On the Lack of a Soft X-Ray Excess from Clusters of Galaxies

APreprinttypesetusingLTEXstyleemulateapjv.6/22/04

ONTHELACKOFASOFTX-RAYEXCESSFROMCLUSTERSOFGALAXIES

JoelN.BregmanandEdwardJ.Lloyd-Davies

DepartmentofAstronomy,UniversityofMichigan,830DavidM.DennisonBuilding,AnnArbor,Michigan48109-1090

DraftversionFebruary5,2008

arXiv:astro-ph/0602527v1 24 Feb 2006ABSTRACT

AsoftX-rayexcesshasbeenclaimedtoexistinandaroundanumberofgalaxyclustersandthisemissionhasbeenattributedtothewarm-hotintergalacticmediumthatmayconstitutemostofthebaryonsinthelocaluniverse.Wehavere-examinedastudyoftheXMM-NewtonobservationsonthistopicbyKaastraetal.(2003)andfindthattheX-rayexcess(ordeficit)dependsuponGalacticlatitudeandappearstobemostcloselyrelatedtothesurfacebrightnessofthe1/4keVemission,whichislargelyduetoemissionfromtheLocalhotbubbleandthehalooftheMilkyWay.WesuggestthatthepresenceofthesoftX-rayexcessisduetoincorrectsubtractionofthesoftX-raybackground.Ananalysisisperformedwherewechoosea1/4keVbackgroundthatissimilartothebackgroundnearthecluster(andforsimilarHIcolumn).WefindthatthesoftX-rayexcesslargelydisappearsusingourbackgroundsubtractionandconcludethatthesesoftX-rayexcessesarenotassociatedwiththetargetclusters.Wealsoshowthatthedetectionsof“redshifted”OVIIlinesclaimedbyKaastraetal.(2003)arecorrelatedwithsolarsystemchargeexchangeemissionsuggestingthattheyarenotextragalacticeither.

Subjectheadings:X-rays:galaxies:clusters;methods:dataanalysis

1.INTRODUCTION

ClustersofgalaxiescontainX-rayemittinghotgas(107-108K)thataccountsfor∼11%ofthetotalmassofthesystemandcontainsmorebaryonsthanthevisiblegalaxies(Allenetal.2002,2003).Therehavebeenmanystudiesofthishotgasaswellassearchesforothercoolergaseouscomponentsandothertypesofemission.Bothexcessabsorptionandemissionhavebeenclaimedtobepresentinclusters,suggestiveofmaterialcolderthantheambienthotclustermaterial,althoughthesehavebeencontroversialissues.Theclaimsaboutexcessabsorp-tionarosefromtheEinsteinObservatorySSSspectra(Whiteetal.1991),wherethesoftemissionwaslessthanwouldbeexpectedfromGalacticabsorptionofthefree-freespectrumofacluster.TheseSSSspectrahadtobecorrectedforthebuildupoficeintheopticalpath,sotherewassomeconcernthatifthecorrectionforicewaswrong,itmightleadtotheobservedeffect.Thisresultwasnotconfirmedwithsubsequentinstruments,suchasROSAT(Arabadjis&Bregman2000)orXMM-Newton(Petersonetal.2003),sowecansafelyconcludethattheoriginalstudywasincorrectandthatthereisnosubstantialabsorbingmedium.

SubstantiallymorecontroversialisthesubjectofanadditionalemissioncomponentatsoftX-rayenergies(0.1-1keV).Formanyclusters,itisclaimedthattheemissiondetectedbytheExtremeUltravioletExplorer(EUVE)andbyfourdifferentX-raytelescopescannotbeexplainedbyclusterfree-freeemissionthatisab-sorbedbycoldMilkyWaygas(e.g.,Lieuetal.(1996,1999,2000);Durretetal.(2002);Kaastraetal.(2003),andreferencestherein).Theyarguethattheemissionbe-comesmoreprominentwithincreasingradiusfromtheclustercenterrelativetotheharderemissionoftheclus-ter,andthatthetemperatureoftheemissionistypi-cally0.1-0.3keV.Theyinterpretthisemissionasbeing

Electronicaddress:ejdavies@umich.edu

eithernonthermal,duetocosmicraysinthecluster(Sarazin&Lieu1998)orthermal,duetogasat1-3×106K(e.g.,Kaastraetal.(2003)).Ifitisthermal,itsmassmaybecomparabletothatinthehotterambientcom-ponent,soitwouldhavecosmologicalconsequences.However,theseworkshavebeencriticizedforsev-eralreasons.Bergh¨oferetal.(2000)(also,Bowyeretal.(2000),andreferencestherein)arguedthatflat-fieldingcorrectionswerenotproperlyappliedtotheEUVEdata,andaftermakingthiscorrection,noexcessemissionisfound,withtheexceptionoftheComacluster.TheROSATdatawereexaminedbyArabadjis&Bregman(2000)whofoundthattheclusterspectracouldbefitwithahotfree-freespectrumplusGalacticabsorptionandthatnoadditionalsoftcomponentwasneeded(ex-ceptfortheComacluster).TheX-rayspectrafromBeppo-SAXwasinvestigatedbyBergh¨ofer&Bowyer(2002),followingthestudybyKaastraetal.(1999)thatAbell2199containedasoftcomponent.Usingadifferentapproachtotheanalysis,Bergh¨oferetal.(2000)foundnoevidenceforanadditionalsoftcomponenteitherinAbell2199,orinAbell1795.ThedifferencesbetweentheseworksandthoseofLieuandcollaboratorshavetodowiththetechnicaldetailsofbackgroundsubtractionandflat-fielding.

Recently,Kaastraetal.(2003)usedXMM-NewtondatatosearchforsoftX-rayexcessemission(0.2keV)inasampleof14galaxiesclusters.Theyfindevidenceforexcessemissioninthespectraofseveraloftheclus-tersandtheyshowthatitisbroadlyextendedacrosstheclusters.TheyattributethisemissiontothepresenceofhotgasininterclusterfilamentsthatcontaintheWarm-HotIntergalacticMedium(WHIM)neartheseclusters.ThedetectionoftheWHIMwouldbeamajordiscoveryandwouldbebestaccomplishedwithXMMduetoitslargecollectingarea.Therefore,weexaminethisresulttounderstandifitissubjecttothecriticismthathavebeenraisedinotherobservations.

2

3

Fig.2.—Theall-skyimageofthe1/4keVX-rayintensity(R12bandfromROSAT),alongwiththelocationsofthegalaxyclustersanalyzedbyKaastraetal.(2003).TheblackcircleshaveanapparentN(HI)deficit(X-rayexcessobjects)andarelocatedinregionsofhigherthanaverageX-raybrightness(notethattwoclusterslieataboutl=10◦,b=50◦).ThewhitecircleshaveanapparentN(HI)excessandgenerallylieinregionswherethebackgroundintensityislowerthanaverage.Thepurple/greycirclesarethetwoclusterswithneitheranexcessordeficit.

Theall-sky1/4keVmap(R12)hasagreatdealofstructureduetowell-knownfeatures,suchastheNorthPolarSpur,andthereisabrighteningtowardthepolesduetothepresenceofaGalactichalo(0.1-0.2keV)alongwiththeLocalBubbleofhotgas(0.1keV),whichisprob-ablyelongatedtowardhighGalacticlatitudes(Fig.2).Uponthisfigure,weshowclustersthathavetoolittleabsorption(fN<-0.3;thesoftexcessobjects),excessabsorption(fN>0.3),andthoseconsistentwithGalac-ticabsorption.MostoftheobjectswithfN<-0.3lieinregionsofenhancedemissioninthemapandthesere-gionsareoftenpartoflargerstructures.Forexample,MKW3s,Abell2052,theVirgocluster,Abell1795,andAbell1835lieonorveryclosetotheNorthPolarSpur,anoldsuperbubble.TheclusterAbellS1101(alsoknowasSersic159-03)liesontheedgeofalargebrightregiontowardtheSouthernGalacticPoleandComacoverstheNorthGalacticPole,anotherlargebrightregion(moreonComabelow).Incontrast,allfourobjectswithfN>0.3(excessabsorption)lieinregionsoflowdiffuseX-rayemission(Abell496isjustafewdegreesawayfromabrightridge).Oneofthesefourobjects,theNGC533cluster,isathighGalacticlatitude(−60◦)andwithsame21cmcolumnastowardthehighlatitudesourcesMKW3sandAbell2052(b=50◦;bothareexcessemis-sionobjects),theprimarydifferencebeingthevaluesoftheGalacticsoftX-raybackground.

3.CONSEQUENCESOFMATCHEDBACKGROUND

SUBTRACTION

Iftheinferencethatthepresenceofthesoftexcessisrelatedtotheremovalofthesoftbackgroundiscorrect,thenthesoftexcessshouldbereducedorvanishwhenamoreappropriatebackgroundisused.ThestandardmethodusedforbackgroundsubtractionofclusterswithXMM-Newtonistousea“blanksky”backgroundcre-atedbytakinganumberofobservationsthatdonotcon-

tainextendedsourcesorbrightpointsources,andstack-ingthemtogethertoobtainameanbackground.Sim-pleprepackagedbackgroundsfortheEPIC-pnandmosinstrumentshavebeenproducedbyLumbetal.(2002)andRead&Ponman(2003)haveproducedanextensivesetofbackgroundsforeachcombinationofinstrumentmodeandfilter.Howevertheseareatbestmeanback-grounds,averagedacrossthesky,andareinrealitythemeanofasmallnumberofbiasedskypositionswhoserelationtotherealall-skymeanisundetermined.SincethelowenergyendoftheX-raybackgroundisthemostspatiallyvariable,duetothegalacticbackgroundandab-sorption,theerrorsintroducedbytheuseofthesemeanbackgroundswillbelargestatlowenergies.

Abackgrounddrawnfrom“blank”fieldswithsimi-larpropertiestothoseofthetargetclusterfields,ratherthanrandomones,shouldmatchtheclusterbackgroundsmuchmoreclosely.Inordertotestthisweconstructed“blanksky”backgroundsmatchedtopropertiesoftheindividualclusterfields.Thetechniqueweusedtoac-complishthisinvolvesidentifyingthreeparametersthatarelikelytoaffecttheobservedbackground:N21cm,R12andparticlebackground.TheR12valuewascalculatedfromameanoffourpointsarrangedaroundtheclus-tercenteratadistanceof2◦.Forallthere-analyzedclustersthevirialradiusislessthan0.5◦sothisshouldeliminateanypossibilityofcontaminationfromclusteremission.“Blanksky”fieldsarethenselectedwithN21cmandR12valuesascloseaspossibletothatoftheclusterfieldtobematched.Theeventlistsarebrokeninto50secondtimeblocksandaparticlebackgroundcalculatedforeachblockusingthefluxmeasuredinthe12-15keVband.Aminimizationisthenperformedbyremoving(andadding)blocksfromthepooluntilthemeanvaluesofthethreeparametersisascloseaspossibletotheval-uesoftheparametersfortheclusterfield.A“blanksky”backgroundisthenconstructedfromtheselectedevent

4

TABLE1

Propertiesofthefourre-analyzedclusterfieldsinorderofincreasinggalacticN21cm.TheR12valuesarethemeanof

fourpoints,2degreesawayfromtheclustercenterindifferentdirections.

Cluster

OBSID

l(deg.)

b(deg.)

z

N21cm(1020cm−2)

R12

Exp.(ks)

Abell17952NH(1020cm2)1.510.504cm)32100

100

200

300

400

0

100

200

300

2AbellS1101Abell1835MKW3SNH(1020400

Radius(kpc)Radius(kpc)

Fig.3.—ComparisonofthefittedNHobtainedbyKaastraetal.(2003)(usinganmean“blanksky”background)forfourclusters(AbellS1101,Abell1835,Abell1795andMKW3s)(opencircles)withthoseobtainedusingbackgroundsmatchedtopropertiesoftheindividualfields(solidsquares).

listblocks.Thisbackgroundneedstohaveanexposuretimesignificantlylargerthanthatofthesourcetoavoiddegradingthedataqualitywithextranoise.

Oneotherproblemwiththesestackedbackgroundsisthatthe“blank”fieldsinevitablycontainlargenumbersofpointsources.Traditionallytheseareexcisedandlargenumbersoffieldsarestackedinordertoreducetheeffectofthemissingdata.Howeveritshouldbenotedthattheeffectsoftheseexcisionsareclearlyvisibleinimagesofprepackagedbackgrounds.Sincethenumberof“blank”fieldsavailableismuchreducedwhentryingtomatchabackgroundtoatargetobservation,weinsteadmaskouttheregionscontainingpointsourcesfromthesourceandbackgroundwhencreatingthespectraforfitting.Inselecting“blank”fieldstobuildabackgroundthereisastrongconstraintthattheyneedtobeasfreeofpointsourcesaspossiblesothatassmallanareaoftheclustermustbemaskedfromthespectralfitting.Thesecon-straintscombinedwiththoseofthebackgroundparam-eters(N21cmandR12)meanthatselectingobservationsforuseinthematchedbackgroundsisacomplexprocess.WeselectfourofthebestcandidatesforapossiblesoftexcessfromKaastraetal.(2003)tostudytheeffectsofusingourmatchedbackgroundtechnique.ThesystemsareAbellS1101(alsoknowasSersic159-03),Abell1835,Abell1795andMKW3s.Thepropertiesfortheseclus-terfieldsareshowninTable1.ForeachoftheseclusterseitherthedatausedbyKaastraetal.(2003),orlongerexposuresifavailable,wereused.Thedatawasreducedintheusualmannerandcleanedbyperformingaiter-ative3-sigmaclippingonthe12-15keVlight-curvetoremoveperiodsofhighparticlebackground.OnlyEPIC-pndatawasusedsinceitistheinstrumentthatreceivesthehighestcountrateandiftheeffect(softexcess)isnotdetectableinasingleinstrumentthenadetectionusingmultipleinstrumentswouldnotbereliablegiventheun-certaintiesincross-calibrationbetweeninstruments.Ta-ble2liststheobservationsusedtoconstructthematchedbackgroundsandmeanpropertiesofthematchedback-grounds.Spectrawereextractedinannuliabouttheclus-tercenterswithaminimumannulussizeof5arcsecgrow-ingwithradiustopreservethesignal-to-noise.Thelatest

5

TABLE2

Propertiesoftheobservationsusedtoconstructthematchedbackgroundsforthefourre-analyzedclusterfields.

Cluster

Abell1795AbellS1101Abell1835

MKW3s

N21cm(1020cm−2)Exposure(ks)

1.16101.5

1.82145.0

2.32192.6

2.32124.5

6

Fig.5.—SpectraofbackgroundstakenfromobservationscloseintimetothoseoffouroftheclustersshowninFigure6ofKaastraetal.(2003):0109060201(MKW3sandAbell2052),0124100101(AbellS1101)and0125300101(Coma).Observation0109060201(MKW3sandAbell2052)showsalargeexcessbetween0.4-1.0keV,particularlyaroundtheOVIIandOVIIIlines,presumablyduetogeocoronalorheliosphericchargeexchange.Thereisalsosomeevidenceofaslightenhancementforobservation0124100101(AbellS1101),especiallyintheCVIlines.

duetothehighersoftX-raybackgroundinthefieldsusedtoconstructourmatchedbackground(sinceAbellS1101hasahighR12andlowN21cm)comparedtothefieldsusedtoconstructthemeanbackgroundofLumbetal.(2002).IfthemeanbackgroundofLumbetal.(2002)isusedforbackgroundsubtractionthesoftbackgroundemissionseeninFigure4willnotberemoved.Thiswillresultinanapparentsoftexcessinthefinalspectrum.

4.OVIIEMISSION

Kaastraetal.(2003)alsoreportred-shiftedOVIIKαlinesfromanumberofclustersintheirsamplewhichwouldalsobeanindicationforforthepresenceofcoolgas.Howeverforonlytwooftheclustersistheresultre-allysignificant,Abell2052(atthe99%confidencelevel)andMKW3s(atthe91%confidencelevel),andthissignificancehangsoftheassumptionthattheemissionisdominatedbythe574eVresonancelineandthe569eVintercombinationline.Iftheemissionisdominatedbythe561eVforbiddenlinethough,thesignificanceisgreatlyreduced.Kaastraetal.(2003)dismissforbiddenlineemissionduetophotoionisationbuttheydonotcon-sideremissionfromheliosphericandgeocoronalchargeexchange.ThisistheresultofcollisionsbetweensolarwindionsandneutralatomsfromISM(heliospheric)andexosphere(geocoronal)andoneofthestrongestexpectedlinesistheOVIIKαforbiddenline(seeSnowdenetal.

(2004),Wargelinetal.(2004)andreferencestherein).Solarwindchargeexchangeemissionisexpectedtovarysignificantlywithtimeandpositiononthesky.Thegeo-coronalemissionisgenerallymuchweakerthatthehelio-sphericemissionexceptduringperiodsofenhancedsolaractivity.

Atthispointitshouldbenotedthattheobserva-tionsofAbell2052andMKW3susedbyKaastraetal.(2003)weretakenwithinadayofeachother(2000-08-21and2000-08-22)andareonlyslightlymorethanadegreeapartonthesky.Totestwhethertheseobservationscouldhavebeenaffectedbysignificantchargeexchangeemissionweselectedobservationsofnon-extendedsourcestakenshortlybeforeorafterwardandpointedinasimilardirection.Backgroundspec-trafortheseobservationsareshowninFigure5.ItcanbeseenthatObservation0109060201,whichwastakenshortlyaftertheobservationsofAbell2052andMKW3sandisabout30degreesontheskyawayfromthem,hasalargeexcessofemissionbetween0.4-1.0keVcom-paredtothebackgroundsfortheobservationscontem-poraneouswiththoseoftheclustersthatdonotshowsignificant“redshifted”OVIIemission.Thiscorrelationextendstothefactthatthebackgroundcontemporane-ouswiththeobservationofComa(theclusterforwhichKaastraetal.(2003)detecttheleastexcessOVIIemis-sion)alsoshowstheleastevidenceofchargeexchange

Fig.6.—TheN21cmimagetheofthe1/4keVX-raycontoursintheareregionsuperimposedoftheComaupontheall-skyininacenterlocalHIofminimum,the5◦diameterwhichiswhiteoftencircle).correlatedTheComaclusterclusterlies(atartifact.

the1/4keVband.Thebrightpixelsatl=55with◦,babrightening=84◦isanemission.Thiswouldseemtostronglyunderminethecasefortheemissionbeingextragalacticinorigin.

5.DISCUSSIONANDCONCLUSIONS

ToinvestigatethecauseofthesoftexcessesobservedinseveralclustersbyKaastraetal.(2003)wehavestud-iedthecorrelationofpropertiesoftheclusterfieldswithKaastra’ssoftexcess/deficitmeasurementsandfindthattheexcess/deficitiscorrelatedwithbothgalacticlati-tudeandthesoftX-ray(R12)background.FromthisweinferthatincorrectsubtractionofthesoftX-rayback-groundisalikelycauseoftheobservedexcess/deficits.Totestthiswere-analyzedthedatafromfourclustersforwhichKaastraetal.(2003)measuresignificantsoftexcesses.Usingbackgroundsmatchedtotheproperties(N21cm,R12andparticlebackground)oftheindivid-ualfieldsweobtainfittedhydrogencolumnsconsistentwiththegalactic21cmcolumnsandconsiderablyhigherthanthosemeasuredbyKaastraetal.(2003).Wethere-foreconcludethatthesoftexcess/deficitobservedbyKaastraetal.(2003)mostlikelytheresultofthefieldbyfieldvariationofthesoftX-raybackgroundthatre-mainsinthedataafteramean“blanksky”backgroundissubtracted.

WehavealsoexaminedtheevidenceforredshiftedOVIIlinespresentedbyKaastraetal.(2003).Thisdetec-tionissignificantonlyintwoclusters,Abell2052andMKW3s,andonlyiftheemissionisdominatedbytheresonanceandintercombinationlines.Howeverweshowthattheobservationsofthesetwoclusters,whichwheretakenwithinadayofeachotherandareadegreeapartonthesky,arelikelycontaminatedwithheliosphericorgeocoronalchargeexchangeemissioncontainingstrongOVIIforbiddenlineemission.Giventhecorrelationbe-tweenourexpectationofchargeexchangecontaminationandthethedetectionof“redshifted”OVIIemissionbyKaastraetal.(2003)weconcludethatthecaseforthe

7

emissionbeingassociatedwithclustersisveryweak.Theonesourcethatseveralauthorsagreeuponashav-inganapparentX-rayexcessistheComacluster,soweexaminewhetherthisistrulyevidenceforaWHIMcomponentincosmicfilaments(Bonamenteetal.2003;Finoguenovetal.2003).EvenontheR12map,onecanseeanenhancementinthisregionofthesky,suggestivethatitisduetotheComacluster(Fig.6).However,thereisalocalminimuminthe21cmcolumndensityatComa(Dickey&Lockman1990;Hartmann&Burton1997),andthiswascertainlynotcausedbytheComacluster.AlocalminimumintheHIskywouldpermitustoseethesoftX-rayemissionfromtheGalactichalomorereadily,causingabrighteninginthesoftX-raymapoftheskyatthatlocation.Itisextremelydifficulttosep-arateaGalactic(halo)brighteninginR12fromthatas-sociatedwiththeComacluster.ThereareclearexamplesofbrighteninginthetheR12fluxatlow21cmcolumnregions,suchasintheLockmanhole(Snowdenetal.1994).Furthermore,the21cmcolumndensityissolowinthisdirection(<1020cm−2)thatitispoorlyknownasthevariousinstrumentalcorrections(fromside-lobes,etc.)becomeasignificantfractionofthesignal(Hartmannetal.1996);thetrue21cmcolumnmaybelowerthantheusualvaluesquoted.Furthercomplicat-ingtheanalysisofComaisthatitsextremelylowNHwouldmakefinding“blanksky”fieldswithsimilarprop-ertiesverydifficult.Therefore,wearecautiousaboutclaimsthattheComaclusterpossessesasoftX-rayex-cessandwenotethatArabadjis&Bregman(2000)wereabletofitafree-freeemissionmodeltoComa(ROSATPSPCdata)withoutanadditionalsoftexcess,buttheyrequiredaGalacticcolumndensity(6×1019cm−2)lowerthanthevaluesofHartmann&Burton(1997)(9×1019cm−2).

DespiteourconcernsontherealityofasoftX-rayex-cessfromclustersofgalaxies,itmightbepossibletoisolatethiscomponentspectrally,whichshouldbepossi-blebecausethissoftexcessisratherbright.IfonecouldshowredshiftedOVIIlineemissionfromtheoutskirtsofaclusterthatwasnotblendedwiththeGalacticfea-ture(orcontaminatedwithchargeexchangeemission),itwouldconstitutestrongevidencefortheWHIMaroundclusters.Currently,theOVIIlineisblendedwiththeGalacticOVIIlineforlowredshiftclustersanddoesnotappeartobepresentinthemoderateredshiftcluster(Abell1835)whosespectrumwouldimplyasoftexcess(fN<-0.3).Theuseofsmaller,higherredshiftclus-terswouldalsohelptoisolatethesoftexcessbecauseonecouldtakealocalbackgroundfromthesamefieldofviewasthatusedtoimagethecluster.

WewouldliketothankJimmyIrwin,RenatoDupke,EricMiller,JohnArabadjis,WiltSanders,andSteveSnowdenforcommentsandadvice.Wewouldalsoliketothanktherefereeforhisusefulcomments.WeacknowledgesupportfromNASAgrantsNAG5-10765,NAG5-13137,andGO1-2147X.ThisresearchhasmadeuseofdataobtainedfromtheHighEnergyAstro-physicsScienceArchiveResearchCenter(HEASARC),providedbyNASA’sGoddardSpaceFlightCenter.WeacknowledgetheuseofNASA’sSkyViewfacility(http://skyview.gsfc.nasa.gov)locatedatNASAGod-

8

dardSpaceFlightCenter.

REFERENCES

Allen,L11

S.W.,Schmidt,R.W.,&Fabian,A.C.2002,MNRAS,334,Allen,MNRAS,S.W.,342,Schmidt,287

R.W.,Fabian,A.C.,&Ebeling,H.2003,Arabadjis,J.S.&Bregman,J.N.2000,ApJ,536,144Bergh¨Bergh¨oofer,fer,T.T.W.W.,&Bowyer,Bowyer,S.,S.&2002,Korpela,ApJ,E.565,2000,L17ApJ,5,695Bonamente,M.,Joy,M.K.,&Lieu,R.2003,ApJ,585,722Bowyer,Dickey,J.S.,M.Bergh¨&Lockman,ofer,T.,&F.Korpela,J.1990,ARA&A,E.2000,ApJ,28,215

8,L135Durret,F.,Slezak,E.,Lieu,R.,DosSantos,S.,&Bonamenta,M.Finoguenov,2002,A&A,A.,390,Briel,397

U.G.,&Henry,J.P.2003,preprint,Hartmann,astro-ph/0309019

D.&Burton,W.B.1997,AtlasofGalacticNeutralHydrogen(Cambridge:CambridgeUniversityPress)

Hartmann,1996,A&A,D.,119,Kalberla,115

P.M.W.,Burton,W.B.,&Mebold,U.Kaastra,J.,Lieu,R.,Mittaz,J.P.D.,Bleeker,J.A.M.,Mewe,Kaastra,R.,Colafrancesco,J.,Lieu,R.,S.,Tamura,&Lockman,T.,Paerels,F.J.1999,F.B.S.,ApJ,&519,denHerder,L119J.W.2003,A&A,397,445

Lieu,—.2000,R.,Bonamente,A&A,3,497

M.,&Mittaz,J.P.D.1999,ApJ,517,L91Lieu,Murphy,R.,Mittaz,E.M.,J.&P.Hwang,D.,Bowyer,C.-Y.S.,1996,Breen,Science,J.O.,274,Lockman,1335

F.J.,Lumb,D.H.,Warwick,R.S.,Page,M.,&DeLuca,A.2002,A&A,Peterson,3,93

J.R.,Kahn,S.M.,Paerels,F.B.S.,Kaastra,J.S.,Tamura,T.,Bleeker,J.A.M.,Ferrigno,C.,&Jernigan,J.G.Read,2003,A.ApJ,M.&590,Ponman,207

T.J.2003,A&A,409,395Sarazin,C.L.&Lieu,R.1998,ApJ,494,L177

Snowden,1182

S.L.,Collier,M.R.,&Kuntz,K.D.2004,ApJ,610,Snowden,Plucinsky,S.P.L.,P.,Egger,Sanders,R.,W.Freyberg,T.,Schmitt,M.J.J.,H.McCammon,M.M.,Tr¨umper,D.,J.,&Voges,W.1997,ApJ,485,125

Snowden,McCammon,S.L.,D.,Hasinger,&Sanders,G.,W.Jahoda,T.1994,K.,ApJ,Lockman,430,601

F.J.,Wargelin,B.J.,Markevitch,M.,Juda,M.,Kharchenko,V.,Edgar,R.,&Dalgarno,A.2004,ApJ,607,596