The XXL Survey

15 Pages·2016·2.12 MB·English

Adami, C., Pompei, E., Sadibekova, T., Clerc, N., Iovino, A., McGee,

S., Guennou, L., Birkinshaw, M., Horellou, C., Maurogordato, S.,
Pacaud, F., Pierre, M., Poggianti, B. M., & Willis, J. (2016). The XXL
Survey: VIII. MUSE characterization of intra cluster light in a z ~ 0.53
cluster of galaxies. Astronomy and Astrophysics, 592, [A7].
https://doi.org/10.1051/0004-6361/201526831
Publisher's PDF, also known as Version of record
License (if available):
CC BY-NC
Link to published version (if available):
10.1051/0004-6361/201526831
Link to publication record in Explore Bristol Research
PDF-document
This is the final published version of the article (version of record). It first appeared online via EDP Science at
10.1051/0004-6361/201526831. Please refer to any applicable terms of use of the publisher.
University of Bristol - Explore Bristol Research
General rights
This document is made available in accordance with publisher policies. Please cite only the
published version using the reference above. Full terms of use are available:
http://www.bristol.ac.uk/red/research-policy/pure/user-guides/ebr-terms/
Astronomy&Astrophysicsmanuscriptno.n308-aaenglish (cid:13)c ESO2016
January5,2016
The XXL Survey VIII: MUSE characterisation of intracluster light in
a z∼0.53 cluster of galaxies (cid:63)
C.Adami1,E.Pompei2,T.Sadibekova3,N.Clerc4,A.Iovino7,S.L.McGee9,L.Guennou6,1,
M.Birkinshaw5,C.Horellou13,S.Maurogordato8,F.Pacaud10,M.Pierre3,B.Poggianti11,andJ.Willis12
1 AixMarseilleUniversite´,CNRS,LAM(Laboratoired’AstrophysiquedeMarseille)UMR7326,F-13388,Marseille,France
2 EuropeanSouthernObservatory,AlonsodeCordova3107,Vitacura,19001Casilla,Santiago19,Chile
3 LaboratoireAIM,CEA/DSM/IRFU/SAp,CEASaclay,91191,Gif-sur-Yvette,France
4 Max-Planck-InstituteforExtraterrestrialPhysics,Giessenbachstrasse1,D-85748,Garching,Germany
5 HHWillsPhysicsLab.,Univ.ofBristol,TyndallAve.,BristolBS81TL,U.K.
6 Institutd’AstrophysiqueSpatiale(IAS),bat.121,F-91405OrsayCedex,France
7 INAF-OsservatorioAstronomicodiBrera,viaBrera,28,I-20121Milano,Italy
8 Laboratoire Lagrange, UMR7293, Universit de Nice Sophia Antipolis, CNRS, Observatoire de la Cte d’Azur, F-06300, Nice,
France
9 SchoolofPhysicsandAstronomy,UniversityofBirmingham,Edgbaston,Birmingham,B152TT,U.K.
10 ArgelanderInstitutfu¨rAstronomie,Universita¨tBonn,AufdemHu¨gel71,D-53121Bonn,Germany
11 INAF-AstronomicalObservatoryofPadova,Italy
12 DepartmentofPhysicsandAstronomy,UniversityofVictoria,3800FinnertyRoad,Victoria,BC,Canada
13 Dept.ofEarth&SpaceSciences,ChalmersUniversityofTechnology,OnsalaSpaceObservatory,SE-43992Onsala,Sweden
Accepted.Received;Draftprinted:January5,2016
ABSTRACT
Aims.Withinacluster,gravitationaleffectscanleadtotheremovalofstarsfromtheirparentgalaxiesandtheirsubsequentdispersal
intotheintraclustermedium.Gashydrodynamicaleffectscanadditionallystripgasanddustfromgalaxies;bothgasandstarscon-
tributetointraclusterlight(hereafterICL).ThepropertiesoftheICLcanthereforehelpconstrainthephysicalprocessesatworkin
clustersbyservingasafossilrecordoftheinteractionhistory.
Methods.ThepresentstudyisdesignedtocharacterisethisICLforthefirsttimeina∼1014M(cid:12)andz∼0.53clusterofgalaxiesfrom
imagingandspectroscopicpointsofview.Byapplyingawavelet-basedmethodtoCFHTMegacamandWIRCAMimages,wedetect
significantquantitiesofdiffuselightandareabletoconstraintheirspectralenergydistributions.Thesesourceswerethenspectroscop-
icallycharacterisedwithESOMultiUnitSpectroscopicExplorer(MUSE)spectroscopicdata.MUSEdatawerealsousedtocompute
redshiftsof24clustergalaxiesandsearchforclustersubstructures.
Results.AnatypicallylargeamountofICL,equivalentini’totheemissionfromtwobrightestclustergalaxies,hasbeendetectedin
thiscluster.Partofthedetecteddiffuselighthasaveryweakopticalstellarcomponentandapparentlyconsistsmainlyofgasemission,
whileotherdiffuselightsourcesareclearlydominatedbyoldstars.Furthermore,emissionlinesweredetectedinseveralplacesof
diffuselight.Ourspectralanalysisshowsthatthisemissionlikelyoriginatesfromlow-excitationparametergas.Globally,thestellar
contributiontotheICLisabout2.3×109 yrsoldeventhoughtheICLisnotcurrentlyformingalargenumberofstars.Ontheother
hand, the contribution of the gas emission to the ICL in the optical is much greater than the stellar contribution in some regions,
butthegasdensityislikelytoolowtoformstars.Theseobservationsfavourrampressurestripping,turbulentviscousstripping,or
supernovaewindsastheoriginofthelargeamountofintraclusterlight.
Sincetheclusterappearsnottobeinamajormergingphase,weconcludethatrampressurestrippingisthemostplausibleprocess
thatgeneratestheobservedICLsources.
Conclusions. This is one of the first times that we are able to spectroscopically study diffuse light in such a distant and massive
cluster,anditdemonstratesthepotentialofMUSEobservationsforsuchstudies.
Keywords.galaxies:clusters
1. Introduction properties (morphology, star formation rate, colour, etc.) from
severalgalaxiesofsimilarmassinthefield(Dressleretal.1980,
Thediffuselightwithinagalaxycluster,oftencalledintracluster
1997; Weinmann et al. 2006). Physical processes could drive
light(ICL),isafossilrecordoftheformation,accretion,andin- thesedifferences,butluckily,asgalaxiesareaccretedandorbit
teractionhistoryofthecluster(e.g.Guennouet.al.,2012).Ithas
within acluster, each physicalprocess may leave characteristic
beenknownforsometimethatgalaxiesinclustershavedifferent low surface brightness signatures. Gravitational effects, includ-
ing mergers and low-energy interactions (Mihos 2004, 2005),
Send offprint requests to: C. Adami e-mail: tidal stripping (Henriksen & Byrd, 1996), and high-speed ha-
[email protected] rassment(Mooreetal.1998),willstriporejectstellarmaterial
(cid:63) BasedonobservationsmadewithESOTelescopesattheLaSilla into the intracluster medium (ICM). This was recently statisti-
andParanalObservatoriesunderprogrammesID191.A-0268and60.A- cally studied within the CLASH sample (Postman et al. 2012)
9302.
1
Article published byEDP Sciences,to be cited ashttp://dx.doi.org/10.1051/0004-6361/201526831
Adamietal.:MUSEcharacterisationsofIntraClusterLightinaz∼0.53clusterofgalaxies
byBurkeetal.(2015),whoconcludedthatthegrowthofstellar shallowpotentialwell,wheretheencountersareslowerthannor-
mass in the ICL is larger than can be provided by the brightest malowingtothelowercharacteristicvelocitydispersion.
cluster galaxy (BCG hereafter) close companions, and that the ThepresentstudyisdesignedtocharacterisetheICLusing
majority of the ICL mass must come from galaxies which fall bothimagingandspectroscopyforthefirsttimeinamassiveand
fromoutsideofthecoreoftheclusters.Thiswouldbeinagree- distant cluster of galaxies. We will investigate the relative con-
ment with De Maio et al. (2015) who used the same sample to tribution of old stars, presumably stripped from the interacting
conclude that the ICL is built up by the stripping of galaxies galaxies, and the in situ intracluster star formation. While we
that are not too faint, and disfavoured significant contribution expect that the first contribution will be dominant, the second
totheICLbydwarfdisruptionormajormergerswiththeBCG. seemstohavebeenobservedinVirgo(Verdugoetal.2015).
Alternatively,hydrodynamicaleffectssuchasrampressurestrip- WefirstdescribetheclusterfromanX-raypointofview.In
pingofcolddiskgasanddust(Gunn&Gott1972)canalsopol- Section3wediscusstheopticalbroadbanddataandthesubse-
lute the ICM with gas that may become the material of future quent detection/characterisation of the diffuse light. Sections 4
in situ star formation. We note, however, that studies such as and 5 describe the Multi Unit Spectroscopic Explorer (MUSE)
Melnicketal.(2012)clearlydonotfavouraveryintenseinsitu measurements and the characterisation of the cluster ICL.
starformation.Excitationprocesseswithinthisgasmayleadto Finally,weconcludeinSection6.
detectableemissionlineswithintheICL. We adopt the standard concordance cosmological model
CurrentobservationsofICLhaveshownthattherearelarge (H0=72kms−1Mpc−1,ΩΛ =0.74,ΩM =0.26).
stellar streams that provide one viable input mechanism to the
ICL(Rudicketal.,2009),butonlyahandfulofsuchstreamsare
known (e.g. Coma, Centaurus, Virgo, Perseus, Norma) and the 2. XLSSC116fromanX-raypointofview
total amount of ICL is almost always equivalent at maximum
Our target is a cluster of galaxies detected in the XXL Survey
to a few L galaxies in massive clusters (e.g. Guennou et al.
(cid:63) (XXL paper I) as a significant extended X-ray source. This is
2012).Fromageneralpointofview,theICLcontributiontothe
notoneofthemostsignificantclusters(onlyC2class,seeXXL
clusterluminosityislessthanafewdozenpercentagepoints.For
paper I), and is not part of the 100 brightest XXL clusters (see
example,usingTypeIasupernovae,Sandetal.(2011)estimated Pacaudetal.2015,hereafterXXLpaperII1).However,itvisu-
in a large cluster sample the intracluster (intergalactic) stellar allyexhibitsaverylargeamountofdiffuselight(seebelow)and
masstobeoftheorderof16%ofthetotalclusterstellarmass.
so has probably experienced an unusual evolution. This source
Gonzalez et al. (2007) found a value of 19% when expressed isprobablyaffectedbynon-thermalemissionbecausethegalaxy
withthesimilarconventionsasinSandetal.(2011).
dominatingtheclusterisdetectedasaradiosource(potentially
MostoftheliteraturestudiesoftheICLarebasedonimag-
coinciding with NVSS J021039-055637, later identified with
ingsurveys.Thereare,however,somenoticeableexceptions,for FIRST data at 02h10m39.7sec -05deg56(cid:48)36.9(cid:48)(cid:48) with a possible
example in Melnick et al. (2012). These authors used a z=0.29 extensionof∼1.5(cid:48)(cid:48)),anditwasconfirmedasaclusteratz∼0.53
clusterofgalaxies(RXJ0054.0-2823)andstackedseveralspec-
(withtheredshiftoftheBCG)inthecourseofthegeneralXXL
trafromfourdifferentplacesinthecluster;however,thisstudy
follow-up(WHTobservations;seeKoulouridisetal.2015,here-
was forced to use the ICL and the BCGs together and to rely
afterXXLpaperXII,fordetailsonobservationsanddatareduc-
onthesurfacebrightnessprofilestodelineatetheboundariesbe-
tion).Itwaspreviouslydetectedasacandidateclusterintheop-
tweenthetwocomponents.TheICLcomponentwasassumedto
ticalwithaphotometricredshiftof0.47byDurretetal.(2011:
be located at more than ∼50 kpc from the BCGs (but a signifi-
W1-0957).
cantcontaminationbyaBCGisstillpossible).Theyfoundthat
themajorityoftheICLstarsinthisclusterareold,butwithhigh
metallicitiesprobablyrelatedtothefactthattheconsideredclus- 2.1. X-raymorphologyoftheXLSSC116cluster
ter has three interacting giant elliptical galaxies. Interestingly,
Figure 1 shows surface brightness contours of the galaxy clus-
they also found a 408±20 km/s velocity dispersion for the ICL
ter in three X-ray bands: [0.3-0.5], [0.5-2] (usual soft energy
very close to the cluster velocity dispersion itself and signifi-
band encompassing most of the Bremstrahlung emission) and
cantly larger than the BCG velocity dispersion. This also puts
[2-10]keV.DespitethelimitednumberofX-raycountsineach
theBCG-independentnatureoftheICLonfirmerground.They
band and the size of the XMM point-spread function (PSF),
finallydetectedextremelyweak[OII]and[OIII]emissionlines
surface brightness contours may reveal a bimodality of the ex-
withintensityratiosthatmightbeconsistentwiththoseofmetal
tended emission. Lower-temperature gas (more prominent in
richHIIregions.
[0.3-0.5] keV) is shifted with respect to the hotter phases. The
The present paper is based on the detection in the XXL [0.5-2]keVcontoursareelongatedandshowan11(cid:48)(cid:48) offsetbe-
Survey (Pierre et al. 2015, hereafter XXL paper I) of a clus-
tweenthesurfacebrightnesspeakandtheBCG(∼70kpcatthe
ter,XLSSC116,atz=0.534withanexceptionalamountofICL
clusters redshift). In order to put error bars on the separation,
comparedtootherstudiedclustersintheliterature;inthiscluster
werandomlyeliminateonethousandtimes50%ofthecountsin
thisisequivalenttoalmosttwobrightestclustergalaxies(BCG
the 0.5-2 keV bandpass and recalculate the centroid each time.
hereafter) in the i’ band, which is nearly ten times larger than
Thedistributionofseparationsinthesetrialsgivesameasureof
normally observed. The existence of such a system is puzzling
theuncertainty.AssumingPoissonstatistics,the1σuncertainty
inthecommonscenariosofICLformation(e.g.Guennouetal. using all counts is ∼5(cid:48)(cid:48). This 11±5 (cid:48)(cid:48) shift is relatively impor-
2012). While ICL is expected to be efficiently formed in clus-
tant(e.g.Adami&Ulmer2000),potentiallyindicatingthatthe
ters/groups, the amount observed in XLSSC 116 is exceptional
cluster has a complex formation history. The other considered
andlikelyindicatesrecentstrongdynamicalinteractionswithin
X-ray bands are also shifted with respect to the BCG position,
thecluster.Suchstronginteractionsmayresultfromahighde-
gree of substructure merging within a young cluster containing 1 XXL 100 brightest cluster data are available in computer read-
a low-temperature intracluster medium. Alternatively, such in- ableformviatheXXLMasterCataloguebrowserhttp://cosmosdb.iasf-
teractions may take place in a merging system with a currently milano.inaf.it/XXL
2
Adamietal.:MUSEcharacterisationsofIntraClusterLightinaz∼0.53clusterofgalaxies
buttheshiftsaresmallerandnotsignificantgiventhedataavail- (seebelow),thatincludesthedensestpartsofthecluster.Itcon-
able.UnlesstheX-rayemissioniscontaminatedbyafaint,unre- tains 806 net counts (thermal + non-thermal emission) in [0.3-
solved,point-source,thispeculiarmorphologymaysupportthe 10]keV.WeusedXSpec(Arnaud1996)tocomputebest-fitval-
picture of different gas phases in the process of merging. We uesandconfidenceintervals.
note,however,thatthetypicalintrinsicXMMPSFwidthwhere We first fit the spectrum with a single-temperature APEC
theclusterislocatedisoftheorderof8(cid:48)(cid:48),resultingina9.4(cid:48)(cid:48)total plasma model, fixing the metallicity abundance to 0.3 solar
uncertaintyatthisplace.HigherresolutionX-raydata(Chandra) (Grevesse & Sauval, 1998), and we obtain a best-fit tempera-
wouldthereforebenecessarytomoreefficientlydiscriminatebe- ture of 2.0+1.3 keV. Since r ∼ 520kpc ≡ 83 (cid:48)(cid:48) at the cluster
tweenthedifferentregionspresentlyconsidered. redshift(fr−o0m.5Sunetal.(2050009)T −r relation,andassuming
500
T =2keV),thistemperaturemeasurementroughlycorresponds
toT .
500
If we perform a similar fit with the metallicity parameter
(abundance) left free, we find a best-fit at T = 2.1+1.3 keV and
−0.7
Ab = 0.3(abundanceinsolarunits),butwecannotassignerror
barstothemetallicityparametergiventheuncertaintiesinthefit
thataretoolarge.
A double-component APEC model fit with different tem-
peratures and normalisations, both at fixed Ab = 0.3 solar,
gives a valley of minima for temperatures T ∼ 2 keV (resp.
1
T ∼ 2keV),correspondingtothepreviouscasewherethenor-
2
malisation of component 2 (resp. component 1) is set to zero.
More interestingly, we find a best-fit solution at T = 0.3 and
1
T = 2.1 keV (the two black dots with error bars in figure 2).
2
Althoughthesignificanceislow,thisfindingagainsupportsthe
caseforabimodalgasdistribution,withonecoolersubstructure
merging into a hotter, 2 keV cluster. According to these spec-
tralmodels,thecoolercomponentaccountsfor10%ofthetotal
systemluminositywithina90(cid:48)(cid:48)(560kpc)projectedradius.
We measure a [0.5-2] keV count-rate of 0.0198 cts.s−1
(±23%)inthisradius,whichtranslatesintoagalacticabsorbed
flux f = 1.8×10−14 ergs/s/cm2 (±23%).Thisvaluecorre-
[0.5−2]
sponds to a rest-frame [0.5-2] keV luminosity L[0.5−2] = 2.0×
500
1043ergs/sandabolometricluminosityLbol =4.7×1043ergs/s.
500
Thepositionoftheclusterintheluminosity-temperatureplaneis
fullycompatiblewithrecentfindingsfromtheliterature(Fig.2).
3. CFHTLSandWIRCAMbroadbanddata:diffuse
lightdetectionandspectralenergydistribution
fitting
3.1. Opticalimagingdata
The cluster is located in the CFHTLS W1 field, providing u*,
g’, r’, i’, and z’ data. We refer the reader to Coupon et al.
(2009) for details. The limiting magnitude of these images is
u* =26.3,g’ =26.0,r’ =25.6,i’ =25.4,z’ =25.0,andthe
AB AB AB AB AB
Fig.1. Upperfigures:X-raymorphologyofXLSSC116asafunction the pixel scale is 0.187(cid:48)(cid:48). In addition, we collected our own
ofenergyband.TheimageisCFHT-LSi(cid:48)withX-raycontoursoverlaid. CFHT/WIRCAM K data (subject of a future paper). Briefly,
s
Crossesindicatethepositionofthesurfacebrightnesspeakineachse- theywereobtainedas2×525secondspointingswitheachpoint-
lectedsub-band.Therightpanelsshowthreepoint-sourcesinthevicin- ing consisting of 21 dithers with 25 seconds of exposure per
ityofXLSSC116withspatialscalesandcontourlevelsmatchingthose dither,foratotalexposuretimeof1050s.Datawerereducedby
ofthemainimage.TheyillustratethesizeoftheXMM-NewtonPSFat the standard TERAPIX pipeline and presented as 1 deg2 tiles
thelocationofthecluster.Wenotethepresenceofapoint-sourceinthe
projected astrometrically onto CFHTLS optical images (same
[2-10]keVbandwestoftheclustercentre.Thissourceismaskedout
pixelscale).ThedepthisclosetoK =22(5σlevel)inaKron-
intheX-rayanalyses.Lowerfigure:zoomonthesamei(cid:48) imagewith sAB
typeaperture.
Serna-Gerbal substructure 1 shown as red squares and substructure 2
shownasbluesquares(seebelow).Galaxyredshiftsarealsoshown. Fig. 3 shows the galaxy colour magnitude relation in a 5(cid:48)
square around the XLSSC 116 cluster centre. We clearly see a
narrow red sequence. Only five galaxies that are spectroscopi-
callyclassifiedasclustermembers(seebelow)arebluerthanthis
red sequence. Of these, two are close to the BCG, well within
2.2. X-rayspectralanalysisoftheXLSSC116cluster
the diffuse light region. The BCG is the brightest cluster mem-
TheX-rayspectrumisextractedinaregionofradius90(cid:48)(cid:48)around ber galaxy, but it is very different from a typical cD galaxy, as
the BCG position. This corresponds to slightly larger than r weshowlaterusingtheMUSEspectra.
500
3
Adamietal.:MUSEcharacterisationsofIntraClusterLightinaz∼0.53clusterofgalaxies
method(theOV WAVpackage,seee.g.Pereira2003;DaRocha
& Mendes de Oliveira 2005). OV WAV is a multiscale vision
model (e.g. Rue´ & Bijaoui 1997). We briefly outline the main
steps of the method. After applying a wavelet transform to an
observation,thetoolidentifiesthestatisticallysignificantpixels
inthewavelettransformspace(atthe5σlevelinourcase).To
define the objects, it groups pixels in connected fields for each
scale. After the construction of an inter-scale connectivity tree,
fieldswithconnectedpixelsacrossthreeormorescalesareiden-
tifiedandassociatedwiththeobjects.
We detected small-scale objects (typically the galaxies) in
theskyimagetoproduceourobjectimage.Weconsideredchar-
acteristicscalesbetween1and512pixelsinwaveletspace(the
1024pixelscaleasinGuennouetal.2012didnotprovideany
additionaldetection).Theobjectimagewasthensubtractedfrom
the sky image to produce the residual image. This residual im-
age includes both hidden features that are typically too faint to
satisfythewaveletfirst-passthresholdingconditionsandthedif-
fuselightpatches,whicharetoofainttobedetectedasobjects.
We did not have to use a second iteration of this process as in
Guennou et al. (2012) as the data were of sufficient quality to
detectallobjectsofinterestinthefirstpass.
Inasecondstepwesearchedforwhatwecallthesignificant
ICLsources,i.e.extendedlowsurface-brightnessfeaturesinthe
Fig.2.PositionofXLSSC116inthesoft-bandluminosityversustem-
residual image. These features were detected in this image by
peratureplane(reddotwitherrorbars).Greypointsaremeasurements
fromtheXXL100brightestclustersample(Gilesetal.2015,hereafter consideringthepixelswherethesignalislargerthan2.5σwith
XXL paper III). The straight line represent the scaling relation from respect to 20 empty areas of the residual image. These sources
XXLpaperIIIfittedtotheXXL100sample(forcedtoself-similarevo- were visually inspected to remove obvious numerical residuals
lution). Black dots with error bars correspond to the two-component of bright saturated Galactic stars or defects due to image cos-
APEC fit to the cluster spectrum, possibly indicating the presence of metics.ThesenumericalresidualsaredescribedinAdamietal.
twostructures.Thelowtemperaturepointhasnotemperatureerror-bar (2013); briefly, they can be modelled by Sinc functions, which
becausewecannotconstrainitowingtothesmallnumberofphotons.
arenegligibleinintensityinthepresentcasebecausetheconsid-
eredpointsourcesarefaint.
Thisexercisewasdoneindependentlywiththeu*,g’,r’,i’,
z’,andK images.WeshowinFigs.4and 5theresultsinthesix
s
availablebands.Essentially,wedetectacomplexsourceofICL
in which several galaxies are embedded. This source (union of
allellipsesinfigs.4and 5andunionofregions1and2infigs.12
and 13)is∼60×180kpcinsize,whichisverylargecomparedto
thetypicalICLsourcesfoundbyGuennouetal.(2012).Indeed,
it is even larger than the plume detected in the Coma cluster
by Gregg & West (1998) or the diffuse light regions in Norma
describedinFumagallietal.(2014).
We measured the source total magnitudes in the several
availablebandsbyintegratingthefluxoftheresidualimagesin
thepinkellipsesinFigs.4and 5.Thesizeoftheseellipseswere
adaptedtothedepthandqualityofeachbandinordertobesten-
closethe2.5σdetectionarea.Wenotethatthefluxesinthedif-
ferentbandswerenotmeasuredinexactlythesameregionsbe-
causethedifferentconsideredbandsdonothavethesamedepth
orquality.Forexample,theu*andK bandsaremuchshallower
s
than other optical bands, so considering exactly the same areas
would lead to including significant noise in our flux estimates.
The magnitudes are compiled in Table 1, and clearly show an
Fig.3.g’-r’versusr’fortheCFHTLSobjectsina5(cid:48) square(allowing extraordinary amount of ICL. We compared these magnitudes
a∼1Mpcradiusattheclusterredshift)aroundtheXLSSC116cluster
to the magnitude of the BCG in each of five optical bands and
centre.Pointsareallthedetectedobjects;filledcirclesarethegalaxies
foundthattheICLhasmorethan2timesthefluxoftheBCGin
withaMUSEmeasuredspectroscopicredshift(red:SGsubstructure1,
g’,r’,i’,orz’.
blue:SGsubstructure2,seeSection4.2).Thelargeblueoctagonisthe
sumofallthediffuselightcomponentsinthecluster. To the best of our knowledge, this is the first time such a
system has been detected. Even the very high concentration of
ICLdetectedatz∼2.1intheCLJ1449+0856cluster(Adamiet
al. 2013) is far from being at the same level. Moreover, ICL in
3.2. Diffuselightdetection
theXLSSC116clusterisclearlyvisibleupto130kpcfromthe
Todetectthediffuselight,weappliedthesametoolasinAdami clustercentre,afeaturethatisalsoquiterareinclusters.
etal.(2013)andGuennouetal.(2012).Thisisawavelet-based
4
Adamietal.:MUSEcharacterisationsofIntraClusterLightinaz∼0.53clusterofgalaxies
Fig.4.Left:originalCFHTLSandWIRCAMimagesintheconsideredbands.Right:residualimagefromOV WAV.Redcontouristhe2.5σlevel
fromtheresidualimage.Bluecontoursstartatthe3σlevel(fromtheresidualimage)andprogressbystepsof1σ.Thesecontoursarecomputed
withtheresidualimagesandreportedonbothresidualandoriginalimages.Thecyancircle(80kpcindiameter)showsoneoftheemptyresidual
fieldsweconsidered.Italsogivesthephysicalsizeofthefigure.Pinkellipsesaretheareaswherewesummedthefluxofthediffuselightinthe
residualimage.ThegreencrossshowstheBCG.Fromtoptobottom:u*,g’,r’bands.
5
Adamietal.:MUSEcharacterisationsofIntraClusterLightinaz∼0.53clusterofgalaxies
Fig.5.Sameasfig.4butfor(fromtoptobottom):i’,z’,K bands.
s
Wewillnowfitspectralenergydistribution(SEDhereafter) 3.3. SEDfitting
modelsonthediffuselight.
ThenextobviousstepistofitaSEDtothemeasuredICLmagni-
tudes.WeassumethattheICLsourceisrelatedtotheclusterand
6
Adamietal.:MUSEcharacterisationsofIntraClusterLightinaz∼0.53clusterofgalaxies
Table1.Magnitudes(notcorrectedforgalacticextinction,whichislowerthan0.1magnitudeinthelessfavourablecase)ofthedetectedICL
sourceandoftheBCGfromu*toK wavelengths.
s
u*AB g’AB r’AB i’AB z’AB K AB
s
ICL 23.58±0.16 21.51±0.14 20.03±0.09 19.40±0.02 19.27±0.03 18.66±0.05
BCG 23.339±0.033 22.429±0.011 21.025±0.006 20.160±0.004 19.823±0.006 17.94±0.05
therefore has a redshift of 0.534 (confirmed in the following). findaspectroscopicallyestimatedSFRof10.6±1.5M /yr.Ifwe
(cid:12)
WeusedtheLePhareSEDfittingtool(Arnoutsetal.1999,Ilbert usethe[OII]flux,theSFRestimateis32.3±0.75M /yr.
(cid:12)
etal.2006) andfixedtheredshift totheknownvalue.We used Given their different stellar population ages and SFRs, the
theCosmossurveytemplates(Ilbertetal.2009)toestimatethe ICLandBCGclearlyhavedifferentstellarpopulations.Thisis
closestgalaxytype,andtheBruzual&Charlottemplates(2003) illustratedinFig.7,wherewehavecomparedourSFRandstel-
to estimate the stellar population age, stellar mass, and the star lar mass estimates with known literature values from Puech et
formationrate(SFR).WealloweddiscretisedE(B-V)valuesbe- al.(2010:emissionlinegalaxies)andLiuetal.(2012:BCGs)at
tween 0 and 1.0 with a 0.025 step. Uncertainty on E(B-V) was similarredshifts.TheICLsourcehasastellarmasstypicalofthe
incorporatedintotheuncertaintiesgiveninTable2.Wealsore- emission line galaxy sample. The XLSSC 116 BCG has a stel-
ferthereadertoIlbertetal.(2010:theirAppendixDandsection larmassclosertotheliteratureBCGs.OurICLsourceseemsto
4.1) for details on how the mass-to-light ratios were taken into exhibitintermediatecharacteristicsthatarebetweenanoldand
accountinthemethod.Forreference,wealsogiveinTable2the passive elliptical galaxy and a younger and more active object.
resultingstellar-mass-to-lightratiosintheKsbandfortheBCG TheXLSSC116BCGdoesnotseemtobeanormalcDgalaxy
andtheICL.WenoteherethattheICLstellar-mass-to-lightra- witharelativelylowstellarmassandahighstarformationrate.
tioisonlyvalidforthetotalamountofICLandnotforitssep- The estimates presented in this section are, however, only
aratecomponents(seebelow).Wealsonotethatwewouldneed basedonmodelfitting.WenowdiscusstheMUSEICLspectra.
deeper Ks images in order to have a more precise value of the
Ks-band stellar-mass-to-light ratio for the ICL, given the very
lowsurfacebrightnessoftheICLfeatures(seee.g.fig.5). 4. MUSEdataanalysis
In Fig. 6 we show the best-fitting SED, and Table 2 gives
4.1. MUSEopticalspectroscopicdataandredshift
thecharacteristicsoftheICL.WeseethattheICLiswellmod-
measurement
elled by an early-type elliptical galaxy SED. We note that this
does not conflict with having ICL emission lines in addition to We were awarded four hours of MUSE Science
thisellipticalgalaxySEDbecausewidebandphotometryisonly Verification Time (# 60.A-9302) in order to observe
weaklysensitivetonarrowbandemissionlines. the cluster using integral-field spectroscopy. MUSE (see
TheageoftheICLstellarpopulationis2.3Gyrwithanun- http://www.eso.org/sci/facilities/develop/instruments/muse.html)
certaintyallowingagesbetween1and6Gyr.TheSFRispoorly isasecondgenerationinstrumentfortheVeryLargeTelescope
constrainedandislikelyoftheorderof5M(cid:12) /yr,butcouldbe (VLT) and is an integral-field spectrograph operating in the
aslowas0.07. visiblewavelengthrange.Weobtainedthedatainthewide-field
A common criticism of ICL searches is that it is often dif- mode with adaptative-optic mode off, simultaneously covering
ficulttodiscriminatebetweentheBCGhaloandtheICLitself. a 1×1 arcmin field from ∼480 to ∼930nm with a final spectral
However,inthepresentcasewenotethatgiventheICLexten- resolution of 1.25Å per pixel. The spatial resolution was of
sion,thereisnowaytohavesuchanextendedBCGhalo.Thus, the order of 0.2(cid:48)(cid:48) before convolution with the seeing, which
it is likely that the diffuse light is only mildly contaminated by variedbetween0.85and1.2(cid:48)(cid:48) duringtheJunerun,andbetween
theBCGhalo. 0.84 and 1.07(cid:48)(cid:48) during the August run. The observations were
We have previously carried out several simulations (pre- centredontheXLSSC116BCGandexecutedinservicemode
sented in Guennou et al. (2012) and Adami et al. (2013)) to be during the nights of 25-26 June and 20-22 August 2014. We
reasonablysurethatourdataarenotheavilypollutedbytheun- obtained four one-hour observing blocks, each consisting of
derlying BCG. However, we perform an additional test in this four exposures separated by 90◦ rotation, to average out the
paper by analysing the BCG in the same way as the ICL (see patterns of the slicers and channels seen by the detector. The
table 2 and Fig. 6). The flux from the BCG was considered in datawerereducedfollowingtherecipesoftheMUSEpipeline,
itsTERAPIX/CFHTLSellipse:3.5×1.4(cid:48)(cid:48) (∼23×10kpc)ellipse version 0.18.2. The June data were taken at a temperature
andwithatrigonometricorientationof311.7degrees.First,we below7◦,whichcausedproblemswiththecorrectidentification
notethattheBCGisnotpassive(besttype:Sd,highSFR).This of the 10th slice of IFU no. 6. To obviate the problem, we
isingoodagreementwiththegalaxybeingaradiosource(prob- followedtheworkaroundrecommendedintheMUSEcookbook
able) or a UV source (possible): NVSS J021039-055637 and and used dedicated trace tables. The reduction steps for each
GALEXASCJ021039.47-055645.5areclosetoitsposition(re- individual exposure included bias, flat-field, wavelength and
spectively at ∼2(cid:48)(cid:48) and 9(cid:48)(cid:48)). We also show the MUSE spectrum flux calibration, and correction for the telluric absorption lines
(described below) of the BCG in Fig. 6. The spectrum clearly using standard stars and geometric correction. Dedicated sky
showsstrongemissionlines.Thisobjectisclearlyformingstars observations were taken in an area adjacent to the cluster and
veryactively.WedonothaveaccesstotheHαlineintheMUSE were reduced following the same recipe and were subtracted
spectra, so we cannot directly measure the SFR from this line. from the corresponding science data. The 16 final cubes were
However, using the Moustakas et al. (2010) Sd galaxies (to be combined using relative RA and DEC offsets, and keeping
consistent with the best-fit type), we estimated an (Hα)/Hβ ra- the whole wavelength range. The spectra were extracted by
tioof4.85±0.7.MeasuringthefluxundertheHβlineinFig.6, summing all the pixels in several elliptical regions of different
translating it into an Hα flux, and using Kennicutt (1998), we extensions using ds9 (see Table 3 for a list of successful
7
Adamietal.:MUSEcharacterisationsofIntraClusterLightinaz∼0.53clusterofgalaxies
Table2.CharacteristicsoftheICLandoftheBCG(type,age,stellarmass,SFR,E(B-V)andstellar-mass-to-lightratiosintheKsband).Wegive
firsttheSED-estimatedSFR,andthenthespectroscopicallyestimatedSFR.
Galaxytype Age log10(StellarMass) log(SFR) E(B-V) logM/LK
s
109Gyr M M /yr M /L
(cid:12) (cid:12) (cid:12) (cid:12)
ICL Ell 2.3[1.1;5.9] 10.7[10.5;10.9] 0.7[-1.2;1.3]/0.11[0.04;0.18] 0.10±0.03 0.69[0.52;1.10]
BCG Sd 9.3[7.9;9.8] 10.9[10.4;11.2] 1.9[1.5;2.7]/1.03[0.96;1.08] 0.72±0.18 0.71[0.25;0.99]
Table3.J2000coordinates,redshifts,i’bandmagnitude,Serna-Gerbal Table4.CharacteristicsofSGdetectedsubstructures.Massestimates
substructurenumberoftheCFHTLSobjectswithasuccessfulMUSE arestatisticallyaffectedby4.61013M(cid:12)uncertainties.
redshiftmeasurement,andspectralqualityflag(seetext).
SGnumber Numberofgalaxies Vel.dispersion Mass
RA(2000) DEC(2000) z i’ SGsub. flag km/s 1013M
(cid:12)
32.6571 -5.9526 0.5363 21.06 1 3 1 20 570 3
32.6556 -5.9416 0.5328 20.31 1 4 2 4 170 0.4
32.6578 -5.9413 0.5308 23.43 1 2
32.6580 -5.9434 0.5263 25.91 1 2
32.6647 -5.9380 0.5299 21.69 1 3
32.6703 -5.9394 0.5315 21.59 1 3
32.6726 -5.9492 0.5334 22.31 1 2
32.6623 -5.9414 0.5394 20.27 2 4 centage of galaxies with a successful redshift measurement, is
32.6607 -5.9449 0.5382 21.40 2 4 90%downtoi’∼20.5and80%downtoi’∼22intheMUSEfield
32.6628 -5.9498 0.5345 23.71 1 2 ofview.
32.6636 -5.9490 0.5288 21.89 1 3
32.6677 -5.9495 0.5314 23.78 1 2
32.6657 -5.9481 0.5394 20.73 2 4 4.2. Structureofthecluster
32.6670 -5.9477 0.5294 22.89 1 2
32.6677 -5.9465 0.5391 23.01 2 2
Given the number of available spectroscopic redshifts, we can
32.6647 -5.9454 0.5327 21.91 1 2
investigatethepresenceofpossiblesubstructuresinthecluster.
32.6643 -5.9447 0.5344 22.31 1 3
To this end, we have applied the Serna-Gerbal (1996, hereafter
32.6655 -5.9433 0.5345 20.16 1 3
SG)hierarchicalmethodalreadyextensivelydescribedinseveral
32.6668 -5.9416 0.5323 21.89 1 3
32.6681 -5.9432 0.5328 21.93 1 2 papers (e.g. Guennou et al. 2014 and references therein). This
32.6677 -5.9424 0.5317 21.74 1 3 methodisquitepowerfulforshowingevidenceofsubstructure;
32.6697 -5.9426 0.5336 20.81 1 4 thecodealsoestimatesthemassofthesubstructures.Massesare
32.6707 -5.9424 0.5338 23.15 1 2 computedthroughabasicversionofthevirialtheorem(neglect-
ingelectromagneticfieldsandonlyusingclassicalestimatorsfor
thegalaxyvelocitydispersion).Theseestimatessufferfromrela-
tivelylargeuncertaintiesoftheorderof4.61013M(cid:12).Thisvalue
redshifts). Narrow- or broadband images can be obtained by
wasstatisticallyestimatedfromthedataofGuennouetal.(2014)
collapsingthefinalcubeintheselectedwavelengthrange.
by comparing the SG estimates and the masses deduced from
The redshifts were obtained by using the EZ redshift mea-
theGiodinietal.(2009)clusterscalingrelation.Guennouetal.
surement code (Garilli et al. 2010) on the final 1D spectra, al-
(2014),however,alsoshowedthatforagivenparent-clusterthe
lowing an additional smoothing of 3 pixels in order to find the
relativemassesbetweensubstructureswerestillreliable.
redshift value more easily when needed. The redshift measure-
mentsweredoneinthesamewayasfortheVIPERSsurvey(e.g. More precisely, the SG hierarchical method calculates the
Guzzoetal.2014)andaredshiftmeasurementqualityflagwas potential binding energy between pairs of galaxies and detects
assigned between 1 and 4. Flag 1 means that we have a 50% substructuresbytakingpositions,magnitudes,andredshiftsinto
chanceofhavingthecorrectredshiftestimate;flag2,75%;flag account. We required at least four galaxies in a given sub-
3,95%;andflag4,morethan99%.Weonlyconsideredobjects structure.TheSGmethoddetectstwosuchsubstructuresinthe
withflags2,3,and4assuccessfulmeasurements.Statistically, XLSSC 116 cluster (see table 4). The first substructure has 20
thismeansthatwemayhavethreeobjectswithanincorrectred- galaxies,anestimatedmassof3×1013M(cid:12),andavelocitydisper-
shiftinTable3. sionof570km/s.Thesecondsubstructureissmaller:4galaxies,
and the estimated mass and velocity dispersion are 4×1012 M
A detailed study of the individual galaxy spectra, and their (cid:12)
and 170 km/s. This second substructure probably corresponds
comparisonwithotherclustersdetectedintheXXLSurvey,will
to the low-temperature component detected in the X-ray data
beundertakeninafuturepaperfollowingthecompletionofthe
(thebrightestgalaxyofthissubstructureisveryclosetotheX-
surveyspectroscopicfollow-up.However,infig.8weshowthe
raypeakofthelow-temperaturecomponent;seefigure1)andis
meanandmedianspectraofallthegalaxies(excludingtheBCG)
at a greater distance (∼2000 km/s) (see blue solid histogram in
spectroscopically measured and found to be inside the cluster.
fig.9).
Eachspectrumwasputintherestframebasedonthemeasured
redshiftandwecomputedthemeanandmedianspectrum.This Galaxiesthataremembersofthetwodetectedsubstructures
givesameanandmedianviewoftheclustergalaxypopulation. are also well mixed inthe cluster red sequence (see fig. 3). We
This spectrum is nearly elliptical, with prominent absorption note that a velocity dispersion of 570 km/s in a relaxed clus-
lines (e.g. H&K), a strong Balmer break, but without any sig- ter would suggest an X-ray temperature on the order of 2 keV
nificant emission lines. We show later that the BCG has a very (Rosati et al., 2002), in good agreement with our X-ray mea-
differentspectrum.Thecompletenesslevel,intermsoftheper- surements.
8
Adamietal.:MUSEcharacterisationsofIntraClusterLightinaz∼0.53clusterofgalaxies
Fig.7. Comparison of our stellar mass and SFR estimates for the
XLSSC116ICLsource(redverticallineforSEDestimationandred
verticaldottedlineforspectroscopicestimation)andBCG(bluevertical
lineforSEDestimationandblueverticaldottedlineforspectroscopic
estimation)withthePuechetal.(2010)emissionlinegalaxies(contin-
uoushistograms)andtheLiuetal.(2012)BCGs(dashedhistograms).
Upperfigure:stellarmass,lowerfigure:SFR.
5. NatureoftheclusterdiffuselightwithMUSEdata
Thefirstquestionwewanttoanswerconcerningthediffuselight
inclustersiswhetherthereisanyindicationoflineemittinggas.
Doestheoldstellarpopulationofdiffuselight(aspreviouslyde-
tected)alsogiverisetointergalacticgasionisation?Orperhaps
somecollisionaleffectscanionisegas?
Thefirstwaytoansweristoreconstructnarrowbandimages
withspectraldatacentredonmajoremissionlinesatthecluster
Fig.6.Spectralenergydistributionfittedontheavailableu*,g’,r’,i’, redshift.GiventheavailableMUSEdata,webuilt[OII](5699-
z’,andKsmagnitudes(3σerrorbarsarealsoshown).Adoptedredshift 5729A),Hβ(7432-7471A),and[OIII](7655-7696A)images
and best-fit model are given in each case. Upper figure: diffuse light,
(lower and upper spectral bounds of these images were chosen
middle figure: BCG. The lower figure is the MUSE spectrum of the
approximatelyasthemeanlinewavelength±2timestheveloc-
BCG.
itydispersionofthemainclusterstructuredetectedbelow).This
gives evidence of a new diffuse light region completely invisi-
ble in any of the broadband images directly south of the BCG
9