Kinematics and stellar populations of low-luminosity early-type galaxies in the Abell 496 cluster

49 Pages·2008·14.72 MB·English

A&A486,85–97(2008) Astronomy
DOI:10.1051/0004-6361:20078709 &
(cid:2)c ESO2008 Astrophysics
Kinematics and stellar populations of low-luminosity early-type
(cid:2),(cid:2)(cid:2),(cid:2)(cid:2)(cid:2)
galaxies in the Abell 496 cluster
I.V.Chilingarian1,2,V.Cayatte3,F.Durret4,C.Adami5,C.Balkowski6,L.Chemin6,T.F.Laganá7,andP.Prugniel8,6
1 ObservatoiredeParis-Meudon,LERMA,UMR8112,61Av.del’Observatoire,75014Paris,France
e-mail:[email protected]
2 SternbergAstronomicalInstitute,MoscowStateUniversity,13Universitetskiprospect,119992Moscow,Russia
3 ObservatoiredeParis-Meudon,LUTH,UMR8102,5pl.JulesJanssen,92195Meudon,France
4 Institutd’AstrophysiquedeParis,CNRS,UMR7095,UniversitéPierreetMarieCurie,98bisBdArago,75014Paris,France
5 LAM,Pôledel’EtoileSitedeChâteau-Gombert,38rueFrédéricJoliot-Curie,13388MarseilleCedex13,France
6 ObservatoiredeParis-Meudon,GEPI,UMR8111,5pl.JulesJanssen,92195Meudon,France
7 InstitutodeAstronomia,GeofísicaeC.Atmosf./USP,R.doMatão1226,05508-090SãoPaulo/SP,Brazil
8 Université de Lyon, 69000 Lyon; Université Lyon 1, 69622 Villeurbanne; Centre de Recherche Astronomique de Lyon,
ObservatoiredeLyon,9Av.CharlesAndré,69561St.GenisLaval;CNRS,UMR5574,ÉcoleNormaleSupérieuredeLyon,Lyon,
France
Received19September2007/Accepted12March2008
ABSTRACT
Context.Themorphology and stellar populations of low-luminosity early-typegalaxies inclustershaveuntil now been limitedto
afewrelativelynearbyclusterssuchasVirgoorFornax.Scenariifortheformationandevolutionofdwarfgalaxiesinclustersare
thereforenotwellconstrained.
Aims.Weinvestigateherethemorphologyandstellarpopulationsoflow-luminositygalaxiesintherelaxedrichnessclass1cluster
Abell496(z=0.0330).
Methods.DeepmultibandimagingobtainedwiththeCFHTMegacamallowedustoselectasampleoffaintgalaxies,definedhere
as objects with magnitudes 18 < r(cid:3) < 22 mag withina 1.2 arcsec fibre (−18.8 < M < −15.1mag). We observed 118 galaxies
B
spectroscopicallywiththeESOVLTFLAMES/GiraffespectrographwitharesolvingpowerR=6300.Wepresentstructuralanalysis
andcolourmapsforthe48galaxiesbelongingtothecluster.Wefitthespectraof46objectswithPEGASE.HRsyntheticspectrato
estimatetheages,metallicities,andvelocitydispersions.Weestimatedpossiblebiasesbysimilarlyanalysingspectraof∼1200early-
typegalaxiesfromtheSloanDigitalSkySurveyDataRelease6(SDSSDR6).Wecomputedvaluesofα/Feabundanceratiosfrom
themeasurementsofLickindices.Webrieflydiscusseffectsofthefixedaperturesizeonthemeasurements.
Results.Forthefirsttime,high-precision estimatesofstellarpopulation properties havebeen obtained foralargesampleof faint
galaxiesinacluster,allowingfortheextensionofrelationsbetweenstellarpopulationsandinternalkinematicstothelow-velocity
dispersionregime.Wehaverevealedapeculiarpopulationofellipticalgalaxiesinthecoreofthecluster,resemblingmassiveearly-
typegalaxiesbytheirstellarpopulationpropertiesandvelocitydispersions,buthavingluminositiesofabout2magfainter.
Conclusions. External mechanisms of gasremoval (rampressure stripping and gravitational harassment) aremore likelytohave
occurredthaninternalmechanismssuchassupernova-drivenwinds.Theviolenttidalstrippingofintermediate-luminosity,early-type
galaxiesintheclustercorecanexplainthepropertiesofthepeculiarellipticalgalaxiessurroundingthecDgalaxy.
Keywords.galaxies:ellipticalandlenticular,cD–galaxies:dwarf–galaxies:clusters:individual:Abell496–
galaxies:kinematicsanddynamics–galaxies:stellarcontent–galaxies:fundamentalparameters
1. Introduction
Understandingtheformationandevolutionofgalaxiesisoneof
themostchallengingtasksinmodernastrophysics,andsubstan-
(cid:3) Based on observations obtained with ESO VLT (program 074.A- tial progresshasbeen achievedin characterisingthe evolution-
0533) and MegaPrime/MegaCam, a joint project of Canada France arypatternofearly-typegalaxies.Massiveellipticals,whichare
Hawaii Telescope(CFHT)and CEA/DAPNIA,at theCanada-France-
found principally in clusters, are known to be already present
HawaiiTelescope(program03BF12),whichisoperatedbytheNational at z (cid:5) 1 (Ziegler 2000, and references therein). Recent large
ResearchCouncil(NRC)ofCanada,theInstitutNationaldesSciences
surveys give strong support to: (1) the downsizing star forma-
de l’Univers of the Centre National de la Recherche Scientifique
tionconcept(thestarformationactivityisseentoprogresswith
(CNRS)ofFrance,andtheUniversityofHawaii.
(cid:3)(cid:3) All the spectral and imaging data used in this paper are avail- time, from high mass galaxies to smaller ones (as first sug-
ablethroughtheVO-ParisDataCentre:http://voplus.obspm.fr/ gested by Matteucci 1994 from variationsof the [Mg/Fe] ratio
DataCollections/Abell0496/ withgalaxyluminosityinellipticals,thenbyCowieetal.1996);
(cid:3)(cid:3)(cid:3) Appendicesareonlyavailableinelectronicformat and (2) the top-down formation where mass assembly occurs
http://www.aanda.org at lower redshifts for lower galaxy masses (Bundy et al. 2006;
ArticlepublishedbyEDPSciences
86 I.V.Chilingarianetal.:Abell496:Kinematicsandstellarpopulations
Cimattietal.2006).However,ifaconsensusexistsforstarfor- galaxy of an intermediate mass lenticular (Chilingarian et al.
mation timescales and chemical evolution, it is not clear even 2007c).
formassiveellipticalsifthemassassemblytimecorrespondsto Abell 496 is a richness class 1 cluster (Abell 1958) of
the star formation episode duration (Bell et al. 2006; Pozzetti cD type (Struble & Rood 1987) at a heliocentric velocity of
et al. 2007; Scarlata et al. 2007). In addition, multiple merg- 9885 km s−1 (z = 0.0330, Durret et al. 2000, and references
ers of smaller galaxies are not the only formation mechanism, therein), or 9707 km s−1 (z = 0.0324) after correcting for
corr
alsoleavingroomforarapidcollapseofgasintheremotepast. infall of the Local Group toward Virgo. For a Hubble constant
Moreover,galaxyevolutioninclustersisexpectedtobedifferent H = 73 km s−1Mpc−1, the distance modulus is 35.70 (d =
0
fromthatinlowerdensityenvironments.Thelocalmorphology- 133 Mpc) and the corresponding scale is 0.627 kpc arcsec−1.
density relation revealed by Dressler (1980), with a very high Abell 496 is a cluster with several hundred measured galaxy
fraction of early-type galaxies found in cluster cores, has been redshifts (Durret et al. 1999). The analysis of the distribution
extended to a larger range of galaxy density from the SDSS of 466 redshifts in the direction of this cluster has revealed
(Goto et al. 2003) and to higher redshift (Capak et al. 2007). the existenceof severalstructuresalong the line of sight;how-
Boththelatterstudyandthedisc-fadingmodeldiscussionforS0 ever, the redshift distribution of the 274 galaxies found to be-
galaxyformationbyChristlein&Zabludoff(2004)indicatethat long to the cluster itself implied that Abell 496 has a regular
the morphologyand star formationcould be affected by differ- morphology and a well relaxed structure (Durret et al. 2000).
entprocesses:theincreaseinearly-typefractionismostlydriven This is confirmedby X-ray data:the X-ray map obtainedfrom
by galaxy interactions and harassment (Moore et al. 1998) or XMM-Newtonobservationsisindeedquiteregular,contraryto
by tidal effects induced by the crossing of the cluster potential most clusters where even if the X-ray emissivity map appears
well (including dynamical friction); the reduction or suppres- regular,thetemperaturemapofthehotgasdoesnot(e.g.,Durret
sion of the star formation is caused by ram pressure gas strip- etal.2005).
ping(Gunn&Gott1972;Abadietal.1999)orbystrangulation, We will presenthere resultsfor 48low-luminositygalaxies
whichpreventsfurthergasaccretionbycuttingofftheouterneu- inAbell496.Theobservationsanddatareductionaredescribed
tralgasreservoir(Larsonetal.1980).Inconclusion,thebulges in Sect.2.We givethephotometricpropertiesoftheAbell496
of lenticularsin clustersare suspected to be the results of tidal sample in Sect. 3 and a description of the spectral fitting ap-
interactions. pliedtorecoverthestellarpopulationandkinematicalproperties
Suchscenariiof morphologicaltransformationsofinfalling inSect.4.Wepresentresultsfromthephotometricandspectro-
galaxiesarealsoinvokedforthetransformationofdwarfirregu- scopicanalysesinSect.5,anddiscussourconclusionsinSect.6.
larsorfaintlate-typespiralsintoearly-typedwarfgalaxiescon-
stitutingthemostnumerousclassofgalaxiesinnearbyclusters.
However, until now both the harassment model and the classi-
2. Observationsanddatareduction
calwindmodelinavirializedprotogalaxyfailtoreproducethe
observedpositionsofdwarfellipticalsinthefundamentalplane 2.1.Imagingobservationsandsampleselection
(Djorgovski & Davis 1987) which links the internal kinemat-
ics and the structural properties (De Rijcke et al. 2005). Geha We obtainedimageswiththe CanadaFranceHawaiiTelescope
et al. (2003) and Van Zee et al. (2004) faced similar difficul- withtheMegacamcamerainthefallof2003(program03BF12,
ties trying to explain the intermediate ages and slightly subso- P.I. V. Cayatte). Megacamcoversa field of 1◦ ×1◦ on the sky,
lar to solar metallicities found for dwarf ellipticals (dEs) from withapixelsizeof0.187×0.187arcsec2.Weobtaineddeepim-
the analysis of Lick indices:no evolutionaryscenario couldbe agesin the u∗, g(cid:3), r(cid:3), andi(cid:3) filters. We reducedthese imagesin
ruledoutorconfirmed.Tocomplicatethesituation,Liskeretal. the usual way (bias and flat field corrections, photometric,and
(2007) have shown that early-type dwarfs including ellipticals astrometriccalibrations)bythestaffoftheTerapixdatacenterat
and dwarf lenticulars(dS0s) do not forma homogeneousclass Institutd’AstrophysiquedeParis(IAP),France.TheSExtractor
of galaxy. At least five subclasses with different morphologi- software was runonthe r(cid:3) image(the imagewith the bestsee-
cal andclusteringpropertiesarefoundin the Virgocluster:the ing) to detect objects and measure their positions and magni-
dE(di)s displaying disc-like features (Lisker et al. 2006a); the tudes.Inparticular,wemeasuredmagnitudeswithina1.2arcsec
dE(bc)s showing a blue centre with recentor ongoingstar for- diameter,toprepareFLAMES/Giraffeobservations(seebelow).
mation(Liskeretal.2006b);thebrightnucleateddwarfsdE(N)s Details on the data reduction of these images can be found in
form an unrelaxed population of disc-shaped dwarfs, which is Bouéetal.(2008).
suspectedtobetheresultoftransformationsofinfallingprogen- We thendiscardedstarsbasedonadiagramofaperturemi-
itors; faint nucleateddwarfs; and all non-nucleateddEs form a nus totalmagnitudeversustotalmagnitudefor r(cid:3) < 21.Above
relaxedpopulationofclassicalspheroidalobjectsthatformedor thismagnitude,wekeptallobjectsinourgalaxysample.Apho-
arrivedintheclusteralongtimeago. tometric redshift code was kindly applied by Ilbert to our cat-
Inordertoclarifytheevolutionarypathofintermediateand aloguein anattempttoeliminate backgroundgalaxies.Finally,
low-mass, early-type galaxies in clusters we have studied the the galaxies observed with FLAMES/Giraffe were taken from
photometric, stellar population and kinematical properties of a this imaging catalogue, with a magnitude within a diameter of
sample of dwarf galaxies in the nearby cluster Abell 496. Our 1.2arcsecinther(cid:3)band(r(cid:3) )inthe[18−22]interval.
1.2
goalistopointoutobjectsthatareclearlytheresultsofthedif- Wetooktheobservedgalaxiesfromthecataloguedescribed
ferentproposedscenariitodefinewhichpropertiesarediscrimi- above, with the following priorities: top priority, objects with
nantfortheformationandevolutionoffaintearly-typegalaxies. 18.0<r(cid:3) <20.75;middlepriority,objectswith20.75<r(cid:3) <
Itismorepromisingtoidentifyandselectthemostprobablecan- 21.5;and1.l2owpriority,objectswith21.5<r(cid:3) <22.Wethus1.o2b-
1.2
didates rather than performing statistical studies on the whole tained118galaxyspectra(somefibershadtobeusedforguide
class of intermediate and low-mass early-type objects. A first stars and sky spectra). A fragment of the r(cid:3)-band image with
stephasbeenthediscoveryofanewcompactellipticalwhichis the galaxies observed spectroscopically with FLAMES/Giraffe
theresultoftidalstrippingbythecentralclusterDominant(cD) isshowninFig.1.
I.V.Chilingarianetal.:Abell496:Kinematicsandstellarpopulations 87
Fig.1. A fragment of the CFHT Megacam r(cid:3)-band image, showing the central part of Abell 496. Galaxies from the spectroscopic sample are
displayedwithdifferentsymbols:(1)confirmedclustermembersasdiamondswiththecolourindicatingthepresenceofembeddedstructures:red,
green, andblueforno, weak,andstrongembedded structuresrespectively; (2)confirmedfieldgalaxiesassquares; (3)galaxieswithunknown
redshiftsastriangles.TheX-raymapobtainedwithXMM-Newtonisshownascontours.
2.2.Spectroscopicobservationsandreduction the Python version of BLDRS – Baseline Data Reduction
Software (girbldrs-1.12) available from http://girbldrs.
We obtained spectra with the ESO Very Large Telescope us- sourceforge.netandwithfunctionsandrecipesdescribedin
ingtheFLAMES/GiraffeinstrumentintheL543.1configuration the BLDRS Software Reference Manual, Doc. No. VLT-SPE-
onthetwonightsof8−10/12/2004.TheFLAMES/Giraffefield OGL-13730-0040(Issue1.12,20/9/2004).Wedidthereduction
of view is 20 arcmin in diameter, with a total numberof fibres with the Image Reduction and Analysis Facility (IRAF). The
of130;eachfiberhasacircularapertureontheskyof1.2arcsec processingincludesbiassubtraction;diffuselightestimation;re-
indiameter.Weusedthe600lines/mmgratingintheLR4setup, moval,localisation,andextractionofthespectra;correctionfor
givinga resolvingpowerof aboutR = 6300in the wavelength fibertransmissionvariations;wavelengthcalibration;divisionby
range5010−5831Å. the continuumlamp spectrum;and sky subtraction.The result-
Weobtainedfourexposuresonthefirstnightfor2700,3300, ingspectraarenotcalibratedinabsolutefluxes.Weusetheflux
2351, and 1699 s. The second night four other exposureshave uncertaintiesprovidedbytheBLDRSforthedataanalysis.
been acquired with the same positioner configuration file and We combinedindividual1Dspectrawiththe RSI IDLsoft-
effective durations of 2700, 3300× 2 and 4200 s. During the wareandwemeasuredredshiftswiththervsao.xcsaopackagein
day, exposures of bias, flat fields, and arc line lamps for the IRAF,usingvariousstellartemplates.Redshiftswerealsomea-
wavelength calibration have been done in the same setup and suredthroughthestellarpopulationsynthesisfitdescribedinthis
with the two separate sets of MEDUSA fibers. The description work,andbothvaluesagreedwithintheiruncertainties.
of the FLAMES/Giraffe instrument can be found in Pasquini We had a very good result for our star-galaxy separation
et al. (2002). We extracted and calibrated the spectra using since we did not observe any stars spectroscopically. On the
88 I.V.Chilingarianetal.:Abell496:Kinematicsandstellarpopulations
other hand, the rejection of backgroundobjects based on pho- explained by the fiber sizes (1.2 arcsec) which are larger than
tometric redshifts was not very efficient, since only 48 out of a normal spectrograph slit width (diffraction limit of the colli-
102 galaxies with measurable redshifts actually belong to the mator)resultinginaΠ-shapedLSF.Thehighspectralresolution
cluster. Of those 48, 46 have sufficientsignal-to-noiseratiosto of FLAMES/Giraffe allows us to measure velocity dispersions
analyse their kinematic andstellar populations.For the two re- as low as 10 km s−1 for data having a signal-to-noise ratio of
mainingobjectswepresentonlythephotometricanalysis. about10perpixel.
Absolute magnitudes were computed using the distance Since our techniquefor extractionof stellar populationand
modulusmentionedabove.Allmagnitudesconsideredthrough- kinematicalparametersisbasedonanon-linear,least-squarefit-
out the paper are corrected for intergalactic extinction accord- tingonmanyparameters,thereisalwaysachancethatthemin-
ing to Schlegel et al. (1998). We corrected for cosmological imisation procedure does not reach the absolute minimum in
dimming,andappliedthe K-correctionandconversionintothe theχ2 space.Inadditionthereareseveraldegeneraciesbetween
B-band(ifneeded)assuminganellipticalgalaxySEDandtrans- theparameters,themostimportantbeingthe(1)age-metallicity;
formationsfromFukugitaetal.(1995). and(2)metallicity-velocitydispersionparameters.Forthisrea-
son,theshapeoftheminimumintheparameterspacebecomes
stronglyextendedalongthelinescorrespondingtothosedegen-
3. Spectralfitting eracies, sometimes exhibiting several local minima, where the
minimisation algorithm can be trapped. To check how critical
To deduce kinematical and stellar population parameters, we
theproblemisinourcase,webuiltχ2maps.
have used direct fitting of the PEGASE.HR (Le Borgne
WeproceededasinChilingarianetal.(2007a)–AppendixA
et al. 2004) synthetic spectra to observed data in pixel space
– fitting only the kinematics and a multiplicative polynomial
(van der Marel & Franx 1993; Cappellari & Emsellem 2004).
continuum for a set of fixed values of ages (t) and metallici-
At the same time, the minimisation procedure returns the pa-
ties (Z)of thetemplates;thusforeverypair ofvalues(t,Z)the
rametersofthepopulationmodel(ageandmetallicityofSimple
bestfittingkinematicalparametersareobtained.Finally,weob-
StellarPopulations,hereafterSSPs)andoftheinternalkinemat-
tained a mapof minimalχ2 valuesin the age-metallicityspace
ics(GaussianLOSVD).
foreach spectrum.Thesemapsallow usto detectpossiblesys-
DetailsofthemethodaregiveninChilingarianetal.(2005,
temic errors on the stellar population parameters derived from
2007d) and Prugniel et al. (2005), and its stability and biases
thenon-linearfitting.
are described in Chilingarian et al. (2007a). In the present pa-
For the 46 spectra the solutions of the fitting coincide with
perwefittheobservationswithsingleSSPscomputedwiththe
theminimaseenonthemaps,therefore,ourminimisationstrat-
Salpeter (1955) IMF. This gives us the SSP-equivalent stellar
egy can be considered as reliable (see figures in Appendix A).
populationparametersthatwe willreferto throughoutthetext. The reduced χ2 values for the spectra (≈0.45), where no tem-
We use a 15th order multiplicative polynomial continuum in platemismatchduetonon-solar[Mg/Fe]ratiosisseen,suggest
the fitting procedure and no additive continuum, as discussed
thatthefluxuncertaintiesprovidedbytheBLDRSareoveresti-
in Appendix B, together with the effects of non-solar [Mg/Fe] matedby∼50percent.
abundanceratiosonthestellarpopulationparameters.
Toobtainreliableandpreciseuncertaintiesofthestellarpop-
ulation parameters, the computations are done in th√e rotated 4. Results
coordinate system√defined as η = (3Z + 2log10t)/ 13;θ = 4.1.Photometricandmorphologicalproperties
(−2Z+3log t)/ 13,wheretheηaxisisparalleltothedirec-
10
tion oftheage-metallicitydegeneracyforintermediate-ageand Weprovideintegratedphotometricparametersandcolourmaps
oldstellarpopulationsnoticedbyWorthey(1994). forthe48galaxiesidentifiedasdefinitivemembersoftheclus-
Since the spectral resolution of FLAMES/Giraffe in the ter.Surfacephotometry,profiledecompositionandfundamental
MEDUSA mode is rather high (R = 6300), the PEGASE.HR relationswillbediscussedinaforthcomingpaper.Aspreviously
models, based on the high-resolution(R = 10000) ELODIE.3 discussed, the galaxies selected inside the virial radius of a re-
empirical stellar library (Prugniel & Soubiran 2001, 2004), re- laxedclusterandinitsredshiftrangehavealargeprobabilityof
maintheonlychoiceifonetriesto:(1)avoidadegradationofthe beingearly-typegalaxies,i.e.,ellipticalsorlenticulars.
spectralresolutionoftheobservedspectra;and(2)useanempir- From the Megacam images, we perform a simplified mor-
icalstellarlibrary.Inordertoacquireunbiasedestimatesofthe phologicalclassificationbyvisualinspection,determiningifthe
velocitydispersions,oneneedstotakeintoaccountvariationsof galaxyisanelliptical,S0,Sa,orlate-typespiral.Morphological
thespectrographline-spread-function(LSF)andtobroadenthe typesarealwaysevaluatedwithascatterofaboutonetypegiven
template spectraaccordingto the LSFshape(strictly speaking, theuncertaintyonthismeasurement,so,insomecases,wekept
accordingtothedifferencebetweentheLSFofthespectrograph two possible types. Thisclassification is reportedin Table 1 of
used to obtain the spectra being analysed and that of the stel- Appendix A. We have foundonly two galaxiesto be of a later
larlibraryusedforspectralsynthesispurposes).Toachievethis, type than Sa and one has been classified as SBa. The 45 other
wefitthetwilightspectraobtainedwithFLAMES/Giraffeinthe galaxies can be considered as real early-type galaxies. Even if
samesetupastheAbell496galaxieswithsolarspectraavailable our sample is not completely representative of the whole pop-
intheELODIE.3library,whichobviouslyhaveexactlythesame ulation of the cluster because some observational constraints
intrinsic LSF as the stars used for the spectral synthesis. The (avoidance of bright objects and some limitations due to the
instrumentalresponseofFLAMES/Giraffeappearedtobe very FLAMES/Giraffe positioner) have been set, the morphological
stable across the fibers. The instrumentalwidth (σinst) changes segregationisobservedasexpectedinaclustercore.
smoothlyfrom19kms−1at5000Åto15kms−1 at5800Å,H3 Since the range in absolute B magnitude [−18.8,−15.0]
remains stable at about −0.01, and H4 at about −0.07 (see the covers the limit between intermediate luminosity and dwarf
definition of Gauss-Hermite parametrization in van der Marel galaxies, we split the sample into three different subsets:
& Franx 1993). Slightly negative values of H4 are trivially the brightest representatives corresponding to the intermediate
I.V.Chilingarianetal.:Abell496:Kinematicsandstellarpopulations 89
luminosity galaxies of Bender et al. (1992) with M < −18.0
B
(among our full sample ten galaxies belong to this subset), a
transition subsample with M between −18.0 and −17.5 mag
B
(theclassificationisgivenasdS0/S0ordE/Edependingwhether
thegalaxyisfoundtobelenticularorelliptical;thissubsetcon-
tains eight galaxies), and finally, the real, early-type dwarf set
with M >−17.5(28objects,i.e.,57percentofthewholesam-
B
ple). Among all three subsets, the galaxiesare at differentpro-
jecteddistancesfromthecD(takenasthecentreofthecluster).
Themorphologicalclassificationseemspertinentforthebright-
est objects,butwe willnotdiscussthe separationbetweendS0
and dE, as suggested by Lisker et al. (2007) who assigned a
commonabbreviation“dE”tothisratherheterogeneousclassof
faintgalaxies.Theyconcludedtheirstudybyevidencingvarious
subclasses of dEs, and we will examine our sample in such a
context.
We have applied an elliptically-smoothedunsharp masking
technique(e.g.Liskeret al.2006a) to the CFHT/Megacamim-
ages of our galaxies to search for embedded structures. Using
differentsmoothingradii(semimajor-axesofellipses)from0.3
to4arcsec,weclassifiedalltheobjectsintothreecategories:no,
weak,andstrongembeddedstructures.Theyincludedbar,disc,
spiral arms, and ring types (see Table A.1). Nine and five ob-
jectswerefoundtohavestrongandweakembeddedstructures,
respectively. We stress that only one galaxy with M > −17.5
B
showsstrongembeddedstructuresandthreeotherfaint−17.5< Fig.2.Absolute Bmagnitude(toppanel)andeffectiveradius(bottom
MB <−16.5objectshaveweakones;ontheotherhand,brighter panel,cf.Kormendyrelation)asafunctionofmeansurfacebrightness
objectsoftenexhibitstrongandcomplexstructures,notobserved within the effective radius. Giant and intermediate-luminosity ellipti-
infainterearly-typedwarfs. cals, power-law and core galaxies are shown in blue, bulges of disc
Among the nine objects with strong embedded structures galaxiesinred,dwarfellipticalsandlenticularsinblack.Wekeepthe
originalmorphologicalclassification(E/dE)forthedatapointscoming
twosetsofthreegalaxiesarefoundrelativelyneareachotherin
projectionbutwith verydifferentradialvelocities.The first set fromtheVirgoClusterACSSurvey,andfromCaonetal.(1993),and
D’Onofrioetal.(1994).Therefore,theyappearbothinblueandblack.
(ACO496J043308.85-130235.6 (G-02), ACO496J043320.35-
Abell496galaxiesareshownasfilledbluecircleswithcrosses.
130314.9(G-06) and ACO496J043321.37-130416.6(G-07))is
located at the northern limit of the X-ray halo (see Fig. 12
of Tanaka et al. 2006); the second set (ACO496J043326.49- colourmapswithsuchaproperty;allofthemareratherflattened
131717.8 (G-13), ACO496J043331.48-131654.6 (G-15) and galaxiescomparedtotheotherdwarfsandcouldperfectlycorre-
ACO496J043333.53-131852.6(G-18)) is in the south near the spondtotheflattenedspheroidinvokedbyLiskeretal.(2006b)
cluster centre, in a region where the X-ray surface brightness toexplaintheirsubclassofdEswithbluecores.Alastsubclass
is seen in excess compared to its azimuthally averaged value defined in the Virgo cluster is obtained by separating the “fea-
and corresponding to a cold front. However, we interpret the tureless” dE class into nucleated and non-nucleated galaxies.
set of three galaxies projected on the cluster core as three ob- Since Abell 496 is more distant than Virgo, the identification
jects only seen close in projection (see notes for each galaxy of a nucleusis more difficultto define with the same precision
in AppendixA). In bothcases, the existenceof a realgroupof in termsof flux and size. From unsharpmasks we can identify
galaxieshastobetested. the galaxies where strong gradients are observed in the central
Concerningthegalaxieswithweakembeddedstructures,one regions; nine dwarf galaxies with M > −18.0 mag present
B
of them is the brightest object found among a group of four brightcompactcomponentsintheircentres,oneofthemhaving
galaxiesinourspectroscopicsample,whichareallinthesame a brightblue core. Compared to the Virgo dwarf study, the es-
regionoftheskyabout400kpcnorthofthecD.Asdiscussedin timatednumberofnucleatedobjectsmissedcouldbelargerbe-
thenotesofAppendixA,ACO496J043348.59-130558.3(G-32), causewecannotbesurethatsmallsizenucleiarenotsmoothed
ACO496J043343.04-130514.1 (G-28), ACO496J043345.67- bytheseeingeffects.Inthefollowingsubsectionwewilldiscuss
130542.2 (G-30) and ACO496J043349.08-130520.5 (G-33) theresultsofthespectroscopicdataintermsofageofthecentral
probablybelongto a realgroupsincetheir radialvelocitiesare stellarpopulation;anothersubclassificationwillbeproposedto
closetoeachother. providetestsforformationandevolutionscenarii.
Finally,eightdwarfgalaxieswith M > −18.0magexhibit Figure 2 presents the updated versions of Figs. 9a,g from
B
spiralarms,bars,rings,oredge-ondiscssimilartothestructural Graham&Guzmán(2003):absolute Bmagnitudeandeffective
propertiesdefinedbyLiskeretal.(2006a)inthedwarfpopula- radiusR versusmean Bsurfacebrightnesswithintheeffective
e
tionoftheVirgocluster,correspondingto24percentforatotal radius(cid:10)μ(cid:11) ,withtheAbell496galaxiessuperimposed.Theplot
e
numberof34dEsinthespectroscopicsampleofAbell496;this contains literature data only for elliptical galaxies and bulges
valueislargerthanthepercentagegivenbyLiskeretal.(2007) of lenticulars/spirals (where the bulge/disc decomposition has
fortheirdE(di)subclass,butitcanbeexplainedbyabiastoward beenmadeintheoriginalpapers).Allintegratedmeasurements
brightgalaxiesin our sample. Anothersubclass pointed outby for spiral and lenticular galaxies have been excluded. Data in
Liskeretal.isthatofdEswithbluecentreswhichexhibitrecent computer-readableformatfordE andE galaxiesfromBinggeli
orongoingcentralstarformation.Threeobjectsareseenonthe & Jerjen (1998), Caon et al. (1993), D’Onofrio et al. (1994),
90 I.V.Chilingarianetal.:Abell496:Kinematicsandstellarpopulations
Faber et al. (1997), Graham & Guzmán (2003), Stiavelli et al.
(2001) and homogenization algorithms for these datasets have
been kindly provided by Graham. We also included photomet-
ricparametersofEanddE/dS0galaxiesfromtheVirgoCluster
ACS Survey(Ferrareseet al. 2006);photometricdata ongiant,
intermediateellipticalgalaxiesandbulgesofspiralsandlentic-
ulars from Bender et al. (1992); photometricparametersof the
SérsiccomponentofM32(Graham2002);anddatafor430el-
liptical galaxies from the HyperLeda1 database (Paturel et al.
2003), with radial velocities below 10000 km s−1 and brighter
thanM =−18.0mag.
B
Thewell-knownstructuraldichotomybetweendiffusedwarf
galaxiesandclassicalellipticalsandbulgesisclearlyvisibleon
both plots in Fig. 2. A sequence of classical elliptical galaxies
startswiththemostluminousclustergalaxiesatfaintersurface
brightnesses and ends with a few compact elliptical galaxies Fig.3. Faber-Jackson relation for ellipticalsand bulges of disc galax-
(M 32-like objects) at high surface brightnesses. At the same ies. Abell 496 objects are shown as blue-filled circles. Giant and
time,diffusedwarfellipticalgalaxiesformaseparatesequence. intermediate-luminosity ellipticalsareinblue, bulgesof discgalaxies
inred,dwarfellipticalsandlenticularsinblack.Theupper endofthe
We notice that the counter-argumentsagaint this interpretation
arrowrepresentingM32correspondstotheHSTSTISmeasurements
ofthestructuraldiagramsexist(e.g.Graham&Guzmán2003).
(Josephetal.2001),andthefilledcirclerepresentsthevalueobtained
It is remarkable that over a large span of absolute magnitudes
fromearlierHSTFOSdata(vanderMareletal.1998),whichwasin
(−18 < M < −13 mag) diffuse galaxies show no correlation
B agreementwithmorerecentground-basedobservations.
betweeneffectivesurfacebrightnessandeffectiveradius:reff al-
ways remains between 0.6 and 2.0 kpc (see bottom panel). A
number of objects, the brightest representatives of our sample, literatureduetoatemplatemismatch(σ-metallicitydegeneracy,
lie in the transition region between the two regimes (“normal” seeChap.1inChilingarian2006).
and“diffuse”ellipticals).
4.3.Stellarpopulationproperties
4.2.Kinematicalproperties
Our fitting method is not sensitive to the presence or absence
The Faber-Jackson relation (Faber & Jackson 1976), reflecting of the Hβ featurein the spectral range:thoughage estimations
the connection between the dynamical and stellar masses, is havehigheruncertaintieswhenHβisnotincluded,theyremain
shown in Fig. 3. We present a compilation of data for dwarf unbiased(seeAppendixBfordetails).
(Geha et al. 2003; van Zee et al. 2004; De Rijcke et al. 2005; To fit the spectra we use the PEGASE.HR synthetic popu-
Matkovic´ &Guzman2005),intermediateluminosity,andgiant lationsbuiltusingtheELODIE.3libraryincludingonlystarsin
ellipticalgalaxiesandbulgesofbrightlenticulars(Benderetal. the nearest solar neighborhood.These stars are known to have
1992). [Mg/Fe]correlatedwith their[Fe/H]metallicities(see Wheeler
Matkovic´ & Guzmán (2005) have analysed a sample of et al. 1989, and references therein). Thus fitting spectra of
mostlydwarfandlow-luminosityellipticalgalaxies(M <−18) galaxies with non-solar [Mg/Fe] for metallicities higher than
B
intheComacluster.InordertohaveaccurateBmagnitudesfor −1.0 dex results in a template mismatch that can bias our es-
thesegalaxies,weretrievedthephotometriccatalogueofgalax- timationsofstellarpopulationparameters.Toquantifypossible
ies in the direction of Coma2 (Adami et al. 2006) and cross- biases and study the effect in detail we used spectra of early-
correlateditwiththeMatkovic´&Guzmáncatalogue.Valuesfor typegalaxiesfromtheSloanDigitalSkySurveyDataRelease6
66matchedobjectshavebeenusedinFig.3. (Adelman-McCarthyetal.2008).SeeAppendixBfordetails.
Someobjectsofoursamplelocatedbelowthebulkofgalax- Toobtain[Mg/Fe]abundanceratiosfortheAbell496galax-
ies in the luminosity range between M = −17 and −19 mag ies and to check whether our age and metallicity estimations
B
are mostly lenticular galaxies with large discs, therefore, their arecorrectweusedstellarpopulationmodelsdealingwithLick
velocitydispersionstendtobelower(e.g.,ACO496J043306.97- indices of magnesium and iron (Thomas et al. 2003). In or-
131238.8(G-01)). der to compute Lick indices we degraded the spectral resolu-
tion to match that needed to compute Lick indices (Worthey
Weobserveapopulationofobjectswithhighervelocitydis-
et al. 1994; Thomas et al. 2003) by convolving the original
persionsthan expectedfortheirluminosities.Mostofthemare
spectra with a Gaussian having a width equal to the square
located in the inner 80 kpc from the cluster centre and exhibit
root of the differences between the squares of the Lick resolu-
quite unusual stellar populations. We will discuss their origin
andevolutionbelow. tion(σLick),LSF(σinst)andvelocitydispersi(cid:2)onvaluesfoundby
We found a number of Abell 496 objects, located sys- thespectralfittingprocedure(σ ):σ = σ2 −σ2 −σ2.
g degr Lick FL g
tematically below dE galaxies in the literature. Since most Oneshouldalsoapplyavelocitydispersioncorrectionforlarge
of the studies (apart from Geha et al. 2003) were based on galaxies,whentheintrinsicbroadeningofthespectrallinesex-
spectroscopy with significantly lower spectral resolution than ceeds the Lick resolution. All the objects in our sample have
FLAMES/Giraffe, we cannot exclude the possibility of sys- relatively low velocity dispersions, therefore these corrections
tematic errors on velocity dispersions for low-σ objects in the werenotnecessary.AsimilarapproachtocomputeLickindices
hasbeenusedbyKuntchneretal.(2006).Thespectralrangeof
1 http://leda.univ-lyon1.fr/ FLAMES/Giraffe in the setup we used and the mean heliocen-
2 http://cencosw.oamp.fr/ tric redshift of the cluster z = 0.0330 allow us to compute the
I.V.Chilingarianetal.:Abell496:Kinematicsandstellarpopulations 91
followingLickindices:Fe , Mgb,Fe ,Fe ,andFe .
5015 5270 5335 5406
Uncertaintiesonthemeasurementswerecomputedaccordingto
Cardieletal.(1998).
There are three sources of systematic errors that can affect
the measurementsofLick indices:(1)data reductionissues re-
sultinginadditiveerrors,suchasproblemswiththediffuselight
or sky subtraction; (2) difference in the spectral resolution be-
tween given observations and the Lick/IDS system; (3) uncer-
taintiesinthedeterminationoftheradialvelocities.
The diffuse light subtraction for FLAMES/Giraffe is done
with a highaccuracythanksto the rathersparse packingofthe
MEDUSA fiber traces on the CCD plane, therefore, it cannot
resultinsignificanterrors.Weobservedtheskysimultaneously
with the galaxies in different parts of the field of view, so it is
possible to assess the quality of sky subtraction by comparing
individual sky spectra obtained in different fibers. No system-
atic difference is observed, we therefore conclude that the sky
subtractionprocedureisreliable.
The radial velocities and velocity dispersions of the
Abell496galaxiesaremeasuredwithhighaccuracy,duetothe
factthattheFLAMES/GiraffeLSFhasalmostinfiniteresolution
intermsofsquaredifferences,comparedtotheLick/IDSsystem.
Fig.4. Mgb – σ (top) and (cid:10)Fe(cid:11)−σ (bottom) relations. Only points
We also notice that in the course of this study to analyze 0 0
Lick indices for both FLAMES/Giraffe and SDSS spectra (see havingΔ(Mgb)<0.5Åareshown.SDSSgalaxieswithagesolderthan
3Gyrareshownaslight-bluedots.Theblack-solidlinerepresentsthe
AppendixB),weusethemodelsbyThomasetal.(2003),rather
correlationbetweenMgbandσ foundbySmithetal.(2006).
0
thanthemodelsbyWortheyetal.(1994).Therefore,converting
the computed values of Lick indices into the Lick/IDS system
throughtheobservationsoftheLick/IDSstarsisnotnecessary.
Moreoverit was notpossible to performthis empiricalconver- the aperture effect may not be very important for giant galax-
sionbecauseLickstarshavenotbeenobserved. ies(asmentionedinGallazzietal.2006)becausetheσgradient
The measurements of the Lick indices are presented in willcompensatethemetallicitygradient;itmaybeimportantfor
AppendixA(TableA.3). dwarfsandintermediateluminosityobjects.
Therehave beena numberof recentstudies (Gallazzietal. Whenmanyclustersofgalaxiesatdifferentredshiftsareob-
2006; Smith et al. 2006; Yamada et al. 2006; Carretero et al. servedwiththesameinstrument(e.g.,Gallazzietal.2006;Smith
2007) addressing correlations between galaxy absorption line etal.2006)theapertureeffectwillincreasethespreadofabsorp-
strengths and velocity dispersions. Most of them are based tionlinestrengthmeasurementsforagivenvelocitydispersion.
on observations obtained with multi-object spectrographs with InFig. 4,we presentthemeasurementsofMgband(cid:10)Fe(cid:11) =
fixed spatial aperture sizes (as in the present study). Although 0.72Fe +0.28Fe versusvelocitydispersion.Wealsoput
5270 5335
thisobservationaltechniqueallowsfortheacquisitionofspectra measurements for ∼700 early-type galaxies older than 3 Gyr
ofmanyobjectsduringasingleexposuretime,ithassomedis- with redshifts z < 0.033 observed in the SDSS on the same
advantageswhenstudyingstellarpopulations.Early-typegalax- plots. We have used values of Lick indices provided by the
ies(giantsanddwarfs)andbulgesofspiralsareknowntohave SDSS. The solid line on the Mgb-σ diagram corresponds to
rather strongmetallicity gradients(see, e.g., Sánchez-Blázquez thebest-fittingrelationfromtheNationalOpticalAstronomical
et al. 2006c) and often evolutionary decoupled nuclei (e.g., Observatory(NOAO)FundamentalPlaneSurvey(NFPS,Smith
Sil’chenko2006;Chilingarianetal.2007b,2008b;Peletieretal. et al. 2006). Measurements for a sample of early-type galax-
2007).Youngand/ormetal-richstellarpopulationsinnucleiwill iesfromSánchez-Blázquezetal.(2006a−c),kindlyprovidedby
not strongly affect the aperture measurements for galaxies lo- Sánchez-Blázquezinacomputerreadableform,areshowninor-
cated 100 Mpc or further, because their spatial sizes are small ange.ThereisrelativelygoodagreementbetweenNFPS,SDSS,
and contributions to the total fluxes within a 1.5 arcsec-wide Sánchez-Blázquezetal.(2006a),andthecorrelationwefindfor
aperturearenegligible,metallicitygradientsmaybiastheresults Abell496galaxies.Ourobjectstendtobeslightlyricherinmag-
quitestrongly. nesium,in contrasttotheSDSSones.Inourcase,wehaveob-
If members of a given cluster of galaxies are analyzed, the servedmostlygalaxiesneartheAbell496clustercore,whilein
scale ofthe gradientwill dependontheir effectiveradii,which thecaseofNFPS,galaxieshavebeenobservedintheperipheral
are rather tightly connected to luminosities. At the same time partsof the clustersaswell, and forSDSS therewas a number
since luminosities of early-type galaxies correlate with central ofgroupandfieldgalaxies(sincewedidnotapplyanyenviron-
velocitydispersions(Faber&Jackson1976),forlow-σobjects mentalselectioncriteria).Smithetal.(2006;seealsoSil’chenko
the same aperture size in average will contain a larger frac- 2006,foranapplicationtolenticulargalaxies)demonstratedthat
tion of the galaxy,leading to an underestimationof the central objectsinthecoresofclusterstendtobeabovetheaverageline,
or effective metallicity. This may cause a change in the slope andconversely,galaxiespopulatinglessdenseenvironmentsare
of the Mgb-σrelation, especially for fainter objects(makingit lessmetal-rich.
steeper). On the other hand, the velocity dispersion also varies We notice four high-σ outliers from the Abell 496 sam-
with radius; however, for low-luminosity and dwarf galaxies, ple significantly above the sequence of Smith et al. (2006)
σ-profiles turn to be almost flat (Simien & Prugniel 2002; in the Mgb-σ plot. Those objects (ACO496J043333.17-
Geha et al. 2002, 2003; van Zee et al. 2004). Therefore,while 131712.6 (G-17); ACO496J043337.35-131520.2 (G-20);
92 I.V.Chilingarianetal.:Abell496:Kinematicsandstellarpopulations
Fig.5.(cid:10)Fe(cid:11)vs.Mgbwithsizesofsymbolsindicatingthevaluesofcen-
tral velocity dispersions. Models from Thomas et al. (2003) for dif-
ferentvaluesof[α/Fe]enrichmentareoverplotted.Onlypointshaving
Δ(Mgb)<0.5Åareshown.
ACO496J043338.22-131500.7 (G-21); ACO496J043341.69-
131551.8 (G-24)) are all located in the innermost part of
Abell 496 and have probably experienced strong tidal harass-
ment(seediscussion).
The (cid:10)Fe(cid:11) versusMgb relation is shown in Fig. 5. The sizes
of symbols reflect values of the velocity dispersion. The data
forearly-typegalaxiesfromSánchez-Blázquezetal.(2006a−c)
areoverplotted.ThemodelsofThomasetal.(2003)for−0.3 <
[Mg/Fe] < +0.5 dex are shown as crosses; each sequence in-
cludes models for an indicated metallicity and ages from 3
to 15 Gyr (bottom-left to top-right). Nearly all objects with
low velocity dispersions exhibit solar [Mg/Fe] ratios (the cor-
responding models are shown in cyan). High-σ objects have
[Mg/Fe]>0.2dexand,ingeneral,tendtohavehighermetallic-
ityvalues.
Fig.6.Relationsbetweencentralvelocitydispersionandstellarpopu-
In Appendix B, we demonstrate that non-solar[Mg/Fe] ra- lationparameters:SSP-equivalentage(top),metallicity(middle),anda
tios bias neither age, nor metallicity estimations obtained by combinationoftheseparametersfreeoftheeffectoftheage-metallicity
spectral fitting in the wavelength range of FLAMES/Giraffe. degeneracy(bottom).Thepresenceofembeddedstructuresinthegalax-
Indeed, the measurements do suffer from the well-known age- iesisindicated.Valuesforearly-typegalaxiesfromSánchez-Blázquez
metallicity degeneracy (see, e.g., Worthey 1994) expressed as etal.(2006a,b)andmeasurementsfor848SDSSgalaxiesareoverplot-
Δt/ΔZ ≈ 3/2or Δ(log t)/ΔZ ≈ 2/3.In AppendixA,we pro- ted.ColoursarethesameasinFig.4.
10
videmapsofχ2 intheage-metallicityspaceforeveryobjectin
oursample.Theelongatedshapeoftheχ2minima,corresponds
exactlytotheexpecteddegeneracy.
In Fig. 6, we compare the age t and metallicity estimates tend to beolderin dense environments,since we find veryfew
versus velocity dispersion for Abell 496 galaxies and for the objectsthereyoungerthan10Gyrhavingσ>60kms−1.
samplesofSDSSobjectsmentionedaboveandearly-typegalax- Metallicityversusσ(middlepanelofFig.6)exhibitsamuch
ies from Sánchez-Blázquez et al. (2006a). We used the values strongercorrelationthanagevs.σ. TheAbell496galaxiesap-
fromSánchez-BlázquezobtainedbyinvertingthegridoftheHβ pear to be almost on the low-σ extension of the sequence of
and (cid:10)MgFe(cid:11) Lick indices. The top panel presents t versus σ. the SDSS galaxies, although there are several faint Abell 496
Abell 496 galaxies are systematically older than objects from dwarfswithratherhighmetallicities.ThebottompanelinFig.6
theSDSSsampleandthespreadofagesisquitelarge.However, shows a combinationof age and metallicity giving an edge-on
wenote(a)theabsenceofyounggalaxieswithhighvelocitydis- view of the age-metallicity degeneracy: 0.4log t + 0.6Z ver-
10
persions;(b)thefactthatsmallgalaxies(withlowvelocitydis- sus velocity dispersion. The correlation becomes much tighter
persions)tendtobeyoungerthanlargeones;and(c)thespread than Z−σ, and loci of Abell 496 and SDSS galaxies follow
of age estimates is larger for low-mass objects. Two explana- the same correlation. The spread of points (standard devia-
tionsforthe offsetbetweenthe Abell496galaxiesandthetwo tion) for a given value of the velocity dispersion is less than
othersamplesonthetoppanelaretheenvironmentaleffectsand 0.1 dex. A number of outliers in the Abell 496 sample are
the sampleselection.We haveselectedtheSDSSgalaxieswith seen: two galaxies in the high-σ area, located in the very cen-
a minimal signal-to-noise ratio, so dwarf galaxies have been tre of the cluster (A496cE and A496g1 using the terminol-
nearly automatically excluded. On the other hand, we confirm ogy from Chilingarian et al. 2007c); and three dwarf galaxies
fromoursample(whichincludesthegalaxiesin the centralre- (ACO496J043324.61-131111.9 (G-08); ACO496J043339.07-
gion of a massive cluster) that intermediate-luminosity objects 131319.7 (G-22); and ACO496J043355.55-131024.9 (G-37)).
I.V.Chilingarianetal.:Abell496:Kinematicsandstellarpopulations 93
This quantity, 0.4log t + 0.6Z, is considered an indicator of
10
theaveragestrengthofabsorptionlinesinthespectrum.
There are seven galaxies in our sample exhibit-
ing relatively young stellar populations (t < 3 Gyr):
ACO496J043321.37-130416.6 (G-07); ACO496J043325.54-
130408.0 (G-12); ACO496J043334.54-131137.1 (G-19);
ACO496J043350.17-125945.4 (G-34); ACO496J043351.54-
131135.5 (G-35); ACO496J043356.18-125913.1 (G-38);
and ACO496J043415.37-130823.5 (G-46). Three of them,
ACO496J043321.37-130416.6 (G-07); ACO496J043356.18-
125913.1 (G-38); and ACO496J043415.37-130823.5 (G-46)
have narrow [OIII] emission lines in the spectra (σ be-
tween 20 and 40 km s−1) and there is no evidence for the
[NI](λ=5199Å)lineinatleasttwoofthem,suggestingongo-
ingstarformationratherthanshockedgas.Thefirsttwo,aswell
asACO496J043351.54-131135.5(G-35),containbluespatially
unresolvedcentralregionsclearly visible on colourmaps.This
canbeconsideredasanargumentforthepresenceofyoungstars
and/orstarformationonlyinthecoresofthegalaxies.Allseven
galaxiesarelocated quitefar fromthe cluster centre(projected
distances are 150−200 kpc for ACO496J043334.54-131137.1
(G-19)andACO496J043351.54-131135.5(G-35)and>450kpc
fortheotherfive).
A significant fraction of galaxies (13 of 46) exhibit
red spatially unresolved cores in the colour maps. Stellar
populations of these objects determined by spectral fit-
Fig.7.Ageand[Mg/Fe]abundanceratiosofgalaxiesasafunctionof
ting are rather old and metal rich, none of the spectra
projecteddistancetotheclustercentre.Onlygalaxieshavinguncertain-
shows emission lines. In some cases (ACO496J043320.35- tieson[Mg/Fe]determinationsbetterthan0.1dexareshowninthebot-
130314.9 (G-06); ACO496J043331.48-131654.6 (G-15); tompanel.Theremaininggalaxiesareshowninblueinthetoppanel.
ACO496J043333.53-131852.6 (G-18); ACO496J043342.10-
131653.7 (G-25); ACO496J043346.71-131756.2 (G-31); and
ACO496J043401.57-131359.7 (G-40)) the red cores represent
the central parts of extended bar-like structures, clearly visible
IntheirstudyofVirgoearly-typedwarfgalaxies,Liskeretal.
on colour maps, which are redder than the outer parts of
(2007,andreferencestherein)donotgivethestellarpopulation
their host galaxies. However, in most of the other red-core
properties of the three subclasses, which are the result of in-
galaxies only the cores show peculiar colours, while the
fallinggalaxytransformation.Ifweputtogethertheeightdwarfs
discs/spheroids look more or less uniform on the maps. In the
correspondingtotheirdE(di)subclassandthefiveotherdwarfs,
caseofACO496J043306.97-131238.8(G-01),anedge-ondisky
which exhibit very young central stellar populations, this sub-
galaxy, dust absorption is responsible for the redder colour of
sample presents the same spatial and velocity distributions as
thecentralregion.
theirleastunrelaxeddwarfgalaxypopulation.Thevelocityhis-
Tryingtorelatethepresenceofembeddedstructures(discs, togram is more spread for the 13 selected dwarfs compared to
spiralarms,bars)tothestellarpopulationparameters,westress that of the 24 remaining early-type dwarfs, including the two
thatthemajorityofthegalaxiesexhibitingembeddedstructures galaxies for which we did not succeed to fit the spectra due
have luminosity-weightedages of between 4 and 10 Gyr (only to low signal-to-noise ratios. If we also take into account their
ACO496J043346.71-131756.2(G-31)andACO496J043403.19- corresponding spatial distributions, we can conclude that the
131310.6 (G-41), with bars and faint spiral arms, are about subclass isprobablynotfullyrelaxed.A cleardifferenceexists
14Gyrold,buttheirabsolutebluemagnitudeisbelow−18mag; in the stellar populationages, since among the 22 dwarfs with
ACO496J043321.37-130416.6 (G-07) is a young star-forming noembeddedstructures,onlysixgalaxieshaveagesbetween3
galaxy)andallofthemhavefeaturesvisibleoncolourmaps.No and9Gyr.
objectswithembeddedstructuresarefoundintheveryinnerpart Thebrighterpartofoursamplecanbecomparedtoexisting
of the cluster (d < 60 kpc) where the luminosity-weighted studies of stellar populations of early-type galaxies. Poggianti
proj
ages of the five galaxies in our sample are above 12 Gyr. In etal.(2001)presentedtheluminosity-weightedagesandmetal-
Fig. 7, the ages and α-enhancements of galaxies at different licities for several dozens of Coma cluster early-type galaxies.
projected distances from the Abell 496 centre are shown. In a From their Fig. 2, it is clear that there are two populations of
transition area (d between 60 and 230 kpc) the age of the galaxies,oldandyoung.TheageoftheoldonesforM >−19.0
proj B
youngest stellar populations decreases when the projected dis- anticorrelateswiththeluminosity.Inoursample,however,wedo
tance increases, as it can be seen in Fig. 7. Half of the galax- nothaveanybrightgalaxywithayoungstellarpopulation.This
ieswithembeddedstructuresarefoundinthisareabutsomeof isprobablyduetotheselectionofobjectsinoursample:(1)we
themcouldhaveadistancetotheclustercentrelargerthantheir tried to avoid galaxies with redshifts already in the literature;
projected distance. In the bottom panel of Fig. 7, the ratio of (2) young galaxies have higher surface brightnesses compared
α-elementsoverironinthetransitionareaisnotaslargeasfor tooldonesforagivensize,therefore,theyhaveabetterchance
theinnermostpartoftheclusterandspreadsoverthesamerange to have already been observed in redshift surveys with limited
asforgalaxiesatlargerprojecteddistances. aperturemagnitudes,andthustobeexcludedfromoursample.
94 I.V.Chilingarianetal.:Abell496:Kinematicsandstellarpopulations
Sánchezetal.(2007)haveobtainedinternalkinematicsand
stellar population parameters for galaxies in the core of the
Abell2218clusterusing3D-spectroscopyoftheclustercentre.
ThebehaviorofageestimatesissimilartothatoftheAbell496
galaxies:forlow-massobjects,agestendtobeyoungerandmore
spreadoutthanforlargegalaxies.
5. Discussion
5.1.OntheoriginofdE/dS0galaxies
We discuss here what we can learn from the stellar population
properties of dE/dS0s in Abell 496 regarding their origin and
evolution. First, what is the explanation for the spread in the
stellaragesandmetallicitiesofourdwarfgalaxysample?Isthis
spreadrelatedtosomedifferencesinthetimeneededforgalax-
iestobeaccretedintotheclusterandtoreachtheclustercore?
Low-massgalaxiesare more sensitivethan massive galaxiesto
bothinternalandexternalprocessesaffectingtheirgascontents
andstarformationrates.Therefore,the“insitu”formationsce- Fig.8.1-σconfidencelevelsofageandmetallicitydeterminationsfor
nario,whichisabletoreproducethevarioustimescalesforstar the Abell 496 galaxies. Objects in the three velocity dispersion bins
formation,could work as well as an accretion scenario. Which are plotted using different colours: red for σ < 30 km s−1, blue for
scenario can explain that star formation occurs in the nuclear 30<σ<60kms−1,andblackforσ>60kms−1.
regionsforanumberofdwarfs?Starformationisdriveninpar-
ticularbythegascontent,andifthegasisrapidlyremovedfrom
its host galaxy, star formation can be stopped in a very short between the dynamical mass and metallicity can be explained
time.Ontheotherhand,aninfallofgasinthegalaxycentrewill by the lower efficiency of star formation in galaxies of lower
inducestarformationinthenuclearregion.Thescenarioshould massesduetomoreefficientsupernovafeedback.However,un-
also reproducethe observed[Mg/Fe]ratio andits relationwith derthisscenario,onewouldnotobserveanyolddwarfellipticals
thevelocitydispersion. becauseyoungerstars,formedinthesecondaryepisodesofstar
Three possible scenarii for the gasremovalusually consid- formation, would dominate the light even though their masses
eredfordEgalaxiesare:(1)supernova-drivenwindsattheearly mightbe low comparedto the massesof the first generationof
stages of galaxy evolution (Dekel & Silk 1986); (2) ram pres- stars. Therefore,one still needs a mechanism to sweep out the
surestrippingbytheintra-clustermedium(e.g.,Marcolinietal. remaininggasfromthelow-massgalaxiesearlierorlaterduring
2003); and(3) tidalharassmentdueto distantandrepeateden- theirlifetime.
counters with other cluster members (Moore et al. 1998). The In the case of Abell 496, [Mg/Fe] ≈ 0 dex for nearly all
first oneis oftenreferredto as“internal”andlattertwo as“ex- low-massobjects(σ <60kms−1),meaningthatthestarforma-
0
ternal”agentsofdEgalaxyformationandevolution. tionepochdurationswereatleast1−2Gyr,theminimalrequired
Theideaofsupernova-drivenwindsisbasedontheassump- timetocompletetheironenrichment(Matteucci1994).Hence,
tion that the gravitational field of dwarf galaxies is not suffi- we cannot consider the scenario of gas removal by supernova-
cientlystrongtokeeptheinterstellarmediumfrombeingswept drivenwindsastheonlyexplanationfortheobservedproperties
out by SN II explosions during the first intense star formation ofdEgalaxies,althoughitallowsustoreproducetheobserved
episode. mass-metallicitycorrelation.
Several different models for galactic winds exist (see Ram-pressurestrippingoflate-typedwarfgalaxies(dIrr)or
De Rijcke et al. 2005, for a detailed review). Simple mod- dwarfspiralsappearstobeanacceptablewaytoremovethegas
els discussed in Yoshii & Arimoto (1987) lead to abrupt gas from dE progenitors.If we assume that late-typegalaxieshave
loss and interruptionof the star formationepisodeaftera short formed outside the central region of the cluster, and later fell
time (107 years). Consequently, later explosions of SNIa on a ontoit,sufficienttimeisleftforironenrichment,sincethetyp-
timescaleofgigayearswillnotcontributetotheironenrichment icalinfalltimeisafewGyr.Ifram-pressurestrippingplaysthe
of the stellar population (Matteucci 1994). Thus, a short star leading role in gas removal, one would expect a large spread
formationepisodewillleadtoanoverabundanceofα-elements of luminosity-weighted ages for low-mass objects that can be
overiron([α/Fe] > 0).Thisphenomenonisobservedinglobu- completely strippedduringtheir first passage throughthe clus-
larclustersand,usually,ingiantearly-typegalaxies(Kuntschner ter centre. This stripping can occur at any moment during the
etal.2006;Sil’chenko2006).IfgasisremovedfromdEgalaxies galaxy lifetime. Due to our small sample we cannotgive a de-
bysupernova-drivenwinds,we wouldexpectto see[Mg/Fe] > cisiveanswertowhetherthespreadofageestimationsinFig.6
0 dex, and it should anticorrelate with the dynamical mass of is the result of ram-pressure stripping of late-type progenitors,
galaxies (or σ). This does not agree with what we see in our or due to the low quality of measurements (including the age-
data. metallicitydegeneracyeffects).
Another study by Chiosi & Carraro (2002), contrary to Smith etal. (2007) pointedoutthatthereis a conspiracyin
Yoshii&Arimoto(1987),predictsverylongandoscillatingstar the sense thatthe distributionof galaxiesin the Z −logt plane
formationhistoriesindwarfgalaxies:supernovaexplosionsdis- at fixed velocity dispersion is aligned with the age-metallicity
perse the gas, stopping star formation, but later the gas cools degeneracy. In Fig. 8, we plotted the relationship between age
down,fallsbackinandanotherstar-formationepisodebegins.In andmetallicityfordifferentintervalsofsigmawiththeSmithet
thiscase,the[Mg/Fe]ratiosdecreaseto∼0dex,thecorrelation al. sample of faint red galaxies in three clusters located inside