APreprinttypesetusingLTEXstyleemulateapjv.04/03/99
THESTELLARCONTENTOFNEARBYSTAR-FORMINGGALAXIES.III.UNRAVELLINGTHE
NATUREOFTHEDIFFUSEULTRAVIOLETLIGHT1
RupaliChandar2,ClausLeitherer2,ChristyA.Tremonti3,DanielaCalzetti2,
AlessandraAloisi2,4,GerhardtR.Meurer5,DuiliadeMello6,7
AcceptedforPublicationintheAstrophysicalJournal
arXiv:astro-ph/0505024v1 2 May 2005ABSTRACT
Weinvestigatethenatureofthediffuseintra-clusterultravioletlightseenintwelvelocalstarburstgalaxies,usinglong-slitultravioletspectroscopyobtainedwiththeSpaceTelescopeImagingSpectrograph(STIS)aboardtheHubbleSpaceTelescope(HST).Wetakethisfaintintra-clusterlighttobethefieldineachgalaxy,andcompareitsspectroscopicsignaturewithSTARBURST99evolutionarysynthesismodelsandwithneighboringstarclusters.OurmainresultisthatthediffuseultravioletlightinelevenofthetwelvestarburstslacksthestrongO-starwindfeaturesthatareclearlyvisibleinspectraofluminousclustersinthesamegalaxies.Thedifferenceinstellarfeaturesdominatingclusterandfieldspectraindicatethatthefieldlightoriginatesprimarilyfromadifferentstellarpopulation,andnotfromscatteringofUVphotonsleakingoutofthemassiveclusters.AcutalongthespatialdirectionoftheUVspectraestablishesthatthefieldlightisnotsmooth,butrathershowsnumerous“bumpsandwiggles.”Roughly30–60%ofthesefaintpeaksseeninfieldregionsoftheclosest(<4Mpc)starburstsappeartoberesolved,suggestingacontributionfromsuperpositionsofstarsand/orfaintstarclusters.ComplementaryWFPC2UVIimagingforthethreenearesttargetgalaxies,NGC4214,NGC4449,andNGC5253areusedtoobtainabroaderpicture,andestablishthatallthreegalaxieshaveadispersedpopulationofunresolved,luminousbluesources.BecausethefieldspectraaredominatedbyBstars,wesuggestthattheindividualsourcesobservedintheWFPC2imagesareindividualBstars(ratherthanOstars),orsmallstarclusters.Weconsiderseveralscenariostounderstandthelackofobservedmassivestarsinthefield,andtheirimplicationsfortheoriginofthefieldstellarpopulation.Ifthefieldstellarpopulationsformedinsitu,thefieldmusteitherhaveanIMFwhichissteeperthanSalpeter(α∼−3.0to−3.5),oraSalpeterslopewithanuppermasscutoffof30–50M⊙.Ifstarformationoccursprimarilyinstarclusters,thefieldcouldbecomposedofolder,fadedclusters,and/orapopulationwhichiscoevalwiththeluminousclustersbutlowerinmass.Weusethesebenchmarkpopulationstoplaceconstraintsonthefieldstellarpopulationorigin.Althoughthefieldprobablyincludesstarsofdifferentages,theUVlightisdominatedbytheyoungeststellarpopulationsinthefield.Ifthefieldiscomposedofolder,dissolvingclusters,weestimatethatstarclusters(regardlessofmass)needtodissolveontimescales7–10Myrtocreatethefield.Ifthefieldiscomposedofyoungclusterswhichfallbelowthedetectionlimitofindividualsourcesinourspectroscopy,theywouldhavetobeseveralhundredsolarmassesorless,inordertobedeficientinOstars,despitetheirextremeyouth.Subjectheadings:galaxies:starburst—galaxies:stellarcontent
1.INTRODUCTION
Nearbystarburstgalaxiesarefundamentaltestinggroundsforkeyquestionsrelatedtobothstarformationandgalaxyevolution,aswellastheinterplaybetweengalaxiesandtheirsurroundings.Adetailedunderstandingofthepropertiesoflocalstarburstsisalsocriticaltofur-therexaminestarforminggalaxiesathighredshift,sincethestarformationratesofthesedistant,younggalaxiesaremostsimilartothosefoundinlocalstarbursts(e.g.,Steideletal.1996).However,wearestilllearningaboutthestellarcontentandpropertiesoflocalUV-brightstar-bursts.Meureretal.(1995)showedthatnearbystarbursts
1Based
areirregularstructuresconsistingofdiffuselydistributedlightinterspersedwithprominentstarclusters,includingbothcompactobjects(withhalf-lightradii<∼10pc),andmoreextendedOBassociations.Theyestimatethatclus-tersonlycontribute∼20%ofthelightseeninultravioletimagesofstarburstgalaxies(Meureretal.1995;Maozetal.1996).Thusthediffuselydistributedlightactuallydominatesthestarburst,comprisingroughly80%ofthetotalintheUV.Toaccountfortheobserved“bimodal”natureofstellarpopulationsinstarbursts,Meureretal.(1995)suggestedthattwomodesofstarformationoccurinhighintensitybursts—prominentclusterformationanddominantdiffuselydistributedstarformation.
onobservationswiththeNASA/ESAHubbleSpaceTelescope,obtainedattheSpaceTelescopeScienceInstitute,whichisoperated
bytheAssociationofUniversitiesforResearchinAstronomy,Inc.underNASAcontractNAS5-26555.
2SpaceTelescopeScienceInstitute,3700SanMartinDrive,Baltimore,Maryland21218;ElectronicAddress:rupali@stsci.edu3StewardObservatory,933N.CherryAve.,Tucson,AZ,857214OnassignmentfromtheSpaceTelescopeDivisionofESA
5JohnsHopkinsUniversity,3400N.CharlesStreet,Baltimore,MD21218
6LaboratoryforAstronomyandSolarPhysics,Code681,GoddardSpaceFlightCenter,Greenbelt,MD207717DepartmentofPhysics,CatholicUniversityofAmerica,620MichiganAvenue,Washington,DC200
1
2
Severalexplanationsbesidesstarformationoccurringina“cluster”modeanda“field”modecanexplaintheob-serveddichotomyofstellarpropertiesinlocalstarbursts.CurrentlyonepopularscenariosuggeststhatthediffuseUVfieldlightiscreatedviadissipationofagingstarclus-ters,whichcontributetheirremainingstarstothefield.Thereareseveralworkswhichindirectlysupportthissce-nario(e.g.,Tremontietal.2001;Lada&Lada2003;Fall,Chandar,&Whitmore2004).
ThestellarcontentofnearbystarburstgalaxieshavebeenstudiedinmoredetailusingmultibandopticalWFPC2photometry.Inadditiontostarclusters,color−magnitudediagram(CMD)analysisofthefieldstarpopulationsuggeststwodistinctoverallpopulations.Greggioetal.(1998)andHarrisetal.(2001)findevidencethattheresolved,diffusestellarpopulationsinNGC1569andM83areolderthantheclusters.Tremontietal.(2001)comparedUVspectroscopyofbothresolvedclustersandthediffusefieldlightinthenearbystarburstNGC5253,withstellarevolutionarymodels.Theyestablishedthatthediffuseintra-clusterlightinNGC5253hasaspectrumlackingthestrongO-starwindfeatureswhichareclearlyseeninanumberoftheclusterspectrainthisgalaxy.Alloftheseobservationsareconsistentwithascenariowherethefieldiscomposedofdissolvedstarclusters.Apossi-blecounter-example,whichsupportstheconceptofstarsformingearlyoninbothclustersandinthefield,istheTosietal.(2001)studyofNGC1705.TheyfoundthatinordertoreproducetheCMDofpointsourcesinthispost-starburstgalaxy,a2–3Myrfieldstarpopulationwasrequired.
However,otherscenarioscanalsoplausiblyexplainthepresenceofdiffuseUVfieldlightinstarbursts.Forexam-ple,continuousstarformationingalaxiescouldresultinapopulationofBstars,whichoriginallyformedinclus-tersoverthelastfew100Myr,anddispersedthroughoutthegalaxyoncetheclustersdissolved.Asecondpossi-bilityisthatafractionofUVphotonsoriginatinginhotstarsembeddedinHIIregionsleakoutandscatteroffdustgrains.In30Doradus,UITobservationssuggestthatscat-teredlightisthemechanismresponsiblefortheUVfield(Chengetal.1992).InHe2-10however,amuchmoredistantstarburst,wewereabletoestablishthatscatteredlightcannotbethedominantmechanismresponsiblefortheobservedUVfield(Chandaretal.2003).Barthetal.(1995)suggestedathirdpossibility,thatthediffuselightinnuclearstarburstringsmaybedominatedby(lowermass),unresolvedstarclusters.Afourthpossibleexplanationisthatindividualmassivestarsmaybeforminginrelativeisolationinthefield,assuggestedforthenuclearregionsofM51(Lamersetal.2002).Finally,inourrecentworkonthestellarcontentofHe2-10,weestablishedthatboththeclustersandintra-clusterregionsshowsignaturesofmassivestars.Fromourdetailedsimulations,wesuggestthatthefieldpopulationwhichformedcoevallywiththeneighboringcompactclusters(∼4–5Myrago)maycomefromlessspatiallyconcentrated“scaledOBassocations”(SOBAs),whicharenotlackinginthemostmassivestars,ratherthanfaint,lowmasscoevalcompactstarclusters.IsHe2-10auniquecase,whereweare“catching”thisgalaxyearlyduringawellcoordinatedburst?Doothergalaxieshavefieldstellarpopulationswhichappearolderthanthoseofneighboringclusters,suchasfoundin
NGC5253?Hereweextendourpreviousstudiesofthediffuseintra-clusterlightinlocalstarburstsusinglongslitSTISUVspectroscopy.TengalaxiesfromourHSTpro-gramGO-9036havehighenoughS/Ninthefieldtoallowustostudytheirstellarcomposition.Tothese,weaddthepublishedresultsofHe2-10andNGC5253foramorecompletelookatthenatureofdiffuseUVlightinstarburstgalaxies.
Thispaperisorganizedasfollows:§2providesback-groundinformationonthegalaxysampleanddescribestheobservations,basicdatareduction,anddefinitionofthefieldregions;§3describesreductionandbasicanal-ysisofcomplementaryarchivalWFPC2imagesforthreeoftheclosesttargetgalaxies(NGC4214,NGC4449,andNGC5253);§4derivesthestellarcontentofthefieldbycomparingwithSTARBURST99models;and§5takesad-vantageofspatialinformationaswellasthespectra,inordertoconstraintheoriginoftheUVfieldlightinoursample.Finally,in§6wesummarizethemainconclusionsofthiswork.
2.THESAMPLE,OBSERVATIONSANDDATAREDUCTION
2.1.GalaxySample
OurHSTprogram,GO-9036,wasdesignedtoobtainahomogeneousfar-andnear-UVspectroscopicdatasetof18nearbystarburstgalaxies.Thetargetgalaxiesspanabroadrangeofmorphologies,chemicalcompositions,andluminosities.Theprimarytargetsineachgalaxywereonetothreeluminousclusters(eithercompact′′orgiantHIIregions);however,theprojected25×0.2′′STISlong-slitalsocoversfainter,serendipitousobjects(whichcouldbestarsorclusters),andregionsofdiffuseUVemissionwherenoobvioussourcesareseeninavailablearchivalFOCandWFPC2images.Thegalaxysample,distances,andareaprojectedintheslitarecompiledinTable1.Additionaldetailsforthegalaxysampleandobservations(includingslitlocations)arepresentedinthefirstpaperinourseries(Chandaretal.2005a;hereafterpaperI).Analysisoftheclustercontentisthefocusofthesecondpaper(Chan-daretal.2005b;hereafterpaperII).Inthisinvestigation,weanalysethediffuseUVemissionseenbetweenclustersintentargetgalaxies.Thesearethegalaxieswherethesignal-to-noise(S/N)oftheextractedfieldspectrumissuf-ficientlyhigh(>∼10)toquantifythedominantstellarpop-ulation.TotheseweaddtheresultsofourpreviousworkinHe2-10(Chandaretal.2003)andNGC5253(Tremontietal.2001),foracomprehensivestudyofdiffuseUVemis-sionin12localstarburstgalaxies.
2.2.DataReduction
Inordertocombinemultiplealignedexposuresofthesametarget,weaddedtogethertheflatfieldedtwo-dimensionalimages(therewerenoshiftsbetweenmultiplegalaxyspectratakenforGO-9036),updatedtheexposuretimeintheheaderstoreflectthetotalintegrationtime,andreprocessedthroughtheCALSTISpipeline.Thispipelinerebinsthespectraandprovidesaglobaldetectorlinearitycorrection,darksubtraction,flatfielding,wave-lengthcalibrationandconversiontoabsolutefluxunits.Forsubsequentanalysisonthefieldregionsweutilizedthesefullycalibratedrectifiedtwodimensionalimagesre-processedthroughtheCALSTISpipeline,whichhavea
linearwavelengthscaleanduniformsamplinginthespatialdirection.Furtherprocessingstepsontheextractedonedimensionalspectraincludedderedshifting(valuestakenfromNED)andrebinningtothewavelengthscaleoftheSTARBURST99models(0.75˚Apixel−1)utilizedforcom-parisonwiththeobservations.Wethencorrectedtheex-tractedspectraforstronggeocoronalemissionatLyαandOIλ1302byassumingthatanouterportionfromthetwodimensionalspectrum,whichistypicallylocatedneartheedgeoftheobservablegalaxy,representsapurelygeocoro-nalspectrumneartheLyαandOIλ1302features.Foreachgalaxy,thespectrumextractedfromneartheendoftheslitistypicallyatleast10×weakerincontinuumfluxthenthefieldregions,anddisplaysveryweakspectralfea-tures(iftheywereobservableatall),indicatingthatthesearesuitableforcorrectinggeocoronalfeatures.
2.3.ExtractionoftheField
Inordertostudythestellarcontentofthegalaxiesinoursample,wefirstdetectedobjects(whichcouldbestarsorclusters)fromprofilestakenalongtheslit(700columnswerecollapsedalongthespatialdirection).Ourdetec-tionprocedureisanalogoustothosegenerallyusedforobjectdetectioninimages.Objectsweredefinedaspeaksalongthespatialdirectionwhichreachedalevelatleastfivetimeshigherthanthestandarddeviation(σ)plusthelocalbackgroundlevel(asecondcategory,faintobjects,whichhave3–5σdetectionsarealsostudiedinpaperII).Thistechniqueensuresthatwearedetectingrealsourcesandnotrandomsurfacebrightnessfluctuations.Inprinci-ple,detectedobjectscanbeeitherstellarclustersorindi-vidualstars.However,individualstarsshouldbeconsid-eredaspartofthediffuseUVpopulation.Therefore,itisveryimportanttoassesswhetherourobjectdetectional-gorithmisdetectingindividualstars,sincewedonotwanttoexcludeindividualOand/orBstarsfromourdefinitionofthefield.Basedonthedepthofourdataanddetec-tionlimits,thepossibilityofexcludingindividualmassivestarsisonlyaconcernfortheclosestgalaxies,thosewithin∼5Mpc.Becauseclustersdimastheyage,thetotallu-minositiesofclustersandindividualO/Bsupergiantscanoverlap,makingitimpracticaltouseasimpleluminositycuttoseparatestarsandclusters.Afurthercomplicationisthatthenarrowslitusedinthisstudyoftendoesnotincludeallofthelightfromanobject,eitherbecausetheobjectisextended,and/orbecauseitisnotcenteredintheslit.Thusitisimportanttoconsiderthetotalintrin-sicluminositiesofdetectedsources,andnotjustthelightfallingintheslit.Toaddressthisconcern,weuseacombi-nation˚ofobjectluminosities(frompaperII)measuredat1500A,whicharestrictlowerlimits,andsizeinformation(fromprofilesalongtheSTISslitandsometimesavailableWFPC2images)toassesswhetherweareindeeddetectingindividualluminousstars.
HereweuseNGC5253asanexampletoexaminethenatureofdetectedsources,sincetheFUVspectrumforthisgalaxyisthedeepestofthethreeclosesttargets.AnOsupergianthasanintrinsicspectralluminosityat1500˚A,Lof∼5×1035ergs−1˚A−1
density
1500,.Thisissim-ilartothelightfallingintheSTISslitfortheweakestdetectedobjects(seee.g.,objects3and11inTremontietal.2001).However,thespatialprofilesfortheseob-3
jectsareclearlyresolved(asaretheprofilesforbasicallyalldetectedobjectsinNGC5253andNGC4214,andmostobjectsinNGC4449),andtheslitlocationsuggeststhatmuchofthefluxfromobject3inNGC5253hasbeenmissed.Infact,acomparisonofthetotalLF170WWFPC2imagesandSTISspectroscopy1500fromavailable(seeTable1inTremontietal.2001)showsthatforthefourbrightestsources,theLofthelightfrom1500fluxmeasuredfromimag-ingwhereallanobjectisaccessible,ishigherbyfactorsof4–15thanthefluxmeasuredfromSTISspectroscopy.Similarargumentsapplyfortheobjectsde-tectedinNGC4214andNGC4449.Weconcludethatthedetectedobjectsingalaxiescloserthan5Mpcarenotindividualluminousstars,butstellarclustersorgroupsofstars.Formoredistantgalaxiesobjectfluxesaretoohightocomefromindividualstars.Therefore,wewillrefertodetectedsourcesinourSTISspectraasclusters.Inouranalysis,wemakeanarbitrarydistinctionineachgalaxybetweenbrightclusters(S/N>10)andfaintclusters(typ-ically4≤S/N≤8).Thesearediscussedfurtherin§5.2.2.Wedefinedthefieldregionsforeachgalaxyintwoways:(1)regionswithin3σofthelocalbackground,and(2)theentiregalaxyminusthe1–3targetedluminousclusters(thoseusedtodeterminetheslitorientation).Thespatialpixelsdefinedasbelongingtothefieldwerethensummedtogiveaone-dimensionalfieldspectrum.Notethatinregionswherethebackgroundchangesrapidlyalongtheslit,itcanbedifficulttodeterminetheappropriatelocallevel.Toaddressthisissueandanylingeringconcernsthatourfirstdefinitioneliminatessmallgroupsofindivid-ualOorBstarsfromtheextractedfield,weanalysedbothextractedfieldspectraforeachgalaxy.Weminimizedcon-taminationfromluminousclustersbyavoidinglightfromthewingsoftheirprofiles.Spatialprofiles,alongwithfieldregionsaccordingtoourfirstdefinitionarepresentedinFigure1.
Notethatanumberofgalaxies,particularlythosewithalargenumberofclustersintheslit(e.g.,NGC3310andNGC4449),alsohavemany“bumpsandwiggles”intheirfieldportions(thosewithin3σofthelocalbackground).Infact,inmostgalaxies,particularlythosewhicharenearby,starclustersarenotembeddedinasmoothbackgroundofUVlight.Thefieldalwaysshowssomestructure.Isthisstructureduetovariablereddeningordiscretestellarpopulations?In§5.4wediscusstheresultsofmeasuringthespatialextentorsizeofnumerousfaint“peaks”visi-bleinstarburstfieldregionsthroughourslits.Wefindingeneralthatasignificantfractionofthesepeaksareunre-solved,andthosewhichareresolvedhavesizeswhicharecomparabletoluminousclustersobservedineachgalaxy.Additionally,wewereabletomatchafewofthebrighterpeaksobservedintheSTISfieldregionswithdiscrete,faintsourcesindeeperarchivalWFPC2images.Takentogether,thesepointssuggestthatatleastsomeofthefaintpeaksseenalongthespatialdirectionareactuallydiscretestellarpopulations(possiblyafewstarsorfaintcompactclusters),ratherthanvariationsinextinctionorrandomfluctuationsinthebackgroundlevel.
4
2.4.SelectionEffects/Biases
Wenotethatthepreciselocationsdefinedasthe“field”ineachgalaxydependonanumberoffactors,suchasto-talexposuretime,galaxydistance,andlocalbackgroundlevels.Thusthe“field”aswehavedefinedit,issubjecttoanumberofselectioneffects/biases.Forexample,deeperobservationsmightallowforadditionalsmallbumpsandwigglesinthespectrumtobeaddedtothe“object”list,astheirS/Nisincreased.However,becausewehavecol-lapsedalargenumberofpixelsalongthespatialdirection,ourdefinitionsaresomewhatrobustagainstthedepthoftheobservations.Ourdefinitionof“field”isalsodistancedependentinthesensethatweareunabletouniformlydefineaminimumobjectmassatagivenagewhichisse-lectedasacluster.Moredistantgalaxiesaremorelikelytohavehighermassclustershidinginthefield.Ifsomeoftheseclustersareveryyoung,thanwewouldexpectthesignatureofmassivestarsintheextractedfieldspectratoincreasewithgalaxydistance.Potentially,someofthelessmassiveobjectsthatwehavedefinedasclusters(thisonlyappliestogalaxiescloserthan∼5Mpc)mayberandomsuperpositionsofseveralfieldOstars,orpossibly“scaledOBassociations”orSOBAs(Hunter1999).
Oursecondextractionofthefield,wherewehaveonlyexcisedtargetedclusters,allowsustodirectlyprobethisissue.In§5.2.2wepresentresultscomparingthestellarcontentinthefieldandclusterspectrawithinstantaneousburstSTARBURST99models.Althoughthesemodelsarenotagoodphysicalrepresentationofthefield,theydoprovideasimplewaytoquantifythemassivestarcontent.Wefoundvirtuallynodifferenceinthebestfitages,andhenceinthemassivestarcontent,inourtwodefinitionsofthefieldforeachgalaxy.
2.5.SpatialProfilesAlongtheSlit
Figure1showstheintegrated1250–1700˚A
fluxalongthespatialdirectionofourlongslitspectra.Inadditiontoallowingustodefineclusters/fieldportions,thesespatialprofilesgivesomeinsightintothestellarvariationsacrosseachgalaxy.InTable2wepresentthetotalfractionofFUVlightfromstarclustersobservedineachpointing.Wesummeduplightfromclusters(alongthespatialdi-rection)andcomparedwiththetotallightalongtheslit.Thefractionoflightinthefieldisassumedtobethatnotfoundinclusters.Wederivevaluesforthefractionoflightinthefieldportionsoflocalstarburstgalaxiesrang-ingfromroughly25%to70%.However,ourslitlocationstargetedUV-brightclusters,andwheneverpossibleorien-tationswerechosentomaximizethenumberofsuchclus-tersintheslit.Thustheparticularlocationsobservedineachgalaxyarebiasedagainstthefield.ThefieldfractionspresentedinTable2representlowerlimitstothefieldcon-tribution(andanupperlimittotheclustercontribution)intheFUV,andareroughlyconsistentwiththe∼50–80%fieldcontributionderivedbyMeureretal.(1995)whenconsideringnearbystarburstsintheirentirety.Below,webrieflydescribethespatialcutsforeachsamplegalaxy.Mkn33:Therearefieldregionsoneithersideof,andbetween,twoluminousUVregionswhichcontainmultipleclusters.
He2-10:Wepreviouslymadeadetailedstudyofstar-burstregionA(Chandaretal.2003).Alongthespatial
directionofourslit,therearefiveindividualcompactclus-tersonanelevatedbackgroundatUV-opticalwavelengths.Thenatureofthisdiffusefieldlightwasinvestigated,andfoundtoincludeasignificantpopulationofmassivestars.Thispopulationcouldariseinalargenumberoflowermasscompactclusters,orfromdiffuseSOBAs.Becausethedensityofputativelowermassclusterswouldhavetobeverylarge,weconcludedthatSOBAsformedcoevallywiththecompactclusterswerelikelythemainstellarcon-tributortothemassivestar-richfield.
NGC1741:Thefieldregioninthisgalaxyincludesabroad,lowlevelcomponent(seeFigure1).
NGC3125:ThereissomediffuseUVemissionobservedbetweenthreebrightclusters.Thefieldincludesanumberoffaintpeaks.
NGC3310:Thespatialcutofthisspiralgalaxyshowsamultitudeofpeaks.Basedonthedefinitiongivenin§2.3,wedetecttenclusters.Thefieldregionincludesanumberofveryfaintpeaks.
NGC4214:InFigure2weshowanenlargementofthespatialprofilealongtheSTISslitforpartofthefieldinNGC4214.Faint,individualpeaksareclearlyseen.ThefieldappearsqualitativelysimilartothatinNGC5253,whereanumberofbumpsandwigglesareseenonabroad,elevatedUVbackground.
NGC4449:ThisgalaxyissimilartoNGC3310inthespatialdirection,wherearelativelylargenumber(12)ofindividualclustersarefound,andthefieldincludesaddi-tional,faintpeaks.
NGC4670:ThisgalaxyhasaprofilesimilartothatseeninHe2-10—thereisasomewhatpeakedplateauofdiffuseemission,withbrightclusterssuperposed.Thereishow-ever,slightlymorestructureseeninthefieldofNGC4670thanthatinHe2-10.
NGC5253:ThisgalaxywasstudiedbyTremontietal.(2001).ThereisanelevatedUVbackground,withanum-berofsmallpeaks(includedinthefield),andlargerpeaks(definedasindividualclusters).
NGC5996:Inadditiontothetargetedcentralcluster,acoupleofotherfainterclustersandfieldregionsareob-servedintheslit.
NGC7552:Therearetwobrightclustersonanele-vatedbackgroundintheslit,withsomefieldportionsin-between.
TOL1924-416:Thereisoneveryluminouscluster,plusonefainterone.Thefieldconsistsofanumberoflowlevelbumpsandwiggles.
3.COMPLEMENTARYARCHIVALWFPC2IMAGES
3.1.DataandReduction
Tofurtherconstrainpropertiesofthefieldregionsstud-iedinthiswork,wedownloadedavailablearchivalWFPC2imagesforourtargetgalaxies.Theseallowthestudyofthestellarpopulationsalongtheslitatdifferentwave-lengths,aswellasprovidingabroaderviewofthestel-larpopulationsinourtargetgalaxies,andnotjustthepointingstargetedbyourSTISlongslitspectra.Becausemuchoftheavailablearchivaldataisnotverydeep,ordoesnotincludeobservationsinseveralfilters,werestricttheWFPC2imageanalysistothethreeclosestgalaxies,NGC4214,NGC4449,andNGC5253,whereindividualstarscaneasilybedetected,andtheUVspectraareleast
subjecttoissuesrelatedtodistancebias.ThesethreetargetgalaxiesallhaveobservationsinfilterswhichcanbeconvertedtoJohnson-CousinsU,V,andI.InTa-ble3,wecompileinformationonthesegalaxiesandonthearchivalWFPC2observationsusedinthiswork.Im-ageswereprocessedthroughtheHSTon-the-flycalibra-tionpipeline(whichautomaticallyselectsthemostup-to-datecalibrationfilesduringprocessing).Multipleimageswerecombinedtoeliminatecosmicrays.
WerantheobjectfindingalgorithmSEXTRACTOR(Bertin&Arnouts1996),tolocateallobjects(individualstars,clusters,andahandfulofbackgroundgalaxies)intheF336Wimages(theclosestmatchtothewavelengthrangeofourSTISobservations).Weusedathresholdof3σabovethelocalbackgroundlevel,whichuponvi-sualinspectionappearedtodetectallobvioussources.Ingeneral,adetectionwhichisthemeanofthelocalback-groundplus3timesthestandarddeviation(σ)ofthisbackground,eliminatesrandomsurfacebrightnessfluctua-tions,andleadstodetectionofactualobjects.TheF336Wfilterhasasignificantredleak,˚causingafractionofanob-ject’sfluxlongwardof4000Atobedetectedinthisfilter.Birettaetal.(2000)showthattheredleakinthisfilterisgenerally<∼5%ofthetotalfluxforstellarpopulationsdominatedbyK3Vorearlier-typestars,whichisappro-priateforthemajorityofobjectsobservedinthethreeclosestgalaxies.However,oneclusterinNGC5253isofparticularinterest,cluster5fromCalzettietal.(1997).Thisisahighlyextincted,veryyoungcluster,whichemitsstrongthermalradioemission,butisalsoseeninopti-cal/ultravioletimages.Notethattheexactphotometryforthissingleobjectdoesnotaffectouroverallconclu-sions.
3.2.Photometry
AperturephotometrywasobtainedusingthePHOTtaskinDAOPHOT(Stetson1987).Ithasbeenestab-lishedthataperturecorrectionsindifferentHSTWFPC2filtersshowverylittledifferencewithobjectsize(Holtz-manetal.1996;Larsen1999).Thus,weusedasmallapertureradius(r=3pixels)todetermineobjectcolors,inordertominimizecontaminationduetoneighboringsourcesandtheimpactofuncertaintiesinthebackgrounddeterminationwhenfainterobjectlightisincluded.How-ever,thereisasignificantfractionoflightoutsidethisra-dius,whichvariesbasedonhowextendedanobjectis,andwhichdirectlyimpactsthetotalmagnitudemeasurements.Inordertomeasureaperturecorrectionsforoursources,wefirstmeasuredthesizesofalldetectedobjectsusingLarsen’s(1999)ISHAPEroutine.Detailsofsizemeasure-mentsaregivenbelow.Sincethesourceshavealargerangeofsizes,whichleadstoverydifferentaperturecorrections,welinkedtheaperturecorrectionstothemeasuredsizeforeachobject.Weusedthepreviouslymeasuredaper-turecorrectionsfromLarsen&Brodie(2000)forISHAPEKING30models(for∆m3−>5and∆m5−>30),whicharecompiledintheirTable1.Thesevaluesaregivenfordis-cretevaluesofthemeasuredFWHMinpixels.Wefitasecondorderpolynomialtothesevalues,andusedthisequationtocalculateaperturecorrectionsforeachobjectindividually.
8see
http://www.noao.edu/staff/dolphin/wfpc2
5
CTEcorrectionswereperformedasdescribedinDol-phin(2000).8ThecorrectedinstrumentalmagnitudeswereconvertedtostandardJohnson-CousinsU,V,andImagnitudes.UsingEquation8andTable7ofHoltzmanetal.(1995),themagnitudeswerederivediterativelyusingWFPC2observationsintwofilters.
3.3.SizeMeasurements
Intrinsicsizesforourentireobjectsampleweremea-suredusingtheISHAPEroutine.Adetaileddescrip-tionofthecodeisgiveninLarsen(1999),alongwiththeresultsofextensiveperformancetesting.Essentially,ISHAPEmeasuresintrinsicobjectsizesbyadoptingananalyticmodelofthesourceandconvolvingthismodelwitha(user-supplied)pointspreadfunction(PSF),andthenadjustingtheshapeparametersuntilthebestmatchisobtained.Kingmodelprofileswithconcentrationpa-rametersofc=30wereconvolvedwithaPSF,andfitin-dividuallytoeachobject.ISHAPEestimatestheFWHMofeachcluster(inpixels),whichwasthenconvertedtothehalf-light(effective)radius,rISHAPEmanual.TheinputPSFeff,asdescribedinthetothisalgorithmiscru-cial.WecreatedaPSFbyhand-selectingstarsintheimage,andthencomparedtheresultswiththoseofasub-sampledTinyTimPSF(whentheTinyTimPSFwasused,convolutionwiththeWFPC2diffusionkernelwasimple-mented).WefoundthatthesizeestimatesfromISHAPEusingthesetwoPSFsdifferedbylessthan20%.FinalmeasurementsweremadeusingtheTinyTimPSF,sincethisiseasilyreproducible.AsinglePSFwasgeneratedforthePCCCD,andonefortheWFCCDs.Weassumethatobjectswith∆FWHM(FWHMmeasured−FWHMPSF)of0.2pixelsorlessareunresolved,pointsources.
4.ANALYSIS
4.1.DerivingStellarProperties
Onewaytoassess(any)differencesbetweenclustersandthefieldinlocalstarburstsistoquantifythestellarcon-tentofeach.Isthefieldlackinginmassivestarsrelativetoneighboringclusters?Theanswertothisquestioncanbeusedasastartingpointtostudythecompositionofgalax-ies,andmakeinferencesconcerningthepropertiesofstarformation.OneofthemainadvantagesofusingUVspec-troscopytoassessthedifferencesbetweenstellarcontentinclustersandthefieldistheshorttimescalesoverwhichmassivestarsevolve.ToillustratethediagnosticpowerofUVspectroscopyinassessingmassivestarcontent,inFigure3wepresentexamplespectrafromvariousSTAR-BURST99modelswhichclearlyshowtheeffectmassivestarshaveoncompositeFUVspectraofstellarpopula-tions,particularlyontheNIVλ1240,SiIVλ1400andCIVλ1550broadPCygniprofiles.
Inordertomakeaquantitativecomparisonbetweenthestellarpopulationsdominatingclustersandfieldregions,wecreatedcomposite“cluster”spectra,whicharetheun-weightedsumofmultipleclusters.Whentherewereanumberofclustersavailable,wecombinedthembylumi-nosity,makinganarbitraryseparationbetweenbrightandfaintclusters,dependingonspecificsofclustercharacteris-tics(luminosities)withinagivengalaxy.Ingeneral,sinceyoungerclusterstendtobemoreluminousthanolderones
6
ofcomparablemass,thiswillroughlydivideclustersbyage.Toderivequantitativeclusterparameters,wethencomparedereddened(reddeningdiscussedbelow)clusterspectratoagridofinstantaneousburstSTARBURST99models(Leithereretal.1999),andrecordedtheresultsinTable4.Instantaneousburstmodelsareappropriateforcompactclusters,sincetheyarecoevalstellarpopulationswithsmallcrossingtimes(<0.1Myr)comparedwiththemodels.Whileinstantaneousburstmodelsarenotanidealphysicalrepresentationoffieldregionsextendingoverhun-dredofparsecs,theyprovideaconsistentwaytocomparethestellarcontentofclustersandthefield.BecauseweareworkingintheUV,thefieldspectra,eventhoughtheyprobablycontainamixofages,willbedominatedbysig-naturesoftheyoungeststellarpopulations.In§5.5.2wecompareourfieldregionswithcontinuousstarformationmodels.
STARBURST99modelshavebeenoptimizedtorepro-ducemanyspectrophotometricpropertiesofgalaxieswithactivestarformation.Detailsoftheinputstellarparam-eterstoSTARBURST99canbefoundinLeithereretal.(1999);herewebrieflysummarizethemodelparametersusedinthiswork.TheSTARBURST99UVspectralli-braryisavailablefor2metallicities:solarandLMC/SMC(∼1/4solar).Ingeneral,weusedthemetallicitymodelwhichprovidestheclosestmatchtoagivengalaxy’sabun-dance(tabulatedinpaper1).However,forsomegalaxiestheabundanceisin-betweenthetwomodels.Forthese,wecomparedtheresultsfrombothabundancemodels,andadoptedtheonewhichresultedinthebetterfit(lowerχ2).Ingeneral,ageestimatesforindividualclustersareingoodagreementbetweenthetwomodels.
Fortheinstantaneousburstmodels,weadoptedastan-dardSalpeter(1955)initialmassfunction(IMF),withloweranduppermassesof1M⊙and100M⊙respec-tively.ArgumentsinfavorofaSalpeterIMFaresum-marizedinLeitherer(1998),andadiscussionofvariousstudiesavailableintheliteraturecanbefoundinTremontietal.(2001).Highmasslossrates,asrequiredbystellarevolutionwereassumed(Meynetetal.1994).
Inordertodetermineintrinsicpropertiesforastellarpopulation,theonedimensionalspectramustbecorrectedfortheeffectof(foreground)Galacticextinction,aswellasforthedustobscurationintrinsictothestarburstitself.WeassumeforegroundextinctionvaluesfromtheSchlegeletal.(1998)maps,andadopttheextinctioncurveofFitz-patrick(1999).Todeterminethereddeninginternaltothestarburstitself,wecomparedtheobservedcontinuumofeachclusterandfieldspectrumwithstellarevolutionarymodels.TheintrinsicFUVspectraldistributionofyoung,unobscuredsinglestellarpopulationsfollowsapowerlaw,witheffectivespectralindexβ,whereFpowerlawexponentfromthatλ∝λβ.Anyde-viationofthepredictedbytheoreticalmodelsisassumedtobeduetotheeffectsofdust.TheintrinsicUVspectralenergydistributionsofanyyoung(<10Myr)starburstpopulationistypicallynearβ∼−2.6.Thefitforβwasperformedoverthespectralregion1240–1600˚A.Toaccountforthespectralfeatures(whicharemostlyinabsorption),weperformedthefitit-erativelywithrejectionthresholdssetat2σforthelowerboundanda3.5σupperbound.
Wedeterminedwhichmodel-ageprovidedthebestfittoourclustersandfieldspectrabycomparingthe(fore-ground+intrinsic)extinctioncorrectedobjectspectrawithinstantaneousburstmodelsfromSTARBURST99withagesfrom1–100Myrinstepsof1Myr.ModelfittingcloselyfollowsthetechniquesdevelopedbyTremontietal.(2001),andthereaderisreferredtothatworkfordetails.Briefly,inordertomaximizethesensitivityofχ2tospec-tralregionsmostsensitivetotheageandstellarcontentofthestarburst,eachpixelinthespectrumisclassifiedasbelongingtoeitherthecontinuum,aninterstellarline,orastellarwindline(basedonthedetailedlineanalysisofdeMelloetal.2000).Thesewerethenassignedweights,withinterstellarlinesbeinggivenaweightofzero,con-tinuumaweightofone,andstellarwindlinesaweightoften.Becausenearlyallofthefeatureswhicharesen-sitivediagnosticsoftheageofastellarpopulationalsohavesomeinterstellarcontamination,theinterstellarcoreoftheNVλ1240,SiIVλ1400,andCIVλ1550lineswerealsomaskedout.ThiseffectivelyeliminatestheISMfromconsideration,andreliesonlyonstellarsignaturestode-rivetheirproperties.Theroutinereturnsthebestfitage(minimumχ2),whichisgiveninTable4.
Toquantifyerrorsassociatedwithourbestfitmodels,weutilizedthebootstraptechnique.Theresidualsofthebestfitwererandomlyresampledandaddedtothemodelspectrum,andthenrerunthroughtheautomatedfittingroutine1000times.Theerrorbarsassociatedwiththe90%confidenceintervalwerederivedfromahistogramofthederivedages.TheageuncertaintiesarealsogiveninTable4.Forclusters,agescanbederivedtowithin±1Myr,highlightingthestrongeffectthatmassive,short-livedstarshaveontheUVspectrum.
4.2.SummaryofModelFittingbyGalaxy
Figure4showscompositeclusterandfieldspectraforeachgalaxyinoursample.TheresultsfromcomparisonwithinstantaneousburstmodelsaregiveninTable4,anddiscussedgenerallybelow.Notethatthefigures,analysis,anddiscussionarebasedonourfirstdefinitionofthefield,wherewehaveincludedallregionswithin3σofthelo-calbackground,sincethisisamoreconservativeapproachanddoesnotincludeobviousstarclusters.However,wehaveverifiedthattheresultspresentedheredonotchangesubstantiallyforthesecondfielddefinitionineachgalaxy.Mkn33:Figure4showsthecompositeclusterandfieldspectraforMkn33.Thefieldhasweaksignaturesofmas-siveOstars,andhasabestfitageof6Myrwhencom-paredwithinstantaneousburstmodels.Thisisslightlyolderthanfoundforthecompositeclusterspectrum.Ingeneral,wefoundthatseparatingclustersintoabrightandfaintcategorykeptanyindividualclusterfromseverelybi-asingtheluminosityweightedageofthecompositeclusterspectra.
He2-10:Studiedinapreviouspaper(Chandaretal.2003),themainstarburstregion(A)showsindividualcompactclustersonanelevatedbackgroundatUV-opticalwavelengths.Thenatureofthisdiffusefieldlightwasin-vestigated,andfoundtoincludeasignificantpopulationofmassivestars.WeconcludedthatalargenumberoflowermasscompactclustersordiffuseSOBAscouldberesponsible.Thetailsofthisdiffuseemissionextendout∼100pcfromstarburstA.Thediffusefieldlightisfainter,butshowsthesamespectralsignatureasthatinstarburstregionA.Therefore,weseethesignatureofmassivestars
forminginthe“field”inthisactivelystarburstinggalaxy.NGC1741:Figure4showsasequenceofthree“com-posite”stellarpopulations.Thetoppanelshowsthemostluminousclusterinourslit,cluster1,thesecondpanelisanaverageoffainterclusters2,3,and4,whilethebottompanelshowsthefieldspectrum.Thereisacleargradientinthestellarpopulationsinthisgalaxy,astracedbythestrengthofthestellarwindlines—thesearestrongestinlu-minouscluster1,weakerinclusters2,3,4andweakerstillinthefield.Theseobservationssuggestanagesequence,whichisquantifiedinTable4.
Cluster1isquiteyoung,withverystrongNVλ1240andSiIVλ1400linesithasabestfitageof3Myr.Theotherthreeclustersallhaveagesof6Myr.Thefieldhastheoldestcompositeageof7–8Myr,de-pendingslightlyonwhichmetallicitymodelisadopted.Regardless,thereisnoobvioussignaturefromthemostmassivestarsinthefieldregion.
NGC3125:Theaveragespectrumofthetwomostlu-minousclustersshowssignificantlystrongerwindlines(inparticularNVλ1240andSiIVλ1400)relativetothefield.ClusterNGC3125-1alsoshowsstrongWRsigna-tures,includingtheNIV]1487˚A,NIV1718˚Alines,andtheHeII10˚AlinetypicallyfoundinWNstars(seeChandaretal.2004foradetailedderivationoftheprop-ertiesforthisunusualcluster).FitstoSTARBURST99instantaneousburstmodelsgivemeanagesof3Myrforbothclusters,and8Myrforthefield.
NGC3310:Thefivebrightestclustershaveameanageof5Myr,andfivefaintoneshaveameanageof7Myr.Thefieldhasabestfitageof8Myr.
NGC4214:Thebrightestclusterhasstrongwindlinesandayoungage(4Myr).Whenweaveragetheindividualfainterclusterspectra,wefindthattheseobjectsarenotonlyfainter,buttheyarealsoolderonaverage,withameanageof7Myr.Ourbestfittothefieldgivesanageof8Myr.
NGC4449:ThemeanclusterspectrumshowsPCygniprofiles,andhasaformalageof5Myr.Thefieldshowslittleevidenceforthepresenceofmassivestars,andhasabestfitageof7Myr.
NGC4670:Themostluminouscluster,NGC4670-1is7Myrold.Threefainterclusters(2,3,4)arecoeval,withagesof5Myr.Notealsothatthesefainterclustersarelo-catedveryclosetooneanother,whilecluster1islocatedslightlyfurtheraway.Thecompositefieldshowstheweaksignatureofmassivestars,andhasameanageof6Myroldwhencomparedwithinstantaneousburstmodels.NGC5253:Tremontietal.(2001)findadifferenceinthemeanagesofthecompositeclusterandfieldspectra.Re-doingtheextractionindependently,wefindameanageforthenineextractedclustersof4Myrandameanageforthefieldregionsof7Myr,similartotheearlierresults.NGC5996:Thebrightestclusterisalsotheyoungest,at5Myr.Thestrengthofthewindlines(particularlytheSiIVλ1400feature)inthecompositefaintclusterspec-trumandthefieldestablishesthatmassivestarsresideinboththefaintclustersandthefield,eventhoughtheseappeartobesomewhatolderthanthebrightestcluster.NVλ1240andCIVλ1550areslightlymorepronouncedinthefaintclustersthanthefield,sothesemaycontainsomewhatmoremassivestarsthanthefield.
7
NGC7552:Thetwobrightclustershaveameanageof5Myr,whilethefieldappearsslightlyolder,at6Myr.TOL1924-416:Thebrightestclusterhasanageof1Myr,whiletheyoungeststellarpopulationcontributingtotheUVfieldlighthasanageof7Myr.
4.3.RadialProfiles
Radialprofilesforobjectsinnearbygalaxiescanbeusedtodeterminewhetheranobjectisextendedornot.Ifanobjectisresolved,itislikelyaclusterorgroupofstars;howeverunresolvedobjectscanbeeitherstars,chancesu-perpositions,surfacebrightnessfluctuations,orcompactclusters.Althoughsizeestimatesaretypicallymadefromimaging,hereweattempttoalsousetheinformationalongthespatialdirectionoftheslittodeterminewhetherclus-tersandbumps/wigglesinthefieldareresolved,andfromtherereachsomeconclusionconcerningthenatureofthediffuseUVlight.ThisisparticularlyvaluablesinceSTISpixelssampleevensmallerscales(0.025′′pix−1)thanthePC(0.0455′′pix−1)CCDonWFPC2.
TomeasuretheFWHMofbumps/wigglesinthefield,wecomparewiththeFWHMmeasurementofabrightstar(GD71)observedwiththesamegratingandslitcombina-tion.WemeasuretheFWHMofstarGD71tobe∼0.08′′,andtakethistorepresentapointsource.Becausethesizeestimatesforthefaintpeaksinthefieldarelikelynotveryaccurate,weonlydeterminewhethertheseappearresolvedornot,relativetothemeasurementofGD71.Wediscusstheresultsofthisexercisein§5.4.
5.DISCUSSION
5.1.IsScatteredUVLightResponsiblefortheField?CouldthescatteringofUVphotonsexplainthepresenceofdiffuseUVemissioninlocalstarburstgalaxies?Inthisscenario,UVphotonsoriginateinyoungstarsembeddedinthemassiveclusters.Sincemostofthenon-ionizingUVra-diationproducedinstarformingregionsescapesthelocalcluster(e.g.Misiriotisetal.2001;Hippeleinetal.2003),itisfreetobescatteredbyinterstellardust.RecentGALEXUVimagesofnearbyspiralgalaxiesclearlyshowthepres-enceofdiffuseUVemissionbetweenspiralarms(e.g.,M101;Popescuetal.2005)andatlargegalactocentricradii(e.g.,M33,M31,andM83;Thilkeretal.2005a,b).Whencombinedwithobservationsatlongerwavelengths,thespectralenergydistributions(SEDs)ofspiralsarewellreproducedbygalaxycomponentswhichincludealayerofdiffuselydistributedcolddust(e.g.,Popescuetal.2000).Far-IRandsubmillimeterimagingofnearbyspiralshavedirectlyshownthatthislayerofcolddustcoverstheentireextentofthedisk(e.g.Haasetal.1998;Hippeleinetal.2003;Popescuetal.2005).InM101,acomparisonofGALEXUVandISOfar-IRimagesshowedthattheratioofUV/far-IRemissionvariesbetweenarmandinter-armregions(Popescuetal.2005).Thesedifferencesareconsis-tentwiththeinter-armUVemissionbeingscatteredlightoriginatingfrommassivestarsinthespiralarms.Simi-larly,thediffusefar-UVemissionobservedbeyondtheHαdiskinM33,M31andM83(Thilkeretal.2005a,b)isin-terpretedasnon-ionisingUVradiationescapingfromtheoutermostHIIregionsandthenscatteringoffdustgrainsfurtheroutinthedisk.
CouldthediffuseUVlightobservedinlocalstarburst
8
galaxiesresultfromsuchscatteringofUVphotons?Scat-teringbydustcanchangetheoriginalspectralshape(i.e.,makethespectrumappearbluer),butcannotaddorsub-tractnarrowspectralfeatures,suchasPCygniprofiles.Inthiswork,wehaveshownthattheyoung,massivestarswhichdominatetheUVfluxfromclustershaveanO-typespectrum,withprominentP-Cygniprofiles.Bycontrast,theUVlightoriginatinginthefieldshowsaB-typestellarspectrum.Becausethespectrafromthetwoenvironmentshavedifferentstellarfeatures,massivestarspoweringtheluminousclustersarenottheoriginalsourceforanyUVphotonsscatteringoffdustgrains.However,iflowermass,B-stardominatedclustersarefoundinthefield,wecannotruleoutthepossibilitythatphotonsleakingoutfromtheselowermassclustersarescattered,andthusproducethefieldUVlight.Regardless,thedifferencesinspectralfea-turesobservedbetweenthefieldandclustersdemonstratesthattheUVemissionineachenvironmentoriginatesinadifferentstellarpopulation.
HowcanourevidencethatthediffuseUVlightinstar-burstoriginatesina(faint)stellarpopulationberecon-ciledwiththeclaimthatthediffuseinter-armUVlightinM101isduetoscatteringoffdustgrains?First,thegeometryanddistributionofdustinlargespiralsmaydiffersubstantiallyfromthatindwarfsystemssuchasthosewhichoftenhoststarbursts.Second,Popescuetal.(2005)pointoutthateveninthemostextremeinterarmregionsdiscretesourcesofUVemissionareobservedinfull-resolutionGALEXimages.Giventhelow-resolutionoftheGALEXimages(6′′FWHM)withrespecttoHST,itisplausiblethatsomefractionofthediffuseinter-armemissioninM101arisesfromadispersedstellarpopulationratherthanfromscatteringduetodust.Forexample,theGALEXimagesarenotsufficientlydeep(anddonothavethespatialresolution)todetectindividualBstars(whichareaplausiblecontributortothediffuseUVfieldlight).Rather,theseobservationsareappropriateforstudyingstellarcomplexeswhichhavesizesofmanytenstohun-dredsofparsecsatthedistanceofM101.AnanalysisofWFPC2imagesforNGC4214,NGC4449,andNGC5253clearlyrevealapopulationofdispersedbluepointsources,whichareconsistentwithapopulationofB-supergiants(§5.3).WesuggestthatdeepUVimagingofinterarmre-gionsinspirals,withsufficientdepthtorevealthepresenceofindividualBstarscouldresolvewhetherthediffuseUVemissioninM101arisesfromadiffuselayerofdust,orfromadiscretestellarpopulation.
5.2.GeneralSummaryoftheStellarContentinField
RegionsofStarbursts
5.2.1.FUVslopemeasurements
InTable5wecompileslopes,β,measuredforthecom-positefieldandclusterspectraaftercorrectionforfore-groundextinction,butwithnocorrectionforextinctionduetothehostgalaxy.Detailsoftheβmeasurementswereprovidedin§4.1.Themainuncertaintiesintheval-uesofβcomefromthefactthatapowerlawdoesnotalwaysprovideagoodfittothedata,mostlyduetolineblanketinginthespectra(althoughwehaveexcludedthemost˚severelylineblanketedspectralregionslongwardof10A).
Ingeneral,theintrinsicslopeofastellarpopulation
becomesredderwithage.However,acomparisonoftheslopesforsinglestellarpopulationsyoungerthan∼10Myr(typicalagesofclustersobservedinourstarburstgalax-ies),withthoseforcontinuousformationmodelswhichfitourdatashowthatbothhavesimilarvalues.Forex-ample,acontinuousstarformationmodelwithSalpeterIMFanduppermasscutoffof30M⊙hasanintrinsicslope,β=−2.5(see§5.5.2foradiscussionofcontinu-ousstarformationmodelsandtheircomparisonwiththedata).Becausethedominant/youngestUVpopulationsintheclustersaresomewhatyoungerthanthoseinthefield(asdiscussedindetailbelow),onemightexpectthattheintrinsicslopefortheclustersissomewhatbluerthanforthefield.
InTable5however,βmeasurementsforthefieldappearcomparabletoorbluer(steeper)forhalf(6/12)ofthefieldspectrawhencomparedwiththecompositeclusterspectrainthesamegalaxy.BecausethespectralfeaturesintheUVallowustopreciselyage-datethestellarpopulationsintheclustersandthefield,weknowthatthedominantstarsinthefieldaregenerallyolderorlessmassivethanthoseintheclusters.Thereforeonemightexpectthatβfortheclusterswouldbebluerthanforthefield.However,ourstudyestablishesthatthefieldisnotcategoricallyredderthanyoungclustersinstarbursts.Amorelikelyexplana-tionfortheobserveddifferenceinslopesbetweenclustersandthefieldisextinctionvariationsbetweenstellarpopu-lationsresidingineachenvironment,withthefieldshow-inglessreddening.BecausetheUV-brightclustersarequiteyoung,itislikelythattheystillhaveatleastsomeresidualnatalISManddustsurroundingthem,leadingtoredderslopemeasurements.
Thebluer,lessreddenedslopesfoundinthefieldre-gionsofsomelocalstarburstswhencomparedwithstarclustershaveimplicationsfortheoverallextinctionesti-matesderivedforhighredshiftgalaxies.Thisisparticu-larlyimportantwhenUVspectraofwholehigh-zgalaxiesarecomparedwithnarrow-slitorsmall-apertureUVspec-traoflocalstar-forminggalaxies.Inferencesofdifferencesinstellarpopulationcompositionsand/orextinctionchar-acteristicsmaystemsfromthefactthattheUVspectrumofthelocalgalaxyisonlytargetingonecomponent(e.g.,onecluster)oftheentirestarformingregion.
5.2.2.CharacterizingtheFieldStellarPopulationWefindthattheoverallstellarcontentinourobservedSTISfieldregionsofnearbystarburstgalaxiescanbedi-videdintothreebroadcategories:theclearpresenceofOstars,aclearlackofOstars,andanintermediateclassshowingaweaksignatureofOstars.He2-10isuniquetooursample,withtheyoungestmeanfieldstellarpopula-tionofanysamplegalaxy,clearlyshowingthesignatureofOstars.Theotherextremeincludes7galaxies,whichshowfieldspectralackinginthemostmassivestars(ab-sentPCygniprofiles),similartothatfoundbyTremontietal.(2001)forNGC5253.ThiscategoryalsoincludesNGC1741,NGC3125,NGC3310,NGC4214,NGC4449,andTol1924-416.Table4showsthatthesehavebestfitagesof≥7Myr,andinspectionofthefieldspectrainFigure4confirmsthelackofmassivestars.ThethirdcategorycomprisesMkn33,NGC4670,NGC5996,andNGC7552,whichhavemeanfieldagesof6Myr,andshowweaksignaturesofOstars.
Theclosestgalaxies(i.e.thosewithin4Mpc;NGC5253,NGC4214,andNGC4449),wherewehavethebestspa-tialresolutionandtheeasiesttimedistinguishinglowlevelclustersfromthefield,allclearlylackbroadlineprofilesofNVλ1240,SiIVλ1400,andCIVλ1550thatarecharac-teristicofOstarwinds,instarkcontrasttothecompositeclusterpopulationsineachgalaxy.Thisresultholdsforbothfielddefinitions,whetherweareexcludingallpeaksabove>∼3σ,oronlyexcludingtargeted,luminousstarclusters.Thereforewearenotconcernedthatourdefini-tionhasaprioriexcludedOstarsfromtheextractedfieldspectra.Notsurprisingly,weseemorevariationinthefieldregionsofmoredistantgalaxies.Some,e.g.,NGC3125,NGC3310,NGC1741,andTol1924-416(allfurtherthan10Mpc),clearlylackthesignatureofmassivestars,andareverysimilartothefieldspectrafortheclosestgalax-ies.Others,suchasMkn33,NGC4670,NGC5996,andNGC7552,againmoredistantthan10Mpc,showweaksignaturesofOstars.Forgalaxiesinthislattergroup,ourfieldextractionlikelyincludessomeyoung,relativelymassiveclusters,whichprovidetheweakOstarsignature.Thisbiaswithdistanceisnotunexpected;however,itisgratifyingthatanumberofthemoredistantgalaxiesinoursampleshowaresultsimilartotheclosesttargets.Weconcludethatthefieldregionsoflocalstarbursts(exceptHe2-10)arenotconducivetotheformationofisolatedmassivestars(assuggestedisoccurringinthenuclearre-gionsofM51;Lamersetal.2002),intherandomfieldregionscoveredbyourlongslitpointings.
Forgalaxieswithmeanderivedfieldagesof≥7Myr,wefindthatthefaintclustersareonly1–3Myryoungerthantheyoungeststellarpopulationscontributingtothefieldregions.Implicationsfortheserelativelysmalldiffer-encesinagebetweenthedominantstellarpopulationsinclustersandthefieldarediscussedin§5.5.4.
Basedontheresultsdescribedabove,He2-10appearstobeuniqueinoursample.Becausethedominantstellarpopulationinthefieldformedcoevallywiththenearbyclusters,weestimatedtheUVcontributionofthecom-pactclustersinstarburstregionAtothetotalUVflux.Assumingthattheclusterswhichdidnotfallinourslithavethesameageasthosewhichwerestudied,wefindthat40%ofthelightcomesfromtheclusters,leavingtheremaining60%originatinginthefield.Thisisstillalowerlimitsincesomeclustersmayhavedissolved,butprovidesafirmlowerlimittotheamountofUVlightcomingfromclustersinthisstarburst.
OurmainresultfromanalysisofthestellarcontentofclustersandthefieldregionsofstarburstsisthatthefieldgenerallyshowsweakerorabsentPCygniprofilescom-paredwithclusters.Thusthefractionofhighmassstarsthatarevisibleissmallerinthefieldthaninthebrightclusters.Herewehavesimplyshowcasedthisresultbypresentingthebestfitagesderivedbycomparinginstan-taneousburstevolutionarysynthesismodelswithourex-tractedspectra.Althoughthisisnotthebestphysicalrepresentationofthefield,itisasimpleexercisewhichillustratesthedifferenceinmassivestarcontent.Thisre-sultinstarburstsissimilartothatfoundforOstarsintheGalaxyandLMC.VandenBergh(2004)usedacatalogofGalacticO-typestars(Maiz-Apellanizetal.2004)andfoundthatO-typestarsinclustersandassociationshaveearliertypes(presumarlylargermassesoryoungerages)
9
thanthoseinthegeneralfield.IntheLMC,Masseyetal.(1995)findsthatalthoughmassivestarsareborninthefield,theyformmorerarelyinthisenvironmentthaninclustersandassociations.
Couldourobservedlackofmassivefieldstarsresultfromstatistics,duetotherelativelysmallareascoveredbyourSTISslit,orpossiblyduetomodeststarformationrates?Inordertoassessthispossibility,weuseourex-tractedspectratoestimatethemeanstarformationrate(
SFRinthefieldregionsoflocalstarbursts,
wecompareourdereddened(forbothforegroundandin-trinsicextinction)fieldluminosityat1500˚A(correctedfordistance)withthatpredictedbyacontinuousSTAR-BURST99modelwithSalpeterIMF.Theparametersusedforthiscalculation,alongwiththederivedfield
SFRfor
massivestarsbyafactorof2.16toextendtheIMFdownto1M⊙,wefindvalueswithinafactor∼2–3forthefivegalaxiesincommonbetweenthetwoworks(NGC3310,NGC4670,NGC5253,NGC7552,andTol1924−416);ahighvalueforthe
10
arymodelpredictionsofBruzual&Charlot(2003)forstarclustersasafunctionofage,withthreeseparatemetallicitymodels.AlthoughlocalstarburstgalaxiestendtohaveLMC/SMC-typemetallicity,thetracksfortheselowermetallicitiespoorlyreproducedobservedcolorsofredgiantstars(Massey&Olsen2003).Theagreementbetweenmodelsandobservationsarebetteratsolarmetal-licity,andthereforeweincludeclustertracksatthismetal-licity.ThebottompanelineachfigureplotsM−Icolor.NoneofthedatahavebeencorrectedVversustheVforanyinternalextinction.
Foreachgalaxy,NGC4214,NGC4449,andNGC5253,theresolvedandunresolvedpointsfallinthesamepa-rameterspace.Ingeneral,theU−Vvs.V−Icolor-colordiagramsofresolvedobjectsdetectedinNGC4214,NGC4449,andNGC5253showamostlyyoungstellarpopulation(<∼10Myr)whencomparedwithstellarevo-lutionarymodelsofstarclusters.Theseresolvedobjectscontinuedowntorelativelyfaintmagnitudes.InspectionoftheWFPC2imagessuggeststhatsomeoftheobjectswhichhave∆elsareactuallyFWHMmeasurementsgreaterthan0.2pix-blendsoftwoormorestars.However,themajorityofresolvedobjectsappeartobeindividualsources.Theseresolvedobjects,whichappeartobegoodstarclustersplusahandfuloffaintbackgroundgalaxies,haveluminositiesextendingdowntoatleastMV=−5,despitetheirverybluecolors.Wealsoseefromthelo-cationintheWFPC2imagesthatindividual,bluepointsourcesaredispersedthroughouteachstarburstgalaxy.Theverybluecolorssuggestthattheseareyoung,mas-sivestars.However,withoutspectroscopy,thedegeneracyinopticalcolorsmakesitdifficulttodeterminewhethertheseareOorBstars.BasedonthefieldsignaturefoundinourSTISspectroscopy,atleastinthesethreegalaxies,thedispersedpopulationofbluepointsourcescannotbeOstars.Weconcludethatthecombinationofphotome-tryandspectroscopysupportsthepresenceofadispersedpopulationofBstarsineachstarburst.
IntheCMDsplottedinthelowerpanels,wedon’tseeevidencefordifferencesintheabsoluteluminositiesofre-solvedandunresolvedsources.Inallthreegalaxies,wefindunresolvedsourceswhichhaveabsoluteVmagnitudesbrighterthan−8,whichisrelativelyrareforindividualstars.Duetotheirhighluminosity,thesemayactuallybesuperpositions,groups,orverycompactclusters,andnotindividualstars.
BecauseitappearslikelythatanyindividualstarsintheWFPC2imagesareBratherthanOstars,whichhavesignificantlylongerlifetimesandcouldeasilyhaveformedinclustersthathavesincedissolved,thedatadonotre-quireon-goingstarformationinthefield,assuggestedforNGC1705byTosietal.(2001)andAnnibalietal.(2003).InNGC1705,theseauthorsfound,viacomparisonofthedatawithsyntheticcolor-magnitudediagramtechniques,thattoaccountfortheblueststarsintheobservedCMD,requiredaburstofstarformation∼3Myrago.Ifthissce-narioisaccurate,itimpliesthatindividualmassivestarsareforminginthefieldregionsofNGC1705,andnotinstellarclusters.Thisconclusiondiffersfromours,sincewefindnoevidenceforveryyoungOstarsforminginthefield,anditispossiblethatthedispersedBstarpopulationorig-inallyformedinclustersmorethan10Myrago.Canthesedifferentconclusionsbereconciled?Onepossibility,isthat
anumberoftheunresolved,blueobjectsinNGC1705areactuallyyoung,lowmassstarclusters.InNGC4214,NGC4449,andNGC5253wefindevidenceforresolved,andhencelikelystarclusters,withbluecolorsdowntofaintmagnitudes.NGC1705isevenfurtherawaythanthesethreegalaxies,withadistancemodulus(m−M)0.26,makingitmoredifficulttosortoutcompact0of28.54±clustersfromindividualstarsbasedonsizemeasurements.HigherangularresolutionstudiesofNGC1705couldsettlethisissue.
5.4.SpatialProfileAnalysisandImplicationsforthe
FieldStellarPopulationInthissection,weuseourcrudefitstolowlevelpeaksinthespatialcuttogleanadditionalinformationconcern-ingthenatureofthediffuselight.Becausethecontribu-tiontothefieldissubjecttobiaseswithgalaxydistance,whererelativelyhighermassclustersmaybe“hiding”inthefieldofmoredistantgalaxies,weagainfocusontheclosestgalaxiesinoursample:NGC4214,NGC4449,andNGC5253alllocatedwithin4Mpc.Forlocalpeakswithinthebackgroundwhichcanbereasonablywellisolatedanddon’tshowobvioussignsofbeingblends,wedeterminedwhethertheywereresolvedorunresolved(precisesizees-timatesattheselowS/Nvalueswouldhavelargeuncer-tainties).Thesethreegalaxiesshowfractionsofresolvedlowlevelpeaksinthefield(typicallywefound∼13–17suchpeakspergalaxy)from∼30%to60%.BasedonourmeasurementofthespatialprofileofthestarGD71,unresolvedobjectshaveaFWHM∼0.08′′.IntheSTISspatialcuts,weconsiderapeak′′toberesolvedifithasaFWHMmeasurementof0.09orgreater.AtthedistancesofNGC4214,NGC4449,andNGC5253,thismeansthatcompactclusterswithFWHMmeasurementssmallerthan∼0.2pcwouldbeconsideredunresolved.Notethatwhencomparedwiththebrighterpeakswhichwehavedefinedasclusters,asmallerrelativefractionofthefaintpeaksinthefieldareresolvedwithinagivengalaxy.
Basedonthefieldspectralsignature,weknowthatthesearenotindividualOstars,althoughtheyarepreciselyinthefluxrangewherewewouldexpecttoseeindividualOstarsincludedinourdefinitionofthefield.CouldwebeseeingthecontinuafromadispersedBstarpopula-tion?TocheckiftheobservationsaredeepenoughtodetectthecontinuafromindividualB-stars,wefirstes-timatedtheapproximatedetectionlimitforcontinuainNGC4214fromtheSTISspectra(few×10−16ergs−1at1500˚16A),NGC4449(few×10−16ergs−1),andNGC5253(10−ergs−1)inourfieldregions.Wethenestimatedthe(unreddened)fluxexpectedforaB0supergiantat1500˚A,assumingatemperatureof30,000K,andamassof17.5M⊙fromKurucz(1993)modelatmospheres.WefindthatinthemostoptimisticscenariowhereasingleB0supergianthasnoreddening,thefluxwouldbejustaboveourdetectionlimit.However,thisscenarioisnotverylikely,sincelowermassBstarsaremuchmorenu-merousthanB0stars,andbecauseitisunlikelythatsuchstarshavenoreddening,asdemonstratedforexample,byacomparisonofspectraltypeandobservedcolorforBstarsintheHumphreys(1978)catalog.
Therefore,weconcludethatwhilesomeoftheweakerunresolvedpeaksinourfieldspectracouldbethecontinua
fromindividualearlyBsupergiants,itisunlikelythatalloftheunresolvedpeakscomefromindividualBstars.ThisdoesnothoweverprecludeapopulationofdispersedBstarsfromcontributingtothecontinuumlevelobservedinourUVspectroscopyofthefield.
BecauseourcalculationssuggestthatindividualBstarsarenotresponsibleforalloftheindividualfaintpeaksseeninthefieldportionsofourSTISspectra,itislikelythatatleastsomeoftheunresolvedpeakswhichweobserveinthefieldregionsoftheselocalstarburstsareeitherchancesuperpositionsofBstars,orcompactstarclusters.Thequestionthenbecomesaretheseyoung,lowmassclusters,oroldermoremassivebutfadedclusters.
WecanalsouseourphotometrytohelpplacelimitsonthepropertiesofstarclusterscontributingtoourSTISspectra,sinceitispossibletomatchindividualsourcesseeninarchivalUbandWFPC2imageswiththeobjectsinourslit.IfwerestricttheobjectsamplesintheWFPC2imagestothosewithin50pcoftheSTISslit,wesetapproximatedetectionlimitsMandNGCV=−5.8,−5.3,−5.9forNGC4214,NGC4449,5253respectively,fromavailableWFPC2imaging.AcomparisonoftheU−Vvs.V−IcolorsofresolvedobjectstothestellarevolutionarymodelsofBruzual&Charlot(2003)suggeststhatthesesourcesarealmostexclusivelyyoung,withlikelyagesofseveralMyr(althoughagesderivedfromintegratedpho-tometry,particularlyforsuchlowmassclusters,aresub-jecttorelativelylargeuncertainties).Byscalingthelu-minosityofa7Myrstarclusterof106M⊙(MV=−15.5accordingtoSTARBURST99models),wefindthatanyyoungclusterswhicharenotobservedinthearchivalWFPC2imageswouldhavemassesofseveralhundredso-larmassesandless(i.e. PopulationOurmainresultisthatthediffuseUVlightinstarburstslacksthestrongO-starwindfeaturesobservedinneigh-boringUV-brightstarclusters.Inthissection,weexplorefourpossiblescenarioswhichcanexplainthisresult: •Starformationoccursinbothclustersandthefield,butthemostmassivestarsinthefieldremaininthedeeplyenshroudedphaselonger,andthustodonotcontributetotheUVflux. •Starformationoccursinboththefieldandclusters,butthereislittleornohighmassstarformationoccuringinthefield.ThishypothesisrequiresthatstarsformedinclustersandthefieldhavedifferentIMFs. •Starformationoccursprimarilyinclusters,andwhatwearecallingthefieldiscomposedofyoung,butlowermassclusters(thelowmassendofthepow-11 erlawclustermassfunction),whichdonotformOstars. •Starformationoccursprimarilyinclusters,andasclustersagetheydissolve,releasingtheirremainingstarsintothefield. In§5.5.1–5.5.4below,wediscusseachscenariointurnincludinggeneralimplicationsforfieldformation.5.5.1.AreDeeplyEnshroudedOStarsHidinginthe Field?Onescenariowhichcouldexplaintheobserveddiffer-enceinmassivestarsignaturesbetweenclustersandthefieldisthepossibilitythatstarformationoccursinbothenvironments,butthatfieldOstarsremainobscured(atUVwavelengths)ontimescalescomparabletotheirlife-times,andthuslongerthansimilarstarsinclusters.Therearetwomechanismswhichinfluencewhenamassivestarbecomesvisibleatoptical/UVwavelengths:(1)radiationandwindsfrommassivestarsblowouttheparentmolecu-larcloud,and(2)OBassociationstendtodriftawayfromtheirparentclouds,sincestarsandgasexperiencedifferentratesoffriction,andthushavedifferentvelocities.Inordertoreproduceboththewindfeaturesandcontinuumslopesobservedinclustersandfieldspectra,individualOstarsformedinthefieldcanneitherblowofftheirnatalmate-rialnordriftawayfromtheirparentcloudontimescalesof∼6–7Myr,otherwisetheirsignaturewouldbeobservedinourfieldspectra.ThisrestrictionwouldnotapplytoOstarsformedinclusters.Thetypicaltraveltime,al-thoughquiteuncertain,hasbeengivenas1–3Myr(deJong&Brink1986;Leisawitz&Hauser1988). Theproblemwiththisscenarioisthatitrequiresfine-tuningofbothdriftandblowouttimescalesforfieldOstars,whichmustinturnbelongerthanthoseforsim-ilarstarsformedinclusters.Therefore,weconcludethatitisunlikelythatdeeplyembedded,individualOstarsareforminginthefieldregionsofthesestarbursts,andwehavejustmissedthembecauselittleUVfluxcanescape. 5.5.2.ConstraintsonFieldIMF Ifstarformationisoccuringinbothfieldandclusters,thelackofmassivestarsobservedinthefieldcouldariseiftheIMFdiffersfromthatinclusters.Wefoundin§5.2.2thatinstantaneousburstSTARBURST99modelsolderthan6–8Myrproducereasonablygoodfitstothefieldspectra.However,theseareunlikelytoberealis-ticrepresentationsofspatiallyextendedfieldregions.Be-causethesefieldregionscoveraminimumlineardistanceof∼100pc,itisunlikelythatallthestarsformedatthesametime.Here,wecomparecontinuousstarformationmodelswiththeextractedfieldspectraanddiscussimpli-cations.Thesemodelstake∼10Myrtoequilibrate,afterwhichtheFUVspectrumchangesverylittleasafunctionofage.Therefore,wefixtheageat50Myr,andassumeastandardSalpeterIMF(slopeα=−2.35),andlowermasscutoffMtroscopyareratherlow=1M⊙(theresultsofourUVspec-insensitivetothelowermasscutofftotheIMF).Therearetwopossiblescenarioswhichcanex-plainthelackofmassivestarsobservedinfieldregions;bothassumeimplicitlythatstarformationisoccurringinsituinthefield.Inthefirstsetofmodels,wefixthe 12 ageofthecontinuousstarformationepisodeat50Myr,assumeaSalpeterIMF,anddeterminewhetherourfieldspectraareconsistentwithanuppermasscutoffwhichdif-fersfromthatforclusters.ThesecondsetofcontinuousstarformationmodelsfitsforthebestIMFslope,withanuppermasscutoffof100M⊙. Forthefirstsetofmodelswherewefitfortheupperstel-larmasscutoff,wehaverecordedthebestfitincolumn2ofTable7.Qualitatively,thelackofmassivestarswhichisexpressedbyolderageswhencomparingwithinstanta-neousburstmodels,translatestouppermasscutoffslowerthan100M⊙.TheeffectofvaryingtheuppermasscutoffisshowninthesecondcolumnofplotsinFigure3.Clearly,astheuppermasscutoffisloweredfrom100M⊙,thesig-natureofmassivestars,particularlythePCygniprofilesforNVλ1240,SiIVλ1400,andCIVλ1550,disappears.Inordertoreproducetheobservedfieldspectra,theupperstellarmasscutoffinthesemodelshastobeloweredfromthe100M⊙usedintheinstantaneousburstmodels,andwhichprovidesagoodfitfortheclusters.Forthefieldspectrainmosttargetgalaxies,particularlyinthenearestgalaxiesandthosedistantoneswhichalsodonotshowthesignatureofmassivestarwindlines,thepreferreduppermasscutoffis30–50M⊙. AsecondpossibilityisthattheIMFofstarsformedinthefieldissteeperthanthoseforminginclusters.Forexample,Masseyetal.(1995)findfieldstarsasmassiveasthoseformedinclustersandassociationsintheMag-ellanicClouds,evenafterensuringthatrunawaystarsareexcluded.Wetestedthispossibilitybycomparingacon-tinuousSTARBURST99modelfixedatanageof50Myr,withloweranduppermasscutoffsof1M⊙and100M⊙respectively,andallowedtheslopeoftheIMF,α,tovarybetweenanormalSalpetervalueof−2.35andthe−5.0foundbyMasseyetal.(1995)forthefieldinportionsoftheMagellanicClouds.Themostimportantinfluenceofthemassivestarsisseeninthewindlines,andsoweonlyweightedNVλ1240,SiIVλ1400,andCIVλ1550inourfits.InthethirdcolumnofplotsinFigure3noteinpar-ticulartherapidlydecreasingstrengthintheNVλ1240line.Wefindingeneralthatthepreferredvalueofαisbe-tween−3.0and−3.5.ThisissteeperthanSalpeter,butnotassteepasfoundfortheMagellanicClouds.Com-parisonofourextractedfieldregionswithcontinuousstarformationmodelsruleoutanIMFslopeassteepas−4.0forourgalaxies.Duetothepossibilitythatourextractedfieldregionsincludestarclusters,andhencesomeaddi-tionalmassivestarswhichwouldresultinaflatterslope,ourvaluesofαrepresentafirmlowerlimittothefieldIMFslope.Therefore,ifstarformationisoccuringinsituinthefield,andnotasaresultofdissolvingoragingstarclusters,thefieldismuchlesslikelytoproducemassivestarsthantheclusterenvironment. 5.5.3.ConstraintsontheContributionofaYoung ClusterPopulationtotheFieldTheultravioletemissionfromthefieldregionssurround-ingstarclustersinourtargetstarbursts,inallcaseswiththeexceptionofHe2-10,lacksthestrongO-starwindlinesofNV,SiIV,andCIVthataresignaturesofthemostmassivestars.Thislackofhighmassstarsisnotduetoun-dersamplingfieldregionsinnearbystarbursts,asthefieldregionscontainaconsiderableamountoflight.Ifwereject thattheIMFslopeandupperstellarmasscutoffoffieldregionsdiffersfromthatfoundinclusters,therearetwore-mainingpossibilitieswhichcanexplainboththeobservedspectralsignatureoftheUVfieldandthestrengthofthefaintpeaksinthefieldregions.Onepossibilityisthatthefieldiscomposedofclusterswhicharecoevalwiththoseobservedinourslit,buthavelowermasses.Thesecondisthatthefieldismadeupofolder/dissolvingstarclus-ters.Thedifferenceinthesetwoscenariosisprimaryoneofage.Althoughgalaxiesalmostcertainlyformclusterswitharangeinbothageandmass,forsimplicity,weas-sumeherethatthepresumedpopulationformedcoevallywiththeluminousclusters.Theseclusterswouldrepre-sentthelowerextensionoftheobservedpowerlawclusterluminosityfunction(seee.g.,Whitmoreetal.2003andreferencestherein). WedeterminedintheprevioussectionthatclusterswithagesofseveralMyrandmassesofseveralhundredsolarmasseswouldnothavesufficientS/Ntobeclassifiedasclustersinourspectra,eveningalaxiescloserthan4Mpc.Furthermore,despitetheiryouth,itisreasonabletoex-pectthatsuchlowmassclustersmaybelackinginthemostmassiveOstars,yetstillbeabletoformBstars.Therearetwocurrentargumentsregardinghowtheup-permassportionofthestellarIMFispopulatedinlowermassclusters.Thefirstsuggeststhatthedistributionofstellarmassesisarandomsamplingoftheinitialstellarmassfunction,whichnaturallywouldresultintheforma-tionofveryfewmassiveOstarsinlowermassclusters.Forexample,arecentstudybyOey,King,&Parker(2004)ofthenumberofOBstarsperclusterorassociation(N∗)intheSmallMagellanicCloud,concludesthatthedistribu-tionofN∗isconsistentwiththesebeingthemostmassivestarsingroupsoflowermassstars.Thesecondargument(e.g.,Weidner&Kroupa2004)suggeststhatthereisafundamentalupperstellarmasslimitwhichdependsonthetotalmassofacluster.FromtheirFigure4,aclusteratthedetectionlimitofourstudywouldhaveanupperstellarmasslimitof∼10M⊙.Weperformedasim-pleexperimenttoinvestigatewhetherstarswithmasses>∼20M⊙(∼Ostars)arelikelytoforminclusterswithmassesofafew100M⊙.Werandomlydrewstarswithmassesbetween0.1and100M⊙,fromadistributionhav-ingaSalpeterslope(−2.35).Outof1000suchsimulatedclusters,roughly200–300hadatleastonestarmoremas-sivethan20M⊙.Therefore,ifyoung,lowermassclustersdominatethediffuseUVlight,ourlimitsimplythattheremustbeaslidingupperstellarmasslimit,whichisrelatedtothetotalmassofthecluster,assuggestedbyWeidner&Kroupa(2004). 5.5.4.ConstraintsontheTimescaleofDissolving ClustersContributiontotheFieldBecausethefieldregionsofstarburstscontainlessmas-sivestarsthantheclusters,afourthpossibilitythatthemaindifferencebetweentheclusterandfieldstellarpop-ulationsistrulyoneofage.Thefieldcouldbecomposedofstarswhichoriginallyformedinclusters,butwhichdis-solvedandreleasedtheirremainingstarstothefield.Ifthefieldistheproductofdissolvingstarclusters,thanwecanplaceconstraintsonthetimescalesthatclustersinthesestarburstsmustdisintegrate.BasedontheyoungestanddominantagesofthefieldcompiledinTable4when 13 comparedwiththoseforstarclusters,wesuggestthatclus-terswhichcontributetothefieldneedtobedestroyedonveryrapidtimescales,oforder7–10Myr. Theevolutionofclustersystemsingeneralincludesanumberofdisruptiveprocesses,suchasmasslossfromstellarevolution,andstellarevaporationduetoexter-nalgravitationalshocksandinternaltwo-bodyrelaxation(e.g.,Fall&Zhang2001).However,thesemechanismsoc-8 curoverrelativelylongtimescales(>∼10yrs)whencom-paredwiththe7–10Myrestimatedabove.ArecentstudyoftheagedistributionofstarclustersintheAntennaegalaxiessuggestsaveryrapiddecline.Fall(2004)andFall,Chandar,&Whitmore(2004)findthatthenumberofclustersasafunctionofagefallsbyafactor∼10bythetimetheclusterpopulationhasreachedanageof10Myr.Thisrapiddeclineisseenfordifferentmassrangesabove3×104M⊙.Numbercountsofembeddedclustersinthesolarneighborhoodwithmasseslowerbyfactors10–102alsoshowasimilarsteepdecline(Lada&Lada2003).Falletal.(2004)suggestthatthisshortdisruptiontimescaleindicatesthatthemajorityof(butnotall)starclustersintheAntennaeendupgravitationallyunbound,evenifthecloudfromwhichtheyinitiallyformedwasinitiallybound.Theionizingradiation,stellarwinds,andsuper-novaeexplosionsfrommassivestarscouldeasilyremoveasignificantfractionoftheISMfromaprotocluster,leav-ingthestarswithinitgravitationallyunboundandfreelyexpanding.Ifthisisthecase,suchclusterswouldbecomeverydifficulttodetectafter∼10Myr.Theveryshorttimescalesforclusterdissolutioninferredfromourdatainlocalstarburstgalaxies,ifdissolvingclustersarere-sponsibleforthediffuseUVfieldemission,areconsistentwiththetimescalesofclusterdisruptionpredictedbythefreeexpansionmodel.TheadvantageofusingUVspec-troscopy,aswehavedonehere,istheabilitytotrackthepresenceofthemostmassivestars,anddirectlytranslatetheseintoagedifferencesbetweenclustersandthefield. 5.6.HighS/NFieldTemplate Thedifferencesobservedinthefieldandclusterspectraofourgalaxysamplehavepotentiallyimportantimplica-tionsfortherest-UVobservationsofgalaxiesathighred-shift.Becauseoftheirsmallangularextent,entiregalaxiesareobservedathighredshift,whilethebestlocalcounter-partsarestudiedinamorepiecemealway.BecausethediffuseUVlightfromstarburststypicallydominatestheoutputfromstellarclusters,itisveryimportanttoin-cludethisasaningredientinspectralsynthesisofgalaxiesathighredshift(althoughSteideletal.1996showthatacomparisonofz∼3LymanBreakgalaxyspectrawiththoseofindividuallocalclustersalsogivereasonablefits).Wehavecombinedtheextractedfieldspectraforfivemetal-poorgalaxies(Mkn33,NGC4214,NGC4449,NGC4670,andNGC5253)whichdonotshowthesig-natureofmassivestars,inordertomakeahighS/Ntem-plateofalowmetallicity“field”spectrum9(Figure6).Typicalabundancesforthegalaxiesare∼1/3–1/5solar,whichisagoodmatchtoabundancesmeasuredforhighredshiftgalaxies(Pettinietal.2000).Thishasacharac-teristicspectrumwhichis“older”thanindividualclusters(i.e.lackingthesignatureofmassivestars),andiswell 9The representedbycontinuousstarformationmodels. 6.SUMMARYANDCONCLUSIONS WehaveusedSTISlong-slitFUVspectraoftwelvelo-calstarburstgalaxiestostudythestellarcontentofthediffuse,UVluminousfieldregionsfoundbetweenpromi-nentstarclusters.TheextractedspectraarecomparedwithSTARBURST99stellarevolutionarysynthesismod-els.He2-10isuniquetoourstudy,inthatitcontainsthesignatureofmassiveOstarsinthefieldregions.Thecompositefieldspectrumisverysimilartothatforfourco-evalclustersinourslit.Weestimatethat40%ofthelightinthefarUVcomesfromobservedcompactstarclusters,providingafirmlowerlimittotheamountofUVlightoriginatinginclustersinthisstarburst. WiththeexceptionofHe2-10,theclustersandfieldregionsintheother11targetgalaxiesexhibitpronounceddifferences.MostoftheUV-brightclustersarequiteyoung(<∼6Myr),andshowthestrongPCygniprofilesfoundinOstars.Theneighboringfieldregionshowever,clearlylackthesewindfeatures.Inparticular,thenearestgalax-ies(NGC4214,NGC4449,andNGC5253),aswellasanumberofmoredistantones,haveB-stardominatedfieldspectra. WeincludeananalysisofUVIWFPC2imagingforthesethreeclosestgalaxies,inordertobetterunderstandthestellarpopulationswhichcontributetothefield.Pho-tometryofobservedsourcesinallthreegalaxiesrevealspopulationsofblueresolvedandunresolvedsources.Thecolorsandluminositiesofresolvedobjectsareindistin-guishablefromthoseofunresolvedsources.Thissuggeststhepresenceoflowmass,resolvedstarclusters,aswellasadispersedpopulationofbluestars.BecauseweseenoevidenceforthepresenceofOstarsinthefieldregionsofthesegalaxiesbasedonour(limitedcoverage)STISspec-troscopy,wesuggestthattheseareadispersedpopulationofBstars. Thespatialprofilesalongtheslitshowthatthefieldre-gionsarenotsmooth,butrathercontainnumerousfaintpeaksandvalleys.AnanalysisofthefaintpeaksinthefieldregionsofourSTISspectrasuggeststhatthesearisefromdiscretestellarpopulations.Roughly30–60%ofthesepeaksareresolved.Ourcalculationsshowhowever,thattheSTISspectraarejustsufficienttodetectthecontinuafromunreddened,individualearlyBsupergiantsinthefield.Therefore,whilethepopulationofdispersedBstarsdiscoveredinthearchivalWFPC2imagesprobablydocon-tributetothefieldspectra,theyareprobablynotresponsi-bleforallofthefaintpeaksweobserve.Weconcludethatasignificantfractionofthesefaintpeaksarelikelysmallgroupsorclustersofstars,regardlessofwhethertheyareresolvedornot. Weexplorefourpossiblescenariostoexplainourobser-vationthatthefieldcontainslowermassstarsthanneigh-boringclusters. (1)Ifstarformationoccursinsituinbothclustersandthefield,andOstarsformedinthefieldstayintheen-shroudedphaselongerthantheircounterpartsinclusters,thenthesestarswouldnotcontributemuchtothefieldUVflux.Thisscenariowouldrequirea“fine-tuning”ofbothseculardriftvelocitiesandblow-outtimescalesfor templatefieldspectraareavailablefromtheSTARBURST99website:http://www.tsci.edu/science/starburst 14 fieldOstars.Thereforewebelievethatitisunlikelythatdeeplyembedded,individualOstarsareforminginthefieldregionsofstarbursts. (2)Ifstarformationoccursinbothfieldandclusterenvironments,thefieldIMFmustdifferfromthatfoundinclusters.IfthefieldIMFhasanormalSalpeterslope,ourdataareconsistentwithanuppermasscutoffof30–50M⊙.Alternatively,thefieldIMFslope,α,issteeperthanSalpeter,withbestfitvaluesof−3.0to−3.5. (3)Ifstarformationisoccuringprimarilyinclustersandassociations,thenthefieldcouldbecomposedofyoung,coevalbutlowermassclusters.Ourphotometryandspec-troscopybothimplylimitsofseveral100M⊙forgroupsorclustersofyoungstarswhichcouldbehidinginourfieldspectra.ThislimitisconsistentwithalackofOstarsonlyifthereisanupperstellarmasslimitwhichscaleswiththetotalclustermass.Soeffectively,thisissimilar toscenario(2). (4)Ifstarformationoccursprimarilyinclusters,buttheseclustersdissolvetocreatetheobservedfield,ouranalysissuggeststhatclustersmustdissolveonveryrapidtimescales,oforder7–10Myr.Thisisconsistentwitharecentlypresentedscenariowheremoststarclusters,al-mostindependentoftotalmass,freelyexpandandrapidlydisrupt. Wethanktheanonymousreferee,whosesuggestionsimprovedthepresentationofthispaper.Wearegrate-fulforsupportfromNASAthroughgrantGO-09036.01-AfromtheSpaceTelescopeScienceInstitute,whichisoper-atedbytheAURA,Inc.,forNASAundercontractNAS5-26555.C.A.T.acknowledgessupportfromNASAgrantNAG58426andNSFgrantAST-0307386,andG.R.M.acknowledgessupportfromNASAgrantNAG5-13083. REFERENCES Annibali,F.,Greggio,L.,Tosi,M.,Aloisi,A.,&Leitherer,C.2003,AJ,126,2752 Barth,A.J.,Ho,L.C.,Filippenko,A.V.,&Sargent,W.L.1995,AJ,110,1009 Bertin,E.,&Arnouts,S.1996,A&AS,117,393Bianchi,L.etal.2005,ApJ,619,L71 Biretta,J.A.,etal.2000,WFPC2InstrumentHandbook,version5.0,STScI,Baltimore Boeker,T.,vanderMarel,R.P.,Mazzuca,L.,Rix,H-W.,Rudnick,G.,Ho,L.C.,&Shields,J.C.2001,aj,121,1473Bruzual,G.,&Charlot,S.2003,MNRAS,344,1000 Calzetti,D.,Meurer,G.R.,Bohlin,R.C.,Garnett,D.R.,Kinney,A.L.,Leitherer,C.,&Storchi-Bergmann,T.1997,AJ,114,1834Chandar,R.,Leitherer,C.,Tremonti,C.A.,&Calzetti,D.2003,ApJ,586,939 Chandar,R.,Leitherer,C.,&Tremonti,C.A.2004,ApJ,604,153Cheng,K.-P.,etal.1992,ApJ,L29Davidge,T.J.19,PASP,101,494 deJong,T.,&Brink,K.1986,inStarFormationinGalaxies,ed.C.J.LonsdalePersson(NASA-CP2466),323 deMello,D.F.,Leitherer,C.,&Heckman,T.M.2000,ApJ,530,251 Dolphin,A.2000,PASP,112,1397 Guseva,N.,Izotov,Y.I.,&Thuan,T.X.2000,ApJ,531,776 Fagotto,F.,Bressan,A.,Bertelli,G.,&Chiosi,C.1994a,A&AS,104,365 Fagotto,F.,Bressan,A.,Bertelli,G.,&Chiosi,C.1994a,A&AS,105,29 Fall,S.M.,2004,inFormationandEvolutionofMassiveYoungStarClusters,ed.H.J.G.L.M.Lamers,A.Nota,&L.J.Smith(SanFrancisco:ASP),inpress(astro-ph/04050) Fall,S.M.,Chandar,R.,&Whitmore,B.C.2004,submittedtoApJLetters Fitzpatrick,E.L.1999,PASP,111,63 Greggio,L.,Tosi,M.,Clampin,M.,deMarchi,G.,Leitherer,C.,Nota,A.,&Sirianni,M.1998,ApJ,504,725 Haas,M.,Lemke,D.,Stickel,M.,Hippelein,H.,Kunkel,M.,Herbstmeier,U.,&Mattila,K.1998,A&A,338,33 Harris,J.,Calzetti,D.,Gallagher,J.S.III,Conselice,C.J.,&Smith,D.A.2001,AJ,122,3046 Heckman,T.M.,Robert,C.,Leitherer,C.,Garnett,D.R.,vanderRydt,R.1998,ApJ,503,6 Hippelein,H.,Haas,M.,Tuffs,R.J.,Lemke,D.,Stickel,M.,Klaas,U.,&Volk,H.J.2003,A&A407,137Holtzman,J.A.,Burrows,C.J.,Casertano,S.,Hester,J.J.,Trauger,J.T.,Watson,A.M.,&Worthey,G.1995,PASP,107,106Humphreys,R.M.1978,ApJS,38,309 Hunter,D.A.1999,inIAUSymp.193,“Wolf-RayetPhenomenainMassiveStarsandStarburstGalaxies”,eds.K.A.vanderHucht,G.G.Koenigsberger,&P.R.J.Eenens(SanFrancisco:ASP),418 Izotov,Y.I.&Thuan,T.X.1999,ApJ,511,639 Kobulnicky,H.A.,Kennicutt,R.C.,Jr.,&Pizagno,J.L.1999,ApJ,514,544 Kurucz,R.L.1993,CD-ROM13,ATLAS9StellarAtmosphereProgramsand2km/sGrid(Cambridge:SmithsonianAstrophys.Obs.) Lada,C.J.,&Lada,E.A.2003,ARA&A,41,57 Lamers,H.J.G.L.M.,Panagia,N.,Scuderi,S.,Romaniello,M.,Spaans,M.,deWit,W.J,&Kirshner,R.2002,ApJ,566,818Larsen,S.S.1999,A&AS,139,393 Larsen,S.S.&Brodie,J.P.2000,AJ,120,2938 Leitherer,C.1998in“TheStellarInitialMassFunction”,eds.G.Gilmore&D.Howell.ASPConferenceSeriesv.142,1998,p61 Leisawitz,D.,&Hauser,M.G.1988,ApJ,332,954Leitherer,C.etal.1999,ApJS,123,3 Maiz-Apellaniz,J.,Walborn,N.R.,Galue,H.A.,&Wei,L.H.2004,ApJS,151,103 Maoz,D.,Barth,A.J.,Sternberg,A.,Filippenko,A.V.,Ho,L.C.,Macchetto,F.D.,Rix,H.-W.,&Schneider,D.P.1996,AJ,111,2248 Massey,P.,Lang,C.C.,Degioia-Eastwood,K.,&Garmany,C.D.1995,ApJ,438,188 Massey,P.,&Olsen,K.A.G.2003,AJ,126,2867 Meurer,G.R.,Heckman,T.M.,Leitherer,C.,Kinney,A.,Robert,C.,&Garnett,D.R.1995,AJ,110,2665 Meynet,G.,Maeder,A.,Schaller,G.,Schaerer,D.,&Charbonnel,C.1994,A&AS,103,97 Misiriotis,A.,Popescu,C.C.,Tuffs,R.J.,&Kylafis,N.D.2001,A&A,372,775 Oey,M.S.,King,N.L.,&Parker,J.W.2004,AJ,127,1632 Pastoriza,M.G.,Dottori,H.A.,Terlevich,E.,Terlevich,R.,&Diaz,A.I.1993,MNRAS,260,177 Pettini,M.,Steidel,C.C.,Adelberger,K.L.,Dickinson,M.,&Giavalisco,M.2000,ApJ,528,96 Popescu,C.C.,Misiriotis,A.,Kylafis,N.D.,Tuffs,R.J.,&Fischera,J.2000,A&A,362,138 Popescu,C.C.etal.2005,ApJ,619,L69Salpeter,E.E.1955,ApJ,121,161 Schlegel,D.J.,Finkbeiner,D.P.,&Davis,M.1998,ApJ,500,525Steidel,C.C.,Giavalisco,M.,Pettini,M.,Dickinson,M.,&Adelberger,K.L.1996,ApJ,462,L17Steson,P.B.1987,PASP,99,191 Thilker,D.A.etal.2005a,ApJ,619,L79Thilker,D.A.etal.2005b,ApJ,619,L67 Tosi,M.,Sabbi,E.,Bellazzini,M.,Aloisi,A.,Greggio,L.,Leitherer,C.,&Montegriffo,P.2001,AJ,122,1271 Tremonti,C.A.,Calzetti,D.,Leitherer,C.,&Heckman,T.M.2001,ApJ,555,322 vandenBergh,S.2004,AJ,128,1880 Weidner,C.,Kroupa,P.2004,MNRAS,348,187 15 Table1 SampleGalaxyProperties Galaxy AdoptedDistance (Mpc) AreaProjectedinSlita (pc) Reference References.—(1)Kobulnicky,Kennicutt,&Pizagno1999(2)Davidge19(3)Guseva,Izotov,&Thuan2000(4)Heckmanetal.1998(5)Pastorizaetal.1993(6)Boekeretal.2001(7)Izotov&Thuan1999 BasedonDistancegiveninColumn2,assuming25′′coverageofthegalaxyandgiventheslitwidthoftheobservations a Table2 FractionofFar-UVLightFromClustersandtheField Galaxy Clusters Field Note.—Valuesdeterminedbysummingupthelightfromspatialplotsforallclustersandcomparingwiththetotallightintheslit.Thespatialplotsaresummedover700pixels,covering∼1250–1700˚A.BecausetheslitlocationswerechosentomaximizethenumberofUV-luminousclustersalignedintheslit,thesefractoinsrepresentlowerlimitstothetotalfractionoffieldlightineachgalaxy. 16 Table3 ArchivalWFPC2Data Galaxy EB−Vaforeground FiltersExposureTime[sec] F336W(U)F555W(V)F814W(I) Foreground,MilkyWayextinctionfromtheSchlegeletal.(1998) maps a Table4 ComparisonofMeanAges(Myr)ofDominantClustervs.FieldStellarPopulations Galaxy BrightClusters FaintClusters Field Note.—TheagesinMyrandassociateduncertaintieshavebeenderivedbycomparingtheobservationswithinstantaneousSTARBURST99models,assumingaSalpeterIMFandupperandlowermasscutoffsof100M⊙and1M⊙respectively.Theun-certaintiesweredeterminedbyusingthebootstraptechniqueasdescribedinTremontietal.(2001) 17 Table5 ComparisonofContinuumSlopesinClustervs.FieldStellarPopulations Galaxy FUVslope(β)a BrightClustersFaintClusters Field ThefarUVslope,β,measuredaftercorrectingforforeground extinction.Typicaluncertaintiesonβmeasurementsare±0.1. a Table6 MeanStarFormationRatesDerivedforSTISFieldRegions Galaxy pixelsa EB−Vb L1500c SFRd #Ostarse Totalnumberofpixelssummedupinthespatialdirectionforeachfieldregion. Reddeningderivedusingthestarburstobscurationcurvefromthedifferenceinthemodelslope(−2.5),andtheobservedslopeofthefieldaftercorrectionforforegroundreddening. −1 Extinctioncorrectedluminosityat1500˚Ainergss−1˚Amea-suredfromtheSTISspectra.WeusetheEB−Vvaluesquotedincolumn3.cb a StarformationrateinM⊙yr−1kpc−2.WehaveassumedastandardSalpeterIMF(α=−2.35),andanuppermasscutofffortheIMFof100M⊙.Theageforthecontinuousstarformationmodelhasbeenfixedat50Myr. d NumberofOstarspredictedbySTARBURST99continuousstarformationmodels(assumingastandardSalpeterIMFandMup=100M⊙)toresideinthefieldregiongiventhemeanSFRgivenincolumn5. e 18 Table7 BestFittingFieldModelsGalaxy Mupa αb Bestfitwhenfieldspectraarecom-paredwithcontinuousformationSTAR-BURST99modelswhichhaveSalpeterIMF,Mlow=1M⊙,andavariableuppermasscutoff. BestfittotheIMFslope,α,incontin-uousformationSTARBURST99model. b a 19 3•10−11 225 He 2−10 Parsecs450675 9001125 1.5•10−11 0.51.0 Kpc 1.52.0 Mkn33 2.5 Summed Flux1•10−11 Summed Flux2•10−111.0•10−11 5.0•10−12 00 200400600800 Pixels along spatial direction Kpc 1.42.84.25.6 N1741 0 10007.0 0 200400600800 Pixels along spatial direction Kpc 0.280.560.841.12 N3125 10001.40 3.0•10−112.5•10−112.0•10−11Summed Flux1•10−118•10−12Summed Flux6•10−124•10−122•10−12 0 1.5•10−111.0•10−115.0•10−12 0 −5.0•10−12 0 200400600800Pixels along spatial direction Parsecs 3256509751300 N3310 10001625 0 6•10−122.0•10−111.5•10−11Summed Flux200400600800 Pixels along spatial direction Parsecs 70140210280 N4214 1000350 Summed Flux4•10−12 1.0•10−11 2•10−12 5.0•10−12 00 200400600800 Pixels along spatial direction 1000 00 200400600800Pixels along spatial direction 1000 A)asafunctionofpositionalongtheslitforeachtargetgalaxyinthisstudy.Fig.1.—Theintegratedflux(from700pixels,∼1250–1700˚ Theportionsofthespectrawhichwereusedtocreatethecompositefieldspectraareplottedinblack.TheSTISMAMAdetectorshaveaplatescaleof0.024′′pix−1. 20 5•10−12 95 190Parsecs285 380 475 N4449 4•10−12xulF3•10−12 demmu2•10−12S1•10−12 0 02004006008001000Pixels along spatial direction 5•10−12 80160Parsecs 240320 400 N5253 4•10−12xulF3•10−12 demmu2•10−12S1•10−12 0 02004006008001000Pixels along spatial direction 1.5•10−11 0.501.00Kpc 1.502.00 2.50 N7552 x1.0•10−11 ulF demmuS5.0•10−12 00 2004006008001000 Pixels along spatial direction Fig.1.—Continued 5•10−12 0.39 0.78Kpc 1.17 1.56 1.95 N4670 4•10−12xulF3•10−12 demmu2•10−12S1•10−12 0 0200400600800 1000Pixels along spatial direction 2.0•10−11 1.152.30Kpc 3.454.60 5.75 N5996 1.5•10−111.0•10−11 5.0•10−12 0 02004006008001000Pixels along spatial direction 2.0•10−11 0.901.80Kpc 2.703.60 4.50 Tol1924−416 1.5•10−11 1.0•10−11 5.0•10−12 00 2004006008001000 Pixels along spatial direction Summed FluxSummed Flux21 4•10−12 N4214 Field 3•10−12Summed Flux2•10−12 1•10−12 0150 200250300Pixels along spatial direction 350 Fig.2.—AnenlargementofthefluxalongthespatialdirectionforaportionofNGC4214(ingrey)containingsomefieldregions(inblack).Thisfigureshowshowthefieldlightisnotsmooth,butcontainsanumberoffaintpeaksandvalleys. 1.00.5 1.0 Rectified Flux0.5 1.00.5 1.00.5 Instantaneous Continuous Continuoust=5 Myr Mup=100M .O α=-2.35 t=6 Myr Mup=50M .O α=-3.0 t=7 Myr Mup=30M .O α=-3.5 t=8 Myr 130014001500Wavelength (A) Mup=10M .O 130014001500Wavelength (A) α=-4.0 130014001500Wavelength (A) Fig.3.—ThefigurepresentsexamplespectrafromvariousSTARBURST99modelsconsideredinthiswork.ThestrengthoftheSiIVλ1400andCIVλ1550windfeatures,plusotherdiagnosticsprovideexcellentconstraintsonthemassivestarcontentinnearbygalaxies.ThefirstcolumnshowstheeffectofagingontheUVspectrumofaninstantaneousburststellarpopulation,withMup=100M⊙.ThesecondsetofpanelsshowsthevariationsinacontinuousstarformationepisodewhentheuppermasscutofffortheIMFisloweredfrom100M⊙to10M⊙.AndthethirdcolumnshowsUVspectraforcontinousstarformationmodelswithdifferentIMFslopes,α.Thedurationforthecontinuousstarformationshownincolumns2and3is50Myr. 22 Flux (10−16 erg s−1 cm−2 Å−1) Composite He 2−10 Clusters Flux (10−16 erg s−1 cm−2 Å−1) 4020 15105 Composite Mkn33 FieldComposite Mkn33 Clusters 755025 201510501200 1082 Composite He 2−10 Field Flux (10−16 erg s−1 cm−2 Å−1) 1300 14001500Wavelength (Å) 16001700 1200 1300 14001500Wavelength (Å) 16001700 Composite NGC3125 Clusters 302010 1510501200 200150 NGC1741 Cluster 1 Flux (10−16 erg s−1cm−2Å−1) 2010 10 NGC1741 Faint Clusters Composite NGC3125 Field Flux (10−16 erg s−1 cm−2 Å−1) 5 1200 Composite NGC1741 Field 1300 14001500Wavelength (Å) 1600 1700 1300 14001500Wavelength (Å) 16001700 15 Composite NGC3310 Clusters 100 105 4020 1200 Composite NGC3310 Field Flux (10−16 erg s−1cm−2Å−1)50 15105 302010 NGC4214 Cluster 1 NGC4214 Faint Clusters 1300 14001500Wavelength (Å) 16001700 1200 Composite NGC4214 Field 1300 14001500Wavelength (Å) 1600 1700 Fig.4.—Unweightedsumofstarclusterspectraandsummedfieldspectraextractedforourtargetgalaxies. Composite NGC4449 Clusters 30 )1−Å 20 2−mc 1−s10 gre 61−15 0 1( xu10 lF5Composite NGC4449 Field 0 12001300 1400150016001700 Wavelength (Å) 20 )115 −Å 2−m10c 1−s 5gre 6 Composite NGC5253 Clusters 1− 0 1( 30 xulF2010 Composite NGC5253 Field 12001300 1400150016001700 Wavelength (Å) Composite NGC7552 Clusters 15 )1−Å 10 2−mc 1−s5 gre 6 1− 0 Composite NGC7552 Field 1( 6 xulF42 1200 1300 140015001600 1700 Wavelength (Å) Fig.4.—Continued 23 80 )160−Å 2−m40c 1−s 20gComposite NGC4670 Clusters re 6 1− 0 1( 30 xulF2010 Composite NGC4670 Field1200 1300 1400150016001700 Wavelength (Å) 4030 2010NGC5996 Cluster 1 )1− Å 2 −m8c 1−s g6re 614−01( 2xulF NGC5996 Faint Clusters 8 6 4 2Composite NGC5996 Field 12001300 140015001600 1700 Wavelength (Å) 75 50)1−Å 2−mc25 1−s gre 6 Composite TOL1924−416 Clusters 1− 0 1( 25 xulF2015 Composite TOL1924−416 Field1200 1300 1400150016001700 Wavelength (Å) 24 -2N4214-2N4449-1-1U-V0resolvedunresolvedU-V0resolvedunresolved1-12-10 1-12-10 MV-8MV0V-I1-22N5253-8-6-4-1-6-4-10V-I12-1U-V0resolvedunresolved1-12-10 MV-8-6-4-10V-I12Fig.5.—Two−colorandcolor−magnitudediagramsforNGC4214,NGC4449,andNGC5253areshown,basedonarchivalWFPC2imaging.ThetoppanelsshowtheU−Vvs.V−Icolor-colordiagrams,correctedforforegroundextinctiononly.Resolved(redcircles)andunresolved(bluecrosses)sourcesareplottedseparately.TheevolutionofsimplestellarpopulationsfromtheBruzual&Charlot(2003)modelsareshownforthreedifferentmetallicities:solar(solidline),1solar(dashed).Thedirectionofthereddening50 678vectorisshownbythearrows,whichrepresentsEB−V=0.1.Agesof10,10,and10areshownasblackfilledcirclesalongthecluster models,startingfromtheupperleft.Inthebottompanels,weshowtheVvs.V−Icolormagnitudediagrams.Overplottedare30M⊙,15M⊙,and9M⊙Padovastellarevolutiontracks(Fagottoetal.1994a,b),with1 25 Fig.6.—RectifiedspectrumoftheunweightedsumoffieldregionsforsixlowmetallicitygalaxieswhichhaveB-stardominatedfieldspectra(MKN33,NGC4214,NGC4449,NGC4670,NGC5253,andNGC1741).
因篇幅问题不能全部显示，请点此查看更多更全内容
Copyright © 2019- huatuo0.cn 版权所有 湘ICP备2023017654号-2
违法及侵权请联系：TEL:199 18 7713 E-MAIL:2724546146@qq.com
本站由北京市万商天勤律师事务所王兴未律师提供法律服务