioned by ref .
this cut was applied to the variables `` @xmath16 '' and `` @xmath17 '' , as displayed again in figure [ fe55 ] . among them
, @xmath17 also provides an energy measurement , while @xmath16 is a mathematical construction to optimize the performance of signal - vs - background differentiation . in this case
, `` background '' corresponds to electronic noise events below the hardware noise edge . versus @xmath19 signals , for both calibration and physics events before acv and crv cuts .
the psd selection is shown .
, width=302 ] _ motivations for the choice of @xmath16 : _ the two pulses sa@xmath49 and sa@xmath50 are correlated for physics signals .
the correlations are different and less strong for noise - events .
the conceptual idea behind the psd cut is that , given the energy of an event is known through measurements by @xmath17 , positive fluctuations of the sa@xmath49-pulses in physics - induced events at a particular time interval and amplitude ranges can be expected .
this is probed by the psd cut .
the selected time interval is optimized for energies near threshold ( @xmath51 ) , such that this interval no longer corresponds to the amplitude - peaks of sa@xmath49 for events at higher energy .
therefore , @xmath16 being non - linear with energy does not jeopardize the validity of the cut .
the reasons behind the non - linear behaviour is well - understood @xmath1 and in fact intentional .
_ is the psd cut arbitrary?_. no .
the application of the psd cut is based on genuine understanding on the behaviour of the detector hardware .
such psd techniques @xmath1 correlations of two signals at two shaping times and , alternatively , of the full and partial integrations of the signals @xmath1 are well - established ones at higher energy in the case of @xmath52/@xmath22 separation in many detector systems .
displayed in figure [ acvtag ] are the survival fraction ( @xmath53 ) of events at e=200@xmath1300 ev with an acv ( anti - compton ) tag versus the relative timing between the acv signals and the ulege triggers .
overlaid are the actual coincidence time interval between the acv and the ulege systems determined independently from hardware timing .
the coincidence window is about 3 @xmath15s .
this is defined by ( a ) the 6 @xmath15s shaping time output sa@xmath49 from ulege which provided the trigger timing ; and ( b ) the increased time - jitters at low energy .
the acv and ulege signals outside this range are accidentals and uncorrelated .
it can be seen that _ only _ ulege events in correct coincidence with the acv - tags give a substantial value of @xmath53 .
the survival fraction for events without acv - tags ( denoted by the data point at negative time in figure [ acvtag ] ) @xmath1 predominantly due to electronic noise at this energy @xmath1 is @xmath54 .
this proves that the psd cut is successfully devised and is performing its intended functions of `` suppressing electronic noise events and selecting the physics - induced events '' . 300 ev with an acv ( anti - compton ) tag versus the relative timing between the acv signals and the ulege triggers .
overlaid are the actual coincidence time interval between the acv and the ulege systems derived independently from hardware timing .
the data point at negative time is due to events without acv - tags , and corresponds to @xmath54 .
, width=302 ] _ selection efficiencies : _ the psd - efficiency for physics - induced events ( @xmath55 ) is related to the measured survival fraction @xmath53 by @xmath56 where @xmath57 and @xmath58 are , respectively , the numbers of physics - induced and noise events in correct coincidence with acv , while @xmath59 is the survival fraction for events triggered by electronic noise . in this experiment , @xmath53 and @xmath60 are measured quantities , @xmath61 at 200@xmath1300 ev and even less at lower energy , while the _ average _ @xmath58 can be evaluated from the non - coincidence samples where @xmath62 .
it can be seen that , in general , @xmath63 so long as there are finite fraction of noise events in the acv - tagged sample .
statistics are limited in the present analysis since only _ in situ _ data were used .
the number of noise events in the coincidence acv - tagged sample is finite but has large uncertainties .
accordingly , the more appropriate approach is to take zero noise - background ( @xmath64 ) , giving rise to the assignment of @xmath65 .
the subsequent upper limits derived would therefore be conservative and less constraining ones .
the variations of @xmath55 with energy under this assignment are displayed in figure 3 of ref .
inevitably , an experiment has to choose various tools from the available pool in the course of its analysis .
we explain why such choices were made for the cases raised in ref .
@xcite , and illustrate the sensitivities to the physics results if alternative schemes would had been selected instead .
it was explicitly stated in ref .
@xcite that the flat @xmath21fe spectrum at low energy is an assumption . under this assumption ,
the psd selection efficiencies were derived .
the results agree well with those obtained by the more rigorous approach with acv - tagged events .
the good agreement suggests validity of the assumption at the present level of accuracy .
accordingly , it is justified to add these results to further constrain ( to reduce uncertainties of ) our knowledge of @xmath55 .
this is the approach adopted in ref .
@xcite .
if the data from the @xmath21fe spectrum would be ignored altogether and the psd efficiencies would be derived exclusively from the acv - tagged events , @xmath55 at e=200@xmath1300 ev would move from ( [email protected] ) to ( [email protected] ) .
the 50% efficiency line would correspond to 224 ev instead of 216 ev , while the @xmath11 limits ( in units of @xmath66 throughout in this section ) at @xmath67 would increase ( become less constraining ) from 0.81 to 0.88 .
a compilation of all quenching factor ( qf ) measurements on germanium is given in figure [ qfdata ] . overlaid
are calculations from the trim software @xcite as well as by the lindhard model @xcite under two parametrizations ( k=0.20 and 0.157 ) .
both schemes have been adopted in various cdm experiments .
it can be seen that the trim results explain well the qf measurements at both low and high energy .
accordingly , we chose to use this scheme in our analysis .
the qf values are less than those evaluated with the lindhard ( k=0.20 ) model , and hence would give rise to more conservative results .
if lindhard ( k=0.20 ) would be used , the qf at 1 kev recoil energy will be increased from 0.20 to 0.21 .
the qf uncertainty estimations of 0.006 in ref .
@xcite can account for this deviation .
this alternative choice will only have minor effects on the exclusion limits , decreasing it ( becoming more constraining ) from 0.81 to 0.80 at @xmath68 the unbinned `` optimal interval method '' as formulated in ref .
@xcite was adopted to derive the exclusion limits .
the unbinned formalism allows the use of all available information in the background spectra and was used in other cdm experiments like cdms and xenon .
no background profile was assumed or subtracted , which is also a conservative approach .
the sensitivities at low @xmath5 under this scheme are driven by the absence of counts between 198 ev and 241 ev .
an alternative method would be to place the background events in different energy bins and follow the formalism of ref .
for instance , choosing 50-ev bins for e@xmath69100 ev ( thereby deliberately filling the hole at 200@xmath1250 ev ) , the @xmath11 limit at @xmath68 would increase ( become less constraining ) from 0.81 to 1.20 .
this reduction in sensitivities is expected since data binning involves loss of information .
we conclude that our choices in these three aspects of the experiment are justified .
the sensitivities of the physics results ( exclusion upper limits ) are dominated by the statistical uncertainties of the background spectra .
the potential effects on them are minor if alternative schemes would have be chosen instead
. 99 f.t .
avignone iii , p.s .
barbeau and j.i .
collar , arxiv : 0806.1314v3 ( 2008 ) .
lin et al . , texono collaboration , arxiv:0712.1645 ( 2007 ) .
avignone iii , p.s .
barbeau and j.i .
collar , arxiv : 0806.1314v2 ( 2008 ) . h.b .
li et al . , reply to arxiv:0806.1341v2 ( 2008 ) .
li et al . , phys . rev .
lett . * 90 * , 131802 ( 2003 ) ; h.t .
wong et al . , phys . rev .
* d 75 * , 012001 ( 2007 ) .
ziegler , _ transport of ions in matter _ , http://www.srim.org ( 1998 ) .
j. lindhard et al .
, medd . * 33 * , 10 ( 1963 ) .
s. yellin , phys . rev . *
d 66 * , 032005 ( 2002 ) .
for example , among many others , g. angloher et al .
, astropart . phys . *
18 * , 43 ( 2002 ) . | an energy threshold of ( 220@xmath010 ) ev was achieved at an efficiency of 50% with a four - channel ultra - low - energy germanium detector each with an active mass of 5 g. this provides a unique probe to wimp dark matter with mass below 10 gev . with a data acquisition live time of 0.338 kg - day at the kuo - sheng laboratory ,
constraints on wimps in the galactic halo were derived .
the limits improve over previous results on both spin - independent wimp - nucleon and spin - dependent wimp - neutron cross - sections for wimp mass between 3@xmath16 gev .
sensitivities for full - scale experiments are projected .
this detector technique makes the unexplored sub - kev energy window accessible for new neutrino and dark matter experiments .
there is compelling evidence from cosmological and astrophysical observations that about one quarter of the energy density of the universe can be attributed to cold dark matter(cdm ) , whose nature and properties are still unknown@xcite . weakly interacting massive particles ( wimp , denoted by @xmath2 )
are the leading candidates for cdm .
there are intense experimental efforts@xcite to look for wimps through direct detection of nuclear recoils in @xmath2n@xmath3@xmath2n elastic scattering or in the studies of the possible products through @xmath4 annihilations .
supersymmetric(susy ) particles@xcite are the leading wimp candidates .
the popular susy models prefer wimp mass(@xmath5 ) in the range of @xmath6100 gev , though light neutralinos remain a possibility@xcite .
most experimental programs optimize their design in the high - mass region and exhibit diminishing sensitivities for @xmath7 , where an allowed region due to the annual modulation data of the dama experiment@xcite @xmath1 further reinforced by the first dama / libra results@xcite @xmath1 remains unprobed .
simple extensions of the standard model with a singlet scalar favors light wimps@xcite .
detectors with sub - kev threshold are needed for probing this low - mass region and studying wimps bound in the solar system@xcite , and non - pointlike susy candidates like q - balls@xcite .
this presents a formidable challenge to both detector technology and background control .
only the cresst - i experiment has set limits@xcite with sapphire(@xmath8)-based cryogenic detector at a threshold of 600 ev .
the kuo - sheng(ks ) laboratory@xcite is located at 28 m from a 2.9 gw reactor core with an overburden of about 30 meter - water - equivalence .
limits on neutrino magnetic moments(@xmath9)@xcite with a 1.06-kg germanium detector(hpge ) at a threshold of 5 kev were reported@xcite .
these data also allowed the studies of reactor electron neutrinos@xcite and reactor axions@xcite . a background level of @xmath10(cpkkd ) at 20 kev , comparable with those of underground cdm experiments , was achieved .
the current goal is to develop detectors with kg - scale target mass , 100 ev - range threshold and low - background specifications for the studies of wimps , @xmath9 and neutrino - nucleus coherent scatterings@xcite .
ultra - low - energy germanium detectors(ulege ) is a matured technique for sub - kev soft x - rays measurements .
they typically have modular mass of 5@xmath110 g while detector arrays of up to 30 elements have been constructed . compared with @xmath8
, ge provides enhancement in @xmath2n spin - independent couplings(@xmath11 ) due to the @xmath12 dependence@xcite , where a is the mass number of the target isotopes .
the isotope @xmath13ge ( natural isotopic abundance of 7.73% ) comprises an unpaired neutron such that it can provide additional probe to the spin - dependent couplings of wimps with the neutrons(@xmath14 ) .
the nuclear recoils from @xmath2n interactions in ulege only give rise to @xmath620% of the observable ionizations compared with electron recoils at the same energy .
the suppression ratio is called the quenching factor(qf)@xcite . for clarity ,
all ulege measurements discussed hereafter in this article are electron - equivalent - energy , unless otherwise stated .
the ulege array consists of 4-element each having an active mass of 5 g@xcite .
standard ultra - low - background specifications were adopted in its construction and choice of materials .
it has identical external dimensions as the 1-kg hpge of ref .
@xcite .
apart from swapping between the two detectors , data taking was performed with all other hardware components , shieldings configuration , electronics and data acquisition ( daq ) systems@xcite kept identical .
the schematic diagram of the experimental setup inside the shieldings is depicted in figure [ ksulege ] .
the ulege signals were provided by built - in pulsed optical feedback pre - amplifiers , and were distributed to two spectroscopy amplifiers at 6 @xmath15s(@xmath16 ) and 12 @xmath15s(@xmath17 ) shaping times and with different amplification factors .
discriminator output of @xmath16 defined the trigger conditions for daq .
the threshold was set to about [email protected] times the rms fluctuations of the @xmath16 signals above the pedestal .
the daq rates for the ulege were about 5 hz , due mostly to electronic noise and agreed well with expectations@xcite .
the @xmath16 , @xmath17 , anti - compton veto ( acv ) and cosmic - ray veto ( crv ) signals were read out by 20 mhz flash analog - to - digital convertors .
random trigger(rt ) events generated at 0.1 hz and uncorrelated with the rest of the system , as well as various system control parameters , were also recorded .
fe source together with x - ray peaks from various materials .
the black histogram represents events selected by psd cuts .
deviations from the expected flat spectra contribute to psd efficiencies .
( b ) bottom : scattered plots of the @xmath18 versus @xmath19 signals , for both calibration and physics events .
the psd selection is shown .
, title="fig:",width=283 ] + fe source together with x - ray peaks from various materials .
the black histogram represents events selected by psd cuts .
deviations from the expected flat spectra contribute to psd efficiencies .
( b ) bottom : scattered plots of the @xmath18 versus @xmath19 signals , for both calibration and physics events .
the psd selection is shown .
, title="fig:",width=283 ] energy measurements were given by @xmath20 defined in the next paragraph .
figure [ fe55]a shows an energy calibration spectrum due to external @xmath21fe sources(5.90 and 6.49 kev ) together with x - rays from ti(4.51 and 4.93 kev ) , ca(3.69 kev ) , and s(2.31 kev ) . photons with energy lower than 2 kev were completely absorbed by the detector window .
the rt - events provided the calibration point at zero - energy .
the rms resolutions for the rt - events and @xmath21fe peaks were about 55 ev and 78 ev , respectively .
the calibration procedures were performed before and after the daq periods .
linearity was checked up to 60 kev with various @xmath22-sources , and between zero and 2 kev with a precision pulse generator .
the energy scale was accurate to @xmath23 , while deviations from linearity were @xmath24 .
the electronic gain drifts , also monitored _ in situ _ by the pulse generator , were @xmath25 . a detector hardware `` noise - edge '' of about 300 ev was achieved .
pulse shape discrimination ( psd ) software was devised to differentiate physics events from those due to electronic noise , exploiting the correlations in both the energy and timing information of the @xmath16 and @xmath17 signals .
displayed in figure [ fe55]b is a scattered plot of the @xmath18 and @xmath19 signals with the psd cut superimposed , where the superscripts p / t denote partial / total integration of the pulses within @xmath15s and @xmath15s , respectively , relative to the trigger instant(t=0 ) .
the noise events were suppressed .
calibration events and those from physics background were overlaid , indicating uniform response .
events selected by psd but with crv or acv tags were subsequently rejected .
the surviving events were ulege signals uncorrelated with other detector systems and could be wimp candidates .
fe - calibration and _ in situ _ data with acv tags . also shown
are the best - fit 1@xmath26 region and the trigger efficiency for physics events recorded by the daq system .
, width=283 ] [ cols="<,^,^",options="header " , ] data were taken with the ulege at ks with different hardware and software configurations .
they provided important input on the background understanding and performance optimizations for future full - scale experiments .
the data set with the best background and threshold has a daq live time of 0.338 kg - day .
the analysis results and the systematic effects at the two energy intervals which defined the sensitivities for @xmath5 below and above @xmath610 gev are summarized in table [ summarytable ] .
are superimposed onto the spectrum shown in linear scales in the inset .
, width=283 ] the ulege data were taken in conjunction with a csi(tl ) scintillator array@xcite for the studies of neutrino - electron scattering .
the combined daq rate was about 30 hz . the daq dead time and the crv+acv selection efficiencies listed in table [ summarytable ] were accurately measured using rt - events@xcite .
the maximum amplitude distributions of the rt pedestals and physics events above the noise - edge of 300 ev were measured .
the corresponding distributions for events @xmath27300 ev were evaluated by interpolations to avoid biased sampling .
the trigger efficiencies depicted in figure [ eff ] correspond to the fractions of the distributions above the discriminator threshold level .
events in coincidence with acv - tags are mostly physics - induced .
the fraction of these events surviving the psd cuts was taken to be the psd efficiency .
this assignment is conservative since the actual efficiency corresponds to the survival fraction of samples _ after _ electronic noise events in accidental coincidence were subtracted , and therefore should be higher . alternatively , under the assumption that the @xmath21fe calibration of figure [ fe55]a would give rise to physics events with a flat spectrum down to the lowest energy relevant to this analysis ( @xmath27100 ev ) , the deviations of the psd - selected events from a flat distribution provided the second measurement .
consistent results were obtained with both approaches , as depicted in figure [ eff ] .
the larger uncertainties of the first method are due to the limited statistics from only the _ in situ _
acv samples .
the efficiencies and their uncertainties adopted for analysis were derived from a best - fit on the combined data set . a threshold of ( 220@xmath010 )
ev was achieved with a psd efficiency of 50% .
n cross - section versus wimp - mass , displaying the ks - ulege limits and those defining the current boundaries@xcite .
the dama - allowed regions@xcite are superimposed .
the striped region is that favored by susy models@xcite .
projected sensitivities of full - scale experiments are indicated as dotted lines .
, width=302 ] the ulege spectrum normalized in cpkkd unit after the crv , acv and psd selections is displayed in figure [ bkgspect ] , showing comparable background as cresst - i@xcite .
listed in table [ summarytable ] are the normalized background rates , indicating that statistical uncertainties dominate over the systematic effects .
the formalisms followed those of ref .
@xcite using standard nuclear form factors , a galactic rotational velocity of @xmath28 and a local wimp density of @xmath29 with maxwellian velocity distribution .
no subtraction of background profiles was made such that the wimp signals can not be larger than the observed event rates .
the unbinned optimal interval method as formulated in ref .
@xcite and widely used by current cdm experiments was adopted to derive the upper limits for the possible @xmath2n event rates . by comparing the observed background in different energy intervals with the expected number of events due to @xmath2n recoils for each @xmath5 ,
the optimal intervals producing the most stringent limits to @xmath11 and @xmath14 were selected .
corrections due to qf , detector resolution and various efficiency factors were incorporated .
the energy dependence of qf in ge was evaluated with the trim software package@xcite .
the uncertainties were taken to be their difference with the statistical best - fit values of the available data@xcite from 254 ev to 200 kev nuclear recoil energy .
exclusion plots on both @xmath30 and @xmath31 planes at 90% confidence level for galactically - bound wimps were then derived , as depicted in figures [ explotsindep ] and [ explotsdep]a , respectively .
the dama - allowed regions@xcite and the current exclusion boundaries@xcite are displayed .
the model - independent approach of refs .
@xcite were adopted to extract limits on the spin - dependent cross - sections .
consistent results were obtained when different @xmath13ge nuclear physics matrix elements@xcite were adopted as input .
the constraints on the effective axial four - fermion @xmath2-proton and @xmath2-neutron spin - dependent couplings@xcite at @xmath5=5 gev are displayed in figure [ explotsdep]b .
the parameter space probed by the @xmath13ge in ulege is complementary to that of the cresst - i experiment@xcite where the @xmath32al target is made up of an unpaired proton instead .
new limits were set by the ks - ulege data in both @xmath11 and @xmath14 for @xmath5@xmath63@xmath16 gev .
the remaining dama low-@xmath5 allowed regions in both interactions were probed and excluded .
the observable nuclear recoils at @xmath5=5 gev and @xmath33=@xmath34(allowed ) and @xmath35(excluded ) are superimposed with the measured spectrum in the inset of figure [ bkgspect ] for illustrations .
it is expected that recent data from the coupp@xcite experiment can place further constraints in the spin - dependent plots of figures [ explotsdep]a&b .
this work extends the bounds on wimps by making measurements in a new observable window of 100 ev@xmath11 kev in a low - background environment .
understanding and suppression of background at this sub - kev region is crucial for further improvement .
measurements are conducted with the ulege at an underground laboratory .
there are recent advances in `` point - contact '' ge detector@xcite which offer potentials of scaling - up the detector mass to the kg - range . preliminary results in dark matter searches were recently reported@xcite .
the mass - normalized external background will be reduced in massive detectors due to self - attenuation@xcite . further reduction in threshold
may be possible with improved junction field - effect - transistors and by correlating signals from both electrodes .
the potential reach of full - scale experiments with 1 kg - year of data and a benchmark background level of 1 cpkkd is illustrated in figures [ explotsindep ] and [ explotsdep]a .
such experimental programs are complementary to the many current efforts on cdm direct searches .
we are grateful to the kims collaboration and the authors of ref .
@xcite for inspiring comments .
this work is supported by the academia sinica pilot project 2004 - 06 , theme project 2008 - 10 , contracts 95 - 2119-m-001 - 028 and 96 - 2119-m-001 - 005 from the national science council , taiwan and 10620140100 from the national natural science foundation , china .
99 m. drees and g. gerbier , j. phys . *
g 33 * , 233 ( 2006 ) , and references therein .
r.j .
gaitskell , annu .
rev .
nucl . part .
sci .
, * 54 * , 315 ( 2004 ) , and references therein . h. h. haber and m. schmitt , j. phys . *
g 33 * , 1105 ( 2006 ) , a. bottino et al .
, phys . rev .
* d 72 * , 083521 ( 2005 ) , and references therein . c. savage , p. gondolo and k. freese , phys
. rev . *
d 70 * , 123513 ( 2004 ) ; p. gondolo and g. gelmini , phys . rev . *
d 71 * , 123520 ( 2005 ) .
r. bernabei et al .
, riv .
nuovo cim .
* 26n1 * , 1 ( 2003 ) .
r. bernabei et al .
, arxiv:0804.2741 ( 2008 ) . x.g .
he et al . ,
mod .
. lett . * a 22 * , 2121 ( 2007 ) .
t. damour and l.m .
krauss , phys .
rev .
lett . * 81 * , 5726 ( 1998 ) ; j.i .
collar , phys .
rev . * d 59 * , 063514 ( 1999 ) .
g. gelmini , a. kusenko and s. nussinov , phys .
rev .
lett . * 89 * , 101302 ( 2002 ) .
g. angloher et al .
, astropart . phys . *
18 * , 43 ( 2002 ) .
h.t .
wong , mod .
phys .
lett . * a 19 * , 1207 ( 2004 ) .
h.t .
wong and h.b .
li , mod .
. lett . * a 20 * , 1103 ( 2005 ) .
h.b .
li et al . , phys .
rev .
lett . * 90 * , 131802 ( 2003 ) ; h.t .
wong et al . , phys . rev . *
d 75 * , 012001 ( 2007 ) .
b. xin et al . , phys . rev . *
d 72 * , 012006 ( 2005 ) .
h.m .
chang et al . , phys . rev . * d 75 * ,
052004 ( 2007 ) . h.t .
wong et al . , j. phys .
. ser . * 39 * , 266 ( 2006 ) .
j.d . lewin and p.f .
smith , astropart . phys .
* 6 * , 87 ( 1996 ) .
j. lindhard et al .
, dan .
vid .
, medd . * 33 * , 10 ( 1963 ) ; j.f .
ziegler , j.p .
biersack and u. littmark , _ the stopping and range of ions in solid _ , pergamon press ( 1985 ) . manufacturer : canberra industries , inc .
w.p .
lai et al . , nucl .
instrum .
methods * a 465 * , 550 ( 2001 ) .
f.t .
avignone iii , p.s .
barbeau and j.i .
collar , arxiv : 0806.1314 ( 2008 ) .
p.j .
statham , x - ray spect .
* 6 * , 94 ( 1977 ) .
h.b .
li et al . , nucl .
instrum .
methods * a 459 * , 93 ( 2001 ) .
s. yellin , phys . rev . *
d 66 * , 032005 ( 2002 ) .
j.f .
ziegler , _ transport of ions in matter _ , http://www.srim.org ( 1998 ) .
k.w .
jones and h.w .
kraner , phys .
rev .
* c 1 * , 125 ( 1971 ) ; k.w .
jones and h.w .
kraner , phys . rev . * a 11 * , 1347 ( 1975 ) ; t. shutt et al . ,
phys .
rev .
lett .
* 69 * , 3425 ( 1992 ) ; y. messous et al .
, astropart . phys . * 3 * , 361 ( 1995 ) .
r. bernabei et al . , phys . lett . * b 389 * , 757 ( 1996 ) ; m. barnabe - heider et al .
, phys . lett .
* b 624 * , 186 ( 2005 ) ; j. angle et al .
, phys .
rev .
lett . * 100 * , 021303 ( 2008 ) ; z. ahmed et al . , arxiv:0802.3530 ( 2008 ) .
a. bottino et al . , phys . lett . * b 402 * , 113 ( 1997 ) .
d.r .
tovey et al . , phys . lett .
* b 488 * , 17 ( 2000 ) . m.t .
ressell et al . , phys . rev . *
d 48 * , 5519 ( 1993 ) ; v.i .
dimitrov , j. engel and s. pittel , phys . rev . * d 51 * , r291 ( 1995 ) .
e. behnke et al . , science * 319 * , 933 ( 2008 ) .
p.a .
barbeau , j.i . collar and o. tench ,
jcap * 09 * , 009 ( 2007 ) .
c.e .
aalseth et al . ,
arxiv:0807.0879 ( 2008 ) . h.b .
li , s.t .
lin , s.k .
lin , a.k .
soma , h.t .
wong@xmath36 + ( texono collaboration ) + @xmath36 corresponding author : [email protected] the authors in a recent paper @xcite raised questions on our estimates of trigger efficiency in the evaluation of the constraints on wimp dark matter in the low - mass ( @xmath37 ) domain with a ultra - low energy germanium ( ulege ) detector at a threshold of 220 ev @xcite .
the discrepancy originates from some misleading terminology in ref .
@xcite .
we address the issue in details in this reply .
we show how the trigger efficiencies are derived and demonstrate that the results of ref .
@xcite are correct . |
the classical helly selection principle ( @xcite ) states that _ a bounded sequence of real valued functions on the closed interval , which is of uniformly bounded ( jordan ) variation , contains a pointwise convergent subsequence whose limit is a function of bounded variation_. this theorem and its recent generalizations for real valued functions and metric space valued maps of one real variable ( @xcite ) have numerous applications in different branches of analysis ( e.g. , @xcite and references therein ) .
extensions of the helly theorem for functions and maps of several real variables heavily depend upon notions of ( bounded ) variation used for these maps , which generalize different aspects of the classical jordan variation of univariate functions and which are known to be quite numerous in the literature ( e.g. , @xcite , and these references are far from being exhaustive on the topic ) . under some approaches to the multidimensional variation ( @xcite ) involving integration procedures
helly - type theorems are rather concerned with the _ almost everywhere _ convergence of extracted subsequences , and no stronger convergence can be expected in this case , but this convergence is far too weak for certain applications ( such as those from @xcite ) . on the other hand , there are definitions of the notion of variation for _ real valued _ functions of several variables ( @xcite ) , which go back to vitali @xcite , hardy @xcite and krause @xcite , such that a complete analogue of the helly theorem holds with respect to the _ pointwise _ convergence of extracted subsequences .
these counterparts of helly s theorem are based on the notion of a ( totally ) _
real valued function of several variables @xcite and an appropriate generalization of jordan s decomposition theorem when a function of bounded variation is represented as the difference of two monotone functions .
the aim of this paper is twofold .
first , we study properties of the _ total variation _ of metric semigroup valued maps of several variables in the approach of vitali , hardy and krause introduced by the first author in @xcite , which extends the classical notion of jordan s total variation for maps of one variable and the notions of the total variation in the sense of hildebrandt @xcite for real valued functions of two variables and leonov @xcite for real valued functions of any finite number of variables .
second , we present two variants of a helly - type _ pointwise _ selection principle for metric semigroup valued maps and maps with values in a reflexive separable banach space .
the main difficulty that we overcome is that for metric semigroup valued maps there is no counterpart of jordan s decomposition theorem , and we have to develop a completely different technique , whose two - dimesional variant is given in @xcite .
the paper is organized as follows . in section [
s : defs ] we present necessary definitions and our two main results , theorems [ t : first ] and [ t : weak ] . in order to get to the proofs of these results as quick as possible , in section [ s : ingred ]
we collect all main ingredients and auxiliary known facts needed for their proofs . in section
[ s : proofs ] we prove theorems [ t : first ] and [ t : weak ] .
the remaining sections [ s : prb][s : pre ] contain proofs of the results exposed in section [ s : ingred ] and used in the proofs of the main theorems .
throughout the paper we adopt and follow the vitali - hardy - krause approach to the notion of variation for maps of several variables in the multiindex notation initiated in @xcite and developed in detail in @xcite ( equivalent approaches in different notation for real functions can be found in @xcite ) .
let @xmath20 and @xmath21 stand for the sets of positive and nonnegative integers , respectively , and @xmath11 .
given @xmath22 , we write @xmath23 for the coordinate representation of @xmath24 , and set @xmath25 , and @xmath26 is defined similarly .
the inequality @xmath27 will be understood componentwise , i.e. , @xmath28 for all @xmath29 , and a similar meaning applies to @xmath30 , @xmath31 , @xmath32 and @xmath33 .
if @xmath27 or @xmath31 , we denote by @xmath34 the rectangle @xmath35=[x_1,y_1]\times\cdots\times[x_n , y_n]$ ] . elements of the set @xmath36 are as usual said to be _ multiindices _ and denoted by greek letters and , given @xmath37 and @xmath38 , we set @xmath39 ( the order of @xmath40 ) and @xmath41 .
the @xmath12-dimensional multiindices @xmath42 and @xmath43 will be denoted simply by @xmath44 and @xmath45 , respectively ( actually , the dimension of @xmath44 and @xmath45 will be clear from the context ) .
we also put @xmath46 ( the set of ` even ' multiindices ) and @xmath47 ( the set of ` odd ' multiindices ) . for elements from the set @xmath48
we simply write @xmath49 . the domain of ( almost )
all maps under consideration will be a rectangle @xmath8 with fixed @xmath50 , @xmath3 , called the _
basic rectangle_. the range of maps will be a _
metric semigroup _
@xmath51 , i.e. , @xmath52 is a metric space , @xmath53 is an abelian semigroup with the operation of addition @xmath54 , and @xmath55 is translation invariant : @xmath56 for all @xmath57 .
a nontrivial example of a metric semigroup is as follows ( @xcite ) : let @xmath58 be a real normed space and @xmath14 be the family of all nonempty closed bounded convex subsets of @xmath59 equipped with the hausdorff metric @xmath55 given by @xmath60 , where @xmath61 and @xmath62 . given @xmath61 , defining @xmath63 as the closure in @xmath59 of the minkowski sum @xmath64 we find that the triple @xmath65 is a metric semigroup .
given @xmath66 , we define the _ vitali - type @xmath12-th mixed ` difference ' _ of @xmath4 on a subrectangle @xmath67 , where @xmath68 and @xmath27 , by ( cf .
@xcite ) @xmath69 for example , for the first three dimensions we have : if @xmath9 , then @xmath70 and @xmath71 , and so , @xmath72 ; if @xmath10 , then @xmath73 and @xmath74 , and so , @xmath75 if @xmath76 , then @xmath77 and @xmath78 @xmath79 , and so , @xmath80 \!&\!&\quad\,\,f(y_1,y_2,y_3)+f(y_1,x_2,x_3)+f(x_1,y_2,x_3)+f(x_1,x_2,y_3)\bigr ) \end{aligned}\ ] ] ( one may draw corresponding pictures to see the points where @xmath4 is evaluated at the left and right hand places of @xmath81 ) .
[ rema ] formally , the value @xmath82 from is defined for @xmath27 . now if @xmath68 , @xmath31 and @xmath83 , then the right - hand side in is equal to zero for any map @xmath84 .
in fact , if @xmath85 for some @xmath86 , then @xmath87 in order to see this , given @xmath88 , we set @xmath89 and note that @xmath90 and , moreover , the map @xmath91 is a bijection between @xmath92 and @xmath93 .
it remains to take into account that @xmath94 for all @xmath95 , because @xmath96 the _ vitali - type @xmath12-th variation _ ( @xcite ) of @xmath97 is defined by @xmath98}^{x[\sigma]}\bigr),\ ] ] the supremum being taken over all multiindices @xmath99 and all _ net partitions _
of @xmath8 of the form @xmath100\}_{\sigma=0}^\kappa$ ] , where points @xmath101=(x_1(\sigma_1),\dots , x_n(\sigma_n))$ ] from @xmath8 are indexed by @xmath102 with @xmath103 and satisfy the conditions : @xmath104=a$ ] , @xmath105=b$ ] and @xmath106<x[\sigma]$ ] for all @xmath107 ( in other words , a net partition @xmath108 is the cartesian product of ordinary partitions of closed intervals @xmath109 $ ] , @xmath110 ) .
note that all rectangles @xmath111}^{x[\sigma]}$ ] of a net partition are non - degenerated , non - overlapping and their union is @xmath8 . in order to define the notion of the total variation of a map @xmath97
we need the notion of variation of @xmath4 of order less than @xmath12 .
following @xcite , we define the _ truncation of a point @xmath38 by a multiindex @xmath49 _ by @xmath112 , and set @xmath113 .
clearly , @xmath114 and @xmath115 , and if @xmath16 , then @xmath116 .
for example , if @xmath117 and @xmath118 , we have @xmath119 and @xmath120\times[a_4,b_4]$ ] .
given @xmath97 and @xmath121 , we define the _ truncated map _ @xmath122 _ with the base at @xmath123 _ by @xmath124 for all @xmath16 .
it follows that @xmath125 depends only on @xmath126 variables @xmath127 $ ] , for which @xmath128 , and the other variables remain fixed and equal to @xmath129 when @xmath130 . in the above example
we get @xmath131 for @xmath132\times[a_4,b_4]$ ] .
now , given @xmath97 and @xmath49 , the function @xmath133 with the base at @xmath134 depends only on @xmath126 variables , and so , making use of the definitions ( [ e : vn ] ) and ( [ e : mdn ] ) with @xmath12 replaced by @xmath126 , @xmath4 replaced by @xmath135 and @xmath8 replaced by @xmath136 , we get the notion of the ( _ hardy - krause - type _
@xcite ) _ @xmath126-th variation _ of @xmath4 , which is denoted by @xmath137 .
the _ total variation _ of @xmath97 in the sense of hildebrandt ( @xcite , ( * ? ? ? * iii.6.3 ) , @xcite if @xmath10 ) and leonov ( @xcite if @xmath11 ) is defined by @xmath138 the summations here and throughout the paper being taken over _
@xmath12-dimensional _ multiindices in the ranges specified under the summation sign . for the first three dimensions
@xmath139 we have , respectively , @xmath140$},\\ \mbox{tv}(f , i_a^b)&=&v_{a_1}^{b_1}\bigl(f(\,\cdot\,,a_2)\bigr ) + v_{a_2}^{b_2}\bigl(f(a_1,\,\cdot\,)\bigr)+v_2(f , i_{a_1,a_2}^{b_1,b_2}),\\ \mbox{tv}(f , i_a^b)&=&v_{a_1}^{b_1}\bigl(f(\,\cdot\,,a_2,a_3)\bigr ) + v_{a_2}^{b_2}\bigl(f(a_1,\,\cdot\,,a_3)\bigr ) + v_{a_3}^{b_3}\bigl(f(a_1,a_2,\,\cdot\,)\bigr)\\[1pt ] & & \quad+v_2\bigl(f(\,\cdot\,,\,\cdot\,,a_3),i_{a_1,a_2}^{b_1,b_2}\bigr ) + v_2\bigl(f(\,\cdot\,,a_2,\,\cdot\,),i_{a_1,a_3}^{b_1,b_3}\bigr)\\[1pt ] & & \quad+v_2\bigl(f(a_1,\,\cdot\,,\,\cdot\,),i_{a_2,a_3}^{b_2,b_3}\bigr ) + v_3(f , i_{a_1,a_2,a_3}^{b_1,b_2,b_3 } ) . \end{aligned}\ ] ] we denote by @xmath141 the space of all maps @xmath97 of _ finite _ ( or bounded ) _ total variation _ ( [ e : tv ] ) .
recall that a sequence @xmath142 of maps from @xmath8 into @xmath14 is said : ( a ) to _ converge pointwise _ on @xmath8 to a map @xmath97 if @xmath143 as @xmath144 for all @xmath16 ; ( b ) to be _ pointwise precompact _
( on @xmath8 ) provided the closure in @xmath14 of the set @xmath15 is compact for all @xmath16 .
our first main result , to be proved in section [ s : proofs ] , is the following helly - type pointwise selection principle in the space @xmath141 : [ t : first ] a pointwise precompact sequence @xmath145 of maps from the rectangle @xmath8 into a metric semigroup @xmath51 such that @xmath146 contains a subsequence which converges pointwise on @xmath8 to a map @xmath147 such that @xmath148 .
this result was announced in @xcite .
it contains as particular cases the results of ( * ? ? ?
* iii.6.5 ) and @xcite ( @xmath10 and @xmath149 ) , @xcite ( @xmath11 and @xmath149 ) and @xcite ( @xmath10 and @xmath14 is a metric semigroup ) . our second main result ( theorem [ t : weak ] below )
is concerned with a weak analogue of theorem [ t : first ] taking into account certain specific features when the values of maps under consideration lie in a reflexive separable banach space .
let @xmath18 be a normed linear space over the field @xmath150 or @xmath151 and @xmath19 be its _ dual _
, i.e. , @xmath152 , the space of all continuous linear functionals on @xmath14 .
it is well - known that @xmath19 is a banach space under the norm @xmath153 , @xmath154 .
the natural duality between @xmath14 and @xmath19 is determined by the bilinear functional @xmath155 defined by @xmath156 for all @xmath157 and @xmath154 , so that @xmath158 , where @xmath159 is the absolute value in @xmath160 . recall that a sequence @xmath161 _ converges weakly in @xmath14 _ to an element @xmath157 ( in symbols , @xmath162 in @xmath14 ) if @xmath163 in @xmath160 as @xmath144 for all @xmath154 ; if this is the case then it is known that @xmath164 . since
a normed linear space @xmath18 is a metric semigroup , the notions of the vitali - type @xmath12-th variation , @xmath126-th variation for @xmath49 and the total variation of a map @xmath97 are introduced as above with respect to the induced metric @xmath165 , @xmath166 .
[ t : weak ] suppose @xmath18 is a reflexive separable banach space with separable dual @xmath19 and @xmath145 is a sequence of maps from @xmath8 into @xmath14 .
if @xmath145 satisfies condition from theorem [ t : first ] and @xmath167 then there exists a subsequence of @xmath145 , again denoted by @xmath145 , and a map @xmath168 satisfying @xmath148 such that @xmath169 this theorem will be proved in section [ s : proofs ] .
it is an extension of a weak selection principle from ( * ? ? ?
* chapter 1 , theorem 3.5 ) for maps of bounded jordan variation of one real variable .
more comments and remarks on theorems [ t : first ] and [ t : weak ] can be found in section [ s : proofs ] .
in this section we collect main ingredients of the proof of theorem [ t : first ] . these are relations between mixed differences of all orders and properties of the total variation ( [ e : tv ] ) . for real valued functions of @xmath12 variables
the main properties of mixed differences of all orders were elaborated in @xcite and for metric semigroup valued maps of two variables in @xcite . for our purposes we need their variants in the multiindex notation , as presented in @xcite with @xmath149 , for maps of @xmath12 variables with values in a metric semigroup .
first , we recall several definitions and results for real valued functions .
a function @xmath170 is said to be _ totally monotone _
( cf . , e.g. , ( * ? ? ? * part ii , section 3 ) , @xcite ) if , given @xmath49 and @xmath68 with @xmath31 , we have : @xmath171 for real valued functions the sum in ( with no factor @xmath172 ) is called the _ @xmath126-th mixed difference _ ( in the sense of vitali , hardy and krause ) of @xmath173 on the rectangle @xmath174 and denoted by @xmath175 ( however , note the difference with in the general case ) . in this case
vitali @xmath12-th variation _
@xmath176 of @xmath177 on @xmath8 is defined as in with the mixed difference at the right - hand side of replaced by @xmath178}^{x[\sigma]})\bigr|$ ] .
the other definitions related to the bounded variation context remain the same as above , and so , we keep the same notation for real valued functions as well .
denote by @xmath179 the set of all totally monotone real valued functions on @xmath8 .
it is known ( e.g. , from the references above ) that if @xmath180 , then @xmath181 , the value at the left - hand side of is equal to @xmath182 , @xmath183 and @xmath184 for all @xmath68 with @xmath31 .
the following helly selection principle in the class @xmath179 is due to leonov ( * ? ? ?
* lemma 3 ) ( for totally monotone functions of two variables it was established in ( * ? ? ?
* iii.6.5 ) and ( * ? ? ? * theorem 3.1 ) ) : [ t : a ] @xmath185an infinite uniformly bounded family of totally monotone functions on @xmath8 contains a sequence , which converges pointwise on @xmath8 to a function from @xmath186 .
it was shown in ( * ? ? ?
* corollary 2 ) that the linear space @xmath187 equipped with the norm @xmath188 , @xmath181 , is a banach space .
this assertion was refined in ( * ? ? ?
* part i , theorem 1 ) : the space @xmath187 is a banach algebra with respect to the norm @xmath189 , and @xmath190 for all @xmath191 . theorem [ t : a ] implies helly s selection principle in the space @xmath187 ( * ? ? ? * theorem 4 ) : _ an infinite family of functions from @xmath192 , which is bounded under the norm @xmath189 , contains a pointwise convergent sequence , whose pointwise limit belongs to @xmath192_. the crucial observation in the proof of this result is that , given @xmath181 , if we set @xmath193 and @xmath194 , @xmath16 , then ( ( * ? ? ?
* theorem 3 ) ) @xmath195 , and jordan s decomposition holds : @xmath196 on @xmath8 ; then theorem [ t : a ] applies to the uniformly bounded families of functions @xmath197 and @xmath198 in the standard way .
now let us consider the case of maps @xmath97 of finite total variation valued in a metric semigroup @xmath51 .
clearly , there is no counterpart of jordan s decomposition for these maps , and so , in order to prove theorem [ t : first ] , we ought to argue in a completely different way .
it will be seen later that , along with theorem [ t : a ] , the following four theorems [ t : b ] through [ t : e ] are the main ingredients in the proof of theorem [ t : first ] ( in a certain sense replacing the arguments involving jordan s decomposition ) . [
t : b ] if @xmath168 , @xmath68 and @xmath31 , then @xmath199 this theorem will be proved in section [ s : prb ] .
it is a generalization of the well - known property of maps of bounded jordan variation of one variable and a counterpart of leonov s ( in)equalities established in ( * ? ?
* theorem 2 and corollary 5 ) for real valued functions of @xmath12 variables ( cf .
also ( * ? ? ?
* part i , lemma 6 and ( 3.5 ) ) ) .
the inequalities in theorem [ t : b ] are also known for metric semigroup valued maps of two variables @xcite . however , in the general case theorem [ t : b ] needs a different proof as compared to the cases of maps of one or two variable(s ) or @xmath149 .
[ t : c ] if @xmath168 , @xmath68 , @xmath31 , and @xmath200 , then @xmath201 [ t : d ] if @xmath168 and if we set @xmath202 , @xmath16 , then for @xmath203 , called the
_ total variation function of @xmath4 _ , we have : @xmath204 and @xmath205 .
these two theorems are extensions of two more properties of the jordan variation for maps of one variable ; in this case is actually the equality known as the _ additivity _ of jordan s variation ( e.g. , ( * ? ? ?
* theorem @xmath206 ) ) . on the other hand , theorem [ t : c ] is a counterpart of chistyakov s inequality ( * ? ? ? * part ii , lemma 8) and theorem [ t : d ] is a generalization of theorem 3 from @xcite and corollary 11 from ( * ? ? ?
* part ii ) given for @xmath149 . for metric semigroup valued maps of two variables
* inequalities ( 11 ) , ( 13 ) and theorem 1 ) .
the proof of theorem [ t : c ] is identical with the proof of lemma 8 from ( * ? ? ?
* part ii ) and the proof of theorem [ t : d ] is identical with the proofs of lemma 9 and corollaries 10 and 11 from ( * ? ? ?
* part ii ) when @xmath149 , and so , they are omitted
. however , it is to be noted that these proofs rely on ( 1 ) equality ( 3.2 ) from ( * ? ? ?
* part i , lemma 5 ) , ( 2 ) lemma 7 from ( * ? ? ?
* part i ) , and ( 3 ) the well - known property of the _ additivity _ of @xmath126-th variation @xmath207 for each for real valued functions of @xmath12 variables . for metric semigroup valued maps assertions ( 1 ) , ( 2 ) and
( 3 ) need a proper interpretation and different , more subtle and hard proofs .
their respective counterparts are presented below as lemmas [ l : un ] , [ l : deux ] and [ l : troix ] . in the first lemma and throughout the paper we use the following short notations : given @xmath49 , the sum over ` @xmath208 ' denotes the sum over ` @xmath95 s.t .
@xmath209 ' , where ` s.t . ' is the usual abbreviation for ` such that ' , and a similar convention applies to the sum over ` @xmath210 ' .
[ l : un ] if @xmath97 , @xmath68 , @xmath31 , @xmath121 and @xmath49 , then @xmath211 in particular , if @xmath134 or @xmath212 , we have , respectively , @xmath213 @xmath214 the proof of lemma [ l : un ] is the same as in ( * ? ? ? * part i , lemma 5 ) ( details are omitted ) : we have to note only that @xmath215 and @xmath216 is even ( odd ) if and only if there exists a unique @xmath217 s.t.@xmath209 , @xmath218 is even ( odd , respectively ) and @xmath219 , and apply definition ( [ e : mdn ] ) where @xmath12 is replaced by @xmath126 . since the total variation is defined via truncated maps with the base at the point @xmath220 , in our next lemma we present a counterpart of chistyakov s equality ( * ? ? ?
* part i , lemma 7 ) exhibiting the relation between the mixed difference @xmath221 and certain mixed differences of maps @xmath222 with the base at @xmath220 for some @xmath223 .
[ l : deux ] if @xmath97 , @xmath49 and @xmath68 with @xmath31 , then @xmath224 the proof of lemma [ l : deux ] is postponed until section [ s : prdeux ] .
the _ additivity _ property of @xmath126-th variation @xmath207 for each @xmath49 , to be proved in section [ s : prtroix ] , is expressed in the following [ l : troix ] given @xmath97 , @xmath68 with @xmath27 , @xmath121 and @xmath49 , if @xmath225\}_{\sigma=0}^\kappa$ ] is a net partition of @xmath34 , then @xmath226}^{x[\sigma]}\lfloor\alpha),\ ] ] where the summation is taken only over those @xmath227 in the range @xmath228 with @xmath29 , for which @xmath128 .
the final ingredient in the proof of theorem [ t : first ] is the _ sequential lower semicontinuity _ of the total variation @xmath229 to be established in section [ s : pre ] : [ t : e ] if a sequence of maps @xmath145 from @xmath8 into @xmath14 converges pointwise on @xmath8 to a map @xmath97 , then @xmath230 .
now we are in a position to prove theorems [ t : first ] and [ t : weak ] .
we divide the proof into four steps for clarity . \1 .
we apply the induction argument on the dimension @xmath12 of the basic rectangle @xmath231 . for @xmath9 theorem [ t : first ] was established in ( * ? ? ?
* theorem 5.1 ) ( and refined in ( * ? ? ?
* theorem 1 ) and ( * ? ? ?
* theorem 1.3 ) ) in the case when @xmath52 is an arbitrary metric space , and for @xmath10 it was proved in ( * ? ? ? * theorem 2 ) . now
, suppose that @xmath232 and theorem [ t : first ] is already established for domain rectangles of dimension @xmath233 .
given @xmath234 , we let @xmath235 be the total variation function of @xmath236 on @xmath8 , i.e. , @xmath237 for all @xmath16 . by theorem [ t : d ] and condition , the sequence @xmath238 is uniformly bounded ( by @xmath239 ) , and so , by theorem [ t : a ] , there exist a subsequence of @xmath240 and the corresponding subsequence of @xmath145 , again denoted as the whole sequences @xmath240 and @xmath145 , respectively , and a function @xmath241 s.t .
@xmath242 it is known ( @xcite , ( * ? ? ? *
iii.5.4 ) , @xcite ) that the set of discontinuity points of any totally monotone function on @xmath231 lies on at most a countable set of hyperplanes of dimension @xmath243 parallel to the coordinate axes .
given @xmath29 , denote by @xmath244 the union of the set of all rational points of the interval @xmath109 $ ] , the two - point set @xmath245 and the set of those points @xmath246 $ ] , for which the hyperplane @xmath247\times\cdots\times[a_{i-1},b_{i-1}]\times\{z_i\}\times[a_{i+1},b_{i+1 } ] \times\cdots\times[a_n , b_n]\ ] ] contains points of discontinuity of @xmath248 .
clearly , the sets @xmath249 $ ] are countable and dense in @xmath109 $ ] , and so , we may assume that @xmath250 .
\2 . in order to apply the induction hypothesis
, we need an estimate on the @xmath251-dimensional total variation of any function @xmath252 from the sequence @xmath145 ` over the hyperplane ' in the sense to be made precise below .
this is done as follows .
let us fix @xmath29 and set @xmath253 , where @xmath44 is the @xmath254-th coordinate of @xmath255 and the other coordinates of @xmath255 are equal to @xmath45 .
note that @xmath256 .
given @xmath257 , we put @xmath258 the map @xmath259 with the base at @xmath260 , truncated by @xmath255 , is defined on the @xmath251-dimensional rectangle @xmath261 and given by : if @xmath16 , then @xmath262 and @xmath263 the @xmath251-dimensional total variation of @xmath264 on @xmath265 is equal to @xmath266 where the summation is taken over @xmath251-dimensional multiindices @xmath267 s.t .
@xmath268 , i.e. , @xmath269 ( this is the only instance and exception when the summation is over @xmath251-dimensional multiindices ) . given @xmath270 , we set @xmath271 , where @xmath44 occupies the @xmath254-th place , and note that @xmath272 .
we have @xmath273 in fact , given @xmath16 , we find @xmath274 and @xmath275 & = & f_{1^i}^{{\overline{a}}}\bigl((a\lfloor1^i)+({\overline{\alpha}}\,\lfloor1^i)[(x\lfloor1^i)-(a\lfloor1^i)]\bigr ) \nonumber\\[2pt ] & = & f_{1^i}^{{\overline{a}}}\bigl([a+{\overline{\alpha}}\,(x - a)]\lfloor1^i\bigr)\nonumber\\[2pt ] & = & f\bigl({\overline{a}}+1^i[a+{\overline{\alpha}}\,(x - a)-{\overline{a}}]\bigr).\label{e : xova } \end{aligned}\ ] ] since @xmath276 and @xmath277 , we get @xmath278&=&{\overline{a}}+1^i(a-{\overline{a}})+1^i\,{\overline{\alpha}}\,(x - a)\\[2pt ] & = & { \overline{a}}+{\overline{\alpha}}\,(x - a)={\overline{a}}+{\overline{\alpha}}\,(x-{\overline{a } } ) , \end{aligned}\ ] ] and so , the value is equal to @xmath279 it follows that the @xmath126-th variation at the right - hand side of is equal to @xmath280 noting that the set @xmath281 is bijective to the set of those @xmath282 , for which @xmath283 , and applying theorem [ t : c ] with @xmath284 , @xmath285 and @xmath286 , we get : @xmath287 & \le&\mbox{tv}(f , i_a^{{\overline{a}}+1^i(b-{\overline{a}})})-\mbox{tv}(f , i_a^{{\overline{a } } } ) \le\mbox{tv}(f , i_a^b).\label{e : est } \end{aligned}\ ] ] thus , given @xmath234 and @xmath29 , setting back @xmath252 , by virtue of , and , we find , for all @xmath257 and @xmath288 : @xmath289 \3 . now , we make use of the diagonal processes .
for @xmath290 and @xmath291 the sequence @xmath292 satisfies the uniform estimate on the rectangle @xmath293 of dimension @xmath243 and , since each map from this sequence is of the form with @xmath294 , then it follows from the assumptions of theorem [ t : first ] that the sequence under consideration is pointwise precompact on @xmath293 . by the induction hypothesis
, the sequence @xmath145 contains a subsequence , denoted by @xmath295 , s.t .
@xmath296 converges pointwise on @xmath293 to a map from @xmath293 into @xmath14 of @xmath251-dimensional finite total variation on @xmath293 . since , by , @xmath297 with @xmath298 and @xmath127 $ ] for @xmath299 , then the pointwise convergence above means , actually , that the sequence @xmath295 converges pointwise on the hyperplane @xmath300\times\cdots\times[a_n , b_n]$ ] .
inductively , if @xmath301 and a subsequence @xmath302 of @xmath145 , which is pointwise convergent on @xmath303 , is already chosen , then the sequence @xmath304 satisfies the uniform estimate on the rectangle @xmath293 , where @xmath236 is replaced by @xmath305 and @xmath306by @xmath307 . moreover , since , as above , the sequence is pointwise precompact on @xmath293 , then , by the induction hypothesis , there exists a subsequence @xmath308 of @xmath302 s.t .
@xmath309 converges pointwise on @xmath293 as @xmath144 to a map from @xmath293 into @xmath14 of @xmath251-dimensional finite total variation on @xmath293 .
again , as above , this pointwise convergence means that the sequence @xmath308 converges pointwise on the hyperplane @xmath310 and , as a consequence , on the set @xmath311 as well .
we infer that the diagonal sequence @xmath312 , which is a subsequence of the original sequence @xmath145 , converges pointwise on the set @xmath313 ; in fact , given @xmath314 , we have @xmath315 for some @xmath316 and @xmath317 , and so , noting that @xmath318 is a subsequence of @xmath308 , we find that @xmath319 converges in @xmath14 as @xmath144 .
let us denote the diagonal sequence @xmath312 extracted in the last paragraph again by @xmath145 .
then we let @xmath320 , @xmath321 and , beginning with the sequence @xmath322 , apply the above arguments of this step . doing this , we will end up with a diagonal sequence , a subsequence of the original sequence @xmath145 , again denoted by @xmath145 , which converges pointwise on @xmath323 .
now suppose that for some @xmath324 we have already extracted a ( diagonal ) subsequence of @xmath145 , again denoted by @xmath145 , which converges pointwise on the set @xmath325 .
then we let @xmath326 and apply the above arguments of this step to the sequence @xmath327 : a subsequence of the original sequence @xmath145 converges pointwise on the set @xmath328 .
in this way after finitely many steps we obtain a subsequence of the original sequence @xmath145 , again denoted by @xmath145 , which converges pointwise on the set @xmath329 .
\4 . finally , let us show that the sequence @xmath145 from the end of step 3 converges at each point @xmath330 .
note that @xmath331 is a point of continuity of the function @xmath248 from s.t .
its coordinates @xmath332 are irrational for all @xmath29 . due to the density of @xmath333 in @xmath8 , the continuity of @xmath248 at @xmath331 and properties of totally monotone functions , given @xmath334 , there exists
@xmath335 with @xmath27 s.t.@xmath336 . by virtue of , choose a number @xmath337 s.t .
@xmath338 and @xmath339 for all @xmath340 . by theorems [ t : b ] and
[ t : c ] with @xmath341 , for all @xmath340 we have : @xmath342 & \le&|\nu_j(y)-\nu(y)|+(\nu(y)-\nu(x))+|\nu(x)-\nu_j(x)|\le3\varepsilon .
\end{aligned}\ ] ] since @xmath343 and , as it was shown in step 3 , the sequence @xmath344 is convergent in @xmath14 , it is cauchy , and so , there exists a number @xmath345 s.t .
@xmath346 for all @xmath347 and @xmath348 .
it follows that if @xmath349 , @xmath350 and @xmath351 , then we have : @xmath352 & \le&3\varepsilon+\varepsilon+3\varepsilon=7\varepsilon .
\end{aligned}\ ] ] thus , the sequence @xmath353 is cauchy in the metric space @xmath14 , and so , since it is also precompact by the assumption , it is convergent in @xmath14 .
it follows from here and the end of step 3 that the sequence @xmath353 converges in @xmath14 at each point @xmath354 , i.e. , the sequence @xmath145 , which is a subsequence of the original sequence @xmath145 , converges pointwise on @xmath8 . let us denote the pointwise limit of @xmath145 by @xmath97 .
then , by virtue of theorem [ t : e ] and assumption , we find @xmath355 and so , @xmath356 .
this completes the proof of theorem [ t : first ] .
[ r : cb ] in theorem [ t : first ] the precompactness of the sets @xmath344 at all points @xmath16 can not be replaced by the closedness and boundedness even at a single point of @xmath8 .
the corresponding examples for maps of one variable are constructed in ( * ? ? ?
* section 3 ) , ( * ? ? ?
* section 5 ) and ( * ? ? ? * section 1 ) and can be easily adapted for maps of several variables .
the proof is adapted for the situation under consideration from the proof of theorem 7 from @xcite .
\1 . in this step
we show that , given @xmath234 and @xmath154 , we have : @xmath357 where the function @xmath358 is given by @xmath359 , @xmath16 , and @xmath239 is the constant from .
in fact , given @xmath49 and @xmath68 with @xmath27 , by virtue of where @xmath360 is replaced by the absolute value @xmath361 in @xmath160 and later on by the norm in @xmath14 , we get : @xmath362 & = & \mbox{md}_{|\alpha|}((f_j)_\alpha^a , i_x^y\lfloor\alpha)\|u^*\| .
\end{aligned}\ ] ] it follows that if @xmath100\}_{\sigma=0}^\kappa$ ] is a net partition of @xmath8 , then @xmath363\lfloor\alpha\}_{\sigma\lfloor\alpha=0}^{\kappa\lfloor\alpha}$ ] is a net partition of @xmath136 , and so , setting @xmath364 $ ] and @xmath365 $ ] in the calculations above , we find @xmath366}^{x[\sigma]}\lfloor\alpha\bigr ) & \le&\sum_{1\lfloor\alpha\le\sigma\lfloor\alpha\le\kappa\lfloor\alpha } \!\!\!\!\!\!\!\!\mbox{md}_{|\alpha|}\bigl((f_j)_\alpha^a , i_{x[\sigma-1]}^{x[\sigma]}\lfloor\alpha\bigr)\|u^*\|\\ & \le&v_{|\alpha|}((f_j)_\alpha^a , i_a^b\lfloor\alpha)\|u^*\| , \end{aligned}\ ] ] the summation over @xmath367 being taken only over those coordinates @xmath227 in the range @xmath228 with @xmath29 , for which @xmath128 .
since @xmath368 is an arbitrary partition of @xmath8 , we get : @xmath369 and so , inequality follows from the definition of the total variation .
moreover , by virtue of , we have : @xmath370 and so , the sequence @xmath371 of functions from @xmath8 into ( metric semigroup ) @xmath160 is pointwise bounded on @xmath8 and , hence , pointwise precompact for each @xmath154 . taking this and into account and applying theorem [ t : first ] to the sequence @xmath371 for any given @xmath154 , we extract a subsequence of @xmath145 , denoted by @xmath372 ( which depends on @xmath373 in general ) , and find a function @xmath374 satisfying @xmath375 s.t .
@xmath376 in @xmath160 as @xmath144 for all @xmath16 .
\2 . making use of the diagonal process and the separability of @xmath19 ,
let us get rid of the dependence of @xmath372 on @xmath154 .
let @xmath377 be a countable dense subset of @xmath19 .
by step 1 , for @xmath378 we get a subsequence @xmath379 of the original sequence @xmath145 and a function @xmath380 satisfying @xmath381 s.t.@xmath382 in @xmath160 for all @xmath16 .
inductively , if @xmath301 and a subsequence @xmath383 of @xmath145 is already chosen , then by virtue of and , we have : @xmath384 and @xmath385 , @xmath16 , for all @xmath234 . by theorem [ t :
first ] , applied to the sequence @xmath386 , there exist a subsequence @xmath387 of @xmath383 and a function @xmath388 satisfying @xmath389 s.t.@xmath390 in @xmath160 as @xmath144 for all @xmath16 . then the diagonal sequence @xmath391 , again denoted by @xmath145 , is a subsequence of the original sequence and satisfies the condition : @xmath392 \3 .
now , given @xmath154 and @xmath16 , let us show that the sequence @xmath393 is cauchy in @xmath160 . taking into account we may assume that @xmath394 for all @xmath316 .
let @xmath334 be arbitrary . by the density of @xmath377 in @xmath19
, there exists @xmath395 s.t.@xmath396 .
by , there exists @xmath397 s.t.@xmath398 for all @xmath399 and @xmath400 .
it follows that for such @xmath401 and @xmath402 we have : @xmath403 & & \quad\,+|\langle f_j(x),u_k^*\rangle-\langle f_{j'}(x),u_k^*\rangle|\\[2pt ] & \le&\|f_j(x)-f_{j'}(x)\|\cdot\|u^*-u_k^*\|+\displaystyle\frac\varepsilon2\\[2pt ] & \le&\displaystyle 2c(x)\frac{\varepsilon}{1 + 4c(x)}+\frac\varepsilon2\le\varepsilon .
\end{aligned}\ ] ] thus , @xmath393 is cauchy in @xmath160 and , hence , there exists an element of @xmath160 denoted by @xmath404 s.t.@xmath405 in @xmath160 as @xmath144 .
in other words , we have shown that for each @xmath154 there exists a function @xmath406 satisfying ( cf .
theorem [ t : e ] and ) @xmath407 ( and so , @xmath374 ) and @xmath408 \4 .
let us prove , i.e. , @xmath409 converges weakly in @xmath14 as @xmath144 for all @xmath16 . by the reflexivity of @xmath14 , we have @xmath410 for all @xmath234 .
define the functional @xmath411 by @xmath412 for all @xmath154 . by virtue of
, we get @xmath413 i.e. , the sequence @xmath414 converges pointwise on @xmath19 to the operator @xmath411 . by the banach - steinhaus uniform boundedness principle , @xmath415 and @xmath416 . setting @xmath417 for all @xmath16 , we find that @xmath97 and @xmath418 for all @xmath154 and @xmath16 , and so , @xmath419 in @xmath14 as @xmath144 for all @xmath16 , which proves .
\5 . it remains to show that @xmath356 and @xmath420 . by
, we have : if @xmath68 with @xmath27 and @xmath49 , then @xmath421 in @xmath14 as @xmath144 , and so , by virtue of and the remarks preceding theorem [ t : weak ] , @xmath422 & = & \liminf_{j\to\infty}\,\mbox{md}_{|\alpha|}((f_j)_\alpha^a , i_x^y\lfloor\alpha).\label{e : wei } \end{aligned}\ ] ] arguing as in step 2 of the proof of theorem [ t : e ] , making use of the inequality , which coincides with ( see p. ) , and taking into account , we get : @xmath423 this completes the proof of theorem [ t : weak ] .
[ r : ca ] note that instead of condition in theorem [ t : weak ] we may assume only that the value @xmath424 is finite .
in fact , it follows from theorem [ t : b ] and condition that , given @xmath16 and @xmath234 , @xmath425
in order to prove theorem [ t : b ] , we need three more lemmas [ l : binco ] , [ l : metsem ] and [ l : funcah ] .
the following equality will be needed in lemma [ l : binco ] ( cf .
* part i , equality ( 3.4 ) ) ) : given two multiindices @xmath426 , we have : @xmath427 where @xmath428 denotes the number of elements in the set @xmath429 and , given @xmath430 , @xmath431 is the usual binomial coefficient ( with @xmath432 ) . also , recall ( cf .
section [ s : defs ] ) that a multiindex @xmath433 is said to be even ( odd ) if @xmath434 ( @xmath435 , respectively ) .
\(a ) since the case @xmath441 is clear , we suppose that @xmath442 . by the binomial formula , @xmath443 , which is equal to @xmath444 and @xmath445
if @xmath446 is even , we change the summation index @xmath447 in these sums and get , respectively : @xmath448 @xmath449 while if @xmath446 is odd , we change the summation index @xmath450 in the first sum and @xmath451 in the second sum and get , respectively : @xmath452 @xmath453 the second equality in ( a ) follows from the equality @xmath454 . . the first equality in lemma [ l : binco](a )
can be written also as @xmath455}^{[(m+1)/2]}c_{m - k}^{\,2i-1-k}=\sum_{i=[(k+1)/2]}^{[m/2]}c_{m - k}^{\,2i - k } , \quad\,\,m\in\mathbb{n},\quad 0\le k\le m-1,\ ] ] where @xmath456 $ ] designates the integer part of @xmath457 . however , we prefer the equality in lemma [ l : binco](a ) since it is more simple and suggestive .
\(b ) by virtue of , the left - hand side of the equality is equal to @xmath458 and the right - hand side of the equality is equal to @xmath459 it remains to put @xmath460 and @xmath461 , note that @xmath462 and apply the equality from the previous assertion ( a ) .
if @xmath51 is a metric semigroup , then , by virtue of the triangle inequality for @xmath55 and the translation invariance of metric @xmath55 on @xmath14 , we have , for all @xmath463 : @xmath464 d(u+u',v+v')&\le&d(u , v)+d(u',v').\label{e : ms.2 } \end{aligned}\ ] ] inequality yields that the addition operation @xmath465 is a continuous map from @xmath466 into @xmath14 . more generally , if @xmath467 , @xmath468 , @xmath469 and @xmath470 as @xmath144 ( convergence of sequences in @xmath14 ) , then @xmath471 .
observe that if @xmath476 for some @xmath316 and @xmath477 , then @xmath478 .
in fact , by the translation invariance of @xmath55 and inequality , we have : @xmath479 applying this observation and equality , we get : @xmath480 in the next lemma we set @xmath485 . also , we stick to the following conventions : ` @xmath486 ' will mean that @xmath487 is omitted in the formula under consideration ( especially in a metric semigroup with no zero ) , and a sum over the empty set is also omitted in any context ( i.e. , @xmath488 ) .
[ l : funcah ] given a map @xmath489 and a multiindex @xmath490 , we have : @xmath491 where @xmath492 if @xmath493 is odd , and @xmath494 if @xmath493 is even , and @xmath495 where @xmath496 if @xmath493 is odd , and @xmath497 if @xmath493 is even .
denote by @xmath498 ( by @xmath499 ) the set of all ` admissible ' @xmath40 s at the left ( right ) hand side of the equality under consideration and , given @xmath500 ( and @xmath501 ) , by @xmath502 ( and by @xmath503)the multiplicity of the term @xmath504 at the left ( and right ) hand sum(s ) .
then the equality can be rewritten as @xmath505 where @xmath506 denotes the sum of terms of the form @xmath504 taken @xmath502 times ( and likewise for @xmath507 ) . in what follows in order to prove , we show that @xmath508 and @xmath509 for all @xmath510 .
let us establish .
we have @xmath513 , i.e. , @xmath514 , and @xmath515 .
the sets @xmath516 and @xmath517 are nonempty ( @xmath518 and @xmath519 ) and @xmath520 . in fact , the inclusion @xmath521 is clear , and so , we let @xmath522 .
since @xmath40 is even , @xmath493 is odd and @xmath523 , there exists @xmath29 s.t.@xmath524 and @xmath525 .
we set @xmath526 .
it follows that @xmath527 , @xmath528 is odd and @xmath209 , and so , @xmath529 .
let us prove . if @xmath534 , then the equality is immediate : the left - hand side is equal to @xmath535 , while the double sum at the right is omitted ( in fact , even @xmath527 implies @xmath536 , and so , no odd @xmath40 s.t.@xmath537 exists ) .
now , if @xmath538 , then @xmath539 and @xmath540 ( disjoint union ) , and @xmath520 .
let @xmath530 . if @xmath531 , then @xmath541 and @xmath542 , and so , by lemma [ l : binco](b ) , @xmath509 . now if @xmath543 , then @xmath544 , and since @xmath496 , then @xmath545 as well .
the conclusion follows as in step 1 .
\3 . in order to prove , we first note that if @xmath546 , then the double sum at the right is omitted and the double sum at the left is equal to @xmath547 .
assume that @xmath548 . then @xmath514 and @xmath549 ( disjoint union ) , and @xmath520
. then @xmath532 and , in particular , @xmath550 . if @xmath543 , then , since @xmath494 , we have @xmath545 , and if @xmath531 , then @xmath551 , and so , by lemma [ l : binco](b ) , @xmath509 .
finally , we prove . since the equality is clear for @xmath546 ( i.e. , ` empty ' equality ) , we assume that @xmath552 .
we have @xmath539 , @xmath553 and @xmath520 . given @xmath530 , we find @xmath531 , @xmath541 and @xmath542 , and so , by lemma [ l : binco](b ) , @xmath509 .
it suffices to prove only the first inequality : the second one follows from the first inequality , and . setting @xmath554 and @xmath555 and taking into account , the first inequality in theorem [ t : b ] can be rewritten equivalently as @xmath556 ( the sum over @xmath557 designates the sum over @xmath49 s.t .
@xmath557 ) , where , given @xmath558 , we set @xmath559 , @xmath560 in order to establish , given integer @xmath561 , we also set @xmath562 and note that @xmath563 suppose that we have already verified equalities if @xmath564 is odd and if @xmath564 is even .
applying lemma [ l : metsem ] , we get inequality , where , by virtue of , we have : @xmath565 now , follows if we sum these inequalities over @xmath566 and take into account .
it remains to verify equalities and .
for this , we apply lemma [ l : funcah ] with @xmath341 and note that @xmath567 .
suppose that @xmath568 is odd . by virtue of
, we have : @xmath569 and by virtue of , we get : @xmath570 which establishes .
now suppose that @xmath568 is even . by ,
we get : @xmath571 and by virtue of , we have : @xmath572 which establishes and completes the proof of theorem [ t : b ] .
[ r : tb ] the left - hand side inequality in theorem [ t : b ] is of interest when @xmath27 .
however , if @xmath31 and @xmath83 , it can be refined in the following way ( cf .
* part i , lemma 6 ) ) : given @xmath68 , @xmath27 , and @xmath200 , we have : @xmath573 in fact , by theorem [ t : b ] , we find @xmath574 where , by virtue of , the mixed difference at the right is equal to @xmath575 if @xmath576 , then @xmath128 and @xmath577 for some @xmath29 , and so , arguing as in remark [ rema ] we find @xmath578 for all even @xmath40 with @xmath209 implying that is equal to zero .
now if @xmath527 , then @xmath579 for any @xmath209 , and so , coincides with the right - hand side of .
first , we show that if @xmath68 , @xmath27 , and @xmath49 , then @xmath950 by virtue of , we have : @xmath951 and a similar equality holds for @xmath236 in place of @xmath4 . applying the inequalities @xmath952 , @xmath463 , and and taking into account the pointwise convergence of @xmath236 to @xmath4 ,
we find @xmath953 & & \quad\,\,\le d\biggl(\sum_{{\mbox{\scriptsize\rm ev\,}}\theta\le\alpha}f_j(\cdots),\sum_{{\mbox{\scriptsize\rm ev\,}}\theta\le\alpha}f(\cdots)\biggr ) + d\biggl(\sum_{{\mbox{\scriptsize\rm od\,}}\theta\le\alpha}f_j(\cdots),\sum_{{\mbox{\scriptsize\rm od\,}}\theta\le\alpha}f(\cdots)\biggr)\nonumber\\ & & \quad\,\,\le\sum_{{\mbox{\scriptsize\rm ev\,}}\theta\le\alpha}d(f_j(\cdots),f(\cdots))+ \sum_{{\mbox{\scriptsize\rm od\,}}\theta\le\alpha}d(f_j(\cdots),f(\cdots))\nonumber\\ & & \quad\,\,=\sum_{0\le\theta\le\alpha}d(f_j(\cdots),f(\cdots))\to0\quad\mbox{as}\quad j\to\infty.\nonumber \end{aligned}\ ] ] [ pg : li ] 2 . in the rest of this proof
we need only the inequality @xmath954 which readily follows from and is applied one more time in the proof of theorem [ t : weak ] ( step 5 ) . if @xmath100\}_{\sigma=0}^\kappa$ ] is a net partition of @xmath8 , then @xmath363\lfloor\alpha\}_{\sigma\lfloor\alpha}^{\kappa\lfloor\alpha}$ ] is net partition of @xmath136 , and so , given @xmath107 , setting @xmath364 $ ] and @xmath365 $ ] in , we find @xmath955}^{x[\sigma]}\lfloor\alpha)&\le & \sum_{1\lfloor\alpha\le\sigma\lfloor\alpha\le\kappa\lfloor\alpha } \liminf_{j\to\infty}\,\mbox{md}_{|\alpha|}((f_j)_\alpha^a , i_{x[\sigma-1]}^{x[\sigma]}\lfloor\alpha)\nonumber\\ & \le&\liminf_{j\to\infty}\sum_{1\lfloor\alpha\le\sigma\lfloor\alpha\le\kappa\lfloor\alpha } \mbox{md}_{|\alpha|}((f_j)_\alpha^a , i_{x[\sigma-1]}^{x[\sigma]}\lfloor\alpha)\nonumber\\ & \le&\liminf_{j\to\infty}\,v_{|\alpha|}((f_j)_\alpha^a , i_{x[\sigma-1]}^{x[\sigma]}\lfloor\alpha).\nonumber \end{aligned}\ ] ] by the arbitrariness of @xmath368 , we infer that @xmath956}^{x[\sigma]}\lfloor\alpha)\le \liminf_{j\to\infty}\,v_{|\alpha|}((f_j)_\alpha^a , i_{x[\sigma-1]}^{x[\sigma]}\lfloor\alpha).\ ] ] we conclude that @xmath957 s. bochner , lectures on fourier integrals ( with an author s supplement on monotonic functions , stieltjes integrals , and harmonic analysis ) , annals of math .
studies , no . 42 , princeton university press , princeton , n.j . , 1959
chistyakov , a banach algebra of functions of several variables of finite total variation and lipschitzian superposition operators .
i and ii , nonlinear anal .
62 ( 3 ) ( 2005 ) 559578 , and 63 ( 1 ) ( 2005 ) 122 .
chistyakov , c. maniscalco , yu.v .
tretyachenko , variants of a selection principle for sequences of regulated and non - regulated functions , in : topics in classical analysis and applications in honor of daniel waterman ( l. de carli , k. kazarian and m. milman , eds . ) world scientific publishing , singapore , 2008 , pp . 4572 .
[ l : hh ] given a map @xmath489 and a multiindex @xmath580 , we have : if @xmath581 is even , then the following two equalities hold : @xmath582 and if @xmath581 is odd , then the following two equalities hold : @xmath583 \1 .
let us prove . if @xmath584 , then @xmath585 is even , and equality is equivalent to the identity @xmath586 . if @xmath536 and if @xmath587 is even , then can be written as @xmath588 which was established in for even @xmath341 .
thus , in what follows we assume that @xmath589 , i.e. , @xmath590 .
we have @xmath591 , where @xmath592 ( so that @xmath593 ) and @xmath594 ( so that @xmath595 ) , and @xmath596 , i.e. , @xmath597 .
we are going to show that @xmath520 .
this equality follows immediately from the definition of @xmath517 and the following two assertions : @xmath598 \theta\in{\mathcal{l}}_2&\iff&\mbox{$\theta$ is even and $ \alpha\lor\theta\ne1$,}\label{e : d.16 } \end{aligned}\ ] ] where @xmath599 ; in particular , and imply that @xmath600 and @xmath601 are disjoint .
let us prove . if @xmath602 , then @xmath603 for some even @xmath604 and , since @xmath581 is even and @xmath605 , then @xmath40 is even , @xmath606 and @xmath607 conversely , if @xmath40 is even and @xmath608 , then @xmath609 or @xmath610 .
setting @xmath611 , we find @xmath603 , where @xmath612 is even and @xmath604 , and so , @xmath602 .
now we establish . if @xmath613
, then @xmath40 is even and there exists odd @xmath614 s.t .
@xmath615 , and so , @xmath616 and @xmath615 imply @xmath617 .
since @xmath618 is odd , @xmath581 is even and @xmath614 , we have @xmath619 .
it follows that @xmath620 and so , @xmath621 .
conversely , if @xmath40 is even and @xmath621 , then there exists @xmath29 s.t .
@xmath622 and @xmath524 . setting @xmath623 with @xmath624 and @xmath625
if @xmath626 , we find @xmath614 , @xmath627 is odd and @xmath615 , and so , @xmath613 . in order to calculate the values @xmath502 and @xmath503 for @xmath530
, we note that , given @xmath628 , we have : @xmath629 in fact , condition @xmath628 is equivalent to condition @xmath630 : @xmath631 and so , if @xmath615 , then @xmath632 , and if @xmath633 , then @xmath634 , and so , @xmath635
. given @xmath529 , by virtue of , we find @xmath636 if @xmath602 , then there exists a unique even @xmath604 s.t.@xmath603 , and so , since @xmath637 , then @xmath638 . at the same time , @xmath639 and so , by the above , @xmath640 as well .
suppose now that @xmath613 .
then , by , @xmath641 or @xmath642 , and so , taking into account and lemma [ l : binco](b ) we find that @xmath643 is equal to @xmath503 .
let us establish .
if @xmath584 , we get an identity , and if @xmath536 and @xmath644 is even , we get equality with even @xmath341 , and so , we suppose that @xmath590 .
we have @xmath591 , where @xmath645 and @xmath646 , and @xmath647 , which , actually , is @xmath648 . we need to verify only that @xmath600 and @xmath601 are nonempty : the rest of the proof of ( including and ) is the same as in step 1 where ` even @xmath40 ' is replaced by ` odd @xmath40 ' . since @xmath655
, there exists @xmath29 s.t .
@xmath622 , and so if we set @xmath623 with @xmath624 and @xmath625 if @xmath626 , then @xmath627 is odd and @xmath614 .
given @xmath656 , @xmath657 , setting @xmath658 with @xmath659 and @xmath660 if @xmath661 , we find @xmath662 is odd and @xmath615 , and so , @xmath613 .
first , we show that @xmath600 and @xmath601 are nonempty . since @xmath649 , @xmath128 for some @xmath29 , and so , setting @xmath658 with @xmath667 and @xmath668
if @xmath626 , we find that @xmath669 and @xmath670 is even , whence @xmath602 . now , since @xmath655 , @xmath622 for some @xmath29 , and
if we set @xmath623 with @xmath624 and @xmath625 if @xmath626 , then we find that @xmath627 is even , @xmath671 is even and @xmath672 , and so , @xmath595 .
second , we assert that and hold ; this will imply that @xmath600 and @xmath601 are disjoint and @xmath520 . in order to prove
, we let @xmath602 .
then @xmath40 is even and @xmath669 , and so , @xmath673 conversely , if @xmath40 is even and @xmath608 , then @xmath609 , and so , @xmath674 implying @xmath675 and @xmath602 .
the proof of follows the same lines as in step 1 if ` ' is replaced by ` ' .
given @xmath529 , taking into account , we have @xmath676 .
if @xmath602 , then @xmath637 , and so , @xmath638 ; in this case @xmath675 , and so , @xmath674 and @xmath640 . now if @xmath613 , then @xmath621 , and so , @xmath642 and , by virtue of lemma [ l : binco](b ) , the value @xmath677 is equal to @xmath503
finally , we establish . if @xmath536 and @xmath644 is odd , we get equality with odd @xmath341 .
assume that @xmath663 .
we have @xmath591 , where @xmath678 ( and so , @xmath593 ) and @xmath679 , and @xmath680 , and so , @xmath648 .
that @xmath601 is nonempty can be seen as follows . since @xmath655 , @xmath622 for some @xmath29 , and
so , if we set @xmath623 with @xmath624 and @xmath625 if @xmath626 , then @xmath614 and @xmath627 is even .
now , since @xmath649 , @xmath681 for some @xmath657 .
if we set @xmath658 with @xmath659 and @xmath660 if @xmath661 , then @xmath662 is odd and @xmath615 , and so , @xmath613 .
assertion with ` @xmath40 is even ' replaced by ` @xmath40 is odd ' is established as in step 3 , while the proof of follows the same lines as in step 1 with ` ' replaced by ` ' .
it follows that @xmath520 .
the proof completes with the last paragraph of step 3 .
the inequality ( actually , equality ) is clear if @xmath584 , and so , we assume that @xmath655 .
the mixed difference at the left - hand side of the inequality is given by , while given @xmath682 , noting that @xmath683 and applying equality we get the following expression for the mixed difference at the right - hand side ( cf .
* part i , expression ( 3.7 ) ) ) : @xmath684 where @xmath685 and @xmath686 . changing the summation multiindex @xmath687 in the sum at the right of the inequality in lemma [ l : deux ]
, we find that it is equivalent to @xmath688 where @xmath689 setting @xmath690 the last inequality can be rewritten as @xmath691 in order to establish , we will apply lemma [ l : metsem ] with @xmath692 and @xmath693 suppose that we have already verified equalities and .
then by lemma [ l : metsem ] , we get inequality , where , by virtue of , @xmath694 summing over @xmath695 and taking into account , we arrive at : @xmath696 assume that @xmath581 is even ; then @xmath481 is odd .
let us verify the first equality in .
for this , we apply equality and calculate the first sum at the left - hand side of .
given even @xmath209 , we have @xmath697 ( cf .
step 1 in the proof of lemma [ l : hh ] ) , and so , by , @xmath698 and @xmath699 , so that the definition of @xmath700 implies @xmath701 applying equality , we get : @xmath702 and the first equality in follows . in a similar manner
we find that the first sum at the left - hand side of is equal to @xmath703 , and , by virtue of , the calculations above show that the second equality in holds as well . now suppose that @xmath581 is odd , and so , @xmath481 ( defined above ) is even . in order to verify the first equality in
, we calculate the first sum at the left - hand side of . given even @xmath40 with @xmath669 , we have ( cf .
step 3 in the proof of lemma [ l : hh ] ) @xmath602 and @xmath608 . moreover ( cf .
* part i , assertion ( 3.9 ) ) ) , there exists a unique @xmath704 s.t .
@xmath604 and @xmath603 ( define @xmath705 by @xmath611 ) . since @xmath612 and @xmath581 is odd , then @xmath705 is odd , and @xmath706 .
it follows that @xmath707 . changing the summation multiindex @xmath708 in the first sum at the left of
, we get : @xmath709 applying equality , we find @xmath710 which proves the first equality in .
similarly , the first sum at the left - hand side of is equal to @xmath487 , and , by virtue of , the calculations above prove the second equality in .
note that if @xmath100\}_{\sigma=0}^\kappa$ ] is a net partition of @xmath8 , then @xmath711}^{x[\sigma]}= \bigcup_{1\le\sigma\le\kappa}\prod_{i=1}^n[x_i(\sigma_i-1),x_i(\sigma_i)]= \prod_{i=1}^n\biggl(\bigcup_{l=1}^{\kappa_i}i_{x_i(l-1)}^{x_i(l)}\biggr)\ ] ] is a union of non - overlapping non - degenerated rectangles @xmath111}^{x[\sigma]}$ ] with the sides parallel to the coordinate axes . in this section
it will be convenient and brief to term the union as in also a _
partition _ of @xmath8 ( by non - overlapping non - degenerated rectangles ) .
if @xmath100\}_{\sigma=0}^\kappa$ ] and @xmath712\}_{\sigma'=0}^{\kappa'}$ ] are two net partitions of @xmath8 , we say that @xmath713 is a _ refinement _ of @xmath368 if @xmath714 .
also , for the sake of convenience we define the _
@xmath12-th prevariation _ of @xmath97 , corresponding to @xmath368 , by @xmath715}^{x[\sigma]}).\ ] ]
it follows that the vitali - type @xmath12-th variation of @xmath4 is given by @xmath716 , where the supremum is taken over all net partitions @xmath368 of @xmath8 .
[ l : etwo ] given @xmath68 with @xmath27 and @xmath718 , we have the following _ partition of @xmath34 , induced by the point @xmath719 : _
@xmath720 where the multiindex @xmath721 is given by @xmath722 0 & \mbox{if } & \mbox{$x_i'\le x_i$ or $ x_i'\ge y_i$ } , \end{array}\right .
\quad\,\,i\in\{1,\dots , n\},\ ] ] and @xmath723 since @xmath727 for all @xmath29 , we have : @xmath728=[x_i , x_i']\cup[x_i',y_i]=i_{x_i}^{x_i'}\cup i_{x_i'}^{y_i } = \bigcup_{\alpha_i=0}^1i_{x_i+\alpha_i(x_i'-x_i)}^{x_i'+\alpha_i(y_i - x_i')},\ ] ] and so ( cf . equation ( 2.5 ) in ( * ? ? ?
* part ii ) ) , @xmath729 the mixed difference at the left - hand side of is given by , and again by virtue of , the mixed difference at the right - hand side of is equal to @xmath730 where @xmath731 and @xmath732 . noting that if @xmath733 , then @xmath734 and @xmath735 , we find @xmath736\scriptstyle 0\le\alpha\le1,\\[-6pt]\scriptstyle\alpha\ne\beta\end{array } } h(\alpha,\beta)+\sum_{{\mbox{\scriptsize\rm ev\,}}\beta\le1}h(\beta,\beta)\nonumber\\ & = & \sum_{{\mbox{\scriptsize\rm ev\,}}\beta\le1}\sum _ { \begin{array}{c}\\[-20pt]\scriptstyle 0\le\alpha\le1,\\[-6pt]\scriptstyle\alpha\ne\beta\end{array } } h(\alpha,\beta)+\sum_{{\mbox{\scriptsize\rm ev\,}}\beta\le1}f(x+\beta(y - x))\nonumber\\ & \equiv&u+u.\nonumber \end{aligned}\ ] ] let us show that the double sum @xmath737 can be represented as @xmath738\scriptstyle 0\le\delta\le\gamma,\\[-6pt]\scriptstyle\delta\ne\gamma\end{array } } c_{\gamma\delta}f\bigl(x+\gamma(x'-x)+\delta(y - x')\bigr)\ ] ] with certian integer factors @xmath739 to be evaluated below .
in fact , given @xmath200 and @xmath740 with @xmath741 , there exist even @xmath742 and @xmath743 , @xmath744 , s.t .
@xmath745 and @xmath746 . in order to see this ,
if @xmath493 is even or @xmath747 is even , we may set @xmath748 and @xmath749 , or @xmath750 and @xmath751 , respectively .
now , if @xmath493 and @xmath747 are odd , then since @xmath741 , we can find @xmath29 s.t .
@xmath752 and @xmath525 , and so , if we set @xmath753 , then @xmath754 , @xmath755 and @xmath756 is even , and it remains to put @xmath757 . given @xmath493 and @xmath747 as above , let us evaluate @xmath739 . since @xmath758 and , given even @xmath618 , the multiindex @xmath743 , @xmath744 , s.t .
@xmath745 and @xmath746 , is determined uniquely by @xmath757 , we have @xmath759 . in a similar manner
, we find @xmath760\scriptstyle 0\le\alpha\le1,\\[-6pt]\scriptstyle\alpha\ne\beta\end{array } } h(\alpha,\beta)+\sum_{{\mbox{\scriptsize\rm od\,}}\beta\le1}f(x+\beta(y - x))\equiv v+v,\ ] ] where @xmath761\scriptstyle 0\le\delta\le\gamma,\\[-6pt]\scriptstyle\delta\ne\gamma\end{array } } d_{\gamma\delta}f\bigl(x+\gamma(x'-x)+\delta(y - x')\bigr)\ ] ] with @xmath762 . by lemma [ l : binco](b ) , @xmath763 , and so , @xmath764 . applying the translation invariance of @xmath55 and inequality ,
we obtain inequality : @xmath765 [ r:2 t ] if @xmath766 in lemma [ l : ethree ] , then all rectangles at the right - hand side of are non - degenerated , i.e. , @xmath767 for all @xmath743 .
moreover , the point @xmath719 gives rise to a net partition @xmath225\}_{\sigma=0}^\kappa$ ] of @xmath34 with @xmath101=(x_1(\sigma_1),\dots , x_n(\sigma_n))$ ] and @xmath768 as follows : we put @xmath769 and , given @xmath29 , we set @xmath770 , @xmath771 and @xmath772 .
we note that if @xmath773 , then @xmath101=x+\sigma(x'-x)$ ] , and if @xmath774 , then @xmath101=x'+(\sigma-1)(y - x')$ ] .
it follows that @xmath775}^{x[\sigma]}.\ ] ] however , in the general case @xmath776 we may also have @xmath777 or @xmath778 , and so , there exists @xmath29 s.t .
@xmath779 or @xmath780 .
thus , since some coordinates of @xmath719 may be equal to the corresponding coordinates of @xmath24 and/or @xmath331 , certain rectangles at the right - hand side of may degenerate into lower - dimensional rectangles , and so , by remark [ rema ] , the mixed difference @xmath781 over these rectangles is equal to zero . in order to exclude these degenerated rectangles from the consideration , we establish the following lemma . [
l : efour ] given @xmath68 with @xmath27 and @xmath724 , we have the following _ partition _ of @xmath34 , induced by @xmath719 : @xmath782 where the multiindices @xmath783 and @xmath784 are defined for @xmath29 by @xmath785 0 & \mbox{if } & x_i < x_i ' , \end{array}\right .
\,\,\,\mbox{and}\,\,\ , \mu_i\equiv\mu_i(x',y)=\left\ { \begin{array}{rcl } 0 & \mbox{if } & x_i'=y_i,\\[2pt ] 1 & \mbox{if } & x_i'<y_i , \end{array}\right.\ ] ] and the following inequality holds : @xmath786
\(iii ) @xmath797 and @xmath794 , and so , if @xmath795 , then @xmath800 and @xmath801 moreover , in all the cases ( i)(iii ) the left endpoint @xmath802 is less than the right endpoint @xmath803 , and so , all the closed intervals above are non - degenerated .
it follows that @xmath804 the point @xmath719 gives rise to a net partition @xmath225\}_{\sigma=0}^\kappa$ ] of @xmath34 as follows : we put @xmath805 and , given @xmath29 , we set @xmath770 and @xmath806 if @xmath807 , and @xmath770 , @xmath771 and @xmath772 if @xmath808
. we note that if @xmath809 , then @xmath101=x+(\sigma+\lambda)(x'-x)$ ] , and if @xmath810 , then @xmath101=x'+(\sigma-1+\lambda)(y - x')$ ] .
also , note that @xmath811 and @xmath812 .
it follows that @xmath813}^{x[\sigma]}.\ ] ] now , we turn to the proof of . by lemma
[ l : ethree ] , inequality holds . clearly ,
if @xmath788 and @xmath789 ( i.e. , @xmath766 ) , then implies .
assume that @xmath814 ( i.e. , @xmath777 ) and suppose that @xmath743 is s.t.@xmath815 .
then there exists @xmath29 s.t .
@xmath793 and @xmath622 , and so , @xmath779 , which implies @xmath816 .
it follows from remark [ rema ] that @xmath817 .
similarly , if we assume that @xmath818 ( i.e. , @xmath778 ) and suppose that @xmath743 is s.t.@xmath819 , then there exists @xmath29 s.t .
@xmath128 and @xmath798 , and so , @xmath780 .
noting that @xmath820 , we find @xmath817 . in this way inequality follows .
suppose that @xmath68 , @xmath27 and @xmath718 .
we set @xmath821 , where @xmath822 is defined in ( the point @xmath823 will play the role of @xmath719 from ) .
we have @xmath824 ; in fact , given @xmath29 , we find : if @xmath825 , then @xmath826 and @xmath827 implying @xmath828 , and if @xmath829 , then @xmath830 or @xmath831 , and @xmath832 implying @xmath833 .
applying with @xmath719 replaced by @xmath823 , we get the following partition of @xmath34 induced by @xmath823 and , hence , by @xmath719 : @xmath834 where @xmath835 and @xmath836 are defined in lemma [ l : efour ] , i.e. , given @xmath29 , we have : @xmath837 0 & \mbox{if } & x_i < x_i '' , \end{array}\right .
\quad\,\mbox{and}\quad\ , \mu_i''=\left\ { \begin{array}{rcl } 0 & \mbox{if } & x_i''=y_i,\\[2pt ] 1 & \mbox{if } & x_i''<y_i .
\end{array}\right.\ ] ] we assert that @xmath838 and @xmath839 . in fact , since @xmath840 , then @xmath839 .
in order to see that @xmath838 , let @xmath29 .
if @xmath826 , then @xmath825 , and so , @xmath841 , which implies @xmath842 and @xmath843 . now if @xmath830 or @xmath831 , then @xmath829 , and so , @xmath832 , which gives @xmath844
now , let us calculate the lower and upper indices in .
we have : @xmath845 and @xmath846 noting that the union in is taken over @xmath847 s.t .
@xmath848 , we get @xmath849 , and so , @xmath850 implying @xmath851 these calculations and observations above prove equality .
let us show that partition is actually induced by @xmath719 . since @xmath718 , by lemma [ l : efour ] , the point @xmath719 induces a partition of @xmath8 of the form : @xmath852 where the multiindices @xmath853 and @xmath854 are defined in lemma [ l : efour ] , i.e. , given @xmath29 , we have : @xmath855 0 & \mbox{if } & a_i < x_i ' , \end{array}\right .
\quad\,\mbox{and}\quad\ , \mu_i'=\left\ { \begin{array}{rcl } 0 & \mbox{if } & x_i'=b_i,\\[2pt ] 1 & \mbox{if } & x_i'<b_i .
\end{array}\right.\ ] ] we assert that for each @xmath433 with @xmath856 there exists a unique @xmath857 with @xmath858 s.t . @xmath859 in order to prove , we define @xmath860 by @xmath861 0 & \mbox{if } & x_i'\ge y_i , \end{array}\right .
\qquad i\in\{1,\dots , n\},\ ] ] and establish equality componentwise . given @xmath29 , we consider the following two cases : ( a ) @xmath791 , and ( b ) @xmath831 . in case ( a )
we have @xmath862 .
first , assume that @xmath792 , and so , @xmath825 .
it follows that if @xmath863 , then @xmath622 or @xmath128 .
if @xmath622 , then we find ( for @xmath864 ) @xmath865=[x_i , y_i]\cap[a_i , x_i']= i_{x_i}^{y_i}\cap i_{a_i+\beta_i(x_i'-a_i)}^{x_i'+\beta_i(b_i - x_i')},\ ] ] and if @xmath128 , then we find ( for @xmath866 ) @xmath867=[x_i , y_i]\cap[x_i',b_i]= i_{x_i}^{y_i}\cap i_{a_i+\beta_i(x_i'-a_i)}^{x_i'+\beta_i(b_i - x_i')}.\ ] ] now , assume that @xmath830 , and so , @xmath829 and @xmath868 .
it follows that if @xmath863 , then @xmath866 implying @xmath728=[x_i , y_i]\cap[x_i',b_i]= i_{x_i}^{y_i}\cap i_{a_i+\beta_i(x_i'-a_i)}^{x_i'+\beta_i(b_i - x_i')}.\ ] ] in case ( b ) we have @xmath829 , @xmath624 and @xmath869 , and so , if @xmath863 , then @xmath128 and @xmath728=[x_i , y_i]\cap[a_i , x_i']= i_{x_i}^{y_i}\cap i_{a_i+\beta_i(x_i'-a_i)}^{x_i'+\beta_i(b_i - x_i')}.\ ] ] let us show that @xmath858
. let @xmath29 . if @xmath870 , then @xmath871 and , since @xmath791 , then @xmath862 . by , @xmath829 , and so , since @xmath863 , then @xmath128 , which implies @xmath872 .
now , if @xmath873 , then @xmath874 and , since @xmath831 , then @xmath624 ( and @xmath829 ) , and so , @xmath872 . finally , if @xmath875 , then @xmath876 and @xmath877 , and so , since @xmath878 , then @xmath879 .
the uniqueness of @xmath880 , for each @xmath856 , is a consequence of the following : if @xmath858 and @xmath881 , then there exists @xmath29 s.t .
arguing as in ( a ) and ( b ) above , we find that the equality can not hold for this @xmath618 .
[ r : abc ] ( a ) if @xmath724 in lemma [ l : etwo ] , then it is easily seen that @xmath883 , and so , equality assumes the form : @xmath884 although this equality looks different from , the two equalities are the same : this is verified as in ( i)(iii ) of the proof of lemma [ l : efour ] .
\(c ) here we consider a certain particular case of and establish conditions on @xmath719 , under which @xmath719 does not induce a ( further ) partition of @xmath34 . in view of , we have : @xmath886 which is also equivalent to @xmath887 clearly , if @xmath888 and @xmath584 , then the left - hand side equality in holds .
conversely , if the left - hand side equality in holds for some @xmath889 , then @xmath890 and @xmath891 , and so , if we suppose that @xmath825 for some @xmath29 , then , by , @xmath826 , and so , @xmath622 and @xmath780 , which is a contradiction .
thus , @xmath888 and @xmath584 .
let @xmath100\}_{\sigma=0}^\kappa$ ] for some @xmath99 and @xmath895 . given @xmath107
, we set @xmath896 $ ] , @xmath897 $ ] and @xmath898 , where @xmath822 is defined in , and note that @xmath899 .
the point @xmath719 induces a partition of @xmath900}^{x[\sigma]}$ ] of the form with @xmath901 and @xmath902 , and so , by virtue of , we get the following partition of @xmath8 , induced by @xmath719 : @xmath903 we denote by @xmath904 the net partition of @xmath8 corresponding to .
moreover , by , for each @xmath107 we have the inequality : @xmath905 with no loss of generality we may assume that @xmath906 : if @xmath907 , i.e. , @xmath908 $ ] for some @xmath909 , then @xmath719 does not affect the partition @xmath368 of @xmath8 in the sense that @xmath910 , and so , @xmath911 . in order to see this , we note that implies @xmath912,x[\sigma'],x[\sigma])=0 $ ] , and so , by remark [ r : abc](c ) , conditions and hold with @xmath913 s.t . @xmath914 0 & \mbox{if } & x_i(\sigma_i')\ge x_i(\sigma_i ) , \end{array}\right .
= \left\ { \begin{array}{ccl } 1 & \mbox{if } & \sigma_i'<\sigma_i,\\[2pt ] 0 & \mbox{if } & \sigma_i'\ge \sigma_i .
\end{array}\right.\ ] ] replacing @xmath368 by @xmath904 in the arguments above , taking @xmath916 and denoting by @xmath917 the partition of @xmath8 induced from @xmath918 by @xmath719 , we get @xmath919 . since @xmath920 is a finite set , we exhaust it by points @xmath719 in a finite number of steps , arrive at the partition @xmath921 of @xmath8 and prove the desired inequality @xmath922 .
first , we establish for @xmath923 , i.e. , @xmath924}^{x[\sigma]}).\ ] ] modulo the notation , there is no loss of generality if we assume that @xmath925 and @xmath285 , so that @xmath225\}_{\sigma=0}^\kappa$ ] is a net partition of @xmath8 .
let @xmath926 be an arbitrary net partition of @xmath8 .
denote by @xmath921 the net partition of @xmath8 induced from @xmath926 by points @xmath225\}_{\sigma=0}^\kappa$ ] , so that @xmath921 is a refinement of @xmath926 .
given @xmath107 , set @xmath927}^{x[\sigma]}$ ] and note that @xmath928 is a net partition of @xmath111}^{x[\sigma]}$ ] , and @xmath929 . then by virtue of lemma
[ l : eone ] , we have : @xmath930}^{x[\sigma]}).\ ] ] since @xmath926 is arbitrary , the left - hand side in is not greater than the right - hand side .
let us prove the reverse inequality .
if @xmath931}^{x[\sigma]})$ ] is infinite for some @xmath107 , then since @xmath111}^{x[\sigma]}\subset i_a^b = i_x^y$ ] , the value @xmath932 is infinite as well .
thus , we suppose that the right - hand side of is finite .
let @xmath334 be arbitrary .
given @xmath107 , by the definition of @xmath931}^{x[\sigma]})$ ] , there exists a net partition of @xmath111}^{x[\sigma]}$ ] , denoted by @xmath933 , s.t .
@xmath934}^{x[\sigma ] } ) -(\varepsilon / c),\ ] ] where @xmath935 .
we denote by @xmath936 the net partition of @xmath8 induced from @xmath225\}_{\sigma=0}^\kappa$ ] by points from @xmath937 .
given @xmath107 , we set @xmath938}^{x[\sigma]}$ ] and note that @xmath939 is a refinement of @xmath933 , and @xmath940 . by virtue of lemma
[ l : eone ] , we find @xmath941}^{x[\sigma ] } ) -\varepsilon\biggl(\sum_{1\le\sigma\le\kappa}1\biggr)\bigl / c,\nonumber \end{aligned}\ ] ] where the factor by @xmath942 is , actually , equal to @xmath45 .
the desired inequality follows if we take into account the arbitrariness of @xmath334 .
, suppose that @xmath49 and @xmath655 .
note that @xmath943 and @xmath944 , and that @xmath225\lfloor\alpha\}_{\sigma\lfloor\alpha=0}^{\kappa\lfloor\alpha}$ ] is a net partition of @xmath945 .
so , replacing @xmath946 by @xmath947 ( so that @xmath948 ) and @xmath4by @xmath125 in , we get : @xmath949 & = & \sum_{1\lfloor\alpha\le\sigma\lfloor\alpha\le\kappa\lfloor\alpha } v_{|1\lfloor\alpha|}\bigl(f_\alpha^z , i_{x[\sigma-1]\lfloor\alpha}^{x[\sigma]\lfloor\alpha}\bigr ) , \nonumber \end{aligned}\ ] ] which is equal to the right - hand side of . | given two points @xmath0 and @xmath1 from @xmath2 with @xmath3 componentwise and a map @xmath4 from the rectangle @xmath5\times\cdots\times[a_n , b_n]$ ] into a metric semigroup @xmath6 , we study properties of the _ total variation _ @xmath7 of @xmath4 on @xmath8 introduced by the first author in [ v.v .
chistyakov , a selection principle for mappings of bounded variation of several variables , in : real analysis exchange 27th summer symposium , 2003 , 217222 ] , which extends the classical notion of c. jordan s total variation ( @xmath9 ) and the corresponding notions in the sense of [ t.h .
hildebrandt , introduction to the theory of integration , academic press , 1963 ] ( @xmath10 ) and [ a.s .
leonov , on the total variation for functions of several variables and a multidimensional analog of helly s selection principle , math .
notes 63 ( 1998 ) 6171 ] ( @xmath11 ) for real valued functions of @xmath12 variables .
the following helly - type pointwise selection principle is proved : _
if a sequence @xmath13 of maps from @xmath8 into @xmath14 is such that the closure in @xmath14 of the set @xmath15 is compact for each @xmath16 and @xmath17 is finite , then there exists a subsequence of @xmath13 , which converges pointwise on @xmath8 to a map @xmath4 such that . _ a variant of this result
is established concerning the weak pointwise convergence when values of maps lie in a reflexive banach space @xmath18 with separable dual @xmath19 .
maps of several variables , total variation , selection principle , metric semigroup , pointwise convergence , weak convergence _ msc classification : _
26b30 , 20m15 , 28a20 |
in the study of the nature of the electroweak phase transition in the early universe the effective potential plays an essential role . since the early studies it was realized that the computation of the effective potential was plagued with difficulties which are mainly related with the bad infrared behavior of the theory near the critical temperature @xcite . in order to avoid the appearance of imaginary terms in the effective potential it was then necessary to improve its perturbative computation by resumming an infinite subset of the leading infrared divergent diagrams , called the ring diagrams .
an important improvement was made by the use of self - consistent methods based on schwinger - dyson equations for the masses .
they can be obtained through a modification of the mass term in the original lagrangian ; a mass parameter is introduced in the beginning and later determined by consistency conditions , the gap equations ( see e.g. @xcite,@xcite ) . in spite of the considerable progress made in this direction ,
there still remain some ambiguities in the calculation of the effective potential at finite temperature which are mainly related with the correct way to perform the resummation of the infrared divergent diagrams .
also when using the kind of methods based on gap equations it will always appear temperature dependent ultraviolet divergencies , which implies that in order to renormalize the theory one should include temperature dependent counterterms . although these counterterms have the same structure as those of the @xmath0 theory , its physical origin is very unclear . in this letter
we show a novel way of computing the one - loop ring - improved effective potential numerically which avoid the above mentioned problems related with renormalization .
it is enough to consider the @xmath1 -theory to explain the method , but for definitness the final calculation is carried out within the standard model .
we begin by reviewing ( in the @xmath1-theory ) the usual way one performs the improvement of the one - loop effective potential at finite temperature .
we add and substract in the original lagrangian the induced thermal mass @xmath2 , where @xmath3 , @xmath4 are the matsubara frequencies . the result of this procedure is an improved perturbation theory in which the original propagator is replaced by the full propagator .
the polarization tensor @xmath5 can be obtained consistently within perturbation theory . due to the fact that it is the infrared behavior the relevant aspect near the phase transition
, it is only necessary to compute @xmath6 in the infrared limit @xmath7 , @xmath8 , @xmath9
. this procedure amounts to a summation of an infinite set of diagrams with the worst infrared behavior , called the ring diagrams , and which are generated by including polarization tensor insertions in the infrared limit in one loop diagrams .
it is not necessary to do that for fermions since in this case the matsubara frequencies are always different from zero .
therefore , the effective potential reads @xmath10 , \nonumber\ ] ] where @xmath11 is the plasma mass ( @xmath12 . the sum over matsubara frequencies can be performed and one obtains @xmath13 \nonumber\ ] ] it is clear that the second term on the r.hs . of this expression contains temperature dependent ultraviolet divergent contributions .
therefore temperature dependent counterterms are needed in order to renormalize this expression . in spite of the similar structure of these counterterms with those of the @xmath0 limit ,
its physical origin is not clear and represents a lack of consistency of the procedure used @xcite .
in fact the renormalization problem is related with the local behavior of the theory and should not be influenced by finite temperature effects , which concerns long distance physics .
nevertheless if one accepts to work with temperature dependent counterterms one can formally derive an expression for the effective potential . to find explicit expressions see ref .
in this section we show a novel way of computing the one - loop ring - improved effective potential which avoids the above mentioned problems related with renormalization . in this approach we will consider the improvement of the effective potential by the summation of rings diagrams , which in this case are generated by including polarization tensor insertions in the infrared limit , in one - loop diagrams with _
zero matsubara frequencies_. to explain the method we start with the expression for the one - loop ring - improved effective potential @xmath14 , \nonumber\ ] ] where the third term on the r.h.s can be separated into the one - loop @xmath15 \nonumber\ ] ] and the ring @xmath16 ^{3/2}-m^3(\phi ) \right)\ ] ] contributions . at first sight , by looking at eqs .
( [ unot ] ) and ( [ ringts ] ) one should be tempted to say that there is a difficulty in using the above equation for the effective potential , because for small enough values of the classical field @xmath17 the scalar mass squared @xmath18 is negative and therefore the potential itself would be a complex expression in that range .
of course this is not so because although both @xmath19 and @xmath20 become complex for small enough values of @xmath17 , their sum , given by equation ( [ pott ] ) , is real provided @xmath21 .
this is not new and people have used this fact to evaluate the effective potential in the high - temperature expansion . in this case
the dangerous terms , i.e. those which become imaginary if @xmath22 cancel each other and everything is safe . but what happens if we want to evaluate the one - loop ring - improved effective potential in an exact way , that is , numerically ? in this case
a calculational problem arises because the scalar mass squared are negative for finite values of @xmath17 . the usual way out of this problem is to carry out the self consistent method outlined in the previous section which , as we have seen , from first principles is not correct , because it mixes the infrared and ultraviolet regimes .
the method we propose is based on eq .
( [ pott ] ) and avoids all the above mentioned difficulties .
the important point to remember is that @xmath23 is real provided @xmath24 .
the one - loop contribution to the effective potential @xmath25 is @xmath26 -\frac { 3}{2}\right ) + 2m^2(\phi ) m^2(\phi _ 0)\right\ } \nonumber \\ & & + \frac{t^4}{2\pi ^2}\int_0^\infty dxx^2\ln \left [ 1-e^{-\sqrt{% x^2+m^2(\phi ) /t^2}}\right ] .\end{aligned}\ ] ] the renormalization is performed here at zero temperature which in turns implies that the required counterterms are temperature - independent . as renormalization prescription
we require that both the classical values of the higgs boson mass @xmath27 and the higgs field at the minimum of the effective potential at @xmath0 , @xmath28 , to be preserved even after quantum corrections are taken into account . in opposition to what happens in the self - consistent approach ,
the renormalization procedure is free of any theoretical inconsistency . if @xmath29 everything is standard and no difficulties appear in the calculation of eq .
( [ pott2 ] ) .
therefore let us consider the case when @xmath30 and define @xmath31 .
thus , by separating the range of integration in the temperature dependent part we have @xmath32 \nonumber \\ & & + \frac{t^4 } { 2\pi ^2 } \int_{\widehat{m}(\phi ) /t}^\infty dxx^2\ln \left [ 1-e^ { -\sqrt{x^2-\widehat{m}^2(\phi ) /t^2}}\right ] .\end{aligned}\ ] ] clearly the second term is real , while the first one is complex and can be explicited as follows .
notice that @xmath33&=&-i\sqrt{% \widehat{m}^2(\phi ) /t^2-x^2}/2+\ln
2+i\frac { \pi}{2 } \nonumber \\ & & + \ln \left [ \sin \left ( \frac { 1}{2}\sqrt{\widehat{m}^2(\phi ) /t^2-x^2}%
\right ) \right ] , \end{aligned}\ ] ] where we have chosen the branch point where @xmath34 thus , @xmath35 \nonumber \\ & & + \frac{t^4}{2\pi ^2}\int_0^{\widehat{m}(\phi)/t}dxx^2\ln \left [ 2\sin \left ( \frac 12\sqrt{\widehat{m}^2(\phi ) /t^2-x^2}\right ) \right ] \nonumber \\ & & + \frac{t^4 } { 2\pi^2}\int_{\widehat{m}(\phi ) /t}^\infty dxx^2\ln \left [ 1-e^ { -\sqrt{x^2-\widehat{m}^2(\phi ) /t^2}}\right ] .\end{aligned}\ ] ] the imaginary part of @xmath36 can be integrated exactly and the result is @xmath37 notice that the second term on the r.h.s .
corresponds to one of the terms appearing in the ring contribution in the scalar sector , see eq .
( [ ringts ] ) . also notice that the presence of the extra term , which does not appear in the ring contribution .
its effect is to cancel the imaginary part of the zero temperature contribution coming from the logarithm ( remember that we are considering the case of negative mass squared ) .
this shows an exact cancellation of the spurious imaginary terms in the effective potential , as it should be the case because for @xmath38 , the effective potential is real .
thus , it is clear that such a cancellation does not depend on the order of the high - temperature expansion .
therefore , the final expression for the effective potential , for the case @xmath30 , is @xmath39 -\frac { 3}{2}\right ) -2\widehat{m}% ^2(\phi ) m^2(\phi_0)\right \ } \nonumber \\ & & + \frac{t^4}{2\pi ^2}\int_0^{\widehat{m}(\phi)/t } dxx^2\ln \left [ 2\sin \left ( \frac { 1}{2}\sqrt{\widehat{m}^2(\phi ) /t^2-x^2 }
\right ) \right ] \nonumber \\ & & + \frac{t^4}{2\pi^2}\int_{\widehat{m}(\phi)/t}^\infty dxx^2\ln \left [ 1-e^{-\sqrt{x^2-\widehat{m}^2(\phi ) /t^2}}\right ] .\end{aligned}\ ] ]
as we said in the introduction , we present our numerical results in the context of the standard model .
the values of the su(2 ) and u(1 ) couplings constants are @xmath40 and @xmath41 , respectively . also , we take @xmath42 gev .
we implement the first ( self - consistent ) approach by following carrington @xcite , that is , by including the whole range of matsubara frequencies in the scalar sector and by considering only the contribution of the zero matsubara frequency in the gauge sector .
in particular , this means that we consider equations of the form ( [ prim ] ) in order to evaluate the temperature dependent part of the effective potential .
more precisely , we use the same set of equations that she used in order to evaluate the effective potential . the way we perform the numerical computation of the effective potential using our approach is to separate the cases where the squared mass of the higgs boson ( respectively the goldstone boson ) is positive or zero from that where it is negative .
in the first case , no problems appear in the numerical evaluation of the effective potential , while we use expressions like equation ( [ potefft ] ) when the scalar masses squared become negative .
the rest of the expressions we use for the evaluation of the effective potential are those given in ref .
@xcite .
figure 1 shows the effective potential at the critical temperature , for a top - quark of mass 180 gev , versus the value of @xmath17 .
the appearence of a first order phase transition is shown .
notice that both results coincide considerable well in a wide range , the differences appearing for values of @xmath43 bigger than the position of the non trivial minimum of the effective potential .
indeed the critical temperature , defined when the minima of the effective potential are degenerate , is the same when calculated through the two different ways .
moreover , the ratio @xmath44 at the critical temperature , which is the relevant parameter to be studied in order to conclude something about the baryon asymmetry washout , is unaltered in both calculations as figure 2 shows . since both methods of evaluation of the effective potential differ by the way how the matsubara frequencies are taken into account ( considering all matsubara frequencies in the first method and only the zero mode in our proposal ) , the results show in a numerical way the irrelevance of those terms contributed by the nonzero matsubara frequencies .
although our results do not differ significatively from those obtained through the application of the self - consistent approach , our method has the advantage of making the renormalization procedure conceptually clear , avoiding the necessity of introducing temperature dependent counterterms .
one of us ( l.v . ) wants to thank o. espinosa for stimulating discussions in earlier stages of this work and acknowledges the hospitality extended by the people of the departamento de fsica of universidad santa mara .
l.v . was supported by a fondecyt postdoctoral fellowship , grant no .
g.p . was supported in part by dicyt no . 049331pa and fondecyt no .
1930067 . .
the one - loop ring improved effective potential at finite temperature u(@xmath17 ) versus @xmath17 is shown , for a higgs mass of 60 gev and a top - quark mass of 180 gev .
circles represent the data from the self - consistent method , while crosses are data from our proposal . | we present a novel way to compute the one - loop ring - improved effective potential numerically , which avoids the spurious appearence of complex expressions and at the same time is free from the renormalization ambiguities of the self - consistent approaches , based on the direct application of schwinger - dyson type equations to the masses . |
highly frustrated magnetic systems are highly susceptible to quantum spin fluctuations and instabilities towards competing ground states .
the triangular heisenberg antiferromagnet with spin @xmath7 constitutes a paradigm of that class of systems . in that context , it is interesting that experimental investigations of the cs@xmath0cucl@xmath1 @xcite and @xmath2-(bedt - ttf)@xmath0cu@xmath0(cn)@xmath3 @xcite materials , which can be both described as a first approximation by a triangular lattice with spatially anisotropic exchange couplings , indicate exotic behaviors .
exotic behaviors include either the realization of a spin - liquid ground state or a magnetically ordered phase with a magnetic excitation dispersion strongly renormalized compared to the classical spin - wave .
indeed , recent numerical studies based on series expansion @xcite have found that frustration manifests itself rather directly in the spin excitations of archetype models of two - dimensional frustrated spin systems .
a softening of the magnon excitations in a broad region of the reciprocal @xmath8 space is observed as a quantity , @xmath9 , which parametrize the level of frustration , is increased . since , within a semi - classical picture ,
the effect of frustration is to bring about the effect of competing ground states , it is expected that the spin excitations out of a semi - classical long - range ordered ground state are direct tell - tale indicators of the presence of frustrating interactions . in this paper
we explore the possibility that frustration can also be investigated via the two - magnon density of states at zero wavevector measured by polarized inelastic magnetic raman scattering .
+ the rest of this paper is organized as follows : in the first section we present the derivation of the scattering operator , then we discuss its form for various polarizations . in the last part we present the associated raman spectra obtained by exact - diagonalization of finite - size clusters .
we first consider the hubbard model on the anisotropic triangular lattice : @xmath10 where @xmath11 is the hopping from a site @xmath12 to a neighboring site @xmath13 , @xmath14 is the spin degree of freedom , and @xmath15 and @xmath16 are the usual annihilation and creation operators . for the case considered here , there are two distinct hopping integrals , @xmath17 and @xmath18 , along the corresponding directions shown in fig .
+ geometry of the lattice and polarization vectors .
_ left : _ the dashed lines represent the coupling @xmath19 , the thick lines are for the exchange coupling @xmath20 .
_ right : _ definition of the polarization vectors by two angles @xmath21 and @xmath22.,width=340 ] at half - filling and in the large-@xmath23 limit , the system is described ( to second order in @xmath24 ) by an effective spatially anisotropic heisenberg hamiltonian : @xmath25 where the different exchange coupling along each direction is due to the difference in the hopping amplitude , _
i.e _ @xmath26 , @xmath27 .
+ raman scattering consists of an incoming photon ( of energy @xmath28 ) scattered into an outgoing photon ( of energy @xmath29 ) , involving different manifold of electronic states having zero or one double - occupancy .
these transitions depend on the polarizations ( referred as * e*@xmath30 and @xmath31 ) of the incoming and outgoing photons . following the early work of fleury and loudon @xcite and shastry and shraiman @xcite
, we derive an effective scattering spin hamiltonian describing this problem .
the @xmath32-component of the electronic hopping current operator is : @xmath33 \ ] ] where @xmath34 is the momentum transfer , and @xmath35 the band - energy . in what follows
we take the most general set of polarization vectors , as shown in the right panel of fig .
[ pola ] : @xmath36 the raman scattering process involves an energy transfer @xmath37 , the momentum transfer being set to zero .
therefore , we restrict ourselves to the assumption @xmath38 , and we have : @xmath39\\
-i\frac{\sqrt{3}}{2}\sin\phi\left[t_2\left(c^\dagger_r c_{r+{\bf e}_2}- h.c\right)+ t_1\left(c^\dagger_r c_{r+{\bf e}_3}- h.c\right ) \right]\\ \ \\ { \bf j}\cdot{\bf e}_{\rm out } = i\cos\theta \left[-t_1\left(c^\dagger_r c_{r+{\bf e}_1}- h.c\right)\right .
\left.-\frac{t_2}{2}\left(c^\dagger_r c_{r+{\bf e}_2}- h.c\right)\right.\\\left.+ \frac{t_1}{2}\left(c^\dagger_r c_{r+{\bf e}_3}- h.c\right)\right]\\ -i\frac{\sqrt{3}}{2}\sin\theta\left[t_2\left(c^\dagger_r c_{r+{\bf e}_2}- h.c\right)+ t_1\left(c^\dagger_r c_{r+{\bf e}_3}- h.c\right ) \right]\\ \end{array}\ ] ] the raman scattering operator is given by the second order formula , @xmath40 and @xmath41 being respectively the initial and final state , @xmath42 being an intermediate state : @xmath43 \end{array}\ ] ] following the same algebra steps as in @xcite , and restricting @xmath40 and @xmath41 to the manifold of singly occupied states , and intermediate states @xmath42 to the manifold of one double occupancy , we use the identity , @xmath44 , and obtain the scattering operator in terms of spin operators .
we note that within our approach the total scattering operator only contains terms of the form : @xmath45 from eq .
( [ scat ] ) we find that the scattering hamiltonian prefactors which will be in front of eq .
( [ scat2 ] ) depend on the polarization vectors orientation .
these prefactors are proportional to the exchange coupling @xmath46 , therefore , in its general form , the total scattering operator will depend on both @xmath20 and @xmath19 .
the two angles @xmath21 and @xmath22 with respect to the @xmath47-axis ( see fig . [ pola ] ) define the polarizations involved in the scattering process .
therefore the scattering operator depends on a projector @xmath48 that defines the polarization set - up .
the raman operator takes the form : @xmath49\\ \ \\
+ \left.j_1 { \bf s}_i\cdot{\bf s}_{i+{\bf e}_3}\left[\cos(\theta+\phi)-\sqrt{3}\sin(\theta+\phi)+4\sin\theta\sin\phi\right ] \right\ } \end{array}\ ] ] which can be written in the compact form : @xmath50 in order to compare with the square - lattice case , we now focus on the following polarization geometries : @xmath51 \\ { \mathcal h}_{lf}(\frac{5\pi}{6},\frac{\pi}{3 } ) & \propto & \sum_{i } j_1 \left[{\bf s}_i\cdot{\bf s}_{i+{\bf e}_1}-{\bf s}_i\cdot{\bf s}_{i+{\bf e}_3}\right ] \end{array}\ ] ] when the diagonal bond @xmath19 is @xmath52 , the first line of eq .
( [ a1gb1 g ] ) is the a@xmath53 raman operator , while the second line gives the b@xmath53 raman operator , both on a square lattice .
most importantly , we note that the scattering operators depend on the ratio @xmath54 if one takes : @xmath55 \\ { \mathcal h}_{lf}(\frac{\pi}{6},\frac{2\pi}{3 } ) & \propto & \sum_{i}\left[j_1{\bf s}_i\cdot{\bf s}_{i+{\bf e}_1}-j_2{\bf s}_i\cdot{\bf s}_{i+{\bf e}_2}\right ] \end{array}\ ] ] we would like to attract the attention of the reader to the fact that the standard parallel polarization does not give a straightforward a@xmath53-like scattering operator as it is the case for the square lattice , but a linear combination of eqs .
[ a1gb1 g ] and [ a1gb1g - j2 ] .
as a first step in exploring the raman scattering in the heisenberg @xmath7 anisotropic triangular lattice antiferromagnet , we perform exact - diagonalizations for the raman operator eq .
( [ operator ] ) .
while raman scattering studies on triangular lattice compounds like cobaltites have been carried out @xcite , these have focused on phonons , and there are yet no systematic study of polarization dependent electronic raman spectra . hence , our aim at this stage is not to obtain a quantitative description of the scattering spectra for a specific material , but rather identify the key qualitative features emerging in the raman spectrum of the anistropic triangular lattice , possibly to motivate raman studies and to connect with recent neutron data @xcite , nmr @xcite and angle - resolved photoemission @xcite .
motivated by the successes of exact - diagonalization to identify the essential qualitative features of the raman spectrum on the square lattice @xcite , and in order to compare with these well - known results , we explore in this study the behaviour of a 16-site cluster as the frustrating @xmath19 coupling is increased from zero .
the results are summarized in fig .
[ raman ] .
the general trend observed in these plots is that a softening progressively develops as the system becomes more frustrated .
exact diagonalizations on a 16-site cluster .
raman spectra for various values of @xmath56 and @xmath4 .
the bottom ( top ) left panel would correspond to @xmath57 ( @xmath58 ) polarization on the square lattice in the case @xmath59.,width=548 ] the bottom left panel of fig .
[ raman ] shows , for @xmath5 , the well known zero a@xmath53 raman scattering on a square lattice , since the scattering operator commutes with the hamiltonian @xcite . as @xmath19 increases non - commuting contribution to the scattering intensity increases and the spectrum develops more structure .
for the bottom right panel , a weak intensity if observe because the scattering operator does not commute with the hamiltonian even for @xmath5 .
+ from the hamiltonian in eq .
[ hamiltheis ] , it is obvious that the ground state will depend on the ratio @xmath4 , ranging from a nel - like order on the square lattice when @xmath5 , to a three sublattice long - range order for @xmath60 @xcite .
these different type of order are characterized by different magnon dispersions , therefore the raman spectra is expected to depend on @xmath4 . in a recent paper , zheng _
@xcite showed , using series expansions , that the magnon dispersion for the anisotropic triangular lattice heisenberg model exhibits a _
roton - like _ minimum .
this local minimum sits at the @xmath61 point for a square lattice with one diagonal bond and is getting softer as the @xmath19-diagonal frustrating coupling increases . in the case of the standard square lattice , for the b@xmath53 channel with crossed polarizations , the electromagnetic field couples to excitations along the @xmath61 direction .
we note from @xcite that this softening is a multi - magnon process that can not be captured to lowest order in a 1/s expansion .
however , since we are using exact - diagonalization which , by its nature , treats processes for the length scale considered exactly , we are able to detect indications of this softening . for specificity , consider the b@xmath53 channel in the top left - hand panel of fig .
a signature of the softening of the magnon excitation is clearly manifest as the frustration ratio @xmath62 is increased from zero , with a shift of the spectral weight , and the main peak moving from @xmath63 when @xmath5 to @xmath64 when @xmath65 .
the results presented here show how frustration can dramatically alter the raman spectrum of an otherwise non - frustrated system .
in particular , we observe a frustration - driven _ spectral downshift _ , the landmark feature of frustrated systems .
this shift is another indicator of the frustration - induced softening of the magnon dispersion recently predicted by others . from a strictly theoretical point of view
, the analysis above sets the stage for a more complete approach .
some exact - diagonalizations on larger cluster and a spin - wave approach including magnon - magnon interactions , along the lines pursued for the square lattice @xcite , are currently being carried out .
support for this work was provided by nserc of canada and the canada research chair program ( tier i ) ( m.g . ) , the canada foundation for innovation , the ontario innovation trust , and the canadian institute for advanced research ( m.g . ) .
m.g . acknowledges the university of canterbury for an erskine fellowship and thanks the department of physics and astronomy at the university of canterbury for their hospitality where part of this work was completed .
10 r. coldea _ et al . _ , phys . rev
. lett . * 88 * , 137203 ( 2002 ) ; r. coldea _ et al .
b * 68 * , 134424 ( 2003 ) | motivated by various spin-1/2 compounds like cs@xmath0cucl@xmath1 or @xmath2-(bedt - ttf)@xmath0cu@xmath0(cn)@xmath3 , we derive a raman - scattering operator _
la _ shastry and shraiman for various geometries . for t=0 ,
the exact spectra is computed by lanczos algorithm for finite - size clusters .
we perform a systematic investigation as a function of @xmath4 , the exchange constant ratio : ranging from @xmath5 , the well known square - lattice case , to @xmath6 the isotropic triangular lattice .
we discuss the polarization dependence of the spectra and show how it can be used to detect precursors of the instabilities of the ground state against quantum fluctuations . |
one of the most exciting challenges in current nuclear physics is the understanding of nuclear systems under extreme conditions @xcite .
the opportunities offered by recent experimental work using beams of exotic nuclei and the corresponding theoretical efforts to describe them are of great interest , especially for nuclear structure physics and nuclear astrophysics @xcite . the decay properties and cross sections for nuclear reactions of radioactive nuclei are fundamental to understand various phases in the stellar evolution , including the energy generation , the nucleosynthesis , and the abundance of elements .
since this information can not be determined experimentally for the extreme conditions of temperature and density that hold in the interior of the star , reliable theoretical calculations for these processes are mandatory .
in particular , the decay properties of proton rich nuclei are fundamental to understand the @xmath1 process ( rapid proton capture nucleosynthesis ) , characterized by very large proton capture reaction rates on proton rich nuclei @xcite .
of special importance in this context are the waiting points like @xmath2kr , where the @xmath1 process is inhibited and the reaction flow has to wait for the relatively slow @xmath0-decay to continue .
the total half - lives of the waiting points determine the speed of nucleosynthesis towards heavier nuclei as well as the isotopic abundances .
decay properties and nuclear structure are intimately related .
it is clear that a precise and reliable description of the ground state of the parent nucleus and of the states populated in the daughter nucleus is necessary to obtain a good description of the decay , and vice versa , failures to describe the decay properties would indicate that an improvement of the theoretical formalism is needed . from a theoretical point of view the physics of exotic nuclei , characterized by very unusual ratios of neutrons and protons ,
can be considered as a test for the already well established models of nuclear structure that are used to describe stable systems .
since the parameters and interactions used in the usual shell model or mean field calculations are determined in order to reproduce the properties of known nuclei , they may not always be appropriate for use in the calculations of nuclei approaching the drip lines .
different microscopic models to describe the @xmath3-strength are present in the literature .
models based on spherical single - particle wave functions and energies with pairing and residual interactions treated in quasiparticle random phase approximation ( qrpa ) were first studied in ref .
deformation was included in ref .
@xcite , where a nilsson potential was used to generate single - particle orbitals .
extensions including woods - saxon type potentials @xcite , residual interactions in the particle - particle channel @xcite , hartree - fock ( hf ) mean field with separable residual interactions treated in tamm dancoff approximation @xcite , selfconsistent approaches in spherical neutron - rich nuclei @xcite and based on an energy - density functional @xcite , can be also found in the literature . in a previous work
@xcite we studied ground state and @xmath3-decay properties of even - even exotic nuclei on the basis of a deformed selfconsistent hf+bcs+qrpa calculation with density dependent effective interactions of skyrme type .
this is a well founded approach that has been very successful in the description of spherical and deformed nuclei within the valley of stability @xcite . in this method
once the parameters of the effective skyrme interaction are determined , basically by fits to global properties in spherical nuclei over the nuclear chart , and the gap parameters of the usual pairing force and the coupling strength of the residual neutron - proton pairing force are specified , there are no free parameters left . both the residual interaction and the mean field
are consistently obtained from the same two - body force .
this is therefore a reliable method , suitable for extrapolations into the unstable regions approaching the drip lines .
it is worth investigating whether these powerful tools designed to account for the properties of stable nuclei are still valid when approaching the drip lines .
one possible way to establish the validity of the known approaches as well as the limits of their applicability , is a systematic investigation of nuclei covering the whole range from stability towards the drip lines .
exploration of series of isotopes moving away from the region of stability would fulfill these requirements allowing us to learn how the adequacy of the description evolves in progressively more unstable nuclei .
following the same criteria as in our previous work @xcite , we apply this formalism to the study of proton rich krypton isotopes , including even - even as well as odd - a isotopes for the first time .
the reasons why this is a region of special interest to study @xmath3-decay have already been stressed in refs .
they include the large @xmath4-values in proton rich nuclei @xcite , the competition of nuclear shapes @xcite that characterizes this mass region , and the possibility to approach systematically the @xmath5 isotope .
thus , we can test the validity of our formalism and look for discrepancies when approaching the drip lines .
the paper is organized as follows .
we first summarize in sect .
2 the method of calculation . mean field and qrpa with residual spin - isospin interactions are introduced and explained separately for both even - even and odd - a nuclei .
gamow - teller strengths , excitation energies , @xmath6 values and half - lives are also discussed . in sect .
3 we present our results .
we discuss similarities and differences in even - even and odd - a nuclei and analyze in detail the @xmath0-decay observables in kr - isotopes , comparing them with the available experimental data .
4 contains the conclusions and some final remarks .
in this section we summarize briefly the theory involved in the microscopic calculations . more details can be found in refs .
@xcite . our method consists in a selfconsistent formalism based on a deformed hartree - fock mean field obtained with a skyrme interaction , including pairing correlations in the bcs approximation .
we consider in this paper the force sg2 @xcite of van giai and sagawa , that has been successfully tested against spin and isospin excitations in spherical @xcite and deformed nuclei @xcite .
comparison to calculations obtained with other skyrme forces have been made in refs .
@xcite , showing that the results do not differ in a significant way .
the single particle energies , wave functions , and occupation probabilities are generated from this mean field .
for the solution of the hf equations we follow the mcmaster procedure that is based in the formalism developed in ref.@xcite and described in ref.@xcite .
time reversal and axial symmetry are assumed .
the single - particle wave functions are expanded in terms of the eigenstates of an axially symmetric harmonic oscillator in cylindrical coordinates .
we use eleven major shells .
the method also includes pairing between like nucleons in the bcs approximation with fixed gap parameters for protons @xmath7 and neutrons @xmath8 , which are determined phenomenologically from the odd - even mass differences through a symmetric five term formula involving the experimental binding energies @xcite .
the values used in this work are the same as those given in ref .
@xcite . for odd - a nuclei , the fields corresponding to the different interactions were obtained by doing one iteration from the corresponding selfconsistent field of the closest even - even nucleus , selecting the orbital occupied by the odd nucleon according to the experimental spin and parity . for those cases where this experimental assignment is not well established we choose the orbital closer to the fermi level .
the chosen state is blocked from the bcs calculation , and we assign to it a pair occupation probability of 0.5 .
the effect of doing several more iterations from the even - even case , in order to see how the extra particle polarizes the core , was studied in ref . @xcite without observing significant changes .
we repeated these calculations here looking for some effect on the gt strength distributions , but again the changes were negligible . according to this procedure our spin and parity assignments are as follows : @xmath9 for @xmath10kr and @xmath11kr ( as determined experimentally @xcite ) , @xmath12 for @xmath13kr from the most recent experimental determination @xcite . for @xmath14kr and @xmath15kr
we take the spin and parity according to our calculations and assign @xmath16 in the first case and @xmath17 or @xmath18 in @xmath15kr depending on the oblate or prolate shape .
we describe the even @xmath19 odd @xmath20 parent nucleus by removing one neutron from the selfconsistent field of the even - even nucleus and the odd @xmath19 even @xmath20 daughter nucleus by removing one proton .
this is a proper way to describe both the parent and daughter nuclei from the same mean field .
taking @xmath21 as the even - even nucleus for reference , the parent odd - a nucleus is @xmath22 and it decays into the daughter nucleus @xmath23 . as an example , to describe the @xmath24 decay of the odd - neutron parent nucleus @xmath13kr ( z=36 , n=37 ) into the odd - proton daughter nucleus @xmath13br @xmath25 , we use the mean field of the even - even nucleus @xmath26kr @xmath27 . in a previous work
@xcite we analyzed the energy surfaces as a function of deformation for all the even - even isotopes under study here .
for that purpose , we performed constrained hf calculations with a quadrupole constraint @xcite and we minimized the hf energy under the constraint of keeping fixed the nuclear deformation .
calculations in this paper are performed for the equilibrium shapes of each nucleus obtained in that way , that is , for the solutions , in general deformed , for which we obtained minima in the energy surfaces .
most of these nuclei present oblate and prolate equilibrium shapes @xcite that are very close in energy .
as we have mentioned , single - particle energies , wave functions , and occupation probabilities in the odd nuclei , are obtained from the converged mean fields of the even - even neighbor .
therefore , calculations in the odd nuclei are done at the corresponding equilibrium shapes of the even - even generator . in this work , besides the odd - nuclei , we also consider a new even - even isotope , @xmath28kr , that was not previously included in the isotope chain studied in refs .
we then show first in fig .
1 the total energy of @xmath28kr as a function of the mass quadrupole moment @xmath29 .
the results correspond to a constrained hf+bcs calculation with the skyrme forces sg2 and sk3 .
as one can see , we obtain two minima in the energy profile , one is oblate and the other is prolate at about the same value of the quadrupole moment .
these two minima are separated by about 1 mev , the oblate being the deepest one with the two forces .
thus , @xmath28kr follows nicely the trend observed for @xmath30kr isotopes in ref .
@xcite , with an oblate shape favored in a shape coexistent isotope . to describe gamow - teller transitions we add to the mean field a spin - isospin residual interaction , which is expected to be the most relevant interaction for that purpose .
this interaction contains two parts , particle - hole ( @xmath31 ) and particle - particle ( @xmath32 ) .
the @xmath31 part is responsible for the position and structure of the gt resonance @xcite and is derived selfconsistently from the same energy density functional ( and skyrme interaction ) as the hf equation , in terms of the second derivatives of the energy density functional with respect to the one - body densities @xcite .
the @xmath31 residual interaction is finally written in a separable form by averaging the landau - migdal resulting force over the nuclear volume , as explained in refs .
the coupling strength @xmath33 is completely determined by the skyrme parameters , the nuclear radius , and the fermi momentum .
the particle - particle part is a neutron - proton pairing force in the @xmath34 coupling channel .
we introduce this interaction in the usual way @xcite , that is , in terms of a separable force with a coupling constant @xmath35 , which is fitted to the phenomenology .
since the peak of the gt resonance is almost insensitive to the @xmath32 force , @xmath35 is usually adjusted to reproduce the half - lives @xcite .
however , one should be careful with the choice of this coupling constant . since the @xmath32 force is introduced independently of the mean field , if @xmath35 is strong enough it may happen that the qrpa collapses , because the condition that the ground state be stable against the corresponding mode is not fulfilled .
this happens because the @xmath32 force , being an attractive force , makes the gt strength to be pushed down to lower energies with increasing values of @xmath35
. a careful search of the optimal strength can certainly be done for each particular case , but this is not our purpose in this work .
instead , we have chosen the same coupling constant ( @xmath36 mev ) for all nuclei considered here .
this value was obtained @xcite under the requirements of improving in general the agreement with experimental half - lives of even - even isotopes in this mass region while being still far from the values leading to the collapse .
the proton - neutron qrpa phonon operator for gt excitations in even - even nuclei is written as @xmath37\ , , \label{phon}\ ] ] where @xmath38 are quasiparticle creation ( annihilation ) operators , @xmath39 are the excitation energies , and @xmath40 the forward and backward amplitudes , respectively . from the qrpa equations
the forward and backward amplitudes are obtained as @xcite @xmath41 \ , , \label{xphon}\ ] ] @xmath42 \ , , \label{yphon}\ ] ] with @xmath43 the two - quasiparticle excitation energies in terms of the quasiparticle energies @xmath44 .
@xmath45 are given by @xmath46 @xmath47 @xmath48 @xmath49 with @xmath50 where @xmath51s are occupation amplitudes ( @xmath52 ) and @xmath53 spin matrix elements connecting neutron and proton states with spin operators @xmath54 explicit expressions for these matrix elements in the cylindrical harmonic oscillator basis can be found in ref .
note that the expressions for the @xmath55 and @xmath56 amplitudes ( [ xphon])-([yphon ] ) defined here differ from those in refs .
@xcite by the contributions due to @xmath32 residual interactions that were not include there .
the solutions of the qrpa equations are obtained by solving first a dispersion relation , which is of fourth order in the excitation energies @xmath57 .
then , for each value of the energy the amplitudes @xmath58 are determined by using the normalization condition of the phonon amplitudes .
the technical procedure to solve these qrpa equations is described in detail in ref .
@xcite . for even - even nuclei the gt transition amplitudes in the intrinsic frame connecting the qrpa ground state @xmath59 to one phonon states @xmath60 , are given by @xmath61 the ikeda sum rule is always fulfilled in our calculations . the qrpa treatment described in the above subsection is formulated for the excitations of the ground state of an even - even nucleus .
the gt transition amplitudes connecting the ground state of an even - even nucleus ( 0qp state ) to all one phonon states in the odd - odd daughter nucleus ( 2qp states ) are given in the intrinsic frame by eq .
( [ ampleven ] ) .
when the parent nucleus has an odd nucleon , the ground state can be expressed as a 1qp state in which the odd nucleon occupies the single - particle orbit of lowest energy .
then two types of transitions are possible , which are represented schematically in fig .
i ) the first type of transitions are phonon excitations in which the odd nucleon acts only as a spectator .
we call them three quasiparticle transitions ( 3qp ) . in the intrinsic frame ,
the transition amplitudes in this case are basically the same as in the even - even case but with the blocked spectator excluded from the calculation , @xmath62 ii ) the other type of transitions are those involving the odd nucleon .
we call them one quasiparticle transitions ( 1qp ) .
we introduce as usual @xcite phonon correlations to the quasiparticle transitions in first order perturbation , taking into account the part of the gt interaction that contains terms linear in the @xmath63 operator .
the transition amplitudes for the correlated states can be found in ref .
@xcite . here , we give the explicit expression for the transition amplitude corresponding to the kr isotopes , that is , a @xmath64-decay of an odd - neutron parent nucleus decaying into an odd - proton daughter nucleus .
@xmath65 \right .
\nonumber\\ & & \left .
- q^u_{\pi\nu } \sum_{\omega_k } m_{+}^{\omega _ { k } } m_{-}^{\omega _ { k } } \left [ e_\pi(\nu,\omega_k ) + e_\nu(\pi,\omega_k)\right ] \right\ } \nonumber \\ & & + 2\kappa ^{pp}_{gt } \left\ { \tilde{q}_{\pi\nu } \sum_{\omega_k } \left [ m_{+}^{\omega _ { k } } m_{++}^{\omega _ { k } } e_\pi(\nu,\omega_k ) - m_{-}^{\omega _ { k } } m_{--}^{\omega _ { k } } e_\nu(\pi,\omega_k ) \right ] \right . \nonumber\\ & & \left . - q_{\pi\nu } \sum_{\omega_k}\left [ m_{+}^{\omega _ { k } } m_{--}^{\omega _ { k } } e_\pi(\nu,\omega_k ) - m_{-}^{\omega _ { k } } m_{++}^{\omega _ { k } } e_\nu(\pi,\omega_k)\right ] \right\}\ , , \label{ampl1qp}\end{aligned}\ ] ] where @xmath66s and @xmath67s are given in eqs .
( [ m-])-([qs ] ) . in this expression
, @xmath68 is in principle given by @xmath69 and a similar expression for @xmath70 changing @xmath71 into @xmath72 .
nevertheless , in order to avoid accidental singularities , we introduce a width @xmath73 mev in the same way as it was done for the first time in ref .
@xcite @xmath74 for simplicity we speak of 1qp , 2qp , and 3qp states , but it should be clear that we always mean qrpa - correlated states .
once the intrinsic amplitudes @xmath75 are calculated in eqs .
( [ ampl3qp])-([ampl1qp ] ) , the gamow - teller strength @xmath76 for a transition @xmath77 can be obtained as @xmath78 \ }
^2 \nonumber \\ & = & \delta_{k_i , k_f } \left [ \left < i_i k_i 1 0 | i_f k_f \right > \left < \phi_{k_f } \left| \beta ^\pm _ 0 \right| \phi_{k_i}\right > \right .
\nonumber \\ & & \left .
+ \delta_{k_i,1/2}(-1)^{i_i - k_i } \left
< i_i -k_i 1 1 | i_f k_f \right >
\left < \phi_{k_f } \left| \beta ^\pm _ { + 1 } \right| \phi_{\bar{k_i}}\right > \right ] ^2 \nonumber \\ & & + \delta_{k_f , k_i+1 } \left < i_i k_i 1 1 | i_f k_f \right > ^2 \left < \phi_{k_f } \left| \beta ^\pm _ {
+ 1 } \right| \phi_{k_i}\right > ^2 \nonumber\\ & & + \delta_{k_f , k_i-1 } \left < i_i k_i 1 -1 | i_f k_f \right > ^2 \left < \phi_{k_f } \left| \beta ^\pm _ { -1 }
\right| \phi_{k_i}\right > ^2 \ , , \label{strodd}\end{aligned}\ ] ] in units of @xmath79 .
to obtain this expression we have used the initial and final states in the laboratory frame expressed in terms of the intrinsic states @xmath80 using the bohr - mottelson factorization@xcite .
( [ strodd ] ) can be particularized for even - even parent nuclei . in this case
@xmath81 , @xmath82 , and @xmath83 .
@xmath84 concerning the excitation energy of the daughter nuclei to which we refer all the gt strength distributions in this paper , we have to distinguish again between the case of even - even and odd - a parents . in the case of even - even systems , this excitation energy is simply given by @xmath85}=\omega -e_{\pi_0 } - e_{\nu_0 } \
, , \label{eexeven}\ ] ] where @xmath86 and @xmath87 are the lowest quasiparticle energies for protons and neutrons , respectively . in the case of an odd - a nucleus we have to deal with 1qp and 3qp transitions
let us consider here our case of an odd - neutron parent decaying by @xmath64 into an odd - proton daughter . for 1qp transitions ,
since the unpaired neutron state is the only neutron state involved in the transition , the excitation energy is @xmath88 on the other hand , in the 3qp case where the unpaired neutron acts as a spectator , the excitation energy with respect to the ground state of the daughter nucleus is @xmath89 this implies that the lowest excitation energy of 3qp type is of the order of twice the neutron pairing gap . in the kr isotopes under study here , @xmath90 mev @xcite .
therefore , all the strength contained in the low excitation energy region , below typically 3 - 4 mev in the odd - a nuclei studied in this paper must correspond to 1qp transitions .
thus , the low excitation energy region ( below twice the neutron pairing gap ) in a @xmath64 decay from an odd - neutron nucleus , basically tells us about the proton spectrum . we should also mention that in odd - a nuclei we have considered the transitions to the rotational states as well .
we consider only those states built on the @xmath91 band heads , which are not forbidden by the selection rules of the gamow - teller operator in the allowed approximation .
this means that we calculate according to eq .
( [ strodd ] ) the gt strength corresponding to transitions from an initial @xmath92 state to final states characterized by i ) @xmath93 when @xmath94 , ii ) @xmath93 and @xmath95 when @xmath96 , and iii ) @xmath93 , @xmath97 , and @xmath98 when @xmath99 .
accordingly , we have added to the excitation energies @xmath100 a standard bohr - mottelson rotational energy @xcite .
the purpose of the inclusion in our calculation of the transitions to the rotational states is to take into account in a more reliable way the fragmentation of the gt strength and the density of excitation energies .
thus , in our calculations we estimate the moment of inertia by using a mean value between the two extreme macroscopic models , rigid rotor ( rr ) and irrotational flow ( if ) model , whose predictions are usually upper and lower boundary values of the experimental moments of inertia .
@xmath6 values are also determined differently depending on the even or odd number of nucleons .
the @xmath6 value is given by @xmath101c^2 $ ] , where @xmath58 s are the nuclear masses ( binding energies of the electrons have been neglected ) . for the @xmath64-decay of an odd - neutron parent
we have @xmath102 where @xmath103 is the fermi level and @xmath104 is the lowest quasiparticle energy @xmath105 .
the @xmath4-value for @xmath64-decay is simply @xmath106 .
these expressions can be compared to that corresponding to an even - even nucleus @xmath21 @xmath107 the half - lives for the @xmath108-decay are obtained using the standard definition ( see for instance ref.@xcite ) , involving sums over all possible final states within the @xmath109window reached in the decay and including standard quenching factors @xcite .
the fermi integrals required for computation of half - lives have been obtained numerically for each value of the charge of the daughter nucleus @xmath19 and the maximum energy available @xmath110 in @xmath0-decay , as explained in ref.@xcite .
in this section we present and discuss the results obtained for the gt strength distributions , half - lives , and summed strengths in the proton rich kr isotopes .
the results correspond to qrpa calculations with the skyrme force sg2 and they have been performed for the nuclear shapes that minimize the hf energy . before discussing the figures we note that the gt strength distributions are plotted versus the excitation energy of the daughter nucleus .
the distributions of the gt strength in figs . 3 , 5 , and 6 have been folded with @xmath111 mev width gaussians to facilitate the comparison among the various calculations , so that the original discrete spectrum is transformed into a continuous profile .
the distributions in those figures are given in units of @xmath79 and one should keep in mind that a quenching of the strength is expected on the basis of the observed quenching in charge exchange reactions and spin @xmath112 transitions in stable nuclei , where @xmath113 is also known to be approximately @xmath114 . therefore , a reduction factor of about two is expected in these strength distributions in order to compare with experiment .
this factor is of course taken into account when comparison to experimental half - lives , summed strengths , and gt strengths is made .
we can see in fig .
3 the effect of the residual interaction treated in qrpa on the uncorrelated calculation ( dotted lines denoted by mf ) . we do that on the example of @xmath115kr for the oblate and prolate shapes that minimize the energy obtained with the skyrme force sg2 .
the coupling strengths of the @xmath31 and @xmath32 residual interactions are @xmath116 mev and @xmath36 mev , respectively .
both @xmath31 and @xmath32 residual interactions reduce the gt strength .
the residual forces produce also a displacement of the gt strength , which is to higher energies in the case of the repulsive @xmath31 force and to lower energies in the case of the attractive @xmath32 force .
these effects are common to even - even and odd - a isotopes .
nevertheless , we observe that the effect of the residual interaction in odd - a nuclei is very small in the low energy tail of the gt strength distribution .
this is especially true for the @xmath32 force .
the reason for that can be understood from the fact that the lowest - lying transitions are affected by the residual force only through the weak correlations with phonons treated in first order perturbation .
this feature has also important consequences when one considers the half - lives because they depend only on the distribution of the strength below the @xmath117energy window .
therefore , the @xmath32 interaction will affect the half - lives differently if we deal with even - even or odd - a nuclei . in fig .
4 we have plotted the half - lives of the @xmath118kr isotopes as a function of the strength @xmath35 . as we can see from this figure the half - lives decrease with increasing values of @xmath35 , but in a different way depending on the isotopes . we can see that the half - lives of odd - a nuclei have a very smooth , almost flat , behavior with the @xmath32 force , while the half - lives of even - even nuclei present a stronger dependence on the @xmath32 force .
decreasing of half - lives with @xmath35 can be understood from the fact that the @xmath32 force is attractive and therefore tends to concentrate the strength to lower energies .
the reason for the smoother decrease of half - lives in the odd - a isotopes is related to the smaller effect of the @xmath32 force in the low - energy tail of the gt strength distribution in odd - a nuclei , previously discussed . the optimum value of @xmath35 to reproduce the half - life depends , among other factors , on the nucleus , shape , and skyrme interaction and a case by case fitting procedure could be carried out .
since the half - lives are practically insensitive to this force in odd - a nuclei , this fit could be restricted to even - even nuclei . in ref .
@xcite we considered this dependence in various even - even isotopes in the mass region @xmath119 ( ge , se , kr , and sr ) , and arrived to the conclusion that a value of @xmath36 mev improves the agreement with experiment in most cases . thus
, we also use in this paper the same coupling strength for the @xmath32 force . to further illustrate the relative importance of the different types of contributions to the gt strength in odd - a nuclei , we show in fig .
5 the gt strengths ( solid ) decomposed into their 1qp ( dotted ) and 3qp ( dashed ) contributions .
the picture emerging from the analysis of this figure is that the gt strength distributions in odd - a nuclei can be divided into two different regions .
one is the energy region below twice the pairing gap , where the individual excitations are determined by the quasiparticle proton ( neutron ) energies in the case of an odd - neutron ( odd - proton ) parent nucleus .
this region is of relevance for @xmath64-decay since it appears within the @xmath4-window .
the other region at higher energies is dominated by 3qp excitations , where the odd - nucleon acts as a spectator .
the strength contained here is much larger than in the low energy region because many more configurations are possible but only in the very proton rich isotopes this is accessible by @xmath64-decay .
6 contains a summary of the results on gt strengths obtained in this work . in this figure
we can see our hf+bcs+qrpa gamow - teller strength distributions predicted by the skyrme force sg2 for the whole kr isotopic chain including odd - a and even - even nuclei .
the trend observed in the gt strength distributions of both odd - a and even - even isotopes is similar .
we can see that the most unstable isotopes have the largest strengths which are also placed at higher energies .
as we move into the stable isotopes by increasing the number of neutrons , the gt resonance appears at lower excitation energies and contains less and less strength .
also the @xmath6 window , represented by the vertical line , becomes smaller and smaller .
we can observe also the similarity between the even - even and odd - a partners . if we compare the strengths for an even - even @xmath120 nucleus with that of the corresponding odd - a @xmath121 we can see that they are about the same once the low energy region of about twice the neutron pairing gap ( between 3 and 4 mev depending on the case ) is suppressed .
we can also observe that the different profiles of the gt strength distributions corresponding to the various shapes of a given isotope can be used in certain cases as a signature of the nuclear shape .
now we concentrate on the energy region below @xmath6 and discuss the possibilities to discriminate between different shapes by @xmath64-decay experiments .
we also compare with the available experimental data .
we perform this detailed analysis by plotting not the folded strengths as it was done in the previous figures , but the individual excitations as they come from the calculation .
the gt strengths have been quenched with the same factor used to calculate the half - lives . in figs .
7 - 9 we can see these results for the even - even isotopes @xmath122kr , respectively . fig . 7 for @xmath2kr
includes the results from our qrpa calculations with the force sg2 with oblate and prolate shapes . below an excitation energy of about 2 mev
we can see that the distribution of the gt strength predicted by the oblate or prolate shapes is qualitatively very similar although the prolate shape gives somewhat larger strength .
we obtain peaks at 0.5 mev and 1.5 mev in both cases and they are in agreement with preliminary data on this nucleus @xcite .
therefore , it will be hard to distinguish between the two shapes .
on the other hand , if we look into the energy range from 2 mev up to @xmath6 , we can see a strong double peak that appears in the oblate case between 2 and 2.5 mev .
the strength in this region is about three times larger than the strength of the first peak at 0.5 mev and it is almost absent in the prolate case .
the appearance or absence of this peak at 2 mev could be the signature of an oblate or prolate shape , respectively .
it is also worth mentioning the huge peak appearing in the prolate case very close to the @xmath6 value .
if it could be seen experimentally , it would be a clear signature in favor of a prolate shape .
we can see in table 1 the total measured gt strength below 1.836 mev @xcite compared to our results with the two shapes , where oblate shape seems to be favored .
8 shows the results in @xmath26kr .
experimental data are from ref .
we can distinguish two regions in this isotope . at energies below 2 mev
we find that the oblate shape predicts much more strength than the prolate shape and the opposite happens beyond 2 mev .
therefore , measuring the gt strength distribution in @xmath26kr up to @xmath6 would help to discriminate between the two shapes .
if the strength is concentrated below 2 mev it will correspond to an oblate shape , while if it is concentrated between 2 and 3 mev it will correspond to a prolate shape .
a comparison of the strength contained below 1 mev is made in table 1 .
the experimental summed strength @xcite lies between the predictions of the two shapes but it is closer to the strength produced by the oblate shape . in fig .
9 we show the results for @xmath123kr corresponding to the spherical and prolate shapes that minimize the energy in this nucleus . experimental data are from ref .
@xcite . in this case , an isolated single peak in the measured strength at very low excitation energy would be the signature of a spherical parent , while a peak close to the @xmath6 limit would be the signature of a prolate parent .
comparison with the available data seems to favor the spherical shape .
this is also true if we compare the total gt strengths contained below 1 mev , as can be seen in table 1 , where the strength generated by the spherical shape is much closer to experiment .
next figures 10 - 13 correspond to the odd - a nuclei @xmath124kr . in fig .
10 we can see our results for the oblate and prolate shapes of @xmath13kr compared to the experimental data from ref . @xcite .
the first thing to mention is that the experimental data have been taken differently in the two energy ranges below and above 3.5 mev .
below 3.5 mev the data have been extracted from direct detection of the gamma rays . beyond this energy
they have been extracted from proton delayed detection and have big errors .
a new effort to measure experimentally all the energy range up to @xmath6 with large efficiency gamma ray detectors is under way at isolde@xcite . until these data become available and confirm the present measurements , the data measured above 3.5 mev
should be considered as partial because part of the strength escapes observation due to the high level density @xcite .
the summed gt strengths in the two energy regions can be seen in table 1 .
the experimental data appear between the predictions of the oblate and prolate shapes , the oblate one being larger .
comparison to the data of the gt strength distributions below 3.5 mev shows that both oblate and prolate shapes produce similar strength in the very low excitation energy range below 0.5 mev , which is compatible with experiment . at higher energies the prolate shape does not generate strength up to 4 mev while the oblate shape generates two bunches at 1 mev and 3.5 mev that compare better with experiment . above 4 mev , the strength increases and
the prolate shapes produces two bumps at 4.5 and 6 mev , while the oblate shape produces a wide peak from 4.5 to 6.5 mev with a huge strength at an excitation energy of 5.5 mev .
it is also worth mentioning that the structure of the gt transitions in the theoretical calculations can be analyzed taking into account the rotational nature of the final states reached by the allowed gt transition .
as we have already mentioned , since in this case we are considering the transitions @xmath125 with @xmath126 , when we reach a final state with @xmath127 , we consider the rotational states with @xmath128 .
the strengths of these states are given by the geometrical clebsch - gordan factors in eq .
( [ strodd ] ) and their excitation energies by the rotational energies .
similar arguments apply to the transitions to @xmath129 , where we consider the rotational states with @xmath130 . for @xmath131 , the only case to account for is @xmath132 . on this basis
we can now see that the two peaks below 0.5 mev in the prolate case are the members of a rotational band @xmath133 and @xmath134 .
similarly we can identify the low - lying structure in the oblate case as the members of several rotational bands .
this can be seen more easily in fig .
11 , where we have plotted the excitation energies in the daughter nucleus @xmath13br of the states reached by @xmath64-decay . in this figure
we have also added by dotted lines the first forbidden transitions that correspond to @xmath135 with parity change , that is , states @xmath136 .
although we have not calculated their gt strength , their excitation energies have been plotted in the figure .
we can see that the agreement with experiment is very reasonable in the oblate case .
we get concentrations of states at the same energies and the total number of states is similar . in this figure
we can follow more easily the structure of the bands .
for example , in the prolate case we can see that the lowest @xmath12 has a @xmath137 associated , the next @xmath138 has two states @xmath12 and @xmath137 associated and so on . a similar analysis has been done in figs .
12 - 13 for @xmath11kr . in fig .
12 we can see the gt strengths in the prolate and spherical cases compared to the experimental data from ref .
@xcite taken up to 2.5 mev .
the summed strengths can be also compared in table 1 .
the spherical shape does not produce any strength in the @xmath6 window except a small bunch of states around 3 mev . on the other hand the prolate shape generates strength at the right position although a little bit smaller than experiment .
this shape also predicts some strength close to @xmath6 , which is not present in the spherical case .
13 shows the experimental low - lying energy spectrum compared to our calculation for the prolate shape . in our results
we can identify the origin of these excitations .
we have a low - lying @xmath139 rotational triplet and a @xmath140 doublet .
they correspond well with the experimental energies
. we should remember that the spacing in the rotational energies is determined by the moment of inertia , and we are using a rough estimate . using a moment of inertia a little bit larger ,
the energy levels become compressed and agree better with the experimental spacing .
we have also added by dotted lines the energies of first forbidden transitions .
experimental information on the decay of @xmath15kr is available for the low - lying states @xcite . in that reference ,
final states for gt decay were reported to the ground state of @xmath15br and to excited states at 9 and 207 kev . however , the spin - parity assignment of these states , as well as the assignment to the ground state of the parent nucleus @xmath15kr are still controversial @xcite .
we can see in fig .
14 the results of our mean field calculations with the force sg2 for the oblate and prolate shapes of @xmath2kr .
the vertical axis is the neutron occupation probability and the horizontal one is the neutron single - particle energy .
the deformed single - particle states are labelled by @xmath141 .
we can see that the picture is compatible with several spin and parity assignments for the odd - neutron isotope @xmath15kr . if we consider the neutron level closest to fermi level in fig .
14 , the spin - parity assignment would be @xmath142 for the oblate shape and @xmath12 for the prolate shape . on the other hand , the assumption taken in ref .
@xcite was @xmath137 , which is also close to fermi level .
we therefore consider these two possibilities for each shape .
we show in fig .
15 the results for the gt strengths from our calculations assuming the above mentioned possibilities for the spin and parity of the parent nucleus , @xmath137 as it was taken in ref .
@xcite , as well as @xmath142 in the oblate case and @xmath12 in the prolate case .
we can also see the spin and parity of each gt excitation .
experimental @xmath6 and @xmath143 values are plotted as circles in fig .
16 for the krypton isotopes considered in this work .
the data have been taken from ref .
@xcite , except for the isotopes @xmath144kr , where we have used the more recent data from ref .
these values are compared to our theoretical results represented by the vertical lines .
the extreme values of these vertical lines correspond to the results obtained from the various shapes using the forces sg2 and sk3 .
this has been done in order to have a better idea of the theoretical spread of the results .
the agreement is in general good for both @xmath6 and half - lives .
this is notorious because there is a very large range of variation ( seven orders of magnitude ) for half - lives .
@xmath6 values are well reproduced with the exception of @xmath28kr , where we obtain a value below experiment .
we can also see that the calculations fail to account for the half - lives of the most unstable nuclei @xmath28kr and @xmath15kr . nevertheless , it should be mention that the calculations correspond to gt transitions neglecting possible contributions from fermi transitions .
but fermi transitions might play a significant role to obtain the total decay rates in nuclei where @xmath145 .
indeed , we have calculated the fermi strength distributions and evaluated the corresponding half - lives .
the calculation of fermi transitions follows closely that of gamow - teller transitions .
one should simply use an isospin - isospin residual interaction with a coupling strength derived from the skyrme force as shown in ref . @xcite and replace the spin matrix elements between proton and neutron states given in eq .
( [ sigma ] ) by the overlaps @xmath146 .
of course , only @xmath147 components will survive .
we find that the half - lives corresponding to fermi transitions are negligible as compared to the gt half - lives except in the isotopes @xmath148kr , where they are comparable and therefore one should not neglect their contribution .
we can see in fig .
16 , the new results for the total half - lives once the contribution from fermi transitions have been included .
they are given by the vertical lines located at the right side in the half - lives of the @xmath148kr isotopes .
the agreement with experiment improves in the cases of @xmath144kr but we are still above the experiment .
nevertheless , it should be mentioned that if we use the experimental @xmath6 value in @xmath28kr , instead of the @xmath6 result from our calculations , we get agreement with the experimental half - life . thus , the discrepancy found in the half - life of @xmath28kr is entirely due to the discrepancy in the @xmath6 value .
we can observe in fig .
16 a nice trend that can be followed by the auxiliary dotted lines joining the experimental data of even - even isotopes on one hand and the odd - a on the other .
the growing @xmath6 trends in the two lines of fig .
16 are dictated by the increase of ( @xmath149 ) as one approaches the proton drip line .
this behavior can be qualitatively understood in the case of @xmath6 from their expressions given by eqs .
( [ qecoddn])-([qeceven ] ) .
as we have mentioned earlier , the only exception to this trend appears in the experimental @xmath6 value of @xmath28kr , which is clearly above the theoretical value . since in this case
, the daughter nucleus @xmath28br is odd - odd @xmath5 , one may argue that the observed jump from the regular trend in @xmath6 and @xmath143 is due to an extra neutron - proton binding , which is not taken into account in the present calculation .
we have applied a selfconsistent deformed hf+bcs+qrpa formalism with density - dependent effective skyrme interactions to the description of the @xmath0-decay properties of proton - rich odd and even kr isotopes .
this approximation has the appealing feature of treating the excitations and the ground state in a selfconsistent framework with basically no free parameters .
our spin - isospin residual interaction contains a particle - hole part , which is derived selfconsistently from the skyrme force , and a particle - particle part , which is a separable force representing a neutron - proton pairing force .
we have analyzed the similarities and differences in the treatment and in the results of even - even and odd - a nuclei .
the low - lying gt response of odd - a nuclei is generated by transitions involving the state of the odd nucleon ( 1qp transitions ) . thus , the excitation spectrum up to twice the pairing gap parameter of neutrons ( protons ) in the @xmath64-decay of and odd neutron ( proton ) parent , gives information on the proton ( neutron ) states .
when this low energy strength in the odd - a nuclei is removed , the resulting gt strength distribution is very similar to that of the even - even neighbor displaced to higher excitation energy @xmath150 . in odd - a deformed nuclei , for each allowed intrinsic gt transition , one has a set of transitions to rotational states with decreasing strength at higher energies .
this is a characteristic of odd - a deformed nuclei , which is not present in the spherical limit .
the gt strength corresponding to transitions to states in a given band is reduced by geometrical factors involving angular momenta .
the energy separation depends on the angular momentum of the transition with a global scale determined by the moment of inertia .
we have found a reasonable agreement with available experimental data .
we have also discussed what can be learned from future comparison of our results with experimental data from isolde that are expected very soon .
this comparison would be interesting for several reasons : i ) from our study of the dependence on the shape of the gt strength distributions we conclude that information on the shape of the parent nucleus can be gained when data in the whole @xmath6 window become available .
we have identified particular narrow regions in the excitation spectra of various nuclei where data could be more conclusive on the nuclear shape ; ii ) comparison to data would be desirable before considering further theoretical refinements , such as the effects of the continuum , the neutron - proton pairing at the mean field level or extensions of the rpa ; iii ) an important point will be to check whether our calculations produce the same level of agreement in all the isotopes or some special characteristics can be found in @xmath5 nuclei . for the moment ,
an interesting feature found is the deviation between theory and experiment in the half - life of @xmath28kr . from the present calculation
, this deviation can be attributed to the difference between experimental and theoretical @xmath6 values , which in turn can be a signature of an extra binding in the @xmath5 odd - odd daughter nucleus . c. dtraz and d.j .
vieira , annu .
nucl . part .
* 39 * ( 1989 ) 407 ; a. mueller and b.m .
sherrill , annu .
nucl . part .
sci . * 43 * ( 1993 ) 529 ; nupecc report on _ nuclear structure under extreme conditions of isospin , mass , spin and temperature _ ,
( 1997 ) .
m. hirsch , a. staudt , k. muto , and h.v .
klapdor - kleingrothaus , nucl . phys .
* a535 * ( 1991 ) 62 ; k. muto , e. bender , and h.v .
klapdor , z. phys .
a * 333 * ( 1989 ) 125 ; h. homma , e. bender , m. hirsch , k. muto , h.v .
klapdor - kleingrothaus , and t. oda , phys .
c * 54 * ( 1996 ) 2972 .
h. flocard , p. quentin , and d. vautherin , phys .
b * 46 * ( 1973 ) 304 ; m. beiner , h. flocard , n. van giai , and p. quentin , nucl . phys . * a238 * ( 1975 ) 29 ; p. quentin and h. flocard , annu . rev .
nucl . part .
* 28 * ( 1978 ) 253 ; p. bonche , h. flocard , p.h .
heenen , s.j .
krieger , and m.s .
weiss , nucl .
a443 * ( 1985 ) 39 .
a. petrovici , k.w .
schmid , and a. faessler , nucl .
* a605 * ( 1996 ) 290 ; a. petrovici , k.w .
schmid , a. faessler , j.h .
hamilton , and a.v .
ramayya , prog . in part . and
nucl . phys .
* 43 * ( 1999 ) 485 ; a. petrovici , k.w .
schmid , and a. faessler , nucl .
a665 * ( 2000 ) 333 .
d. lode , w. pessara , h. ohlsson , and e. roeckl , z. phys .
* 260 * ( 1973)253 ; t. paradellis , a. houdayer , and s.k .
mark , nucl .
* a201 * ( 1973 ) 113 ; b. singh and d.a .
viggars , nuclear data sheets * 42 * ( 1984 ) 233 .
j. bea , ph.d .
thesis , universidad de valencia ( 1995 ) ; j. bea , b. rubio , a. gadea , l. garcia - raffi , j. rico , j.l .
tan , m.j.g .
borge , o. temblad , s. skoda , j. eberth , h. keller , r. kirchner , and e. roeckl , _ detailed study of the @xmath11kr beta decay properties _ , gsi scientific report ( 1993 ) 55 .
* table 1 .
* comparison of the gt strengths contained below some given excitation energy between experimental measurements ( @xcite for @xmath2kr , @xcite for @xmath13kr , @xcite for @xmath26kr , @xcite for @xmath11kr , and @xcite for @xmath123kr ) , and theoretical calculations | @xmath0-decay properties of proton - rich odd - a and even - even krypton isotopes are studied in the framework of a deformed selfconsistent hartree - fock calculation with density - dependent skyrme forces , including pairing correlations between like nucleons in bcs approximation .
residual spin - isospin interactions are consistently included in the particle - hole and particle - particle channels and treated in quasiparticle random phase approximation . the similarities and differences in the treatment of
even - even and odd - a nuclei are stressed .
comparison to available experimental information is done for gamow - teller strength distributions , summed strengths , and half - lives .
the dependence of these observables on deformation is particularly emphasized in a search for signatures of the shape of the parent nucleus . |
massive stars are a driving force of physical and chemical evolutions of their host galaxies .
supernova explosions and their remnants have been intensively studied for decades , but the pre - explosion effects are equally important ; the integrated mass , momentum , and energy releases over the lifetime of an o star can be comparable to those by a supernova explosion at the end of their lives @xcite .
diffuse x - ray emission is generated as a consequence of shocks by stellar winds impinging on the interstellar medium ( ism ) .
therefore , we can quantitatively study the effects of energy dissipation and the chemical enrichment of interstellar space by early - type stars through the spectroscopy of x - ray emission in h regions .
@xcite presented a self - similar solution of stellar winds interacting with the ism .
they showed a single o7 star forms a hot ( @xmath110@xmath3 k ) bubble by the shock , which can be observed as soft extended ( @xmath110 pc ) x - ray emission .
@xcite claimed the first unambiguous detections of such emission in the h regions m17 and the rosette nebula using the advanced ccd imaging spectrometer ( acis ; @xcite ) onboard the chandra x - ray observatory @xcite . with a @xmath140 ks integration time of m17 ( obsid@xmath4972 ) ,
diffuse soft x - ray emission was detected apart from 886 point sources above @xmath110@xmath5 erg s@xmath6 @xcite .
@xcite showed the entire structure of the soft x - ray diffuse emission using the position - sensitive proportional counter ( pspc ; @xcite ) onboard rosat @xcite .
they measured the total x - ray luminosity and compared to the wind - blown bubble models with and without heat conduction .
they concluded that only the bubble without heat conduction can account for the observed x - ray luminosity of @xmath7 erg s@xmath6 .
the magnetic field may be responsible for suppressing the heat conduction and mass evaporation between the hot gas and cold ism @xcite .
the results obtained by these high - resolution imaging studies are generally consistent with the wind - blown bubble models
. however , the observational results are not still accurate enough to compare to the theoretical works .
it is often ambiguous whether the observed diffuse emission is from wind - blown bubbles or from supernovae , which give rise to diffuse emission with a similar x - ray spectral hardness and luminosity in a similar spatial scale .
the largest uncertainty stems from the lack of spectral analysis based on line diagnostics in a spatially - resolved manner .
resolving emission lines is crucial to examine whether the spectrum is thermal and to determine the temperature and the chemical composition of the plasma .
for example , the intensity ratio of k@xmath0 lines between o and o and that between ne and ne are steep functions of the plasma temperature at 110 mk @xcite . in between the o and ne k@xmath0 complex , fe l series lines dominate the spectrum .
the metallicity of these elements is one of the factors to determine the x - ray luminosity expected from a wind - blown bubble @xcite .
the anomaly in the o and fe abundance ratio can be used to discriminate different types of supernovae @xcite if the emission is of a supernova origin .
the previous studies using rosat and chandra were incapable of resolving these lines , limiting their ability to diagnose the plasma emission . the x - ray imaging spectrometer ( xis ; @xcite ) onboard suzaku @xcite has a superior spectral resolution , a low background , and a large effective area , which are particularly suited for spectroscopy of extended x - ray emission .
the capability to resolve key elements with sufficient statistical significance has been illustrated by several initial studies on extended emission in h regions @xcite .
@xcite resolved various emission lines from the diffuse emission in the carina nebula .
based on the low nitrogen - to - oxygen ratio and the spatial variation of the fe and si abundances , they suggested that the diffuse emission originates not from wind - blown bubbles but from one or multiple old supernova remnant(s ) .
m17 is a galactic h region at a distance of @xmath11.6 kpc @xcite .
@xcite identified nine o stars in the central ob association with near infrared spectroscopy
. a much larger number of young ob stars are suggested by near infrared photometry @xcite .
the earliest system is a binary of two o4o5 stars .
the age of the cluster is estimated to be @xmath81 myr based on the h - r diagram @xcite .
strong winds and radiation from the central ob association sculpted the ambient matter to form a -shaped cloud , which was traced by molecular and atomic hydrogen lines @xcite .
the diffuse x - ray emission found by chandra and rosat has an asymmetric morphology with respect to the ob association and fills the cavity of the molecular cloud toward the negative galactic latitude @xcite .
m17 is suitable for x - ray studies of hot bubbles in h regions for being proximate and having a very high contrast of the diffuse emission against point sources . from the chandra study @xcite
, the diffuse emission is more intense than the integrated emission of resolved point sources by more than ten - fold at 1 kev . despite the limited spatial resolution of suzaku ,
therefore , the obtained spectrum is not seriously contaminated by unresolved point sources .
suzaku observed m17 on 2006 march 1114 in the first announcement of opportunity observing cycle .
suzaku has xis and hard x - ray detector ( hxd ; @xcite ) instruments .
we concentrate on the xis data in this paper , which has a sensitivity for the soft emission studied here .
the xis is equipped with four x - ray ccds .
three of them ( xis0 , 2 , and 3 ) are front - illuminated ( fi ) ccds and the remaining one ( xis1 ) is a back - illuminated ( bi ) ccd .
each ccd chip has a format of 1024@xmath91024 pixels and is composed of four segments of 256@xmath91024 pixels .
fi and bi ccds are superior to each other in the hard and soft band responses , respectively .
they are mounted at the focus of four independent x - ray telescopes ( xrt ; @xcite ) .
the detectors are sensitive in the energy range of 0.212.0 kev with an initial energy resolution of @xmath165 ev in the full width at half maximum and a total effective area of @xmath11360 @xmath10 at 1.5 kev .
an xis field of view covers a @xmath118@xmath918 region with a half power diameter of @xmath12 .
the radioactive sources of illuminate two corners of each of the four ccds for calibration purposes .
the performance of xis is subject to degradation due to the radiation damage in the orbit @xcite . as of the observation date ,
the energy resolution in the full width at half maximum is @xmath190 ev at 1.5 kev .
the relative energy gains among xis chips and segments have a systematic uncertainty of @xmath15 ev .
an unknown contaminant accumulates on the optical blocking filters of the xis in the orbit . as a result ,
the effective area at the soft band end has diminished significantly , making the carbon and nitrogen features difficult to detect below @xmath10.5 kev .
this effect is included in the auxiliary response .
the observation was conducted using the normal clocking mode with a frame time of 8 s. data ( revision 1.2 ) were screened to remove events during the south atlantic anomaly passages , at elevation angles below 4@xmath11 from the earth rim , and at elevation angles below 10@xmath11 from sunlit earth rim .
we constructed plots of the raw count rate versus the elevation angles to find that these criteria maximize the exposure time with negligible contaminating emission .
after the filtering , the net integration time is @xmath1110 ks .
( 85mm,85mm)f1a.eps ( 85mm,85mm)f1b.eps figure [ fg : f1 ] shows the xis images of the study field in the ( a ) soft ( 0.51.5 kev ) and ( b ) hard ( 1.55.0 kev ) bands .
the two band - limited images appear strikingly different . in the hard band ,
the image is dominated by the emission from the ob association .
we also see the excess emission from a group of protostars m17 north @xcite . in the soft band ,
the extended emission emerges in the eastward of the ob association , as was claimed by @xcite and @xcite .
the xis observation was centered at the most intense part of the extended emission at ( r.a .
, decl . ) @xmath1 ( , ) in the equinox j2000.0 , while the acis observation was at the ob association @xcite .
the xis and acis images have a similar size ( @xmath118 and @xmath117 arc - minute square , respectively ) with a @xmath170% overlapping area ( figure [ fg : f1]b ) . the astrometry of the xis frame was registered using 1wgaj1820.601615 found in both images .
the o8 star @xcite is bright , isolated , and point - like in the acis image ( cxouj182035.87161542.5 ; @xcite ) , thus serves as a good astrometric calibrator .
we shifted the xis frame by @xmath117 to the north so that the position matches with that by the acis observation .
the chandra frame is accurate to @xmath105 in the astrometry .
we first examine the spectrum of the entire diffuse emission .
we extracted the source spectrum from a rectangular region and the background spectrum from a region devoid of intense diffuse emission ( figure [ fg : f1]a ) .
because the off - axis angles of the source and background regions are different , we processed the raw spectra in the following way before subtracting the background from the source : ( 1 ) the non x - ray - background ( nxb ) spectrum was subtracted , which was constructed from night earth data at the same extraction region .
the nxb of xis is a function of the geomagnetic cut - off rigidity .
we therefore compiled night earth observations such that the cut - off rigidity distribution becomes the same with that of the m17 observation .
( 2 ) the vignetting was corrected by multiplying the effective area ratios between the source and background regions for each energy bin of the background spectrum .
this takes into account the accumulating contaminant on the xis optical blocking filter .
the merged fi and the bi spectra are shown in figure [ fg : f2 ] .
in the merged fi spectrum , we added the three fi spectra to increase the photon statistics , because the redistribution matrix functions ( rmfs ) and the auxiliary response functions ( arfs ) are essentially the same for these chips . on the other hand , we handled the bi spectrum separately for its different response .
( 85mm,85mm)f2.eps we resolved the emission lines clearly for the first time , which include k shell lines of o , ne , and mg as well as l shell lines of fe .
this indicates that the emission is of a thermal origin .
the o line is much stronger than the o line and the ne line is so than the ne line if we taking the energy dependence of the efficiencies into account .
these line ratios alone infer that the plasma temperature is in the range of 2.54 mk @xcite even without spectral model fittings . in order to investigate the spatial difference of the plasma properties
, we divided the source region into three sub - regions ( a , b , and c in figure [ fg : f1]a ) based on the morphology of the diffuse emission .
we constructed the spectra from each region and subtracted the background in the same manner for the entire emission .
the merged fi and the bi spectra in each sub - region are shown in figure [ fg : f3 ] .
( 85mm,240mm)f3.eps we fitted the 0.41.8 kev spectra with a thin - thermal plasma model at a collisional equilibrium ( the apec model ; @xcite ) convolved with the interstellar absorption @xcite .
the abundances of the noticeable elements ( o , ne , mg , and fe ) were free parameters .
those of the other elements were fixed at 0.3 solar , which is canonically used in x - ray spectroscopy in star - forming regions ( e.g. , @xcite ) .
we used the rmfs ( version 2006 - 08 - 01 ) of the observation month and generated arfs using a ray - tracing simulator ( ` xissimarfgen ` version 2006 - 08 - 28 ; @xcite ) assuming that the emission is uniform across a 15 radius circle centered at the optical axis . in order to compensate for the possible uncertainty in the energy gain calibration
, we introduced an additional fitting parameter ( offset ) .
the resultant offset values were 14 ev , which are within the current calibration limitation .
a single temperature model yielded acceptable fits for all spectra . the best - fit hydrogen - equivalent column density ( @xmath12 ) , plasma temperature ( @xmath13 ) , metallicity ( @xmath14 , @xmath15 , @xmath16 , and @xmath17 ) , the average surface brightness ( @xmath18 ) , and the luminosity ( @xmath19 ) in the 0.52.0 kev band
are summarized in table [ tb : t1 ] .
the best - fit models are shown in figure [ fg : f3 ] .
we attempted different plasma models with multiple temperatures ornon - equilibrium ionization , but did not obtain improved fits .
we therefore consider that a single temperature model at a collisional equilibrium is adequate .
we also constructed spectra of the three discrete sources ( 1wgaj1820.61615 , m17 north , and the ob association in figure [ fg : f1]b ) .
the source signals were accumulated from elliptical regions of 2035 axis lengths , while the background signals were from adjacent regions free of sources .
the spectra are shown in figure [ fg : f4 ] .
all the spectra are characterized by hard emission full of k@xmath0 emission lines of highly ionized ions , which include mg , mg , si , si , s , s , ar , ar , ca , and fe .
( 85mm,240mm)f4.eps we can have a crude estimate for spectral models from the intensity of these lines .
the spectra of 1wga j1820.61615 and the ob association show a mg line stronger than mg as well as fe and a hard continuum up to @xmath18 kev , requiring at least two thermal components of different temperatures .
m17 north shows a mg line and no prominent mg line , thus a single temperature model would be adequate .
we fitted these spectra using the attenuated thin - thermal plasma model similarly for the diffuse emission .
the merged fi and the bi spectra were simultaneously fitted for 1wgaj1820.61615 and the ob association .
only the merged fi spectrum was fitted for the m17 north because its bi spectrum is strongly contaminated by the calibration source .
we first fitted the spectra with a one - temperature model .
an additional component with a different temperature was added if the fitting was rejected due to systematic residuals . as the crude estimate , the spectra of 1wgaj1820.61615 and the ob association required a two - temperature model , while that of the m17 north was fitted by a one - temperature model .
the best - fit models and parameters are shown in figure [ fg : f4 ] and table [ tb : t2 ] , respectively .
llccc pars . & units & + & & ( a ) & ( b ) & ( c ) + @xmath12 & ( 10@xmath20 @xmath21 ) & 4.8 ( 4.45.2 ) & 4.6 ( 4.45.2 ) & 4.3 ( 3.74.6 ) + @xmath13 & ( kev ) & 0.25 ( 0.240.27 ) & 0.24 ( 0.220.25 ) & 0.27 ( 0.250.28 ) + @xmath14 & ( solar ) & 0.10 ( 0.060.14 ) & 0.15 ( 0.110.18 ) & 0.13 ( 0.100.16 ) + @xmath15 & ( solar ) & 0.20 ( 0.170.24 ) & 0.32 ( 0.250.39 ) & 0.22 ( 0.190.26 ) + @xmath16 & ( solar ) & 0.10 ( 0.090.13 ) & 0.12 ( 0.070.16 ) & 0.12 ( 0.090.16 ) + @xmath17 & ( solar ) & 0.10 ( 0.080.13 ) & 0.19 ( 0.140.20 ) & 0.12 ( 0.110.14 ) + @xmath18 & ( 10@xmath22 erg s@xmath6 @xmath21 arcmin@xmath23 & 2.16 ( 2.122.21 ) & 1.34 ( 1.311.38 ) & 1.55 ( 1.521.58 ) + @xmath19 & ( 10@xmath24 erg s@xmath6 ) & 1.3 & 1.0 & 1.2 + @xmath25/d.o.f & & 155.5/146 & 121.8/114 & 176.9/129 + llccc pars . & units & 1wgaj1820.601615 & m17 north & ob association + @xmath12 & ( 10@xmath26 @xmath21 ) & 1.7 ( 1.42.0 ) & 2.4 ( 2.12.7 ) & 1.3 ( 1.21.4 ) + @xmath27 & ( kev ) & 3.8 ( 3.14.7 ) & 2.7 ( 2.43.2 ) & 4.0 ( 3.94.1 ) + @xmath28 & ( kev ) & 0.56 ( 0.470.65 ) & ... & 0.59 ( 0.560.62 ) + @xmath15 & ( solar ) & 0.3 & 0.3 & 0.51 ( 0.340.71 ) + @xmath16 & ( solar ) & 0.49 ( 0.240.87 ) & 0.3 & 0.51 ( 0.380.65 ) + @xmath29 & ( solar ) & 0.30 ( 0.160.58 ) & 0.3 & 0.53 ( 0.390.71 ) + @xmath30 & ( solar ) & 1.07 ( 0.581.72 ) & 0.3 & 1.17 ( 0.981.38 ) + @xmath31 & ( solar ) & 0.3 & 1.67 ( 0.723.00 ) & 1.44 ( 1.001.90 ) + @xmath32 & ( solar ) & 0.3 & 0.3 & 0.46 ( 0.010.90 ) + @xmath17 & ( solar ) & 0.33 ( 0.190.48 ) & 0.36 ( 0.220.50 ) & 0.28 ( 0.250.31 ) + @xmath33 & ( 10@xmath34 erg s@xmath6 @xmath21 ) & 3.9 ( 3.84.0 ) & 7.4 ( 7.17.7 ) & 58.3 ( 57.858.8 ) + @xmath19 & ( 10@xmath35 erg s@xmath6 ) & 3.2 & 4.9 & 30.0 + @xmath36/d.o.f . & & 100.2/134 & 66.4/76 & 473.3/405 +
the spectra of the diffuse emission in figures [ fg : f2 ] and [ fg : f3 ] are contaminated by other sources of emission .
we evaluate the levels of contamination by unresolved point sources , the galactic ridge x - ray emission ( grxe ) , the cosmic x - ray background ( cxb ) , and the local hot bubble ( lhb ) , and argue that their contributions to the background - subtracted spectra are negligible .
about 70% of the xis field is covered in the chandra observation ( figure [ fg : f1]b ) , which has a much better spatial resolution and sensitivity for faint point sources @xcite . among the three sub - regions ,
the region ( a ) has a complete coverage by chandra .
we therefore use this sub - region as a representative to evaluation the levels of various contaminations .
first , we examine the contribution of unresolved point sources .
we extracted 19 point sources from the acis data in the sub - region ( a ) , constructed the composite spectrum , and fitted it with a thermal plasma model .
the spectrum of each source is too poor to fit individually , so we assume that all sources have the same spectral shape that best describes the composite spectrum .
using their positions , flux , and the assumed spectral shape , we generated their xis events using a ray - tracing simulator ( ` xissim ` ; @xcite ) . in figure
[ fg : f5 ] , we compare the integral of the simulated spectra of unresolved point sources ( ps1 ) to the observed diffuse spectrum ( pluses ) . in the displayed observed spectrum
, we subtracted the nxb spectrum but not the background spectrum in the neighboring region .
the point source contribution accounts for @xmath18% of the emission in the 0.41.8 kev band .
( 85mm,60mm)f5.eps to have an estimate of the maximum contamination to the lines , we repeated the same procedure using the model with the maximum allowable abundance values in the best - fit model for simulating unresolved point source events . the resultant integrated spectrum ( ps2 )
is also shown in figure [ fg : f5 ] . at the ne line at 0.92 kev ,
the contribution by the point sources accounts only for @xmath3710% of the observed emission .
second , we derive the grxe contribution , which is ubiquitous along the galactic plane at galactic longitudes ( @xmath38 ) of @xmath39 45 @xcite .
we refer to a chandra result @xcite for the spectral shape in the soft band and to a rossi x - ray timing explorer result @xcite for the surface brightness at the position of m17 ( @xmath40 15 ) .
the grxe contribution to the observed emission is estimated to be @xmath13% and is shown in figure [ fg : f5 ] .
third , for the cxb contribution , we consulted the suzaku xis observation of the north ecliptic pole @xcite both for the spectral shape and the surface brightness .
the cxb was observed and fitted by a power - law model .
we convolved the model with the xis responses and found that the contribution is @xmath12% . finally , we constrain the contribution by the lhb emission .
we estimate its surface brightness to be @xmath14@xmath910@xmath41 counts s@xmath6 arcmin@xmath42 in the pspc r1 and r2 bands @xcite .
we assume that the spectrum is a thin - thermal plasma ( the raymond - smith model ; @xcite ) with a temperature of 0.1 kev . with the derived emission measure of @xmath12.8@xmath43 @xmath44 pc
, the lhb contributes @xmath12% of the observed emission .
none of the above components play a significant role in the observed diffuse spectrum .
moreover , we removed most of them by subtracting a background spectrum from a neighboring region .
the background events show no indication of time variability caused by solar flares @xcite .
we conclude that the background - subtracted xis spectrum represents the spectrum of the uncontaminated diffuse emission quite well .
the background spectrum accounts for @xmath135% of the source spectrum of the sub - region ( a ) .
besides the nxb contributing @xmath17% of the source flux , the sum of unresolved point sources , grxe , cxb , and lhb contributes @xmath115% .
therefore , emission with a flux of @xmath113% of the source flux is additionally included in the background spectrum .
the contribution from the bright sources outside of the extraction region is negligible .
we attribute the remaining emission to the diffuse emission in the background region ( figure [ fg : f1]a ) .
if this is the case , the flux estimate of the diffuse emission ( table [ tb : t1 ] ) is underestimated by @xmath113% .
we compare our results with the previous works using rosat @xcite and chandra @xcite .
we derived that the diffuse spectrum is explained by an absorbed single temperature thin - thermal plasma model of @xmath45 0.25 kev and @xmath46 4.5@xmath910@xmath20 @xmath21 .
the total luminosity ( 0.52.0 kev ) in the combined ( a)@xmath47(b)@xmath47(c ) region is @xmath13.5@xmath910@xmath24 erg s@xmath6 ( table [ tb : t1 ] ) .
the rosat study @xcite shows that the total luminosity of the diffuse emission is @xmath12.5@xmath910@xmath24 erg s@xmath6 . the smaller estimate than our result is more noticeable if we consider that the rosat value was derived from a larger area and in a wider energy range ( 0.12.4 kev )
this stems from an underestimate of the extinction . without a sufficient spectral resolution to resolve lines ,
two different models were not disentangled in the rosat pspc spectra ; one is a low plasma temperature with a large extinction ( @xmath48 kev and @xmath49 @xmath21 ) and the other is a high plasma temperature with a small extinction ( @xmath10.7 kev and @xmath50 @xmath21 ) .
@xcite derived the luminosity based on the latter , but our spectroscopy shows that the former should be in the case .
the best - fit xis values of @xmath19 and @xmath12 are consistent with those presented in the chandra study @xcite , in which @xmath51 3.4@xmath910@xmath24 erg s@xmath6 ( 0.52.0 kev ) , and @xmath52 ( 4 @xmath2 1)@xmath910@xmath20 @xmath21 .
however , the plasma temperatures are different between the two studies . in chandra ,
the primary component of @xmath53 0.6 @xmath2 0.1 kev and the secondary component of @xmath10.13 kev were claimed . in suzaku , however , we confirmed that a single temperature component of @xmath10.25 kev is adequate from the diagnosis of resolved oxygen and neon lines . from the discrete sources , we detected k@xmath0 emission lines from highly ionized ions ( figure [ fg : f4 ] ) .
the spectrum of the ob association is comprised of hundreds of point sources , but the emission from an o4o5 binary dominates the spectrum .
@xcite claimed that both components of the binary ( sources 543 and 536 ) have plasma temperatures exceeding 10 kev .
however , the strong fe k@xmath0 line at 6.7 kev and the weak fe k@xmath0 line at 7.0 kev in the xis spectrum ( figure [ fg : f4]c ) do not support such high temperatures .
a high signal - to - noise ratio spectrum by xis enabled us to conduct spatially - resolved spectroscopy of the diffuse emission .
the plasma temperature and the chemical composition are uniform , except possibly for a larger metallicity in the sub - region ( b ) .
the chandra study @xcite also show no evidence for spatial variation of plasma temperature .
the observed uniformity indicates that the entire plasma is at a thermal equilibrium in the observed spatial scale of @xmath15 pc , unless the plasma is patchy at equilibria locally by magnetic confinement .
the global equilibrium is reasonable considering the fact that the plasma sound crossing time ( @xmath12@xmath910@xmath54 yr ) is much smaller than the time scale of the system ( @xmath110@xmath55 yr ) , thus the constant pressure is achieved @xcite . here
, we used the plasma volume and the electron density as @xmath130 pc@xmath56 and @xmath11 @xmath57 , respectively , by assuming that the plasma distribution has a conical shape with its apex at the ob association and with a filling factor of 1 . given the uniformity of the plasma temperature and pressure , we speculate that the density is also spatially uniform .
the observed surface brightness , however , is different among the three sub - regions ( figure [ fg : f1]a , table [ tb : t1 ] ) .
it is @xmath11.6 times more intense in the sub - region ( a ) than ( b ) .
this is not attributable entirely to the different extinction , as @xmath12 is larger in the sub - region ( a ) than in ( b ) . because the plasma has a uniform temperature and density , we speculate that the difference of the surface brightness is likely due to the different line - of - sight depths or different filling factors . with the derived plasma volume and the density , the total mass of the plasma is @xmath11 @xmath58 .
this is comparable to the integrated ejecta mass by stellar winds at a mass loss rate of @xmath110@xmath59 @xmath58 for @xmath110@xmath55 yr and agrees with the estimates by the previous works @xcite .
we also speculate that the swept - up and evaporated ism does not make a significant contribution to the plasma mass .
the chemical composition of the diffuse emission is revealed for the first time in this study .
the metallicity of the diffuse emission is 0.10.3 solar ( table [ tb : t1 ] ) , which is significantly lower than those derived for the discrete sources ( 0.31.5 solar ) .
the metal abundances relative to each other in the diffuse emission are consistent with solar among o , mg , and fe @xcite , but ne is enhanced in all three sub - regions by a factor of @xmath12 ( table [ tb : t1 ] ) .
this is also evident in the simultaneous spectral fits of the three sub - regions , in which we tied the abundance values of these elements .
the resultant values are @xmath60 0.12 ( 0.110.13 ) , @xmath61 0.22 ( 0.210.25 ) , @xmath62 0.10 ( 0.080.12 ) , and @xmath63 0.12 ( 0.110.13 ) solar .
such ne enhancement from other metals is widely seen in coronally active stars @xcite for unknown reasons .
we consider that the ne enhancement is an intrinsic feature of the diffuse plasma , and is not influenced by the contamination of ne - enhanced point sources spectra ( [ sect : s4 - 1 ] ) .
one explanation for the anomaly is that the poorly - constrained solar ne abundance is underestimated by a factor of a few @xcite .
this would account for the observed ne enhancement in the m17 diffuse plasma as well .
the lack of a clear spatial variation of the chemical composition comprises a sharp contrast to the diffuse emission in the carina nebula @xcite , where the different abundance patterns across a similar spatial scale suggest the supernova origin for the emission . in the diffuse emission in m17 ,
the o and fe ratio is consistent with the solar abundance , which is another line of evidence against the supernova interpretation .
if a supernova has occurred in m17 , it should have been caused by a star earlier than the earliest ( o4o5 ) star in the ob association .
such a massive source causes a core - collapse type supernova .
it would have yielded a measurably larger ratio of o against fe than the solar value by a factor of a few @xcite .
we conducted a spectroscopic study of the soft diffuse x - ray emission in m17 using the xis onboard suzaku .
high signal - to - noise ratio spectra of the diffuse emission were obtained in a spatially - resolved manner .
half a dozen of emission lines were resolved clearly for the first time , which include k@xmath0 lines from highly ionized o , ne , and mg and l lines from fe .
these lines are clear evidence for the thermal origin of the diffuse emission .
based on the chandra data of an overlapping field and on the previous works in the literature , we confirmed that the background - subtracted diffuse spectrum by xis is barely contaminated by unresolved point sources , grxe , cxb , and lhb emission .
we showed that the diffuse spectra are explained by a single temperature plasma model of @xmath45 0.25 kev , @xmath64 3.5@xmath910@xmath24 erg s@xmath6 ( 0.52.0 kev ) , and @xmath464.5@xmath910@xmath20 @xmath21 .
the temperature and the chemical composition of the diffuse plasma is spatially uniform , indicating that the plasma is at a thermal equilibrium .
the apparent difference in the surface brightness is probably due to the difference in the line - of - sight depth or in the filling factors , and not in the plasma density .
the abundance is obtained individually for the detected elements , which are consistent with 0.10.3 solar values .
the enhancement of ne against other metals by a factor of @xmath12 is seen in the diffuse plasma , which may be explained by a upward revision of the solar ne abundance .
the lack of spatial variations in the chemical composition comprises a sharp contrast to the diffuse emission in the carina nebula , where different abundance patterns across a similar spatial scale suggest the supernova origin for the emission .
together with the o to fe ratio consistent with the solar value , this gives evidence against the interpretation that a supernova is the cause of the diffuse emission observed in m17 .
the authors thank takashi hosokawa for useful discussion . y.h . and
acknowledge financial support from the japan society for the promotion of science .
the work is supported by the grants - in - aid for the 21st century center of excellence program `` center for diversity and universality in physics '' and for the program number 18204015 from the ministry of education , culture , sports , science and technology of japan .
the research made use of data obtained from the data archive and transmission system ( darts ) , provided by the plain center , isas / jaxa .
anders , e. , & grevesse , n. 1989 , geochim .
acta , 53 , 197 audard , m. , gdel , m. , sres , a. , raassen , a. j. j. , & mewe , r. 2003 , , 398 , 1137 brinkman , a. c. , et al .
2001 , , 365 , l324 brogan , c. l. , & troland , t. h. 2001 , , 560 , 821 broos , p. s. , feigelson , e. d. , townsley , l. k. , getman , k. v. , wang , j. , garmire , g. p. , jiang , z. , & tsuboi , y. 2007 , , 169 , 353 cappellari , m. , & copin , y. 2003 , , 342 , 345 chrysostomou , a. , brand , p. w. j. l. , burton , m. g. , & moorhouse , a. 1992 , , 256 , 528 chu , y .- h . ,
chang , h .- w .
, su , y .-
, & mac low , m .-
m . 1995 , , 450 , 157 diehl , s. , & statler , t. s. 2006 , , 368 , 497 drake , j. j. , & testa , p. 2005
, , 436 , 525 dunne , b. c. , chu , y .- h . , chen , c .- h .
r. , lowry , j. d. , townsley , l. , gruendl , r. a. , guerrero , m. a. , & rosado , m. 2003 , , 590 , 306 ebisawa , k. , et al .
2005 , , 635 , 214 felli , m. , massi , m. , & churchwell , e. 1984 , , 136 , 53 fujimoto , r. , et al . 2007 , , 59 , s133 garmire , g. p. , bautz , m. w. , ford , p. g. , nousek , j. a. , & ricker , g. r. , jr . 2003 , , 4851 , 28 getman , k. v. , et al .
2005 , , 160 , 319 hamaguchi , k. , et al . 2007 , , 59 , s151 hanson , m. m. , howarth , i. d. , & conti , p. s. 1997 , , 489 , 698 henning , t. , klein , r. , launhardt , r. , lemke , d. , & pfau , w. 1998 , , 332 , 1035 imanishi , k. , nakajima , h. , tsujimoto , m. , koyama , k. , & tsuboi , y. 2003 , , 55 , 653 ishisaki , y. , et al .
2007 , , 59 , s113 jiang , z. , et al .
2002 , , 577 , 245 kaneda , h. , makishima , k. , yamauchi , s. , koyama , k. , matsuzaki , k. , & yamasaki , n. y. 1997 , , 491 , 638 kastner , j. h. , huenemoerder , d. p. , schulz , n. s. , canizares , c. r. , & weintraub , d. a. 2002 , , 567 , 434 kokubun , m. , et al .
2007 , , 59 , s53 koyama , k. , et al .
2007 , , 59 , s23 lada , c. j. , depoy , d. l. , merrill , k. m. , & gatley , i. 1991 , , 374 , 533 leitherer , c. , robert , c. , & drissen , l. 1992 , , 401 , 596 liefke , c. , & schmitt , j. h. m. m. 2006 , , 458 , l1 maggio , a. , flaccomio , e. , favata , f. , micela , g. , sciortino , s. , feigelson , e. d. , & getman , k. v. 2007 , , 660 , 1462 mitsuda , k. , et al . 2007 , , 59 , s1 morrison , r. , & mccammon , d. 1983 , , 270 , 119 nielbock , m. , chini , r. , jtte , m. , & manthey , e. 2001 , , 377 , 273 nomoto , k. , hashimoto , m. , tsujimoto , t. , thielemann , f .- k . ,
kishimoto , n. , kubo , y. , & nakasato , n. 1997 , nuclear physics a , 616 , 79 ogura , k. , & ishida , k. 1976 , , 28 , 35 pfeffermann , e. , et al .
1987 , , 733 , 519 povich , m. s. , et al .
2007 , , 660 , 346 raymond , j. c. , & smith , b. w. 1977 , , 35 , 419 revnivtsev , m. , molkov , s. , & sazonov , s. 2006 , , 373 , l11 serlemitsos , p. j. , et al .
2007 , , 59 , s9 smith , r. k. , brickhouse , n. s. , liedahl , d. a. , & raymond , j. c. 2001 , , 556 , l91 snowden , s. l. , egger , r. , finkbeiner , d. p. , freyberg , m. j. , & plucinsky , p. p. 1998
, , 493 , 715 stelzer , b. , & schmitt , j. h. m. m. 2004 , , 418 , 687 sugizaki , m. , mitsuda , k. , kaneda , h. , matsuzaki , k. , yamauchi , s. , & koyama , k. 2001 , , 134 , 77 takahashi , t. , et al .
2007 , , 59 , s35 townsley , l. k. , feigelson , e. d. , montmerle , t. , broos , p. s. , chu , y .- h . , & garmire , g. p. 2003
, , 593 , 874 trmper , j. 1982 , advances in space research , 2 , 241 tsujimoto , t. , nomoto , k. , yoshii , y. , hashimoto , m. , yanagida , s. , & thielemann , f .- k .
1995 , , 277 , 945 tsujimoto , m. , hyodo , y. , & koyama , k. 2007 , , 59 , s229 tucker , w. h. , & gould , r. j. 1966 , , 144 , 244 weaver , r. , mccray , r. , castor , j. , shapiro , p. , & moore , r. 1977 , , 218 , 377 weisskopf , m. c. , brinkman , b. , canizares , c. , garmire , g. , murray , s. , & van speybroeck , l. p. 2002
, , 114 , 1 white , n. e. , giommi , p. , & angelini , l. 1994 , , 6100 , 1 wilson , t. l. , fazio , g. g. , jaffe , d. , kleinmann , d. , wright , e. l. , & low , f. j. 1979 , , 76 , 86 | we present the results of a suzaku spectroscopic study of the soft extended x - ray emission in the h region m17 .
the spectrum of the extended emission was obtained with a high signal - to - noise ratio in a spatially - resolved manner using the x - ray imaging spectrometer ( xis ) .
we established that the contamination by unresolved point sources , the galactic ridge x - ray emission , the cosmic x - ray background , and the local hot bubble emission is negligible in the background - subtracted xis spectrum of the diffuse emission .
half a dozen of emission lines were resolved clearly for the first time , including k@xmath0 lines of highly ionized o , ne , and mg as well as l series complex of fe at 0.51.5 kev .
based on the diagnosis of these lines , we obtained the following results : ( 1 ) the extended emission is an optically - thin thermal plasma represented well by a single temperature of @[email protected] mk , ( 2 ) the abundances of elements with emission lines in the diffuse spectrum are 0.10.3 solar , while those of bright discrete sources are 0.31.5 solar , ( 3 ) the metal abundances relative to each other in the diffuse emission are consistent with solar except for a ne enhancement of a factor of @xmath12 , ( 4 ) both the plasma temperature and the chemical composition of the diffuse emission show no spatial variation across the studied spatial scale of @xmath15 pc . |
the main focus of this work is a stochastic growth model in @xmath1 dimensions , called the polynuclear growth ( png ) model .
it belongs to the kpz ( kardar - parisi - zhang @xcite ) universality class and it can be described as follows ( see figure [ figpng ] ) . at time @xmath2 , the surface is described by an integer - valued height function @xmath3 , @xmath4 .
it thus consists of up - steps ( @xmath5 ) and down - steps ( @xmath6 ) .
the dynamics has a deterministic and a stochastic part : + ( a ) up- ( down- ) steps move to the left ( right ) with unit speed and disappear upon colliding , + ( b ) pairs of up- and down- steps ( nucleations ) are randomly added on the surface with some given intensity .
+ the up- and down - steps of the nucleations then spread out with unit speed according to ( a ) .
the png model can be interpreted in several different ways , see @xcite for a review . on a macroscopic scale
the surface of the png model grows deterministically , i.e. , @xmath7 is a non - random function . however , on a mesoscopic scale fluctuations grow in time .
this is called roughening in statistical physics and extensive numerical studies have been made @xcite .
since the png model is in the kpz universality class , the fluctuation of the surface height is expected to live on a @xmath8 scale and non - trivial correlations are to be seen on a @xmath9 scale .
therefore , to have an interesting large time limit , we have to rescale the surface height as @xmath10 one of the most natural initial conditions for png is the flat initial condition , i.e. , @xmath11 for all @xmath12 .
we consider nucleations occurring with translation - invariant intensity .
we refer to the png model with such initial condition as _
flat png_. in this case , by mapping the flat png to a point - to - line directed percolation model it was proven @xcite that the one - point distribution is , in the @xmath13 limit , the goe tracy - widom distribution @xmath14 , first discovered in random matrix theory @xcite .
however , no information on joint height distributions at more than one point has been previously known .
_ new results .
_ the main results of this paper are precisely the computation and asymptotic analysis of these joint distributions . in particular , we prove the convergence of the height rescaled as in ( [ eqfirst ] ) to the airy@xmath0 process in the @xmath13 limit ( see section [ subsectscalinglimit ] for a definition of the process ) .
the airy@xmath0 process has been discovered in the context of the asymmetric exclusion process @xcite .
our result , stated in theorem [ thmxitai ] , is obtained by first determining an expression for the joint distributions for finite time @xmath2 ( proposition [ proppng ] ) and then taking the appropriate scaling limit .
proposition [ proppng ] is actually just a particular case of theorem [ thmcpng ] , where we determine joint distributions along any space - like paths ( as in minkowski diagram ) , for which fixed time is a special case .
the scaling limit is analyzed at this level of generality , thus theorem [ thmxitai ] holds for any space - like paths .
in contrast to previous works on the subject , our approach does not rely on the so - called rsk correspondence ( rsk for robinson - schensted - knuth ) , which was successfully applied for corner growth models , but does not seem to be well suited for the flat growth . on the way of getting the results for the flat png , we consider a discrete time version of it , the gates - westcott dynamics @xcite .
this model is closely related to the totally asymmetric simple exclusion process ( tasep ) in discrete time with parallel update and alternating initial conditions .
the corresponding results for this model are theorem [ thmkalt ] for the joint distributions along space - like paths , and theorem [ thmxitoai ] for the convergence to the airy@xmath0 process in the scaling limit . for the tasep ,
the extreme situations of space - like paths are positions of different particles at a fixed time and positions of a fixed particle at different time moments ( tagged particle ) . the space - like extension for tasep
is based on the previous paper @xcite .
_ previous works on png .
_ another type of initial conditions for the png model has been analyzed before .
it is the corner growth geometry , where nucleations occur only inside the cone @xmath15 .
the limit shape @xmath16 is a semi - circle , and the model is called _ png droplet_. in this geometry , the limit process has been obtained in @xcite ; it is known as airy@xmath17 process ( previously called simply airy process ) .
the approach uses an extension to a multilayer model ( inherited from the rsk construction ) , see @xcite .
the multilayer method was also used in other related models @xcite . also , for the flat png it was used to connect the associated point process at a single position and the point process of goe eigenvalues @xcite .
results on the behavior for the png droplet along space - like paths can be found in @xcite . for a very brief description of the previously known results on tasep fluctuations
see the introductions of @xcite . in section [ sectmodel ]
, we introduce our models and state the results . in section [ sectjointdistr ] , we give an expression of the transition probability of the discrete tasep as a marginal of a determinantal signed point process . in section [ sectjointspacelike ] the fredholm determinant expression for the joint distributions is obtained .
the argument substantially relies on the algebraic techniques of @xcite . in section [ sectpotxitoai ]
, we consider the scaling limit of the parallel tasep . in section [ sectpotcpng ] ,
the continuous time png model is considered . in section [ sectpotxitai ]
, we consider the scaling limit for the continous png model .
a. borodin was partially supported by the nsf grants dms-0402047 and dms-0707163 .
ferrari is grateful to h. spohn for useful discussions .
the work of t. sasamoto is supported by the grant - in - aid for young scientists ( b ) , the ministry of education , culture , sports , science and technology , japan .
we start from the discrete time tasep with parallel update
. then we will make the connection with a discrete version of the png , from which the continuous time png is obtained .
we consider discrete time tasep with parallel update and alternating initial conditions , i.e. , particle @xmath18 has initial position @xmath19 , @xmath20 . at each time step , each particle hops to its right neighbor site with probability @xmath21 provided that the site is empty .
the particle positions at time @xmath2 is denoted by @xmath22 .
the dynamics of a particle depends only on particles on its right .
this fact allows us to determine the joint distributions of particle positions also for different times , but restricted to `` space - like paths '' . to define what we mean with `` space - like paths '' , we consider a sequence of couples @xmath23 , where @xmath24 is the number of the particle and @xmath25 is the time when this particle is observed . on such couples
we define a partial order @xmath26 , given by @xmath27 a space - like path is a sequence of ordered couples , namely , @xmath28 the reason of the name `` space - like '' will be clear in the large time limit , where everything becomes continuous .
then space - like is the same concept as in the minkowski diagram .
the border cases for space - like paths are fixed time ( @xmath29 ) and fixed particle number ( @xmath30 ) .
the next theorem concerns the joint distributions of particle positions .
[ thmkalt ] let particle with label @xmath18 start at @xmath19 , @xmath20 . consider a space - like path @xmath31 .
for any given @xmath32 , the joint distribution of the positions of the first @xmath32 points in @xmath31 is given by @xmath33 where @xmath34 . the kernel @xmath35 is given by @xmath36}\nonumber \\ & + & \widetilde{k}((n_1,t_1),x_1;(n_2,t_2),x_2),\end{aligned}\ ] ] where @xmath37 and @xmath38 where @xmath39 ( resp .
@xmath40 ) is any simple loop , anticlockwise oriented , with @xmath41 ( resp .
@xmath42 ) being the unique pole of the integrand inside the contour .
* remark . * in the limit @xmath43 under the time scaling by @xmath44 the discrete time tasep converges to the continuous time tasep , and theorem [ thmkalt ] turns into a special case of proposition 3.6 of @xcite , where a more general continuous time model called pushasep was considered .
starting from theorem [ thmkalt ] we can analyze large time limits .
the limit process is the so - called airy@xmath0 process introduced in @xcite , which we recall here .
[ defin ] define the extended kernel , @xmath45 the _
airy@xmath0 process _ ,
@xmath46 , is the process with @xmath32-point joint distributions at @xmath47 given by the fredholm determinant , @xmath48 where @xmath49 .
theorem [ thmkalt ] allows us to analyze joint distributions of particle positions for situations spanning between fixed time and fixed particle number ( the tagged particle problem ) .
one way to parametrize such situations is via a space - like path .
we thus consider an arbitrary smooth function @xmath50 satisfying @xmath51 see figure [ figspacelikepath ] .
the requirement @xmath52 reflects @xmath53 .
[ c]@xmath54 .
its slope is , in absolute value , at most @xmath55.,title="fig:",height=151 ] then , we choose couples of @xmath56 on @xmath57 , where @xmath58 is a large parameter .
the case of fixed time , say @xmath59 , is obtained by setting @xmath60 , while fixed particle number , say @xmath61 , by @xmath62 with some constant @xmath63 . from kpz scaling exponents @xcite
, we expect to see a nontrivial limit if we consider positions at distance of order @xmath64 .
thus , the focus on the region around @xmath65 is given by @xmath66 , i.e. , setting @xmath67 instead of @xmath68 and , by series expansions , we scale time and particle number as @xmath69 the kpz fluctuation exponent is @xmath70 , thus we expect to see fluctuations of particle positions on a scale of order @xmath71 .
therefore , we define the rescaled process @xmath72 by @xmath73 here the mean speed of particles , @xmath74 , is determined to be @xmath75 from the subsequent asymptotic analysis but can be known beforehand from the stationary measure for density @xmath76 @xcite .
this process has a limit as @xmath77 given in terms of the airy@xmath0 process .
[ thmxitoai ] let @xmath72 be the rescaled process as in ( [ xit ] ) .
then @xmath78 in the sense of finite dimensional distributions .
the vertical ( fluctuations ) and horizontal ( correlations ) scaling coefficients are given by @xmath79 * remark .
* a similar result for the pushasep with alternating initial condition has been proved in theorem 2.2 of @xcite . as mentioned in the introduction , the discrete tasep with parallel update is related to a surface growth model from which the polynuclear growth model in continuous time can be obtained as a limit .
let @xmath80 and @xmath81 denote the time and the one - dimensional space coordinate respectively , and let @xmath82 be the height of the surface at time @xmath2 and at position @xmath83 .
let us introduce a dynamics of @xmath82 as follows .
initially , at time @xmath84 , the surface is flat ; @xmath85 , for all @xmath86 .
right after each integer time @xmath87 , there could occur a nucleation with width 0 and height 1 with probability @xmath88 ( @xmath89 ) independently at each integer position @xmath83 such that @xmath90 is even .
each nucleation is regarded as consisting of an upstep and a downstep and each upstep ( resp .
downstep ) moves to the left ( resp .
right ) with unit speed .
this is a deterministic part of the evolution .
when an upstep and a downstep collide , they merge together .
see the solid line in figure [ figgw ] for an example until @xmath91 .
the dynamics of the growth model , if we focus only on half - odd times ( @xmath92 ) , is the same as one considered in @xcite , i.e. , a discretized version of the gates - westcott dynamics @xcite .
it is known that in an appropriate @xmath93 limit this growth model reduces to the standard continuous time png model @xcite . to see the connection to the discrete tasep ,
let us focus on integer times @xmath94 and positions ( @xmath95 ) from now on and represent the surface as consisting of elementary upward slopes @xmath96 and downward slopes @xmath97 as indicated by dashed lines in figure [ figgw ] . at @xmath84 , even ( resp .
odd ) @xmath83 s are taken to be the center of the upward ( resp .
downward ) slopes .
then the dynamics of the surface is described as follows : at each time step the surface grows upward by unit height deterministically and then each local maximum ( @xmath98 ) of slope turns into a local minimum ( @xmath99 ) independently with probability @xmath100 . if we interpret an upward ( resp .
a downward ) slope as a site occupied by a particle ( resp .
an empty site ) , this is equivalent to the discrete time tasep with parallel update under the alternating initial condition .
.,height=188 ] the relation between the surface height @xmath82 and the position of the tasep particle is given by @xmath101 and is understood as follows .
on the plot of the surface at some fixed time @xmath2 , draw also the initial surface at @xmath102 .
see figure [ figgwtasep ] for an expample .
then , from the correspondence between the growth model and the tasep , the surface at time @xmath2 can be regarded as the particle positions . in this
plot particles move along the down - right direction as indicated .
the left hand side of ( [ hx ] ) is equivalent to the condition that the tasep particle corresponding to @xmath103 has already reached @xmath83 . since the axis of the particle number @xmath104 is in the down - left direction , the value of @xmath104 corresponding to @xmath103 is @xmath105 .
this consideration results in the relation ( [ hx ] ) . from the relation ( [ hx ] ) the joint distributions of the height of the growth model is readily obtained through @xmath106 combined with theorem [ thmkalt ] .
when @xmath93 , the tasep particles move almost deterministically and the surface @xmath82 grows slowly , when a particle decides not to jump ( with probability @xmath88 ) .
the continuous time png model is obtained by taking @xmath93 while setting space and time units to @xmath107 ( the @xmath108 is chosen to have nucleations with intensity @xmath108 like in @xcite ) .
denote by @xmath109 and @xmath110 the position and time variables in the continuous time png model .
the png height function @xmath111 is then obtained by the limit @xmath112 here the minus sign on the right hand side is put for a convenience .
the results below do not depend on this sign because of the symmetry of the model in consideration . the joint distribution of the surface height at time @xmath110 is given as follows .
[ proppng ] consider @xmath32 space positions @xmath113 .
then , the joint distribution at time @xmath110 of the heights @xmath114 , @xmath115 , is given by @xmath116 where the kernel is given by @xmath117 where @xmath118 and @xmath119 are the modified bessel functions and the bessel functions , see e.g. @xcite .
the last indicator function is obvious if one thinks about the png model .
in fact , the height at position @xmath109 at time @xmath110 depends on events lying in the backward light cone of @xmath120 on @xmath121 .
thus , when @xmath122 , the backwards light cones of @xmath123 and @xmath124 do not intersect in @xmath121 , which implies that the two height functions are independent .
the fredholm determinant then splits into blocks .
the result of proposition [ proppng ] is actually a specialization of a more general situation which follows from the tasep correspondence . in the tasep ,
the space - like paths @xmath50 we had for particle numbers and times become the paths @xmath125 the condition @xmath126 implies that @xmath127 $ ] , i.e. , these are space - like paths as in special relativity oriented into the past . by the symmetry of the problem ,
one can consider also space - like paths locally oriented into the future , just looking at the process in the other direction .
denote by @xmath128 such a path on @xmath129 , i.e. , @xmath130 , then @xmath68 and @xmath54 are given by the relations @xmath131 and the joint distributions of the surface height along the path @xmath128 are expressed as in theorem [ thmcpng ] .
[ thmcpng ] let us denote by @xmath132 .
then , the joint distributions of @xmath133 , @xmath115 , is given by @xmath134 where the kernel is given by @xmath135 where @xmath118 and @xmath119 are the modified bessel functions and the bessel functions .
the condition @xmath136 means that @xmath137 or @xmath138 . in the first term
, for @xmath139 , the condition @xmath140 is satisfied for @xmath132 .
also , notice that when @xmath141 , the first term of the kernel goes to @xmath142 .
the last result of this paper is the large time behavior of the flat png .
the large parameter denoted by @xmath58 is proportional to time @xmath110 . using the function @xmath128 , we consider @xmath143 , see figure [ figpngpath ] .
[ l]@xmath110 [ l]@xmath144 for continuous time png .
@xmath58 is proportional to the png time @xmath110.,title="fig:",height=132 ] since the system is translation invariant , we focus around the origin , i.e. , we look at the png height at @xmath145 the surface height grows with the speed equal to @xmath108 .
thus , for large time @xmath110 , the macroscopic height will be close to @xmath146 .
fluctuations live on a @xmath71 scale .
consequently , we define the rescaled height process @xmath147 by @xmath148 the large @xmath58 ( thus large time too ) behavior of @xmath147 is given in terms of the airy@xmath0 process as stated below .
[ thmxitai ] let @xmath147 be the rescaled process as in ( [ xpt ] ) .
then , in the limit of large @xmath58 , we have @xmath149 in the sense of finite dimensional distributions .
the scaling coefficients @xmath150 and @xmath151 are given by @xmath152 for @xmath153 , i.e. , fixed time , this was conjectured to hold in @xcite .
let @xmath154 denote the transition probability of the parallel tasep with @xmath155 particles starting at @xmath84 at positions @xmath156 .
this is the probability that the @xmath155 particles starting from positions @xmath157 at @xmath84 are at positions @xmath158 at @xmath2 .
consider a determinantal signed point process on the set by setting the measure @xmath159 where @xmath160 the function @xmath161 defined by @xmath162 and where we used the convention , @xmath163 .
the following proposition states that the one time transition probability of the tasep is given as a marginal of the signed measure ( [ wn ] ) . [ gp ]
let us set @xmath164 .
then @xmath165 where summation is over the variables in the set , @xmath166 varying over @xmath167 .
note that @xmath168 is actually symmetric with respect to permutations of variables with same upper index , so the ordering in ( [ eq3.5 ] ) is used for singling out the minimal @xmath169 .
* remark . *
similar representations for the transition probability of continuous time tasep , discrete time tasep with sequential update and pushasep have been obtained in @xcite . in the different parts of the proof of proposition [ gp ]
, we will use several properties of the function @xmath170 , which are listed below .
[ propf ] @xmath171 here `` @xmath172 '' represents the convolution : @xmath173 .
lemma [ propf ] these are proven by using the definition ( [ eq2.3 ] ) and ( [ f ] ) .
the first step in the proof of proposition [ gp ] is the following lemma .
[ det_f ] let us set @xmath174 and @xmath175 .
then , for any antisymmetric function @xmath176 , @xmath177 where @xmath178 and @xmath179 lemma [ det_f ] from the antisymmetry of @xmath180 and of the determinant , ( [ df_gdf ] ) is equivalent to @xmath181 since a basis of the antisymmetric functions is made of the antisymmetric delta functions and the relation to prove is linear in @xmath180 , it is enough to consider @xmath182 for fixed @xmath183 . here @xmath184 is the group of all permutations of @xmath185 . for
this special choice of @xmath180 , the left hand side of ( [ df_gdf2 ] ) is @xmath186 times the single determinant , @xmath187 we have the following three cases .
+ ( a ) @xmath188 : the second row gives @xmath189 which is proportional to the first row .
therefore in this case the lhs is zero .
\(b ) @xmath190 : the second row is @xmath191 . subtracting @xmath192 times the second row from the first row one obtains @xmath193 \(c ) @xmath194 : the first column is @xmath195 .
thus the determinant is @xmath196 .
the result in each case agrees with @xmath186 times the rhs of ( [ df_gdf2 ] ) and hence the lemma is proved .
let @xmath197 denote the number of @xmath198 s s.t .
@xmath199 . using the above lemma with @xmath200 in which case
@xmath201 reduces to @xmath202 , we have the following result .
[ gtilde ] with @xmath164 , one has @xmath203_{1\leq i , j\leq n}\nonumber \\ & = : & \widetilde{g}(x_1,\ldots , x_n;t).\end{aligned}\ ] ] lemma [ gtilde ] for simplicity , we denote @xmath204 for @xmath205 . from the definitions ( [ wn ] ) , the lhs of ( [ gdet ] ) writes @xmath206 applying lemma [ det_f ] with @xmath207 , @xmath208 , @xmath209 , @xmath210 , @xmath211 and @xmath212 we obtain @xmath213 heine s identity , @xmath214_{1\leq i , j\leq n},\ ] ] allows us to rewrite the last summation in ( [ phif1 ] ) as @xmath215 we repeat the procedure up to a total of @xmath216 times in column @xmath198 and we get @xmath217_{1\leq i , j\leq n}. \label{phif2}\ ] ] the proof of the lemma is finished using ( [ fi ] ) , ( [ pf3 ] ) and @xmath218 .
proposition [ gp ] we need to prove @xmath219 this statement was also proved in @xcite by the bethe ansatz techniques .
our proof is by induction in @xmath2 .
we start by showing that the initial conditions agree , i.e. , @xmath220 , that is , @xmath221_{1\leq i , j\leq n } = \prod_{n=1}^n \delta_{x_n , y_n}. \label{ic}\ ] ] we first show that lhs of ( [ ic ] ) is zero if @xmath222 . if @xmath223 , since @xmath224 , one has @xmath225 .
then , from ( [ pf4 ] ) we have @xmath226 , i.e. , the first column of lhs of ( [ ic ] ) is zero .
similarly , if @xmath227 , since @xmath228 , one has @xmath229 .
then , from ( [ pf5 ] ) we have @xmath230 , i.e. , the first row of lhs of ( [ ic ] ) is zero .
this agrees with rhs of ( [ ic ] ) also being zero if @xmath222 .
now let us assume @xmath231 .
there are two cases .
\(a ) @xmath232 . in this case , since @xmath233 , one has @xmath234
. then the first column of lhs of ( [ ic ] ) is @xmath235 and hence the determinant is equal to @xmath236_{2\leq i , j\leq n}$ ] .
\(b ) @xmath237 .
first let us see that lhs of ( [ ic ] ) is zero when @xmath238 .
we have @xmath239 . if @xmath240 , we have @xmath241 , for @xmath242 , and @xmath243 , for @xmath242 .
then the first and the second row of lhs of ( [ ic ] ) are both of the form , @xmath244 where @xmath172 represents an arbitrary number and hence the determinant is zero .
hence lhs of ( [ ic ] ) is zero if @xmath238 . on the other hand ,
when @xmath245 , the upper - left @xmath246 submatrix of the determinant is @xmath247 whose determinant is @xmath248 .
repeating the same procedure , at each step one has either case ( a ) or ( b ) .
the final result is that @xmath249 , for @xmath250 , otherwise the determinant in lhs of ( [ ic ] ) is zero .
moreover , when @xmath249 , @xmath250 , denote by @xmath251 the values of @xmath198 such that @xmath252 . then lhs of ( [ ic ] ) is equal to @xmath253 with @xmath254_{1\leq i , j\leq n}\ ] ] finally using ( [ pf6 ] ) , ( [ pf7 ] ) , we obtain an explicit form of the matrix . to compute its determinant
it is enough to develop along the first row .
the determinant of the @xmath255 minor is @xmath256 , while the one of the @xmath257 minor is @xmath258 because the minor is the same as the @xmath255 minor except the first column is multiplied by @xmath259 .
all the other minors have determinant zero , because the first two column are linearly dependent .
thus , @xmath260 , and since @xmath261 , it follows that @xmath262 this ends the part of the proof concerning initial conditions .
next we prove that ( [ gg ] ) holds for @xmath263 if it does for @xmath2 . since this is true for @xmath84 , by induction it will be true for all @xmath264 .
@xmath265 satisfies the tasep dynamics , thus @xmath266_{1\leq i , j\leq n } .\nonumber\end{aligned}\ ] ] here @xmath267 and in the second equality we have used the assumption of the induction .
we rewrite @xmath268 using ( [ pf1 ] ) and ( [ pf2 ] ) as follows . for @xmath269 from @xmath55 to @xmath155 : + ( a ) if @xmath270 , then we use ( [ pf2 ] ) to the @xmath271th column .
then , the new term with the @xmath272 factor in front cancels out because it is proportional to its left column of the determinant .
+ ( b ) if @xmath273 , then we just use ( [ pf1 ] ) .
+ with these replacements we get @xmath274_{1\leq i , j\leq n},\nonumber\end{aligned}\ ] ] where @xmath275 [ c]@xmath276 [ c]@xmath277 [ c]@xmath278 [ c]@xmath279 [ c]@xmath2 [ c]@xmath263 [ b][b][0.8]@xmath88 [ b][b][0.8]@xmath55 [ b][b][0.8]@xmath280 [ c][][0.8]@xmath281 [ c][][0.8]@xmath282 ) .
the dots represents empty places , while a line leaving / arriving to a point is an occupied position . in ( a ) and
( b ) , on the left ( resp .
right ) we indicate the weights different from @xmath55 of lhs ( resp .
rhs ) of ( [ eq2.38 ] ) . in ( c ) and ( d ) the bottom and top lines of two blocks at distance @xmath108 at time @xmath263 are represented , for the cases corresponding to ( a ) and ( b ) for the top block.,title="fig:",width=529 ] comparing ( [ gw ] ) and ( [ gtw ] ) , it is enough to show @xmath283 this indeed holds and can be seen by checking case by case .
we illustrate it using figure [ figic ] .
first consider a block of particles , say @xmath32 of them at time @xmath263 .
there are two possibility of reaching this situations in one time step , as indicated in figure [ figic ] ( a ) and ( b ) .
the products of all the weights on the right and on the left are the same , i.e. , ( [ eq2.38 ] ) holds for a single block of particles . if two blocks of particles at time @xmath263 are at distance at least @xmath108 , they are independent during one time step .
we just have to check that ( [ eq2.38 ] ) holds for two blocks at distance @xmath108 at time @xmath263 .
case ( a ) is illustrated in ( c ) and the weights are unchanged for both blocks . case ( b ) is illustrated in ( d ) .
this time , the @xmath88 on the top line of the second block becomes a @xmath55 , but this is compensated by an extra factor @xmath88 on the left .
[ thmjointcorr ] let us consider particles starting from @xmath284 and denote @xmath285 the position of @xmath198th particle at time @xmath2 .
consider a sequences of particles and times which are space - like , i.e. , a sequence of @xmath32 such couples @xmath286 .
the joint distribution of their positions @xmath287 is given by @xmath288 where @xmath289 . here
@xmath290 is the extended kernel with entries @xmath291 where @xmath292},\nonumber\end{aligned}\ ] ] and the functions @xmath293 are given by @xmath294 the functions @xmath295 are determined by the two conditions @xmath296 and @xmath297 the paths @xmath298 are any simple loops anticlockwise oriented including @xmath41 , @xmath42 and no other poles .
theorem [ thmjointcorr ] this is the analogue of proposition 3.1 of @xcite .
the first step to obtain this proposition , was lemma 4.4 of @xcite . in the simple case where in the model studied in
@xcite all particles have the same jump rates , i.e. , @xmath299 for all @xmath18 , then lemma 4.4 of @xcite and our proposition [ gp ] have exactly the same structure . for the comparison of the two , we need just the following identifications . [
cols="^,^",options="header " , ] where @xmath300 from ( [ pf3 ] ) it then follows that @xmath301 satisfies @xmath302 the identity ( [ eq3.6 ] ) corresponds to lemma 4.3 of @xcite . to obtain proposition 3.1 of
@xcite one needs to prove lemma 4.5 of @xcite which needs as ingredients only lemma 4.3 of @xcite .
these lemmas are the only ingredients used to obtain theorem 4.1 and then theorem 4.2 of @xcite .
this last theorem concerns some general determinantal measures ( a generalization of lemma 3.4 in @xcite ) , whose specialization to the pushasep model is proposition 3.1 of @xcite .
all these steps go through unchanged in our case provided we make the above identifications . to obtain the precise expressions of theorem [ thmjointcorr ]
, we use an integral representation for @xmath303 , namely , @xmath304 from this we get @xmath305 also , the @xmath306 from @xcite now becomes @xmath307 the only step we have still to prove is that the space where we need to do the orthogonalization is actually @xmath308 .
this follows from lemma [ lemmaspaceortho ] below .
[ lemmaspaceortho ] we have the identity @xmath309 for some constant @xmath310 .
lemma [ lemmaspaceortho ] the proof is made by induction .
assume that ( [ eq3.10 ] ) is true for some @xmath104 , which is the case for @xmath311 .
then we prove that ( [ eq3.10 ] ) holds for @xmath312 .
consider @xmath313 we will prove ( [ eq3.10 ] ) for @xmath314 instead of @xmath315 ; the statements are clearly equivalent by shifts @xmath316 . set @xmath317 then
, for @xmath318 , @xmath319 and @xmath320 is a polynomial of degree @xmath321 in @xmath83 ( by evaluating the residue at @xmath322 ) , while for @xmath323 , @xmath324 ( because the residue at @xmath325 gives zero ) .
therefore @xmath326 for the proof by induction , we apply heine identity , @xmath327 the computation of the convolution leads to @xmath328 .
theorem [ thmjointcorr ] holds for general fixed initial conditions .
we want to apply it to the alternating initial condition . for that we first have to do the orthogonalization with the result given in the next lemma .
[ psiphi ] for initial conditions @xmath329 , @xmath330 , we have @xmath331 and @xmath332 where , as before , @xmath21 .
in particular , @xmath333 .
lemma [ psiphi ] the formula for @xmath334 is just obtained by substituting the initial conditions into ( [ eq2.4 ] ) .
now we prove that the orthonormality relation ( [ ortho ] ) holds . for @xmath335 ,
the pole at @xmath336 in @xmath337 is not present , and for @xmath338 , @xmath339 because the residue at @xmath42 vanishes .
thus @xmath340 where we have the constraint on the integration paths @xmath341 . the last term ( the sum )
equals @xmath342 now the pole at @xmath343 has disappeared and instead of it there is a simple pole at @xmath344 .
thus , the integral over @xmath345 is just the residue at @xmath344 , leading to @xmath346 where we used the change of variable @xmath347 .
lemma [ psiphi ] together with theorem [ thmjointcorr ] leads to the kernel for the alternating initial condition .
[ knalt ] for @xmath348 , the kernel @xmath290 in theorem [ thmjointcorr ] is given by @xmath349 } \nonumber \\ & & + \widetilde k((n_1,t_1),x_1;(n_2,t_2),x_2)\end{aligned}\ ] ] where @xmath350 is given by ( [ phi ] ) and @xmath351 here @xmath39 ( resp @xmath40 ) is any simple loop , anticlockwise oriented , which includes the pole at @xmath352 ( resp .
@xmath343 ) , satisfying @xmath353 and no point of @xmath39 lies inside @xmath40 .
proposition [ knalt ] we substitute ( [ psif ] ) and ( [ phif ] ) in the kernel ( [ eqkernelfinal ] ) . since @xmath354 for @xmath355 , we can extend the sum over @xmath269 to @xmath325 .
we can take the sum inside the integrals if the integration paths satisfy .
then we compute the geometric series and obtain @xmath356 at this point both simple poles @xmath344 and @xmath357 are inside the integration path @xmath298 , but the integrand does not have any pole anymore at @xmath336 . thus we will drop the @xmath41 in @xmath298 . separating the contribution from the pole at @xmath344
we get @xmath358 moreover , we also have @xmath359 thus the last two terms of ( [ eq2.55 ] ) and ( [ eq2.56 ] ) cancel out , leading to ( [ eq2.53 ] ) . with proposition [ knalt ] we almost obtained theorem [ thmkalt ] .
what remains to do is to focus far enough into the negative axis , where the influence of the finiteness of the number of particles is not present anymore .
there the kernel is equal to the kernel for the initial conditions @xmath360 , @xmath20 .
theorem [ thmkalt ] the kernel for the flat case is obtained by considering the region satisfying @xmath361 where the effect of the boundary in the tasep is absent . here the pole at @xmath343 vanishes .
computing the residue at @xmath357 in proposition [ knalt ] gives the kernel ( [ kt ] ) up to a factor @xmath362 which we cancel by a conjugation of the kernel .
from theorem [ thmkalt ] we have that @xmath363 .
we have such a situation but with @xmath364 . with this change of variable , we get @xmath365 where @xmath366 ( here we did not write explicitly the @xmath367 entries ) . taking into account the scaling ( [ eqscaling1 ] ) , we thus have to analyze the rescaled kernel @xmath368 with @xmath369 , @xmath75 .
in particular , we have to prove that , for @xmath370 fixed , @xmath371 ( or a conjugate kernel of it ) converges to the kernel @xmath372 uniformly on bounded sets and have enough control ( bounds ) on the decay of @xmath373 in the variables @xmath374 such that also the fredholm determinant converges . in order to have a proper limit of the kernel as @xmath77 , we have to consider the conjugate kernel @xmath375 given by @xmath376 the new kernel does not change the determinantal measure , being just a conjugation of the old one .
so , in the following we will determine the limit of @xmath377 as @xmath77 .
[ propunifcvg ] for @xmath370 fixed , according to ( [ eqscaling1 ] ) , set @xmath378 , \\
n_i & = n(u_i),\quad t_i = t(u_i).\end{aligned}\ ] ] then , for any fixed @xmath379 , we have @xmath380 uniformly for @xmath381 ^ 2 $ ] , with the kernel @xmath382 given by ( [ eqkernelexpanded ] ) and the constants @xmath383 and @xmath384 given by ( [ eqkappa0 ] ) .
proposition [ propunifcvg ] first we consider the first term in ( [ k ] ) .
we thus consider ( [ phi ] ) with the above replacements and conjugation .
this term has to be considered only for @xmath385 .
the change of variable @xmath386 leads then to @xmath387 with @xmath388 and @xmath389 the function @xmath390 has a critical point at @xmath391 .
the series expansions around @xmath391 are @xmath392 where @xmath393.\ ] ] to prove convergence of ( [ eq4.6 ] ) we have to show that the contribution coming around the critical point dominates in the @xmath77 limit .
we do it by finding a steep descent path , we say that @xmath128 is a steep descent path if ( 1 ) @xmath394 is maximal at some @xmath395 : @xmath396 for @xmath397 and ( 2 ) @xmath394 is monotone along @xmath398 except , if @xmath128 is closed , at a single point where @xmath399 is minimal .
] for @xmath390 passing by @xmath391 .
consider the path @xmath400 .
then , on @xmath39 , @xmath401 , @xmath402 and @xmath403 thus @xmath39 is a steep descent path for @xmath390 .
now we set @xmath404 .
then , the real part of @xmath405 is maximal at @xmath391 and strictly less then @xmath406 for all other points on @xmath39 .
therefore , we can restrict the integration from @xmath39 to @xmath407 .
for @xmath408 small , the error made is just of order @xmath409 with @xmath410 ( @xmath411 as @xmath412 ) . in the integral over @xmath413
we can use ( [ eq4.8 ] ) to get @xmath414 we use @xmath415 to control the difference between ( [ eq4.10 ] ) and the same expression without the error terms . by taking @xmath408 small enough and the change of variable @xmath416
, we obtain that this difference is just of order @xmath417 , uniformly for @xmath374 in a bounded set . at this point
we remain with ( [ eq4.10 ] ) without the error terms .
we extend the integration path to @xmath418 and this , as above , gives an error of order @xmath409 .
thus we have @xmath419 therefore , uniformly for @xmath374 in bounded sets , @xmath420 with @xmath421 .
now we have to consider the second term in ( [ k ] ) .
notice that this time the restriction @xmath385 does not apply .
set @xmath422 .
then @xmath423 with @xmath424 , \notag \\
f_1(z ) & = ( \pi'(\theta)+1)\big [ u_1 ( ( 1-\sqrt{q})\ln(-z)+\sqrt{q}\ln((1+(1-q)z)/q ) ) \notag\\ & \hspace{5em } -u_2((1-\sqrt{q})\ln(1+z)-\ln(1+(1-q)z))\big ] \notag\\ & + ( 1-\pi'(\theta))(u_1-u_2)\ln\left(\frac{(1+z)(-z)}{1+(1-q)z}\right ) , \end{aligned}\ ] ] @xmath425 \notag\\ & +
\xi_1\ln(-qz/(1+(1-q)z))-\xi_2\ln(1+z ) , \notag \\
f_3(z ) & = -\ln(-z(1+(1-q)z)).\end{aligned}\ ] ] the function @xmath426 has a double critical point at @xmath391 and the series expansions around @xmath391 of the @xmath427 s are given by @xmath428 with @xmath429 given in ( [ eqkappa ] ) and @xmath430 the leading contribution in the @xmath77 limit will come from the region around the double critical point . the first step is to choose for @xmath431 a steep descent path for @xmath426 .
first we consider @xmath432 , @xmath433 $ ] .
the only part in @xmath434 which is not constant along @xmath431 is the term @xmath435 multiplied by @xmath436 , @xmath437 .
simple computations lead then to @xmath438 this expression is strictly less than zero along @xmath431 except at @xmath439 , provided that the last term is strictly positive for @xmath440 .
this is easy to check because the last term reaches his minimum at @xmath441 . solving a second degree equations
, we get that on @xmath442 it is strictly positive and at @xmath443 is zero .
thus , the path @xmath431 is steep descent for @xmath426 .
but close to the critical point , the steepest descent path leaves with an angle @xmath444 .
therefore , consider for a moment @xmath445 . by symmetry
we can restrict the next computations to @xmath318 .
we have to see that @xmath446 is maximum at @xmath447 and decreasing for @xmath448 .
we have @xmath449 the term in the second line is always positive for all @xmath318 . to see this , remark that it is a polynomial of third degree which goes to @xmath325 as @xmath450 and at @xmath447 is already positive and has positive slope .
therefore one just computes its stationary points and , if reals , takes the right - most one .
there , the term under consideration turns out to be positive , which concludes the argument .
consequently , @xmath451 is also a steep descent path .
we choose a steep descent path @xmath39 as follows .
we follow @xmath451 starting from the critical point until we intersect it with @xmath431 , and then we follow @xmath431 . since @xmath39 is steep descent for @xmath426 , we can integrate only on @xmath452 . the error made by this cut
is just of order @xmath453 for some @xmath454 ( with @xmath455 as @xmath456 ) . around the critical point
we use the series expansions ( [ eq4.15 ] ) .
thus we have @xmath457 we want to cancel the error terms .
the difference between ( [ eq4.20 ] ) and the same expression without the error terms is bounded using @xmath458 , applied to @xmath459 .
then , this error term becomes @xmath460 for some @xmath461 depending on @xmath408 .
as @xmath456 , the @xmath462 .
thus , for @xmath408 small enough , we have @xmath463 . by the change of variable @xmath416 we obtain that ( [ eq4.22 ] )
is just of order @xmath417 .
thus we have @xmath464 the extension of the path @xmath413 to a path going from @xmath465 to @xmath466 accounts into an error @xmath467 only .
we do the change of variable @xmath468 and we define @xmath469 then , @xmath470 where @xmath471 is any path going from @xmath465 to @xmath466 .
the proof ends by using the airy function representation ( [ airy ] ) .
[ propbounddiffusion ] let @xmath24 , @xmath25 , and @xmath472 be defined as in proposition [ propunifcvg ] .
then , for @xmath473 fixed and for any @xmath474 , the bound @xmath475 holds for @xmath58 large enough and @xmath476 independent of @xmath58 .
proposition [ propbounddiffusion ] we start with ( [ eq4.6 ] ) .
the difference now is that the contribution coming from large @xmath477 can be of the same order as the one from @xmath405 .
we consider as path @xmath478 .
the difference is that now we choose @xmath479 as follows .
for an @xmath480 with @xmath481 and @xmath482 , @xmath483 by ( [ eq4.50 ] ) and ( [ eq4.51 ] ) , @xmath39 is a steep descent path for @xmath405 plus the term proportional to @xmath484 in @xmath485 .
so , integrating on @xmath486 instead of @xmath39 we do only an error of order @xmath409 times the value at @xmath487 , for some @xmath410 .
thus @xmath488 on @xmath413 , the @xmath489-dependent term in @xmath490 is equal to zero . with the same procedure as in proposition [ propunifcvg ]
one shows that the integral is bounded by a constant , uniformly in @xmath58 .
it remains to estimate the first factor in ( [ eq4.31 ] ) . with our choice ( [ eqrho ] ) , we need just series expansions of @xmath390 and @xmath491 around @xmath479 .
namely , by ( [ eq4.8 ] ) @xmath492 first consider the case @xmath493 .
we replace @xmath479 given in ( [ eqrho ] ) into ( [ eq4.32 ] ) and get that the sum of the two contributions in ( [ eq4.32 ] ) writes @xmath494 @xmath495 comes from @xmath496 , while the @xmath417 from @xmath497 .
then , by taking @xmath480 small enough and @xmath58 large enough , we get @xmath498 in the case , @xmath499 , we also replace the appropriate @xmath479 given in ( [ eqrho ] ) into ( [ eq4.32 ] ) .
we explicitly use the bound @xmath500 to bound @xmath501 .
then , we obtain the following bound for the sum of the two contributions in ( [ eq4.32 ] ) , @xmath502 by taking a fixed @xmath480 small enough and then @xmath58 large enough . finally , for @xmath503 , the same result holds in a similar way .
[ propboundmainterm ] let @xmath24 , @xmath25 , and @xmath472 be defined as in proposition [ propunifcvg ] .
let @xmath379 fixed .
then , for given @xmath370 and @xmath504 , the bound @xmath505 holds for @xmath58 large enough and @xmath476 independent of @xmath58 .
proposition [ propboundmainterm ] for @xmath506 $ ] it is the content of proposition [ propunifcvg ] . thus we consider @xmath507 ^ 2 $ ] .
define @xmath508 .
then we consider a slight modification of ( [ eq4.14 ] ) , namely @xmath509 with @xmath510 and @xmath511 as in ( [ eq4.16])-([eq4.16b ] ) , @xmath512 as @xmath513 in ( [ eq4.16b ] ) but with @xmath514 and @xmath515 replaced by @xmath516 , and finally @xmath517 is set to be equal to @xmath518 in ( [ eq4.16 ] ) plus the term @xmath519 we also chose @xmath520 . in the proof of proposition [ propunifcvg ] we already proved that @xmath39 is a steep descent path for @xmath426 for the values @xmath521 $ ] . also , since @xmath522
, @xmath523 is also decreasing while @xmath524 is increasing .
the precise choice of @xmath479 is @xmath525 for some small @xmath526 which can be chosen later .
let us define @xmath527 then , since @xmath39 is a steep descent path for @xmath528 , @xmath529 where @xmath530 , for a small @xmath531 .
the expansion around @xmath487 leads to @xmath532 with @xmath533 which is strictly positive for @xmath479 chosen as in ( [ eqrho2 ] ) .
also , @xmath534 for some bounded @xmath535 ( we do not write it down explicitly since the precise formula is not relevant ) .
therefore , the last term in ( [ eq4.42 ] ) is bounded by @xmath536 with @xmath537 . by choosing @xmath408 small enough and independent of @xmath58 , and then @xmath58 large enough , the error terms
can be replaced by @xmath76 , and the integral is then bounded by the one on @xmath538 .
thus @xmath539 in the worse case , when @xmath540 , which happens when @xmath541 , we have @xmath542 , which dominates @xmath535 for @xmath543 large enough .
therefore we have shown that @xmath544 it thus remains to find an bound on @xmath545 .
we have , by ( [ eq4.15 ] ) , @xmath546(1+{\mathcal{o}}(-\rho - z_c))}.\ ] ] in the case @xmath547 , we then obtain @xmath548 for @xmath549 , @xmath550 . finally , when @xmath551 , we have @xmath552 by first choosing @xmath526 small and then @xmath58 large enough . theorem [ thmxitoai ] the proof of theorem [ thmxitoai ] is the complete analogue of theorem 2.5 in @xcite .
the results in propositions 5.1,5.3,5.4 , and 5.5 in @xcite are replaced by the ones in proposition [ propunifcvg ] , [ propbounddiffusion ] , and [ propboundmainterm ] .
the strategy is to write the fredholm series of the expression for finite @xmath58 and , by using the bounds in propositions [ propbounddiffusion ] and [ propboundmainterm ] , see that it is bounded by a @xmath58-independent and integrable function .
once this is proven , one can exchange the sums / integrals and the @xmath77 limit by the theorem of dominated convergence .
for details , see theorem 2.5 in @xcite .
in this section we prove theorem [ thmcpng ] . by theorem [ thmkalt ] ,
the right hand side of ( [ hxn ] ) with @xmath553 can be written as fredholm determinant of the kernel @xmath554 with @xmath290 given in ( [ k ] ) . by the change of variable @xmath555 , one obtains the fredholm determinant of the kernel @xmath556 with this preparation , we now go to the proof of theorem [ thmcpng ] .
theorem [ thmcpng ] we have to analyze the kernel ( [ k ] ) with entries @xmath557 and take the limit @xmath93 with @xmath558 fixed .
the scaling of @xmath472 might look different from the one in ( [ eqheightrescaling ] ) but , as we can see below , ( [ eq5.10 ] ) with the last one replaced by @xmath559 gives the same limiting kernel .
as @xmath93 , the kernel does not have a well defined limit and , as usual , we first have to consider a conjugate kernel .
more precisely , we define @xmath560 what we have to prove is @xmath561 first we prove the pointwise convergence and then we obtain bounds allowing us to take the limit inside the fredholm determinant .
consider the term coming from ( [ phi ] ) . by the change of variable @xmath386
, we get @xmath562 and , by inserting ( [ eq5.10 ] ) , one obtains @xmath563 consider @xmath564 for some @xmath565 fixed .
then , we can fix the path @xmath39 independent of @xmath88 , and the @xmath93 limit is easily obtained .
it results in @xmath566 where we applied ( [ bes4 ] ) .
it is the turn of the term coming from ( [ kt ] ) .
we do the change of variable @xmath567 and then we insert ( [ eq5.10 ] ) .
the result is @xmath568 if @xmath569 , then for @xmath88 small enough , the result is identically equal to zero , because the pole at @xmath336 .
if @xmath570 , then the result is also zero , because the residues at all other poles , @xmath571 , @xmath572 , and @xmath325 vanishes . in the other case , when @xmath573 , the apparent pole at @xmath574 is actually not there .
so , we can choose a @xmath39 independent of @xmath564 for some @xmath565 .
then , we can simply take the limit @xmath93 of the integrand , which leads to @xmath575 where in the last step we made the change of variable @xmath576 and applied ( [ bes3 ] ) . to have convergence of the fredholm determinants we still need some bounds for large values of @xmath577 .
for @xmath88 small enough , say @xmath578 $ ] for some @xmath565 , we can set in ( [ eq5.11 ] ) @xmath579 in the case @xmath580 , and @xmath581 in the case @xmath582 .
then , we get the bound @xmath583 for some finite constant @xmath584 independent of @xmath88
. moreover , in ( [ eq5.12 ] ) we can choose @xmath585 , which leads to the bound @xmath586 with @xmath587 independent of @xmath564 .
these two bounds are enough to have convergence of the fredholm determinants .
the strategy is exactly the same as in the proof of theorem [ thmxitoai ] .
we analyze the kernel ( [ pngkernel ] ) with the scalings @xmath588 ( see ( [ eq1.24 ] ) and ( [ xpt ] ) ) .
the strategy of he proof is the same as that for theorem [ thmxitoai ] and hence we only give the main differences .
first we consider the first term in ( [ pngkernel ] ) . from ( [ pngdiff ] )
it is rewritten in the from ( [ eq4.6 ] ) with @xmath589 replaced by @xmath590 the critical point of @xmath405 is now @xmath591 .
the series expansions around @xmath391 are @xmath592 the steep descent path can be taken to be @xmath593 .
then the same arguments as in the proof of theorem [ thmxitoai ] give the first term in ( [ eqkernelexpanded ] ) .
next we consider the second term in ( [ pngkernel ] ) . from ( [ pngmain ] )
it is rewritten in the form ( [ eq4.14 ] ) with @xmath594 replaced by @xmath595 their series expansions around @xmath596 are @xmath597 the steepest descent path is taken to be @xmath598 with @xmath599 . using these one obtains the second term in ( [ eqkernelexpanded ] ) .
the bounds for the diffusion terms and the main term of the kernel are also proved in the same way as those of propositions [ propbounddiffusion ] and [ propboundmainterm ] .
in this appendix we list some integral representations of the bessel functions and the modified bessel functions ( we use the conventions of @xcite ) .
@xmath600 a. borodin and g. olshanski , _ stochastic dynamics related to plancherel measure _ , ams transl . : representation theory , dynamical systems , and asymptotic combinatorics ( v. kaimanovich and a. lodkin , eds . ) , 2006 , pp . 922 . | we consider the polynuclear growth ( png ) model in 1 + 1 dimension with flat initial condition and no extra constraints .
the joint distributions of surface height at finitely many points at a fixed time moment are given as marginals of a signed determinantal point process .
the long time scaling limit of the surface height is shown to coincide with the airy@xmath0 process .
this result holds more generally for the observation points located along any space - like path in the space - time plane .
we also obtain the corresponding results for the discrete time tasep ( totally asymmetric simple exclusion process ) with parallel update . |
the star formation rate ( sfr ) is the key measure that connects the structure formation in the universe and the actual observable light that is emitted by the galaxies .
the plot of sfr as a function of redshift , known as the ` madau plot ' , depicts the evolution of the formation rate of stars in galaxies . using a large - scale hydrodynamical simulation , it was shown by blanton et al .
( 1999 ) and nagamine et al .
( 1999 ) that the gas which falls into gravitational potential - wells gets shock - heated , and the higher overdensity regions become no longer preferred sites for galaxy formation as the temperature increases towards the present . here
we present the effects of this phenomena from different directions by analysing the sfr as a function of overdensity of luminosity .
details of the simulation can be found in the above two papers .
the sfr is a direct output of our hydrodynamical simulation .
figure 1 shows the sfr divided into the quartiles of the light - overdensity distribution in v - band at z=0 .
the v - band luminosity of each stellar particle in the simulation was obtained using the latest isochrone synthesis model of gissel99 ( bruzual & charlot 1999 ) with the metallicity of each particle taken into account .
it is clearly seen that the stellar particles in higher overdensity regions form earlier in time than those in lower overdensity regions .
this is due to the effect which we described in the introduction ; the gas in high overdensity regions is shock - heated as it falls into the potential - well , thus further star formation is prohibited by the high temperature of the gas .
hence the steeper turn - off of the sfr between z=1 and 0 . in lower overdensity regions , the gas is less heated than the higher overdensity regions and the peak of the sfr is at lower redshift .
observationally , this is well known as the ` morphology - density relation ' ( eg . , dressler 1980 ) . in clusters of galaxies ,
the old non - star - forming early - type galaxies dominate , whereas in the field , star forming late - type galaxies are seen more often .
also , the recent observations of galaxies in voids by grogin & geller ( 1999 , 2000 ) find that the galaxies in the lowest density environments show stronger star formation than those in higher density region , which is consistent with what we find in figure 1 .
\1 . blanton , m. et.al . , 1999 , submitted to _ apj _ , astro - ph/9903165 .
bruzual , g. a. and charlot , s. , 1999 , in preparation .
3 . dressler , a. , 1980 , _ apj _ , * 236 * , 351 .
grogin , n. a. and geller , m. j. , 1999 , _ aj _ in press , astro - ph/9910073 .
grogin , n. a. and geller , m. j. , 2000 , _ aj _ in press , astro - ph/9910096 .
nagamine , k. , cen , r. , j. p. ostriker , 1999 , submitted to _
apj _ , astro - ph/9902372 . | we use a large - scale @xmath0cdm hydrodynamical simulation to assess the dependence of the cosmic star formation rate ( sfr ) on overdensity of luminosity . |
a hypergraph @xmath1 is a set system @xmath20 defined on a finite universe @xmath21 ( called the vertex set ) .
the sets in the set system are called _
hyperedges_. we consider @xmath0-uniform hypergraphs , that is , hypergraphs whose hyperedges all have @xmath0 elements .
we say that a hypergraph is two - colorable , or has property b , if there is an assignment of colors @xmath22 to the vertex set @xmath21 such that every hyperedge @xmath23 has a vertex colored red and a vertex colored blue , that is , no hyperedge is _
monochromatic_. note that the hypergraph two - coloring problem can be viewed as a constraint satisfaction problem where the clauses are of a specific kind ( the ) . in the special case of graphs ( which we may view as two - uniform hypergraphs )
two - colorability is easy to characterize and establish : the graph is two - colorable if and only if it does not have an odd cycle ; one can find a two - coloring in linear time and @xmath24 space with random access to the input ; if the edges are streamed one at a time , a two - coloring can be constructed with @xmath25 space and @xmath26 bit operations . for the case of hypergraphs , the situation is not as simple . using the probabilistic method , erds @xcite showed that any @xmath0-uniform hypergraph with fewer than @xmath27 hyperedges is two - colorable : a random two - coloring is valid with positive probability ; furthermore , this randomized method can be derandomized using the method of conditional probabilities .
erds @xcite later showed that there are @xmath0-uniform hypergraphs with @xmath28 vertices and @xmath29 hyperedges that are _ not _ two - colorable .
both these bounds remained unchanged for some time , until it was first improved by beck @xcite , and further improved by radhakrishnan and srinivasan @xcite , who showed that any hypergraph with fewer than @xmath30 hyperedges is two - colorable .
they also provided a polynomial - time randomized algorithm ( and its derandomization ) for coloring such hypergraphs .
the algorithm has _
one - sided _ error , i.e. , with some small probability @xmath31 that can be made arbitrarily small , it declares failure , but otherwise it always outputs a valid two - coloring .
recently , cherkashin and kozik @xcite showed that a remarkably simple randomized algorithm achieves this bound .
erds and lovsz @xcite have conjectured that any hypergraph with fewer than @xmath32 hyperedges is two - colorable . in
related work , achlioptas et al .
@xcite studied when , in terms of the number of hyperedges , does a _ randomly _
chosen @xmath0-uniform hypergraph stop being two - colorable , and relate this to similar questions for random @xmath0-sat .
the above results are formally stated in terms of bounds on the number @xmath15 , the minimum number of hyperedges in a _ non - two - colorable _ hypergraph . in its original description
, the delayed recoloring algorithm of radhakrishnan and srinivasan @xcite assumes that the entire hypergraph is available in memory , its vertices and hyperedges and may be accessed quickly .
unfortunately , the number of hyperedges can be exponentially larger than the number of vertices .
for example , the number of hyperedges can be as high as @xmath33 even when there are just @xmath34 vertices , and it may be unrealistic to expect the entire hypergraph to be available in the main memory . in this paper , we ask if the performance of the delayed recoloring algorithm can be replicated when the work space of the algorithms is limited to ( say ) a polynomial in the size of vertex set .
to study this question , we consider the hypergraph two - coloring problem in the streaming model , where the data arrives as a stream or is stored in external memory , and an algorithm with limited local work space analyzes it by making a small number of sequentially passes over it .
the resources whose use one tries to minimize in this model are : the number of passes over the data , the amount of local memory required , and the maximum processing time for any data item . on receiving the complete data set , the algorithm must decide as fast as possible either to start another pass or stop with an output
. the algorithm could be either deterministic or randomized , and in the latter case , its output needs to be accurate with high probability over the algorithm s internal coin tosses .
a number of important algorithms have been developed in the streaming model of computation , e.g. , estimating frequency moments @xcite and heavy hitters @xcite of a data stream .
these algorithms find applications in the context of network routing , where large amount of data flows through a router but each of them has limited memory .
further motivation for our work comes from the _ semi - streaming model _ for graphs that has recently been considered widely in the literature .
proposed by muthukrishnan @xcite , this model looks at solving fundamental problems on graphs such as max - cut , @xmath35-connectivity , shortest paths , etc .
( see , for e.g. the works @xcite and references therein ) when the edges are streamed in one at a time .
the algorithm only has @xmath36 bits of workspace , and limited number of passes over the edge - stream , which means that it can not store the entire graph in memory .
our model for hypergraphs is an extension of this setting , when we have a stream of _ hyperedges _ coming in .
the parameter of interest to us is the _ uniformity _ @xmath0 of the hypergraph , and we assume the number of vertices is @xmath37 : this is the natural regime in which all the results known for two - coloring hypergraphs are stated .
the hyperedges are made available one at a time as a list of @xmath0 vertices each , where every vertex is represented using @xmath38 bits .
furthermore , we assume that @xmath37 bits ( equivalently , @xmath39 bits are available in the algorithm s work space .
notice that the number of hyperedges could now be _ exponential _ in @xmath0 ( or subexponential in @xmath40 ) , and hence we allow for a space that is @xmath39 , rather than close to linear in the number of vertices like in the semi - streaming model for graphs .
in the first part of the paper , we investigate the space requirements of _ deterministic _ streaming algorithms for the hypergraph two - coloring problem .
it seems reasonable to conjecture that any deterministic two - coloring one - pass algorithm must essentially store all the hyperedges before it can arrive at a valid coloring .
let @xmath3 be the minimum space used by a deterministic ( one - pass ) streaming algorithm that on receiving an @xmath41-uniform hypergraph @xmath4 on @xmath5 vertices and @xmath6 hyperedges produces a proper two - coloring of @xmath4 .
[ thm : result1 ] @xmath8 .
note that when @xmath42 , the graph is guaranteed to have a two - coloring .
however , the above theorem shows if the number of hyperedges is large , a two - coloring ( though it is guaranteed to exist ) can not be found efficiently by a deterministic streaming algoirithm .
lower bounds for space bounded computations often follow from lower bounds for associated communication complexity problems .
the above result is also obtained using this strategy .
however , the communication complexity problem turns out to be somewhat subtle ; in particular , we are not able to directly reduce it to a well - known problem and refer to an existing lower bound .
we conjecture that no deterministic algorithm can do substantially better if it is allowed only a constant number of passes over the input .
note , however , that proving this might be non - trivial , because the corresponding two - round deterministic communication problem has an efficient protocol ; see section [ sec : detub ] for details .
we show that a version of the delayed recoloring algorithm can be implemented in the streaming model , and provides essentially the same guarantees as the original algorithm .
let @xmath9 be the minimum space used by a randomized ( one - pass ) streaming algorithm that on receiving an @xmath0-uniform hypergraph @xmath4 on @xmath5 vertices and @xmath6 hyperdges with probability at least @xmath43 ( say ) produces a proper two - coloring of @xmath4 ( or declares failure ) .
we suppose that each vertex is represented using @xmath38 bits , and each edge is represented as an @xmath0-tuple of vertices . [
thm : randomizedonepass ] @xmath44 .
furthermore , the corresponding randomized algorithm maintains a coloring of the vertices encountered , and updates this coloring in time @xmath45 per hyperedge .
if the hypergraph has at most @xmath46 hyperedges , then with high probability the two - coloring is valid . if the two - coloring being maintained valid
, the algorithm declares failure ( the algorithm never outputs an invalid coloring ) .
the upper bound for property b , @xmath47 , was shown by erds by exhibiting a hypergraph with @xmath48 hyperedges that is not two - colorable ; this hypergraph needed @xmath19 vertices .
we ask if there are such non - two - colorable hypergraphs with @xmath49 vertices .
[ thm : verticesnontwocolorable ] 1 .
let @xmath50 be such that @xmath11 .
an @xmath0-uniform hypergraph with at most @xmath51 vertices and at most @xmath52 hyperedges is two - colorable .
the corresponding two - coloring can be found using an efficient randomized streaming algorithm : @xmath53 .
if @xmath54 , there is a graph with @xmath51 vertices and @xmath55 hyperedges that is not two - colorable .
setting @xmath56 in theorem [ thm : verticesnontwocolorable](a ) above implies that for any @xmath0-uniform hypergraph with @xmath57 vertices and @xmath58 hyperedges , there is a randomized one - pass streaming algorithm that outputs a two - coloring with high probability .
this algorithm requires @xmath59 space at any instant and @xmath34 processing time after reading each hyperedge .
a comparison of this to the result of erds @xcite , which gave a construction of a non - two - colorable hypergraph with the same @xmath60 hyperedges , but on @xmath28 vertices shows that the number of vertices being @xmath61 was crucial , and anything significantly smaller would not have worked .
furthermore , theorem [ thm : verticesnontwocolorable ] ( b ) recovers erds bound when @xmath62 , and generalizes it to make explicit the dependence of the erds upper bound on @xmath15 on the vertex - set size . [
[ organisation - of - the - rest - of - the - paper ] ] organisation of the rest of the paper : + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + in section [ sec : notation ] we introduce the notation . in section [ sec : deterministiclower ] , we establish theorem [ thm : result1 ] , showing the limitation of deterministic streaming algorithms . in section [ sec : streaming - alg ] , we prove theorem [ thm : randomizedonepass ] by showing how a version of the off - line delayed recoloring algorithm of @xcite ( see section [ sec : off - line - delayed - recoloring ] ) can , in fact , be implemented efficiently in the streaming model . in section [ boundeduniverse ] ,
we prove the results in theorem [ thm : verticesnontwocolorable ] regarding non - two - colorability .
we conclude with some remarks and open problems .
our @xmath0-uniform hypergraphs will be denoted by @xmath63 , where @xmath21 is the set of vertices , and @xmath64 is the set of hyperedges of @xmath4 . for a hypergraph @xmath1
, we use @xmath5 for @xmath40 , and @xmath6 for @xmath65 . in our setting
, @xmath5 will typically be a small polynomial in @xmath0 and @xmath6 will be exponential in @xmath0 . for any @xmath66
, we use the notation @xmath67
\coloneqq \{1 , \ldots , k\}$ ] .
a valid _ two - coloring _ of the hypergraph @xmath4 is an assignment @xmath68 that leaves no edge monochromatic , i.e. , @xmath69 , @xmath70 such that @xmath71 .
a hypergraph that admits a valid two - coloring is said to be _ two - colorable , _ or equivalently , to have _
property b_. we will be interested in space - efficient streaming algorithms for finding two - colorings of hypergraphs . consider a hypergraph with edge set @xmath72 .
the hyperedges are made available to the algorithm one at a time in some order @xmath73 . to keep the problem general
, we will not assume that the vertex set is fixed in advance .
the algorithm will become aware of the vertices as they arrive as part of the stream of hyperedges .
we will assume that each vertex is encoded using @xmath38 bits .
the goal is to design a space - efficient algorithm ( deterministic or randomized ) that can output a valid coloring for the entire hypergraph once all the hyperedges have _
passed_. we may allow multiple passes over the input data .
we call the algorithm a @xmath74-pass streaming algorithm , if it outputs a valid coloring after making @xmath74 passes over the input stream . by space
- efficient , we mean that the algorithm uses @xmath37 internal workspace .
we will assume that @xmath0 is large , say at least 100 .
denote by @xmath75 the minimum space used by a deterministic @xmath74-pass streaming algorithm that on receiving an @xmath0-uniform hypergraph @xmath4 on @xmath5 vertices and @xmath6 hyperedges produces a proper two - coloring of @xmath4 .
similarly , @xmath76 is the minimum space used by a randomized @xmath74-pass streaming algorithm that on receiving an @xmath0-uniform hypergraph @xmath4 on @xmath5 vertices and @xmath6 hyperedges with probability at least @xmath43(say ) produces a proper two - coloring of @xmath4 ( or declares failure ) .
when @xmath74 is omitted from the subscript , it is assumed @xmath77 .
we first show lower bounds in the deterministic setting .
we recall the routine translation of an efficient streaming algorithm to a communication complexity protocol @xcite , with a view to proving lower bounds .
there are two computationally unbounded players alice and bob , who both know of a relation @xmath78 .
alice receives an input @xmath79 and bob gets @xmath80 ; in the beginning , neither player is aware of the other s input .
their goal is to exchange bits according to a fixed _ protocol _ and find a @xmath81 so that @xmath82 .
the communication complexity of @xmath83 is the minimum number of bits that alice and bob exchange in any valid protocol for the worst case input pair @xmath84 .
several generalizations of this model can be defined with @xmath85 players , we will define our specific model below . for more details on communication complexity in general , please consult the book by kushilevitz and nisan @xcite . to show lower bounds for hypergraph two - coloring
, we define the class of communication problems @xmath86.-uniform , we will not explicitly state @xmath0 as a parameter ] [ defn : hyper - comm - game ] for @xmath87 , an instance @xmath88 has @xmath89 players @xmath90 .
each player @xmath91 has a subset @xmath92 of some hypergraph @xmath93 , with @xmath94 and @xmath95 .
the communication is done in sequential order : starting with @xmath96 , @xmath91 sends a message to @xmath97 for @xmath98 .
this sequence of communication constitutes a round . in a multiple - round protocol
, @xmath99 may start a new round by sending back a message to @xmath100 , who would continue communication in the above order . in a valid @xmath74-round protocol @xmath101 , some @xmath102
in the course of the @xmath74-th round will output a coloring @xmath103 that is valid for @xmath4 .
denote by @xmath104 the communication sent by @xmath102 in round @xmath105 on instance @xmath106 .
we define the @xmath74-round communication complexity of @xmath86 as follows : @xmath107 , l\in[r ] } } \quad \left\vert \pi(i , p_i , l ) \right\vert\ ] ] in this definition , we consider the maximum communication by any single player instead of total communication because this quantity is related more closely to the space requirement of streaming protcols ( see proposition [ prop : streamingtocommunication ] below ) . if the maximum communication by any player is @xmath108 bits , then the total communication is bounded above by @xmath109 bits for an @xmath74 round protocol . besides , for the values of @xmath89 and @xmath74 that we consider , this @xmath110 blowup is immaterial .
note that when @xmath111 and @xmath77 , we get the two - player one - round model ( where we traditionally call @xmath100 as alice and @xmath112 as bob ) : alice sends a message @xmath113 to bob depending on her input and bob outputs a coloring looking at @xmath113 and his input .
our lower bound for hypergraph coloring in the streaming model will rely on the following well - known connection between streaming and communication complexity introduced in @xcite .
[ prop : streamingtocommunication ] for any @xmath114 , we have : @xmath115 thus , to establish theorem [ thm : result1 ] , it is enough to show an appropriate lower bound on @xmath116 .
we start with the two - player case , which already introduces most of the ideas .
[ thm : oneroundlowerbound ] @xmath117 .
note that the above theorem gives a non - trivial lower bound only when @xmath118 .
ideally , we would expect a lower bound that is linear in @xmath6 , for all values of @xmath6 . consider a valid one - round protocol for the two - player problem where alice sends a message @xmath113 from a set @xmath119 of possible messages .
let the input to the protocol be @xmath120 , where @xmath121 and @xmath122 are hypergraphs , each with @xmath6 hyperedges on a common vertex set @xmath123 $ ] .
for every hypergraph @xmath124 that bob receives , he must output a coloring @xmath125 based on some deterministic function @xmath126 . for @xmath127 , define @xmath128 it is easy to see that bob may identify the message @xmath113 with the list @xmath129 ; on receiving @xmath113 , he must find a proper coloring for @xmath122 from @xmath129 .
thus , for every @xmath130 , we have the following .
completeness for bob : : : every possible input hypergraph to bob ( i.e. all hypergraphs on @xmath6 hyperedges ) should have a valid coloring in @xmath129 . soundness for alice
: : : let @xmath131 . then every @xmath132 should be valid for every hypergraph @xmath133 .
we will show that these two conditions imply the claimed lower bound on @xmath134 . given a coloring @xmath103 ,
we define its _ shadow _ @xmath135 to be the set of all possible hyperedges that are monochromatic under @xmath103 .
the shadow of a list @xmath136 of colorings is @xmath137 note that in the above definition , the shadow @xmath135 collects _ all possible _ monochromatic hyperedges under @xmath103 , so it depends only on the coloring @xmath103 , and not on any hypergraph .
similarly , @xmath138 is also a collection of hyperedges and does not depend on any hypergraph ; in particular , if a hypergraph @xmath121 is monochromatic under every coloring in @xmath136 , then none of @xmath121 s hyperedges can appear in @xmath138 . in the following ,
assume that @xmath0 is large .
for every coloring @xmath103 , we have @xmath139 : one of the two color classes @xmath103 has at least @xmath140 vertices .
it follows that @xmath141 we next observe that the _ completeness condition _ for bob imposes a lower bound on @xmath142 . for every @xmath127 , [ cl : shadowunion ] @xmath143
: suppose the claim does not hold , that is , @xmath144 choose a random hypergraph @xmath124 by choosing @xmath6 hyperedges randomly from @xmath142 .
we will say that the hypergraph _ hits _
@xmath145 , if at least one of its hyperedges falls in @xmath135 , otherwise we say it _ misses _ @xmath103 .
for each @xmath132 , we have , using the bounds ( [ eq : oneshadow ] ) and ( [ eq : shadowassumption ] ) , that @xmath146 & \leq \left(1-\frac{|\delta(\chi)|}{|\delta(l(m))|}\right)^{q}\\ & \leq \left(1-\frac{{n}^2}{{q}}\right)^{q}\\ & \leq \exp(-{n}^2)\end{aligned}\ ] ] since there are at most @xmath147 colorings , the union bound yields : @xmath148 \leq 2^{{n}^2 } \exp(-10 { n}^2 ) \ll 1\ ] ] thus there exists a hypergraph @xmath4 with @xmath6 hyperedges that hits every coloring @xmath132 , that is every coloring in @xmath129 is invalid for @xmath4 .
this , however , violates completeness for bob , proving the claim .
we can now complete the proof of the theorem .
consider a random hypergraph @xmath4 for alice , obtained by choosing each of its hyperedges uniformly at random from the set of all hyperedges .
since alice sends some @xmath130 for _ every _ hypergraph , the _ soundness condition _ for alice implies that @xmath149 \leq where we used claim [ cl : shadowunion ] to justify the first inequality .
taking logarithms on both sides yields the desired lower bound latexmath:[$\log as remarked earlier , the above communication lower bound implies that the space required by a deterministic streaming algorithm to find a valid coloring is exponential in @xmath0 even for hypergraphs that have very simple randomized coloring strategies ; it does not yield any such lower bound for hypergraphs that have fewer than @xmath151 hyperedges . in order to overcome this limitation ,
we generalize the analysis above to the @xmath89-player hypergraph coloring problem .
we will see later that for @xmath152 there do exist efficient two - player one - round protocols , so we could not have proved our lower bounds while restricting our attention to the two - player setting .
theorem [ thm : result1 ] will follow from the following theorem : let @xmath153 . @xmath154^\frac{1}{k}\right)\ ] ] recall that the protocol has @xmath155 players , @xmath156 .
player @xmath102 receives a hypergraph @xmath157 with @xmath6 hyperedges over the vertex set @xmath158 $ ] .
the communication starts with @xmath100 , who sends a message @xmath159 of length @xmath160 to player @xmath112 ; in the @xmath161-th step , @xmath102 sends a message of length @xmath162 to @xmath97 . in the end , @xmath163 produces a coloring for the hypergraph @xmath164
. it will be convenient to view this coloring as a message sent by player @xmath163 , and set @xmath165 ( the number of bits needed to describe a coloring ) . in a @xmath166-player protocol ,
after the messages sent by the first @xmath161 players have been fixed , we have a list of colorings that may still be output at the end ; we use the following notation to refer to this list ( recall that @xmath167 is a coloring ) : @xmath168 in particular , by considering the situation at the beginning of the protocol ( when no messages have yet been generated ) , we let @xmath169 .
let @xmath170 , and for @xmath171 , let @xmath172 and @xmath173 here the minimum is taken over all possible sequences of first @xmath161 messages that arise in the protocol .
in particular , @xmath174 corresponds to the union of shadows of all colorings ever output by the protocol , and @xmath175 corresponds to the shadow of the output corresponding to the transcript ( that is , the shadow of the last message , which is a coloring ) .
[ claim : generalshadow ] @xmath176 inequality ( [ cl : a ] ) is immediate from the definition of @xmath174 . for claim ( [ cl : b ] ) , we use ideas similar to those used in the proof of claim [ cl : shadowunion ] for two - player protocols . fix @xmath177 .
we will show that @xmath178 is at least the right hand side of ( [ cl : b ] ) .
pick a random hypergraph @xmath4 ( which we will consider as a possible input to @xmath97 ) as follows : choose @xmath6 hyperedges independently and uniformly from the set @xmath179 .
when @xmath4 is presented to @xmath97 , it must respond with a message @xmath180 .
none of the colorings that @xmath163 produces after that can include any edge of @xmath4 in its shadow , that is , @xmath181 .
for each valid choice @xmath113 for @xmath180 , we have @xmath182 and @xmath183 \leq \left(1-\frac{s_{i+1}(m_1,\ldots , m_i , m)}{s_{i}(m_1 , \ldots , m_i))}\right)^{{q } } \leq \left(1-\frac{s_{i+1}}{s_{i}(m_1 , \ldots , m_i))}\right)^{{q}}.\ ] ] thus , @xmath184 \leq 2^{\ell_{i+1}}\left(1-\frac{s_{i+1}}{s_{i}(m_1 , \ldots , m_i))}\right)^{{q}}\ ] ] this yields @xmath185 , giving @xmath186 . by minimizing over valid sequences @xmath187 , we justify our claim .
claim [ cl : c ] follows from the fact the shadow of every coloring has at least @xmath188 hyperedges . by combining parts ( [ cl : a ] ) and ( [ cl : b ] )
, we obtain @xmath189 the theorem follows from this by using ( [ cl : c ] ) , noting that @xmath190 , and for @xmath191 : @xmath192 } \ell_i$ ] . [ restatement of theorem [ thm : result1 ] ] every one - pass deterministic streaming algorithm to two - color an @xmath0-uniform hypergraph with at most @xmath6 hyperedges requires @xmath193 bits of space . setting @xmath194 in the previous result
immediately yields that for hypergraphs on @xmath195 vertices and at most @xmath196 hyperedges , the communication required is @xmath197 .
the lower bound for streaming algorithms then follows from proposition [ prop : streamingtocommunication ] . in the previous section
, we derived our lower bound for the deterministic streaming algorithms by invoking _
multi - player _ communication complexity , because the two - player lower bound did not give us a non - trivial lower bound for hypergraphs with fewer than @xmath151 hyperedges .
in this section , we first show an upper bound in the two - player setting , which shows that it was essential to consider the multi - player setting in order to get the stronger lower bound .
next , we consider two - round protocols , for they are related to two - pass streaming algorithms .
we show below , perhaps surprisingly , that the two - round two - player deterministic communication complexity for the problem is _
@xmath37_. however , we do not have a streaming algorithm with a matching performance .
@xmath198 alice and bob will base their protocol on a special collection of lists of colorings @xmath199 , with @xmath200 .
suppose alice s hypergraph is @xmath121 and bob s hypergraph is @xmath122 .
the protocol will have the following form .
alice : : alice sends an index @xmath161 of a list @xmath201 such that every coloring in @xmath202 is valid for @xmath121 .
bob : : bob outputs a coloring @xmath203 that is valid for @xmath122 .
we next identify some properties on the collection of lists that easily imply that the protocol above produces a valid two - coloring . 1 .
a collection of lists @xmath204 is good for bob , if for every hypergraph @xmath122 , in list @xmath136 in @xmath204 there is a valid coloring for @xmath122 in @xmath136 .
a collection of lists @xmath204 is good for alice , if for every hypergraph @xmath121 , there is list @xmath136 in @xmath204 , such that coloring in @xmath136 is valid for @xmath121 .
note that good collections are defined differently for alice and bob . with this definition ,
the existence of a collection of lists that is good for both alice and bob would furnish a one - way communication protocol with @xmath205 bits of communication , and establish the theorem .
such a _ good _ collection is shown to exist in lemma [ goodlistsexist ] below .
[ goodlistsexist ] suppose @xmath121 and @xmath122 are restricted to have at most @xmath151 hyperedges .
then , there exists a collection of @xmath206 of @xmath200 lists , each with @xmath207 colorings , which is good for alice and bob .
we pick @xmath204 by picking @xmath74 lists randomly : list @xmath202 will be of the form @xmath208 , where each coloring is chosen independently and randomly from the set of all colorings .
we will show that with positive probability @xmath204 is a good collection of lists .
we will separately bound the probability that @xmath204 fails to satisfy conditions ( a ) and ( b ) above .
we may restrict attention to hypergraphs that have exactly @xmath209 hyperedges . 1 .
we first bound the probability that @xmath204 is not good for bob .
consider one list @xmath201 .
@xmath210 & = & \pr[\exists h_b \text { such that none of the } k \text { colorings are valid for } h_b ] \\ &
\leq & { { { v}\choose { n } } \choose { q}}\left(\frac{{q}}{2^{{n}-1}}\right)^{k}.\end{aligned}\ ] ] since @xmath204 is a collection of @xmath74 lists , we have @xmath211 & = & \pr[\exists l_i \in \mathcal{l } \text { is not good for bob } ] \nonumber \\ & \leq & r { { { v}\choose { n } } \choose { q}}\left(\frac{{q}}{2^{{n}-1}}\right)^{k } \\
& \leq & 2^ { ( { n}{q}\log { v}+ \log r ) -k ( { n}- \log { q}-1 ) } .
\label{ineq : notgoodforbob}\end{aligned}\ ] ] to make the right hand side of ( [ ineq : notgoodforbob ] ) small , we will later choose @xmath89 such that @xmath212 2 .
we next bound the probability that @xmath204 is not good for alice .
fix a hypergraph @xmath121 for alice , and consider @xmath213 .
@xmath214 & \geq & \left(1-\frac{{q}}{2^{{n}-1}}\right)^{k}\\ & \geq & \exp\left(-\frac{{q}k}{2^{{n}-1}-{q}}\right ) . \quad \mbox{(using $ 1-x \geq \exp\left(-\frac{x}{1-x}\right)$)}\end{aligned}\ ] ] since @xmath204 is a collection of @xmath74 such lists chosen independently and there are @xmath215 choices for @xmath121 , we have @xmath216 & \leq & { { { v}\choose { n } } \choose { q } } \left(1- \exp\left(-\frac{{q}k}{2^{{n}-1}-{q}}\right)\right)^r \nonumber \\ & \leq & \exp\left ( { q}n \ln { v}- r \exp\left(-\frac{{q}k}{2^{{n}-1}- { q}}\right)\right ) .
\label{ineq : notgoodforalice}\end{aligned}\ ] ] to make the right hand side of ( [ ineq : notgoodforalice ] ) small , we will choose @xmath74 such that @xmath217 now , suppose @xmath218 .
then one can verify that if we set @xmath219 and @xmath200 , then ( [ eq : boundonk ] ) and ( [ eq : boundonr ] ) both hold for all large @xmath0 .
it follows that the required collection of lists exists .
next we present our efficient two - player , two - round communication protocol for hypergraph two - coloring .
[ thm : two - round - ub ] @xmath221 we will prove theorem [ thm : two - round - ub ] by exploiting the connection between circuit complexity and _ karchmer - wigderson games _ @xcite . given a monotone boolean function @xmath222 , the karchmer - wigderson communication game @xmath223 between two players alice and bob is the following : alice gets an input @xmath224 , and bob gets an input @xmath225 , and they both know @xmath126 .
the goal is to communicate and find an index @xmath226 $ ] such that @xmath227 and @xmath228 ( such an index exists since @xmath126 is monotone ) .
the communication complexity of this game is denoted by @xmath229 ( or specifically @xmath230 for an optimal @xmath74-round protocol ) .
karchmer and wigderson make the following connection : [ thm : kw - game ] if a monotone boolean function @xmath126 has a depth-@xmath231 , size @xmath108 circuit , then @xmath232 .
the proof of theorem [ thm : two - round - ub ] will be based on the circuit complexity of _ approximate majority functions_. define the partial function @xmath233 on a subset of @xmath234 as follows : @xmath235 when @xmath236 , and @xmath237 when @xmath238 .
a function @xmath222 is an approximate majority function if it satisfies : @xmath239 when @xmath240 , and @xmath241 when @xmath242 .
we need the following lemma .
[ lem : reduction - to - maj ] for any approximate majority function @xmath243 , we have : @xmath244 we will prove this by reducing @xmath245 to @xmath223 , for any approximate majority function @xmath126 . for convenience , let @xmath246 .
suppose alice receives @xmath247 , and bob receives @xmath248 .
let @xmath249 be an indicator vector for alice that marks which of the @xmath250 possible two - colorings of @xmath21 ( arranged in some canonical order ) are valid for @xmath121 , and let @xmath251 be the corresponding indicator vector for bob . since we know that both @xmath252 have @xmath253 hyperedges , a randomly chosen coloring will color either hypergraph with probability @xmath254 , which implies @xmath255 and @xmath256 . to find a coloring valid for both @xmath121 and @xmath122 , alice and bob
just have to find an index @xmath257 $ ] such that @xmath258 .
alice now complements her input to consider @xmath259 . since @xmath260 , @xmath261 , whereas @xmath262 . clearly , playing the karchmer - wigderson game for any approximate majority function @xmath126 on @xmath250 bits , @xmath263 would find a index @xmath161 where @xmath264 , and consequently yield a valid coloring for @xmath265 .
hence , a small depth-3 monotone circuit that computes @xmath266 on its domain would yield an efficient 2-round protocol for @xmath245 , using theorem [ thm : kw - game ] .
ajtai @xcite showed that such a circuit exists , and viola @xcite further showed uniform constructions of such circuits .
[ thm : appmaj - circuit]@xcite there exist monotone , uniform @xmath267-sized depth-3 circuits for @xmath233 on @xmath250 input bits .
the proof of theorem [ thm : two - round - ub ] is now immediate from theorems [ thm : kw - game ] , [ thm : appmaj - circuit ] and lemma [ lem : reduction - to - maj ] , since @xmath5 is polynomial in @xmath0 .
in this section , we present streaming algorithms that come close to the performance of the randomized off - line algorithm of radhakrishnan and srinivasan @xcite .
we first point out why this algorithm , as stated , can not be implemented with limited memory .
next , we show how an alternative version can be implemented using a small amount of memory .
this modified version , however , does return colorings that are not valid ( though it does this with small probability ) ; we show how by maintaining a small amount of additional information , we can derive an algorithm that returns and valid coloring with high probability or returns failure , but never returns an invalid coloring . in order to describe our algorithms , it will be convenient to use @xmath268 to denote vertices in the hypergraph . also , in this section , we will explicitly use @xmath40 to denote the size of the vertex set @xmath21 , and @xmath65 to denote the number of hyperedges in the hypergraph .
the vertex set of the hypergraph will be identified with @xmath269 $ ] , and @xmath38 will denote the number of bits required to represent a single vertex in @xmath21 .
radhakrishnan and srinivasan @xcite showed that by introducing delays in the recoloring step of an algorithm originally proposed by beck @xcite one can two - color hypergraphs with more hyperedges than it was possible before .
[ thm : radhakrishnan - srinivasan ] let @xmath1 be an @xmath0-uniform hypergraph with at most @xmath270 hyperedges .
then @xmath4 is two - colorable ; also a proper two - coloring can be found with high probability in time @xmath271 . input @xmath1 . for all @xmath272 ,
independently set @xmath273 to @xmath274 or @xmath275 with probability @xmath276 .
let @xmath277 be the set of hyperedges of @xmath4 , that are monochromatic under @xmath278
. for all @xmath272 , independently set @xmath279 to be @xmath280 with probability @xmath281 , and @xmath282 with probability @xmath283 .
let @xmath284 be one of the @xmath285 permutations of @xmath21 , chosen uniformly at random .
@xmath286 is obtained from @xmath287 by retaining the colors of all vertices except perhaps @xmath288 .
if @xmath289 , and some hyperedge containing @xmath288 was monochromatic under @xmath278 and all its vertices still have the same color in @xmath287 , then @xmath290 is flipped to obtain @xmath291 .
output @xmath292 .
the term _ off - line _ refers to the fact that we expect all hyperedges to remain accessible throughout the algorithm .
the term _ delayed recoloring _ refers to the fact that vertices are considered one after another , and are recolored only if the initially monochromatic edge they belonged to has not been set right by a vertex recolored earlier . for a coloring @xmath103 ,
let @xmath293 be the set of hyperedges that are colored entirely blue under @xmath103 , and let @xmath294 be the set of hyperedges colored entirely red under @xmath103 . in @xcite ,
the following claims are established for this algorithm under the assumption that @xmath295 . in the following ,
the parameter @xmath281 denotes the probability that the bit @xmath296 is set to @xmath280 in algorithm [ alg : off - line ] .
@xmath297 & \leq & \frac{2}{10 } \sqrt \frac{{n}}{\ln { n } } ( 1-p)^{n}\label{ineq : staysmono } \\ \pr[{\cal b}(\chi_f ) \setminus { \cal b}(\chi_0 ) \neq \emptyset ] & \leq & \frac{2{n}p}{100 \ln { n } } ; \label{ineq : becomesblue } \\ \pr[{\cal r}(\chi_f ) \setminus { \cal r}(\chi_0 ) \neq \emptyset ] & \leq & \frac{2{n}p}{100 \ln { n}}. \label{ineq : becomesred}\end{aligned}\ ] ] the first inequality ( [ ineq : staysmono ] ) helps bound the probability that an initially monochromatic hyperedge does not change our attempts at recoloring ; the second inequality ( [ ineq : becomesblue ] ) considers the event where an initially non - blue hyperedge become blue because of recoloring ; the third inequality ( [ ineq : becomesred ] ) similarly refers to the event where an initially non - red hyperedge becomes red because of recoloring .
these events together cover all situations where @xmath298 turns out to be invalid for @xmath4 .
now , set @xmath299 so that we have @xmath300 thus , the above algorithm produces a valid two - coloring with probability at least @xmath276 .
furthermore , since the entire graph is available in this off - line version , we can efficiently verify that the final coloring is valid . by repeating the algorithm @xmath301 times
, we can ensure that the algorithm produces a valid coloring with probability at least @xmath302 , and declares failure otherwise ( but never produces an invalid coloring ) .
[ [ implementation - in - the - streaming - model ] ] implementation in the streaming model : + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + now , suppose the hyperedges arrive one at a time , each hyperedge is a sequence of @xmath0 vertices , each represented using @xmath38 bits . in time @xmath303 per edge , we may extract all the vertices .
the corresponding colors and bits for recoloring can be generated in constant time per vertex .
generating the random permutation @xmath284 is a routine matter , we maintain a random permutation of the vertices received so far by inserting each new vertex at a uniformly chosen position in the current permutation ; using specialized data structure , the total time taken for generating @xmath284 is @xmath304 .
overall , the the algorithms can be implemented in time @xmath305 ( the @xmath306 hides some @xmath307 factors ) .
the space required is rather large because we explicitly store the hyperedges .
observe , however , that we need store only the hyperedges that are monochromatic after the initial random coloring .
thus , the algorithm can be implemented using @xmath308 bits of space on the average .
the space and time requirements may then be considered acceptable , but this implementation does not allow us to determine if the coloring produced at the end is valid ; moreover , it is not clear how we may reduce the error probability by repetition . in the next section
, we give an implementation that does not suffer from this deficiency . in this section ,
we modify the randomized streaming algorithm of the previous section so that it uses @xmath309 bits of space and takes time comparable to the off - line algorithm .
this version maintains a running list of vertices as suggested above , but it assigns them colors immediately , and with the arrival of each edge decides if some of its vertices must be recolored .
the algorithm is derived from the off - line algorithm .
it maintains enough information to ensure that the actions of this streaming algorithm can somehow be placed in one - to - one correspondence with the actions of the off - line algorithm .
input : the hypergraph @xmath1 as a sequence of hyperedges @xmath310 .
read @xmath311 .
set @xmath312 to @xmath274 or @xmath275 with probability @xmath276 . set @xmath279 to @xmath280 with probability @xmath281 and @xmath282 with probability @xmath283 .
insert @xmath313 at a random position in @xmath284 .
[ linegeneratechi ]
let @xmath313 be the first vertex ( according to the current @xmath284 ) such that @xmath314 .
flip @xmath315 .
output @xmath292 .
we wish to show that algorithm [ alg : streaming ] succeeds in constructing a valid coloring with probability at least @xmath276 . to justify this
, we compare the actions of this algorithm and the off - line delayed recoloring algorithm stated earlier , and observe that their outputs have the same distribution .
recall that @xmath278 was the initial coloring generated in the algorithm [ alg : off - line ] .
let @xmath316 be the corresponding coloring for algorithm [ alg : streaming ] : that is , let @xmath317 be the color @xmath313 was first assigned when the first hyperedge containing @xmath313 appeared in the input .
similarly , let @xmath318 be the sequence of bits generated by the above algorithm , and @xmath319 be the final permutation of vertices that results .
notice @xmath320 and @xmath321 have the same distribution . in algorithm
[ alg : off - line ] , once @xmath278 , @xmath322 and @xmath284 are fixed , the remaining actions are deterministic .
that is , @xmath323 is a function of @xmath320 .
similarly , once we fix ( i.e. , condition on ) @xmath321 , the final coloring @xmath324 is fixed .
[ samecoloring ] for all @xmath103 , @xmath322 and @xmath284 , we have @xmath325 .
suppose @xmath326 for some vertex @xmath313 .
one of them must equal @xmath273 .
we have two cases .
suppose @xmath327 : : : let us also assume that @xmath328 is a hyperedge that necessitated @xmath313 s recoloring in the off - line algorithm .
this implies that @xmath314 , and @xmath329 for the vertices @xmath330 of @xmath328 that appeared before @xmath313 in @xmath284 . now in the streaming algorithm above , when @xmath328 is considered , we would find that @xmath328 is monochromatic in @xmath278 , and flip @xmath315 unless it is already flipped .
once flipped , the color of @xmath313 will not change again .
thus , @xmath331a contradiction .
suppose @xmath332 : : : let @xmath328 be the hyperedge that necessitated the recoloring of @xmath313 in the above streaming algorithm .
this implies that @xmath314 , and all vertices @xmath330 of @xmath328 that appear before @xmath313 in permutation @xmath284 have @xmath329 .
but in such a situation , the original off - line algorithm will find @xmath328 monochromatic when @xmath313 is considered , and will flip its color .
thus , @xmath333a contradiction .
we conclude that ( [ eq : offlineerror ] ) applies to algorithm [ alg : streaming ] as well .
let @xmath1 be an @xmath0-uniform hypergraph with at most @xmath334 hyperedges .
then , with probability at least @xmath276 algorithm [ alg : streaming ] produces a valid two - coloring for @xmath4 .
note , however , algorithm [ alg : streaming ] lacks the desirable property that it outputs only valid colorings . a straight - forward check does not seem possible using @xmath335 ram space .
we show below that with carefully storing some vertices of a few hyperedges , we can in fact achieve this using only @xmath336 ram space . as remarked before the events considered in inequalities ( [ ineq : staysmono ] )
, ( [ ineq : becomesblue ] ) and ( [ ineq : becomesred ] ) cover all situations when the algorithm might return an invalid coloring .
so , in order never to return an invalid coloring , it suffices to guard against these events . 1 . to ensure that some vertex in every initially monochromatic hyperedge @xmath328 does change its color
, we just need to verify that @xmath314 for some @xmath337 .
this we can ensure by examining @xmath278 and @xmath322 just after they are generated in line [ linegeneratechi ] of algorithm [ alg : streaming ] .
2 . to ensure that no initially non - blue hyperedge @xmath328 has turned blue
, we will will save the red vertices of @xmath328 whenever there is potential for them all turning blue , and guard against their turning blue .
consider the hypergraph @xmath338 with hyperedges @xmath339 we will then verify that no hyperedge in @xmath338 is entirely blue in the end .
3 . to ensure that no initially non - red hyperedge @xmath328 has turned red , we consider the corresponding hypergraph @xmath340 with hyperedges @xmath341 and verify in the end that no hyperedge in @xmath340 is entirely red in the end . it remains to show how @xmath338 and @xmath340 can be stored efficiently .
each hyperedge will be stored separately by listing all its vertices .
thus , the expected sum of the sizes of the hyperedges in @xmath338 is @xmath342 & = & \sum_{h \in e(h ) } 2^{-{n } } \sum_{i=1}^{n}i { { n}\choose i } { p}^i \\ & \leq & |e|~ { n}p 2^{-{n } } \sum_{i=0}^{{n}-1 } { { { n}-1 } \choose { i-1 } } { p}^{i-1 } \\ & \leq & |e|~{n}p 2^{-{n } } ( 1+p)^{{n}-1}.\end{aligned}\ ] ] clearly , the same bound applies to @xmath340 .
we will choose the value of @xmath281 , so that with high probability @xmath343 and @xmath343 are both at most @xmath0 ( if either of them exceeds @xmath0 , the algorithm terminates with failure ) , while also ensuring that the right hand sides of ( [ ineq : staysmono ] ) , ( [ ineq : becomesblue ] ) and ( [ ineq : becomesred ] ) stay small .
as before set @xmath344 . to the failures accounted for in ( [ eq : offlineerror ] ) , we must account for @xmath343 or @xmath343 exceeding @xmath0 .
now , @xmath345
\leq \frac{1}{10 } \sqrt \frac{{n}}{\ln { n } } \frac{\ln { n}}{2 } \sqrt{\frac{{n}}{\ln { n } } } \leq \frac{{n}}{20}$ ] , and by markov s inequality @xmath346 \leq \frac{1}{20}$ ] .
similarly , @xmath347
\leq \frac{1}{20}$ ] .
thus , the revised algorithm fails to deliver a valid coloring with probability at most @xmath348 , while it uses at most @xmath349 bits of space .
the above results deal with the cases when a upper bound on the number of hyperedges of @xmath4 guarantees that it has a valid two - coloring . using the lovsz local lemma
one can show that @xmath0-uniform hypergraphs where no hyperedge intersects more than @xmath350 others has a two - coloring @xcite .
note that this does not require a bound on the number of hyperedges of @xmath4 , but only one on the number of intersections of any one hyperedge with others .
the algorithms of moser and tardos @xcite can now be used to recover a valid two - coloring of @xmath4 . as in the previous case , though the algorithm a - priori requires access to the entire hypergraph in an offline fashion , we can modify it to adapt it to the streaming setting .
the details are straightforward , and we omit them here to state only the result that derives from the parallel algorithm in @xcite .
if each hyperedge of a hypergraph @xmath4 intersects at most @xmath351 other hyperedges , then a two - coloring of the hypergraph can be found by a streaming algorithm that makes at most @xmath352 passes over the input stream and uses space @xmath335 .
compared to the moser - tardos algorithm , which works when the number of intersections is @xmath353 , this works only until intersections of up to @xmath354 .
furthermore , note that it requires @xmath355 passes over the input stream , in contrast to previous algorithms that required just one pass .
the upper bound @xmath47 was shown by erds by considering a random graph on @xmath19 vertices .
we ask if hypergraphs with a comparable number of hyperedges but supported on fewer vertices ( perhaps @xmath356 ) can be non - two - colorable .
we show that that the choice of @xmath19 above is essentially optimal .
[ thm : fewvertices ] let @xmath357 be a parameter , and @xmath4 be an @xmath0-uniform hypergraph on @xmath5 vertices and @xmath6 hyperedges with @xmath358 vertices .
then , a random two - coloring that partitions the vertex set into two almost equal color classes is a valid coloring with probability at least @xmath359 .
the following corollary is obtained by setting @xmath360 .
every @xmath0-uniform hypergraph with at most @xmath361 vertices and at most @xmath362 hyperedges is two - colorable .
the corollary shows that erds construction of non - two - colorable hypergraphs having @xmath363 hyperedges on a vertex set of size @xmath364 is nearly optimal in terms of the number of vertices .
suppose @xmath365 $ ] .
consider the coloring @xmath103 obtained by partitioning @xmath366 $ ] into two roughly equal parts and coloring one part @xmath274 and the other part @xmath275 .
every hyperedge has the same probability @xmath281 of being monochromatic under this coloring .
the theorem will follow immediately from the claim below using a union bound .
@xmath367 .
proof of claim : if @xmath368 , then this probability is @xmath282 . in general , the probability that a hyperedge is monochromatic is @xmath369 we then have @xmath370 now , @xmath371 this shows that @xmath367 and concludes the proof of the claim . to recover the statement of theorem [ thm : fewvertices ] , we union - bound over the set of hyperedges , and observe that @xmath372 . [
[ a - streaming - algorithm ] ] a streaming algorithm : + + + + + + + + + + + + + + + + + + + + + + theorem [ thm : fewvertices ] ( more generally , proposition [ prop : fewvertices ] at the end of this section for @xmath89-coloring ) gives rise to a straightforward randomized streaming algorithm for two - coloring ( @xmath89-coloring ) assuming that the number of vertices in the hypergraph is known in advance .
for example , suppose we are given an @xmath0-uniform hypergraph on @xmath373 vertices and at most @xmath374 hyperedges .
we just pick a random coloring @xmath103 that assigns colors red and blue to roughly the same number of vertices , and verify that no edge is monochromatic under @xmath103 .
the algorithm uses @xmath375 bits of space , and returns a valid coloring with probability @xmath276 and fails otherwise .
the probability of failure can be reduced to @xmath31 by running the algorithm @xmath376 times in parallel .
we further examine the trade - off between the number of vertices and the number of hyperedges .
let @xmath377 be the minimum number of hyperedges in a non - two - colorable hypergraph with @xmath13 vertices .
note that when the number of vertices @xmath378 , _ every _
@xmath0-uniform hypergraph is trivially two - colorable by any coloring that colors @xmath379 vertices blue and the rest red , and so we will take @xmath380 . the above proof ( of theorem [ thm : verticesnontwocolorable](a ) ) shows : @xmath381 using arguments similar to those used by erds @xcite , we will now obtain a comparable upper bound on @xmath377 .
this makes the dependence of two - colorability on the vertex set - size explicit .
suppose @xmath382 is an integer that is at least @xmath383 .
then , there is an @xmath0-uniform hypergraph with @xmath250 vertices and @xmath384 hyperedges that is not two - colorable .
thus , @xmath385
. our non - two - colorable @xmath0-uniform hypergraph will be obtained by picking the required number of hyperedges at random from @xmath386 $ ] .
fix a coloring @xmath103 .
then , a randomly drawn hyperedge is monochromatic with probability at least @xmath387 let @xmath4 be the hypergraph obtained by picking @xmath6 hyperedges uniformly and independently from @xmath388 \choose { n}}$ ] . the probability that @xmath103 is valid for @xmath4 ( i.e. , none of its hyperedges is monochromatic under @xmath103 ) is at most @xmath389 since there are at most @xmath390 colorings , the probability that some coloring @xmath103 is valid for @xmath4 is at most @xmath391 if @xmath392 this quantity is less than @xmath280 .
our claim follows from this .
the above results can be generalized easily to @xmath89-colorings .
let @xmath393 be the minimum number of hyperedges in a non-@xmath89-colorable hypergraph with at most @xmath13 ( @xmath394 ) vertices .
we then have the following bounds ( we omit the proofs ) .
[ prop : fewvertices ] @xmath395 .
the lower bound we obtain on the space requirements of one - pass streaming algorithms is optimal ( up to @xmath37 ) factors .
we present an efficient two - player two - round deterministic communication protocol for two - coloring @xmath0-uniform hypergraphs with up to @xmath396 hyperedges , but we do not know if there is a corresponding streaming algorithm . as mentioned in the introduction , it is easy to come up with a deterministic streaming algorithm that works in @xmath40 passes using the method of conditional expectations . even in the two - player communication setting
it would be interesting to determine if the protocol can accommodate up @xmath397 hyperedges , perhaps even @xmath398 hyperedges .
our two - coloring algorithm for hypergraphs with @xmath399 vertices does not improve on the bound provided via the delayed recoloring algorithm when @xmath50 is small , say , @xmath400 . we believe it should be possible to combine our argument and the delayed recoloring algorithm to show that if the number of vertices is @xmath49 then we can two - color hypegraphs with strictly more than @xmath46 hyperedges .
s. shannigrahi .
coloring , embedding , compression and data structure problems on uniform hypergraphs _ ph.d .
thesis , school of technology and computer science , tata institute of fundamental research , 2011_. n. alon , y. matias and m. szegedy .
the space complexity of approximating the frequency moments . _ proceedings of the twenty - eighth annual acm symposium on theory of computing ( stoc 96 ) pages 20 - 29 , 1996 .
_ joan feigenbaum , sampath kannan , andrew mcgregor , siddharth suri , jian zhang .
_ on graph problems in a semi - streaming model _ , proc .
31st international colloquium on automata , languages and programming , ( icalp ) , v. 3182 : 531 - 543 , 2004 . | we consider space - efficient algorithms for two - coloring @xmath0-uniform hypergraphs @xmath1 in the streaming model , when the hyperedges arrive one at a time .
it is known that any such hypergraph with at most @xmath2 hyperedges has a two - coloring @xcite , which can be found deterministically in polynomial time , if allowed full access to the input .
* let @xmath3 be the minimum space used by a deterministic ( one - pass ) streaming algorithm that on receiving an @xmath0-uniform hypergraph @xmath4 on @xmath5 vertices and @xmath6 hyperedges @xmath7 , produces a proper two - coloring of @xmath4 . we show that @xmath8 .
* let @xmath9 be the minimum space used by a randomized ( one - pass ) streaming algorithm that on receiving an @xmath0-uniform hypergraph @xmath4 on @xmath5 vertices and @xmath6 hyperedges with high probability produces a proper two - coloring of @xmath4 ( or declares failure ) .
we show that @xmath10 . *
we show that for any @xmath11 , @xmath12 ; in particular , this shows that every @xmath0-uniform hypergraph with at most @xmath13 vertices and @xmath14 hyperedges is two - colorable .
the above results are inspired by the study of the number @xmath15 , the minimum possible number of hyperedges in a @xmath0-uniform hypergraph that is _ not _ two - colorable .
it is known that @xmath16 and @xmath17 .
the lower bound ( due to radhakrishnan and srinivasan @xcite ) has a corresponding algorithm to deterministically produce the two - coloring ; the upper bound due to erds is obtained by picking about @xmath18 hyperedges randomly from a vertex set of @xmath19 vertices .
our first result shows that no efficient deterministic streaming algorithm can match the performance of the algorithm in @xcite ; the second result shows that there is , however , an efficient randomized algorithm for the task ; the third result shows that that if the number of vertices is substantially smaller than @xmath19 , then every non - two - colorable hypergraph has significantly more than @xmath18 hyperedges . |
this summary is the write - up of a quick and spontaneous tour through the experimental results presented at the conference , rather than a deep discussion of the experimental situation of qcd in general . among the accelerators , which presently provide results
, hera concentrates most on topics related to qcd , which is at the heart of interests of the hsqcd organisers at gatchina . and , in spiritu loci
, many analyses presented at this conference consider the limitations of dglap evolution , ask whether bfkl evolution might be more appropriate .
but there is progress in many other respects as well , e.g. parton density analyses get more precise and more general , similarly , the data on charm and beauty production in @xmath0 interactions improve rapidly , the spin structure of the proton is studied in more and more detail , rhic continues to provide interesting results on hadronic high density states .
although not all experiments , which produce relevant results in context of qcd were represented , a wide spectrum of interesting results was displayed .
qcd is generally discussed in the frame of the standard model .
i will therefore begin with searches for physics beyond it .
d. south @xcite reported on searches at hera for physics beyond the standard model ( sm ) which was not really shaken yet , although in @xmath1 the h1 collaboration observes more events than expected with isolated leptons and large missing transverse momentum and large @xmath2 ( @xmath3 gev ) , the transverse momentum of the summed hadronic final state .
however in this kinematic range the statistics is still rather small , with 15 events with isolated electrons or muons for a sm expectation of @xmath4 events .
this excess is not confirmed by the zeus collaboration and is not observed in @xmath5 data .
it is certainly very important to take as much data as possible , but it might be difficult to get full clarity in the remaining hera running time .
the same is true for the multi - lepton final states , where 6 multi - electron states with a di - electron invariant mass above 100 gev were recorded by h1 , 3 @xmath6 events and 3 @xmath7 events with sm expectations of @xmath8 and @xmath9 events , respectively .
with increasing statistics , hera enters a phase of mature electro - weak physics analyses .
the available nc and cc data give the opportunity to demonstrate nicely the unification of electro - weak physics ( see @xcite ) .
the absence of right handed weak currents could be demonstrated by the the dependence of the cc cross section on the polarisation of the incident @xmath10 beam ( see figs . in @xcite ) .
the propagator mass in cc reactions is measured with a precision of about 2 gev and found to be consistent with the time like @xmath11 mass determined in @xmath12 and @xmath13 .
finally , first results are presented for light quark-@xmath14 couplings @xcite which supplement the lep and tevatron measurements . since its beginning
, hera is the world supplier for information on the proton structure at low @xmath15 . after the discovery of the rise of @xmath16 towards small @xmath15 @xcite , this rise is described with continuously increasing precision in terms of parton density functions ( pdfs ) ( see @xcite ) . now
the hera collaborations provide pqcd analyses of their nc and cc data with the aim to include as little other data as possible , thereby avoiding additional systematic uncertainties .
in particular , due to the sensitivity of the cc reaction @xmath17 to the valence @xmath18 quark distribution , pdf fits without inclusion of @xmath19 scattering data are possible , thereby avoiding nuclear corrections .
the zeus collaboration provided also a pdf fit @xcite using jet data , in addition to the inclusive data , thereby exploiting their sensitivity to the gluon density via the boson - gluon fusion interaction .
the pdf fits recently provided by the hera collaborations are consistent with the results of the global fitters @xcite .
still increasing precision can be expected at small @xmath15 from the final analysis of the full hera i data including results at large @xmath20 which are sensitive to the longitudinal structure function @xmath21 and therefore to the gluon density .
in addition new techniques are explored to improve the nc measurements at very high @xmath15 @xcite .
thus , further progress in the determination of parton densities can be expected in future .
complementary to inclusive @xmath0 scattering ( previous section ) , the convolution of proton pdfs with the hard interaction processes can be studied in detail by the observation of jets in deep inelastic scattering ( dis ) and photoproduction @xcite .
the high precision hera data at large @xmath22 and large transverse momenta provide , besides detailed comparisons with pqcd calculations , results on the strong coupling constant @xmath23 which turn out to be consistent and competitive with results from inclusive @xmath0 or @xmath24 data . as mentioned already above
, the inclusion of @xmath25 and dis jet data in pdf fits improves the determination of the gluon density in the proton at medium and large @xmath15 , where the inclusive nc data have only little sensitivity . in most of the phase space ,
the experimental uncertainties of the available jet data are smaller than the theoretical ones .
extra complications appear in the theoretical description of dijet production in dis at low @xmath22 and large @xmath26 . here
a resolved virtual photon contribution is expected , where a partonic fraction of the virtual photon enters the hard interaction .
this is supported by the data which are not described in this region by nlo calculations with only direct photon interactions .
such a resolved component improves also the description of forward jet " data @xcite . by requiring @xmath27 and jet angles as close to the proton beam as possible with the h1 and zeus calorimeters , a phase space is selected , where the data might discriminate the calculations based on dglap evolution or ccfm or bfkl evolutions .
less jet production is obtained in the @xmath28 ordered dglap evolution than in unordered emissions like in the cdm model , which might be close to bfkl expectations , or the ccfm approach ( angular ordering ) .
indeed , the dglap based models produce too small cross sections at small @xmath29 compared to the measurements of h1 and zeus
. however the best descriptions are obtained in conceptually quite different models like cdm with unordered emissions , and the dglap based rapgap model , where the strict @xmath28 ordering is broken by a resolved photon component .
interesting to note that for a more special event selection , requiring besides the forward jet two additional central jets , the resolved rapgap model also fails , in contrast to cdm @xcite .
impressive data on inclusive single jet production in @xmath30 were presented by the d0 collaboration @xcite .
the @xmath31 distributions agree beautifully with pqcd in nlo .
angular correlations demonstrate very well that there is no need for higher orders .
however the large effect of the experimental uncertainty of the jet energy scale on the cross section measurement allows no conclusions on pdfs to be drawn .
the observation of diffractive phenomena in dis at hera gives hope to reach , in terms of qcd , full understanding of the regge exchange processes , which dominate the strong interactions processes at high energies .
so far , the efforts to describe the colour singlet exchange in diffractive interactions by pdfs are partially successful , as diffractive inclusive data , diffractive jet production and diffractive charm production can indeed be described in a common approach @xcite .
these pdfs , however , can not be naively applied to diffractive @xmath30 data , not even to diffractive photoproduction of jets . in the latter case ,
the predictions , which are based on the pdfs deduced from inclusive diffraction , are above the data by about a factor two .
this may indicate that in photoproduction , as in @xmath30 interactions , the gap signature of the diffractive interaction is more often destroyed than in dis by additional parton - parton interactions .
more results are expected from hera and the hadron colliders which may help to describe diffractive phenomena coherently .
new results were presented on @xmath32 photo- and electro production at low @xmath33 , on @xmath32 production in dis at high @xmath33 , as well as high @xmath33 @xmath34 photoproduction , and on @xmath35 production in dis @xcite .
photo and electro - production of vector mesons ( vm ) is an interesting testing ground for qcd due to the interplay of soft processes described by regge theory and hard interactions described by pqcd . for the latter different hard scales
may be involved , @xmath22 , the vm mass , the momentum transfer @xmath33 .
the presented results show that whenever hard scales are involved , the data can not be described by soft pomeron exchange .
vector mesons are particularly suited for detailed qcd calculations as the measured decay angular distributions provide information on the helicity structure of the interaction . such an analysis was presented for @xmath36 at high @xmath33 @xcite .
@xmath32 production in 920 gev proton - carbon interactions where presented by the hera - b collaboration which extend the available @xmath37 distributions far into the target region @xcite ( see also the discussion @xcite ) .
results on deeply virtual compton scattering ( dvcs , @xmath38 ) were presented by the hermes @xcite , h1 and zeus collaborations @xcite . the reaction @xmath38 , which , in contrast to vm production , does not require knowledge of a non - perturbative wave function , gives in principle access to generalised non collinear parton densities .
hermes obtains dvcs amplitudes by measuring asymmetries due to the interference of bethe heitler ( bh ) @xmath39 emissions from electrons with the dvcs process .
there are already first results on @xmath40-beam - spin and on beam - charge asymmetries and on longitudinal and transverse target spin asymmetries , both on @xmath41 and @xmath18 targets .
more data are expected with improved control of the elasticity of the final state with the new recoil detector of hermes .
h1 and zeus have not yet exploited the bh - dvcs interference , which vanishes when integrated over the phase space of the measurements , but provide dvcs cross section measurements which are described by nlo pqcd calculations and various colour dipole models .
the measured energy dependence of the cross section is characteristic of processes where a hard scale is involved .
precise measurements of the @xmath33 dependence allow the theoretical predictions to be normalised .
it remains to hope that the future data will distinguish better between the different theoretical approaches .
heavy flavour production in electro and photoproduction is expected to be well described by pqcd due to the large scale given by the heavy quark mass .
it was therefore surprising , that in spite of this advantage , beauty @xmath42 production cross sections were notoriously above the pqcd nlo predictions @xcite .
this is still the case , but it is interesting to note that the available more inclusive measurements with large acceptance , which use @xmath43 identification by life time tags , are consistent with the predictions .
the contributions of charm and beauty to the inclusive proton structure function @xmath16 are substantial , about 20% to 30% for charm and about 0.3% to 3% for beauty , in the covered kinematic region @xcite .
fragmentation properties of @xmath44 quarks were studied by h1 in dis and with more statistics in @xmath25 by zeus @xcite . by and large ,
the findings are consistent with @xmath24 results , supporting universality " of @xmath44 quark fragmentation .
results on hadroproduction of open beauty near threshold were presented by hera - b @xcite which are reasonably consistent with fermilab data and pqcd calculations which include nnlo soft gluon corrections @xcite .
the hermes experiment @xcite produces a wealth of results on the many measurable effects of the proton spin in electron nucleon scattering when besides the scattered electron one hadron is detected . in particular the effects of the nucleon spin polarisation
transverse to the electron beam direction ( transversity " ) lead to sophisticated analyses which distinguish between the effect of the quark polarisation on the @xmath31 transverse to it in fragmentation ( collins effect ) , and the effect of the struck quarks @xmath31 in the polarised nucleon on the @xmath31 in fragmentation ( sivers effect ) , which gives access to the quark angular momentum . for @xmath45
the collins moments are negative but larger in absolute than expected .
the sivers moments show a striking difference between @xmath46 and @xmath45 , the latter being consistent with zero , the first definitely positive which points to the valence @xmath47 content of the @xmath48 .
the expected future data and further theoretical analyses might eventually lead to a clear understanding of the proton spin . ] .
many results exist on the polarisation of hyperons produced in hadron beams , without clear theoretical understanding @xcite .
in particular , the results obtained by wa98 in the 340 gev @xmath49 beam on the polarisation of inclusively produced @xmath50 and @xmath49 baryons originally came as a surprise with a large unexpected positive polarisation of the @xmath51 s ( normal to the production plane ) .
meanwhile explanations exist in terms of recombination of a quark with a strange di - quark @xcite .
however no coherent theoretical approach to the observed detailed polarisation phenomena exists yet .
food for theorists ! " ( w.h .
siebert @xcite )
rhic data show a strong azimuthal anisotropy of the particle flow in au au collisions at @xmath52 gev @xcite . the particle flow is increased in the direction transverse to the reaction plane .
it is understood to be sensitive to the initial geometric overlap of the colliding nuclei as well as the later expansion driven by the initial pressure .
the effect is seen for all final state hadrons and the results suggest @xcite early formation of the anisotropy in a still partonic phase . to clarify the mechanism , it is particularly interesting to measure the energy flow for direct photons . as not strongly interacting , direct photons penetrate through hadronic clouds and are known to be not suppressed in heavy ion collisions @xcite in contrast to hadrons . with present statistics the data on prompt photons are not yet conclusive @xcite .
a new result of a pentaquark search was reported @xcite from the svd-2 detector using the 70 gev proton beam at ihep . in the reaction
@xmath53 a bump is seen at the @xmath54 mass @xmath55 ( syst . )
mev with a width @xmath56 mev , which can be interpreted as the signal of @xmath57 pentaquark @xmath58 .
the estimated significance is 5.6 @xmath59 .
the many negative and positive evidences for pentaquarks were discussed in a review by v. popov @xcite .
l. gladilin @xcite reported on the results from hera which are not yet clear , with zeus seeing a @xmath58 signal in contrast to h1 and h1 seeing a candidate for a charmed pentaquark @xmath60 at a mass of 3100 gev which is not confirmed by zeus and other experiments .
negative results were reported for @xmath61 @xcite . also other non standard hadronic states were discussed , namely possible tensor glue balls with masses around 2000 mev @xcite the great time of rapid pentaquark discoveries seems to be over , but there are still many interesting signals which need final confirmation or other explanations .
the approved * f*acility for * a*ntiproton and * i*on * r*esearch ( fair ) at gsi , darmstadt ( germany ) is expected to provide in 2013 very intensive beams ( e.g. primary beams of @xmath62 up to 35 gev / u and @xmath63 protons of 30 gev ) @xcite .
besides various radioactive secondary beams , a @xmath64 beam will be available for the planned experiment panda which foresees a rich program on hadron spectroscopy , in particular for searches of glue balls or exotic hadrons and detailed studies of charmonia and production of open charm .
it is seen as an advantage of @xmath30 compared to @xmath24 interactions that the formation of states is not restricted to the virtual photon quantum numbers @xmath65 .
thus , fair will be after 2013 an interesting facility for hadron physics in the low and medium energy range .
there is still intensive experimental activity addressing qcd related questions .
hera , the tevatron , rhic will provide important information before the lhc will produce results .
we experimentalists have not yet presented results which need bfkl beyond any doubt , however some final state analyses prefer non @xmath28 ordered gluon emissions . at next hsqcd
we will know considerably more .
i thank all the organisers , in particular victor kim and lev lipatov for a stimulating conference with very lively discussions in the sessions and the breaks .
i thank emmanuelle perez for reading the draft and helpful comments .
d. m. south [ h1 collaboration ] , these proceedings ; arxiv : hep - ex/0602028 .
yongdok ri , these proceedings ; ibid .
j. gayler ; a. aktas _ et al . _
[ h1 collaboration ] , phys .
b * 634 * , 173 ( 2006 ) [ hep - ex/0512060 ] ; [ zeus collaboration ] , hep - ex/0602026 .
a. aktas _ et al .
_ [ h1 collaboration ] , phys . lett .
b * 632 * ( 2006 ) 35 .
s. chekanov _ et al . _
[ zeus collaboration ] , eur .
j. c * 42 * ( 2005 ) 1 .
a. d. martin , r. g. roberts , w. j. stirling and r. s. thorne , eur .
j. c * 23 * ( 2002 ) 73 .
j. pumplin , d. r. stump , j. huston , h. l. lai , p. nadolsky and w. k. tung , jhep * 0207 * ( 2002 ) 012 .
m. sutton , these proceedings .
s. s. adler _ et al . _ [ phenix collaboration ] , phys .
* 94 * ( 2005 ) 232301 .
s. s. adler _ et al .
_ [ phenix collaboration ] , phys .
* 96 * ( 2006 ) 032302 .
v. popov , these proceedings ; a. kubarovsky and v. popov [ svd collaboration ] , hep - ex/0510006
. v. popov , these proceedings . | a summary of experimental contributions to hsqcd 2005 is given .
this includes results from the four hera experiments h1 , zeus , hermes and hera - b , as well as results from the tevatron ( d0 ) , the cern hyperon beam ( wa98 ) , from rhic ( phenix ) , and ihep ( svd ) .
i present a short survey of the points appearing most relevant to me . |
the generalized seniority @xcite or broken pair @xcite framework provides an approximation ( or truncation ) scheme for nuclear shell model calculations . if strong pairing correlations are present , the generalized seniority scheme can represent these correlations in a space of greatly reduced dimensionality relative to the full shell model space .
the approach also serves as the basis for mappings from the shell model to bosonic collective models @xcite .
the underlying premise of generalized seniority is that the ground state of an even - even nucleus can be well approximated by a condensate built from _ collective _ @xmath2 pairs , which are defined as a specific linear combination of pairs of nucleons in the different valence orbitals , each pair coupled to angular momentum zero .
( the conventional seniority approach , in contrast , considers pairs within the different orbitals separately and does not directly address the correlations between orbitals @xcite . ) for a certain very restricted class of interactions @xcite , generalized seniority describes an exact scheme for obtaining certain states : the @xmath3 ground state is exactly of the @xmath2 condensate form , and the first @xmath4 state involves exactly one broken @xmath2 pair .
however , for generic interactions , the generalized seniority approach ( or , equivalently in this context , the broken pair approach ) constitutes a truncation scheme for the shell model space , in which the ground state and low - lying states are represented in terms of a condensate of collective @xmath2 pairs together with a small number @xmath5 ( the _ generalized seniority _ ) of nucleons not forming part of an @xmath2 pair .
the generalized seniority scheme is closely related to the bcs scheme with quasiparticle excitations : the @xmath2 condensate has the same form as a number - projected bcs ground state , and the model space with generalized seniority @xmath5 is identical to the space of number - projected bcs @xmath5-quasiparticle states @xcite .
the essential difference is that in the generalized seniority scheme diagonalization is carried out on states of definite nucleon number , that is , after projection rather than before projection .
although the generalized seniority approach has long been applied in various contexts ( _ e.g. _ , refs .
@xcite ) , it has not been systematically benchmarked against calculations carried out in the full shell model space .
only recently have limited comparisons been made , for certain even - mass light @xmath6 isotopes @xcite and even - mass light @xmath0 isotopes @xcite . otherwise , extensive previous studies with generalized seniority bases , as reviewed in ref .
@xcite , have instead compared the generalized seniority results with experiment .
such comparisons do not disentangle the question of how accurately the truncated calculation approximates the full - space calculation from the largely unrelated question of how physically appropriate the assumed interaction and model space are for description of the particular set of experimental data .
these comparisons were also mostly based on schematic interactions , _
e.g. _ , pairing plus quadrupole , phenomenologically adjusted to a small set of experimental observables .
the purpose of the present work is to establish a benchmark comparison of the results obtained in a generalized seniority truncated model space against those obtained in the full shell model space , for a full major shell and with realistic interactions .
in particular , we consider the @xmath0 isotopes ( @xmath7@xmath8 ) , in the @xmath9-shell model space , with the fpd6 @xcite and gxpf1 @xcite interactions .
both even - mass and odd - mass isotopes are considered , with generalized seniority @xmath10 , that is , at most one broken @xmath2 pair .
truncation of the model space according to the generalized seniority drastically reduces the dimensionality of the shell model space .
the full system of valence nucleons is effectively replaced by a much smaller system , consisting of just the unpaired nucleons , either @xmath11 or @xmath12 in the present calculations .
shell model calculations in the full valence space are now well within computational reach @xcite for semimagic nuclei .
therefore , for semimagic nuclei , the immediate implications of the generalized seniority scheme are _ conceptual _ , that is , to the interpretation of the shell model results in terms of collective pairs , and therefore also indirectly to assessing the plausibility of generalized seniority as the basis for boson mapping , rather than to extending _ computational _ capabilities . however , if one moves away from semimagic nuclei , to nuclei which simultaneously have large numbers of valence protons and neutrons , full shell model calculations are still computationally prohibitive . generalized seniority might therefore be of direct computational value in making calculations for these nuclei tractable , provided that the seniority - violating nature of the proton - neutron interaction @xcite does not necessitate an impractically large number of broken pairs for an accurate description of these nuclei . the construction of the generalized seniority basis and other technical aspects of the calculational method are outlined in sec .
[ sec - methods ] .
the calculations for the @xmath0 isotopes in the generalized seniority scheme are then described ( sec .
[ sec - results - over ] ) , and detailed comparisons with the full shell model results are made for the level energies ( sec .
[ sec - results - energy ] ) , orbital occupations ( sec .
[ sec - results - occ ] ) , and electromagnetic observables ( sec .
[ sec - results - trans ] ) .
in order to define the generalized seniority basis , let us first introduce some basic notation .
let @xmath13 be the creation operator for a particle in the shell model orbital @xmath14 , with angular momentum projection quantum number @xmath15 .
the angular - momentum coupled pair creation operator is then @xmath16 for a pair of angular momentum @xmath17 .
here we follow standard angular momentum coupling notation for spherical tensors .
the collective @xmath2 pair of the generalized seniority scheme is defined by @xmath18 where @xmath19 runs over the active orbitals , and @xmath20 .
this operator creates a linear combination of pairs in different orbitals @xmath19 , with respective amplitudes @xmath21 .
a basis state within the generalized seniority scheme consists of a `` condensate '' of collective pairs , together with @xmath5 additional nucleons not forming part of a collective @xmath2 pair .
the number @xmath5 is the _ generalized seniority _ of the state . in the present calculations , we consider semimagic nuclei
, so only like valence particles ( here , all neutrons ) are present .
however , it should be noted that a generalized seniority basis can be defined equally well for nuclei with valence particles of both types via a proton - neutron scheme , that is , by taking all possible products of proton and neutron generalized seniority states with generalized seniorities @xmath22 and @xmath23 .
further discussion of the basis may be found in refs .
the notation and methods used in the present work are established in detail in ref .
@xcite . for even - mass nuclei
, the @xmath2 condensate state , with @xmath24 , is defined as @xmath25 .
this state has @xmath26 valence nucleons ( _
i.e. _ , @xmath27 is the number of valence pairs ) and angular momentum @xmath28 .
it can be shown @xcite that the @xmath2-pair condensate state is simply the number - projected bcs ground state , and the amplitudes @xmath21 are related to the standard bcs occupancy parameters @xmath29 and @xmath30 by @xmath31 . the @xmath32 model space for angular momentum @xmath17 , in turn ,
is spanned by states of the form @xmath33 , that is , with @xmath34 @xmath2 pairs and one `` broken '' collective pair . for odd - mass nuclei , there must be at least one unpaired nucleon . adding a single nucleon to the @xmath2 condensate yields the @xmath35 state @xmath36 , with @xmath37 valence nucleons .
a different such state is obtained for each choice of valence orbital @xmath19 for the added nucleon , and the resulting state has angular momentum @xmath38 . in the context of bcs theory ,
these states are number - projected one - quasiparticle states @xcite .
the @xmath39 model space is spanned by states of the form @xmath40 , obtained by breaking one @xmath2 pair . before calculations can be carried out in the generalized seniority model space , an orthonormal basis
must be constructed , and matrix elements of the hamiltonian must be obtained with respect to this basis .
the states used just above to define the generalized seniority model space are not normalized .
moreover , for @xmath41 , they are not mutually orthogonal , and , for @xmath42 , they are linearly dependent , _
i.e. _ , constitute an overcomplete set .
however , a suitable basis is obtained by a gram - schmidt procedure , which yields orthogonal , normalized , and linearly - independent basis states as linear combinations of the original basis states , _
e.g. _ , for @xmath32 , @xmath43 here @xmath44 is simply a counting index , labeling the orthonormal states , and the @xmath45 coefficients are determined in the gram - schmidt procedure , from the overlaps of the original nonorthogonal basis states , _
e.g. _ , @xmath46 for @xmath32 or @xmath47 for @xmath39 , which are calculated as described below .
the size of the resulting basis is the same as for the shell model problem with only @xmath5 particles in the same set of orbitals , regardless of the number of pairs .
increases to the point where fewer than @xmath5 vacancies remain among the active orbitals , the dimensionality of the problem is instead that of the shell model problem defined by the remaining number of holes . ]
the dimensions for the present @xmath9-shell calculations are summarized in table [ tab - dim ] .
rrrrrrrr & & + + @xmath17&@xmath48&@xmath11&@xmath49&@xmath50&@xmath51&@xmath52&@xmath53 + dimension&@xmath49&@xmath54&@xmath55&@xmath56&@xmath57&@xmath58&@xmath59 for a semimagic nucleus , the valence shell contains only like nucleons , and the two - body nuclear hamiltonian in this proton space or neutron space may then be expressed as @xcite @xmath60 where the @xmath61 are the single - particle energies , the @xmath62 are number operators for the orbitals , the @xmath63 are like - nucleon normalized , antisymmetrized two - body matrix elements , and the phase convention @xmath64 is used . to construct the hamiltonian matrix for diagonalization
, matrix elements of the one - body and two - body terms appearing in the hamiltonian must be obtained . several approaches @xcite
have been developed for evaluating matrix elements in the generalized seniority basis , together with the overlaps required ( as discussed above ) for the orthogonalization process .
the present calculations have made use of the recurrence relations derived in ref .
matrix elements are first calculated with respect to the original nonorthogonal , unnormalized , and overcomplete generalized seniority basis .
these matrix elements are then transformed to the orthonormal basis via ( [ eqn - gram - schmidt-2 ] ) or ( [ eqn - gram - schmidt-3 ] ) .
note that the occupations @xmath62 in ( [ eqn - h ] ) are not diagonal in the generalized seniority basis , so matrix elements must explicitly be obtained for @xmath62 as a one - body operator .
after diagonalization , the same set of recurrence relations is used for the evaluation of matrix elements of elementary multipole operators @xmath65 , needed for the calculation of one - body densities and , from these , observables .
fundamental to the definition of the generalized seniority model space is the choice of values for the coefficients @xmath21 appearing in the collective pair ( [ eqn - defn - s ] ) .
these coefficients enter into the computation of the overlaps and matrix elements of the generalized seniority states . for even - mass nuclei ,
the coefficents are commonly chosen variationally , so as to minimize the energy functional @xmath66 @xcite . for odd - mass nuclei , prior calculations ( _ e.g. _ , ref .
@xcite ) have commonly taken the @xmath21 values from the neighboring even - mass nuclei . here , we have deduced the coefficients for the odd - mass nuclei directly , by variationally minimizing the energy expectation @xmath67 for the @xmath35 state .
the result for the @xmath21 parameters may be expected to depend upon the particular choice of quasiparticle for the @xmath35 state , _
i.e. _ , the orbital for the creation operator @xmath68 . comparisons of these prescriptions for the present @xmath9-shell calculations are given in sec .
[ sec - results - over ] .
in the following , we consider the semimagic @xmath0 isotopes , treated as consisting of neutrons in the @xmath9-shell model space ( _ i.e. _ , the @xmath69 , @xmath70 , @xmath71 , and @xmath72 orbitals ) .
the calculations cover the entire sequence of isotopes possible within this set of orbitals , namely , @xmath73 ( _ i.e. _ , @xmath74{ca}$]@xmath75{ca}$ ] ) .
the generalized seniority treatment may thus be traced from the beginning of the shell ( filling of the isolated high-@xmath76 @xmath77 orbital ) , across the subshell closure , to the end of the shell ( filling of closely - spaced lower-@xmath76 orbitals ) .
both the fpd6 @xcite and gxpf1 @xcite interactions are used in the calculations , as representative realistic interactions for the @xmath9 shell .
the primary interest here is systematic comparison of _ calculational _ results in truncated and full spaces , so the same interactions and model space are used throughout , even though for the highest - mass isotopes a more _ physically _ relevant description would likely require inclusion of the @xmath78 orbital , as well as modification of the interactions .
the even - mass @xmath0 isotopes are considered in the @xmath32 generalized seniority model space , and the odd - mass @xmath0 isotopes are considered in both the @xmath35 and @xmath39 model spaces .
that is , for both sets of isotopes , at most one @xmath2 pair is broken .
these results are benchmarked against results obtained in the full shell model space , calculated using the code nushellx @xcite . near the beginning and end of the shell , where there are few particles or few holes ,
the generalized seniority calculation is strictly equivalent to the full shell model calculation .
that is , for the @xmath79-particle or @xmath79-hole nuclei ( @xmath80 and @xmath81 ) , the @xmath35 model space is identical to the full model space ; for the @xmath11-particle or @xmath11-hole nuclei ( @xmath82 and @xmath83 ) , the @xmath32 model space is identical to the full model space ; for the @xmath12-particle or @xmath12-hole nuclei ( @xmath84 and @xmath85 ) , the @xmath39 model space is identical to the full model space , _ etc . _ in the following analysis , we consider the level energies ( sec .
[ sec - results - energy ] ) , orbital occupations ( sec .
[ sec - results - occ ] ) , and electromagnetic observables ( sec .
[ sec - results - trans ] ) .
the focus is on the ground state and lowest - lying excited states , as these are expected to require the fewest broken pairs for their description .
specifically , the lowest @xmath28 , @xmath11 , and @xmath49 states are taken for the even - mass @xmath0 isotopes , along with the first excited @xmath28 state .
for the odd - mass @xmath0 isotopes , the lowest states of @xmath86 , @xmath51 , @xmath52 , and @xmath53 are considered .
these correspond to the @xmath76-values of the orbitals in the @xmath9 shell and thus are the angular momenta which can arise in the one - quasiparticle ( @xmath35 ) description .
the coefficients @xmath21 appearing in the collective @xmath2 pair , as obtained according to the variational procedures described in sec .
[ sec - methods ] , are summarized in fig .
[ fig - alpha ] .
the coefficients obtained under the usual procedure , using the @xmath24 condensate for the variation , are shown in fig .
[ fig - alpha](a ) , while those obtained obtained by minimizing @xmath87 for the @xmath35 states are shown in fig .
[ fig - alpha](b d ) , for states built by creating a quasiparticle in the @xmath77 , @xmath88 , and @xmath89 orbitals . for the one - quasiparticle state based on the @xmath90 orbital does not uniquely determine the @xmath21 coefficients ( in general , this is true for any @xmath91 orbital ) . ]
the fpd6 interaction is taken for illustration , but similar results are obtained with gxpf1 . only _ ratios _ of @xmath21 values are relevant , since the overall scale is set by the conventional normalization @xmath92 for the @xmath2 pair @xcite . throughout the shell ,
the amplitude for the @xmath77 orbital dominates , followed by that of the @xmath88 orbital .
these are the orbitals with the lowest single - particle energies , ( @xmath77 ) , @xmath93 ( @xmath88 ) , @xmath94 ( @xmath90 ) , and @xmath95 ( @xmath89 ) . for gxpf1 ,
the corresponding values are approximately @xmath96 ( @xmath77 ) , @xmath97 ( @xmath88 ) , @xmath98 ( @xmath90 ) , and @xmath99 ( @xmath89 ) .
] respectively , so the result is consistent with natural filling order .
notice that the amplitudes for the orbitals other than @xmath77 dip sharply at the @xmath77 subshell closure ( @xmath100 ) .
the results may be contrasted with the situation for an ideal generalized seniority _ conserving _ interaction , as considered by talmi @xcite , for which the @xmath21 values would be constant across the shell .
the results obtained by the different prescriptions in fig . [ fig - alpha ] are not seen to differ from each other in any substantial qualitative fashion . for the odd - mass @xmath0 isotopes , only the results of calculations carried out using coefficients obtained from the variation involving the @xmath77 quasiparticle [ as in fig .
[ fig - alpha](b ) ] will be used in the following discussions .
the other choices lead to differences in the quantitative details but give similar overall results .
let us begin with the even - mass @xmath0 isotopes , by considering the energy eigenvalue of the @xmath28 ground state , shown in fig .
[ fig - ca - even - energy ] , _ i.e. _ , the valence shell contribution to the nuclear binding energy .
it is worth first noting some properties of the @xmath32 model space for @xmath28 .
this space is spanned by the four states @xmath101 , where @xmath19 runs over the four @xmath9-shell orbitals .
space , for @xmath28 , lies entirely within the conventional seniority zero space .
the nucleons in the pair @xmath102 are coupled pairwise to angular momentum zero , and thus carry zero _ conventional _ seniority , but these nucleons do not participate in the collective @xmath2 pair of ( [ eqn - defn - s ] ) , so they do still contribute to the _ generalized _ seniority .
] however , when the coefficients @xmath21 appearing in the @xmath2 pair are chosen so as to minimize the energy of the @xmath24 condensate state @xmath103 , as described in sec .
[ sec - methods ] , the ground state obtained by diagonalization in the @xmath32 space is still simply this condensate ( see appendix ) .
thus , as far as the ground state is concerned , the @xmath32 results are identical to those for the @xmath24 condensate or , equivalently , to the results of number - projected bcs , with variation after projection . the ground state energy eigenvalue itself is shown in fig .
[ fig - ca - even - energy](a , b ) , as a function of @xmath104 , for the fpd6 and gxpf1 interactions .
the results obtained in the generalized seniority @xmath32 model space and in the full shell model space are overlaid in these plots . however , the deviations are so small that the results are essentially indistinguishable , when viewed on the @xmath105 energy scale necessary to accomodate the eigenvalues .
the two - neutron separation energy [ @xmath106 , shown in fig . [ fig - ca - even - energy](c , d ) , reveals finer details , in particular differences between the _ interactions _ ( comparing the two panels ) , but the generalized seniority and full shell model results are still largely indistinguishable at this scale .
a distinctive feature of generalized seniority as an _ exact _ symmetry , in the sense of talmi @xcite , is that the ground state energies vary quadratically across the shell , and the separation energies are therefore strictly linear in @xmath104 , insensitive to any subshell closures @xcite .
however , as observed in ref .
@xcite , when generalized seniority is simply used as a variational approach , there is no such constraint , and it is possible to obtain subshell effects .
therefore , it is worth noting the jump in @xmath107 at the @xmath77 subshell closure ( @xmath100 ) in the generalized seniority calculations [ fig .
[ fig - ca - even - energy](c , d ) ] , in agreement with the full shell model calculations .
to more clearly compare the calculated ground state energy in the generalized seniority @xmath32 model space with that in the full model space , we consider the residual energy difference @xmath108 , obtained by subtracting the full space result from the generalized seniority result , shown in fig .
[ fig - ca - even - energy](e , f ) .
this difference may be considered as the missing correlation energy , not accounted for in the @xmath2-pair condensate description of the ground state . by the variational principle , the quantity @xmath108 must be nonnegative , since the @xmath32 space is a subspace of the full model space .
the residual vanishes where the @xmath32 and full calculations are equivalent , at @xmath82 and @xmath83 ( see sec .
[ sec - results - over ] ) . for the fpd6 interaction [ fig .
[ fig - ca - even - energy](e ) ] , the residual energy difference grows more or less smoothly towards mid - shell , where it is @xmath109 . for the gxpf1 interaction [ fig .
[ fig - ca - even - energy](f ) ] , the residual energy differences are generally smaller by a factor of @xmath110 , peaking at @xmath111 . for both interactions , there is a small ( @xmath112 ) dip in the residual at the @xmath77 subshell closure ( @xmath100 ) .
this is consistent , albeit not dramatically , with the hypothesis of ref .
@xcite , that the generalized seniority description should improve at subshell closures . for the energy eigenvalues of the ground state and other low - lying states ,
the deviations of the generalized seniority @xmath32 model space results from those obtained in the full space are summarized in table [ tab - ca - energy ] .
the values given are averages ( root - mean - square ) of the deviations across the full range of neutron numbers .
it may be noted that the deviations are consistently smaller for the gxpf1 interaction than for the fpd6 interaction .
rdddddddd @xmath17 & & & & & & & & + fpd6&0.31&0.48&0.62&0.88&0.45&0.31&0.30&0.48 + gxpf1&0.13&0.25&0.34&0.63&0.13&0.20&0.13&0.21 + excitation energies @xmath113 for the first @xmath114 and @xmath49 states , calculated relative to the @xmath28 ground state , are shown in fig .
[ fig - ca - even - ex ] .
although the deviations of @xmath115 noted above for the eigenvalues ( table [ tab - ca - energy ] ) are comparatively small on the @xmath116 scale of these eigenvalues [ fig .
[ fig - ca - even - energy](a , b ) ] , they are significant on the few-@xmath117 scale of the excitation energies . the broad features of the evolution of @xmath113 across the shell are reproduced within the @xmath32 model space .
for instance , for the @xmath114 state [ fig .
[ fig - ca - even - ex](a , b ) ] , spikes are obtained at the @xmath77 subshell closure ( @xmath100 ) and @xmath88 subshell closure ( @xmath118 for fpd6 or @xmath119@xmath120 for gxpf1 ) .
quantitatively , the excitation energy calculated for the @xmath114 state deviates from that calculated in the full model space by at most @xmath121 for fpd6 or @xmath122 for gxpf1 . for the @xmath123 state ,
the largest deviations obtained are @xmath124 for fpd6 or @xmath125 for gxpf1 . for both states ,
the excitation energies calculated in the @xmath32 model space are systematically higher than those calculated in the full model space , even though this direction for the deviation is not guaranteed by any variational principle .
returning to the @xmath28 states , the question arises as to whether or not the first _ excited _ @xmath28 state can be reasonably reproduced within the @xmath32 space . the calculated excitation energy for this state is shown in fig .
[ fig - ca - even - ex](e , f ) .
the general expectation @xcite is that @xmath126 or higher contributions should be important for an accurate description . the description of the excitation energy [ fig .
[ fig - ca - even - ex](e , f ) ] within the @xmath32 model space is qualitatively reasonable , but it is also quantitatively less accurate than for the yrast states ( see table [ tab - ca - energy ] ) . the calculation within the @xmath32 model space reproduces the main features of the @xmath104 dependence of the first excited @xmath28 energy : roughly constant @xmath127 for @xmath128 , followed by a drop to @xmath129 at @xmath118 for the fpd6 interaction [ fig .
[ fig - ca - even - ex](e ) ] , or a transient dip in the case of the gxpf1 interaction [ fig .
[ fig - ca - even - ex](f ) ] .
the largest deviation is @xmath130 .
however , from the occupations ( sec .
[ sec - results - occ ] ) , it will be seen that physically significant differences arise between the nature of the excited @xmath28 state obtained in the @xmath32 model space and in the full model space , in the lower part of the shell .
the excitation energy is , once again , systematically calculated higher in the generalized seniority model space than in the full model space . for the odd - mass @xmath0 isotopes ,
the level energies calculated in the generalized seniority @xmath35 and @xmath39 model spaces are shown in fig .
[ fig - ca - odd - energy ] .
let us begin by examining the energy eigenvalue for the lowest @xmath86 state [ fig .
[ fig - ca - odd - energy](a , b ) ] .
this is the ground state ( both calculated and experimental ) for @xmath131 , where nucleons in the @xmath77 subshell dominate the structure .
the @xmath35 calculation constitutes the most extreme approximation within the generalized seniority framework attempting to treat the lowest - energy state as a one - quasiparticle state based upon the @xmath77 orbital .
it is seen that the energy obtained in the @xmath35 calculation differs from that is the full space very noticeably ( several @xmath117 ) for a range of @xmath104 values above the @xmath77 subshell closure ( @xmath132 for fpd6 [ fig .
[ fig - ca - odd - energy](a ) ] or @xmath133 for fpd6 [ fig . [ fig - ca - odd - energy](b ) ] ) .
the energy obtained in the @xmath39 space , however , is indistinguishable from the result in the full space , on this scale .
we therefore again consider residual energy differences [ fig .
[ fig - ca - odd - energy](c , d ) ] .
similar comments apply to the energies of the lowest states of @xmath134 , @xmath52 , and @xmath53 , calculated in the @xmath35 and @xmath39 model spaces , for which the residuals are shown in fig .
[ fig - ca - odd - energy](e j ) , although the range of @xmath104 under which the @xmath35 approximation deviates most differs for the different states . for all states ,
the residual energy difference for the @xmath39 calculation identically vanishes at @xmath84 and @xmath85 , where the @xmath39 and full calculations are equivalent , and similarly for the @xmath35 calculation at @xmath80 and @xmath81 .
the differences between energies calculated in the @xmath39 space and the full shell model space are again typically smaller for the gxpf1 interaction than for the fpd6 interaction .
for the @xmath86 state , the residual reaches @xmath135 for the fpd6 interaction [ fig .
[ fig - ca - odd - energy](c ) ] but is never larger than @xmath136 for the gxpf1 interaction [ fig .
[ fig - ca - odd - energy](d ) ] .
given the sharp @xmath104 dependences observed in fig .
[ fig - ca - odd - energy ] , global averages of the deviations across the shell provide only a very crude measure of the level of agreement between @xmath39 and full space calculations .
nonetheless , the quantitative results are summarized in table [ tab - ca - energy ] .
the range of neutron numbers over which the @xmath35 approximation provides a reasonable reproduction of the full space results , for each different @xmath17 , can be roughly interpreted in terms of the natural filling order of the orbitals ( see also the discussion of occupations below in sec . [ sec - results - occ ] ) .
first , it should be noted that , although the @xmath35 state @xmath137 involves a superposition of different possible occupation numbers for the orbital @xmath19 , since the @xmath138 operator adds particles in _ pairs _ to each orbital , the state will involve contributions only with an _ odd _
occupation for the orbital @xmath19 .
the one - quasiparticle description therefore requires at least one _ particle _ in the given orbital @xmath19 , but also at least one _ hole _ in that orbital .
for instance , for the @xmath86 state to be one - quasiparticle in nature requires the presence of at least one particle but also one hole in the @xmath77 orbital .
this is naturally the situation early in the shell , but retaining a vacancy in the @xmath77 orbital becomes increasingly energetically penalized as the shell fills .
once more nucleons are present than can be accomodated in the @xmath77 orbital , the one - quasiparticle state is subject to the single - particle energy cost of promoting at least one nucleon out of the @xmath77 orbital , to leave an @xmath77 hole .
it would thus be natural to expect the one - quasiparticle state to lie several @xmath117 higher in energy than other configurations .
the residual energy difference between the one - quasiparticle state and the lowest @xmath86 state in the full space does indeed jump to several @xmath117 after the @xmath77 subshell closure [ fig .
[ fig - ca - odd - energy](d ) ] , but only at neutron numbers somewhat larger than @xmath100 .
this may be understood in terms of the impossibility of generating @xmath86 with nucleons purely in the next available orbital , @xmath88 .
the configurations competing with the one - quasiparticle state therefore are also subject to a single - particle energy penalty for promoting nucleons to yet higher orbitals ( see sec .
[ sec - results - occ ] for the orbitals actually involved ) .
similar interpretations may be given for the @xmath35 energies for the other @xmath17 values .
for the @xmath134 state to be one - quasiparticle in nature requires at least one particle in the @xmath89 orbital .
since the @xmath89 orbital is highest in the shell , configurations involving a particle in this orbital only become favored by single - particle energy considerations above @xmath139 . for both interactions ,
the @xmath35 energy residual is indeed lowest at the very end of the shell [ fig .
[ fig - ca - odd - energy](g , h ) ] , dropping below @xmath140 for @xmath141 . for the @xmath142 state , natural filling of
the @xmath88 orbital occurs midshell , for @xmath143 .
the @xmath35 energy residual for the @xmath142 state is lowest in this mid - shell region , although the reduction is not sharply confined to these particular neutron numbers [ fig .
[ fig - ca - odd - energy](k , l ) ] . for the @xmath144 state ,
the interpretation of the @xmath35 energy residual [ fig .
[ fig - ca - odd - energy](k , l ) ] is less clear .
natural filling would occur in a very limited region just above midshell ( @xmath145 ) . on the other hand , it is also relatively difficult to generate energetically favored shell model configurations with @xmath146 to compete with the one - quasiparticle configuration .
for instance , at the beginning of the shell , no pure @xmath147 configuration has @xmath144 , so any competing configuration in the full model space will likewise involve promoting at least one particle out of the @xmath77 subshell .
the occupations of the @xmath9-shell orbitals provide a simple , direct measure of the structure of the shell model eigenstates .
the occupations may also be considered as experimental observables , through their connection to spectroscopic factors .
rdddddddd @xmath17 & & & & & & & & + fpd6&0.03&0.09&0.11&0.29&0.06&0.04&0.05&0.09 + gxpf1&0.015&0.03&0.08&0.26&0.015&0.03&0.02&0.14 + while the basis states used in traditional shell model calculations have a definite number of nucleons in each orbital , by construction , the generalized seniority basis states do not .
indeed , the @xmath2-pair condensate has a bcs - like distribution of occupations for the different orbitals .
nonetheless , the orbital occupations for states obtained in the generalized seniority scheme can readily be calculated much as any other one - body observable ( sec .
[ sec - methods ] ) , as the expectation values @xmath148 .
the orbital occupations for the even - mass @xmath0 isotopes are shown in fig .
[ fig - ca - even - occ ] , for the same states as considered in sec .
[ sec - results - energy ] , calculated both in the generalized seniority @xmath32 model space and in the full shell model space . for the lowest @xmath28 , @xmath11 , and @xmath49 states [ fig .
[ fig - ca - even - occ](a f ) ] , the curves obtained in the @xmath32 and full spaces are nearly indistinguishable .
for the ground state in particular ( recall the @xmath32 ground state is just the @xmath2-pair condensate ) , the generalized seniority approximation reproduces the mean occupation to within @xmath149 nucleon for all the orbitals , across the entire shell .
the _ typical _ deviations in @xmath148 for the ground state are in fact even much smaller , @xmath150 for the fpd6 interaction or @xmath151 for the gxpf1 interaction .
the deviations are only modestly larger for the @xmath114 and @xmath49 states , as summarized in table [ tab - ca - occ ] .
qualitatively , the trend appears to be that the generalized seniority calculations smooth the evolution of the occupations as functions of @xmath104 , relative to the calculations in the full space .
such is observed if one examines the difference between the curves obtained in the @xmath32 and full spaces for the occupations of , for instance , the @xmath88 or @xmath89 orbitals at @xmath152 for the @xmath114 state , under fpd6 [ fig .
[ fig - ca - even - occ](c ) ] , or for the @xmath88 orbital at @xmath153 for the @xmath123 state , under either interaction [ fig .
[ fig - ca - even - occ](e , f ) ] . inspection of the occupations for the ground state and excited @xmath114 state below the @xmath77 subshell closure ( @xmath100 ) in fig .
[ fig - ca - even - occ](a d ) indicates that these are nearly pure @xmath147 configurations , with just trace occupations of the other @xmath9-shell orbitals . in the limit of a pure @xmath147 configuration
, the generalized seniority description reduces to the conventional seniority description in the @xmath77 shell , indeed , a classic domain for application of seniority @xcite .
that is , the generalized seniority @xmath24 state is just the conventional zero - seniority state , the generalized seniority @xmath32 space contains just the conventional seniority @xmath11 state of each @xmath17 , _ etc .
_ generalized seniority thus only becomes fully distinct from conventional seniority when multiple orbitals are simultaneously significantly occupied , above @xmath100 .
the calculated occupations for the first excited @xmath28 state are shown in fig .
[ fig - ca - even - occ](e , f ) . as already noted in sec .
[ sec - results - energy ] , the generalized seniority @xmath32 calculation is taking place in a very low - dimensional space .
the @xmath32 results are seen to track those obtained in the full space reasonably well across the shell , but with notable differences below the @xmath77 subshell closure ( @xmath100 ) .
in particular , these indicate differing structural interpretations for the excited state in the @xmath32 model space and the full space in the lower part of the shell . in the @xmath32 space for @xmath28 , the nucleons must couple to zero angular momentum _ pairwise _
( recall the @xmath101 basis states ) and thus occupy orbitals in even numbers .
therefore , a @xmath154 excited state is obtained at @xmath155 and @xmath156 [ see the @xmath88 curve for @xmath32 in fig .
[ fig - ca - even - occ](e , f ) ] , as the next most energetically favored configuration after the approximately-@xmath157 ground state .
however , the structure in the full model space is found to involve promotion of only a single nucleon to the @xmath88 orbital [ see the @xmath88 curve for the full space calculation in fig .
[ fig - ca - even - occ](e , f ) ] .
the remaining nucleons in the @xmath77 orbital must therefore be in a @xmath142 configuration , which has conventional seniority @xmath12 , so that total @xmath28 may be obtained .
however , above the @xmath77 subshell closure , agreement of the occupations in the @xmath32 and full spaces is much closer .
the occupations of the orbitals in the odd - mass @xmath0 isotopes are shown in fig .
[ fig - ca - odd - occ ] , as calculated in the generalized seniority @xmath39 model space and the full shell model space .
the lowest @xmath86 and @xmath51 states are taken for illustration .
the results obtained in the generalized seniority @xmath39 model space again closely track those found in the full model space .
the quantitative deviations are again summarized in table [ tab - ca - occ ] , and again those for the gxpf1 interaction are smaller than for the fpd6 interaction . as discussed in sec .
[ sec - results - energy ] , the @xmath35 one - quasiparticle description varies greatly with @xmath104 in its success at reproducing the energy of the lowest state of each @xmath17 , and this variation may qualitatively be interpreted in terms of the single particle energy cost of generating a quasiparticle in the relevant orbital .
the orbital occupations in fig .
[ fig - ca - odd - occ ] provide an immediate indication of whether or not a given state might be dominantly one - quasiparticle in nature , if we recall that the orbital containing the quasiparticle must have odd occupation , and therefore contain at least one particle but also one hole .
for instance , for the @xmath86 state , recall that the @xmath35 result provides a reasonable description of the energy only for @xmath158 for the fpd6 interaction [ fig .
[ fig - ca - odd - energy](c ) ] or for @xmath159 for the gxpf1 interaction [ fig . [ fig - ca - odd - energy](d ) ] .
examining the occupation of the @xmath77 orbital in the @xmath86 state [ fig .
[ fig - ca - odd - occ](a , b ) ] , it is seen that , indeed , the occupation is only consistent with an @xmath77 quasiparticle for @xmath158 for the fpd6 interaction [ fig .
[ fig - ca - odd - occ](a ) ] , after which the @xmath77 orbital completely fills , but that a hole remains in the @xmath77 orbital for @xmath159 for the gxpf1 interaction [ fig .
[ fig - ca - odd - occ](b ) ] . for the @xmath134 state [ fig .
[ fig - ca - odd - occ](c , d ) ] , the situation is less obvious .
the calculated occupations admit the _ possibility _ of one - quasiparticle structure for @xmath160 for fpd6 [ fig . [
fig - ca - odd - occ](c ) ] or @xmath161 for fpd6 [ fig .
[ fig - ca - odd - occ](c ) ] , since these are the ranges over which the @xmath89 orbital has an occupation of at least one , but recall that the @xmath35 calculation provides a reasonable description of the energy only for @xmath162 [ fig .
[ fig - ca - odd - energy](g , h ) ] .
one may also directly examine the occupations obtained in the @xmath35 calculations ( not shown in fig .
[ fig - ca - odd - occ ] ) , and they are found to track the results from the full space calculations very well over the ranges of @xmath104 just described and poorly outside of these ranges .
the matrix elements of electromagnetic transition operators ( which we will consider in spectroscopic terms , as electromagnetic moments or transition strengths ) probe the extent to which various correlations are preserved when the nuclear calculation is carried out in a space of restricted generalized seniority . practically , the accuracy with which these observables are reproduced is of prime interest if generalized seniority is to be used as a truncation scheme for shell model calculations .
matrix elements of the @xmath163 and @xmath164 operators directly follow from the one - body densities calculated in the generalized seniority scheme ( sec .
[ sec - methods ] ) much as in a conventional shell model calculation .
it should be noted that the fpd6 and gxpf1 interactions are defined only in terms of two - body matrix elements between orbitals labeled by quantum numbers @xmath165 , without reference to any specific form for the radial wave functions .
a particular choice must be made if electromagnetic transition observables are to be computed .
we adopt harmonic oscillator wave functions , with the blomqvist - molinari parametrization @xmath166
@xcite for the oscillator energy .
since only neutrons are present for the @xmath0 isotopes , the overall normalization of the @xmath163 matrix elements is set by the neutron effective charge , which is taken as @xmath167 . for the calculation of @xmath164 matrix elements ,
the free - space neutron @xmath168-factors are used . for the even - mass @xmath0 isotopes , we consider the electric quadrupole moment @xmath169 , electric quadrupole reduced transition probability @xmath170 , and magnetic dipole moment @xmath171 , as representative electromagnetic observables for the low - lying states .
the values are shown in fig .
[ fig - ca - even - trans ] , calculated both in the @xmath32 and full shell model spaces .
for the quadrupole moment [ fig .
[ fig - ca - even - trans](a , b ) ] , the calculation in the @xmath32 space qualitatively tracks the results for the full space , in particular , the alternations in sign as a function of neutron number . however , under the fpd6 interaction [ fig .
[ fig - ca - even - trans](a ) ] , the quadrupole moment obtained in the @xmath32 space is consistently much smaller in magnitude than in the full space , roughly by a factor of two .
the difference under the gxpf1 interaction [ fig .
[ fig - ca - even - trans](b ) ] is less marked , but the quantitative agreement is still crude .
the quadrupole moment obtained in the @xmath32 space is smaller by @xmath172 , that is the average deviation from the full - space result is @xmath173 ( taking root - mean - square averages to better accomodate signed quantities ) , on quadruople moments averaging @xmath174 .
this attenuation of the calculated quadrupole moment is perhaps not surprising , given that the generalized seniority scheme is expected to be restricted in its ability to reproduce quadrupole correlations @xcite .
however , for the @xmath175 strength [ fig .
[ fig - ca - even - trans](c , d ) ] , which is often taken as a proxy for the quadrupole deformation , the agreement between the values obtained in the @xmath32 model space and in the full shell model space is much closer than for @xmath169 . here
the deviations under fpd6 are @xmath176 [ averaging @xmath177 , on @xmath178 values averaging @xmath179 , or for gxpf1 only @xmath180 [ averaging @xmath181 , on @xmath178 values averaging @xmath182 .
thus , intriguingly , generalized seniority seems more capable of incorporating the correlations necessary for reproducing @xmath175 than for reproducing @xmath169 .
the magnetic dipole moment of the first @xmath114 state [ fig .
[ fig - ca - even - trans](e , f ) ] evolves in a complicated manner with neutron number , involving multiple reversals in sign , and these are well reproduced by the calculations in the generalized seniority @xmath32 model space .
( the most noticeable discrepancy arises for the gxpf1 interaction at @xmath183 [ fig .
[ fig - ca - even - trans](f ) ] , also the point of largest deviation in the @xmath4 excitation energy [ fig .
[ fig - ca - even - ex](b ) ] for this interaction , but unremarkable in terms of occupations [ fig .
[ fig - ca - even - occ](d ) ] . )
quantitatively , the deviations are comparable for both interactions , for fpd6 @xmath184 ( averaging @xmath185 , on moments averaging @xmath186 ) or for gxpf1 @xmath187 ( averaging @xmath188 , on moments averaging @xmath189 ) . for the odd - mass @xmath0 isotopes ,
let us consider the electromagnetic moments @xmath190 and @xmath191 of the first @xmath86 state .
the values are shown in fig .
[ fig - ca - odd - trans ] , calculated in the generalized seniority @xmath35 and @xmath39 model spaces and in the full shell model space .
for both these moments , the evolution calculated in the @xmath39 model space closely tracks that obtained in the full model space , with isolated discrepancies .
for the quadrupole moment [ fig .
[ fig - ca - odd - trans](a , b ) ] , the deviations for fpd6 are @xmath187 ( averaging @xmath192 , on moments averaging @xmath193 ) or for gxpf1 @xmath194 ( averaging @xmath195 , on moments averaging @xmath196 ) .
notice the much better agreement obtained for this quadrupole moment in the @xmath39 space than for @xmath169 in the @xmath32 space .
for the dipole moment [ fig .
[ fig - ca - odd - trans](c , d ) ] , the deviations for fpd6 are @xmath184 ( averaging @xmath195 , on moments averaging @xmath197 ) or for gxpf1 @xmath198 ( averaging @xmath199 , on moments averaging @xmath200 ) .
the quadrupole moment @xmath190 obtained in the one - quasiparticle ( @xmath35 ) description [ fig .
[ fig - ca - odd - trans](a , b ) ] can be understood in terms of the single particle value for the @xmath77 orbital and conventional seniority arguments .
for the one - quasiparticle state @xmath137 , only the orbital @xmath19 can contribute to the quadrupole moment ( simply by angular momentum selection ) , and this orbital carries a conventional seniority of @xmath79 from the unpaired particle .
conventional seniority in a @xmath201 configuration gives a simple linear variation of the quadrupole moment with @xmath202 across the @xmath76-shell , vanishing midshell ( the quadrupole operator is part of a rank-@xmath79 tensor with respect to quasispin @xcite ) .
this conventional seniority result can not be directly applied to the generalized seniority one - quasiparticle state , unless the occupation @xmath62 is approximately sharp in this state .
recall that this is indeed the case for the @xmath77 orbital , under the present interactions , where the occupations below @xmath100 are , to very good approximation , @xmath203 [ fig .
[ fig - ca - odd - occ](a , b ) ] .
thus , across the @xmath77 subshell , the quadrupole moment for the @xmath86 , @xmath35 state [ fig .
[ fig - ca - odd - trans](a , b ) ] varies linearly from the @xmath77 single - particle value to the @xmath77 single - hole value .
then , above the subshell closure , the @xmath35 state has , to very good approximation , @xmath204 , _
i.e. _ , exactly one hole .
thus , the @xmath35 quadrupole moment plateaus at the @xmath77 single - hole value . the range of neutron numbers over which the one - quasiparticle calculation provides a reasonable approximation to the quadrupole moment is @xmath205 for the fpd6 interaction or @xmath206 for the gxpf1 interaction , and again for @xmath207 .
this does not quite correspond to the ranges found in secs .
[ sec - results - energy ] and [ sec - results - occ ] from energies and occupations , _
i.e. _ , @xmath158 for the fpd6 interaction or @xmath159 for the gxpf1 interaction .
the dipole moment @xmath191 obtained in the one - quasiparticle description [ fig .
[ fig - ca - odd - trans](c , d ) ] is , more simply , constant ( the dipole operator is scalar with respect to quasispin @xcite ) and has the single - particle schmidt value .
the departure of the full space result from the schmidt value is modest over the range @xmath158 for the fpd6 interaction or @xmath159 for the gxpf1 interaction , and consists of a smooth linear evolution with @xmath104 ( increasing from @xmath208 to @xmath209 ) , before jumping abruptly at the ends of these ranges .
this may be interpreted as reflecting the same one - quasiparticle nature observed in the energies ( secs .
[ sec - results - energy ] ) and occupations ( sec . [ sec - results - occ ] ) over these ranges .
from the comparisons carried out in this work , it is found that calculations in a highly - truncated , low - dimensional ( table [ tab - dim ] ) generalized seniority model space , with just one broken pair , can reproduce energy , occupation , and electromagnetic observables for low - lying states with varying but in some cases remarkably high fidelity to the results obtained in the full shell model space .
these results were obtained for _ semimagic _ nuclei in the @xmath9 shell , under two different realistic interactions .
deviations in energies ( table [ tab - ca - energy ] ) vary from the @xmath210 range to the @xmath211 range for the states considered , which , while small compared to the @xmath212 binding energies , is nonnegligible for the evaluation of few-@xmath213 excitation energies .
nonetheless , the evolution of excitation energies is reasonably well - reproduced across the shell . for level occupations of the low - lying states ( table [ tab - ca - occ ] ) , accuracies in the few - percent range are obtained . with the notable exception of @xmath169 in the even - mass nuclei , electric quadrupole and magnetic dipole observables
are reproduced to @xmath214 or better ( sec .
[ sec - results - trans ] ) . given the distinct improvement of results from the @xmath35 space to the @xmath39 space ( for the odd - mass nuclei ) , the most natural extension is to generalized seniority spaces with two broken pairs ( @xmath126 for even - mass nuclei and @xmath215 for odd - mass nuclei ) . aside from the conceptual interest of generalized seniority as a means of interpreting shell model results in a bcs pair - condensate plus quasiparticle framework
, real computational benefits will be obtained if generalized seniority can also be successfully applied as an accurate truncation scheme for nuclei in the interior of the shell , when significant numbers of both valence protons and neutrons are present .
the obvious challenge is the seniority - nonconserving , or pair - breaking , nature of the proton - neutron interaction @xcite , since a pair broken in the conventional seniority scheme also implies breaking of a generalized seniority @xmath2 pair .
the approach is likely to be more advantageous for weakly - deformed nuclei ( in large model spaces ) than for strongly - deformed nuclei .
seniority decompositions of shell model calculations @xcite suggest seniorities @xmath216 should be sufficient for a variety of weakly - deformed nuclei .
such values are consistent with the possibility of successful calculation in a generalized seniority model space with two broken pairs for both protons and neutrons .
it was systematically observed that the calculations in the generalized seniority model space more accurately match those in the full model space for the gxpf1 interaction than for the fpd6 interaction , typically by a factor of @xmath110 .
it would be valuable to have a systematic quantitative understanding of the deviations expected for a given interaction , given some appropriate quantitative measures characterizing the interaction .
the question has been addressed in the context of _ random _ two - body interactions , in terms of the random ensemble parameters @xcite .
in particular , it appears to be important that the energy spacing scale of the single particle energies be large compared to the scale of the two - body matrix elements @xcite .
it might therefore be relevant that the spread of the single particle energies is indeed slightly larger for gxpf1 than for fpd6 .
alternatively , since the generalized seniority approach is based upon the dominance of pairing correlations , it is worth investigating the possibility that the decomposition of realistic interactions into pairing and non - pairing ( _ e.g. _ , quadrupole ) components through the use of spectral distribution theory , as carried out in ref .
@xcite , could yield relevant measures .
+ we thank f. iachello , p. van isacker , s. de baerdemacker , s. frauendorf , and a. volya for valuable discussions and b. a. brown and m. horoi for generous assistance with nushellx .
this work was supported by the research corporation for science advancement under a cottrell scholar award , by the us department of energy under grant no .
de - fg02 - 95er-40934 , and by a charg de recherche honorifique from the fonds de la recherche scientifique ( belgium ) .
computational resources were provided by the university of notre dame center for research computing .
in this appendix , a simple demonstration is provided to establish the property noted in sec .
[ sec - results - energy ] , that the @xmath28 ground state obtained in the @xmath32 model space is simply the @xmath24 condensate state , provided the @xmath21 coefficients have been chosen according to the variational prescription described in sec .
[ sec - methods ] .
it is convenient to first modify the normalization convention on the @xmath21 coefficients , from that given in sec .
[ sec - results - over ] , so as to instead give the state @xmath103 unit normalization .
then the energy functional in the variational prescription simplifies to @xmath217 and is subject to the constraint @xmath218 . since @xmath219 ,
the lagrange equations for the extremization problem are of the form @xmath220 , with a separate equation obtained for each orbital @xmath19 .
since the states @xmath101 span the @xmath32 space , this is simply the condition that @xmath221 within the @xmath32 space .
that is , @xmath103 is an eigenstate of the hamiltonian , and , in practice , it is the ground state .
note that this result also establishes the equivalence of the `` iterative diagonalization '' prescription for determining the @xmath21 coefficients , proposed in ref .
@xcite , to the variational prescription , since the iterative diagonalization prescription determines the @xmath21 coefficients so as to decouple the @xmath2 condensate from the rest of the @xmath32 space , exactly as found above for the variational prescription . | the generalized seniority scheme has long been proposed as a means of dramatically reducing the dimensionality of nuclear shell model calculations , when strong pairing correlations are present . however , systematic benchmark calculations , comparing results obtained in a model space truncated according to generalized seniority with those obtained in the full shell model space , are required to assess the viability of this scheme . here ,
a detailed comparison is carried out , for semimagic nuclei taken in a full major shell and with realistic interactions .
the even - mass and odd - mass @xmath0 isotopes are treated in the generalized seniority scheme , for generalized seniority @xmath1 .
results for level energies , orbital occupations , and electromagnetic observables are compared with those obtained in the full shell model space . |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.