APJ,INPRESS PreprinttypesetusingLATEXstyleemulateapjv.11/26/03 CHANDRATEMPERATUREMAPSFORGALAXYCLUSTERSWITHRADIOHALOS F.GOVONI1,2,M.MARKEVITCH3,A.VIKHLININ3,L.VANSPEYBROECK3,4,L.FERETTI1,ANDG.GIOVANNINI1,2 Full-resolutionversionathttp://hea-www.harvard.edu/ maxim/papers/rhalos_tmaps ∼ ApJ,inpress ABSTRACT WeanalyzeChandratemperaturemapsforasampleofclusterswithhigh-qualityradiohalodata,tostudythe originoftheradiohalos. ThesampleincludesA520,A665,A754,A773,A1914,A2163,A2218,A2319,and 1E0657–56. We presentnewtemperaturemapsforallbuttwo ofthem(A520andA754). Alltheseclusters exhibitdistortedX-raymorphologyandstronggastemperaturevariationsindicatingongoingmergers. Some 4 clusters, e.g., A520, A665, 1E0657–56, exhibitthe previouslyreportedspatialcorrelationbetweenthe radio 0 halo brightness and the hot gas regions. However, it is not a general feature. While most mergers are too 0 messytoallowustodisentangletheprojectioneffects,wefindclearcounterexamples(e.g.,A754andA773) 2 where the hottest gas regions do not exhibit radio emission at the present sensitivity level. This cannot be n explainedby projectioneffects, and thereforearguesagainstmergershocks— at least those relatively weak a onesresponsiblefortheobservedtemperaturestructureinmostclusters—asthemainmechanismforthehalo J generation. Thisleavesmerger-generatedturbulenceas a more likelymechanism. The two clusterswith the 1 clearestradiobrightness–temperaturecorrelation,A520and1E0657–56,arebothmergersinwhichasmall 2 densesubclusterhasjustpassedthroughthemain cluster, verylikely generatingturbulencein itswake. The maximumradiobrightnessandthehotgasarebothseeninthesewakeregions. Ontheotherhand,thehalos 1 in1E0657–56andA665(bothhigh-velocitymergers)extendintotheshockregionsinfrontofthesubclusters, v wherenostrongturbulenceisexpected.Thus,inhigh-velocity(M≃2- 3)mergers,bothshockandturbulence 1 accelerationmechanismsmaybesignificant. 2 4 Subject headings: galaxies: clusters: general — galaxies: clusters: individual (A520, A665, A754, A773, 1 A1914,A2163,A2218,A2319,1E0657–56)—intergalacticmedium—X-rays: galaxies: 0 clusters—Radiocontinuum 4 0 / 1. INTRODUCTION proposed (see, e.g., recent reviews by Enßlin 1999; Sarazin h 2001;Brunetti2002;Petrosian2002andreferencestherein). The baryonic content of the galaxy clusters is dominated p - by the hot(T ∼2- 10 keV) intergalacticgaswhose thermal Wewillbrieflyreviewtheoreticalargumentsin§4. o emission is observed in X-rays. Many clusters also exhibit On the observationalside, the numberof knownradio ha- r diffuseradiosourcesthathavenoapparentconnectiontoany los is already sufficientto start lookingfor correlationswith t other cluster properties in order to elucidate their accelera- s individualcluster galaxy. In this paper, we are interested in a tion mechanism. Halos are typically found in clusters with theextremelylow-brightness,large-scalesourcesclassifiedas : significant substructure in the X-ray brightness which indi- v radio halos. They often span the whole cluster and are un- catesmergeractivity(e.g.,Feretti1999;Buote2001). Halos i polarized. Other diffuse radio sources, not considered here, X are present in rich clusters, characterized by high X-ray lu- includestronglyelongatedrelics, whicharehighlypolarized r andoftenfoundattheclusterperiphery,andmini-haloswhich minositiesandtemperatures. Thepercentageofclusterswith a halosin a completeX-ray flux-limitedsample (thatincludes arefoundaroundthepowerfulcentralradiogalaxy;forrecent systemswithL >5×1044ergs- 1inthe0.1- 2.4keVband) reviewsseee.g. Kempneretal. (2003),Feretti(2003). X is≃5%. ThehalofractionincreaseswiththeX-rayluminos- While the cluster X-ray emission is due to thermal elec- ity,to≃33%forclusterswithL >1045ergs- 1(Giovannini, tronswithenergiesofseveralkeV,theradiohaloemissionat X ∼1GHzisproducedbysynchrotronradiationofrelativistic Feretti, & Govoni1999a). Theradiopowerofa halo, if one electronswith energiesof∼10GeVin magneticfieldswith ispresent,stronglycorrelateswiththeclusterluminosity,gas B≃0.5- 1µG.Theseelectronsshouldcoexistwiththether- temperature(e.g.,Liangetal.2000;Colafrancesco1999;Fer- etti1999),ortotalmass(Govonietal.2001a).Inanumberof mal population. Their origin is still uncertain; the difficulty well-resolvedclusters,aspatialcorrelationbetweentheradio in explaining their presence arises from the combination of thelargesizesofhalos(r∼1Mpc)andtheshortsynchrotron haloandX-raybrightnessisobserved(Govonietal.2001b). lifetimeoftheseelectrons(107- 108yrs).Oneneedsamecha- These observations indicate that radio halos are closely re- latedtotheintra-clusterthermalgas, itshistoryandenerget- nismbywhichtheseelectronsarelocallyandsimultaneously ics. However,detailsofthisconnectionneedfurtherinvesti- (re-)accelerated over the halo volume. Several such mecha- gations. nismsoffeedingenergytotherelativisticelectronshavebeen Because the halos appear to be related to cluster merg- 1IstitutodiRadioastronomiadelCNR,ViaGobetti101,I-40129Bologna, ers, cluster gas temperature maps, which contain informa- Italy tion on the merger geometry, stage, and velocity, can pro- 2 Dipartimento diAstronomia, Università diBologna, ViaRanzani1,I- vide further information on their origin. Indeed, Chandra 40127Bologna,Italy studiesofA2163,A665(Markevitch&Vikhlinin2001,here- 3 Harvard-Smithsonian Center for Astrophysics, 60 Garden St., Cam- after MV01), 1E0657–56 (Markevitch et al. 2002a, here- bridge,MA02138 after M02a) and A520 (Markevitch et al. 2002b, hereafter 4Wehonorthememoryofourlatecolleague 2 GOVONIETAL. warepackagefordataprocessing,XSPECandA.Vikhlinin’s TABLE1 ZHTOOLStofittheoverallspectraandanalyzeimages,and X-RAYDATA ourowncodetoderivethetemperaturemaps. Name z ObsID exposure, Observation TheX-rayobservationsareidentifiedbytheirIDnumbers ksec date (ObsID)inTable1,wherethecleanexposuretimeandtheob- servingdatearelisted. Allobservationswereperformedwith A520........... 0.1990 528 9.3 2000Oct10 theACIS-Idetectorwhichcoversafieldofviewof16′×16′. A665........... 0.1819 531 8.5 1999Dec29 ExceptforA1914,A2163,and 1E0657–56, forwhichwe 3586 23.8 2002Dec28 usethelongerofthetwoavailableexposures(ofverydifferent A754........... 0.0542 577 39.1 1999Oct30 length),whenmorethanoneobservationexistedforthesame A773........... 0.2170 533 10.4 2000Sep5 object,theobservationswerecombined. 3588 8.3 2003Jan25 A1914......... 0.1712 3593 17.6 2003Sep3 2.1. X-raydatapreparation A2163......... 0.2030 1653 69.6 2001Jun16 For each observation, standard screening was applied to A2218......... 0.1756 553 5.2 1999Oct19 thephotonlist. We removedbadpixelsandcolumns,events 1454 10.9 1999Oct19 withASCAgrades1,5and7,andcosmicrayafterglows. The 1666 22.8 2001Aug30 backgroundwasmodeledusingacompositeblank-fieldback- A2319......... 0.0557 3231 14.0 2002Mar15 ground dataset corresponding to the period of the observa- 1E0657–56..... 0.296a 3184 69.6 2002Jul12 tions,cleanedandnormalizedasdescribedbyMarkevitchet Col.1:Clustername; al.(2003b). Inparticular,wefirstexcludedflareperiods,ex- Col.2:Redshift(Struble&Rood1999;aTuckeretal.1998) tracting the X-ray light curves from the source-free parts of Col.3:ChandraobservationIDnumber; the field and requiring that the 0.3–12 keV backgroundrate Col.4:Cleanexposuretime; be within a factor of 1.2 of the nominal background. After Col.5:Observingdate. theflareexclusion,wenormalizedtheblank-skybackground datasettotheclusterobservationbytheratioofcountsathigh energies (in the PHA interval of 2500–3000ADU; approxi- mately10–12keV)whichisfreefromtheskyemission.This M02b) revealed a possible spatial correlation between the correctionwasalwayssmall, within10%oftheexposurera- high-temperaturecluster regions and the diffuse radio emis- tio. IfanobservationwasperformedinVFmode,andifthe sion. A665 showed a spatial correlation between the radio correspondingVF-modebackgrounddatasetexisted(ObsIDs halobrightnessandthelocationofapossiblebowshock,ap- 1666,1653,3184,3231,3586,3588,3593),additionalparti- pearing to support a merger shock origin for the relativistic clebackgroundreduction(Vikhlinin2001)wasappliedtothe haloelectrons. clusterandbackgrounddata. Clearly,anX-rayandradiostudyofagreaterclustersam- The soft Galactic background in the direction of the ana- pleisneededtoseewhetheraspatialcorrelationbetweenthe lyzedclusters(Snowdenetal.1997)isnotmuchdifferentthan radiohaloandthegastemperatureisacommonfeature. We intheregionsincludedintheblank-skybackground—sono presentChandraX-rayimagesandtemperaturemapsforsuch special treatmentwas required— for all clusters but A2163 asample. andA2319. ForA2163,treatmentofthesoftGalacticexcess We make the radio – X-ray comparison for the clusters is described in MV01; the central region of A2319 that we A520, A665, A754, A773, A1914, A2163, A2218, A2319, analyzehereissobrightthatsuchanexcessdoesnotmatter. and 1E0657–56. Chandramapsfor A754 and A520are re- Ingeneral,ourtemperaturemapsare limitedto thebrightest producedfromMarkevitchet al. (2003a)and M02b, respec- regionsoftheclusterswherethebackgroundaccuracyisnot tively. For A2163 and 1E0657–56, we present temperature critical. mapsfromnew observations,deeperthan those analyzedby MV01 and M02a; the full derivation of these maps will be 2.2. X-rayspectra presented by Markevitch et al. (in prep.) A higher-accuracy temperaturemap forA665is derivedin thisworkbyadding For consistency checks, we first derived the average tem- anewlongobservationtotheoneusedinMV01. ForA773, peraturesof ourclusters, extractingspectra fromcircularre- A1914,A2218,andA2319,wederivethefirsthigh-resolution gions of r=0.8 Mpc (or smaller if such a region did not fit temperaturemapshere. within the field of view). This radius encompasses most of Theradiodataweretakenfrompreviouslypublishedwork thecluster emission. Pointsourceswere excluded. The cor- (A2319: Feretti, Giovannini, & Böhringer 1997; A665, responding spectral Redistribution Matrix Files (RMF) and A2218: Giovannini&Feretti2000;1E0657–56: Liangetal. Auxiliary Responce Files (ARF) were generated using the 2000;A520,A773:Govonietal.2001a;A2163:Ferettietal. A. Vikhlinin’s software and were weighted by the cluster 2001;A754,A1914:Bacchietal.2003). brightness. The large quantum efficiency (QE) spatial non- WeuseH0=70kms- 1Mpc- 1,Ω0=0.3andΩΛ=0.7. uthneifAorCmISityfoccaaulspeldanbey-C1T1I0iCn)twheasdamtaodperiloedrtuosiFnegbth2e00f0or(mwuitlha 2. X-RAYANALYSIS from Vikhlinin (2000); for other data, the standard QE uni- formitymaps were used. ARFs for the ACIS-I observations WeusedarchivalandourproprietaryChandraACISdatato deriveX-rayimagesandgastemperaturemaps.ACIS5hasan alsoincludedthetime-andposition-independentfudgefactor energybandof0.3–9keVandcombines1′′angularresolution of0.93atE<1.8keVtoaccountforthebackside/frontside andamoderateenergyresolution. We usedthe CIAO6 soft- illuminatedCCD flux discrepancyin the currentcalibration. The time-dependent,position-independentcorrectionfor the 5http://asc.harvard.edu/cal soft absorption caused by contamination buildup (Plucinsky 6http://cxc.harvard.edu/ciao etal.2003)wasincludedintheARFs. TEMPERATUREMAPSFORRADIOHALOCLUSTERS 3 3.1. Abell520 TABLE2 FITSTOAVERAGECLUSTERSPECTRA Opticalspectroscopicdata onA520suggestthatthisclus- ter is undergoing strong dynamical evolution (Proust et al. Name T, abund. radius, NH, 2000). A short Chandra observation (Table 1) placed this keV ′′ 1020cm- 2 X-ray-luminouscluster in the ACIS chip I3. The X-ray ob- A520........... 7.1 0.7 0.24 0.12 245 7.80 servationconfirmsthatA520isinthemiddleofamerger. In ± ± A665........... 8.2 0.5 0.27 0.07 260 4.24 Fig. 1, we compare an X-ray image and a temperature map ± ± A754........... 10.0 0.3 0.30 0.05 540 4.36 (fromM02b)derivedfrom this Chandraobservation,a VLA ± ± A773........... 7.5 0.8 0.34 0.11 230 1.44 radiohaloimageat1.4GHzfromGovonietal.(2001a),and ± ± A1914......... 10.9 0.7 0.24 0.08 275 0.95 anopticalDigitizedSkySurvey(DSS)7plate. ± ± A2163......... 12.4 0.7 0.20 0.03 240 18.7 3.5 The most prominent feature in the X-ray data is a dense, ± ± ± A2218......... 6.7 0.5 0.10 0.07 270 3.24 compact—butclearlyextended—coolgasclumpsouthwest ± ± A2319......... 10.1 0.4 0.30 0.05 540 7.93 of the center. Apparently, it has just passed straight through ± ± 1E0657–56..... 13.9 0.7 0.25 0.05 180 4.60 themainclusterfromthenortheast,accompanyingapossible ± ± groupofgalaxies(althoughnotcenteredonanyonegalaxy). Col.1:Clustername; Thetemperaturemapshowsthatthebrightgastrailithasleft Col.2:Clustertemperature; behind is cool, and reveals a hot strip of apparently shock- Col.3:Abundancerelativetosolar; heated gas along this trail. Most of the radio halo emission Col.4:Radiusofthespectrumextractionregion appearsto follow this hotstrip, and notthe coolbrighttrail. Col.5:Galacticabsorptioncolumn(Dickey&Lockman1990), The X-ray image hints at a possible bow shock in front of exceptforA2163and1E0657–56,seetext. the dense clump, coincident with the southwest edge of the radio halo, but a longer observation (planned for late 2003) isneededtodeterminethenatureofthatfeature. Thiscluster Spectra were fit in the 0.8- 10 keV energy band with the is one of the best examples of a spatial correlation between the radio halo brightnessand the gas temperature. It is also MEKALmodel(Kaastra1992)byfixingtheabsorptioncol- oneofonlytwoclustersinoursamplewithasimpleandclear umn to the Galactic value (Dickey & Lockman 1990). Two mergergeometry(theotherbeing1E0657–56). exceptionsareA2163and1E0657–56whereitisknowntobe significantlydifferent,higherforA2163(e.g.,Elbaz,Arnaud, &Böhringer,1995)andlowerfor1E0657–56(e.g.,Lianget 3.2. Abell665 al.2000). ForA2163,itwasfitasafreeparameter,whilefor TheX-rayimageofA665stronglysuggeststhatthecluster 1E0657–56,theLiangetal.’sbest-fitROSATvaluewasused, isundergoingamerger(e.g.,Jones&Saunders1996;Gómez, whichisingoodagreementwithourdata. Hughes,&Birkinshaw2000). Themergerapparentlyoccurs Theresultingaveragetemperaturesandabundances,atthe in the direction of the elongation of the galaxy distribution 90%confidence,aregiveninTable2togetherwiththeextrac- (Geller&Beers1982;Beers&Tonry1986).Indeed,aChan- tion radii and absorptioncolumns. All the obtained temper- dratemperaturemaprevealedlargevariationsandapossible ature are in good agreement with previously derived values shockwithM≈2intheexpectedlocation(MV01). Alumi- foundintheliterature. nousradiohaloinA665wasdiscoveredbyMoffet&Birkin- shaw(1989),confirmedbyJones&Saunders(1996)andfur- 2.3. Temperaturemaps therstudiedbyGiovannini&Feretti(2000). Temperatureimagesoftheclustercentralregionswerede- TheclusterwasobservedbyChandratwice (seeTable1). rived as described in Markevitch et al. (2000a) and MV01. Here we derive a more accurate temperature map than that Specifically, we excludedpoint sources, extracted imagesin presentedinMV01,byincludingtherecentlongerexposure. severalenergybands,subtractedthebackground,dividedthe InFig.2,theresultingmapandanACISimagearecompared resulting images by the exposure maps, and smoothed them to the 1.4 GHz VLA image (Giovannini& Feretti 2000)and byaGaussianfilterwithvariablewidth(sameinallbands)to anopticalDSSplate. Thenewmapconfirmsthepresenceof get usefulstatistical accuracyover the interesting brightness hot—mostprobablyshock-heated—gassouthorinfrontof range. A temperatureineach pixelofthe mapwasobtained the coolcore. The new image, however, still does not show byfittingvaluesineachpixeloftheseimageswithathermal anycorrespondingdensityfeatureassharpasthebowshock plasma,fixingNH totheGalactic (orother,seeabove)value observed in 1E0657–56. The edge of the cool core is also andtheelementabundanceto0.3solar. Toverifytheresult- notassharpasinsomeotherclusters,sothemergerprobably ingtemperaturemaps,wefitspectrafromseveralinteresting proceedsatanangletotheskyplane. regionsofeachcluster using XSPEC. The outlyingareasof TheX-raybrightnesspeakisoffsetfromthecentralgalaxy, thetemperaturemaps,wherenoisestartstodominate(usually indicating gas motion in the core. Indeed, the new temper- wherethe1σerrorsaregreaterthan∼2keV),werecutoff. ature map reveals complex temperature structure inside the core as well as in its wake, includingtwo streams (or trails) 3. RESULTS ofcoolgasintheNWandNEdirectionsfromthecore,anda BelowwepresentourX-rayimagesandtemperaturemaps, hotstreamnorthofthecore. discusshowtheyelucidatethemergerstateandgeometry,and The radio halo is very extended and elongated in the SE- comparethemtotheradiohaloimages. Foreachcluster,we NWdirection,whichistheapparentmergerdirectionandthe present a temperature map with an X-ray and a radio con- X-rayelongation. AsnotedbyMV01,the“leadingedge”of tour plot overlaid, an optical image with an X-ray contour thehaloextendsbeyondthecoolcoreandcoincideswiththe plot overlaid, and an X-ray image with a radio contour plot overlaid. 7http://archive.stsci.edu/dss 4 GOVONIETAL. shock region. MV01 presented an image of the ratio of the collisionofthegascloudshasgeneratedashock-heatedarea radio brightness to the square root of the X-ray brightness, there,seeninthetemperaturemap. whichhastheapproximatemeaningofthedensityofrelativis- The radio halo does not follow the X-ray brightness, nor tic electrons to thermal electrons (assuming a uniformmag- temperaturedistribution–moreover,thehottestclusterregion neticfield). Itsmaximumcoincidedwiththesouthernshock is not at all bright in the radio (although a more sensitive, region. lower spatial resolution radio image shows a faint extension mostly southward; Govoni et al. 2001a). Instead, the radio 3.3. Abell754 haloiscenteredintherelativelycoolregionbetweenthetwo OpticalandX-raystudies(e.g.,Fabricantetal.1986;Bird galaxysubclusters. 1994; Slezak, Durret, & Gerbal 1994; Zabludoff & Zarit- 3.5. Abell1914 sky 1995; Henry & Briel 1995; Henriksen & Markevitch 1996;Blitonetal.1998;DeGrandi&Molendi2001)showed The presence of a radio halo in A1914 was suggested by that A754 is undergoinga violent merger. Markevitch et al. Giovanniniet al. (1999b)from an NVSS search. It was de- (2003a)notedthatitisdifficulttoexplainthecomplexdetails tectedbyKempner&Sarazin(2001)intheWesterborkNorth- revealed by the Chandra X-ray image and temperature map ernSkySurvey(WENSS)at0.3GHz,andconfirmedbyBac- (whichwereproducehereinFig.3)withasimpletwo-cluster chietal.(2003)indeepVLAobservationsat1.4GHz. mergermodel. Krivonosetal. (2003)noteda possibleweak A1914wasobservedbyChandratwice,butbecauseofthe bowshockeastofthebrightcoreintheROSATimage;itislo- relatively short useful exposure of the first observation(Ob- catedapproximatelyattheleftmostX-raybrightnesscontour sID 542), only the second one is used here. In Fig. 5, we inFig.3. comparethe radioimage fromBacchi et al. (2003)with our AradiohaloinA754wasdiscoveredbyHarris,Kapahi,& ChandraX-rayimageandtemperaturemap,andaDSSplate. Ekers(1980)andrecentlyconfirmedbyKassimetal.(2001) Thistemperaturemap, thefirstforA1914,immediatelypro- throughradio observation at 0.3 GHz. A hard X-ray excess vides a likely merger scenario, not at all obvious from the atE &45keVwithrespecttothermalemissionwasalsore- X-ray and optical images alone. In this scenario, the NE- ported (Fusco-Femianoet al. 2003). The most likely expla- SWarch-likehotregionthroughtheclustercenter,whichco- nationisinverseComptonemissionfromthesamerelativistic incides with a distinct component in the X-ray image, is a electronsresponsibleforthediffuseradioemission, shock between the two infalling subclusters. One of them InFig. 3, we comparetheChandraX-rayimageandtem- has arrived from the southeast, where the map shows cool perature map with a DSS optical plate and a 1.4 GHz VLA gas probablystripped from that subcluster. Its gas core was imagefromBacchietal. (2003). Theradiodiffuseemission partlyshockedandstoppedbythe collisionaroundthe posi- iscomplexandveryextended.Bacchietal.(2003)notedthat tionofthepresentX-raybrightnesspeak.Thisgasiscurrently itconsistsoftwolargecomponents,roughlycoincidentwith squirtingsidewaysalongthehotarch,creatingtheprominent the two main galaxy concentrations. The temperature map coolerelongationtotheeast. Itsgalaxiesappeartohavepen- shows strong spatial variations, such as a large hot area in etratedfurther,insidetheshock-heatedgas,andarenowseen the south and southwestand coolgasat the tip of the bright asa northeasterngalaxyconcentration. Thebrightestcluster tongue-likestructure.Thisstructureappearstobetheremains galaxywasprobablyatthecenteroftheothercollidingsub- ofthecoreofoneofthesubclusters,presentlyflowinginthe cluster,whicharrivedfromthewest. Itscoolercorewasleft northeastdirection. Theeasterndiffuseradioemissionislo- behindandiscurrentlyseenasawesterncoolextension,some catedaroundtheeasterninterfacebetweenthetongueandthe ofitpossiblysquirtingalongthewesternsideoftheshocked surroundinggas. Thewesternradioemissionextendsbeyond region, similarly to the situation on the southeastern side of the boundaryofour temperaturemap. Thatregionwas cov- thecluster. Ifthismergerhadasmallerimpactparameterand eredbytheASCAmap(Henriksen&Markevitch1996)which occurred closer to the sky plane, we would now see a clas- showedhotgas;moredetailswillsoonbeprovidedbyXMM sical picture from the simulations with two subclusters just (P.Henry&A.Finoguenov,privatecommunication).Signifi- beforecorepassageandashockedregionbetweenthem(e.g., cantly,thelarge,hotsouthernregioninthepresentmapdoes Schindler&Müller1993andmanylaterworks). notexhibitradioemissionatthepresentsensitivitylevel. ThediffuseradioemissioninA1914isunusualinthatithas a distinct, bright, elongatedregion, approximatelyalong the 3.4. Abell773 presumedpathtraveledbytheeasternsubcluster.Itisaccom- An irregular X-ray shape of A773 was noticed in the paniedbya moretypical, diffuse,low-brightnesshalointhe ROSAT data (e.g., Pierre & Starck 1998; Rizza et al. 1998; clustercenter. Thebrightfeatureisclearlyextended(Bacchi Govoni et al. 2001a). A radio halo in this cluster was sug- etal.2003).Itwouldbeinterestingtodetermineifthisregion gested by Giovannini, Tordi, & Feretti (1999b) from the isphysicallydistinctfromtherestofthehalo.Anupperlimit NRAOVLASkySurvey(NVSS,Condonetal.1998)andcon- onitspolarizationis3%(Bacchietal.2003),whichexcludes firmedbyGovonietal.(2001a)throughadedicatedVLAob- thepossibilitythatitisarelicprojectedontotheclustercen- servation at 1.4 GHz. In Fig. 4, we compare this radio im- ter. The remaining low-brightness region of the radio halo agewithanopticalDSSplate, andourtemperaturemapand approximatelycoincideswiththehotcentralregionoftheX- Chandraimage.ThisisthefirsttemperaturemapforA773.It rayclusterandmayevenfollowthepresumedstreamsofthe revealsstrongtemperaturevariationsinthe6- 12keVrange. gas of the two subclusters. However, a detailed comparison The optical image clearly shows two galaxy subclusters, requiresbetteraccuracyandaremovaloftheradiosourcesin one in the center of the X-ray emission and another at the thesouthandtheresolutionofthenatureofthebrightfeature. easternoutskirt. Takingintoaccountthetemperaturemap,it 3.6. Abell2163 appearsasthoughttheeasterngalaxygroupiscurrentlyexit- ing the merger site, having shed its gas due to ram pressure A2163isamongthehottestandmostX-rayluminousclus- at the entry into the main cluster at its southwest side. The ters known (e.g., Arnaud et al. 1992). The presence of an TEMPERATUREMAPSFORRADIOHALOCLUSTERS 5 extended,powerfulradiohaloinA2163wasfirstreportedby Trèvese,Cirimele,&DeSimone2000)suggestedthatitcon- Herbig & Birkinshaw (1994). The ROSAT image (e.g., El- sistsoftwocomponentssuperimposedalongthelineofsight, bazetal.1995),crudeASCAtemperaturemap(Markevitchet with a subcluster around the second-brightest galaxy pro- al. 1994)and the previouslypublishedChandratemperature jected about 10′ northwest of the cD galaxy (seen in Fig. map(MV01)clearlyshowthattheclusterisamerger. How- 8a). Thatsubclusterisseen asa coolX-rayextensioninthe ever,evenwiththetemperaturemap,itsgeometryisdifficult ROSAT imageandthecrudeASCAtemperaturemap(Marke- todetermine. vitch1996). A2319exhibitsanextendedandpowerfulradio InFig.6, wepresentanewtemperaturemapderivedfrom halo(Harris&Miley1978)withanirregularmorphologywell a deeper Chandra exposure by Markevitch et al. (in prepa- correlatedwiththeX-raybrightness(Ferettietal.1997). ration). It is in agreement with but more accurate than the WeuseaChandraobservationofA2319toderivethefirst MV01map.TheX-rayimageandthenewmapshowstreams detailedtemperaturemapofthiscluster,showninFig.8along ofhotandcoldgasflowingindifferentdirections,aswellas withourX-rayimageanda1.4GHzVLAimagefromFeretti apossibleremnantofacoolgascore. Itstilldoesnotclarify et al. (1997). The most prominentfeature seen in the X-ray the geometry of the merger, however. This, the absence of data is a sharp cold front southeast of the cD galaxy, such anysharpfeaturesinthehigh-statisticsX-rayimage,andthe asthosediscoveredbyChandrainmanyothermergingclus- optical spectroscopic data (Soucail et al., in preparation) all ters (e.g., Markevitchet al. 2000a; Vikhlinin, Markevitch & pointtothepossibilitythatthemergerisoccurringatalarge Murray2001). The centralcoolgascloud is clearlymoving angletotheskyplane. southeast with respect to the ambient hotter gas (or, equiva- ThetemperaturemapiscomparedwithanopticalDSSplate lently, the hotter gas is flowing around it to the northwest). andthe radiohalo imagefromFeretti etal. (2001). Thelat- ThecDgalaxyisneitheratthecentroidnoratthecoolestspot ter combines two images with different FWHM resolutions, ofthecluster,suggestingthatthecoolgascoreismovingin- 30′′ (innerblackcontours)andthemoresensitive 45′′×60′′ dependentlyofthisgalaxy.ThecDgalaxyappearstohaveits (outerbluecontours),respectively,toshowthelargeextentof owngasdensitypeakonasmallerlinearscale,itselfdisplaced theradiohalo. Thehaloextension2- 3′ eastofcentercoin- eastwardfromthegalaxy. (Wenotethatthisbrightnesspeak cideswiththehottestregionofthecluster;however,asnoted fallsexactlyontheACIS-Icentrallow-exposurespot,sothe inMV01,thisregionalsocoincideswiththeX-raybrightness apparent temperature dip there may be an artifact.) We can extension. In general, the halo brightnessfollows the distri- alsoseeacoolarmextendingaroundtheclustercenterfrom butionoftheX-raybrightnessquitewell(Ferrettietal.2001). thetip ofthecoldfrontinthegeneraldirectionofthenorth- western subcluster. It may either be a tail of that subcluster, 3.7. Abell2218 orgasstrippedfromthecoldfrontcloud,unrelatedtothesub- InA2218,strongandweaklensingmassreconstructionre- cluster. Overall,thepicturesuggestsalaterstageofamerger, vealedtwodistinctmasspeaksinsidetheclustercore,approx- wellpasttheinitialencounter. imately around the two brightest galaxies along the NW-SE Theradiohalo followsremarkablycloselythe distribution directionseen in the opticalplate Fig. 7a (e.g., Abdelsalam, ofthecoolgasinthecore,exceptfortwolow-brightnessex- Saha,&Williams1998;Squiresetal.1996;Kneibetal.1995, tensionsintothehottergasnortheastandsouthwestofthecold 1996).ROSATimagessuggestedthattheclusterisnotrelaxed front.TheradiohaloismoreextendedsouthwestoftheX-ray (e.g.,Kneibetal.1995;Squiresetal.1996;Markevitch1997), brightnesscentroid—towardthecoolergasthatweobserve althoughatlargescales,itappearedrelativelysymmetric. there. TheclusterwasobservedwithChandrathreetimes. Anal- ysis of the first two exposureswas presented by Markevitch 3.9. 1E0657–56 etal.(2000b)whofoundacentralpeakintheradialtemper- 1E0657–56 is one of the hottest, most luminous clusters atureprofile,andMachaceketal.(2002)whoreportedanaz- known(Tuckeretal.1998)whichalsocontainsthemostlumi- imuthallyasymmetrictemperaturestructureinthecore. Here nousradiohalo(Liangetal.2000). AnearlyChandraobser- weaddamorerecent,longerexposureandderivethefirstde- vationrevealedaspectacularbowshockpropagatinginfront tailedtemperaturemapofA2218. ItisshowninFig.7along of a dense, cool bullet-like subcluster exiting the site of the withtheX-rayimage,a1.4GHzVLAimagefromGiovannini collisionwithabiggersubcluster(M02a). Thoseauthorsde- &Feretti(2000),andanopticalDSSplate. rivedanapproximatetemperaturemapofthismergerandno- ThemapconfirmstheresultsfromtheearlierChandraanal- ticedthattheradiohalobrightnesspeaksatthehottestcluster ysesandrevealsstrongasymmetrictemperaturevariationsin region. the5- 10keVrange. Togetherwiththerelativelysymmetric In Fig. 9, we compare the Liang et al. (2000) 1.3 GHz X-rayimage,suchanirregularbutcentrallypeakedtempera- radioimagewith a new, moreaccuratetemperaturemapob- turemapsuggeststhatA2218isatalatermergerstage,when tainedfromadeeperChandraexposurebyMarkevitchetal. the violentgasmotionsare starting to subside, as seen, e.g., (inpreparation). The newmap andthe X-rayimagesuggest inthesimulationsbyRoettiger,Stone,&Mushotzky(1998). thatthesmallersubclusterhasarrivedatthecollisionsitefrom The relatively small radio halo in A2218 (Giovannini & thesoutheast. Mostofitsoutergaswasshockedandstripped Feretti2000)isslightly(byabout40′′)offsetfromtheX-ray during the collision with a bigger subcluster (whose largest brightnesspeakand doesnotshowa particularlystrongcor- galaxies are now seen in the east). This stripped gas, to- relationwiththetemperaturemap—infact,thehottestspot gether with the shocked gas from the other subcluster, form in the cluster is notseen in the radio, as in A754and A773. the north-south bar-like structure seen in the X-ray image, Theradioimagehasarelativelylowsensitivity,however. whichisprobablyapancakeinprojection. Shockscouldnot penetrateand stop the dense core of the subcluster, and it is 3.8. Abell2319 now continuing to the west, preceded by a bow shock with Optical analyses of the bright nearby cluster A2319 (e.g., M≈3.5. Thissubclustershouldbegeneratingvigoroustur- Faber & Dressler 1977, Oegerle, Hill, & Fitchett 1995, bulenceinitswake. 6 GOVONIETAL. TheradiohalopeakisclearlyoffsetfromtheX-raybright- strongprojectioneffects,ofcourse). nesspeak(intheregionthatexcludesthebullet),andinstead On the other hand, the gas turbulence, although not di- is centered in the hottest cluster region. The halo’s eastern rectly observable at present, is expected to exist throughout part is elongated along the presumed infall trajectory of the the merging clusters, including shock-heated and cooler gas subcluster. Interestingly,thewesternpartofthehaloextends regions(e.g.,Sunyaev,Norman&Bryan2003). Thereforein allthewaytothebowshock. theturbulencescenario,thereshouldbenostrongcorrelation betweentheradiobrightnessandthetemperature. 4. DISCUSSIONANDCONCLUSIONS Theanalysispresentedhereisa qualitativeattempttodis- SubstructureintheX-rayimagesandthegalaxyspatialand tinguishbetweentheshockandturbulenceaccelerationmech- velocity distributions, as well as complex gas temperature anisms by means of a systematic comparison of radio halo structure, are signatures of cluster mergers. All these prop- mapswiththetemperaturemapsforasampleofhaloclusters ertiesarecommoninclusterscontainingradiohalos. Onthe withgoodradioandX-raydata. otherhand,wedonotknowofextendedradiohalosinrelaxed Allclustersstudiedhererevealclearsignsofongoingmerg- clusters.Therefore,theavailabledatasuggestthatradiohalos ers and the accompanying strong spatial temperature varia- arerelatedtomergers. tions. Althoughinmostoftheseclusters,themergergeome- From the energetic grounds, mergerscan indeed dissipate tryisambiguousandthelikelyprojectioneffectscomplicate enoughkineticenergy—simultaneouslyoveramegaparsec- comparisonwiththeradiodata,wecandrawseveralconclu- scale volume — for the maintenance of a radio halo. How- sions. ever, it is not clear how exactly the relativistic particles are We confirmthe previouslyreportedspatial coincidenceof accelerated. In-situ acceleration (or re-acceleration) of rela- bright radio features with the high temperature regions in tivistic electrons during a merger can occur in shocks (e.g. A665, A2163, and 1E0657–56, which was based on the Sarazin1999,Fujita&Sarazin2001)orinthegasturbulence shorter X-ray observations. 1E0657–56 and A520 are the (e.g. Schlickeiser,Sievers,&Thiemann1987;Brunettietal. bestexamplesofthistemperature-radioconnection.Thisspa- 2001;Ohno,Takizawa,&ShibataS.2002,Fujita,Takizawa, tial“coincidence”isnotquitea“correlation’—theseclusters &Sarazin2003). alsoexhibitbrightradioemissionfromsomeofthe coolgas Therearetheoreticalargumentsagainsttheshockhypothe- regions, but in the presense of projection effects, exact cor- sis. Mostimportantly,arelativelystrongshockwithM>4- 5 relationisnotexpected. AsdiscussedbyMV01,theleading is believed to be necessary for generation of an observable edgeofthe“limb-brightened”(Jones&Saunders1996)halo halo (e.g., Brunetti 2002). Such high Mach numbersshould inA665coincideswiththeapparentlocationofapossiblerel- be very rare, as most merger shocks should have M ∼1 at ativelystrongshockaheadofthefast-movingcoolcore. This the clustercenter. Gabici& Blasi (2003)arguedthatshocks observationappeartoarguefortheshockaccelerationmech- expected in mergersof clusters with comparablemasses are anism. too weak to result in significant non-thermal emission (the On the other hand, in A2319, which is at a very similar expected Mach number increases with the mass ratio of the merger stage to A665, we do not detect a particularly hot, merging subclusters). Moreover, according to Miniati et al. shock-heatedregion in front of the moving cool core, so its (2001),suchradioemissionwouldlookmorelikeradiorelics Machnumbershouldbelow. Yet,theclusterexhibitsaradio at the cluster peripheryratherthan as radiohalos. However, halowhose brightestpartfollowsratherclosely the distribu- physics of collisionless shocks in clusters is not well under- tionofthecoolergasinthecore(althoughatalowerbright- stood, and so at present even relatively weak mergershocks ness level, the halo’s NE-SW extension appears to coincide cannotbecompletelyruledoutasanaccelerationmechanism. withtheoutlyinghotterregions). Indeed, comparisonof the radio halo and gas temperature Moreover, in A754 and A773, the hottest cluster regions mapsforafewmergingclusters(MV01,M02a,M02b)hinted donotshowradioemission,atleastatthepresentsensitivity, attheirspatialcorrelation,whichcouldbeeasilyexplainedif whilethereisradioemissionelsewhereintheseclusters. The electrons were accelerated in shocks. A currently propagat- same appearstrue for the smallhalo in A2218, althoughthe ingorjust-passedmergershockwouldleaveclearimprintin sensitivity of the radio data here is poor. While radio emis- the cluster gastemperaturemaps in the form of hotregions. sionfromcoolerareascanbeexplainedbyprojectioneffects, Aftertheshockpassage,regionsofshock-heatedgasquickly given that the X-ray-derivedprojected temperatureis biased expandadiabaticallyandcomeintopressureequilibriumwith towardthe denser, coolergasonthe line of sight, projection thesurroundings. Large-scalegasmotionsduringthemerger cannot explain the absence of the radio emission from the subsequently mix gases of different temperature, resulting hottest cluster regions. We believe that this is a strong ob- in patchytemperaturestructure (as seen in simulations, e.g., servationalargumentagainst merger shocks, at least the rel- Roettiger, Stone, & Burns(1999),Takizawa 2000,Ricker & ativelyweakonesexpectedinmostmergers,asthemainac- Sarazin2001)whichpersistsforaconsiderabletimeintheab- celerationmechanism. sence ofthermalconduction(Markevitchet al. 2003a). Any This leaves turbulence as a more viable mechanism. In- relativistic electronsacceleratedas the shockpasses through deed,allclustersinoursamplewherethemergergeometryis thegas,willbepreventedfromdiffusingfarfromtheirorigin tractable—1E0657–56,A520,A2319,andA665—exhibit bythemagneticfields(thesamefieldsthatsuppressthermal relativelysmall, dense, coolmovingcloudswhich are likely conductionanddiffusionofthermalelectrons,e.g.,Vikhlinin togenerateturbulenceintheirwake. Theradiohalosareob- etal.2001;Markevitchetal.2003a),andwillfollowthebulk served along the path of these moving clouds. In A754, the motionoftheirhostgas. Thus,ifrelativisticelectronsareac- easternhaloliesalongthebrightcoolX-raytongue,whichis celeratedbymergershocks,andassumingauniformmagnetic likely to be the flow of dense gaspossibly generatingturbu- field (or, at least, a field uncorrelated with the gas tempera- lenceatitsinterface.InA773,thepassageoftheeasternmass ture),oneexpectsstrongspatialcorrelationbetweentheradio peakthroughthe gasofthemainsubclustermayalsogener- halobrightnessandthehottestgasregions(intheabsenceof ateturbulenceinthebrightradioarea.Theremainingclusters TEMPERATUREMAPSFORRADIOHALOCLUSTERS 7 havesufficientlyuncertainmergergeometriestobeconsistent radio halo emission would be invaluable for pinpointingthe withthepossibilityofstrongturbulenceintherightregions. electronaccelerationsites (because the halo spectrum steep- However,withturbulencealone,itisdifficulttoexplainthe ensastherelativisticelectronsloseenergy).Suchdatawould observedextensionoftheradiohaloin1E0657–56aheadof be especially illuminating in clusters such as 1E0657–56, thegasbulletallthewaytothebowshock,asimilarextension A665 (Feretti et al., in preparation), and A520, where both intotheshockregionaheadofthecoreinA665,andpossibly turbulenceandshocksarelikelytobepresentandcanbesepa- inA520.Turbulencecausedbythesecorescannotprecedethe ratedspatially.Also,fortheseclusters,higher-resolutionhalo fast-moving core, especially in such a high-velocity merger imagesarerequiredtodetermineifthefrontedgesofthehalos as 1E0657–56. Thus, in systems like these, at least some indeedcoincidewiththeshockfronts. of the relativistic electrons should be accelerated in shocks. Incidentally,1E0657–56hasashockwiththehighest-known Machnumber(M≈3.5,M02a)andA665appearstocontain FGthanksthehospitalityoftheHarvard-SmithsonianCen- arelativelyhigh-Mshockaswell(M≈2,MV01). terforAstrophysicswheremostofthisworkwasdone. Sup- Weconcludethatinmostclusters,theradiohaloelectrons portwasprovidedbyNASAcontractNAS8-39073,Chandra areprobablyacceleratedbyturbulence,butinthoserarecases grantsGO2-3164XandGO2-3165X,andtheSmithsonianIn- when shocks with M ≃2- 3 are present, these shocks also stitution. We thank M. Murgia for the use of the Synage++ appear to contribute in the acceleration. In this regard, we program. notethatKempner&David(2003)performedasimilarradio– X-ray comparisonfor anotherhalo cluster, A2744, and con- NOTE ADDED IN PROOF: After this paper was submit- cluded that both turbulence and shock acceleration may be ted, an independentlyderivedChandra temperaturemap for presentinthatcluster. A2319 has been published by O’Hara, Mohr, & Guerrero Inconclusion,wenotethatmapsofthespectralindexofthe (2004).Itisingoodagreementwithourmap. REFERENCES Abdelsalam,H.M.,Saha,P.,&Williams,L.L.R.,1998,AJ,116,1541 Henriksen,M.J.,&MarkevitchM.L.,1996,ApJ,466,L79 ArnaudM.,HughesJ.P.,FormanW.,JonesC.,Lachieze-ReyM.,Yamashita HenryJ.P.,&BrielU.G.,1995,ApJ,443,L9 K.,&HatsukadeI.,1992,ApJ390,345 HerbigT.,&BirkinshawM.,1994,BAAS,26,1403 Bacchi,M.,Feretti,L.,Giovannini,G.,&Govoni,F.,2003,A&A,400,465 Kaastra,J.S.,1992,“AnX-RaySpectralCodeforOpticallyThinPlasmas” Beers,T.C.,&Tonry,J.L.,1986,ApJ,300,557 (InternalSRON-LeidenReport,updatedversion2.0) BirdC.M.,1994,AJ107,1637 Kassim,N.E.,Clarke,T.E.,Enßlin,T.A.,Cohen,A.S.,&Neumann,D.M., BlitonM.,RizzaE.,BurnsJ.O.,OwenF.N.,&LedlowM.J.,1998,MNRAS 2001,ApJ,559,785 301,609 Kempner, J.C.,Blanton, E.L.,Clarke, T.E.,Enßlin, T.,Johnston-Hollitt, Brunetti,G.,Setti,G.,Feretti,L.,&Giovannini,G.,2001,MNRAS,320,365 M.,&Rudnick,L.,2003,Proc.“TheRiddleofCoolingFlowsinGalaxies Brunetti G., 2002, Proc. “Matter and Energy in Clusters of Galaxies”, andClustersofGalaxies”,eds.T.H.Reiprich,J.C.Kempner,&N.Soker Taiwan, ASP Conf. Series, eds. S. Bowyer & C.-Y. Hwang, in press (astro-ph/0310263) (astro-ph/0208074) Kempner,J.C.,&David,L.P.,2003,MNRAS,submitted(astro-ph/0310185) Buote,D.A.,2001,ApJ,553,L15 KempnerJ.C.,SarazinC.L.,2001,ApJ,548,639 Colafrancesco S.,1999,Proc.RingbergworkshoponDiffuseThermaland Krivonos,R.A.,Vikhlinin, A.A.,Markevitch, M.L.,&Pavlinsky,M.N., Relativistic PlasmainGalaxyClusters,ed.H.Böhringer,L.Feretti,&P. 2003,AstronomyLetters,29,425 Schuecker,MPEreport271,269 KneibJ.P.,MellierY.,PelloR.,Miralda-EscudeJ.,LeBorgneJ.-F.,Böhringer Condon,J.J.,etal.1998,AJ115,1693 H.,&PicatJ.-P.,1995,A&A,303,27 DeGrandiS.,&MolendiS.,2001,ApJ551,153 Kneib,J.P.,Ellis,R.S.,Smail,I.,Couch,W.J.,&Sharples,R.M.1996,ApJ, DickeyJ.M.,&LockmanF.J.,1990,ARA&A,28,215 471,643 ElbazD.,ArnaudM.,&BöhringerH.,1995,A&A293,337 Liang,H.,Hunstead,R.W.,Birkinshaw,M.,&Andreani,P.,2000,ApJ544, EnßlinT.A.,1999,inTheUniverseatlowradiofrequencies,Symp.IAU199, 686 Pune(India),inpress,(astro-ph/0001433) Machacek, M.E.,Bautz,M.W.,Canizares, C.,&Garmire,G.P.,2002,ApJ Faber,S.M. &Dressler,A.,1977,AJ,82,187 567,188 Fabricant D.,Beers T.C.,Geller M.J.,Gorenstein P.,HuchraJ.P.,&Kurtz Markevitch,M.,Yamashita,K.,Furuzawa,A.,&Tawara,Y.,1994,ApJ,436, M.J.,1986,ApJ308,530 L71 FerettiL.,GiovanniniG.,&BöhringerH.,1997,NewAstronomy,2,501 Markevitch,M.,1996,ApJ,465,L1 FerettiL.,1999,inTheUniverseatlowradiofrequencies,Symp.IAU199, Markevitch,M.,1997,ApJ,483,L17 Pune(India),inpress(astro-ph/0006379) Markevitch,M.,etal.,2000a,ApJ,541,542 FerettiL.,Fusco-FemianoR.,GiovanniniG.,&GovoniF.,2001,A&A,373, Markevitch, M., Vikhlinin, A., Mazzotta, P., & VanSpeybroeck, L., 106 2000b, Proc. “X-ray astronomy 2000”, Palermo, September 2000, eds. Feretti L., 2003, XXI Texas Symposium on Relativistic Astrophysics , R. Giacconi, L. Stella, S. Serio, ASP Conf. Ser., 234, CDROM Florence,Italy(astro-ph/0309221) (astro-ph/0012215) Fujita,Y.&Sarazin,C.L.,2001,ApJ,563,660 Markevitch,M.,&Vikhlinin,A.,2001,ApJ563,95(MV01) Fujita,Y.,Takizawa,M.,&Sarazin,C.L.,2003,ApJ,584,190 Markevitch, M., Gonzalez, A. H., David, L., Vikhlinin, A., Murray, S., Fusco-Femiano, R.,Orlandini, M.,DeGrandi, S.,Molendi, S.,Feretti, L., Forman,W.,Jones,C.,&Tucker,W.,2002a,ApJ567,L27(M02a) Giovannini,G.,Bacchi,M.,&Govoni,F.,2003,A&A398,441 Markevitch,M.,Vikhlinin,A.,&Forman,W.R.,2002b,Proc.“Matterand Gabici,S.&Blasi,P.,2003,ApJ,583,695 EnergyinClustersofGalaxies”,Taiwan,ASPConf.Series,eds.S.Bowyer Geller,M.J.,&Beers,T.C.,1982,PASP,94,421 &C.-Y.Hwang,301,37(M02b),(astro-ph/0208208) Giovannini,G.,Feretti,L.,&Govoni,F.,1999a,inTheUniverseatlowradio Markevitch,M.,etal.,2003a,ApJ,586,L19 frequencies,Symp.IAU199,Pune,inpress,(astro-ph/0006380) Markevitch,M.,etal.,2003b,ApJ583,70 Giovannini,G.,Tordi,M.,&Feretti,L.,1999b,NewAst.4,141 Miniati,F.,Jones,T.W.,Kang,H.,&Ryu,D.2001,ApJ,562,233 Giovannini,G.,&Feretti,L.,2000,NewAstronomy,5,335 Moffet,A.T.&Birkinshaw,M.,1989,AJ,98,1148 Gómez,P.L.,Hughes,J.P.,&Birkinshaw,M.,2000,ApJ,540,726 Oegerle,W.R.,Hill,J.M.,&Fitchett,M.J.,1995,AJ,110,32 Govoni, F., Feretti, L., Giovannini, G., Böhringer, H., Reiprich, T.H., & O’Hara, T. B., Mohr, J. J., & Guerrero, M. A. 2004, ApJ, submitted Murgia,M.,2001a,A&A,376,803 (astro-ph/0311402) Govoni,F.,Enßlin,T.A.,Feretti,L.,&Giovannini,G.,2001b,A&A,369, Ohno,H.,Takizawa,M.,&Shibata,S.2002,ApJ,577,658 441 Petrosian,V.,2002,Proc.“MatterandEnergyinClustersofGalaxies”,ASP Harris,D.E.&Miley,G.K.,1978,A&AS,34,117 Conf.Series,eds.S.Bowyer&C.-Y.Hwang,inpress,(astro-ph/0207481) Harris,D.E.,Kapahi,V.K.,&Ekers,R.D.,1980,A&AS,39,215 Pierre,M.,&Starck,J.L.,1998,A&A330,801 8 GOVONIETAL. Plucinsky,P.P.,etal.,2003,Proc.SPIE,4851,89 SquiresG.,etal.,1996,ApJ,461,572 Proust,D.,etal.,2000,A&A355,443 Struble,M.F.,&Rood,H.,1999,ApJ,125,35 Jones,M.,&Saunders,R.,1996,inRöntgenstrahlungfromtheUniverse,ed. Sunyaev,R.A.,Norman,M.L.,&Bryan,G.L.,2003,AstronomyLetters,in H.U.Zimmermannetal.(MPEReport263),553 press,(astro-ph/0310041) Ricker,P.M.&Sarazin,C.L.,2001,ApJ,561,621 Takizawa,M.,2000,ApJ,532,183 Rizza,E.,Burns,J.O.,Ledlow,M.J.,Owen,F.N.,Voges,W.,&Bliton,M., Trèvese,D.,Cirimele,G.,&DeSimone,M.,2000,ApJ,530,680 1998,MNRAS,301,328 Tucker,W.,etal.,1998,ApJ,496,L5 Roettiger,K.,Stone,J.M.,&Mushotzky,R.F.,1998,ApJ493,62 Vikhlinin, A., 2000, Chandra calibration memo, Roettiger,K.,Stone,J.M.,&Burns,J.O.1999,ApJ,518,594 http://asc.harvard.edu/cal/Links/Acis/acis/Cal- prods/qe/12- 01- 00 Sarazin,C.L.,1999,ApJ,520,529 Vikhlinin, A., 2001, Chandra calibration memo, Sarazin,C.L.,2001,in“MergingProcessesinClustersofGalaxies”,edited http://asc.harvard.edu/cal/,”ACIS”,“ACISBackground”,“VFmode” byL.Feretti,I.M.Gioia,andG.Giovannini,inpress,(astro-ph/0105418) Vikhlinin,A.,Markevitch,M.,&Murray,S.S.,2001,ApJ,551,160 Schlickeiser,R.,Sievers,A.,&Thiemann,H.,1987,A&A,182,21 Zabludoff,A.I.,&Zaritsky,D.,1995,ApJ,447,L21 Schindler,S.,&Müller,E.,1993,A&A,272,137 Slezak,E.,Durret,F.,&Gerbal,D.,1994,AJ,108,1996 Snowden,S.L.,etal.,1997,ApJ,485,125 TEMPERATUREMAPSFORRADIOHALOCLUSTERS 9 FIG.1.—A520.(a):X-raycontoursoverlaidontheopticalDSSimage.The0.8- 4keVX-rayimageisadaptivelysmoothed;contoursarespacedbyafactor of√2.(b):X-raycontourplotoverlaidonthetemperaturemap(colors).(c):Theisocontourmapat1.4GHzofthecentralregionofA520overlaidontheX-ray image(colors).TheradioimagehasaFWHMof15′′ 15′′.Thecontourlevelsare:0.06,0.12,0.24,0.48,0.96,1.92,3.84,7.68mJy/beam.(d):Radiocontours overlaidonthetemperaturemap(colors);crossesmar×ksomeradiosourcesunrelatedtothehaloemission. 10 GOVONIETAL. FIG.2.—A665.(a):X-raycontoursoverlaidontheopticalDSSimage.The0.8- 4keVX-rayimageisadaptivelysmoothed;contoursarespacedbyafactor of√2.(b):X-raycontourplotoverlaidonthetemperaturemap(colors).(c):Theisocontourmapat1.4GHzofthecentralregionofA665overlaidontheX-ray image(colors). TheradioimagehasaFWHMof52′′ 42′′. Thecontourlevelsare: 0.20.40.81.5361225mJy/beam. (d):Radiocontoursoverlaidonthe temperaturemap(colors);crossesmarksomeradiosou×rcesunrelatedtothehaloemission.