ebook img

The role of AGN in the colour transformation of galaxies at redshifts z~1 PDF

0.73 MB·English
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview The role of AGN in the colour transformation of galaxies at redshifts z~1

Mon.Not.R.Astron.Soc.000,000–000(0000) Printed2February2008 (MNLATEXstylefilev2.2) The role of AGN in the colour transformation of galaxies at redshifts z ≈ 1 A. Georgakakis1⋆, K. Nandra1, R. Yan2, S. P. Willner3, J. M. Lotz4,5, C. M. Pierce4, 8 0 M. C. Cooper2,6, E. S. Laird1, D. C. Koo7, P. Barmby3, J. A. Newman8, J. R. Primack9 0 6 2 A. L. Coil n a 1AstrophysicsGroup,BlackettLaboratory,ImperialCollege,PrinceConsortRd,LondonSW72BZ,UK J 2DepartmentofAstronomy,UniversityofCalifornia,Berkeley,CA94720 3Harvard-SmithsonianCenterforAstrophysics,60GardenStreet,MailStop65,Cambridge,MA02138,USA 4 4DepartmentofPhysics,UniversityofCalifornia,SantaCruz,1156HighStreet,CA95064,USA 1 5NationalOpticalAstronomicalObservatories,950N.CherryAvenue,Tucson,AZ85719,USA 6StewardObservatory,UniversityofArizona,933N.CherryAve.,Tucson,AZ85721-0065,USA ] h 7UCO/LickObservatoryandDepartmentofAstronomy&Astrophysics,UniversityofCalifornia,SantaCruz,CA95064,USA p 8LawrenceBerkeleyNationalLaboratory,Berkeley,CA94720,USA - 9DepartmentofPhysics,UniversityofCalifornia,SantaCruz,1156HighSt.,SantaCruz,CA95064,USA o r t s a 2February2008 [ 1 v ABSTRACT 0 6 WeexploretheroleofAGNinestablishingand/ormaintainingthebimodalcolourdistri- 1 butionofgalaxiesbyquenchingtheirstar-formationandhence,causingtheirtransitionfrom 2 the blue to the red cloud. Important tests for this scenario include (i) the X-ray properties . 1 of galaxies in the transition zone between the two clouds and (ii) the incidence of AGN in 0 post-starbursts, i.e. systems observed shortly after (< 1Gyr) the termination of their star- 8 formation.WeperformthesetestsbycombiningdeepChandraobservationswithmultiwave- 0 lengthdata fromthe AEGISsurvey.StackingtheX-rayphotonsatthe positionsof galaxies : v (0.4 < z < 0.9)notindividuallydetectedatX-raywavelengthssuggestsapopulationofob- i scuredAGNamongsourcesinthetransitionzoneandintheredcloud.TheirmeanX-rayand X mid-IR propertiesare consistent with moderately obscured low-luminosity AGN, Compton r thicksourcesoramixofboth.Morphologiesshowthatmajormergersareunlikelyto drive a theevolutionofthispopulationbutminorinteractionsmayplayarole.Theincidenceofob- scured AGN in the red cloud (both directdetections and stacking results) suggests that BH accretionoutlivestheterminationofthestar-formation.Thisisalsosupportedbyourfinding thatpost-starburstgalaxiesatz ≈ 0.8andAGNareassociated,inagreementwithrecentre- sultsatlow-z.Alargefractionofpost-starburstsandredcloudgalaxiesshowevidenceforat leastmoderatelevelsofAGNobscuration.ThisimpliesthatifAGNoutflowscausethecolour transformationofgalaxies,thensomenucleargasanddustcloudseitherremainunaffectedor relaxtothecentralgalaxyregionsafterquenchingtheirstar-formation. Key words: Surveys– galaxies: active – galaxies:bulges – galaxies: evolution– galaxies: interactions–cosmology:observations 1 INTRODUCTION massesofthesemonstersaretightlycorrelatedtothestellarveloc- itydispersionofthehostgalaxybulges(e.g.Ferrareseetal.2000; Amajorrecentdevelopmentinextragalacticastronomyisthedis- Gebhardtetal.2000),suggestingthattheformationandevolution covery that most of the spheroids inthe local Universe contain a of spheroids and the build-up of the super-massive BHs at their super-massive black hole (BH; e.g. Magorrian et al. 1998). The centresareinterconnected.Thisinterplaymayholdthekeyforun- derstandingsomeoftheobservedtrendsingalaxyevolution. ⋆ MarieCuriefellow Oneofthefundamentalpropertiesofgalaxiesistheirbimodal 2 Georgakakiset al. distributioninrest-framecolour.Thisbimodalityisobservedtobe- et al. 2007). The red optical colours of these systems are most yond redshift z 1 and is believed to hold important clues on likelybecauseofoldstarsandnotdustystar-formation.Although ≈ galaxyassembly(e.g.Stratevaetal.2001;Baldryetal.2004;Bell thereareexamplesofdeeplyburiedAGNinstar-forminggalaxies et al. 2004; Weiner et al. 2005; Cirasuolo et al. 2007). Evolved (e.g.Genzeletal.1998;CidFernandesetal.2001;Franceschiniet spheroidal galaxies are found in the “red-cloud” of the colour- al.2003;Georgakakisetal.2004;Alexanderetal.2005),thefre- magnitudediagram(CMD),star-formingsystemsdefinethe“blue quencyofthesesystemsamongtheobscuredX-raypopulationat cloud”, while the region between these overdensities is sparsely z 1isnotyetclear.Forexample,Rovilosetal.(2007)arguethat ≈ populated. Thisbimodal distributionislikelytobe thecombined anon-negligiblefractionofX-rayselectedAGNatz 1haveµJy ≈ resultofevolutioneffectsandexternalfactorssuchasinteractions radio continuum emission (1.4GHz) that is consistent with star- and mergers (e.g. Bell et al. 2004, 2006a, 2006b; Blanton 2006). formationinthehostgalaxy.Theseauthorshowever,donotfinda In the simplest interpretation, the CMD is consistent with a pic- strongtrendbetweenX-rayobscurationandtheincidenceoffaint turewherethegalaxystar-formationistruncated,byamechanism radioemission,likelytobeassociatedwithstarbursts. that remains to be identified, leading to the ageing of the stellar Aplausibleinterpretationoftheevidenceaboveisthatcurrent populationandtherapidtransitionfromthebluetotheredcloud. X-raysurveysarebiasedagainsttheearlyphaseofBHevolution. AGNareproposedtoplayamajorpartinthisprocessbyregulating Atthisinitialstageoftheirformation,AGNmaybedeeplyburied thestar-formationingalaxiescausingtheirtransitionincolouror under star-forming clouds and/or low luminosity because theBH maintainingthemintheredcloud. massisstillsmall,althoughgrowingrapidly.Ifthereissuchapop- Modelling work indeed suggests that the energy released by ulationofobscuredand/orintrinsicallyfaintyoungAGNbelowthe AGN is sufficient toeither heat up or blow away the cold gas of detection threshold of current surveys, it may be possible to find galaxies (e.g. Silk & Rees 1998; Fabian 1999; King 2003), irre- them using stacking analysis. Alternatively, AGN may not be re- versiblyalteringtheirevolution.Twomainroutesareproposedby sponsibleforquenchingthestar-formationingalaxiesandproduc- which AGN feedback affects the host galaxy. The first one op- ingthebimodalityof theCMD.A keytestofthisscenarioisthe erates during the luminous and high accretion rate stage of BH incidenceofAGNinpost-starburstgalaxies,i.e.systemsobserved growth, which issuggested tooccur during major mergers(Hop- shortlyaftertheterminationof their star-formation(< 1Gyr).In kinsetal.2005,2006a,b).Thesecatastrophiceventsalsoproduce thelocalUniverse(z . 0.1),anumber ofstudiespointtoanas- nuclear starbursts that obscure the central engine for most of its sociationbetweenopticallyselectedAGNandpost-starbursts(e.g. active lifetime. When the AGN becomes sufficiently luminous it Kauffmannetal.2004;Goto2006;Yanetal.2006).Atz 1,close ≈ drivesoutflows,whichsweepawaythenucleargasanddustclouds, tothepeakoftheAGNdensityintheUniverse(e.g.Bargeretal. therebyquenchingthenuclearstar-formation.Followingthisstage, 2005;Hasingeretal.2005),thereisstillverylimitedinformation theAGNalsodeclinesastheBHrunsoutofaccretingmaterialand onthelinkbetweenX-rayselectedAGNandpost-starbursts. eventuallyswitchesoff. We address the issues above using data from the All- Inadditiontothecoldgasaccretionmodeabove,simulations wavelength Extended Groth strip International Survey (AEGIS; also include a second AGN feedback recipe, which isinvoked to Davisetal.2007).X-raystackingisemployedtosearchfordeeply suppress the accretion of hot gasonto the central galaxy of large buriedand/orlow-luminosityAGNamongopticallyselectedgalax- darkmatterhaloes(e.g.Crotonetal.2006; Okamotoetal.2007; iesintheredshiftinterval0.4<z <0.9toexploretheroleofBH Cattaneoetal.2007).Thismodeoperatesinsystemswherethesu- accretion in establishing the bimodal colour distribution of these permassiveBHshavealreadybeenbuiltthroughmergersandarein systems. We also study the X-ray properties of post-starbursts as placeatthecentralgalaxyregions.Inthispicture,asgalaxiesenter identified from Keck spectra at z 0.8 to investigate the link ≈ themassivedarkmatterhaloesofgroupsorclusters,theircoldgas between recent star-formation events and active BHs. We adopt supplyiscutoffleadingtopassiveevolutionandtheirmigrationto H0=70kms−1Mpc−1,ΩM =0.3andΩΛ =0.7. theredcloud(Dekel&Birnboim2006;Crotonetal.2006).How- ever, cooling flows in these dense environments could produce a reservoirofcoldgasinthecentralgalaxyregions,therawmaterial 2 MULTIWAVELENGTHOBSERVATIONS for starburst and QSO activity. Low-level BH accretion is there- fore invoked to counterbalance cooling flows by producing low- Details about the AEGIS datasets can be found in Davis et al. luminosityAGNthatheatthebulkofthecoolinggasandthereby (2007).Inthispaperweuse(i)thedeepChandraobservationsto suppressinganysubsequentstar-formationinthesegalaxies. exploretheX-raypropertiesofopticalgalaxieswithspectroscopic InthetwoAGNfeedbackprescriptionsabove,theoutflowsce- redshiftsfromtheDEEP2survey,(ii)theSpitzerIRACandMIPS nario during the luminous and high accretion rate phase of BH 24µmdatatoselectsubsamplesbasedonmid-IRcolour,and(iii) growth, assigns AGN a central role in establishing the bimodal thehighresolutionHST/ACSimagingformorphologicalstudies. colour distributionof galaxiesby directly causing their transition TheX-raydataarefromtheChandrasurveyoftheExtended fromthebluetotheredcloud.ThedistributionofX-rayselected GrothStrip(EGS).Theobservationsconsistof8ACIS-Ipointings, AGNintheCMDappearstobebroadlyconsistentwiththispicture. each witha total integration timeof about 200kssplit in at least AGNarepreferentiallyassociatedwithgalaxiesintheredcloud,at 3shorter exposures obtained at different epochs. Thedatareduc- the red limit of the blue cloud and in the transition zone in be- tion, source detection and flux estimation are described in detail tween.TheseassociationsaresuggestiveofAGN-driventruncation byNandraetal.(2007,inpreparation)andarebasedonmethods ofthestar-formation(Nandraetal.2007).However,moredetailed presentedbyNandraetal.(2005).Briefly,standardreductionsteps studyofthepropertiesofAGNrevealanumberofinconsistencies aretakenusingtheCIAOversion3.2dataanalysissoftware.After withthemergerscenario.ObscuredAGNarepredominantlyhosted mergingtheindividualobservationsintoasingleeventfile,wecon- bygalaxiesintheredcloud(Nandraetal.2007;Rovilos&Geor- structedimagesinfourenergybands0.5-7.0keV(full),0.5-2.0keV gantopoulos2007),withprominentbulgesandlittlemorphological (soft), 2.0-7.0keV (hard) and 4.0-7.0keV (ultra-hard). The count evidence for ongoing major mergers (Grogin et al. 2003; Pierce rates in the above energy intervals are converted to fluxes in the TheroleofAGNinthecolourtransformationof galaxiesatredshiftsz ≈ 1 3 Figure2.Mid-IRcolour-colourplot.Forclarityweusedifferentpanels.Left:redcloudgalaxiesintheredshiftrange0.4 0.9.Middle:bluecloudgalaxiesin − thesameredshiftinterval.Inbothpanelsweonlyplotsourcesdetectedinall4IRAC/MIPSbandsusedtoestimatecolours.Right:expectedtracksofdifferent galaxytypesintherangez = 0 1usingobservedmid-IRSEDsfromtheSINGSdescribedbyDaleetal.(2007).NGC1097Sy1:long-dasheddotted; − NGC5408Im:short-dasheddotted;ScspiralgalaxyNGC628:dotted;SabspiralNGC4450:continuous;ellipticalgalaxyNGC584:longdashed.Thedots markthepositionofz=0andthecrossescorrespondtoz=0.4. The DEEP2redshift survey uses the DEIMOS spectrograph (Faberetal.2003)onthe10mKeck-IItelescopetoobtainredshifts forgalaxiestoR = 24.1mag.Theobservational setupusesa AB moderately high resolution grating (R 5000), which provides a velocity accuracy of 30kms−1 and a≈wavelength coverage of 6500–9100A˚.Thisspectralwindowallowstheidentificationofthe strong [OII] doublet 3727A˚ emission line to z < 1.4. We use DEEP2galaxieswithredshiftdeterminationssecureatthe>90% confidencelevel(qualityflagQ>3;Davisetal.2007). The Spitzer IRAC and MIPS-24µm observations cover the central 10 120arcmin2 subregion of the EGS (Barmby et al. × 2006). The integration time is about 2.7hr and 1200s per point- ing for the IRAC and MIPS observations, respectively. The 5σ flux density limit for point sources is 0.9, 5.8 and 83µJy at 3.6, 8.0 and 24µm, respectively. There are about 57400, 13600 and 6300detectionstothelimitsabove.Thesesourcesarematchedto theDEEP2photometriccataloguebyfindingthenearestneighbour withinasearchradiusof1.0arcsec. HSTimagesoftheEGSweretakenwiththeAdvancedCam- era for Surveys (ACS) in the V (F606W, 2260s) and I (F814W, 2100s)filtersovera10.1 70.5arcmin2 strip(Lotzetal.2007). × Figure1. Rest-frameU−BcolouragainstB-bandabsolutemagnitudefor The5sigmalimitingmagnitudesforapointsourceareVF606W = DEEP2galaxies(smallbluedots)intherange0.4 < z < 0.9estimated 28.14(AB)andIF814W =27.52(AB).Forextendedsourcesthese byWillmeretal.(2006).Post-starburstgalaxiesareshownwiththe(red) limitsareabout2magbrighter.Atotalof15797galaxieswerede- triangles.ThecontinuouslineisdefinedbyWillmeretal.(2006)toseparate theredfromtheblueclouds. tectedto IF814W = 25mag. These werematched to theDEEP2 photometriccatalogueusingan1.0arcsecmatchingradius. standardbands0.5-10,0.5-2,2-10and5-10keV,respectively.The limitingfluxineachofthesebandsisestimatedtobe3.5 10−16, 1.1 10−16,8.2 10−16,and1.4 10−15ergs−1cm−2×,respec- 3 SAMPLESELECTION × × × tively.TheX-raycataloguecomprisesatotalof1318sourcesover Themainsample usedinthispaper consistsof opticallyselected 0.63deg2toaPoissondetectionprobabilitythresholdof4 10−6. galaxies with DEEP2 spectroscopy and redshifts in the interval × Atz 1the2-10keVlimitabovecorrespondstoanX-raylumi- 0.4<z <0.9.Thisredshiftrangeistominimisecolourdependent nosity≈ofabout3 1042ergs−1(Γ=1.4),i.e.typicalofSeyferts. biasesintheselectionofgalaxiesintroducedbythemagnitudelimit × For theoptical identification weuse theDEEP2photometric cat- oftheDEEP2spectroscopicsurvey,R < 24.1mag.Redcloud AB alogues and the Likelihood Ratio(LR) method (e.g. Brusa et al. sourcesdropbelowthesurveylimitatlowerredshiftsthanintrin- 2007).Atotalof903sourceshavecounterpartstoR =25mag sicallybluergalaxies(e.g.Gerkeetal.2007).Thiseffectbecomes AB (LR>0.7),theapproximatelimitoftheDEEP2photometricsur- increasinglysevereatz&1astheR-bandstraddlestherest-frame vey. UV.Atz = 0.9thesampleiscompletetoM = 20mag.The B − 4 Georgakakiset al. lowerredshiftcut,z = 0.4,istoavoidbiasesassociatedwiththe Table1.FractionofX-raydetectionsindifferentsubsamples smallvolumesampledbytheAEGISbelowthislimit. There are4814 sources inthe maingalaxy sample of which Sample NTOT NX fX about70percentoverlapwiththeSpitzerIRACandMIPSsurveys (%) oftheAEGISand30percenthaveI-bandmorphologicalinforma- (1) (2) (3) (4) tion fromthe HST/ACSobservations. The morphological sample red0.4<z<0.7 548 36 6.5±1.1 wascompiledusing thecriteriaofLotzetal.(2007) byselecting red0.7<z<0.9 490 20 4.1±1.0 redpost-starburst 34 5 14.7±7.0 galaxiesbrighter thanIF814W = 25mag withaveragesignal-to- noiseratioperpixel>2.5andpetrosianradius>0.3arcsec.About valley0.4<z<0.7 105 4 3.8±1.9 valley0.7<z<0.9 126 8 6.3±2.3 65 per cent of the sources detected on theHST/ACSimages ful- valleypost-starburst 9 1 11±11 filthesecriteria.Almostallofthem(97percent)overlapwiththe IRAC/MIPSarea. blue0.4<z<0.7 2007 24 1.2±0.2 blue0.7<z<0.9 1769 28 1.5±0.3 TheCMDofthemaingalaxysampleispresentedinFigure1. bluepost-starburst 10 2 20±15 TheblueandredcloudsaredefinedusingtherelationofWillmer etal.(2006).GalaxieswithU Bcolours 0.05aboutthisrela- Thecolumnsare:(1):Sampledefinition.Notethatthereisoverlapbetweentheredorthebluecloud − ± andthevalley.Theredshiftinterval0.7 <z < 0.9iscomparableintermsofselectioneffects tiondefinethevalleybetweenthetwoclouds.Willmeretal.(2006) topost-starburstgalaxies;(2):NTOT isthetotalnumberofDEEP2galaxiesinthesample;(3): findthattheredshiftevolutionoftheredgalaxypopulationU B NXcorrespondstothenumberofDEEP2galaxieswithX-raycounterparts;(4):fXisthefraction colour is not strong in DEEP2.As a result a line with fixed−nor- ofX-raydetectionsinthesample,i.e.fX =NX/NTOT.Theerrorsareestimatedassuming Poissonstatistics. malisationisadequatefordefiningtheblueandredcloudsoverthe redshiftintervalofoursample.Thenextsectionsdiscusstheprop- ertiesofgalaxiesinnarrow slicesoftheCMD.Thesearedefined torunparalleltotheWillmeretal.(2006)lineseparatingthetwo cloudsandareparametrisedbytheirdistanceinU Bcolour,∆C, − fromthatline. The post-starburst sub-sample was drawn from the main galaxysamplefollowingthemethoddescribedbyYanetal.(2006, 2007 in prep.). The continuum-subtracted DEEP2optical spectra werefitwithacombinationofanold(K-component)andayoung (A-component) stellar population SED. Post-starburst candidates are defined as systems with f > 0.25, where f is the frac- A A tion of light contributed by the young stellar component around 4500A˚ inthelineardecompositionofthespectrum.Theequivalent widthofHβ,aftertheremovalofthestellarabsorption,wasthen usedtoidentifyresidual/ongoingstar-formationactivitywithinthe sample. Yan et al. (2006) showed that, for high redshift galax- ies, this line is a more reliable star-formation indicator than the [OII]3727line,whichisoftenassociatedwithAGNactivity.Post- starbursts(K+As)aredefinedassystemswithlittleHβ emission, EW(Hβ) < 5f 1 (total of 44). Because of the wavelength A − coverageoftheDEEP2spectroscopy, post-starburstscanbeiden- tified only in the redshift slice 0.68 0.88. In this interval, the − DEEP2spectralwindow includestheHβ lineaswellastherest- framewavelengthrange 3900 . λ . 4900A˚, whichisessential Figure3. Meanhardnessratioforgalaxiesintheblue/redcloudsandthe forthedecompositionofthespectrumintoanoldandayoungstel- valley.Withineachofthese3groups,opencirclescorrespondtoallsources lar components. Figure1 shows theposition of post-starbursts in inthegroup,squaresarefor24µmbrightgalaxies,andtrianglesrepresent theCMD.Mostofthemareintheredcloud. 24µmfaintsystems.Thehardnessratiouncertaintiesareestimatedassum- Themid-IRSEDofgalaxiesresultsfromacombinationofcir- ingPoissonstatistics.Inthecaseofnodetectioninthehardband,the3σ upperlimitinthehardnessratioisplotted.Thehorizontallinescorrespond rus radiation, stellar emission, dusty star-formation, and possibly totheexpectedhardnessratioofapower-lawX-rayspectrumwithspectral hot dust associated withanAGN.Samplesselectedat 24µm are indexfromtoptobottomΓ = 1.0,1.4and1.9,respectively. Despitethe generallydominated bydust enshrouded galaxies,bothAGNand errorbars,thereisevidenceforhardeningofthemeanX-rayspectrumfrom starbursts(e.g.Yanetal.2004).Inthispaper,wewillusethe24µm thebluetotheredcloud.24µmbrightsourceshaveharderspectraineach emission to identify such systems and to study them separately group. fromquiescentgalaxies.Figure2presentstheSpitzerIRAC/MIPS coloursofDEEP2/AEGISsources.Alsoshowninthisplotarethe colour tracks for different galaxy types using a selection of ob- servedSEDsfromtheSINGSprogram(Daleetal.2007).Undoubt- edlytherearelargeintrinsicvariationsinthemid-IRpropertiesof mid-IR colours (log(f24/f8.0) < 0.4) or no detection at 24µm galaxies of a given type resulting in a range of mid-IR colours. definesthe“24µm–faint”sample.Thelattersampleislikelytoin- Nevertheless,thegeneraltrendinFigure2isforquiescentsystems clude“24µm–bright”sourcesthatarebelowthedetectionlimitof to occupy the lower part of the plot and star-formation or AGN the24µm observations. Asexpected most of the24µm selected activitytoproducereddermid-IRcolours.Inthispaperwedefine sourceslieabovethelog(f24/f8.0) > 0.4cutinFigure2.Quies- “24µm–bright”sourcesasthosewithlog(f24/f8.0)>0.4.Bluer centgalaxiesareneverthelessalsopresentinthatsample. TheroleofAGNinthecolourtransformationof galaxiesatredshiftsz ≈ 1 5 Figure4. Meanhardnessratioestimatedbystackingopticalgalaxieswithin Figure 5. Hardness ratio against redshift. The curves are the expected differentslicesoftheCMD.∆Cisdefinedasthedifferencebetweenthe tracks for the ARP220 dusty starburst (Ptak et al. 2003), the Compton colourofthegalaxyandthelineseparating thebluefromtheredclouds thickAGNNGC1068(Mattetal.1999)andNGC6240(Ptaketal.2003) (Willmeretal.2006;seeFig.1).Thecolourslicesareparalleltothisline and a simple absorbed power-law model with Γ = 1.9 and NH = andtheirpositionintheCMDdependsonMB.Theverticaldottedlines 2×1022cm−2.Theshort-dashed line corresponds topower-law X-ray definethevalleybetweentheredandtheblueclouds.Thehorizontaldashed spectrumwithΓ=1.9,typicalofunabsorbedAGN.Thefilled(red)circle linesarethesameasinFigure3.ThelineofΓ = 1.4correspondstothe corresponds tored24µm-bright galaxies with∆C > 0.15.Theopen − meanhardnessratiooftheX-raydetectedAGN.24µm-brightand24µm- (blue)circleisforblue24µm-brightgalaxieswith∆C< 0.15. − faintsystemsareshownwiththesquaresandthetrianglesrespectively.The horizontal error bar ofeach point corresponds to the width of the CMD slicewithinwhichgalaxiesarestacked.Thehardnessratiouncertaintiesare whereT andBarethetotal(source+background)andbackground estimatedassumingPoissonstatistics.Inthecaseofnodetectioninthehard counts respectively. The 2arcsec radius includes only a fraction band,the3σ upperlimitinthehardnessratioisplotted(arrowspointing of the source photons at each position. We account for the re- downward). maining flux when estimating count-rates and fluxes by applying ameanaperturecorrectiondeterminedbyaveragingtheexposure- timeweightedPSFcorrectionsforindividualsources. 4 THEX-RAYSTACKINGMETHOD Themean X-ray properties of sources which arenot individually detectedtothelimitoftheX-raysurveyareexploredusingstack- inganalysis(e.g.Nandraetal.2002; Lairdetal.2005). Weused 5 RESULTS afixedradiusaperturetoextractandtosumtheX-rayphotonsat 5.1 X-raypropertiesofgalaxiesintheredcloudandinthe the positions of optically selected galaxies. Sources that are sep- transitionzone arated from an X-ray detection by less than 1.5 times the local 90per cent EncircledEnergyFraction(EEF)radius areexcluded Table1summarisesthefractionofX-raydetectionsamonggalaxy fromthestacking.Thisistoavoidcontaminationofthestackedsig- samplesselectedatseveralpositionsintheCMD.TheX-rayiden- nalfromphotonsassociatedwiththePointSpreadFunction(PSF) tificationrateishigher intheredcloud, inagreement withrecent wingsofX-raydetections.Also,inordertoexcluderegionswhere studiesonthehostgalaxypropertiesofX-rayselectedAGN(Nan- theChandraPSFislargeandthesensitivityislow,wedonotex- draetal.2007). tract photons from pointings where a particular source is located Inthissection,wesearchforevidenceofobscuredand/orlow- morethan9arcminawayfromthecentreofthedetector.Thisoff- luminositysystemsbelowtheX-raydetectionthresholdbystack- axisanglecutoffischosentomaximisethestackedsignal,although ingopticallyselectedgalaxiesindifferentregionsoftheCMD.Of ourresultsarenotparticularlysensitivetothisparameter.Forthe particularinterestisthevalley,whereAGNaresuggested toplay extractionradius,weexperimentedwithaperturesintherange1– animportantroleingalaxyevolution.Theresultsfordifferentsub- 3arcsecandselected2arcsecastheoptimalradiusthatmaximises samples are summarised in Table 2 and are plotted in Figure 3. thesignal-to-noiseratioofthestackedX-rayphotons. Thereissomeevidenceforaprogressivehardeningofthestacked Toassessthesignificanceofthestackedsignal,weestimated signalfromthebluetotheredcloud.Thisisfurtherdemonstrated the local background of a particular source by averaging the X- in Figure 4 which plots the hardness ratio of galaxies in CMD rayphotonswithina50arcsecradius(100pixels)andthenscaling sliceswhichrunparalleltotheWillmeretal.(2006)lineseparating totheareaoftheextractionaperture.Whendeterminingthelocal thebluefromtheredclouds.Althoughtheerrorbarsofindividual backgroundweclippedregionsaroundX-raydetectionsusingara- pointsarelarge,thereisasystematictrendwherebythemeanX- dius 1.5 times larger than the 90 per cent EEF. The significance ray spectrum of 24µm-bright sources becomes harder, reaching (in sigma) of the stacked signal is estimated by (T B)/√B, Γ 1.2, for ∆C > 0.15, i.e. for sources around the valley − ≈ − 6 Georgakakiset al. iesbluerthan∆C < 0.15isconsistentwithstar-formation(i.e. − X-raybinariesandhotgas),althoughwecannotexcludethepossi- bilityoflow-luminosityunobscuredAGN. TheobscuredAGNinterpretationforthe24µm-brightpopu- lationwith∆C > 0.15isalsosupportedbyFigure6,whichplots − the 2-10keV X-ray luminosity against the 24µm luminosity, in comparisonwiththerelationbetweenthesetwoquantitiesforstar- forminggalaxiesadaptedfromRanallietal.(2003).Sourceswith ∆C > 0.15inthisfigureareX-rayluminouscomparedtothis − relation,suggestingthattheirstackedX-raysignalisdominatedby AGN.Thesameconclusion applies ifgalaxiesaresplitintofiner colourbins,e.g.upperbluecloud( 0.15<∆C < 0.05),valley − − ( 0.05 < ∆C < 0.05) andredcloud(∆C > +0.05). These − − subsamples have mean X-ray and 24µm luminosities similar to theoverall∆C > 0.15populationandthereforearealsoX-ray − luminouscomparedtotheexpectationfromstar-formation.These results are in contrast to the mean X-ray luminosity of 24µm- bright galaxies with ∆C < 0.15, which is consistent with the − star-formationrelation.ThesesuggeststhatX-raybinariesandhot gasdominatetheX-rayemissionofthispopulation. Figure6. 2-10keVX-rayluminosityagainst24µmluminosity.Thecon- tinuouslineistheLX −νL24 relationforstar-forminggalaxiesadapted 5.2 X-raypropertiesofpost-starbursts fromRanallietal.(2003).Thedashedlinescorrespondtothe1sigmarms An interesting trend in Table 1 is the higher fraction of X-ray envelopearoundthisrelation.The(red)triangle correspondstothemean sourcesinpost-starbursts.Intheredcloudinparticular,wherethe νL24 andLX for24µm-brightgalaxieswith∆C > −0.15.Similarly, majorityofpost-starburstsarefound,15percentofthesesystems the(red)filledcircleisfor24µm-brightgalaxieswithopticalcoloursbluer than∆C< 0.15.Theuncertaintiesintheluminositiesofthesetwopop- areassociatedwithX-raysources.IncontrasttheX-rayidentifica- ulationsares−mallerthanthesizeofthepointsandarenotplottedforclarity. tionrateisonly2percentfortheoverallopticalgalaxypopulation Theopen(black)circlesareAEGISX-rayAGNselectedinthehardX-ray andabout4percent forgalaxiesintheredcloud.Inorder toas- band(2-10keV).Alsoshownisthelocalsampleoflow-luminosityAGN sessthesignificanceoftheexcess,weresampledtheopticalgalaxy (open blue squares), which includes Compton thick candidates (crossed populationtoconstructsubsampleswithsizeequaltothenumber bluesquares),fromTerashimaetal.(2002).Themid-IRluminosityofin- ofred-cloudpost-starburstsandwithsimilarM andU Bdis- B dividualAEGISX-raysources,νL24µm,isestimatedfromthe24µmflux tributions.ThefractionofX-rayidentificationsineachsu−bsample densitybyadoptingtheSEDofNGC1068forthek-correction.Theconclu- is registered and the experiment was repeated 10000 times. The sionsarenotsensitivetothisassumption.Also,forAEGISX-raysources X-raydetectionrateof therandom subsamples islower thanthat theLX(2−10keV)iscalculatedfromthe2-10keVobserved-frameflux ofpost-starburstsin98percentoftheexperiments.Theexcessof using a typical intrinsic AGN spectrum of Γ = 1.9 (Nandra & Pounds 1994). This effectively produces absorption-corrected fluxes for sources X-ray sources inthis population istherefore significant at the 98 withcolumndensities NH < 1023cm−2 atz ≈ 1.FortheTerashima percentlevel.Post-starburstsarealsoanon-negligiblecomponent etal.(2002)low-luminosityAGNweusetheobscurationcorrectedlumi- oftheX-raypopulation.Intheredshiftinterval0.68 . z . 0.88, nositylistedbytheseauthors,exceptinthecaseofComptonthickcandi- 21 10%(5/24)oftheX-raysourcesintheredcloudhavepost- ± dates,identifiedbytheequivalentwidthoftheFeKline,adoptingthecrite- starburstopticalspectra.Forcomparisonthefractionofredcloud rionEW>900eV.The24µmluminosityoftheTerashimaetal.(2002) galaxies that are post-starbursts is 7 per cent (34/490). These re- sourcesisestimatedusingtheIRAS25µmfluxdensity.Themean24µm sults,althoughlimitedbysmallnumberstatistics,tentativelysug- luminosityforthesubsamplesusedinthestackingisestimatedbytakingthe gestalinkbetweenAGNandthepost-starburststageofgalaxyevo- averageoftheνL24µmofindividualsources.ThemeanLX(2−10keV) lutionatz 0.8.Thelinkisalsosupportedbythemeanstacked isestimatedusingtheX-rayfluxdeterminedbythestackinganalysisand ≈ X-raypropertiesofpost-starburstsnotindividuallydetectedtothe themeanredshiftofthesourcesineachsubsample. limitof AEGISChandra survey. The post-starburst population is marginallydetectedinboththesoftandthehardspectralbandsata andintheredcloud.Bluergalaxieshavehardnessratioupperlim- significancelevel& 3σ(Table2,column7).Althoughthestacked its(3σ)consistent withsoftermeanX-rayspectra,Γ > 1.4.The signalisnotdominatedbyasinglesourcejustbelowtheX-rayde- corresponding upper limitsfor the24µm-faint subsample do not tectionthreshold,smallnumberstatisticsareaconcern.Theresults providestrongconstraints,butalsosuggestrelativelysoftspectral inTable2,takenatfacevalue,areconsistentwithΓ 0.7.This ≈ properties.Inanycase,thereisnoevidencethatthemeanhardness is much harder than the X-ray spectrum of star-forming galaxies ratioofthispopulationincreaseswith∆C atthesamelevelasfor (Figure5)orunobscuredAGN(Γ 1.9;Nandra&Pounds1994) ≈ 24µm-brightgalaxies. suggesting that absorption issuppressing thesoft X-ray emission In order to improve the statistics we split the 24µm- fromthecentralengineinthesesystems. bright sample at ∆C = 0.15 and stacked separately sources − bluer/redder than this limit. The resulting mean hardness ratio is 5.3 Opticalmorphology:interactionsandstar-formation plotted as a function of redshift in Figure 5. The mean hardness ratioofthe24µm-brightpopulationwith∆C > 0.15isharder In the previous sections, we found evidence for AGN activity − thantheARP220starbursttemplateandinbetteragreement with among 24µm-bright galaxies with ∆C > 0.15 (which in- − theobscuredAGNmodels.Incontrast,thehardnessratioofgalax- cludes post-starbursts). This cut includes the reddest part of the TheroleofAGNinthecolourtransformationof galaxiesatredshiftsz ≈ 1 7 Figure7. GiniagainstM20diagrams.Theregionsoftheparameterspaceoccupiedbydifferentgalaxytypesaredemarcatedwiththedashedlines.Forclarity sourcesatdifferentpartsoftheCMDareplottedindifferentpanels:upperleftsourceswith∆C > +0.05,upperright 0.05 < ∆C < +0.05(valley), − lowerleft 0.15<∆C< 0.05(upperbluecloud).Inallthesepanelsopentrianglesare24µmbrightsourcesandthedotsarefor24µmfaintgalaxies.In − − thelowerrightpanelweshowthepositionofX-raydetectedAGNontheGini–M20diagram. bluecloud,thevalley,andtheredcloud.Theopticalmorphology parameter, theasymmetry, isalsoused tosearch for disturbances of these sources should be a powerful diagnostic of their nature. thatmightindicaterecent/ongoinginteractions(e.g.Abrahametal. Aretheyinteracting/mergingsystems?Dotheyshowevidencefor 1996;Conseliceetal.2000).Thisparametermeasuresthedegree star-formation?Forclassifyinggalaxiesintodifferentmorpholog- towhichthelightofthegalaxyisrotationallysymmetric. ical types, we use the HST/ACS data to estimate the Gini coeffi- cient,whichmeasurestheclumpinessofasource,andthesecond Figure 7 plots the Gini–M20 parameters for galaxies with ∆C > 0.15. While the 24µm-bright population with ∆C > momentofthebrightest20%pixelsofthegalaxy,M20(Lotz,Pri- − 0.15 includes systems classified as ongoing mergers, these are mack&Madau2004),whichmeasuresthecentralconcentrationof − about 7 per cent of the population. The majority of the 24µm- agalaxy.DifferentHubbletypesareseparatedintheGini–M20di- brightsourcesinFigure7,about70percent,havespiralorirreg- agramandthemorphologicalclassificationbasedonthesetwonon- ular morphology, Sb or later. Blue star-forming regions in these parametricestimatorsremainsrobustathighredshift.Anadditional galaxiesareindeedresolvedbytheHST.The24µmemissionof 8 Georgakakiset al. Table2.Stackingresults Sample N <z> band T B S/N photonrate HR fX logLX (σ) (10−9) (10−17) (erg/s) (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) red cloud(∆C > 0.0) soft 741 389.9 17.9 9.6 0.9 1.7 0.2 40.4 all 881 0.67 ± 0.16 0.09 ± hard 1178 892.4 9.6 7.0 1.1 − ± 6.4 1.0 40.9 ± ± soft 169 80.8 9.8 10.7 1.9 1.9 0.3 40.4 24µm-bright 177 0.69 ± 0.01 0.14 ± hard 286 187.5 7.2 10.8 2.4 ± 10.7 2.4 41.2 ± ± soft 451 235.2 14.1 9.5 1.2 1.6 0.2 40.4 24µm-faint 513 0.66 ± 0.27 0.12 ± hard 676 537.3 6.0 5.5 1.4 − ± 4.7 1.2 40.8 ± ± valley(−0.05 < ∆C < +0.05) soft 155 81.8 8.1 9.5 2.0 1.6 0.3 40.4 all 188 0.68 ± 0.23 0.21 ± hard 238 187.1 3.7 6.0 2.4 − ± 5.3 2.1 40.9 ± ± soft 85 41.0 6.9 11.4 2.9 2.0 0.5 40.5 24µm-bright 89 0.70 ± 0.09 0.22 ± hard 113 92.3 4.2 9.5 3.5 − ± 9.0 3.3 41.1 ± ± soft 42 26.1 3.1 5.8 3.0 1.0 0.5 40.3 24µm-faint 61 0.68 ± <0.15 ± hard 61 60.7 0.0 <7.8 <6.9 <41.1 bluecloud(∆C < 0.0) soft 2009 1469.1 14.1 3.9 0.4 0.6 0.1 40.0 all 3331 0.66 ± 0.52 0.15 ± hard 3565 3375.7 3.3 1.2 0.5 − ± 0.9 0.4 40.2 ± ± soft 538 269.0 15.8 9.7 1.1 1.6 0.2 40.4 24µm-bright 572 0.68 ± 0.37 0.12 ± hard 745 613.6 5.3 4.4 1.2 − ± 3.6 1.0 40.8 ± ± soft 1018 823.7 6.8 2.4 0.5 0.4 0.1 39.8 24µm-faint 1811 0.66 ± < 0.24 ± hard 1910 1897.2 0.3 <1.4 − <1.2 <40.3 ∆C > −0.15 soft 1017 560.5 19.3 8.7 0.7 1.5 0.1 40.3 all 1270 0.67 ± 0.17 0.08 ± hard 1643 1279.8 10.2 6.2 0.9 − ± 5.7 0.8 40.9 ± ± soft 300 153.6 11.8 9.5 1.4 1.7 0.2 40.4 24µm-bright 332 0.69 ± +0.00 0.12 ± hard 516 352.7 8.7 9.5 1.7 ± 9.4 1.7 41.1 ± ± soft 531 294.4 13.8 8.2 1.0 1.4 0.2 40.3 24µm-faint 648 0.66 ± 0.31 0.13 ± hard 810 671 5.3 4.35 1.2 − ± 3.7 1.0 40.8 ± ± ∆C < −0.15 soft 1733 1298.5 12.1 3.5 0.5 0.6 0.1 40.0 all 2942 0.66 ± < 0.49 ± hard 3100 2988.3 2.0 <1.2 − <0.9 <40.2 soft 397 196.2 14.3 10.3 1.3 1.7 0.2 40.5 24µm-bright 417 0.67 ± 0.54 0.16 ± hard 515 448.4 3.2 3.1 1.4 − ± 2.4 1.1 40.6 ± ± soft 938 764.6 6.3 2.3 0.5 0.4 0.1 39.8 24µm-faint 1676 0.66 ± < 0.21 ± hard 1776 1763 0.3 <1.5 − <1.2 <40.3 soft 20 13.7 1.7 <7.5 <1.4 <40.3 post-starburst 26 0.76 >+0.35 hard 44 25.9 3.6 15.5 7.1 17.8 8.2 41.4 ± ± Thecolumnsare:(1):Sampledefinition.Theredshiftintervalis0.4 0.9forallsamplesexceptforpost-starbursts;(2):numberofsourcesusedforstacking − afterexcludinggalaxiesthatlieclosetoorassociatedwithX-raydetections; (3):meanredshiftofthesample;(4):spectralbandwhereX-rayphotonsare stacked:soft:0.5–2keVandhard:2–7keV;(5):Totalcounts(source+background)withintheextractionradius;(6):Backgroundcounts;(7):significanceof thedetectedsignalestimatedfromtherelation(T B)/√Bandexpressedinbackgroundstandarddeviations;(8):Photoncountrateinunitsof10−9s−1 − correctedforapertureeffectsandtheChandraresponse.Inthecaseofnon-detectionthe3σupperlimitislisted;(9):Hardnessratiodefinedas(H-S)/(H+S) whereH,Sarethehardandsoftbandphotoncountratesrespectively. TheerrorsareestimatedassumingPoissonstatistics;(10):Rest-frameX-rayfluxin unitsof10−17ergs−1cm−2estimatedinthe0.5-2and2-10keVspectralbandsadoptingthemeanpower-lawspectralindexconsistentwiththeHR.Inthe caseofnon-detectionthe3σupperlimitislisted;(11):Logarithmofrest-frameluminosityinunitsofergs−1. thesesystemsisalsolikelyassociatedwithyoungstars.Moreover, phologicalevidenceabovesuggeststhatdisksrepresentasubstan- visualinspectionoftheHST/ACSimagesfurthershowsthatmany tialfractionofthe24µm-brightgalaxieswith∆C > 0.15and − 24µm-brightsourcesintheE/S0/SaregionoftheGini-M20 dia- abouthalfoftheX-raydetectedAGN.Thissuggeststhat(i)some gramalsohavedisksinadditiontothedominant bulge,i.e.early levelofstar-formationislikelytakingplaceinthehostgalaxyand type spirals. Similarly,visual inspection of the HST/ACSimages (ii)recentmajormergersareunlikelytohaveplayedacentralrole oftheX-raydetectedAGN,alsoplottedinFigure7,suggeststhat intheevolutionofthesesources. about half of these systems are associated withspirals. The mor- Althoughmajormergersdonotappeartobefrequentamong TheroleofAGNinthecolourtransformationof galaxiesatredshiftsz ≈ 1 9 24µm-brightgalaxieswith∆C > 0.15tidaldisruptions,ormi- − normergersmayplayaroleintheirevolution.Figure8compares thedistributionof theasymmetryparameter forthe24µm-bright andthe24µm-faintgalaxiesinthe∆C > 0.15partoftheCMD. − ThepanelsinthisfigurecorrespondtodifferentHubbletypesbased on the Gini-M20 classification of Figure 7. In the E/S0/Sa and Sb/bc panels, the 24µm-bright sample is offset to higher asym- metries compared to the 24µm-faint galaxies. For later Hubble types, Sc/d/Irr, there is littledifference between the two subsam- ples. A Kolmogorov-Smirnov tests shows that the likelihood of theobserveddifferencesifthe24µm-brightand24µm-faintsam- ples were drawn from the same parent population is < 10−6, 7 10−6, and 0.15 for the E/S0/Sa, Sb/bc and Sc/d/Irr classes, × respectively.Forcomparison,X-raydetectedAGN,mostofwhich areintheE/S0/SapartoftheGini-M20diagram,arealsooffsetto higher asymmetriescomparedtoearly-type24µm-faintgalaxies. Thisevidencesuggeststhatminorgravitationalencounters playa roleintheevolutionofboththe24µm-brightsystemswithearly- type optical morphology (E/S0/Saand Sb/bc) andsome of X-ray detected AGN associated with disks. Alternatively clumpy star- formationinthediskofgalaxiescanalsoproducethesameeffect, i.e. higher asymmetry parameter distribution. Differentiating be- Figure8.Normaliseddistributionoftheasymmetryparameterfor24µm- tweenthetwointerpretationsisnoteasy,anditislikelythatclumpy bright (hatched red histogram) and 24µm-faint galaxies (open blue his- star-formationandminorgalaxyinteractionsarelinked(e.g.Bellet togram)with∆C > −0.15.Thedifferentpanelscorrespondtodifferent al.2005). morphologicaltypesdefinedusingtheGini-M20diagramshowninFigure 7.X-raydetectedAGNintheAEGISareplottedinthetoppanel. For late type galaxies (Sc/d/Irr) there is no statistically sig- nificant difference in the asymmetry parameter distribution be- tween 24µm-bright/faint sources. The two populations however, havedistinctM distributions,suggestingdifferencesintheirstel- B larmass.ThisisshowninFigure9,wherelate-type24µm-bright galaxieswith∆C > 0.15 have mean absolute B-band magni- − tudeM 20.8mag.Thisismoreluminousthantheaverage B ≈ − M 20.0magfor24µm-faintsourcesofsimilarmorpholog- B ≈− icaltype.Thetwopopulationsalsohavesimilarrest-framecolour distributionssuggestingsimilarmass–to–lightratios,M/L.Adopt- inglogM/L = +0.1,consistent withthecoloursofthegalaxies inthesample(Bell&deJong2001),theabovemeanopticallumi- nositiestranslatetostellarmassesof 4 1010and2 1010M⊙ ≈ × × for24µm-brightandfaintsourcesrespectively. A plausible interpretation of the results above is that AGN activity requires a massive BH and some gas to fuel it (Kauff- mann et al. 2003). Massive galaxies often host large BHs, while star-formation is an indication of gas availability. The 24µm- brightgalaxieswith∆C > 0.15aretypicallyluminous,M . B − 20mag, and therefore most likely massive. The morphological − evidenceabovealsoindicatesstar-formationinalargefractionof thesesystems.Thisiscontraryto24µm-faintgalaxies,whichare either quiescent and/or less massive than 24µm-bright sources. Galaxyencounters,notnecessarilymajormergers,maystillbere- Figure9.Normalised distribution ofthe absolute B-band magnitude for quiredinthepictureabovetodisturbthegastothegalaxycentre 24µm-bright(hatchedredhistogram)and24µmfaintgalaxies(openblue andalsototriggertheformationofstars. histogram)withlate-typeopticalmorphology(Sc/d/Irr)and∆C> 0.15. − 5.4 ThenatureofAGNactivityintransitionzoneandred AGN.TheX-rayundetectedgalaxieslieinthesameregionofthe galaxies parameter space with moderately obscured low-luminosity AGN TheX-rayhardnessratio& 0.1estimatedinsections5.1and5.2 and Compton thicksources, suggesting that theobserved stacked − for24µm-brightgalaxieswith∆C > 0.15andpost-starbursts signal may be associated with one or the other type of AGN ac- − can be attributed to low-luminosity Compton thin AGN, Comp- tivity,orpossiblywithamixofboth.Ifthemid-IRluminosityof ton thick sources (e.g. Figure 5), or radiatively inefficient accre- 24µm-bright galaxies with ∆C > 0.15 is, on average, domi- − tionflows(e.g.Brandetal.2005).Figure6showsthattheaverage natedbystar-formation,andnotbydustheatedbythecentralen- 24µm-brightbutX-rayundetectedgalaxywith∆C > 0.15has gine,thenthemoderatelyobscuredlowluminosityAGNscenario − X-rayluminosityaboutafactor10belowthetypicalX-raydetected ismostprobable.Wenotehowever,thatstar-formationinthehost 10 Georgakakiset al. galaxy is also likely to contribute or even dominate the mid-IR 6 DISCUSSION emissionof boththeTerashimaet al.comparison sampleandthe X-raysourcesintheAEGISsurvey.InthecaseofComptonthick activity,theintrinsicAGNluminosityisexpectedtobe100–1000 times brighter than the observed one (e.g. Iwasawa et al. 1997). DeeplyburiedAGNwithluminositiesapproachingthoseofQSOs 6.1 AGNandtransitionzonegalaxies (L 1044ergs−1) among the red galaxies in this study can- X ≈ In the nearby Universe, z . 0.1 there has been evidence for an not be ruled. Such objects are postulated to match the spectrum associationbetweenopticallyselectedAGN,recentstar-formation of the diffuse X-ray background (e.g. Gilli et al. 2007), but, cur- andgalaxiesinthetransitionzonefromthebluetotheredcloud. rently,fewhavebeensecurelyidentified(e.g.Georgantopoulos& Kauffmannetal.(2003)foundthatluminousAGNintheSDSSare Georgakakis 2007; Tozzi et al. 2006). Selection methods that in- hosted by early-type bulge-dominated galaxies that have stopped clude mid-IR wavelengths, like the one used here (24µm-bright forming stars, but only in the recent past. Martin et al. (2007) and∆C > 0.15),arecapableofdetectingthispopulation(e.g. − extended thisstudy by constructing thenear-UV/optical CMD of Lacyetal.2004;Sternetal.2005;Martinez-Sansigreetal.2006; galaxiesat z < 0.1 usingdata fromtheGALEXMedium Imag- Donleyetal.2007).X-raystackinganalysishasindeedconfirmed ingSurveyandtheSDSS.ThedensityofSeyfert-2sinthatsample, that at least some of these methods produce samples with hard identified by diagnostic emission line ratios, peaks in the valley meanX-rayspectrathatareconsistentwithheavilyobscuredAGN betweenthetwoclouds.Ourworkextendstheseresultstohigher (e.g. Daddi et al. 2007; Fiore et al. 2007; Georgantopoulos et al. redshift,z 0.7, byfindingevidence forobscured AGNamong 2007). ≈ galaxies in the vicinity of the valley of the CMD and in the red cloud.SimilarresultsalsoapplytotheX-raydetectedAGNpopu- lation.X-raysourceswith∆C > 0.15haveanaveragehardness − ratioof 0.09 0.07,whileforthosewith∆C < 0.15themean − ± − hardnessratiois 0.31 0.08. − ± The incidence of AGN among transition zone galaxies sug- geststhat BH accretion may playa rolein theevolution of these Brand et al. (2005) also reported a hard stacked X-ray sig- systems.ThislinkhasbeenhighlightedrecentlybyHopkinsetal. nal for red galaxies in the range 0.3 < z < 0.9 selected in the (2007)whoshowthatthebuilduprateoftheredcloudmassfunc- XBootes field of the NOAO Deep Wide-Field Survey (Kenter et tionisingoodagreementwithpredictionsbasedontheQSOlumi- al.2005).Theseauthorssuggestedthatthedetectedsignalcanbe nosityfunction.ThiscalculationassumeseitherafixedEddington interpreted as unabsorbed emission from a radiatively inefficient accretionrateforQSOsandthelocalBH-hostmassempiricalre- accretion(suchasADAFs;Narayan&Yi1994,1995).Weexplore lationoramorephysicallymotivatedmodelfortheQSOevolution thispossibilitybycomparingthemeanEddingtonratioof24µm- basedonsimulationswhereBHaccretionistriggeredbygas-rich brightgalaxieswith∆C > 0.15withthatofM87,thebeststud- − galaxy mergers (Hopkins et al. 2006a, b). Hopkins et al. (2007) ied example of a system undergoing radiatively inefficient accre- alsofoundthattheobservedrateofgalaxymergersandtheirdis- tion.ThecentralsupermassiveBHofthisgalaxyhasbeenshown tributioninstellarmassatdifferentredshiftsarebroadlyconsistent tohaveverylowradiativeefficiency,correspondingtoanEdding- ton ratio of 10−7 (Di Matteo et al. 2003). The mean optical withthegrowthrateoftheredcloudmassfunction.Theevidence ≈ abovesuggestsanassociationbetweenAGNactivity,thetransition absolutemagnitudeof24µm-brightgalaxieswith∆C > 0.15is − ofgalaxiestotheredcloud,andmergers.Thesecatastrophicevents M 20.8mag, whichcorresponds toamean stellarmassof B 4 1≈010−M⊙ foramass–to–lightratiooflogM/L = +0.1,con- canleadtothemorphologicaltransformationofgalaxiesandatthe × sametimeofferanefficientwayforchannelinggastothenuclear sistentwiththeaveragecoloursofthegalaxieswith∆C > 0.15 − regionsofthegalaxy,triggeringthegrowthofthecentralBH. (Bell&deJong2001).Assumingthatthebulkofthestellarmass isassociatedwiththegalaxybulgeandusingtherelationbetween Although our analysis broadly supports an association be- BHmassandbulgedynamical massofHa¨ring&Rix(2004), we tweenAGNandgalaxiesonthemovetotheredcloud,theevidence estimateaBHmassof6 107M⊙.Thisisanupperlimitasthe foracausallinkbetweenthetwoislessclear.TheX-rayspectraof × estimatedstellarmassofthe24µm-brightpopulationmayhavea AGNintheredcloud(bothdetectionsandstackingresults)implyat largecontributionfromthediskofthesegalaxies,whichhasbeen leastmoderateamountsofobscurationinalargefractionofthese ignored inthis exercise. Themean 2-10keV X-ray luminosity of systems. This is counter-intuitive to a picture where AGN-driven thispopulation(Table2)correspondstoabolometricluminosityof outflowsblowawaythegasanddustcloudsfromthenucleargalaxy L = 3.5 1042ergs−1,adoptingthebolometriccorrectionof regionsandsuggeststhatsomegasanddustcloudseitherremainat bol × Elvisetal.(1994).ThisisalowerlimitinthecaseofComptonthick orrelaxtothecentreofthegalaxyafterthequenchingofthestar- activity.TheBHmassandtheL estimatesabovetranslatetoa formationtoformatorus.Additionally,majormergersarenotsup- bol lowerlimitfortheEddingtonratioof 5 10−4,morethan3dex portedbythedata.Minormergersortidaldisruptionsmaystillplay ≈ × higherthanthecorrespondingquantityforM87.IfthehardX-ray aroleintheevolutionof24µm-brightgalaxieswith∆C > 0.15. − signal of 24µm-bright galaxies with∆C > 0.15 isassociated Recentstudiesontheevolutionofthestar-formationsincez 1 − ≈ witharadiativelyinefficientaccretionmode, theyhavetobesig- alsosuggestthatadeclineinthemajormergersisnotthedominant nificantlylessextremethanM87.Forcomparison,X-raydetected physicalmechanismdrivingtheevolution.Gasexhaustionindisk AGN atz 1have Eddington ratiosintherange 10−4 10−1 galaxiesthatformstarsinaquiescentmodemaybemoreimportant ≈ − (e.g. Bundy et al. 2007; Babic et al. 2007). The mean Eddington (e.g.Belletal.2005;Wolfetal.2005;Melbourne,Koo&LeFloch ratio of the 24µm-bright galaxies with∆C > 0.15 liesat the 2005; Noeske et al. 2007a, b). The gasdepletion scenario isalso − low-endoftheintervalabove,suggestingthattheyaredominated consistentwiththeobscuredAGNactivityobservedingalaxiesin bylowluminosityand/orComptonThickAGN. thetransitionzoneandintheredcloud.

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.