First detection of Lyman continuum escape from a local starburst galaxy

cESO2008󰀎

FirstdetectionofLymancontinuumphotonescapefromalocal

starburstgalaxy

arXiv:astro-ph/0601608v2 27 Jan 2006ObservationsoftheluminousbluecompactgalaxyHaro11withtheFar

UltravioletSpectroscopicExplorer(FUSE).I.⋆

6¨NilsBergvall1,E.Zackrisson2,B-GAndersson3,D.Arnberg4,J.Masegosa5,andG¨oranOstlin

1

2

3

45

6

Dept.ofAstronomyandSpacePhysics,Box515,S-75120Uppsala,Sweden

e-mail:nils.bergvall@astro.uu.se

Dept.ofAstronomyandSpacePhysics,Box515,S-75120Uppsala,Swedene-mail:erik.zackrisson@astro.uu.se

DepartmentofPhysicsandAstronomy,JohnsHopkinsUniversity,3400NorthCharlesStreet,Baltimore,MD21218,USAe-mail:bg@pha.jhu.edu

Dept.ofAstronomyandSpacePhysics,Box515,S-75120Uppsala,SwedenInstitutodeAstrofisicadeAndalucia,Granada,Spaine-mail:pepa@iaa.es

StockholmObservatory,SCFAB,SE-10691Stockholm,Swedene-mail:ostlin@astro.su.se

Received7July2005/Accepted2November2005

ABSTRACT

Context.Thedominatingreionizationsourceintheyounguniversehasnotyetbeenidentified.Possiblecandidatesincludemetalpoordwarfgalaxieswithstarburstproperties.

Aims.Weselectedanextremestarburstdwarf,theBlueCompactGalaxyHaro11,withtheaimofdeterminingtheLymancontinuumescapefractionfromUVspectroscopy.

Methods.SpectraofHaro11wereobtainedwiththeFarUltravioletSpectroscopicExplorer(FUSE).AweaksignalshortwardsoftheLymanbreakisidentifiedasLymancontinuum(LyC)emissionescapingfromtheongoingstarburst.Fromprofilefittingtoweakmetallineswederivecolumndensitiesofthelowionizationspecies.AdoptingametallicitytypicaloftheHIIregionsofHaro11,thesedatacorrespondtoahydrogencolumndensityof∼1019cm−2.ThisrelativelyhighvalueindicatesthatmostoftheLyCphotonsescapethroughtransparentholesintheinterstellarmedium.WethenusespectralevolutionarymodelstoconstraintheescapefractionoftheproducedLyCphotons.

Results.AssuminganormalSalpeterinitialmassfunctionweobtainaLymancontinuumescapefractionoffesc∼4–10%.Wearguethatinahierarchicalgalaxyformationscenario,theupperlimitwederivefortheescaperateallowsforasubstantialcontributiontocosmicreionizationbystarburstdwarfgalaxiesathighredshifts.

Keywords.Galaxies:evolution-formation-starburst-dwarfs,Ultraviolet:galaxies,Cosmology:diffuseradiation

1.Introduction

AccordingtoWMAPdata,theepochofreionizationstartedataredshiftofz≈20(Kogutetal.2003).ObservationsofthehighopacityintheLymancontinuum(hereafterLyC)emis-sionofgalaxies(Beckeretal.,2001,Fanetal.2002)andtheresultsfromtherecentstudiesoftheintergalacticmedium

2NilsBergvalletal.:FirstdetectionofLymancontinuumphotonescapefromalocalstarburstgalaxy

ithasbeenproposedthatminiquasarsfedbyintermediate-massblackholesmayhavebeenimportant(Madauetal.2004).Atlowerredshifts,observationsofasampleofLyman-breakgalaxiesatz∼3.4(Steideletal.2001)seemtoshowthatgalaxiescouldaccountforasignificantfractionofthephotonfluxrequiredforreionization.Theobservationsindicateacon-tributionofLyCphotonsafactorof∼5higherthanforquasarsatthesameepoch.Theseresultshavehowevernotbeencon-firmedbyothersimilarstudies.Onthecontrary,ithasbeenargued(Fern´andez-Sotoetal.2003)thatamaximumof4%oftheLyCfluxcouldescapefromluminousgalaxiesatred-shifts1.9Wehavesearchedforlocaldwarfswithpropertiessimilartoyoungdwarfgalaxies,thatcouldbeusedastestobjects.Afewsuchstudieshavealreadybeencarriedoutbutonlypro-videdupperlimitstotheescapefraction(e.g.,Giallongoetal.1997;Devriendtetal.1998;Shulletal.1999).However,pre-dictedequivalentwidthsoftheBalmeremissionlinesinstar-burstsareinsomecaseslargerthanobserved(e.g.Bergvall1985,Bresolinetal.,1999,Mas-Hesse&Kunth1999,Moyetal.2001,Stasinskaetal.2001).Severalpossibleexplana-tionstothiswerediscussedbyMas-Hesse&Kunth(1999).Onepossibilityisthatthemostmassivestarsremainhiddenintheirparentalcloudsuntilmostofthefuelhasbeenconsumed.ThiswouldmimicatruncationoftheIMFathighmasses.Mas-HesseandKunthalsodiscussedthepossibilitythatalargefractionoftheionizingphotonsareabsorbedbydustintheHIIregions.Toobtainagreementwiththeobservations,alargeamountofthephotons,∼60%wouldhavetobeabsorbed.ThispossibilityhasrecentlybeeninvestigatedbyHirashitaetal.

(2003)wheretheescapefractionoftheLyCphotonsfromtheHIIregionitselfisestimatedfrommodelcomparisons.Basedoninputdatafromasampleofstarforminggalaxiestheyfindthat,inthemeanabout40%ofthephotonsaretrapped,whichislessthanwhatwouldbeneededaccordingtoMas-HesseandKunth.Butthescatterishighsoinsomecasesthiscertainlywouldbeapossibility.Hirashitaetal.seemtofindthattheescapefractionishigherisgalaxieswithpropertiessimilartoHaro11(e.g.theBCGESO338-IG04).Thereforeanalterna-tive,orcomplementary,explanationforthelowfluxofthere-combinationlines,maybethatasignificantfractionoftheLyCphotonsisinfactescapingfromthegalaxy.Thisisthepossi-bilitythatwestudyinthispaper.

HIIregionsinnearbygalaxiesappeartoleak40–50%oftheLyCphotonsproducedbythehotstarsintothediffuseinterstellarmedium(Fergusonetal.1996;Oeyetal.1997).AlthoughitcannotbecompletelyexcludedthatthegasmaybeionizedbysingleOstarslocatedinsitu,acomparisonoftheobservedemissionlinestrengthswiththeoreticalmod-els(Iglesias-P´aramo&Mu˜noz-Tu˜no´n2002;Wood&Mathis2004)givesupporttotheleakageinterpretation.AnevenmoreextremecaseofadensityboundedHIIregionaroundamas-sivestarclusterwasreportedbyLeithereretal.(1996).Theyestimatethatabout75%oftheLyCradiationisleakingintothediffuseambientmedium.Dependingonthedistributionofthegas,thelocationwherethesephotonsareabsorbedmaybeverydifferentfromcasetocase.InsomesituationsthediffuseionizedgasmaybemuchmoreextendedthantheHIIregionproducingthephotonsandwillbeignoredbytheobserverwhoconsequentlywillunderestimatethestarformationrate.

TheotheralternativeforthediscrepancybetweentheoryandobservationsoftheBalmeremissionlinestrengthsisthatthedustisunevenlydistributedacrossthestarburstregionsothattheyoungerregionsaremorereddenedthantheolderones(Calzettietal.1994).ThiscouldreducetheequivalentwidthofHαwithuptoafactorof≈2.Theonlywaytoreallydis-tinguishbetweenagloballeakageintotointergalacticmediumandtheothertwoalternativesistosearchforleakingphotonsshortwardoftheLymanlimit.

Inpreviousstudiesofstarburstdwarfs(Bergvall1985,

¨¨Bergvall&Ostlin2002,Ostlinetal.1999,2001)wehave

focusedonafewluminous(–19>MB>–21)metalpoorbluecompactgalaxies(BCGs).TheyhavepropertiessimilartotheLBGsatintermediateredshifts(e.g.Guzmanetal.2003).HerewereportonsuchastudyusingFUSE,theFarUltravioletSpectroscopicExplorer.

1.1.ThetargetgalaxyThetargetgalaxyofourchoiceisHaro11(=ESO350-IG38)atα2000=00h36m52.5s,δ2000=-33◦49′49′′.Fromtheemission-linespectrumwederiveaheliocentricradialvelocityofv0=6180kms−1,correspondingtoadistanceof86Mpc,assumingaHubbleconstantof72kms−1Mpc−1.Wehavepreviouslystud-iedthisgalaxyoverawidefrequencyrange,fromUVtoradio

¨(Bergvall&Olofsson1986,Bergvalletal.2000,2002,Ostlin

etal.,1999,2001).Ithasuniquepropertiesthatmakesitoneof

NilsBergvalletal.:FirstdetectionofLymancontinuumphotonescapefromalocalstarburstgalaxy3

themostrelevantlocalobjectsforourpurpose.Thechemicalabundancesarelow,theoxygenabundancebeing∼20%solar(relativetotherevisedsolaroxygenabundancebyAsplundetal.2004b).

AnestimateofthetypicalhydrogenmassofaBCGcanbeobtainedfromthecorrelationbetweengasmassandmajordiameter(Gordon&Gottesman1981).Haro11hasamajordiameterof11kpcatµB=25magarcsec−2indicatingthattheHImassshouldbeapproximately2×109M⊙.Despitesev-eraleffortsatVLA,NancayandParkes(allunpublished)wehavenotbeenabletodetectHIinthisgalaxy.WeestimatetheupperlimittoM(HI)∼108M⊙.Theextraordinarystrong[CII]λ158µline,observedwithISO(Bergvalletal.2000),showsthatalargepartoftheHIgasislocatedinphotodis-sociationregions(PDRs).WehaveestimatedthecontributionfromH2andHIIgastobetween108and109M⊙each.Thiswouldleavearemarkablysmallfractionofthetotalgasmassinatomicform.Apossibleexplanationisthatthebulkofthegasinthehalothatnormallyisinneutralstatehasbecomeionizedbythestarburst.OurHαobservationsshowthatthemainbodyisindeedsurroundedbyanextendedhaloofionizedgas.TheconditionsforLyCphotonescapethusappeartobeunusuallyfavourable.ItisinterestingtonotethatthenearbyBCGPox186,sometimescalled’ultracompact’,isalsoHIquiet(Begum&Chengalur2005).

HSTGHRSUVspectroscopicobservationsofHaro11werecarriedoutbyKunthetal.(2003)whofoundaquitecom-plexabsorptionsystemwithvelocitycomponentsofLyαinab-sorptiononbothsidesoftheemissionline.Thusitappearsthattherearebothoutflowsandinfallofgasinthecentredirection.FromtheabsorptionlinesKunthetal.estimatecolumndensi-tiesandfindvaluesashighaslognH∼20cm−2.Theapertureusedissmallhowever(2′′circulardiameter)andcoversonlytheverycentralpartofthegalaxy.Ouropticalspectroscopy(Ostlin¨etal.2005)revealsvelocitygradientsofbothgasandstarsaslargeas200kms−1acrosstwoofthebrightconden-sationsintheverycentre,indicatinganextremelyhighmassdensityinthisregion.

2.Observationsandreductions

2.1.IUEobservationsIn1984weobservedthegalaxyinlowdispersionmodewiththeSWPUVspectrographoftheIUEspacecraft.Theapertureusedwas10′′x20′′.AlowresolutionspectrumisshowninFig.1.Belowwewillusetheslopeofthespectrumasaninputtothederivationofthedustextinction.

WewillalsousetheIUEspectrumtodeterminetheC/Oabundanceratio,basedontheCIII]λ1908,andOIII]λ1663lines.Thestrongestline,alsorepresentingthedominatingion-izationstageisCIII]λ1908whiletheOIII]lineisnosier.ThustheerrorinthedeterminationwillbesignificantbutitstillwillbeusefulasacheckoftheC/OratiodeterminedfromtheFUSEabsorptionlines(seesect.3.1).TheatomiccollisionstrengthsaredeterminedfromtheelectrontemperaturebyBergvall&Ostlin¨derived

(2002)T≈13400K.Fromtherelative

lineintensitieswethenobtainN(C+)/N(O++)∼0.3±0.1.The

6

Haro11 IUE

4

2

0

1200140016001800

Fig.1.SpectrumofHaro11obtainedwiththeIUEspace-craftinthelowresolutionSWPmode.Theapertureusedwas10′′×20′′.Badpixelshavebeenflaggedwithcrosses.

errorisdominatedbythephotonstatisticsandhasbeenesti-matedfromthenoiseintheregionaroundtheline.AssumingionizationcorrectionfactorsICF(C/C+)∼1.4(Stasi´nska1990)

andICF(O/O++)∼1.5(Bergvall&Ostlin¨2002),wethenfinally

obtainN(C)/N(O)=0.46+0.15

−0.10.

2.2.FUSEobservationsThedatawereobtainedonOct12,2001,usingtheFUSElowresolutionmodeLWRSwithanapertureof30′′x30′′.Thetotalintegrationtimeontargetwas16ks.12ksoftheobservationswerecarriedoutduringorbitalnight.ThedatawerereducedusingCalFUSEv3.0,withdefaultprocessingparameterforallspectralsegments.AsthebackgroundlevelinFUSEdataisquitelowandhencecannotbereliablymeasuredindetailforshortobservations,thestandardFUSEbackgroundcorrectionisbasedonasetoftemplatefileswhicharescaledtoaveragesoverthewholedetectorinanygivenobservation.Thesetem-platefilesconsistofcoaddeddatafrommany,longobserva-tionsofblankskyacquiredthroughoutthemission.TheFUSEdetectorsconsistofmicrochannelplateswithelectronicpixelswhoseexactlocationsdependontheelectricalfieldstructureandread-outelectronicsofthedetector(Sahnowetal.,2000).Artificialcountenhancementsoccurattheedgesofthedetec-torduetonon-uniformitiesinthefieldandhaveaverysteepdependenceonthedistancefromtheedge.Theexactmappingofthedetectortolocationintheelectronicdetector-imagede-pendsonseveralfactorsincludingthetotalbrightnessofthede-tectorillumination.Becauseofthesteepnessofthisedgeeffect,minormismatchesinthetargetandbackgroundimagescancausesignificanterrorsinthebackgroundsubtraction(Sahnow,2005,privatecommunication).AstheSiCspectrumfallsveryclosetotheloweredgeofdetector1B,anyrelativechangeinthestrengthofthisedgeeffectcomparedtotherestoftheback-groundwillcauseproblemsinthebackgroundcorrection,par-ticularlyforweaksourcessuchasHaro11.Tocompensatefor

4NilsBergvalletal.:FirstdetectionofLymancontinuumphotonescapefromalocalstarburstgalaxy

thiseffect,weresettheareasandrelativeweightingusedtodeterminethescalingofthebackgroundtemplateinthereduc-tionoftheSiC1Bdata.WithoutanyfurthermodificationsthederivedSiC1BspectrumthenagreeswellwiththeSiC2Aspec-trum.TheregionsofthelowestsignalinboththeSiC1BandtheLiF1Asegmentswerehoweversignificantlybelowthezerolevel,indicatingthattheproperbackgroundonthewholewasfainterthanthesubtractedtemplatebackground.Afterextrac-tion,calibrationandcombinationoftheindividualsegments,aconstantwasthereforeaddedtothebackgroundtocompensateforthis.Aswearemorestronglyconstrainedbylowsignal-to-noisethanspectralresolution,wearenotsignificantlyaf-fectedbytheslightdifferenceindispersionsolutionbetweensegments,andtheconsequentloweringofresolutioncausedbythiscoaddition.

3.Results

Fig.3showsthefullspectrumwithafewofthestrongestabsorptionandemissionlinesindicated.StronggeocoronalLymanemissionlinesareseenontopoftheabsorptionspec-trum.Thenarrowspikeat1168.7ÅisascatteredHeIsolaremissionline.Fromtheabsorptionlineswecandefinetwovelocitysystems,oneoriginatingintheMilkyWayandtheotherinHaro11.Fig.2showstheHI21-cmlinespectruminthedirectionofHaro11(l=328◦,b=-83◦),obtainedfromtheLeiden/Dwingeloosurvey(Burton&Hartmann1994).Ouranalysisshowsthatthespectruminthevelocityrange-450kms−1–400kms−1canbeapproximatedwithtwocomponentswithsmallnegativevelocitiesrelativetotheLSRasshowninTable1.

Fig.2.The21-mlineinthedirectionofHaro11obtainedfromtheLeiden/Dwingeloosurvey(Burton&Hartmann1994).VLSRistheradialvelocityrelativetothelocalstandardofrestandTAistheantennatemperature.

ThespectralfeaturesidentifiedtobelongtotheMilkyWayISMareindicatedwithhatchedverticallinesinFig.3.TheseabsorptionlineshavealinewidthofFWHM∼30kms−1in

Table1.ComponentsofthelocalHIspectruminthedirectionofHaro11.Thecolumnsshowthecentralvelocityrelativetothelocalstandardofrest,peakantennatemperature,fullwidthathalfmaximumandthederivedcolumndensity.

V(centre)T(peak)FWHMNkms−1

K

kms−1

cm−2

agreementwiththeFWHMgiveninTable1.ThenarrownessofthelinesisimportantsinceitensuresthatwecanfindtheproperleveloftheLymancontinuumofHaro11betweenthepositionsofthegalacticLymanabsorptionlines.ThespectrumofHaro11isrichinspectralfeaturesanddominatedbyabsorp-tionlinesfromOandBstars.PossibledetectionsofmolecularhydrogenlinesintheLymanserieshavealsobeenindicated.TheheliocentricsystemicvelocityofHaro11,asdefinedfromtheopticalHIIemission(unpublished)andthestellarCaIItripletabsorptionlines,weightedfromtheUVmappingofthe

galaxy(Ostlin¨etal.2005),isv=6180kms−1.Thisvelocity

agreeswiththepositionofthePVλ1118Å(λobs=1141Å)line.ThePVlineoriginatesfromhotdwarfstarsonthemainsequence.Ifthesestarsaremetal-richandmassivetheytendtoshowPCygniprofileswhileatlowermetallicities,asinourcase,thefeatureismainlyphotosphericandnarrowandcanbeusedasasystemicvelocityfixpoint.Thuswehavetwoinde-pendentderivationsofthesystemicvelocityofthestellarpop-ulationseeninthefar-UVspectralregion.Strongdeviationsfromthisvaluewouldindicatein-oroutflowsofgas.

Thestrongestabsorptionlines,likethoseoftheLymanse-ries,aredarkbutnotsaturatedasisnormallythecaseforstar-burstdwarfs(e.g.Heckmanetal.2001).Thisinitselfisanindi-cationthattheopticaldepthintheLyCislowerthannormalforthistypeofgalaxies.ManyoftheabsorptionlinesoriginatinginHaro11arebroadandshowPCygniprofiles,e.g.theOVIλλ1032,1038(redshiftedtoλ1053,1058Å)lines.MostoftheselinesareprobablyformedinthephotospheresorinthewindsofOandBstarsbuttheremayalsobeacontributionfrommoreglobaloutflows.Afteraroughcorrectionforthestellarcon-tribution(seenextsection),someofthestrongestabsorptionfeaturesshowasymmetries,indicatinghigh-velocityoutflows.Thisaspectwillbebrieflydiscussedbelowandmoreindetailinaforthcomingpaper,focusedontheabsorptionlinespectrum.

InFig.4wedisplaytheshortwavelengthpartofthespec-trum.Thisspectralregioniscoveredbytwodetectorsegments,1Band2A,producingtwoindependentspectraSiC1BandSiC2A.Wavelengthsbelow917Åareonlycoveredbyseg-ment1B.Thisisalsothedetectorsegmentwiththesmall-esteffectiveareaofthetwo.Thus,inthisregionthesignalisnoisierthanatlongerwavelengths.Whatisparticularlyinter-estingfromFig.4isthatbelowtheLymanlimitofHaro11thereclearlyseemstobeaweakbutsignificantsignalintheLymancontinuum.ShortwardsoftheLymanlimitofthelocalrestframethesignaldiminishesgradually,asexpectedfromtheincreasingdensityoftheLymanabsorptionlinesofthelo-calISM.Theemissionisseeninthe2-dimensionaldisplayof

NilsBergvalletal.:FirstdetectionofLymancontinuumphotonescapefromalocalstarburstgalaxy5

Fig.3.FUSEspectrumofHaro11.Themajorabsorptionfeatureshavebeenindicatedwithsolidlines(Haro11opticalrestframe)andhatchedlines(localframe).ThespectrumalsoshowsstronggeocoronalLymanemissionlinesandscatteredsolarHeIatλ1169Å.Thedatahavebeenrebinnedwithaboxcarofsize0.26Å.thespectralregionasafaintnarrownoisystructureinbothSiCchannels.WethereforefeelconfidentthattheLyCexcessisnotduetocalibrationproblems.Thelevelofthesignalhow-everisnotsowelldetermined,mainlybecauseofuncertaintiesinthecalibration,alowsignalandpossiblyinfluencesofauro-ralemissionlinesandemissionandabsorptionlinesfromtheMilkyWaywhichwewillnowdiscuss.

Haro11isatagalacticlatitudeof-83deg.andthusinterstellarabsorptionisnotaseriousproblem.FromSchlegelatal.(1998)weobtainAB=0.049mcorrespondingtoALy−limit∼0.18m(Mathis1990).TheLymanemissionlinesfromtheterrestrialairglowareindicatedinFig.5.Forcom-parisonweshowinFig.6boththespectrumbasedonnight-and-daydataandthespectrumbasedonnightdataonly.WecanclearlyidentifythegeocoronalLymanemissionlines.ShortwardsoftheLymanlines,atmuchlowerintensities,CIlinesmayweaklyinfluencethecounts(Feldmanetal.2001).In

thefollowingdiscussionwewillusethe900ÅrestframefluxdensityofHaro11asareferenceoftheLyClevel.Thischoiceallowscomparisonswithpreviousstudiesofothergalaxies(e.g.Leithereretal.1995,Deharvengetal.2001,Inoueetal.2002).Inthelocalframethispartofthespectrumhasbeenred-shiftedtoλ=920Å.TobeatdownthenoiseinthespectrumwedeterminetheLyClevelovertheline-freeregionsintheinter-val920-925Å,sincetheLyCsignalisnotexpectedtovarysignificantlyoversuchashortinterval.Aftercorrectingforgalacticextinction,weobtainf900=1.1±0.1×10−17Wm−2Å−1.Theuncertaintyduetothecomplexityinthedeterminationofthebackgroundisprobablysignificantlylargerthantheformalerror.Intheworstcaseitcouldamounttooftheorderof∼50%ofthesignal.

6NilsBergvalletal.:FirstdetectionofLymancontinuumphotonescapefromalocalstarburstgalaxy

Fig.6.Comparisonbetweennightanddaydata.Thelowerpanelshowsthe1σnoiselevel,correctedtothebinsizeusedintheplot.

Fig.4.PartofspectrumshowingtheLymancontinuumsignal.TheLymanlimitinthelocalframe(hatchedlines)andintheframeofHaro11(solidlines)havebeenindicated.

Fig.5.AirglowlinesandtherestframeLymanseries.

3.1.Spatialdistributionoftheabsorbers.WementionedintheintroductionthatHaro11appearstobedevoidofneutralhydrogen.ThiscouldindicatethattheLyCphotonsareleakingoutfromatruncatedStr¨omgrensphere,i.e.thattheemissionis’densitybounded’.Thehydrogenion-

izationcrosssectionattheLymanlimitis≈8×10−18cm2.ThismeansthattheLyCisessentiallyblackatcolumnden-sitiesofafewtimes1017cm−2,iftheneutralgasisdiffuselydistributed.Asdiscussedinsect.1.1wehavederivedanupperlimitofthemassofneutralhydrogeninHaro11of108M⊙.

NilsBergvalletal.:FirstdetectionofLymancontinuumphotonescapefromalocalstarburstgalaxy7Wecanusethisnumbertoestimatethemaximumcolumnden-sityofneutralhydrogenundertheassumptionthattheHIhasthesamespatialdistributionasthestarsinHaro11.Thestel-larscalelengthishV∼2.5kpc(Bergvall&Ostlin¨2002).ThisgivesanHIcolumndensityinthedirectionofthecentreofN(HI)=5×1019cm−2whichcanberegardedasanupperlimit,considerablylargerthanthelimitingcolumndensityforLymanleakage.InthissimplemodelHIwouldbecometransparentat4-5scalelengths(i.e.∼10kpc)fromthecentre,wherethehostgalaxycomponentstartstodominate.Butweknowthattheneutralgasisclumpedsowewilladaptthesimplestalternativetothehomogeneousmodel-the’picket-fence’model.HereitisassumedthattheLyCradiationisleakingoutthroughholesintheambientgas.InprinciplewecancarryoutprofilefittingtotheLymanlinesinabsorptiontoderivetheHIcolumnden-sities.TheselinesarehowevernotveryusefulsincethereisamixedcontributionfromOBstarphotospheres,ISMabsorptionandlineemissionfromtheHIIregions.Wecanmakeanesti-mateoftheinfluenceofstellarfeaturesfrommodelsofspec-traofyoungstellarpopulationsbyRobertetal.(2003).Thesemodelsarebasedonobservedstellarspectraoverarangeofmetallicities.Themetal-poormodelsarevalidforagesbelow15Myr.Weinterpolatedthemodelspectratoametallicityinagreementwithwhatweobtainedfromtheopticalspectra,i.e.anoxygenabundanceof∼18%solar(basedonthesolaroxygenabundancederivedbyAsplundetal.2004b).WealsoassumedaSalpeterinitialmassfunction(IMF)andaconstantstarfor-mationrate(SFR).Fig.7showsthebestfittothesemodels,aburstofthehighestageinthelibrary,14Myr.TheopticalspectraofthecentralburstshowstrongW-Rfeaturesand,aswementionabove,PCygniprofilesareseenintheFUSEspec-trum.Thesedonotfittothemodelspectrumshowninfig.7.AbetterfitcanbeobtainedifweassumeanincreasingSFRwiththeUVemissiondominatedbymassivestarsofanageofafewMyr.ModificationsofthestandardIMFcanfurtherim-provethefit.TheremayalsobeasignificantcontributiontothePCygnicomponentfromthesupernovageneratedsuperwind.However,theseissuesarenotofmainimportancehere.

Theinfluenceofyoungstarsinthespectrumisstrong.ThepresenceofOVIanditsPCygniprofileindicatesthepresenceofOsupergiants(e.g.Gonz´alezDelgadoetal.1997).Thustheburstmustbeactiveorstoppednotmorethan3-4Myrsago.ForacontinuousSFRwedonotexpectthespectrumtochangemuchovertheexpectedmaximumdurationofatypicalburst,∼108yrandthereforethecomparisoninFig.7shouldbequal-itativelycorrect.Themodelspectrumhasonlypartlybeencor-rectedforinterstellarabsorptionbuttheinfluenceontheLyβlineismarginalsothelineprofileshouldbeentirelydominatedbythestellarphotosphericcontribution.ThisisconfirmedalsofromacomparisonwithpredictedLyβabsorptionlinestrengthsinmodelsofOstarspectra(Gonz´alezDelgadoetal.1997).ThusitisobviousthatthesefeatureshaveastronginfluenceandthustheLymanlinescannotbeusedtoderivethecolumndensitywithoutamorecarefulcorrectionfortheinfluenceofthestellarfeatures.However,ifwecanderivecolumndensitiesoflowionizationstagesofheavyelementswithknownabun-dancesrelativetohydrogenwecanobtainastricterconstraintontheHIcolumndensitysincetheatomsandionsinthese

1.5010201025103010351040Fig.7.TheFUSEspectrumofHaro11(solidline)alongwiththebestfittingstellarspectrum(dottedline)fromRobertetal.(2003).ThewavelengthscalehasbeencorrectedtotherestframeofHaro11,assumingasystemicheliocentricvelocityof6180kms−1.ThemodelspectrumisbasedonaSalpeterIMF,astellarmassrangeof1–120M⊙,ametallicityof18%solarandacontinuousSFR.Theageis14Myr.stagestoafirstapproximationshouldcoexistwiththeneutralhydrogen.

Wewilluseafewratherwelldefinedandisolatedlowion-izationlines,CIIλ1036,OIλ1039,SiIIλ1021,ArIλ1048,FeIIλ1097andλ1125.Wecannotexcludethefactthattheremaybeaweakcontributiontosomeoftheselinesfromthestellarphotospheresbutthebulkoftheabsorptionshouldorig-inateintheISM.Theionizationpotentialsoftheselinesareslightlyhigherthan1Ryd.Itisknownfromstudiesofthelocalinterstellarmediumthatelementswithionizationpotentialsbe-lowthatofhydrogenandtheionizationpotentialofthesecondionizationstageabove,existpredominantlyinthesinglyion-izedstagewherehydrogenisneutral.ItisthereforepossibletousealsotheselinestoobtaingoodestimatesofN(HI)un-dercertainconditions.ThelinesoftheneutralatomsOandArcandirectlybeusedtoderivecolumndensitiesiftheatomsareeffectivelyshieldedfromionizingphotonsbytheneutralhy-drogen.InregionswhereHIispartiallyionized,argonmaybemorefullyionizedasaconsequenceofithighcrosssectiontophotonabsorption.Ifso,wewouldunderestimatethecolumndensitybasedonArI.ThisseemstobeconfirmedfromtheresultsinTable2.Oxygenbehavesdifferentlysincetheioniza-tionfractioniscoupledtothatofhydrogenviaresonantcharge-exchangereactions(Jenkinsetal.2000).

Theopticaldepthofaspecificspeciesisgivenby:τ(λ)=

πe2

fλ

󰀁

τ(v)dv(2)

8NilsBergvalletal.:FirstdetectionofLymancontinuumphotonescapefromalocalstarburstgalaxywhereλ,thewavelengthoftheline,isgiveninÅandvisthevelocityinkms−1.

WederivedthecolumndensitiesusingtheOwenssoftwareobtainedfromtheFrenchFUSEteam.Thelineminimumcor-respondstotheopticallyderivedradialvelocity.ThetypicalvirialvelocityofBCGsofthisluminosityis<100kms−1.Thewidthoftheabsorptionlinesaremuchlargerthanthis.Thisisprobablyduetooutflowsfromthestarburstandtoalesserex-tentalsotoinfallofneutralgas.Wehavetoadoptthefitofthelineprofilestothisconditionandtheprogrammeweusecon-venientlyallowsonetofitlinesbelongingtoseveralvelocitycomponentssimultaneously.Ourapproachwastotestanum-berofsituationsinwhichweassumedthecloudensembletoconsistofbetween3and9differentvelocitysystems,inallcasescoveringthewholewidthofthelines.InthecaseswherethenumberofcomponentswerelowthiswascompensatedbythesoftwarebyincreasingtheDopplerbroadeningparameter,b.Thebparameterwasallowedtovarybetween30and180kms−1.FUSEhasaresolutionof∼20kms−1andonemightworrythattherecouldbeasignificantnumberofunresolvedsaturatedcomponentswithlowbparametersinthesightline.FromFUSEcurve-of-growthofneutralandlowionizationab-sorptionlinesystemsintheMagellanicbridgeoneobtainsbparametersintherange15-20kms−1(Sembachetal.1979,Lehner2002).TheH2linesgivesmallerbvalues,indicatingastrongerconfinementandlowerturbulenceoftheseclouds.WedonotexpecttheconditionsinHaro11tobemorequietthanintheMagellanicbridgeandthereforeweshouldbesafewithalowervalueofthebparameterofthelowionizationlinesof30kms−1.Intheexamplewediscussbelow,however,weallowtheparametertovaryfreely,utilizingthefullresolutionofFUSE.Inonesetupofparameterswefixedtherelativevelocitydiffer-encesbetweenthedifferentcomponentstothesamevalue,sothatthefullwidthofthelineswascovered.Wealsotestedfitswhereboththevelocitydifferencesandthelinewidthswereal-lowedtovaryfreely.Inallcasesweassumedagastemperatureof300K.Theabsorptionlinesfromallatomsandionsunderstudywereincludedsimultaneouslyinthefits.Theoptimumfitswerefoundtocontainbetween5-7components.Ideallyonewouldexpecttherelativestrengthsofthecomponentstobethesameforeachelement.Asthecomponentsgetweakerandnoiserhowever,theresultsbecomemoreandmoreuncer-tain.Thuswhilethecentral3-4componentsshowagoodagree-ment,thefaintercomponentsmaydeviatesignificantlyfromonespeciestotheother.Thegoodnessoffit(lowesttotalχ2residual/numbersofdegreesoffreedom)showedsmallvaria-tionsbetweenfitstodifferentnumbersofsubcomponents.Asanexample,oneofbestresultsforthecarbonlineisshowninFig.8.

FromTable2weobtainanoxygentocarbonratioofN(C)/N(O)∼0.5.ThisagreeswiththevaluederivedfromtheIUEspectra(seesect.2.1).Thuswefeelconfidentthatthepro-cedureweappliedabovegivesreliableresults.FromthefitswederivedthecolumndensitieslistedinTable2.WehavepreviouslyobtainedtheoxygenabundanceofHIIregionsinHaro11,log(N(O)/N(H))+12=7.9.Wewillassumethatthisisanupperlimitoftheoxygenabundanceintheneutralregion.Applyingtheconversiontohydrogencolumndensitiesfrom

1.50105610581060Fig.8.ThebestfreefittotheobservedCIIλ1036linebasedon5velocitycomponentswithdifferentDopplerwidthsandcomponentseparations.Thestrongestcomponentiscenteredonthesystemicvelocityasderivedfromtheopticaldata.Table2.ColumndensitiesoflowionizationspeciesSpeciesλlogNσlogNthisnumberweobtainalowerlimittothehydrogencolumndensity,logN(HI)=19.6.Wehavenocorrespondingdetermi-nationoftheremainingelementsbutifweassumetherela-tiveabundancestobethesameasthesolar(Asplundetal.,2004a)weobtainalowerlimittologN(HI)ofbetween19.6and20.3withthelowervalueforthestrongcarbonline.Theserelativehighvaluesarenotconsistentwithadiffusedistribu-tionofHIsinceitwouldprohibitleakage.Weconcludethatthephotonsmustbeleakingoutthroughtransparentwindowsinthegaseoushaloina”picket-fencefashion”.Thiswouldalsomeanthattheestimatedcolumndensitiesareslightlytoolow,duetothedilution.Apossiblescenarioexplainingwhyweob-servealeakageisthatthemergingprocesshasallowedyoungstarclusterstoseparatefromhighdensitygaseousregionsonatimescaleshorterthanthelifetimeoftheburst,i.e.inashortmomentofstronggravitationaldisturbances.

Inthefollowinganalysiswewillassumethatthepicket-fencemodelisthebestapproximationofthesituation.

3.2.ExtinctioncorrectionsInthenextsectionwewillderiveanestimateoftheLyCescapefractionbasedonboththeobservedHαfluxandthecontinuumfluxesat900and960ÅrestwavelengthofHaro11,f900andf960.Thefluxeswereintegratedoverspectralregionsofwidths5and2Årespectively.Asmentionedabove,the900Åcontin-

NilsBergvalletal.:FirstdetectionofLymancontinuumphotonescapefromalocalstarburstgalaxy9

uumwavelengthwaschosentoallowadirectcomparisonwithpreviousstudiesofLymanleakage.The960ÅwavelengthwaschosenforspectralevolutionarymodelcomparisonssinceitrepresentsaspectralregionreasonablyclosetotheLymanlimitbutunaffectedbyLymanabsorptionlinewingsandotherstronglines.Theobserveddataneedtobecorrectedforgalacticandinternaldustextinction.Thegalacticextinctioninthedi-rectionofHaro11isAB=0.049(Schlegeletal.1998).The900and960Årestwavelengthsareredshiftedto∼920and980Åintheobservedspectrum.AtthesewavelengthAB=0.049correspondstoA920=0.177andA980=0.175(Mathis1990).Wewillnowdiscusstheinternalextinctioncorrections.

Inthepicket-fenceapproximation,asisfurtherdetailedinthenextsection,wewillassumethattheLyCphotonsarees-capingthroughdustfreewindows.Theonlygoalforthere-mainingpartofthissectionisthustoobtainavalueoftheex-tinctioncorrectiontothe960Åregion,A960.FromtheHα/Hβratioofthecentral4′′x4′′,correctedforunderlyingabsorption

¨(BergvallandOstlin2002),weobtain,assumingtheCalzetti

etal.(1994)extinctionlaw,aninternalextinctioninBofAB=1.08m±0.07m.TheFUSEapertureis30′′×30′′,i.emuchlargerthantheapertureoftheopticalspectroscopy.Thefluxinsidethe4′′×4′′apertureisonlyabout10%ofthetotalfluxsoanestimateoftheglobalextinctionbasedonthesedataisquiteuncertain.IfweassumethatthedustcolumndensityisproportionaltotheBbandfluxweusethesmallaperturedatatoderivetotalextinctioninBtobeAB=0.66m±0.08m.Theer-rordoesnotincludetheuncertaintyofthemodelofthedustdistribution.

WecanalsoestimatetheextinctionfromtheobservedslopeoftheUVcontinuuminHaro11,usingamethoddevelopedbyCalzettietal.(1994,1995)andMeureretal.(1995).Calzettietal.(1994)investigated39starburstandbluecompactgalax-iestostudytheinterstellarextinctionintheUV.TheobservedspectraldistributioninthisregionmaybeapproximatedbyapowerlawF(λ)∝λβ.TheyfoundthattheopticaldepthobtainedfromtheHα/Hβratiocorrelatedwiththeslopeβandderivedavalueofβinadustfreeenvironment,β0=-1.7.TheyalsofoundthattheopticaldepthcalculatedfromtheHα/Hβratiowasafactorof2largerthanthatderivedfromthecontinuumunder-lyingtheemissionlines.Thiswasinterpretedasaconsequenceofasystematicallyhigherextinctioninyoungstarformingre-gionsascomparedtoolderregions.

Later,Meureretal.(1995)foundacorrelationbetweenβandtheratiobetweenthefarinfraredfluxandtheUVfluxat2200Å.ThiscorrelationagreedverynicelywithaforegrounddustscreenmodelfollowingtheCalzettietal.extinctionlaw,ifthegalaxieshadasimilarintrinsicUVcontinuumdistribu-tion.Theobservedslopecanthereforebeusedtoderiveanesti-mateoftheextinctionbycomparingtothe”standard”dustfreeslope.ThevaluederivedbyMeureretal.β0=-2.5±0.5,isbluerthanwhatCalzettietal.found.ThisisprobablyduetothefactthattheMeureretal.samplecontainedmorelow–luminositystarburstdwarfsthatwerecompletelydominatedbytheyoungpopulationwhilethemoremassivegalaxiesintheCalzettietal.samplehadaslightlylargercontributionfromolderstars.TheextinctionofHaro11thereforedependsonwhichvaluewechoose.Haro11isinvolvedinaveryintenseglobalstarburst

andismetalpoor.Ontheotherhanditisquiteluminousandmostlikelyhasquiteacomplexstarformationhistorysoweexpectamixtureofstellarpopulationsofdifferentages.ThisputsHaro11somewhereinbetweentheMeureretal.sampleandtheCalzettietal.sample.WewillthereforeassumethatitsdustfreeUVcontinuumisrepresentedbytheweightedmeanoftheMeureretal.valueandtheCalzettietal.value.Bothhaveasimilarobservedscatteraroundthemeansothevaluewewilladoptisβ0=-2.1±0.2.

TheUVslopeinHaro11,obtainedfromourIUEobser-vations,coveringthewavelengthinterval1250Å–1980Åisβ=-1.4±0.1.(Inthepicket-fencemodelitisincorrecttousetheslopewithoutcorrectingforthecontributionfromtheunobscuredpart,butthiswouldmakeanegligiblediffer-encehere.)AsanadditionalcheckofthisvaluewecanusetheobservedfluxratioFFIR/FUVofHaro11andapplythemodelmostfavouredbyMeureretal.tocalculatethepre-dictedvalueofβ.FFIRisderivedfromIRASdataandtheap-proximationfromLonsdale&Helou(1985).ThisresultsinFFIR=1.26·10−14(2.58F60+F100)Wm−2=2.74·10−13Wm−2.Thefluxat2200Åwasobtainedfromanextrapolationofthefittotheshorterwavelengthregion.ThisgivesusFFIR/FUV=5.9.Fromthisvalueweobtainβ=-1.6±0.1,agreeingnicelywiththeformervaluewederived.β=-1.4correspondstoanextinctioninBofAB=0.83,whichisquiteclosetothevalueweobtainfromtheHα/Hβratio(AB=0.66).WewillassumeavalueofAB=0.7±0.2.

Thefinalvalueoftheinternalextinctionwasobtainedintwosteps.FirstweadoptedtheCalzettietal.extinctionlaw,rel-evantinthewavelengthregion1250Å–8000Å,todeterminethereddeningcorrectionat1250Å.ThenweusedtherecentdeterminationoftheextinctioncurveintheSMCbyCartledgeetal.(2005)togofromthereto960Å,theregionwewillusetosamplethestellarcontinuum.WethusfinallyadoptaninternalextinctionA960=2.9m±0.3.

Wewillnow,withtheuseofspectralevolutionarymodels,deriveanestimateoftheupperandlowerfluxescapelimits.

3.3.Estimateoftheescapefraction3.3.1.Comparisonwithspectralevolutionarymodels

Undertheassumptionofasimplegeometryforthedistributionofstars,gasanddust,limitsontheglobalLyCescapefractionfromHaro11maybeinferredfromtheobservedLyCflux.Inprinciple,theLyCleakagecouldbecausedeitherbyholesinthegas(e.g.supernovachimneys),byatotalgasmasstoolowtoformacompleteStr¨omgrensphere,orbyacombinationthereof.SincethehighcolumndensitiesofneutralhydrogenderivedinSect.3.1disfavouratruncatedStr¨omgrensphereforHaro11,wewillhereestimatetheescapefractionassumingthattheescapeisexclusivelyduetoanisotropicdistributionofholes,whichareassumedtobedevoidofgas(neutralandionized)aswellasdust.

FollowingDeharvengetal.(2001),wedefinethetotalLyCescapefractionfescasfesc=

L900,obs

10NilsBergvalletal.:FirstdetectionofLymancontinuumphotonescapefromalocalstarburstgalaxy

whereL900,obsrepresentstheobserved(escaping)luminosity(WÅ−1)atawavelengthcorrespondingto900Åintherestsystemofthegalaxy,andL900,⋆representsthecorrespondingluminosityproducedbythestellarcomponentbeforeattenua-tionbygas.Undertheapproximationsadoptedhere,L900,⋆isgivenbythesumoftwocomponents:L900,⋆=L900,obs+L900,abs,

(4)

whereL900,absrepresentstheluminositylostthroughgasab-sorptioninsidethegalaxy.L900,absisinturnrelatedtothenum-berofLyCphotonsabsorbedbygasNLyC,absthroughL900,abs=NLyC,abs×10−k1,

(5)

wherek1isamodel-dependentparameter.Undertheassump-tionofcaseBcombination,thenumberofabsorbedLyCpho-tonscanthenberelatedtotheextinction-correctedHαlumi-nosityLHαthroughNβ

LyC,abs=

λHαeffLHα,(6)

Hβ

wherejHα/jHβistheintrinsicHα/Hβlineratio,αBistheCase

Brecombinationcoefficientandαeff

cientfortheHβline.TheseHβistheeffectiverecombina-tioncoeffithreeparametersdependonthetemperatureofthenebulaand,tolesserextent,onthedensity.Byassuminganelectrondensityof100cm−3andfit-tingacubicsplinetothejHα/jHβ,αBandαeff

weestimatethatfortheHelectronβvaluesgivenbyOsterbrock(19),tempera-tureofHaro11(T≈13400K;Bergvall&Ostlin¨2002):NLyC,abs≈8.56×1011LHα.

(7)

Combinationoftheequationsabove,withtherelationL=4πD2fbetweenluminosityL,HαfluxfHα,fluxdensityfλanddistanceD,gives:ffesc=

900,obs

L),(9)

900,abs

weusetheZackrissonetal.(2001,hereafterZ01)spectralevo-lutionarymodel.Withthismodel,agridofburst-likespectralevolutionarysequencesisgenerated,definedbythemodelpa-rametervaluesspecifiedininTable3.Inthecaseofastan-dardSalpeterIMF(dN/dM∝M−α;α=2.35withuppermasslimitMup=120M⊙)therangeofk1valuesallowedbythisgridofstarformationscenariosbecomesk1=13.23–13.24.WhenthefullrangeofIMFvariationsconsidered(α=1.35–3.35andMup=20–120M⊙)isused,thisconstraintrelaxestok1=12.70–13.58.

Whenthesevaluesofk1arecombinedwiththepreviouslydiscussedmeasurementsoff900,obs=1.06±0.09×10−17Wm−2Å−1andfHα=4.5±0.5×10−15Wm−2,limitsonfescmaybe

inferred,resultingin0.04≤fesc≤0.10forthestandardIMF

and0.01≤fesc≤0.11forthemoregenerousrangeofIMFscenarios.

ThesefescestimatesrelyheavilyontheconversionfromLHαtoL900,abs,andareonlyvalidundertheassumptionthattheHαlineisdominatedbyphotoionizationbystars,i.e.thation-izationcontributionfromshocksoranactivegalacticnucleusarenegligible.WehavealsoassumedthedusttobeexclusivelylocatedoutsidetheStr¨omgrensphere,sothatnoLyCphotonsarelostduetointernaldustextinction.Potentialfar-UVopac-itysourcesoutsideHaro11,otherthandustintheMilkyWay,havefurthermorebeenneglected.

Aconsistencytestmaybecarriedoutbycomparingtheob-servedslopeofthespectrumacrosstheLymanlimit,tothecor-respondingslopepredictedbymodels.Todothis,wemeasurethecontinuumfluxdensityatarestwavelengthof960Å,wherethespectrumshouldbecompletelyunaffectedbygaseousab-sorption.AftercorrectionforGalacticextinction,theobservedfluxdensityatthiswavelengthisf960,obs=7.4±0.4×10−17Wm−2Å−1.Insect.3.2thecorrectionfactorforinternalex-tinctionwasestimatedtobey960=14+5

−3.Giventheassumptionthattheholesinthenebulaarecompletelydevoidofbothgasanddust,andhencethatthedustcorrectionappliesonlytothepartsofthenebulawherethereisnoLyCleakage,theescapefractionmaythenbeestimatedfrom:fesc=

k2f900,obs

f.(11)

900,⋆

Althoughtheexactvalueofk2isnotknownforHaro11,alowerlimitonfesccanbeimposedfromtheminimumk2pre-dictedbytheplausiblerangeofstarformationscenarios.FromthemodelgriddefinedinTable3,weinfermin(k2)=1.78forthestandardIMFandmin(k2)=1.42otherwise.Thistrans-latesintofesc≥0.025+0.012+0.009

iscompletely−consistent0.010andfesc≥0.019withtheLyC−0.003,respectively,whichescapefractionderivedfromtheHαflux.

Tofurthertesttherobustnessoftheseresults,twootherspectralevolutionarymodels–Starburst99v.4(Leithereret

al.1999)andPEGASE.2´(Fioc&Rocca-Volmerange1999)–

havebeenusedtoassessthemodel-sensitivityofthek1andk2parameters.Wefind,thatwhenidenticalinputparametersareusedinallcodes,verysimilarresultsareproduced,indicatingthatourlimitsonfescdonotcriticallydependontheuseofanyspecificmodel.

4.Discussion

Aswementionedintheintroductionithasbeenarguedthatstar-forminggalaxiesmayyieldtherequiredLyCemissivitiestoreionizetheuniverseatz>5-6.Thisapproachhastheadvan-tageoverassumingAGNsassourcesinthatitexplainstheearlypollutionofheavyelementsobservedintheIGMatthesered-shifts(e.g.Cowieetal.1995).Moreover,analysisoftheoptical

NilsBergvalletal.:FirstdetectionofLymancontinuumphotonescapefromalocalstarburstgalaxy11

Table3.ThegridofZ01evolutionarysequences.Thegridcon-sistsofallpossiblecombinationsoftheparametervalueslistedbelow.Eachevolutionarysequencerunsfromagesof0to100Myrinshorttimesteps.

5.Conclusions

WereportonthediscoveryofaweaksignalintheLymancon-tinuumoftheluminousstarburstgalaxyHaro11.ItisthefirsttimeLymancontinuumleakageisfoundinagalaxyinthelocalenvironment.Usingthederivedcolumndensitiesoflowion-izationspeciesweestimatethecolumndensityofneutralhy-drogentobe∼1019cm−2,whichistoohightoallowLymancontinuumphotonescapeifthegasisdiffuselydistributed.WeconcludethattheLymancontinuumradiationisescapingthroughtransparentwindowsoftheISM.AssumingaSalpeterIMF,weestimateatotalescapefractionof0.04Acknowledgements.ThisworkhasbeendoneusingtheprofilefittingprocedureOwens.fdevelopedbyM.LemoineandtheFUSEFrenchTeam.WeareindebtedtoCarmelleRobertforhelpwiththeaccesstotheLavalSBsyntheticspectrallibrary.Wearepleasedtothankourreferee,DanielKunth,formanyvaluablecommentsandsuggestionstoimprovementsofthemanuscript.KjellOlofssonisthankedforhisobservationalcontributiontotheIUEdata.ThisworkwassupportedbytheSwedishScienceCOuncilandtheSwedishSpaceBoard.

IMF:dN/dM∝M−α

Mup:TheuppermasslimitoftheIMFSFH=Starformationhistory.

c=ConstantstarformationrateduringthesubsequentnumberofMyr.

e=Exponentiallydecliningstarformationrate,

SFR(t)∝exp(−t/τ),withane-foldingdecayrate(τ)equaltothesubsequentnumberofMyr.Z⋆:Themetallicityofthestellarpopulation

depthattheHIandHeIILyαcontinuumbreaksindifferenten-vironmentsatz<4indicatesthatthemajorionizationsourceindenseregionsisQSOswhileinvoidregionstheradiationap-pearstobesofter,givingroomforasubstantialcontributionfromstarburstgalaxies.

Severalsemi-analyticalmethodshavebeenusedtomodelthereionizationepoch.Duetothestronginhomogeneitiesoftheradiationfieldintheearlystagesofthereionization,thecomputationsarehoweverquitedemandingandcomplex(seeRazoumovetal.2002forasummaryofthetechnicalprob-lems)andrealisticsimulationshavenotbeenpossibleuntilrecently.In2003,Sokasianetal.(2003)usedafastcodeca-pableofexploringtheimpactoflow-massstarburstgalaxiesinhighspatialresolution.TheyfocusedontheevolutionoftheinteractionbetweentheIGMandstarforminggalaxiesofnormalproperties,e.g.similartoHaro11,intheredshiftin-tervalz∼20toz∼6.Oneoftheresultswasthattheprevi-ousestimatesoftherequiredphotonfluxrateneededtocom-pletethereionizationwasoverestimatedby63%ifassumedthatonlygalaxieswithmasses>109h−1M⊙wouldaccountforthereionization.Byincludingsourceswithmassesdownto∼4×107h−1M⊙theimpactofthelargepopulationofdwarfstarburstgalaxiescouldberealisticallyassessed.Thecalcula-tionsdemonstratedthatthesoftionizationsourcesoftypestar-burstdwarfsgivetherequiredphotonfluxprovidedthatthees-capefractionisfesc≥0.20.Thisnumberistwicetheupperlimitoffesc,gasthatwederivedforHaro11.Theestimateishow-everbasedonanormalstellarpopulation.Atredshiftsz>10popIIIstarscoulddominatethepopulation(e.g.Choudhury&Ferrara2005).InthesestarstheemissivityintheUVmightbelargerthaninpopIIbyafactorof2(Tumlinson&Shull2000;Brommetal.2001).Thus,althoughfescwouldnotchange,thenumberofproducedphotonsperunitmasswouldincreasesig-nificantly.Thusitseemsfeasiblethat,atleastatz󰀁10,dwarfstarburstsofthemassofHaro11andlowercouldprovideasubstantialcontributiontotheearlyphaseofthereionizationprocess.

References

Asplund,M.,Grevesse,N.,Sauval,A.J.,2004a,in”Cosmicabun-dancesasrecordsofstellarevolutionandnucleosynthesis”,ed.:F.N.Bash&T.GBarnes.ASPconf.series,inpress.

Asplund,M.,Grevesse,N.,Sauval,A.J.,AllendePrieto,C.,

Kiselman,D.,2004b,A&A417,751Becker,R.H.,etal.,2001,AJ122,2850

Begum,Ayesha,Chengalur,JayaramN.,2005,MNRAS362,609Bergvall,N.,1985,A&A146,269

Bergvall,N.,Olofsson,K.,1986,A&AS,,469

¨Bergvall,N.,Masegosa,J.,OstlinG.,Cernicharo,J.,2000,A&A359,

41

¨Bergvall,N.,Ostlin,G.,2002,A&A390,1

Bresolin,F.,Kennicutt,R.C.,Garnett,D.R.,1999,ApJ510,104Bromm,Volker,Kudritzki,RolfP.,Loeb,Abraham,2001,ApJ552,

4

Burton,W.B.,Hartmann,Dap,1994,Ap&SS217,1

Calzetti,D.,Kinney,A.L.,Storchi-Bergmann,T.,1994,ApJ429,582Calzetti,D.,Bohlin,R.C.,Kinney,AnneL.,Storchi-Bergmann,T.,

Heckman,TimothyM.,1995,ApJ443,136

Cartledge,S.I.B.,Clayton,G.C.,Gordon,K.D.,Rachford,B.L.,

Draine,B.T,Martin,P.G,Mathis,J.S.,Misselt,K.A.,Sofia,U.J.,Whittet,D.C.B.,Wolff,M.J.,2005,AJ630,355Choudhury,T.Roy,Ferrara,A.,2005,MNRAS361,577

Cowie,L.,Songaila,A.,Kim,T.,Hu,E.M.,1995,AJ109,1522Deharveng,J.-M.,Buat,V.,LeBrun,V.,etal.,2001,A&A375,805Devriendt,JulienE.G.,Sethi,ShivK.,Guiderdoni,Bruno,Nath,

BimanB.,1998,MNRAS298,708Fan,X.,etal.,2001,AJ122,2833

12NilsBergvalletal.:FirstdetectionofLymancontinuumphotonescapefromalocalstarburstgalaxy

Fan,X.,etal.,2002,AJ123,1247White,RichardL.,Becker,Robert

H.,Pentericci,Laura,Rix,Hans-Walter,2002,AJ123,1247

Feldman,PaulD.,Sahnow,DavidJ.,Kruk,JeffreyW.,Murphy,

EdwardM.,Moos,H.Warren,2001,JGR106,8119

Ferguson,AnnetteM.N.,Wyse,RosemaryF.G.,Gallagher,J.S.,III,

Hunter,DeidreA.,1996,AJ111,2265

Fernandez-Soto,A.,Lanzetta,K.M.,Chen,H.-W.,2003,MNRAS

342,1215

Fioc,M.,Rocca-Volmerange,B.,1999,astro-ph/9912179

Giallongo,E.,Fontana,A.,Madau,P.,1997,MNRAS2,629

GonzalezDelgado,RosaM.,Leitherer,Claus,Heckman,Timothy,

1997,ApJ4,601

Gordon,D.,Gottesman,Guzm´an,R.,Ostlin,¨S.T.,1981,AJ86,161

G.,Kunth,D.,Bershady,M.A.,Koo,D.C.,

Pahre,M.A.,2003,ApJ586,L45

Heckman,T.M.,Sembach,K.R.,Meurer,G.R,Leitherer,C.,

Calzetti,D.,Martin,C.L.,2001,ApJ,558,56

Hirashita,H.,Buat,V.,Inoue,A.K.,2003,A&A410,83Hurwitz,M.,Jelinsky,P.,Dixon,W.V.D.,1997,ApJ481,L31Iglesias-P´aramo,J.,Mu˜noz-Tu˜no´n,C.,2002,MNRAS336,33

Inoue,A.K.,Iwata,I.,Deharveng,J.-M.,Buat,V.,Burgarella,D.,

2005,A&A435,471

Jenkins,E.B.,Oegerle,W.R.,Gry,C.,Vallerga,J.,Sembach,K.R.,

Shelton,R.L.,Ferlet,R.,Vidal-Madjar,A.,York,D.G.,Linsky,J.L.,Roth,K.C.,Dupree,A.K.,Edelstein,J.,2000,ApJ538,L81

Kitayama,T.,Yoshida,N.,Susa,H.,Umemura,M.,2004,ApJ613,

631

Kogut,A.,Spergel,D.N.,Barnes,C.,Bennett,C.L.,Halpern,M.,

Hinshaw,G.,Jarosik,N.,Limon,M.,Meyer,S.S.,Page,L.,Tucker,G.S.,Wollack,E.,Wright,E.L.,2003,ApJS148,161Kunth,Daniel,Mas-Hesse,J.M.,Terlevich,E.,Terlevich,R.,

Lequeux,J.,Fall,S.Michael,1998,A&A334,11Lehner,N.,2002,ApJ578,126

Leitherer,C.,Ferguson,H.C.,Heckman,T.M.,Lowenthal,J.D.,1995,

ApJ4,L19

Leitherer,C.,Vacca,W.D.,Conti,P.S.,Filippenko,A.V.,Robert,C.

andSargent,W.L.W.,1996,ApJ465,717

Leitherer,C.,Schaerer,D.,Goldader,J.D.etal.,1999,ApJS123,3Lilly,Simon,Schade,David,Ellis,Richard,LeFevre,Olivier,

Brinchmann,Jarle,Tresse,Laurence,Abraham,Roberto,Hammer,Francois,Crampton,David,Colless,Matthew,Glazebrook,Karl,Mallen-Ornelas,Gabriela,Broadhurst,Thomas,1998,ApJ500,75

Lonsdale,C.J.,Helou,G.,1985,”Cataloguedgalaxiesandquasars

observedintheIRASsurvey”,Pasadena:JPL

Madau,Piero,Haardt,Francesco,Rees,MartinJ.,1999,ApJ514,8Madau,P.,Rees,M.J.,Volonteri,M.,Haardt,F.,Oh,S.P.,2004,ApJ

604,484Mall´en-Ornelas,Gabriela,Lilly,SimonJ.,Crampton,David,Schade,

David,1999,ApJL518,L83

Mas-Hesse,J.Miguel,Kunth,Daniel,1999,A&A349,765Mathis,J.S.,1990,ARA&A28,37

Meiksin,Avery,White,Martin,2003,MNRAS342,1205Mesinger,Andrei,Haiman,Zoltn,2004,ApJ611,69

Meurer,G.R.,Heckman,T.M.,Leitherer,C.,Kinney,A.,Robert,C.,

Garnett,D.R.,1995,AJ110,2665MatthewD.,Leitherer,Claus,LowenthalJ.,1997,AJ114,

Miller,J.S.,Mathews,W.G.,1972,ApJ172,593

Moy,E.,Rocca-Volmerange,B.,Fioc,M.,2001,A&A,365,347Oey,M.S.,Kennicutt,R.C.,Jr.,1997,Ostlin,¨MNRAS291,827G.,Amram,P.,Masegosa,J,Bergvall,N.,Boulesteix,J.,1999,A&AS137,419

Ostlin,¨G.,Amram,P.,Bergvall,N.Masegosa,J.,Boulesteix,J.,2001,A&A374,800Ostlin¨etal.,2005,inpreparation

Osterbrock,D.E.19,Astrophysicsofgaseousnebulaeandactive

galacticnuclei,UniversityScienceBooks

Razoumov,AlexeiO.,Norman,MichaelL.,Abel,Tom,Scott,

Douglas,2002,ApJ572,695

Robert,Carmelle,Pellerin,Anne,Aloisi,Alessandra,Leitherer,Claus,

Hoopes,Charles,Heckman,TimothyM.,2003,ApJS144,21Savage,BlairD.,Sembach,KennethR.,1991,ApJ379,245Schlegel,D.J.,Finkbeiner,D.P.,Davis,M.,1998,ApJ500,525Seaton,M.J.,1979,MNRAS187,73

Sembach,KennethR.,Howk,J.Christopher,Savage,BlairD.,Shull,

J.Michael,2001,AJ121,992

Shull,J.Michael,Roberts,David,Giroux,MarkL.,Penton,StevenV.,

Fardal,MarkA.,1999,AJ118,1450

Sokasian,A.,Abel,T.,Hernquist,L.,Springel,V.,2003,MNRAS344,

607Stasi´nska,G.,1990,A&AS,83,501Stasi´nska,G.,Schaerer,D.,Leitherer,C.,2001,A&A,370,1

Steidel,CharlesC.,Pettini,Max,Adelberger,KurtL.,2001,ApJ6,

665

Tumlinson,Jason,Shull,J.Michael,2000,ApJ528,L65Yan,Haojing,Windhorst,RogierA.,2004,ApJ600,L1Wood,K.,Mathis,J.S.,2004,MNRAS353,1126

Zackrisson,E.,Bergvall,N.,Olofsson,K.,Siebert,A.,2001,A&A

375,814(Z01)

Zurita,A.,Beckman,J.E.,Rozas,M.,Ryder,S.,2002,A&A386,801