X-ray emission from the extended disks of spiral galaxies

1.6 MB·English

by  R. A. Owen

Mon.Not.R.Astron.Soc.000,1–17(2009) Printed27January2009 (MNLATEXstylefilev2.2)
X-ray emission from the extended disks of spiral galaxies
R.A. Owen, R.S. Warwick
9 X-ray&ObservationalAstronomyGroup,Dept.ofPhysics&Astronomy,UniversityofLeicester,UniversityRoad,LeicesterLE17RH,U.K.
0
0
2

n 27January2009
a
J

7 ABSTRACT
2
WepresentastudyoftheX-raypropertiesofasampleofsixnearbylate-typespiralgalax-
]
iesbasedonXMM-Newtonobservations.Sinceourprimaryfocusisonthelinkagebetween
A
X-rayemissionandstarformationinextended,extranucleargalacticdisks,wehaveselected
G
galaxieswithnearface-onaspectandsufficientangularextentsoastobereadilyamenableto
. investigationwiththemoderatespatialresolutionaffordedbyXMM-Newton.Afterexcluding
h
regionsineachgalaxydominatedbybrightpointsources,westudyboththemorphologyand
p
- spectral properties of the residual X-ray emission, comprised of both diffuse emission and
o the integrated signal of the fainter discrete source populations. The soft X-ray morphology
r
generallytracestheinnerspiralarmsandshowsastrongcorrelationwiththedistributionof
t
s UV light, indicative of a close connection between the X-ray emission and recent star for-
a
mation.Thesoft(0.3–2keV)X-rayluminositytostarformationrate(SFR)ratiovariesfrom
[
1−5×1039ergs−1(M yr−1)−1,withanindicationthatthelowerrangeofthisratiorelates
⊙
1 to regionsof lower SFR density.The X-ray spectra are well matched by a two-temperature
v thermalmodelwithderivedtemperaturesoftypically∼0.2keVand∼0.65keV,inlinewith
3
publishedresultsforothernormalandstar-forminggalaxies.Thehotcomponentcontributesa
6
higherfractionofthesoftluminosityinthegalaxieswithhighestX-ray/SFRratio,suggesting
2
alinkbetweenplasmatemperatureandX-rayproductionefficiency.Thephysicalproperties
4
of the gas present in the galactic disks are consistent with a clumpy thin-disk distribution,
.
1 presumablycomposedofdiffusestructuressuchassuperbubblestogetherwiththeintegrated
0
emissionofunresolveddiscretesourcesincludingyoungsupernovaremnants.
9
0 Keywords:
: Galaxies:ISM–galaxies:spiral–X-rays:galaxies
v
i
X
r
a
1 INTRODUCTION ralgalaxiesisbelievedtoariseingasheatedbyshocksgenerated
as a result of supernova explosions and stellar wind interactions
ThefirstdetailedstudiesoftheX-rayemissionfromnearby late- (Chevalier&Clegg 1985; Stricklandetal. 2000; Suchkovetal.
type galaxies were carried out with the Einstein observatory and 1994;Grimesetal.2005).ThisissupportedbyanalysisoftheX-
revealed both individual luminous X-ray sources, as well as an rayspectraofindividualshell-typesupernovaremnants(SNRs)of
extended, often complex, underlying structure (Fabbiano 1989). reasonableage,whichareoftencharacterisedbyplasmatempera-
In a number of near edge-on star-forming systems. such as M82 turesintherange0.2-0.8keV,similartothevaluesobtainedwhen
(Watsonetal. 1984), evidence was also found for hot outflowing studying larger-scale structures in galactic disks. Stricklandetal.
winds.ThesuperiorspatialresolutionandenhancedsoftX-rayre- (2004a) showed that the luminosity of the diffuse X-ray emis-
sponse of ROSAT greatly extended this work, leading to the de- sion observed in star-forming galaxies is proportional to the rate
tection of substantial numbers of discrete sources as well as es- of mechanical energy injection into the ISM from young stars.
tablishing a much clearer picture of the diffuse emission compo- Mas-Hesse,Oti-Floranes&Cervino (2008) further found that the
nents (e.g. Readetal. 1997). More recently the advent of XMM- ratioofthesoftX-rayluminositytofarinfraredluminosity,taking
NewtonandChandra hasextendedour knowledgeofthediscrete thelatterasaproxy forthestar-formationrate(SFR),isstrongly
X-ray source populations and the ultrahot ISM in nearby galax- dependentontheefficiencywithwhichmechanicalenergyiscon-
ies to unprecedented luminosity and surface brightness thresh- vertedintosoftX-raysandalsotheevolutionarystatusofthestar
olds(e.g.Stricklandetal.2004a;Colbertetal.2004;Pietschetal. formationepisode.Thisallarguesforacloseconnectionbetween
2004;Fabbiano2006;Yangetal.2007;Bogda´n&Gilfanov2008; softX-rayemissionandstarformationinlate-typegalaxies.
Baueretal.2008).
A recent study of a sample of 12 nearby spiral galax-
Thediffuse X-rayemission associated withthedisksof spi- ies using Chandra (Tyleretal. 2004) has shown strong cor-
2 R.A.Owen &R.S. Warwick
Table1.Propertiesofthesampleofgalaxies.
Galaxy HubbleTypea Distanceb D25a Inclinationb ForegroundNH c Nucleus
(Mpc) (′) (◦) (1020cm−2)
NGC300 SA(s)d 2.0 21.8 30 3.6 -
M74 SA(s)c 11.0 10.5 7 4.8 -
NGC3184 SAB(rs)cd 11.6 7.2 <24 1.0 -
M51 SA(s)bc 8.4 7.3 20 1.8 Sy2
M83 SAB(s)c 4.5 12.9 24 3.9 Starburst
M101 SAB(rs)cd 7.2 26.9 18 1.1 -
a-Hubbletypeandthemajor-axisdiameter(D25)fromtheRC3catalogue(deVaucouleursetal.1991).
b-See§4.2forreferences.
c-N valuesbasedonDickey&Lockman(1990).
H
Table2.DetailsoftheXMM-Newtonobservationsofthefivenewgalaxiesinoursample.
Galaxy ObservationID StartDate Filtera Targetco-ordinatesb Usefulexposure(ks)
(yyyy-mm-dd) pn/MOS1/MOS2 RA(J2000) Dec(J2000) pn MOS1+2
NGC300 0112800201 2000-12-26 M/M/M 00h54m55.0s −37◦41′00′′ 29.4 65.1
0112800101 2001-01-01 M/M/M 40.4 88.4
0305860401 2005-05-22 M/M/M 27.2 64.9
0305860301 2005-11-25 M/M/M 30.4 68.2
M74 0154350101 2002-02-02 T/T/T 01h36m23.9s +15◦45′13′′ 23.3 49.9
0154350201 2003-01-07 T/T/T 23.5 50.2
NGC3184 0028740301 2001-11-02 T/T/T 10h18m40.0s +41◦25′11′′ 19.9 49.9
M51 0112840201 2003-01-15 T/T/T 13h29m51.9s +47◦10′32′′ 18.8 40.6
0212480801 2005-07-01 M/M/M 20.4 56.7
0303420101 2006-05-20 T/T/T 32.3 77.3
0303420201 2006-05-24 T/T/T 14.1 43.0
M83 0110910201 2003-02-18 T/M/M 13h37m00.4s −29◦52′04′′ 18.3 38.2
a-T=thinfilter,M=mediumfilter.
b-Assumedpositionofthegalacticnucleus.
relation between soft diffuse X-ray emission and sites of re- sourcesandthediffusegalacticemission(§4).Onthebasisofpub-
cent star formation in spiral arms traced by mid-infrared and lished source luminosity functions, we estimate the likely contri-
Hαemission. A similarinvestigation hasnot been completed us- butionof trulydiffuseemission tothisresidual signal. Wego on
ing XMM-Newton, but a number of studies of individual spiral toinvestigatespectralpropertiesofboththebrightsourceandthe
galaxies have been reported (Pietschetal. 2001; Takahashietal. unresolved(residual)components(§5).Wealsoconsiderthemor-
2004;Trudolyubovetal.2005;Warwicketal.2007),whichagain phologyoftheresidualemissioninthecontextofthestarformation
demonstrate the strong linkage between tracers of star formation evidentinthegalacticdisks(§6).Finallywediscusstheimplication
anddiffuseX-rayemission. Furthermore, adual-component ther- ofourresults(§7)andbrieflysummariseourconclusions(§8).
malmodel,withcharacteristictemperaturesof∼0.2keVand0.6–
0.7 keV, very often provides a good description of the spectrum
of the diffuse X-ray emission. Studies of the point source X-ray
2 GALAXYSAMPLEANDOBSERVATIONS
luminosityfunction(XLF)inspiralgalaxies(Tennantetal.2001;
Soria&Wu2003;Colbertetal.2004)haveshownthatthebright-
In this paper we investigate the large-scale spatial and spectral
est point sources contain the bulk of the integrated point source
properties of the X-ray emission emanating from the extended
luminosity, suggesting that, provided a sufficient number of the disks of six nearby late-type spiral galaxies. The galaxies in
brightestpointsourcesareexcluded,itshouldinprinciplebepossi-
question are NGC300, M74, NGC3814, M51, M83 and M101,
bletoprobetheunderlyingstructureofthegalaxytorelativelylow
brief details of which are given in Table 1. In a previous paper
levelsofsurfacebrightness. (Warwicketal.2007;hereafterW07),wereportedtheresultsofan
In this paper, we focus on the spectral and morphological XMM-Newton study of one of these galaxies, namely M101, and
properties of the diffuse X-ray emission emanating from the ex- hereweapplyasimilarmethodologytoextendthesample,thereby
tendeddisksoflate-typenormalgalaxies,asdeducedfromXMM- allowing the intercomparison of the galaxy X-ray characteristics
Newtonobservations.In§2,weidentifyasampleofnearbygalax- basedonXMM-Newtonobservations.
ies with close to face-on disks of sufficient angular extent so as WhendealingwithX-rayspectralimagingdataofonlymod-
tobereadilyamenableforstudywithXMM-Newton.Wethenout- erate spatial resolution (as afforded by XMM-Newton), the pres-
linethedatareductionmethodsemployed(§3).Usingimageanal- ence of a number of very bright point sources has the potential
ysis,wesub-dividethetotalX-rayluminosityofeachgalaxyinto toobscurethepropertiesoftheunderlyinggalaxy.Sinceourgoal
thecontributionofspatiallyresolvedbrightsourcesandaresidual is to study the latter rather than the former, we have limited our
componentcomprisingtheintegratedemissionofunresolvedpoint galaxy sample to nearby face-on systems of Hubble type Sbc-Sd
X-rayemissionfromspiralgalaxies 3
with major-axisD25 extent greater than 7′, thereby ameliorating 3.2 Spatialmaskingofbrightsources
the source confusion problem. (Wehave excluded M33 from our
study since this galaxy has a D25 diameter of 70′and thus ex- Inordertoproduceaspatialmasktosuppressthebrightsourcecon-
taminationitwasnecessary toconstruct acatalogue ofthebright
tends well beyond the EPIC field of view; M33 is the subject of
sourcesassociatedwitheachgalaxy.Thestartingpointwasapre-
a specific XMM-Newton programme utilising multiple pointings
liminarysource-listforeachgalaxyderivedfromthesecondXMM-
(Pietschetal.2004;Misanovicetal.2006).
Newton Serendipitous Source Catalogue (2XMM; Watsonetal.
DetailsoftheXMM-NewtonEPICobservationsofthefivenew
2008), encompassing those sources located withinthe galaxy ex-
additionstoourgalaxysamplearesummarisedinTable2.
traction region (as defined above). Next we determined the net
count-rate of each source in the broad-band pn+MOS image1.
Thosesourcesforwhichthemeasuredcountratesexceededade-
finedcount-rate threshold (specifictoeach galaxy) werethen in-
3 DATAREDUCTION
cludedinthefinal“master”source-listforthegalaxy.Inthecaseof
ThedatareductionwasbasedonSASv8.0.Thedatasetswereini- bothM83andNGC300twoadditionalsources,notincludedinthe
tially screened for periods of high background by accumulating 2XMMcatalogue,butclearlypresentintheimagesatcount-rates
full-field10–15keVlightcurves.MOSdatawereexcludedduring abovethethresholdwereadded byhand. Similarlyinthecaseof
periods when the10-15 keV count rateaveraged over 100 s bins M51oneadditionalsourcewasadded.Thenumberofsourcesin-
exceeded0.2cts−1,andthepndataweresimilarlyexcisedwhen cludedineachmastersource-listisgiveninTable3,togetherwith
thepncountrateexceeded2cts−1.AssummarisedinTable2,the thethreshold setting(quotedastheequivalent unabsorbed source
exposuretimesaftereventfilteringwithinindividual observations luminosity in the 0.3–6keV band on the basis of a source spec-
rangedfrom14to40ksinthepncamera(withtypicallyslightly trumconsistingofpower-lawcontinuumwithΓ=1.7absorbedby
greaterexposureintheMOScameras). theGalacticforegroundN ).Atthisstagewealsodeterminedthe
H
In W07, we discussed how the use of an appropriate spatial countratesforeachsourceinthesoft,mediumandhardbandsby
mask greatly reduced the “contamination” arising from the pres- applying theabove processtotheappropriate pn+MOS sub-band
enceofrelativelysmallnumbersofluminouspointsourcesinthe images.
diskofM101.Hereweemployarefinementofthismaskingtech- Thenextstepinvolvedthecreationofthespatialmask.Tothis
niqueinouranalysisoftheXMM-Newtonobservationsoffivefur- end, weproduced asimulatedimageof thepoint sourcespresent
thertargetgalaxies. inthebroad-bandpn+MOSobservation.TheMOS-andpn-camera
PSFsweremodelled foreachenergybandasasetofsub-images
representingthePSFoutto1′fromthesourceposition.Asimulated
3.1 ImageConstruction sub-imageofeachsource(inthreebandsandtwoinstruments)was
thenproducedbyscalingtheappropriatePSFimagesoastomatch
Thestartingpointofouranalysis,foreachgalaxyandforeachin-
theactualimagedata(intermsofthecountratecontainedwithin
dividual observation, was thecreation of pn, MOS 1and MOS 2
thenominalsourcecell).Theresultantsetofsub-imageswerethen
imagesand associated exposure maps inthreeenergy bands: soft
co-addedwithsuitablespatialoffsettingsoastosimulatethesur-
(0.3–1keV), medium (1–2keV) and hard (2–6keV). In all cases,
facebrightnessdistributioninthebroad-bandpn+MOSimagede-
thepixelsizewassetat4′′×4′′.Forboththeimageanalysisand rivingfrombrightpointsourceswithinthegalaxyfield2.Aspatial
thesubsequentspectralanalysis,weutilisedsingleanddoublepixel
maskwasthenderivedfromthisimagebyapplyinganappropriate
eventsinthepncamera(pattern0–4)andsingletoquadrupleevents
surfacebrightnesscut.
(pattern0–12)intheMOScameras.Theresultingimageswerethen
Figure3 shows thesimulated “bright source image” derived
flat-fielded by subtraction of a constant background particle rate
using the above procedure for M83. In this example, the bright
(estimatedfromcountratesrecordedinthecornersoftheCCDde-
sourcemaskwasobtainedbyapplyingasurfacebrightnesscutto
tectorsnotexposedtothesky)anddivisionbytherelevantexposure the image at a level of 0.07 pn+MOS1+MOS2 ct ks−1 pixel−1,
map.Atthesametime,spuriousdatafrombadpixels,hotcolumns
leading to a “spillover” fraction, (i.e., the fraction of the bright
andchipgapswereexcluded.Wheretherewasmorethanoneob-
sourcesignalnotcontainedwithinthemaskedregion)of9%.Simi-
servation, the images pertaining to a given instrument and band
larsurfacebrightnesscutswereappliedtotheothergalaxiesresult-
wereatthispointcombinedintoasinglemosaicedimage.Foreach
inginspill-overfractionsintherange9-12%.
band, the resultant flat-fielded MOS 1 and MOS 2 images were
The derived spatial masks were used to separate the “bright
combinedtogivea“MOS”imageandfinallya“pn+MOS”image
sourceregion”fromthe“residualemissionregion”inboththespa-
wasalsoproducedbysummingtheindividualpnandmosimages
tialandspectralanalysis.Inordertostudythespatialdistributionof
(withappropriate scaling inregions affected by chip gaps, where
theresidualemission(see§6),asimulatedimagerepresentativeof
therewasincompletecoverageinoneoftheinstrumentchannels).
Figures 1 & 2 show the soft-band images derived for each
ofthefivenew galaxies,together withthecorresponding datafor
M101fromW07.InthecaseofM74,NGC3184,M51andM83our 1 Thepn+MOS1+MOS2count-rateabovethelocalbackgroundwasdeter-
analysispertainstocircularregionscentredonthegalacticnucleus minedwithinacellof16′′radiuscentredonthesourceposition.Acorrec-
of diameter equal to the major-axis D25 dimension, whereas for tionfactorwasthenappliedtoallowforthesignalcontainedinthewingsof
thepointspreadfunction(PSF)extendingbeyondthenominalsourcecell.
NGC300,giventherelativelylowX-raysurfacebrightness,were-
2 Asafinaliteration,countrateswereextractedfromthesimulatedimages
strictourattentiontoacentral10′diameterregion(0.46oftheD25 withina16′′ cellcentredoneachsourcepositionandcomparedtothose
extent).SimilarlyforM101theX-rayanalysisreportedinW07ap-
obtainedfromtheactualdata,withanydifferencesattributabletothePSF
pliestoacentral20′diameterregion(0.74oftheD25 dimension). spreadingofsignalfromadjacentsources.Basedonthisanalysis,thenomi-
Hereafter werefer tothese galaxy-centred circular regions asthe nalsourcecountrateswereadjustedtocompensateforthissourceconfusion
“X-rayextractionregions”. effectandanewversionofthesimulatedimageswasproduced.
4 R.A.Owen &R.S. Warwick
Figure 1. Comparison of the soft X-ray and far-ultraviolet images of three galaxies. Left-hand panels: Adaptively smoothed versions of the XMM-
Newton pn+MOS images inthe soft(0.3-1 keV) band. The amplitude scaling in all cases is logarithmic with the contour levels representing factor two
stepsinthesoftX-raysurfacebrightness.Right-handpanels:TheGALEXFUV(λ =2267A˚)imageonthesamespatialscaleastheX-raydata(except
eff
forNGC3184wheretheUVdataarefromtheXMM-NewtonOpticalMonitorintheUVW1(≈ 2680A˚)band.Theamplitudescalingisagainlogarithmic.
ThecontoursshowthesoftX-raymorphologyforcomparisonpurposes.Toppanels:NGC300.Thecirclehasadiameterof10′.Middlepanels:M74.The
circlehasadiameterof10.5′.Bottompanels:NGC3184.Thecirclehasadiameterof7.2′.
X-rayemissionfromspiralgalaxies 5
Figure2.AsforFig.1.Toppanels:M51.Thecirclehasadiameterof7.3′.Middlepanels:M83.Thecirclehasadiameterof12.9′.Bottompanels:M101.
Thecirclehasadiameterof20′.
6 R.A.Owen &R.S. Warwick
Figure3. Leftpanel:Themodelled“brightsource”imageconstructedforM83fromasetofcount-ratescaledPSFsub-images.Asourcemaskwasproduced
bythresholdingthisimageatasurfacebrightnesslevelsuchthat91%ofthesourcesignalwascontainedwithintheresultingmask.Rightpanel:Thesoft-band
imageoftheresidualemissioninM83obtainedfromthefullimagebysubtractingthebrightsourcemodelandthenapplyingthespatialmask.Muchofthe
“contamination”duetobrightsourcesissuppressedbythisprocess.Inbothpanelsthecirclehasadiameterof12.9′.
Table3.CharacteristicsofthebrightX-raysourcepopulationineachgalaxy.
Galaxy X-rayRegiona ThresholdLXb Numberin NumberofhighLXsources
(′) (1037ergs−1) SourceList (LX>5×1038ergs−1)
NGC300 10.0 0.2 22 0
M74 10.5 5.0 20 4
NGC3184 7.2 6.0 18 4
M51 7.3 4.0 26 6
M83 12.9 2.0 40 2
a-Diameterofthe“X-rayextractionregion”.
b-NominalLXthresholdappliedinthe0.3–6keVbandindefiningsample.
thesoftbanddatawassubtractedfromthecorrespondingpn+MOS from ideal, because the sky X-ray background in the soft band
image and thesource mask imposed - see Figure3.Thisprocess variesfrom fieldtofield, and thisis avital component of theto-
servestotosuppressmuchofthe“contamination”arisingfromthe tal background signal we wish to subtract. Instead, we extracted
brightestsourcesinthegalaxy,includingthebulkofthespillover spectra from both an annulus surrounding the defined galaxy re-
into the residual galaxy region. Note that a slightly different ap- gionandalsofromthecornerregionsofeachdetectornotexposed
proachisnecessaryinthespectralanalysis;herewetakeaccount to the sky. By using an appropriate scaling of these spectra, us-
ofthebrightsourcesignalspillingintotheresidualemissionregion ingthemethoddetailedinW07,anappropriatemodelbackground
by including an appropriate fraction of the bright source spectral spectrumcanthenbeproduced.TheSAStoolsarfgenandrmfgen
modelinthespectralfitoftheresidualemission-see§5.2. wereusedtoproduceappropriateAuxiliaryResponseFile(ARF)
andResponseMatrixFile(RMF)filesforthesourceandresidual
galaxyregions.Finally,thecountsrecordedinadjacent(raw)spec-
tral channels were summed to give a minimum of 20 counts per
3.3 SpectralExtraction
spectralbininthefinalsetofspectra.
The soft-band images shown inFig. 1 &2 demonstrate theexis-
tenceof anextended emission component inaddition tothepop-
ulation of bright point sources. On the basis of the approach de-
scribed earlier, we extracted the integrated spectrum of both the 4 OVERVIEWOFTHEGALACTICX-RAYPROPERTIES
brightsourceregion(boundedbythespatialmask)andtheresidual
4.1 Thecontributionofluminouspointsources
emissionregion(correspondingtothefullX-rayextractionregion
lessthemaskedarea). Byapplicationofthespatialmaskdescribedearlier,weareableto
The relatively large extent and complicated shapes of the measure the count rates (net of background) in the soft, medium
bright source and residual emission regions makes the evaluation andhardbandimagesassociatedwithboththe“bright sourcere-
of the background subtraction somewhat more complicated than gions” and the “residual emission regions” (hereafter we refer to
with most XMM-Newton data. In this context, the use of “blank- thelatterasthe“residualgalaxy”).Table4summarisestheresults
sky”XMMfieldstoproduceabackground spectrumwouldbefar foreachofthefivenewgalaxiesinoursample.Herewehavecon-
X-rayemissionfromspiralgalaxies 7
Table4.ContributionofpointsourcestothetotalX-rayluminosity.
Galaxy Spillover/Area Component LX(1039ergs−1)
Factors(%) (0.3-1keV) (1-2keV) (2-6keV) (0.3-6keV)
NGC300 12/7 BrightSources 0.19 0.07 0.25 0.51
UnresolvedSources [0.02] [0.01] [0.03]
ResidualGalaxy 0.05 - - 0.05
TotalMeasured 0.24 0.07 0.25 0.56
M74 11/24 BrightSources 1.4 1.3 3.0 5.7
UnresolvedSources [0.5] [0.4] [0.8]
ResidualGalaxy 1.4 0.7 - 2.1
TotalMeasured 2.8 2.0 3.0 7.8
NGC3184 9/22 BrightSources 1.2 1.4 3.4 6.0
UnresolvedSources [0.6] [0.5] [0.9]
ResidualGalaxy 2.0 0.5 - 2.5
TotalMeasured 3.2 1.9 3.4 8.5
M51 12/24 BrightSources 6.8 3.8 6.5 17.1
UnresolvedSources [0.5] [0.4] [0.8]
ResidualGalaxy 6.3 0.9 - 7.2
TotalMeasured 13.1 4.7 6.5 22.3
M83 9/30 BrightSources 2.5 2.0 3.1 7.6
UnresolvedSources [0.5] [0.4] [0.9]
ResidualGalaxy 3.3 0.7 - 4.0
TotalMeasured 5.8 2.7 3.1 11.6
vertedthemeasuredcountratestoequivalentfluxesinthebandon duetothelowSFRdensityinthisGalaxyratherthanbeingsimply
thebasisofaspecificspectralmodelandthentransposedfromflux aconsequence ofamorestringentsourcerejectionthreshold. For
toluminosityusingthedistancesquotedinTable1.Forthebright M74,NGC3184,M51andM83abroadlysimilarsourcedetection
sourceregionsweassumeaspectralmodelconsistingofapower- thresholdtransposestoluminositythresholdsforsourceexclusion
lawcontinuumwithphotonindexΓ = 1.7modifiedbythefore- rangingfrom2.0×1037 ergs−1(forM83)upto6.0×1037 ergs−1
ground absorption in our Galaxy. For the residual galaxy we use (forNGC3184).Thelikelycontributionofunresolvedsourcesto
the model which best fitsthe actual spectrum of this component, theresidualgalaxyemissionineachcaseisinvestigatedbelowand
asderivedin§5.2.ThequotedX-rayluminosities(LX)havebeen acorrectiontoafixedluminositythresholdisappliedlater inthe
corrected for foreground absorption, that is they are unabsorbed analysis(see§7).
values.
TheLXvaluesreportedinTable4havebeencorrectedfortwo Asafurtherexerciseitispossibletoestimatethelikelycon-
furthereffects.Thefirstisthespilloverofbrightsourcefluxintothe
tribution to the residual galaxy emission of an unresolved popu-
residual galaxy region (discussed earlier). The second correction
lation of somewhat less luminous sources with spectral charac-
relatestothefact thatthebrightsource maskobscuresasizeable
teristicsand distribution similar to the sources represented in the
fractionofthegalaxydisk.Tocorrectforthis,wehavescaledupthe
brightsourcelist.HerewemakeuseofrecentresultsfromChan-
residual galaxycomponent byafactor derived frominterpolating
dra, which show that the discrete source luminosity function ap-
theazimuthally-averagedbrightnessdistribution(asrepresentedby
propriate to the disk regions of spiral galaxies typical takes the
the radial distributions of the X-ray surface brightness discussed
form of a power-law with a slope in the range −0.5 to −0.8 (in
in§6)intothemaskedregions.Thespilloverfractionandthearea
the cumulative form). For example, Tennantetal. (2001) quote a
correctionfactorapplicabletoeachgalaxyarelistedinTable4.
slope of −0.5 for the disk sources in M81, whereas Doaneetal.
The results in Table 4 demonstrate that whereas the set of (2004) measure −0.6 for NGC 3184 and Penceetal. (2001) re-
very luminous sources represented by our bright source sample port−0.8forM101.Soria&Wu(2003)examinethelogN–log
provide the dominant contribution to the total galactic X-ray lu- SrelationforrelativelyfaintHMXBsinthediskofM83andfind
minosityabove2keV,atlowerenergies,andparticularlybelow1 a broken power-law provides a reasonable fit to the data, with a
keV, thesplit between thebright source and residual galaxy con- low-luminosityindexof−0.6.AlsoColbertetal.(2004)measure
tributionsismorebalanced,albeitwiththeexceptionofNGC300. asimilarrangeofpower-lawslopesforthesourceluminosityfunc-
Forthelattergalaxy,wewereabletosetamuchlowerthreshold tionwithinarelativelylargesampleofspiralgalaxies.Theflatness
forbrightsourceremovalthanwaspossiblefortheothergalaxies ofthisformmeansthatrelativelysmallnumbersofveryluminous
inthesample(onaccountoftherelativelyproximityofNGC300 sources provide the bulk of the integrated luminosity residing in
-seeTable1).InotherwordsahigherfractionoftheintegratedX- discretesources.Furthermore,inthecontextofprovidingarough
rayfluxfrompointsourceswasresolvedinthisgalaxy.However, estimateof the contribution of unresolved sources tothe residual
weshowlater(see§6)thatthelowsurfacebrightnessinferredfor galaxysignal,itimpliesthatwecan,inprinciple,extrapolatebelow
theunderlyingdiffuseX-rayemissioninNGC300ismostlylikely oursourcedetectionthresholdtoarbitrarilyfaintlevels(assuming
8 R.A.Owen &R.S. Warwick
ofcoursethattheslopeoftheluminosityfunctionremainsconstant (Freedmanetal.2001).ManylargeHIIregionsarevisiblethrough-
overanappropriatelywiderangeofsourceluminosity). out the galaxy, and there is evidence that the galaxy has experi-
Herewemaketheassumptionthattheslopeofthesourcelu- enced many episodes of star formation. However, star formation
minosityfunctionis−0.6acrossthesetofgalaxiesandoverawide in the central regions appears to be suppressed in recent times,
luminosityrange.Wethenusethenumberofbrightsourcesactu- with a small population of stars present in the central arcminute
allyobservedinthegalaxyfromthethresholdluminosityuptoa ofthegalaxywithagesbelow1Gyr(Davidge1998).Thereisclear
fixeduppercut-off(sethereasLX=5×1038ergs−1-seeTable evidence of spiral structure, although thecontrast between “arm”
3), to define the normalization of the source luminosity function and“off-arm”regionsinthegalaxyislow.InX-raysNGC300has
for that particular galaxy. We first calculate the integrated emis- beenstudiedbyROSAT(Read&Pietsch2001)andmorerecently
sionofunresolvedsourcesinthebroad0.3-6keVbandandthen, by XMM-Newton (Carpanoetal. 2005). Carpanoetal. (2005) re-
onthebasisofthebrightsourcespectralmodelnotedearlier,split portthepresenceofdiffuseemissioninthegalacticdiskofNGC
this intosoft-, medium- and hard-band contributions. The entries 300withLX<1038ergs−1andtemperaturesof≈0.2and≈0.8
inTable4undertheheading“unresolvedsources”summarisesthe keV. In the current paper we determine the aggregate luminosity
resultsonagalaxybygalaxybasis.Inthesoftband,whichisthe ofthe20brightestpointsourceswithinthecentral5′ ofNGC300
focusofourattentionindescribingtheX-raymorphologyandthe to be LX≈ 5×1038ergs−1, in agreement with earlier studies.
linkage of X-ray emission to star formation, the potential contri- We estimate the X-ray luminosity of the residual emission to be
bution of unresolved, but relatively luminous, source populations ≈ 5×1037ergs−1 (0.3–1 keV). The surface brightness of this
to the measured residual galaxy emission ranges from as low as componentisextremelylowbutappearstotraceacentralbar-like
∼10%inM51upto∼40%inNGC300.Byimplicationthebulk structureandtoextendouttoatleastadistanceof2.5kpc(4′).
oftheresidualgalaxyluminosityinthisbandmustresideinsome
combination of: (i) truly diffuse X-ray emission associated with
4.2.2 M74
thegalaxydiskpresumablyresultingfromtheenergyinputofsu-
pernovaandstellarwindsand(ii)theintegratedemissionofpop- M74 (NGC 628) is an SA(s)c spiral seen almost face-on (i =
ulationsofrelativelysoftspectrum,lower-luminositysourcessuch 7◦; Shostak&vanderKruit 1984) at a distance of 11 Mpc
assupernova remnants, cataclysmicvariables, activebinaries and (GildePazetal.2007).Thenuclear regionof thegalaxyischar-
stellarcoronalsources,populationswhicharenotwellrepresented acterised by strong HII emission with no evidence for the pres-
inthebright-source(i.e.,highLX)samples.Thereisalsoapoten- ence of an AGN. A grand-design twin-armed spiral structure
tial contribution from a putative extended galactic halo, which is is observed. In X-rays M74 has been studied by both Chan-
oftenseeninedge-ongalaxies;however theface-onaspectofthe dra (Kraussetal. 2005) and XMM-Newton (Soria&Kong 2002;
presentgalaxiessuggeststhataveryextendedhalocomponentwill Soria,Pian&Mazzali2004),albeitwithlittleattention,todate,on
bedifficulttodistinguishfromthecomponentsconfinedtothedisk. thediffuseX-raycomponent.Inthepresentanalysisweconfirmthe
presenceofanextensivediscretesourcepopulationinM74withan
aggregateluminosityinthe20brightestpointsourcesamountingto
4.2 X-raycharacteristicsoftheindividualgalaxies 5.7×1039ergs−1.TheunderlyingresidualX-rayemissionhasa
Figures1&2comparethegalaxymorphologyobservedinsoftX- luminosityof2.1×1039ergs−1.Thiscomponenttracestheinner
rayswiththatmeasuredinthefarultravioletFUV(λ ≈ 2267 spiralarmstructureandisevidentouttoadistanceof9kpc(3′),
eff
A˚)bandbyGALEX3,althoughinthecaseofNGC3184weusethe atwhichpointitfallsbelowthebackgroundlevel.
XMM-NewtonOpticalMonitorUVW1(≈2680A˚)imagerecorded
inthesameobservationastheX-raydata(seeTable2)tomakethis
4.2.3 NGC3184
comparison.InallcasesthetheUVimageswerelightlysmoothed
with a circular gaussian mask with σ = 4′′. In a number of the NGC3184 is an SAB(rs)cd barred spiral seen at low inclination
galaxies, particularly M74, M83 and M101, the X-ray emission (i < 24◦; Lyon-Meudon Extragalactic Database [LEDA]4) at a
traces segments of the inner spiral arms which are delineated as distanceof11.6Mpc(Leonardetal.2002).Thenuclearregionof
verypronouncedfeaturesintheFUVimages.Inallcasesthecen- the galaxy isdominated by HII emission with no compelling ev-
tralconcentrationofthesoftX-rayemissionanditsfall-offwithin- idence for an AGN (Doaneetal. 2004). Grand-design twin spi-
creasinggalactocentricradiusisquitewellmatchedtothesurface ral arms are observed, although the contrast in brightness be-
brightnessdistributionmeasuredintheFUVchannel.Thisgeneral tween“arm”and“off-arm”regionsinthegalaxyislow.InX-rays
correlationofsoftX-rayandFUVlightimmediatelyestablishesthe NGC3184hasbeenstudiedbyChandrawithanemphasisonboth
closeconnectionofasignificantcomponentoftheX-rayemission thepointsourcepopulation(Colbertetal.2004)andthediffuseX-
withrecent star-formationinthese late-type systems. Weexplore rayemittinggas(Doaneetal.2004;Tyleretal.2004).Thediffuse
thelinkagebetweentheX-rayemissionandstar-formationinmore X-rayemissionshowsastrongcorrelationwithregionswhichare
detailin§6.Herewebrieflycommentoneachofthesegalaxies,in brightinHα(Tyleretal.2004).Doaneetal.(2004)reportthepres-
thecontextoftheirextendedX-rayemittingcomponents. ence of diffuse X-ray emission concentrated in areas of younger
stellarpopulationsandstar-formingregions,withasurfacebright-
ness five times greater in spiral arm regions than in off-arm re-
4.2.1 NGC300
gions. The spectrum of the diffuse emission in the galactic disk
NGC300 is an SA(s)d spiral seen at low inclination (i = canbewellfittedwithatwo-temperaturethermalmodelwithkTof
30◦; deVaucouleursetal. 1991) at a distance of 2.0 Mpc 0.13keVand0.43keV(Doaneetal.2004).Inthepresentstudywe
measuretheaggregateluminosityofthe18brightestpointsources
3 The GALEX images were obtained from the public archive at
http://galex.stsci.edu/GR2/ 4 athttp://leda.univ-lyon1.fr
X-rayemissionfromspiralgalaxies 9
tobe≈6×1039ergs−1,consistentwithearlierChandraestimates fromthebrightnuclearregionoutalongthetwinspiralarmstoa
(Colbertetal.2004).TheresidualX-rayemissionhasanX-raylu- galactocentricdistanceofabout4kpc(3′).ExtendedX-rayemis-
minosityof2.5×1039ergs−1 andextendsfromthenucleusout sionofrelativelylowsurfacebrightnesscanthenbefurthertraced
toadistanceof8kpc(2.5′). outto∼8kpc.TheX-rayluminosityoftheresidualcomponentis
4×1039ergs−1.
4.2.4 M51
M51 (NGC5194) is an SA(s)bc spiral seen almost face- 5 SPECTRALANALYSIS
on (i = 20◦; Tully 1974) at a distance of 8.4 Mpc
(Feldmeier,Ciardullo&Jacoby 1997). It hosts a low-luminosity Usingthemethodologyoutlinedin§3.3,spectraldatasetswereob-
activegalacticnucleus,whichshowsopticalemissionlinesandis tainedforboththebright-sourceandtheresidual-galaxyregionsfor
classifiedasaSeyfert2(Stauffer1982;Ho,Filippenko&Sargent fourgalaxies,namelyM74,NGC3184,M51andM83.Forsimplic-
1997). M51 is interacting tidally with a companion SB0/a ity,webasedouranalysisonasingleobservationforeachgalaxy,
galaxy (NGC5195), and both systems are seen to con- namelytheobservationwiththedeepestpnexposure(seeTable2).
tain starburst activity. There is grand-design twin-arm spi- Duetolimitedsignal-to-noiseintheMOSchannels, wefocuson
ral structure present. In X-rays M51 has been studied by the spectra derived from the EPIC pn camera, except in the case
Einstein (Palumboetal. 1985), ROSAT (Marstonetal. 1995; ofM83,wherebothpnandMOSspectrawerefullyutilisedandin
Ehleetal. 1995), ASCA (Terashimaetal. 1998) and more re- M74wherebothdatasetswereemployedinthespectralinvestiga-
centlybyChandra(Terashima&Wilson2004;Colbertetal.2004; tionofitsbright-sourceregion.Unfortunatelytheresidual-galaxy
Tyleretal.2004)andXMM-Newton(Dewanganetal.2005).Stud- emissioninNGC300wastoofaintfordetailedspectralanalysisto
ies have also been carried out at higher X-ray energies by Bep- bemerited.Thespectralfittingwascarriedoutusingthesoftware
poSAX(Fukazawaetal.2001).ThediffuseX-rayemissioniswell packageXSPECVersion12.4.
correlatedwiththeradiocontinuumandmid-infraredemissiondis-
tributed in the spiral arms establishing a clear link with star for-
5.1 Spectraofthebright-sourceregions
mation (Ehleetal. 1995; Tyleretal. 2004). Terashima&Wilson
(2004) find that the diffuse X-ray emission in the southern ex- The spectra representative of the bright-source regions in M74,
tranuclear cloud is well fitted with a single-temperature thermal NGC3184, M51 and M83 were well fitted with either a simple
model with kT of 0.58 keV. The current analysis confirms the power-lawcontinuumorapower-lawcontinuumplusanadditional
general properties of the galaxy established by the ROSAT and solar-abundancethermalplasmacomponent(theMEKALmodelin
Chandra observations. The aggregate luminosity of the brightest XSPEC).Thebest-fitphotonindiceswereallintherange1.55–1.9
25pointsourcesinM51is1.7×1040ergs−1 (0.3–6keV),28% withthetemperatureofthethermalemissiontypically≈0.65keV.
ofwhichisattributabletothenuclearsource.Ouranalysisreveals Fig.4showsthemeasuredspectra,thebest-fittingmodelandresid-
residual soft X-ray emission extending from the bright nucleus ualχ2 tothesebestfits.Similarly,Table5summarisesthedetails
alongthespiralarmsouttoadistanceof5kpc(2.5′).Lowersur- ofthebest-fittingmodels.Aftercorrectingforthefluxspilloverbe-
facebrightnessemissionfromtheextendeddiskisobservedoutto yond the spatial mask, the inferred luminosities in the integrated
aradialdistanceof9kpcinalldirections,withfurtherX-rayemis- bright-sourcespectrawerefoundtobeingeneral agreementwith
sionjoiningM51toitscompaniongalaxy.TheX-rayluminosityof theestimatesquotedpreviouslybasedontheimageanalysis.The
theresidualcomponentis7.2×1039ergs−1. photon index observed is typical of the spectra of HMXB and
LMXB found in the Local Group (Grimmetal. 2007). The ther-
malemissionassociatedwiththebright-sourceregionsisofsimilar
4.2.5 M83
temperaturetothe“hot”componentobservedintheresidual-galaxy
M83 (NGC 5236) is an SAB(s)c barred spiral seen at low in- emission(seebelow).However,wherethiscomponentisdetected
clination (i = 24◦; Talbot,Jensen&Dufour 1979) at a dis- inthebright-sourcespectra,namelyinNGC3184,M51andM83,
tance of 4.5 Mpc (Thimetal. 2003). It harbours an active nu- itappearsataproportionatelyhigherlevelthanmightbeinferred
clear region undergoing a violent starburst, a circumnuclear star- basedonasimpleareascalingwithrespecttotheresidual-galaxy
burstandgrand-design twin-armedspiralstructure.Therearenu- emission.
merous sites of active star-formation coincident with the spi-
ral arms (Vogleretal. 2005). In X-rays M83 has been stud-
ied by Einstein (Trinchierietal. 1984), ROSAT (Ehleetal. 1998; 5.2 Spectraoftheresidual-galaxyregions
Immleretal. 1999), ASCA (Okada,Mitsuda&Dotani 1997 and
Previouslyweestimatedthat9-12percentofthecountsfromthe
more recently by Chandra (Soria&Wu 2002; Soria&Wu2003;
bright sources overspill into the residual-galaxy regions. In car-
Colbertetal.2004;Tyleretal.2004).Onanarcminutescaleboth
rying out the spectral analysis of the latter we have correct for
theROSATandChandra X-rayobservations traceapparently dif-
this spillover effect, by including an appropriate fraction of the
fuseemissionassociated withtheinner spiral-arms(Immleretal.
best-fit bright-source model as a fixed component of the spectral
1999;Soria&Wu2003;Tyleretal.2004).Immleretal.(1999)re-
model. Visual inspection of the residual-galaxy spectra revealed
port the diffuse emission associated with a cluster of bright HII
almost no emission above 1.5 keV in any of the galaxies. It was
regionsembeddedinthesouth-westernspiralarmashavingatwo-
further found that thermal models comprising either one or two
temperature thermal characteristic withkT of 0.26 keV and 0.95 solar-abundance5 MEKAL components provided a reasonable fit
keV.Inthepresent analysiswemeasuretheaggregateluminosity
ofthebrightpointsourcesinM83tobeL =7.6×1039ergs−1,
X
whichincludesacontributionof3.1×1039ergs−1fromthecen- 5 Ingeneralthequalityofthespectrawereinsufficienttoallowusefulcon-
tral region. The XMM-Newton soft band image traces emission traints to beplaced onthe metal abundances inthe thermal plasmas. As
10 R.A.Owen &R.S. Warwick
Figure4.TheEPICspectraforfourgalaxies.Left-handpanels:Thespectraofthebright-sourceregions.Right-handpanels:Thespectraoftheresidual-galaxy
regions.Inallcasesthesolidlinecorrespondstothebest-fitspectralmodel(seetext).Theχ2residualswithrespecttothebest-fittingmodelarealsoshown.