HotHorizontalBranchStarsintheGalacticBulge.I.1,2
RuthC.Peterson3,DonaldM.Terndrup4,ElaineM.Sadler5,andAlistairR.Walker6
ABSTRACT
Wepresentthefirstresultsofasurveyofbluehorizontalbranch(BHB)starsintheGalacticbulge.Inthisexploratorystudy,candidateswith15≤V≤17.5coveringawiderangeinB−VcolorwereselectedfromCTIOSchmidtUBVphotometry.Bluespectrawererecordedat2.4˚AFWHMresolutionfor164starsina1.3sq.dg.field∼7.5◦fromtheGalacticcenter.Radialvelocitiesweremeasuredforallstars.ForstarswithstrongBalmerlines,wedevisedandappliedaspectroscopictechniquetodeterminestellartemperatureTeff,gravitylogg,andmetallicity[Fe/H]independentofreddening.ThereddeninganddistancetoeachstarwerethenfoundfromUBVphotometry.Reddeningprovedhighlyvariable,withE(B−V)rangingfrom0.0to0.55aroundameanof0.28.TheB−VcolorsofcoolHBstarsofsolarmetallicityreddenenedbyE(B−V)≥0.3overlapthoseofforegroundmain-sequencestars,buttheU−Bvs.B−VdiagramdistinguishesthesegroupsuntilE(B−V)>0.5.
Forty-sevenBHBcandidateswereidentifiedwithTeff≥7250K.SevenhavethegravitiesofPopulationistars,threeareambiguous,and37areHBstars,includingperhapsadozenRRLyraes.TheunambiguousBHBstarsareallcoolerthan9000K.Theyspanawidemetallicityrange,fromsolarto1/300solar.ThewarmerBHB’saremoremetal-poorandlooselyconcentratedtowardstheGalacticcenter,whilethecooleronesareofsomewhathighermetallicityandaresituatedclosertothecenter.Wedetecttwocoolsolar-metallicityHBstarsinthebulgeofourownGalaxy,thefirstsuchstarsknown.Stillelusivearetheirfainterhotcounterparts,themetal-richsdB/Ostarsstronginultravioletlight.
–2–
Subjectheadings:stars:abundances,stars:horizontal-branch,stars:variable:other,Galaxy:center,dust,extinction
1.Introduction
Accordingtoclassicalcalculationsofsingle-starevolution(e.g.,Rood1973),thecolorofastarthathasleftthegiantbranchtobecomeacore-helium-burningstaronthehorizontalbranch(HB)dependsprimarilyonitsageandmetallicity.Atlowermetallicities,awiderangeincolorisbothfoundandpredictedamongBHBstarsifmodestmasslossisassumed.Amongpopulationsofsolarmetallicityorhigher,onlyHBstarsredderthantheRRLyraeinstabilitystripshouldbeproducedwithinaHubbletime.However,themostmetal-richglobularclustersintheGalaxy,suchasNGC6388andNGC6441neartheGalacticcenter,doshowahandfulofBHBstars(Richetal.1997).ThesearemostlycoolorwarmBHB’s(Moehler,Sweigart,&Catelan1999),withfewifanyofthehottest,faintesttypes,thesubdwarfsdBstarsandtheirrelativelyraresuccessorsthesdOstars.Surprisingly,significantnumbersofsdB/O’sarefoundinmetal-richpopulationsintheMilkyWay;evenmoresurprisingly,theyappeartooutnumbercoolandwarmmetal-richBHB’s.Inthefield,whereitisdifficulttodistinguishBHB’sofintermediatetemperaturefromPopulationiAstars,fewifanymetal-richcoolBHB’sareknown–Gray,Corbally,&Philip(1996),forexample,uncoveredonlyonepossibleBHBwith[Fe/H]≥−0.7,outof39BHB’snewlyidentified.Incontrast,therearedozensoffieldsdB/OstarswhosegravitiesindicateHBstatus,andwhoseradialvelocitiesindicateathickdiskratherthanhalopopulation(Safferetal.1994,1997).Amongopenclusters(reviewedbyFriel1995),themoststrikingoccurrenceofsdB/OstarsisfoundinNGC6791,withametallicity3–4timessolar(Peterson&Green1998).InadditiontotwocoolerBHB(orbluestraggler)members,theclusterharborsfourorfivesdB/Ostars(Liebert,Saffer,&Green1994)whosemembershipislikelygiventheirspatialconcentrationtowardstheclustercenter(Kaluzny&Udalski1992).
Hot,metal-richBHBstarsalsoappeartobepresentinmetal-richextragalacticsystems.AsreviewedbyO’Connell(1999),ellipticalgalaxiesandearly-typespiralbulgescommonlyshowanupturninintegratedlightbelow2000˚A,whereoldmain-sequenceturnoffstarshaveverylittleflux.BoththesmoothspatialdistributionandthecontinuousspectraldistributionoftheUVupturnindicatethatitisnotcausedbyyoungOandBstars,butratherbyhotoldstars,sdB/O’sandthebrighterbutrarerbluepost-AGBstars(Dorman,O’Connell,&Rood1995;Brownetal.1997).Amongellipticalgalaxies,thesizeoftheUVupturntendstoincreasewithincreasinggalacticmetallicity(Faber1983;Bursteinetal.1988;Longoetal.1989),althoughthestrongestcorrelationiswithlineindicesbasedonlightelementsratherthanironitself,thescatterislarge,andthereisnocontinuitybetweenthesesystemsandglobularclusters.
–3–
ThepresenceofBHBstarsinhigh-metallicitypopulationsandthereversalintheircolordistributionraisemajorquestionsastotheirorigin.BothobservationandtheorysuggestthatBHBstarsinmetal-richsystemsmaybeproducedbychannelsinadditiontothoseoperatinginmetal-poorsystems.Theoreticalproductionofmetal-richsdB’sfromsingle-starevolutioncanbeachievedwitharapidincreaseat[Fe/H]>0ineitherheliumabundance(Bressan,Chiosi,&Fagotto1994;Yi,Demarque,&Oemler1998)ormassloss(Dorman,O’Connell,&Rood1995;D’Cruzetal.1996).Itmayalsobeaccomplishedwithdeepmixing,andthusaccompaniedbylight-elementenhancements(Sweigart1997;Kraftetal.1998).However,binarymasstransfermayplayadominantroleatthehotendoftheBHB,sincealargefractionoffieldsdB/Ostarsarefoundtobebinaries(Allardetal.1994;Green,Bowers,&Saffer2000).Bluestragglers,thoughtalsotohaveformedbymasstransfer,maythereforebeinvolvedintheproductionofhotHBstars.Constraintsonthesemechanismsmightbeplacedbydeterminingthecolordistributionandlight-elementratiosofmetal-richversusmetal-poorBHB’s,andbydeterminingwhetherbluestrag-glersofsimilarmetallicityareinvariablypresent.ThebestplacetodothisisintheMilkyWaybulgeitself.ThebulgeistheonlyMilkyWaypopulationsizableenoughtosupportlargenum-bersofBHBstars.Unlikethecentersofexternalgalaxies,itissufficientlynearbythatstarsasfaintasthesdB’smayberesolvedindividually.Itresemblesellipticalgalaxiesandspiralbulgesinstellardensityandstar-formationhistory(Whitford1978;Frogel&Whitford1987;Terndrup,Fro-gel,&Whitford1990;Houdashelt1995)aswellasinarelativelyhighabundanceoflightelements(McWilliam&Rich1994;Sadler,Rich,&Terndrup1996).Itsmetallicitygradient(Terndrup1988;Frogeletal.1990;Tyson&Rich1993;Tiede,Frogel,&Terndrup1995)providesanaturaltestbedofhowmetallicitydrivesstellarevolution.
ConsequentlywehaveundertakenasurveyoftheBHBpopulationinfourwindowsofthebulgealongandnearitsminoraxis.BHBcandidatesareselectedfromUBVphotometry.Forarepresentativesubset,follow-upmoderate-resolutionspectroscopyisprovidingstellarparameters(confirmingBHBstatusandestablishingreddening),plusironandmagnesiumabundancestoashighatemperatureaspossible.ThebasicgoalisastatisticallycompletesurveyofstarsonorneartheBHBacrossalltemperatureswheretheyoccur,from7250Kto35,000K.WeaimforthesamedegreeofcompletenessforallBHBstars,regardlessoftheirtemperatureandmetallicity.SuchasurveywouldshowimmediatelywhethercoolBHB’sandsdB/Ostarsexistatallinthebulge,assuggestedatthecoolendfromitsRRLyraestars(Walker&Terndrup1991).KnowledgeofthenumbersofhotBHBstarsineachfield,inconjunctionwiththemetallicityandtemperaturedistributionofcoolerBHBstarsinthesameregion,shouldhelpgreatlyindisentanglingtherelativeinfluenceofthevariousproductionfactorsnotedabove.
WebeginwiththispilotstudyofasingleregionlimitedinmagnitudetothecoolendoftheBHB,whoseresultswebrieflysummarizehere.SpectrawereobtainedduringthecommissioningphaseoftheTwo-DegreeField(2DF)spectrographoftheAnglo-AustralianTelescope.Forty-sevenstarswhoseBalmer-lineprofilesindicatedtemperatures≥7250Kwereanalyzedbycomparingtheirspectrawithagridoftheoreticalspectra.Reddeningwasthenfoundfromthemodelcolors
–4–
astabulatedbyKurucz.Weshowthatreddeningvariesdramaticallyfromstartostarwithinthisfield,with0≤E(B−V)≤0.55,sothatcolorsalonecannotdetermineTeffandlogg.However,theUBVcolor-colordiagramhelpstodistinguishBHBstarsfrommain-sequenceturnoffinterlopersaslongasE(B−V)0.50.
Thirty-sevenofthehotstarsprovedtohavetemperaturesandgravitiesindicatingapositiononthehorizontalbranch.TwomoreofthehotteststarsmightbeeitherBHBormain-sequencestars.NoneoftheunambiguousBHBstarsinthissampleprovestohaveTeff≥9000K,whichweattributetothesparsenessofoursampleatbluecolorsandfaintmagnitudes.WeestimateroughlyadozenofourcoolesthotstarstobeRRLyraes,pulsatingvariableslocatedjustredwardoftheBHBstarsonthehorizontalbranch.ThehotterBHBstarsinthissampletendtobemoremetal-poorandmorespatiallyextendedthanthecoolerones.WecommentonthepossibleimplicationsforBHBandRRLyraeproductioninmetal-richandmetal-poorbulgepopulations.FiveHBstarsarediscernedwith[Fe/H]≥−0.5inthebulgeitself,thefirstsuchstarsfound.SearchestofaintermagnitudesareplannedtoreachthehotsdBstarsbelievedresponsiblefortheUVupturn.
2.PhotometricObservationsandDataReduction
Duringaseven-nightrunin1995May-June,weusedtheCurtisSchmidtTelescopeatCerroTololoInteramericanObservatory(CTIO)toobtainUBVimagesofatotalof11fieldsofthebulge:fiveoverlappingfieldsbetween−6◦and−8◦alongtheminoraxis,andthreedisplaced3◦oneachsideof−6◦.Thefielddiscussedherewasobservedon1995May31UT,locatedat(ℓ,b)=(−3.325,−6.731),withcentralcoordinatesα=18h05m17s,δ=−35◦10′51.00′′(2000.0).Wewillcallthisour“target”field,ascomparedtoacalibrationfielddiscussedbelow.
TheCCDdetectorprovidedascaleof2.024′′pixel−1inV,forafieldofviewof1.15◦onaside,or1.3sq.deg.Theexposuresweretakeninthesequence8×90secinV,8×300secinB,and10×300secinU.Thelimitingmagnitudeswereabout18.7,19.2,and18.2forV,B,andUrespectively;about60,000starspersquaredegreeweredetectedwithbothBandVphotometry.Thebasicprocessingoftheframesinthetargetfieldconsistedofoverscanandzerocorrections,anddivisionbymediantwilightskyflats.ThisstepwasperformedusingtheccdredpackageoftheIRAFreductionpackage,availableathttp://iraf.noao.edu/.
StellarpositionsandmagnitudesweremeasuredoneachindividualCCDframeusingDAOPHOTII(Stetson1987).Initialpositionsweredeterminedbyfindingstarsonacombinedframeineachfilter,wheretheindividualimageswereshiftedtoacommoncoordinatesystemandthencombinedusingasigmaclippingalgorithmtoreducecosmicrays.Thephotometry,however,wasperformedontheuninterpolatedimages,asthestellarprofileswereundersampled.ThePSFs,whichshowedsignificantvariationwithposition,weredeterminedforeachframeusingbetween50and200bright,uncrowdedstars;thepositionsofthePSFstarswerecheckedtoinsurethattheywereuniformlydistributedovereachframe.ThePSFwasmodeledtohaveFWHMsinxandywhichvaried
–5–
quadraticallywithpositionintheframe.TherelativemagnitudesweredeterminedwiththeALL-STARroutineinDAOPHOTIIinasinglepass(i.e.,noadditionalstarsweresubsequentlyfoundonPSF-subtractedimagesandaddedtothelistoftrialpositions).
Theresultinglistsofmagnitudesandpositionswerethentransformedtoacommoncoordinatesystem,definedbythepositionsononeoftheVframes.Thecorrectionincludedtermsforpixelscale(identicalinxandy)andforrotation;thelattercorrection,however,wasinsignificant.ThenextstepwastocombinetheseveralmagnitudesineachfilterintothreelistsofaverageinstrumentalmagnitudesinUBV.Themethodwastopickoneoftheexposuresineachfilterasastandard,thentodetermineasingleoffsettobringthemagnitudescalesintoagreement.Thesemagnitudeoffsetsshowedsmallbutsignificantvariations,indicatingthatthetransparencyvariedduringtheUBVseries.Whenthemagnitudeoffsetsweredeterminedforalltheframesinonefilter,theseoffsetswereappliedandtheresultingphotometrywasaveraged.Specifically,theweightedaverageandweightedstandarddeviationwerecomputedforeachstar,wheretheweightsweredeterminedfromtheerrorsinasinglemeasurementreportedbyDAOPHOT.Thestandarddeviationofthe(x,y)positionsgaveameasureofthepositionalaccuracy,whichwasabout1/30pixelforthebrightstarsanddecreasingtoabout1/3pixelatthedetectionlimit.Starswhichweredetectedinatleastthreeframesoneachfilterwerekeptandtheothersdiscarded.
PhotometriccalibrationwasbasedonaseriesofUBVimagesobtainedforanoverlappingfieldonthephotometricnightof1995June1UT.Thesecalibrationframeswereoffsetfromthetargetfieldby7′inrightascensionand2′indeclination,inordertocalibrateseveralSchmidtfieldsatonce.Theexposuresequencewas2×90secinU,2×60secinB,and2×30secinV.Basicimageprocessing,theextractionofinstrumentalmagnitudesandaveragingofthephotometryproceededasforthetargetfield.Onthesamenight,wealsoobservedmanyfieldswithE-regionstandardsin(Graham1982),obtainingaperturephotometryofthestandardsandcomputingtransformationequationsinV,B−V,andU−Bincludingairmasscorrections.Wethenappliedthesetransformationsandcomputedthezeropointstobringthedeeperphotometryfromthetargetfieldontothesamescale.
AnastrometrictransformationwasobtainedfromtheVframesbyusingthesearchengineathttp://www.cfht.hawaii.edu/∼bernt/gsc/gscbrowser.htmltoidentify271guidestarsfromHubbleSpaceTelescope.Anastrometricsolutionwasfoundwhichincludedasingleframescaleinα,δ,arotationangle,andquadratictermsinrowandcolumnposition.Theresultingsolutionhadanr.m.s.scatterof0.4′′inrightascensionanddeclination(i.e.,about1/5pixel),adequateforourmultifiberspectroscopy.
Figure1showsthecolormagnitudediagram(CMD)inV,B−Vforourtargetfield.Starsobservedspectroscopicallyareindicatedaslargesymbols,withfilledcirclesrepresentingthosestarsfoundtobe7250Korhotterintheanalysisdescribedbelow.Smalldotsmarkthepositionsof17%oftheremainingstars.ThesamedesignationsaremaintainedinFigure2,acolor-colordiagramplottingU−Bvs.B−Vinthevicinityofthemain-sequenceturnoffofforegroundstars.As
–6–
discussedin§6,thisdiagramhelpsdiscriminatelightly-reddenedforegroundstarsfromhotterbutmoreheavilyreddenedstars,includingtheHBstarsofinterest.
3.SpectroscopicObservationsandDataReduction
Inthisexploratorywork,ourselectionofcandidateswasbasedontheCMDonly,toensureacompletelyunbiassedsample.Fromwell-populatedregionsoftheCMD,wedrewstarsbroadlydistributedincolorastargetsforfiberspectroscopy.Themagnitudeandcolorlimits,14.8≤V≤17.8andB−V≤1.2,includeHBstarswithinandbeyondthebulgeoverawiderangeofmetallicityandreddening.Fibertanglingandminimumproximity(≥11.5′′inthiscase)ruledoutsomecandidatesoncehigher-prioritytargetsareselected;wegavehigherprioritytobluerstars.On1997July3UT,theTwo-DegreeField(2dF)instrumentattheAnglo-AustralianTelescope(Lewis,Glazebrook,&Taylor1998;Smith&Lankshear1998)wasusedtoobtain6×1800secexposureswithasinglefiberconfigurationonthetargetfield.188fibers2.16′′indiameterwereplacedwithinits2◦fieldofview.Atotalof164fiberswereassignedtostellartargets,and24assignedtoskyregionschosentobefreeofresolvedstarsbyinspectionoftheBandVCCDframes.The1200BgratinginthestandardAATsetwasemployedinfirstorder,yieldingaresolutionof1.11˚Apixel−1andawavelengthcoverageof3834–4973˚A.Theeffectiveresolution,judgedfromthewidthoftheemissionlinesintheterrestrialsky,isabout2.0pixels,or150kms−1atthecentralwavelengthofthespectra.Spectralsynthesisindicates167kms−1or2.2pixelsforthecoaddedtargetobservations.
Thespectrawereextractedusingthe2DFDRreductionpipelinedescribedbyBailey&Glaze-brook(1999).Cosmicrayswereremovedfromtheindividualexposuresusingasigma-clippingalgorithm,andthecleanedimagesweresummedtoproduceanimagewithaneffectiveexposuretimeof10,800sec.ThewavelengthscalewasdeterminedfromemissionlinesinaCuArspectrum.Theextractedspectrawerescaledtocorrectforthenonuniformfiberefficiencies.Theskyspectrawerethenaveragedtogetherandsubtractedfromeachoftheobjectspectra.
Themeancountsperpixelintheextractedspectracorrelatedtightlywithmagnitude,fromabout8500countsatV=15.5to2000countsatV=17.5,verifyingtheaccuracyoftheastrometry.Thebrightestandfaintestspectrahavesignaltonoiseof∼125and30per2.2-pixelresolutionelementat4800˚A,decreasingtoabouthalfthisat3900˚Aforthehotstars.
4.RadialVelocities
Tomeasureradialvelocities,thespectrawererebinnedtoalog-wavelengthscaleandinspectedbyeye.Theyfellintotwodistinctclasses—F-starspectrawithnarrowabsorptionlinesofhydrogenandmetals(94stars),andBHB-likespectradominatedbystrong,broadBalmerabsorptionlines
–7–
(38stars).15spectrawereclassifiedasintermediatebetweenthesetwoclasses,andabout20spectrahadlowS/Nandwereomittedfromfurtheranalysis.
Thevelocitiesweremeasuredfromthescrosscross-correlationroutineoftheFigaroreductionpackage(Shortridge1995).Weconstructedtworadial-velocitytemplatesbyco-addingspectrawithgoodS/N.Onewasthesumof42F-starspectra,theotherthesumof11BHB-likespectra.Theradialvelocityforeachstarwasfoundbycross-correlationwiththeappropriatetemplate;forthe“intermediate”stars,weusedbothtemplatesandtooktheaveragevalue.VelocitiesarerelativetotheF-startemplate,andhavearandomerrorof5–10kms−1,dependingonexposurelevelandlinestrength.Thisestimateisbasedontheagreementwithvelocitiesdeterminedbyvisualcomparisonwithsyntheticspectra.Moststars’observedspectramatchedwellafter10kms−1wassubtractedtobringthetabulatedvelocitiesontoageocentricscale.Forstars168,75,110,118,190,and57,anadditional10kms−1neededtobesubtractedfromthetabulatedvelocitiestomatchthesynthesis;forstars126and192,10kms−1hadtobeaddedback;forstars52and66,anadditional30kms−1hadtobesubtracted.
ThevelocitydispersionsoftheF-starandBHB-stargroupsaresignificantlydifferent.(The“intermediate”starshadavelocitydistributionsimilartotheBHBstars,andsotheywereaddedtotheBHBgroup).The94starswithF-starspectrahaveameanvelocityof−3.9kms−1andadispersionof51.0kms−1,whilethe53BHBandintermediatestarshaveameanvelocityof−2.9kms−1andadispersionof120.9kms−1.Thereisnoobvioustrendofradialvelocitywithapparentmagnitude,andthemeanmagnitudeofthetwogroupsisverysimilar.Onthebasisoftheirkinematics,then,weidentifytheF-stargroupasmembersoftheforegoundGalacticdisk,forwhichweexpectavelocitydispersionof40–95kms−1dependingontheirmeandistance(e.g.,Terndrup,Sadler,&Rich1995),andtheBHBgroupasprimarilybulgestars.
5.SpectrumAnalysis
Spectrumanalysiswasbasedonvisuallymatchingeachobservedspectrumtoasequenceofabinitiocalculationsbasedonmodelatmospheresandalineparameterlist.Theprogramused,SYN-THE(Kurucz&Avrett1981),wasdownloadedfromtheKuruczWebsiteathttp://cfaku5.harvard.edu,andmodifiedbyS.AllenandB.DormantorunonanUltra-30attheUniversityofVirginia.WealsodownloadedtheKurucz(1991)gridofATLAS9modelsandcolors,andthelinelistsgf0400.100andgf0050.100dated25-May-98.AsdescribedbyKurucz(1995),theydifferfromthoseofKu-rucz&Bell(1995)inincorporatingformanyiron-groupspeciesallnewlyidentifiedandrevisedtransitionswhoseenergylevelsweremeasuredrecentlybyNaveetal.(1994)andotherlaboratorygroups.Moreover,theyincludeonlyatomiclinesidentifiedinthelaboratory.Thepredictedlinesareofnouseherebecausetheirwavelengthsaregoodtoonly∼10˚A(beinggeneratedbyKuruczfromhissemi-empiricalatomicmodels,inwhichlaboratory-basedenergylevelswereextrapolatedtohighexcitation).Theiromissionhasnoeffectonderivedparameters,sinceveryfewunidentifiedlinesappearintheseopticalstellarspectra,whicharegenerallyweaker-linedthantheSun.
–8–
Modelmetallicitiesrangedfrom1/300solartotwicesolar:[Fe/H]=−2.5,−2.0,−1.5,−1.0,−0.5,0.0,and+0.3.Thelight-elementabundanceratioswereheldfixedat[α/Fe]=0.0(solar)for[Fe/H]≥−0.5,[α/Fe]=+0.2for−1.0and−1.5,and[α/Fe]=+0.4for[Fe/H]≤−2.0,withO,Mg,Si,Ca,andTiabundancessoadjusted.Thiswasdonebecausemostofthelight-elementlinesareblendedwithiron-peaklinesatthisresolution;athigherresolutionthe[α/Fe]ratiocouldbedeterminedindependently.Inallcases,theironabundancewasreducedby0.15dextothecurrentlyacceptedsolarvalue,andthemodelionizationequilibriawererecalculated.
Eachstellarspectrumwascalculatedfrom3555to4965˚Aataresolutionof300,000,andbroadenedbyamacroturbulentvelocityvm=1.5kms−1,amicroturbulentvelocityχt=2.0kms−1,andaGaussianinstrumentalprofileofFWHM167kms−1(2.4˚Aat4400˚A)whichbestmatchedtheobservedspectra.Fortwostarsnotedin§8,theobservedspectralfeaturesrequiredadditionalbroadening.
Withfewexceptions,weanalyzedonlythosestarswithtemperatureTeff≥7250K,thatofthemetal-richstar2-17situatednearthecoolendoftheBHBinNGC6791,eitheraBHBorabluestraggler(Peterson&Green1998).Ourspectralcalculationsshowedthatattheresolutionofthesedata,thecentralfluxofHβis39%thatofthecontinuumforacoolBHBstarat7250Kwithlogg=2.5.TobesuretoidentifyallstarswithTeff≥7250Kinthepresenceofnoiseorabinarycompanion,weacceptedforfurtherconsiderationanyspectrumwitharelativecentralHβfluxof52%orless.OnestarwithbroadbutshallowBalmerlineswasalsoincluded.WehadtriedusingmorestandardcriteriabasedonthebreadthoftheHβwingsat80%(or70%)fluxlevelandthetotalHβequivalentwidth,butbothapproachesfailedatthisresolutionandwiththiswiderangeofmetallicity,becauseatomic-lineblendsinsolar-metallicitystarsmimicmoderatelystrongerBalmerwingsandgeneratelargerBalmerequivalentwidths.Asillustratedbelow,thisblendingisworseforHγandprohibitivefortheotherBalmerlines.
Becauseofthisblending,settingthestellartemperatureTeffwasdoneusingthecentral∼5–8˚AoftheprofilesofHβandHγ,excludingasnotedbelowthecentral1–2˚Aofthecore(withthelargervaluesreferringtohotterstars).Thisapproachistotallyinsensitivetoreddening,andwefoundittobeindependentof[Fe/H]aswellandlargelyinsensitivetologg.AninitialmodelwaschosentoapproximatelymatchbothBalmerandmetallines;itsspectrumwascalculatedandshiftedbythestellarvelocity,thencomparedtotheobservedspectrum.Anewmodelwasselectedtoaddressanymismatch(withdifferent[Fe/H]andvelocityifnecessary)andtheprocedurerepeateduntiltheagreementwassatisfactory.Afewcasesneverreachedthisgoal;asdiscussedbelow,mosthaveanomalousBalmerlineprofilesandarelikelytobeRRLyraestars.Analternativepossibilityofahighheliumabundance(withheliumandhydrogenequalbynumber)wasproposedbyGray,Corbally,&Philip(1996)fortwofieldBHBstarsthattheyfoundtohaveBalmerlinestoobroadfortheBalmerjump.
OnceTeffwasset,thebreadthoftheBalmerlinewingswasusedtodeterminethegravity,byfittingthewingsoftheHβandHγprofiles.Finaladjustmentwasmadeto[Fe/H]tobestreproduce
–9–
thestrengthsoftheatomiclines,againasjudgedbyeye.Weanalyzedfullyatotalof60stars,includingfour,seven,andtwowithTeff=7000K,6750K,and6500Krespectivelyastemplatestoweedouttheothercoolerspectra.
ReddeningwasderivedforeachstarwithUBVphotometry.ThevalueofE(B−V)tothenearest0.05magwhichbestreproducedtheU−BandB−VcolorswasfoundfromthetabulationbyKurucz(1991)ofUBVcolorsvs.E(B−V).ForeachstarwhoseparametersplaceitneareithertheBHBorthemainsequence,anMVvaluewasassignedaccordinglyandadistancethendeterminedfromtheobservedphotometryanddeducedE(B−V).
Table1listsbasicobservationaldataandderivedstellarparametersforeachofthe47starsfoundtohaveTeff≥7250K.Thefirstcolumngivesthe2dFfibernumber,bywhichwesubsequentlyrefertoindividualobjects.Thenextcolumnslistquantitiesdeterminedduringthephotometricandastrometricreductions:thestarID,position,Vmagnitude,B−Vcolor,andU−Bcolorandtheiruncertainties,the1σdeviationofthemeasurementfromanindividualframeaboutthemean.NextappearstheradialvelocityVradinkms−1ontheinternal2dFscale.ThefollowingcolumnsgivethededucedTeff,logg,and[Fe/H](designatedZ),andthefinalcolumnsgiveparametersdeducedbyfittingthecolorsofthemodelatthereddeninglistedtothoseobservedforthestar.DisthedistancefromtheSuninkpc;MV,theabsolutevisualmagnitude;E,thereddeningE(B−V);andδUandδB,thedifferencebetweentheobservedandthemodelU−BandB−Vcolors.InFigure3,weshowthefitsachievedforstarsspanningawiderangeinparametersandinS/N.Eachpanelcomparestheobservedandcalculatedspectraforthesamesixstarsinadifferentwavelengthregion.Ineachcomparison,theheavylineistheobservationandthelightlinethecalculation;successivespectraaredisplacedupwardsby20%.Themodelatmosphereparametersarelistedontherightwiththefibernumberofthestarobserved.
WenoteimmediatelythattheverycenterofeachBalmerlineisalwaystoodeepinthecalcu-latedspectra.WehaveverifiedthatthecoreandonlythecoreoftheBalmerlinesismismatchedinBHBstars,usingechellespectraoffieldBHBstars.Forexample,forHD130095,wematchtheHγlineprofileto2%orbetteruntil+/-1.2Aofthecenter,wheretheresidualfluxhasdroppedbelow30%ofthecontinuum(Petersonetal.,inpreparation).IntheSun,thecoreofHαisformedwellintothechromosphere,whereCaIIemissionarises(Dupree,Sasselov,&Lester1992).Wethusattributethiscoremismatchtouncertaintiesinthesuperficiallayersofcoolstarswithconvectionandchromospheres,andignorethelinecoreindeterminingstellarparameters.
InadditiontothestrongBalmerlines,numerousweaklinesofheavierelementsappearinFigure3.Theirstrengthandnumbersgenerallyincreasetowardsbluewavelengthsandincoolerstars.ThustheBalmer-lineprofiles,especiallyHδ,areseentobecontaminatedbyweakatomiclinesinthecool,metal-richstars,precludingtheuseofstandardmeasuresofBalmer-linebreadthtoderiveTeffandlogg.ThatourownproceduresucceedsisindicatedbythegoodnessoffitofthespectrainallpanelsofFig.3,inparticularHδinpanela.
Toexploreuniquenessofthedeterminations,alternativefitsweretried.InFigure4,asecond
–10–
calculationisshownfortheHγandHβregionsofthespectrumofeachstarofFigure3.Thesamemodelparameterswereassumedexceptthatsurfacegravitywastakentobelogg=3.5(4.0inthebottomspectrum).NotshownisathirdcalculationinwhichthesamegravitieswereadoptedasinFigure4,butTeffwaschangedby250Kinanattempttocompensateforthechangeinlinewingsintroducedbythechangeinlogg.TheresultingfitstoHγandHβwerethesameorworseasinFig.4.
ComparingFigures3band4a,andFigures3dand4b,indicatesthatgravitiesarewell-determinedforthehotterstars,butby8000K,thesensitivityofHγandHβtoachangeof0.5dexinlogghasbecomeminimal.Atlowertemperatures,theweakBalmerprofilesmakethegravitydeterminationdifficultathighmetallicityandlowS/N.ThetopthreespectraillustratetheincreasingdifficultyofestablishingloggatTeff≤7500K.
Nonetheless,thebestfitsareclearlybetterinbothBalmerlinesthanthealternativefits,inallbutthetopspectrum.Eventhere,thebestfitismarginallybetter.Moreover,bluewardof4315˚Ainthetopspectrumonly,thebandheadoftheCHmoleculeisevidentastheabsorptionunmatchedbythespectralsynthesis.Becausemolecularlinesbecomestrongonlyincool,metal-rich,and/orhighsurface-gravitystarssuchasthisone,theywerenotincludedinthespectralsynthesislinelist.TheappearanceofCHheresupportsthelowTeffandhighloggwededuceforthisstar.
Wealsorancomparisonstotesttheeffectofthetreatmentofconvection.Castelli,Gratton,&Kurucz(1997)haveshownthatturningoffconvectiveovershootinKuruczmodelschangesthetemperatureofdeeplayersincoolstars.ColorsandBalmer-lineprofilesareaffected,especiallyatlowmetallicitiesandinlateAstarswhereconvectivetransportisstrongest.Peterson,Dorman,&Rood(2000)findthatCastellietal.modelsbutnotKurucz(1995)modelssimultaneouslymatchtheHαprofileandthemid-UVfluxdistributioninmetal-poorturnoffstarswithTeff<7000K.Wethuscalculatedadditionalspectraforstars57and98usingCastellietal.models(downloadedfromtheKuruczwebsite).AtTeff=7500K,logg=3.0,andsolarmetallicity,nochangeintheHβlineprofilewasseen.AtTeff=7250Kand[Fe/H]=−1.0,thewingsofHβbecametoostrongtomatchstar57atanylogg.AtTeff=7000Kandlogg=3.0agoodfitwasfoundforHβ,butthewingsofHδweretooweak.Unfortunately,ourphotometryshowsthatthisstarisvariable(§7).Furthertestswithhigher-qualitydatafornonvariablestarsareneededtoelucidatetheeffectsofthetreatmentofconvectiononBalmer-lineprofilesincoolBHBstars.
Fromcomparisonssuchasthis,weestimateuncertaintiesasfollows.TheuncertaintyinTeffforthestarswithTeff≥7500Kis250K,andinloggis0.5dex,providedHδisalsomatchedinstarswithTeff≤8000K.Formetal-poorstarsof7250K,thesameerrorsapplyingravity,towhichtheBalmer-linewingsarestillsomewhatsensitive,butthetreatmentofconvectionmayincreaseerrorsinTeff.Forstarsof7250Kwithnear-solarmetallicity,evolvedstarscanbedistinguishedfrommain-sequencestarsviatheCHband.
AnadditionalcheckonthevalidityofthededucedatmosphericparameterswasmadefromtheresidualsinB−VandU−B.ForstarswithTeff≥8000K,the1σstandarddeviationoftheB−V
–11–
andU−Bcolorsismarginallygreaterthantheaverageuncertaintiesofmeasurement.However,severalcoolerstarsshowsignificantlylargerdiscrepanciesandlargererrorsinstellarparameters.ThesemaybeRRLyraevariables.Wediscussthesebelow,afterexaminingtheconsequencesofthelargerangeinreddening.
6.Reddening
Forthetotalof36HBstarswithUBVphotometry,E(B−V)=0.28,andthe1σdeviationaboutthemeanis0.14.However,reddeningdiffersdramaticallyamongthehotstars:valuesfromE(B−V)=0.0to0.55occurthroughoutthe1.3sq.deg.field.AsimilarmeanandrangeinE(B−V)arefoundamongonlythosestarswithTeff≥8000K,almostallofwhichareBHBstars,indicatingthatRRLyraevariabilityisnottoblame.
Withsucharangeinreddening,UBVphotometrybyitselfcannotbeusedtosimultaneouslydetermineTeff,logg,and[Fe/H]ofastar,sinceB−VandU−Barebothsensitivetoallfourparameters.ManyrecentstudiesofBHBstarsinthefieldhalo(e.g.,Kinman,Kraft,&Suntzeff1994;Wilhelm,Beers,&Gray1999)simplyassumeareddeningvalue,usuallyE(B−V)<0.10.Whilethismaybeappropriateforthehalo,itisclearlyunwarrantedhere,andpresumablyelsewhereatlowgalacticlatitude.Incontrast,fittinghighS/Nspectraof2.4˚AFWHMresolutiontotheBalmerlinesHβ,Hγ,andHδdoesprovideareliablewaytodisentangletheparametersforstarswith7250K≤Teff<10,000K.Themethodworksbecausetheblendsareexplicitlymodelled,theinnerfew˚AoftheBalmerlinesoutsidethecoregenerallyreflectsTeff,andthewingsthengivelogg;reddeningfollowsfromthecolors.
Whenmoderatelyreddened,themoremetal-richhotstarsneartheredendoftheBHBhaveB−Vcolorswhichoverlapthoseofmain-sequenceforegroundstars.ThisisevidentinFigure1,wherethehotstars(filledcircles)arepredominantlybluewardofthemain-sequencecrushbutincludemanystarswithinit.Eveninthepresenceofmoderatereddening,however,theU−Bvs.B−Vcolor-colordiagramaidsinseparatinghotstarsfromsolar-metallicityturnoffstars.ThisseparationofhotandcoolstarsisillustratedinFigure2.Empirically,themajorityofthehotstarshaveabluerB−VcoloratagivenU−Bthanthecoolerones(opencircles).(Manyofthemostdeviantfilledcirclesaretheprobablevariableslistedbelow,andtheopencirclesintheBHBregionmaybeRRLyraeswhichwerecoolerthan7250Katthetimeofthespectroscopicobservations.)Thetheoreticalmodelcolorsconcur,asshownbythecurves.Theyweregeneratedfromthetabulatedcolorsofmodelswith[Fe/H]=0byinterpolatinginloggalongthezero-ageBHB,fromTeff,logg=7250K,2.75to13,000K,4.1.ThethreecurvesrepresentreddeningsofE(B−V)=0.0,0.3(themeanforthefield),and0.5;thereddeningvectorisgivenbythearrow,whoseheadscorrespondtothelatterreddenings.Thefigureillustratesthatthecolor-colordiagramprovidesdiscriminationagainstfieldmain-sequencestarsforBHBstarsofalltemperaturesuntilE(B−V)exceeds0.5,whenBHB’scoolerthan∼15,000Karesuperimposedonthem.Athigherreddening,itmightbethoughtthatIbandcolorswouldhelp,butthemodelcolorsindicatethis
–12–
istrueonlyiftheImagnitudeismeasuredto±0.05mag.JHKphotometrywouldalsohelpinprinciple,butrequiresdifferentdetectorswhichcurrentlyhaveasmallerfieldofview,andleadstoworsecrowdingbyredderstars.
7.RRLyraeInterlopers
AsnotedinthereviewofRRLyraestarsbySmith(1995),thepulsationalcyclethatdrivesRRLyraevariabilityleadstotemperaturesofRRLyraestarsnearmaximumlightthatoverlapthoseofstableBHBstars.ThespectraltypeatmaximumofRRLyraesisA7–A8,correspondingto∼8000K.ThusasignificantfractionofRRLyraesisanticipatedamongourcoolerHBsample.OneindicationoftheirpresenceistheoccasionallargesizeofdeviationsinU−Band/orB−Vcolorseenforstarscoolerthan8000KinTable1.Ofthehotterstars,onlystar162showsadeviation>0.12,butstar163doessoat7750K,andsixcoolerstarsdoaswell.AnotherindicationcomesfromanomalousBalmerlines.Star162showsdistinctBalmer-coreemission,andanincreasingfractionofcoolerstarshaveBalmerlineswithasymmetricprofilesorfilled-incoresorverybroadwings.Themajorityofthestarswithlargephotometricdeviationsshowsuchanomalies.Wesuspectthatthesearesignsofpulsationalvariability.Metal-poorRRLyraestarsoftypeabareknowntoshowBalmer-lineemissionduringtherisetomaximumlight(Preston1964),duetoshocksformedbypulsation;themilderpulsationoftheovertonecvariablesmightatleastfillintheBalmercores.Thewingsarealsoprobablyaffectedbythedynamicalstateofthestar,leadingtopoorfitsduringsomephases.However,duringintermediatephaseswhentheycloselyresemblestableBHBstars,RRLyraevariableswillnotbedistinguishablespectroscopically,asborneoutbythegoodfitstostar57inFigure3.
Wethussoughtphotometricevidenceforvariability.Theindividualmeasurementsinthephotometricdatasetshouldshowvariationsinmagnitudeandincolor,whichshouldreddenasmagnitudedeclines.However,thetargetfieldobservationsspanonly18,40,and52mininU,B,andVrespectively,muchshorterthantheRRLyraeperiodsof5–27hr(Smith1995).Insteadwecomparedthemeanmagnitudesfoundforallthosehotstarsobservedinboththetargetfieldandthecalibrationfieldobtainedthenextnight,27–29hrlater.Threestarsshoweddefinitevariability:57,124,and132.Allarelow-gravitystarswithTeff=7250K,theonesmostlikelytobeRRLyraes.Roughlyonlyonequarterofthevariablesshouldbedetectedinapairofobservationsmadeanightapart,duetoimproperphasingandtherelativeconstancyofabvariablesnearminimumlight.ThestarsinoursamplemostlikelytobeRRLyraesfromtheabovecriteriaare162,163,189,150,166,57,124,159,70,66,and132.Overall,weestimatethatRRLyraesmightcompriseasmanyasaquarterofourlow-gravitystarswithTeff=8000–8250K,halfwith7500–7750K,andthree-quarterswithTeff=7250K.
–13–
8.
ResultsandDiscussion
Ofthe47starsfoundtohaveTeff≥7250K,gravitiesloggindicatethatninearepossiblePopulationicore-hydrogen-burningstars.(Atenth,star163withtheunusualgravitylogg=3.5atTeff=7750K,islikelytobeanRRLyrae.)Ofthenine,threestarshaveTeff=7250Kandlogg=4.0,andarelikelytobeforegroundyoungstarsgiventheirdistances.FourclearlyareyoungPopulationistars,withlogg=4.0andTeff<9000K.Twooftheseshowclassicalpeculiarities(Jaschek&Jaschek1990):91isaSr-Cr-EuApstar,withstrongSriilinesandsuitableB−Vcolors,and46hasveryweakCaiiforitsBalmer-lineandmetallic-linestrengths,asistypicalofAmstars.Stars75and91arebroad-lined.Ifduetorapidrotation,impliedvsinivaluesare∼110kms−1and∼190kms−1respectively,valuesinkeepingwithnormalmain-sequenceAstars(Jaschek&Jaschek1990).Twostars(48and182)withTeff≥9000Kandlogg=3.5couldbeeitherforegroundyoungstarsorPopulationiievolvedstars.Ifthelatter,neitherisinthebulge,butratherinfrontofit.
Thirty-sevenstarsareevolvedPopulationiicore-helium-burningstarsontheHB:thosewith8500K≥Teff≥7250Kand3.0≥logg≥2.5.Asdiscussedabove,amongthecoolerstarswemustallowforRRLyraes,whichareHBstarsbutnotBHBstars.WeestimatethattwostarswithTeff≥8000K,fivewith7500–7750K,andsevenoftheninewith7250Kshouldbeconsideredvariables.Thisleavesthenumberofbona-fideBHBstarsaround23.
Inoursample,nounambiguousBHBstarsappearwithTeff>8500K.TheBalmerlinestrengthsreachamaximumnear9500K,thatofthehotteststarwehavediscerned,sothepossibilityexiststhatwehaveerroneouslyassignedhotterstarstothecoolerside.Wedonotthinkthishashappened,becauseinhotstarsboththeBalmerlineprofilesandthestrengthsofweakatomiclinesarewellreproducedatthetemperaturesassigned.
ItisprematuretoconcludethathotterBHBstarsareabsentfromthebulge,however,becauseofthesparsecoverageofthespectroscopicdatasetatthecolorsandmagnitudesappropriateforstarshotterthan10,000KreddenedbyE(B−V)=0.3.WemayjudgethisfromtheNGC6791CMDandthemodelcolors.Chaboyer,Green,&Liebert(1999)matchisochronestotheNGC6791CMD,deriving0.08≤E(B−V)≤0.13and13.30≤(m−M)V≤13.45.Indeed,thecolors(Montgomery,Janes,&Phelps1994;Kaluzny&Rucinski1995)of2-17,thecoolBHBcandidateinNGC6791,arematchedtoafewhundredthsinU−B,B−V,andV−IbythoseofthemodelwithTeff=7250K,logg=3.0,and[Fe/H]=+0.3whenE(B−V)=0.15istaken.AdoptingMV=1.1forthisstar,forwhichV=15.0,alsoreproduces(m−M)V=13.45.Starsinthebulgewouldappear1magfainteratthesamereddening,or1.5magfainterinVifE(B−V)=0.3.ThusastarwithTeff=11,500Kandlogg=4.0wouldhaveV=16.0inNGC6791,butwouldhaveV=17.5inthebulgeatE(B−V)=0.3,andB−V=0.20andU−B=−0.06basedonthemodels.Wehaveobtainedveryfewspectraofstarsthisfaintandthisblue(seeFigures1and2),andsoshouldnotbesurprisedtofindnonethishot.
TheHBstarsuncoveredbyouranalysisspanawidemetallicityrange,−2.5≤[Fe/H]≤+0.0.
–14–
ThisisconsiderablybroaderthanthemetallicitydistributionforRRLyraestarsinBaade’sWindowat−4◦(Walker&Terndrup1991),whichissharplypeakednear[Fe/H]=−1.0anddropsquicklyfor[Fe/H]>−0.9.TheHBmetallicitydistributionisalsobroaderthanthatofKgiants,butintheoppositesense:theK-giantmeanis[Fe/H]=−0.11,andfewKgiantsarefoundbelow−1.0(Sadler,Rich,&Terndrup1996).
ThetrueBHBstarswith8500K≥Teff≥8000Kspanaratherwiderangeofdistances.ExcludingasRRLyraestarsbothstar162and(arbitrarily)star118(at8000K),wefindfor14BHBstarsameandistanceof8.9kpcwithanindividualdeviationσi=4.1kpc.Theirradialvelocitydispersionis100.5kms−1withameanof22.0kms−1.Theyaremetal-poor,averaging[Fe/H]=−1.59±0.13withσi=0.49dex.ThehotterBHBstarsinoursamplethusaredominatedbyametal-poor,extended-bulgepopulationreminiscentofthehalo.
ThemeanmetallicityofthecoolerHBstarsisdifficulttojudgebutisprobablyhigher.Themeanandstandarddeviationforall19low-gravitystarswithTeff≤7750Kis[Fe/H]=−1.29±0.18andσi=0.77.ForthetenstarsremainingafterexcludingthemostlikelyRRLyraeslistedabove,wefind[Fe/H]=−1.10±0.25andσi=0.78dex.BecauseofthelargevelocityvariabilityofRRLyraes,exceeding100kms−1forabtypes,itisprematuretoexaminethecoolHBstars’velocitydispersion.
TherearefiveHBstarswith[Fe/H]≥−0.5,includingtwowithsolarmetallicity.OnlyonehasTeff>7500K.Thefivestarsarelocatedanaverage8.2kpcaway,witha1σdispersionof2.9kpc.Thustheyareindeedbulgestars.Wecannotexcludethepossibilitythatboththesolar-metallicitystarsareRRLyraes,giventheir7500Ktemperatures.Thisisnotlikely,however,inviewofthepaucityofmetal-richcRRLyraevariables(Smith1995),aswellasthegoodfitstotheirspectraandcolors.Moreover,thevelocitiesofthesetwostarsarelow,−56and−45kms−1.Ourworkhasthusidentifiedsolar-metallicityHBstarsinthebulgeforthefirsttime.
CoupledwiththedichotomyabovebetweentheRRLyraeandK-giantmetallicitiesinthebulge,thissuggeststhatthewarmandcoolBHBstarsatthisangulardistancefromtheGalacticcentermaybedominatedbytwodifferentprogenitorpopulations.Oneismetal-poorandtheothermetal-rich,andtheirHB-staroutputdiffersconsiderably.Themetal-poorpopulation,lessconcentratedtotheGalacticcenter,seemstoproducemanywarmandcoolBHBstarsperKgiant.Themetal-richpopulation,moreconcentratedtowardsthegalacticcenter,seemstoproducefarfewerHBstarsperKgiant,mostlycoolones.
Moreover,themetal-poorpopulationmightformRRLyraesinpreferencetostableBHBstarsoflowtemperature,whilethemetal-richpopulationmightdotheopposite.Thisissuggestedfromthedearthofmetal-richRRLyraesinboththebulgeandthesolarneighborhood,andbythelowtemperatureof7250K(Peterson&Green1998)foundforthenonvariablemetal-richstar2-17inNGC6791,ifatrueBHB.Thusmetal-poorandmetal-richRRLyraes,likemetal-poorandmetal-richsdB’s,mightbeformedthroughasomewhatdifferentassortmentofpathwaysthancoolBHBstars.Suchapossibilityisalsosuggestedbymeasurementsofsurfacerotationvsini.Although
–15–
noneofthe27fieldRRLyraestarsmeasuredbyPeterson,Carney,&Latham(1996)showedvsini≥10kms−1,overhalfofthecoolandwarmBHBstarsdoso,inbothglobularclustersandthefield(Peterson,Tarbell,&Carney1983;Peterson1985;Peterson,Rood,&Crocker1995).Globular-clusterBHBstarswithTeff≥12,000Kalsoshowverylowvsinivalues(Behretal.1999,2000).However,toclarifythepathwaysofformationofRRLyraesandsdB’sinmetal-richandmetal-poorpopulations,wemustanalyzeamuchlargersampleofbulgeHBstarsinwhichRRLyraesaredetectedphotometrically,andwhichgoesblueenoughandfaintenoughtodetectBHBstarshotterthan12,000K.
ReachingthesdB’s–thehottest,faintestBHBstars–requiresgoingtomagnitudeswellbeyondthoseofourcurrentsurvey.Giventheaveragemagnitude WeareindebtedtoS.AllenofU.C.SantaCruzandB.DormanofGoddardforwritingscriptstoconverttheVAXversionofKuruczcodesintoUNIXversions,andthankE.M.GreenandJ.Liebertforusefuldiscussions.RCPandDMTgratefullyacknowledgesupportfromNSFgrantAST-9900582toAstrophysicalAdvancesandNSFgrantsAST-9157038,INT-9215844,andAST-9820603toOhioStateUniversity.ChristopherJ.Burkeassistedwiththereductionofthephotometry. REFERENCES Allard,F.,Wesemael,F.,Fontaine,G.,Bergeron,P.,&Lamontagne,R.1994,AJ,107,1565Bailey,J.A.,&Glazebrook,K.1999,2dFUserManual,Anglo-AustralianObservatory,available fromhttp://www.ast.cam.ac.uk/AAO/2df/manual.htmlBehr,B.B.,Cohen,J.G.,McCarthy,J.K.,&Djorgovski,S.G.1999,ApJ,517,L135 Behr,B.B.,Djorgovski,S.G.,Cohen,J.G.,&McCarthy,J.K.,Cote,P.,Piotto,G.,&Zoccali, M.2000,ApJ,inpressBressan,A.,Chiosi,C.,&FagottoF.1994,ApJS,94,63 Brown,T.M.,Ferguson,H.C.,Davidsen,A.F.,&Dorman,B.1997,ApJ,482,685 –16– Burstein,D.,Bertola,F.,Buson,L.M.,Faber,S.M.,&Lauer,T.R.1988,ApJ,328,440Castelli,F.,Gratton,R.G.,&Kurucz,R.L.1997,A&A,318,841Chaboyer,B.,Green,E.M.,&Liebert,J.1999,AJ,117,1360 D’Cruz,N.L.,Dorman,B.,Rood,R.T.,&O’Connell,R.W.1996,ApJ,466,359Dorman,B.O’Connell,R.W.,&Rood,R.T.1995,ApJ,442,105Dupree,A.K.,Sasselov,D.D.,&Lester,J.B.1992,ApJ,387,L85Faber,S.M.1983,HighlightsAstron.,6,165Friel,E.D.1995,ARA&A,33,381 Frogel,J.A.,Terndrup,D.M.,Blanco,V.M.,&Whitford,A.E.1990,ApJ,353,494Frogel,J.A.,&Whitford,A.E.1987,ApJ,320,199Graham,J.A.1982,PASP,94,244 Gray,R.O.,Corbally,C.J.,&Philip,A.G.D.1996,AJ,112,2291Green,E.M.,Bowers,T.C.,&Saffer,R.A.2000,inpreparation.Houdashelt,M.1995,thesis,OhioStateUniversity Jaschek,C.,&Jaschek,M.1990,TheClassificationofStars(Cambridge:CambridgeU.Press)Kaluzny,J.,&Rucinski,S.M.1995,A&AS,114,1Kaluzny,J.,&Udalski,A.1992,ActaAstr.,42,29 Kinman,T.D.,Kraft,R.P.,&Suntzeff,N.B.1994,AJ,108,1722 Kraft,R.P.,Sneden,C.,Smith,G.H.,Shetrone,M.D.,&Fulbright,J.1998,AJ,115,1500Kurucz,R.L.1991,inPrecisionPhotometry:AstrophysicsoftheGalaxy,ed.A.G.D.Philip,A. R.Upgren,&K.A.Janes(Schenectady:Davis)Kurucz,R.L.1995,inA.S.P.Conf.Ser.81,LaboratoryandAstronomicalHighResolutionSpectra, ed.A.J.Sauval,R.Blomme,&N.Grevesse(SanFrancisco:ASP),583Kurucz,R.L.,&Avrett,E.H.1981,S.A.O.Spec.Rpt.,391 Kurucz,R.L.,&Bell,B.1995,KuruczCD-ROMNo.23(Cambridge,MA:SmithsonianAstro-physicalObservatory)Lewis,I.J.,Glazebrook,K.,&Taylor,K.1998,Proc.SPIE,3355,828 –17– Liebert,J.,Saffer,R.A.,&Green,E.M.1994,AJ,107,1408Longo,R.,Stalio,R.,Polidan,R.S.,&Rossi,L.1989,ApJ,339,474McWilliam,A.,&Rich,R.M.1994,ApJS,91,749 Moehler,S.,Sweigart,A.V.,&Catelan,M.1999,A&A,351,519Montgomery,K.A.,Janes,K.A.,&Phelps,R.L.1994,AJ,108,585 Nave,G.,Johansson,S.,Axner,O.,Ljungberg,P.,Malmsten,Y.,&Baschek,B.1994,Phys. Scripta,49,581O’Connell,R.W.1999,ARA&A,37,603Peterson,R.C.1985,ApJ,294,L35 Peterson,R.C.,Carney,B.W.,&Latham,D.W.1996,ApJ,465,L47Peterson,R.C.,Dorman,B.,&Rood,R.T.2000,inpreparation.Peterson,R.C.,&Green,E.M.1998,ApJ,502,L39 Peterson,R.C.,Rood,R.T.,&Crocker,D.A.1995,ApJ,453,214Peterson,R.C.,Tarbell,T.D.T.,Jr.,&Carney,B.W.1983,ApJ,265,972Preston,G.W.1964,ARA&A,2,23Rich,R.M.,etal.1997,ApJ,484,L25Rood,R.T.1973,ApJ,184,815 Sadler,E.M.,Rich,R.M.,&Terndrup,D.M.1996,AJ,112,171Saffer,R.A.,Bergeron,P.,Koester,D.,&Liebert,J.1994,ApJ,432,351 Saffer,R.A.,Keenan,F.P.,Hambly,N.C.,Dufton,P.L.,&Liebert,J.1997,ApJ,491,172Shortridge,K.1995,TheFigaro http://www.aao.gov.au/figaro/ Data Reduction System, available at Smith,G.,&Lankshear,A.1998,Proc.SPIE,3355,905 Smith,H.A.1995,RRLyraeStars(Cambridge:CambridgeU.Press)Stetson,P.B.1987,PASP,91,191 Sweigart,A.V.1997,inThirdConferenceonFaintBlueStars,ed.A.G.D.Philip,J.Liebert,& R.A.Saffer(Schenectady:Davis),3 –18– Terndrup,D.M.1988,AJ,96,884 Terndrup,D.M.,Frogel,J.A.,&Whitford,A.E.1990,ApJ,357,453Terndrup,D.M.,Sadler,E.M.,&Rich,R.M.1995,AJ,110,1774Tiede,G.P.,Frogel,J.A.,&Terndrup,D.M.1995,AJ,110,2788 Tyson,N.D.,&Rich,R.M.1993,inGalacticBulges,IAUSymp.No.153,ed.H.DeJonghe&H. J.Habing(Dordrecht:Kluwer),133Walker,A.R.,&Terndrup,D.M.1991,ApJ,378,119Whitford,A.E.1978,ApJ,226,777 Wilhelm,R.,Beers,T.C.,&Gray,R.O.1999,AJ,117,2308Yi,S.,Demarque,P.,&Oemler,A.1998,ApJ,492,480 –19– Fig.1.— Color-magnitudediagraminVvs.B−Vforthefieldstudiedhere.Thelargesymbolsrepresentstarsobservedspectroscopically,withfilledpointsforthehotstarslistedinTable1,andopencirclesfortheothers.Smallpointsareplottedfor17%ofallotherstarswithVandBphotometry,toshowtherelationshipofthestarsobservedspectroscopically.Main-sequencecontaminationbecomessignificantredwardofB−V=0.5.Fig.2.— Color-colordiagraminUBV.Thelargesymbolsareasdescribedforthepreviousfigure.Smallpointsareplottedfor17%ofallotherstarswithU,B,andVphotometry.Theconcentrationofpointscorrespondstothemain-sequenceturnoffoftheforegrounddisk.ThecurvesrepresentthecolorsofBHBstarswith[Fe/H]=0fromTeff,logg=7250K,2.7to13,000K,4.1,reddenedbyE(B−V)=0.0,0.3,(themeanforthefield),and0.5.ThereddeningvectorE(U−B)=0.70×E(B−V)ofKurucz(1991)isdepictedbythearrow,withheadsatthelatterreddenings.ExceptwhereTeff≤7500K,theBHBmodelUBVcolorsataparticularTeffandloggchangeby<0.01magingoingto[Fe/H]=+0.3or−0.5.ThisillustratesthatthecolorsofBHBstars,regardlessofmetallicity,remaindistinguishablefromthecolorsofmain-sequencestarscoolerthan7000Kwithnear-solarmetallicityuntilE(B−V)>0.50.Fig.3.— Observedspectra(heavyline)arecomparedtocalculations(lightlines)forsixstarsofvarioustemperaturesandmetallicities.Thenormalizedspectralfluxisplottedagainstwavelengthin˚A,withsuccessivespectradisplacedupwardsby20%ofthecontinuum.TheBalmerlinesHδ(4101˚A),Hγ(4340˚A),andHβ(4861˚A)arethestrongestfeaturesshowninthespectrum.ManyweakmetallinesareseensuperimposedontheBalmerlinewings,especiallyatlowertemperatures,highermetallicities,andbluerwavelengths.Attheupperrightofeachspectrum,themodelatmosphereparametersTeff,logg,[Fe/H],and[α/Fe]usedforthebest-fitcalculationarelistednexttothenumberofeachstar.(a)4060–4230˚A.(b)4230–4400˚A.(c)4400–4570˚A.(d)4780–4950˚A.Fig.4.— AsinFigure3,butadoptingmodelsthatdifferby0.5dexinlogg.(a)4230–4400˚A.(b)4780–4950˚A. Table1. # ID RA (2000) Dec Astrometry,Photometry,andStellarParametersforHotStars V σ B−V σ U−B σ Vrad Teff logg Z D MV E δU δB –20–Table1—Continued # ID RA (2000) Dec V σ B−V σ U−B σ Vrad Teff logg Z D MV E δU δB Note.—Units:RAinh,m,s;Decin ◦,′,′′;TeffinK;Vradinkms−1relativetotheinternalF-startemplate;Dinkpc. –21–13 14 15V161718−.5 0B − V .51 −.50.51−.5 U − B0 .5B − V 11.5 因篇幅问题不能全部显示，请点此查看更多更全内容