A new superwind galaxy: XMM-Newton observations of NGC 6810

0.44 MB·English

by  David K. Strickland

Mon.Not.R.Astron.Soc.000,1–12(2006) Printed5February2008 (MNLATEXstylefilev2.2)
A new superwind galaxy: XMM-Newton observations of
NGC 6810
David K. Strickland⋆
Department of Physics & Astronomy, The Johns Hopkins University, 3400 North Charles Street, Baltimore, MD 21218, U.S.A.
7
0
0
Accepted ...,Received...;inoriginal...
2

n
a ABSTRACT
J

2 We present the first imaging X-ray observation of the highly inclined (i = 78◦)
2 SabSeyfert2galaxyNGC6810usingXMM-Newton,whichrevealssoftX-rayemission

that extends out to a projected height of ∼7 kpc away from the plane of the galaxy.
1
The softX-rayemissionbeyondthe opticaldiskofthe galaxyis mostplausibly extra-
v
planar, although it could instead come from large galactic radius. This extended X-
0
3 ray emission is spatially associated with diffuse Hα emission, in particular with a
6 prominent 5-kpc-longHα filament on the north-westof the disk. A fraction.35%of
1 the total soft X-ray emission of the galaxy arises from projected heights |z|> 2 kpc.
0 Within the optical disk of the galaxy the soft X-ray emission is associated with the
7 star-forming regions visible in ground-based Hα and XMM-Newton Optical Monitor
0 near-UV imaging. The temperature, super-Solar α-element-to-iron abundance ratio,
h/ softX-ray/Hαcorrelation,andX-raytofar-IRfluxratioofNGC6810areallconsistent
p with local starbursts with winds, although the large base radius of the outflow would
- make NGC 6810 one of the few “disk-wide” superwinds currently known. Hard X-
o ray emission from NGC 6810 is weak, and the total E = 2 – 10 keV luminosity and
r
t spectral shape are consistent with the expected level of X-ray binary emission from
s
the old and young stellar populations. The X-ray observations provide no evidence of
a
: any AGN activity. We find that the optical,IR and radioproperties of NGC 6810are
v all consistent with a starburst galaxy,and that the old classificationof this galaxy as
i
X a Seyfert 2 galaxy is probably incorrect.
r Key words: Galaxies: halos – Galaxies:individual: NGC 6810 – Galaxies:Seyfert –
a
Galaxies: starburst – X-rays: galaxies.
1 INTRODUCTION the minor axes of spiral galaxies, noting that the high mi-
nor axis velocity dispersion and kinematics were sugges-
NGC 6810 is a early-type spiral galaxy (morpholog-
tive of an outflow. The moderately high inclination2 of
ical type Sab(s):sp, Tully 1988) also classified as a NGC 6810, i = 78◦, is advantageous for studies of extra-
Seyfert 2 (NASA Extragalactic Database1; see also
planar emission and is similar to other galaxies with well-
Kirhakos & Steiner 1990b). Despite being relatively nearby
studied superwinds such as NGC 253, NGC 3079 and M82
(D ∼ 27 Mpc, see below) it has not been the target of (i ∼ 76,78,85◦ respectively, using averages of the values
much observational study. What little literature on NGC
given in Strickland et al. 2004).
6810 exists suggests that a powerful galactic-scale out-
Superwinds may play important roles in galaxy forma-
flow(orsuperwind,Heckman et al.1990)emanatesfrom it.
tion and evolution (see e.g. Heckman 2003; Veilleux et al.
Hameed & Devereux (1999) present a Hα image of NGC
2005, and references therein), but examples of starburst-
6810aspartofastudyofstarformation ratesinearlytype
driven winds that are nearby, bright enough and have
spirals, but do not comment on the ∼ 5 kpc-long emission
line filament rising at an angle of ∼ 45◦ out of the plane
of the galaxy that is visible in their image. Coccato et al.
(2004) observed NGC 6810 in a survey of ionized gas along 2 WederiveaninclinationforNGC6810ofi=78◦ basedonthe
opticalangularsizeparameterD25andR25valuesgiveninTully
(1988). The minor to major axis angular size ratio q =10−R25,
⋆ E-mail:[email protected] fromwhichfollowstheinclinationi=cos−1p([q2−q2]/[1−q2])
0 0
1 NED:http://nedwww.ipac.caltech.edu/ whereq0=0.2fortypeSabgalaxies(Haynes &Giovanelli1984).
2 D. K. Strickland
Figure1.SmoothedXMM-NewtonEPICPNandMOSimagesofNGC6810.Panelsathroughcareexposure-correctedandbackground-
subtracted combined MOS1+MOS2 images in the E = 0.3 – 1.1, 1.1 – 2.8 and 2.8 – 8.0 keV energy bands respectively. The dashed
ellipse represents the inclination-corrected d25 ellipse of NGC 6810 (see Tully 1988). The coordinate system shown is the offset from
the2MASSnear-IR center ofNGC 6810 atα=19:43:34.4, δ=-58:39:20.6(J2000.0). Each ofthe grey-scaleimageshas been adaptively
smoothedandisshownonasquare-rootintensityscale.Theoverlaidcontoursarefromimagessmootheduniformlywitha2-dimensional
GaussianmaskofFWHM =15′′.Contours beginatasurfacebrightness ofΣ=4.0×10−7 MOScounts s−1arcsec−2 andincreasein
factorsoftwo.Thelowestcontoursareequivalentto4.7σ,4.6σand4.3σinthethreeenergybands.Panelsdthroughfaretheequivalent
imagesfromthePN detector. Herethecontour levels beginat Σ=9.6×10−7 PNcounts s−1arcsec−2,andincreaseinfactors oftwo.
Thelowestcontoursareequivalentto3.0σ,4.7σ and4.1σ inthethreeenergybands.Extended, possiblyextra-planar,X-rayemissionis
mostclearlydetected intheE=0.3–1.1keVsoftX-rayband.
suitable inclinations for detailed study are moderately 2 OBSERVATIONS AND DATA ANALYSIS
rare3. The existence and properties of AGN-driven super-
To test these hypotheses NGC 6810 was observed with
winds are even less well constrained (Colbert et al. 1996;
XMM-Newtonfor∼49ksstartingon2004April25(Obser-
Levenson et al. 2001b; Colbert et al. 2005). Given its prox-
vation ID 0205220101).
imityandSeyfert2classification wejudgedthatNGC6810
We processed the data using version 6.5 of the XMM-
deserved investigation.
Newton SAS software, rerunning the EPIC MOS and PN
All currently known superwinds display extended soft
andOpticalMonitor(OM)analysispipelineswiththelatest
diffuse X-ray emission from gas at a temperature logT ∼
calibration data. Heasoft version 5.3.1 and Ciao 3.2 were
6.5,commonlyinclosespatialproximitytothewarmionized
alsousedduringdatareduction.Therecommendedstandard
gasatlogT ∼4(Dahlem et al.1998;Strickland et al.2004;
eventgradefiltersandprotonflarescreeningwasappliedto
Tu¨llmann et al.2006).IfthereisasuperwindinNGC6810
eachdataset.Afterfilteringtheremainingexposuretimesin
thereshouldbea5–10kpc-scalesoftX-rayhaloaroundthe
the XMM observation were 25.5 ks and 44.1 ks for PN and
galaxy. The presence of the Seyfert nucleus should also be
each MOS instrumentsrespectively.
clear in thehard X-ray energy band,either via direct AGN
Spectra and response files were created using SAS and
X-rayemissionorbyareflectioncomponentandfluorescent
the methods described in the on-line SAS Data Analysis
iron line emission.
Threads4.SpectralfittingwasperformedusingXspec(ver-
sion 11.3.1t).
Simultaneous broad-band optical imaging in V, B, U,
3 Withinadistanceof30Mpcthereareapproximately14galax- UVW1 (λ ∼ 245 – 320 nm) and UVM2 (λ ∼ 205 – 245
ieswithinclinationsi&60◦withknownorsuspectedsuperwinds,
nm)bandswasobtainedusingtheOMinstandardimaging
excluding NGC 6810. In order of increasing distance these are
NGC 253, NGC 1482, NGC 1511, NGC 1569, NGC 1808, NGC
2146,M82(NGC3034),NGC3044,NGC3079,NGC3628,NGC
4631,NGC4666,NGC4945andNGC5775. 4 Seehttp://xmm.vilspa.esa.es/sas/new/documentation/threads/
XMM-Newton observations of NGC 6810 3
mode(Mason et al.2001).Theseobservationsachievedtyp- Theprofilesdisplay thetotalsurface soft X-raybright-
icallimitingmagnitudesfora5σ detectionof19.5 (V),20.7 ness, the surface brightness of the expected background
(B), 20.1 (U), 20.1 (UVW1) and 19.4 (UVM2). Regions of (both the genuine X-ray and particle background) and the
theimages neartoNGC6810aremarred byelliptical stray background-subtractedemissionassociatedwithNGC6810.
light features caused by internal reflections within the OM A version of the SAS processing task edetect chain, al-
telescope. tered to correctly treat the presence of diffuse emission in
CopiesoftheR-bandandcontinuum-subtractedHαim- and around NGC 6810, was used to detect X-ray sources
ages ofNGC6810 presentedin Hameed & Devereux(1999) andcreatemapsofthetotalbackgroundandexposuretime
were obtained from theNASAExtragalactic Database. As- overtheentirefieldofviewofeachoftheMOS1,MOS2and
trometric solutions for these images were calculated based PN instruments. The surface brightness profiles were cre-
on the V-band OM images, which match up with Digital ating using this location-specific background and exposure
Sky Survey images to within . 1′′. Pointing offsets during time information, excluding statistically significant X-ray
theOM observations were typically also .1′′. sources (other than NGC 6810 itself) and detector-specific
NGC6810isnotbrightenoughanX-raysourcetopro- chip-gaps, bad CCD pixels and bad columns. The values of
videusefulX-rayspectrawiththegratingspectrometerson the MOS and PN backgrounds produced by this method is
XMM-Newton (the RGS). The RGS data are not discussed within ∼ 1% of the average local background values used
further. for the images displayed in Fig. 1 and the X-ray spectra
discussed in § 3.4. Each bin in the surface brightness pro-
files was initially the sum of the data in a region of width
2′.′5 parallel totheaxisof interest andtotal length 76′′ per-
3 X-RAY RESULTS pendiculartothataxis.Subsequentlyconsecutivebinswere
rebinnedwiththeaimofachievingaminimumof20counts
3.1 The spatial distribution of the X-ray emission
perfinal bin.
Images in E = 0.3 – 1.1, 1.1 – 2.8 and 2.8 – 8.0 keV en- In an effort to maximize the signal-to-noise for the
ergybandswerecreatedfromtheEPICPNandMOSdata. faintest emission the data from the PN and both MOS de-
These energy bands were chosen based on the X-ray spec- tectorswascombined,anditisthesecombinedprofilesthat
trum of NGC 6810 discussed below. The images from each areshown inthetoppanelsofFig. 2.Ingeneralthesurface
MOSdetectorwerecombinedforthepurposesoffurtherim- brightnessprofilesproducedfromthePN,MOS1andMOS2
age analysis. In each energy band thebackground intensity detectorsindividually were verysimilar toeach other(with
was estimated as the mean surface brightness on the same the PN and MOS2 profiles being the most similar to each
CCDchipasNGC6810,afterexcludingthedatawithinara- other), justifying their combination. The profiles from the
diusof30′′ ofanydetectedpoint-likeX-raysourceorwithin MOS1 detector were the most different, and are shown in
90′′ of NGC 6810. A simple uniform-level background sub- thelowerpanelsofFig.2,butstilldisplaythesamegeneral
traction is appropriate given the relative compactness and features.
brightness of the X-rayemission from NGC 6810. ThesoftX-rayminorandmajoraxissurfacebrightness
Figure.1showsadaptivelysmoothedPNandcombined profilesforNGC6810aresimilarinformtothediffuseemis-
MOS images of NGC 6810 overlaid with surface brightness sionprofilesfoundinChandraX-rayobservationsofhighly-
iso-contours from images smoothed with a FWHM = 15′′ inclinedstarburstgalaxies(Strickland et al.2004),although
Gaussian mask.Displaying thePN andcombined MOS im- notethatintheXMM-Newtondatawecannotrobustlyre-
agesseparatelyishelpfulwhenstudyingfaintextendedemis- move X-ray emission from the point sources in NGC 6810.
sion in that it allows us to ascertain whether certain fea- The surface brightness drops rapidly with increasing dis-
tures of interest are genuine and which may be statisti- tance from the nucleus of the galaxy along both minor and
cal or smoothing artifacts. The lowest contour levels dis- major axes. The minor axis profiles also clearly show the
playedintheseimagesarebetween3–4.7σ abovetheback- asymmetry in the soft extended emission between the east
ground, given the FWHM = 15′′ uniform smoothing (see andwestsidesofthedisk,presumablycausedbyabsorption
Hardcastle 2000). within thedisk of NGC 6810.
InthesesmoothedimagesthesoftextendedX-rayemis- Noobviousendoredgetothethisextendedsoft X-ray
sion can be traced out to a projected angular separation of emission can be discerned from this data. At low signal-
z ∼ 60 – 70′′ (∼ 8 – 9 kpc) from the plane of NGC 6810. to-noise the emission appears to extends out to projected
The emission appears to arise from within the optical disk distances of |z| or |r| ∼10 – 15 kpc (∼ 75 – 115′′) from
interiortoaradiusof∼50′′ (6.5kpc)fromthecenterofthe the center of the galaxy. However it is possible that, rather
galaxy. The centroid of the soft X-ray emission is slightly than genuine emission from these locations, we are instead
offset to the north west of the center of the galaxy. In the seeingemissionfromthebrightcentralregionsofNGC6810
hard X-ray band (E =2.8 – 8.0 keV) a single point-like re- in the very extended, but low surface brightness, wings of
gionofemission,roughlycoincidentwiththecenterofNGC theXMM-Newton PSF.However,at intermediatedistances
6810, is visible. fromthenucleustheextendedemissionmustpredominantly
Smoothingwill artificially increasetheapparentextent be genuine, as the central core of the PSF is too compact
oftheX-rayemission. Tobetterquantifytheprojected size (FWHM∼6′′,HalfEnergyWidth∼12′′)toberesponsible
ofthesoftX-rayemission1-dimensionalprofilesoftheX-ray for the ∼ 1′ diameter region of bright soft X-ray emission
surface brightness along the major and minor axes of NGC (e.g. Fig. 1 or Fig. 2).
6810 were created (see Fig. 2), excluding theemission from With no clear edge and the effect of the PSF it is im-
X-ray sources not obviously associated with NGC 6810. possibletoquantifythesizeoftheextendedsoftX-rayemis-
4 D. K. Strickland
Figure 2.SoftX-ray(E=0.3–1.1keV)surfacebrightnessprofilesalongtheminorandmajoraxesofNGC6810.X-raysourcesother
than NGC 6810 have been excluded. The total surface brightness is shown as open diamonds (green), with the expected background
(frombothX-raysandparticles)shownasopencircles(red).Thebackground-subtracteddataisshownasfilledcircles(blue).Allofthe
valueswereevaluated inslicesoftotaldepth 76′′ (∼10kpc)perpendiculartotheaxisofinterest.Panelsaandbdisplaythecombined
MOSandPNdata,whilepanelscandddisplaydatafromMOS1alone.GapsinthedataareduetogapsbetweentheCCDchips.The
errorbarsrepresent1σuncertainties.IntheminoraxisprofilethenegativeaxisliestotheeastofthediskasseeninFig.1.Inthemajor
axisprofilethenegativeaxisextends towardthenorthfromthecenter ofthegalaxy.
sioninamannerthatisnotobservation-specific.Incommon theoffsetoftheopticalandUVnucleusfromtheIRmaybe
withobservationsofextra-planarradioemission fromspiral duetoobscuration.Alternativelythe2MASSpositionmight
galaxies (e.g. Dahlem et al. 1995, 2006) we will quote the beinerror,inwhichcasethetruenucleusofNGC6810lies
extentastheobservedsizeatafixedstatisticalsignificance, at α=19:43:34.3, δ=-58:39:17.3 (±1′.′0) (J2000.0).
here evaluated as the size at S/N = 3 in the background- With respect to the nucleus or the nearby UV/Hα/X-
subtracted surface brightness profiles. Underthis definition ray central peak the larger-scale UV emission is the most
theE =0.3–1.1keVemissionextendstoprojectedheights asymmetrically distributed. What appears to be a UV-
ofz∼−7.0(west)andz∼+6.9kpc(east)alongtheminor bright arc partially cups the nuclear emission to the west
axis,andalongthemajoraxistoprojectedradiiofr∼−6.3 and south, with a total major-axis diameter of ∼30′′, with
(north) and r∼+6.5 kpc(south). scatteredfainterUVsourcesatslightlylargerradius.Inthe
Hα image the UV arc and exterior fainter UV sources ap-
peartoallbepartofaroughlyrectangularHα-brightregion
3.2 Comparison to UV and optical images alongPosition Angle∼167◦,(offset from thePAoftheop-
tical disk by ∼ 9◦) and of total length ∼ 50′′. The north
IfwecomparetheX-rayimagestoground-basedopticaland
western edge of this region is the base of the most promi-
near-UVimagesfromtheOpticalMonitoritbecomesappar-
nentHαfilament,whichextendsat least ∼40′′ (∼5.3kpc)
entthatthesoft diffuseX-rayemission isclosely associated
from theplane of thegalaxy (see Fig 3i).
withthestrongeststarformingregionsinthedisk,andwith
warm-ionized gas such as theHαfilament that projects be- At low surface brightness the outer edge of the optical
yond theoptical disk of thegalaxy (Fig. 3). disk sports a ring of Hii regions coincident with an optical
The most intense soft X-ray emission occurs at the re- dust lane (Fig. 4). This ring is not visible in the soft X-ray
gion ofbrightest optical and near-UVemission. Thisregion images, even though X-ray emission is detected at regions
is slightly offset, by ∼ 3′.′5, to the NNW of the nuclear po- of similar Hα surface brightness outsideof theoptical disk.
sition from the 2MASS near-IR observations (2MASS Ex- ThespatialresolutionoftheEPICPNandMOSinstru-
tended Objects. Final Release, shown as the white cross in ments is insufficient to allow X-ray emission from compact
panelsatocofFig.3).Thisoffsetislargerthanthe.1′′ac- sourcestobeseparatedfromtrulydiffuseemission.Theex-
curacyoftheOM,opticaland2MASSIRimageastrometry. tendedsoftX-rayemissionweobserveislikelytobemainly
Strongdust lanes throughout theoptical disk of thegalaxy diffusethermalemissionbutthiscannotbeprovenwiththe
are visible in the 3 color composite image shown in Fig. 4, existing data. Observationswith the Chandra X-ray Obser-
and even more so in an archival F606 HST WFPC2 image vatory would certainly allow a better separation of the ex-
whichpartiallycoversthenuclearregionofNGC6810.Thus tendedemissionintodiffuseandpoint-likeX-raysources,in
XMM-Newton observations of NGC 6810 5
Figure 3.Near-UVandoptical imagesofNGC6810. Panelsatocshowthecentral 90′′ asseenintheUV(UVM2filter),continuum-
subtractedHαandsoftX-rayemission.Thecoordinatesofthenucleus,asfoundinthe2MASSsurvey,isshownasawhitecross.Panels
d through f are XMM-Newton Optical Monitor images in the UVM2, UVW1 and U filters respectively, overlaid with the contours of
E=0.3–1.1keVsoftX-rayemissionfromtheadaptivelysmoothMOSimage.Contourlevelsareatthesamesurfacebrightnesslevelsas
inFig.1a.NotethestraylightartifacttothesouthwestofNGC6810.Panelsg,handiarethenarrowbandRandcontinuum-subtracted
Hα+[Nii]imagestakenbyHameed&Devereux(1999).TheHα+[Nii]imageisshowninpanelshandiontwodifferentintensityscalesto
presentboththehighandlowsurfacebrightnessopticalemission.Thewavelengthrangeofeachfilterisgiveninunitsofnm.Thedashed
ellipserepresentstheinclination-correctedd25 ellipseofNGC6810.ThestrongestspursofextendedsoftX-rayemissionseenbeyondthe
optical disk of the galaxy are clearly associated with Hα emission, in particular at the filament emerging from the northern-most UV
brightregionintheUVM2imageandextending ∼40′′ (∼5.3kpc)tothenorth-west.
particularinandaroundtheUVandHαbrightstarforming the galaxy may indeed be extra-planar emission at heights
regions. of several kiloparsecs along the minor axis, or it might be
emission at large galactic radius but still within the plane
ofthegalaxy.Ascalediagramofthesescenariosisshownin
Fig. 5.
3.3 Emission region geometry
With an inclination of i = 78◦ the exact location of the IfmuchoftheX-rayandHαemissionisassociatedwith
extended soft X-ray emission is ambiguous with respect to a starburst-driven wind then it can be seen from Fig. 5a
the true disk plane of NGC 6810. In particular the soft X- that the apparent size of the emission features (i.e. seen in
ray emission seen in projection beyond the optical disk of projection) is similar to but larger than their true vertical
6 D. K. Strickland
Figure4.(a)Three-colorcompositeimageofNGC6810,withthecontinuum-subtractedHαimage(red),adaptively-smoothedE=0.3–
1.1keVMOS1+MOS2image(green)andR-bandimage(blue)allshownonahistogram-equalizationscale.Theassociationbetweenthe
Hα filament on the NW side of the disk and the edge of the soft X-ray emission is apparent. (b) A UV/optical three-color composite
imageofNGC6810,shownonthesamespatialscaleas(a).Theground-basedopticalR-bandimage(red),OMU-band(green)andOM
UVW1images(blues)areallshownonahistogram-equalizationscale.Notetheprominentdustlanesinthediskandtotheeastofthe
nucleus.TheannularstructuretotheSWofNGC6810seenintheUandUVW1imagesisanartifactofstraylightintheOMtelescope.
extent.Intheexample shown thetrueradius of thebase of Asemissionfromlargegalacticradiusislesslikelythan
theoutflowisassumed tober′ =6.5kpc.Thetruevertical extra-planar emission we will assume for the remainder of
extent of the X-ray emission z′ ∼ 4 kpc can be calculated thepaperthattheemissionbeyondtheopticaldiskofNGC
given that the apparent projected height z ∼ 7 kpc is the 6810 is extra-planar.
sum of the projected extent of inclined disk (h = r′sinφ, Note that with the existing data we can not prove un-
where φ=90−i) and the projected size of the edge of the ambiguouslythatthisisthecase.Sensitivehigh-spatialres-
outflow lcosθ1/2, where θ1/2 is the half opening angle in olution X-ray observations with Chandra might be used to
the truncated conical geometry we assume, and that z′ = distinguish between the two scenarios. If the emission ge-
lsin(θ1/2+φ). ometry is similar to that shown in Fig. 5a then we would
InthiscasemuchofthesoftX-rayemission(E .2keV) expect to see evidence of additional absorption of the X-
withintheopticalconfinesofthediskandbeyonditislikely ray emission near the eastern optical edge of the disk, as
to be genuinely-diffuse thermal emission. Indeed fitting the the X-ray emission seen in projection in this region would
EPIC PN and MOS spectrum of theentire galaxy confirms need to pass through the intervening disk of NGC 6810 to
thepresenceof asoft thermalcomponent(see below).Note reachus(anexampleofthiseffectcanbeseeninNGC253,
that only a minorfraction of thesoft X-rayemission comes e.g. Pietsch et al. 2000). We would also need to know more
from regions beyond the optical disk of the galaxy, and it abouttheHigasdistribution inNGC6810 beforeassessing
shouldbenotedthatthetotalX-rayspectrumpresentedbe- whethersuchatestwerefeasible.Himappingwouldalsobe
lowisnotnecessarilyrepresentativeofthis(possibly)extra- advantageous in providingmore accurate determinations of
planar X-ray emitting material. theinclination angle of thedisk.
If the emission seen at a projected height above the One explanation for the position angle offset between
disk mid-plane z is not actually extra-planar then it must thelarge-scaleopticaldiskofNGC6810andtheUV,Hαand
lieatalargegalacto-centric radius,r′ =z/cos(i),asshown softX-raybrightinnermost6kpcisthatthisisastellarbar.
in Fig. 5b. For the optically-derived inclination of i = 78◦ Thepresenceofbars,andtheirassociation with thebaseof
for NGC 6810 then the north-western Hα filament, with a the wind in superwind galaxies, has been remarked upon
projected height ofz ∼5kpc,would actually extendoutto before(Strickland et al.2004).Howevertheexistingdatais
r′∼22kpc.Thesoft X-rayemission extendstoaprojected equallyconsistentwiththeUVarcandHα-brightregionbe-
height of z ∼ 7 kpc, implying a true radial extent of r′ ∼ ingpartially obscured innerspiralarmsoracircum-nuclear
34 kpc in this model. These values are large compared to star forming ring.
the stellar radius of the galactic disk. Bright Hα and/or For the purpose of comparing the presumed extra-
soft X-ray emission is not typically observed so far beyond planar emission with other highly inclined (i & 70◦)
the stellar disk of spiral galaxies unless it is extra-planar star forming galaxies observed with Chandra and XMM-
(Bland-Hawthorn et al. 1997; Read et al. 1997; Tyler et al. Newtonanestimateofthefractionofthetotal(background-
2004). subtracted)X-raycountrateintheE =0.3–1.1keVenergy
The estimates of the true radial and vertical sizes are bandarisingatminoraxisheights|z|>2kpcisgiveninTa-
sensitive to the assumed geometry, in particular uncertain- ble1.Thetotalcountratewasassessedinrectangularregion
tiesintheinclinationofthegalaxy.Forexample,anerrorof oftotalwidth(alongthemajoraxisofthegalaxy)2.′39and
only 5◦ in our adopted inclination would change the radial total(minor-axis)height2.′54(∼20kpc)centeredonthenu-
extentoftheX-rayemission inthesecondcase givenabove cleus. Both MOS and PN data give a similar answer, with
by between 30 and 70% from thevalueestimated above. ∼ 35% of the soft band X-ray emission coming from the
XMM-Newton observations of NGC 6810 7
Figure5.AdiagramillustratingtwopossibledistributionsforthesoftX-rayemitting(thickbluedashedlines)andHαemittingmaterial
(thinreddashedlines)inandaroundNGC6810.Thepanelsrepresentslicesperpendiculartotheplaneofthesky,i.e.alongthelinethe
sight.ThesolidellipserepresentsthestellardiskofNGC6810,withamajor-axisdiameterof18.9kpcandaxialratio5:1(seebelow),at
theadoptedinclinationofthediski=78◦.InbothpanelstheplaneoftheskyismarkedbythelineDD′.ThetrueminoraxisofNGC
6810isparalleltothelineEE′,andthetruemajoraxisisparalleltoFF′.ThescalebarAA′ representsaprojectedverticalheightof2
kpc.TheHαfilamentonthenorthwestsideofdiskextendsouttoaprojectedheightfromthemid-planeof5kpc(representedbyBB′),
whilethe softX-rayemissionisdetected out toaprojected distance of7kpc fromthemid-plane(CC′)In panel (a) theassumption is
thatsoftX-rayandHαemittingplasmaaredistributedwithinandabovethediskinageometrysimilartothatofthematerialobserved
in galaxies with starburst-driven winds. The extra-planar emissionis most visible on a truncated conical surface that marks the outer
edge of an outflow with a half-opening angle of θ ∼ 25◦ (based on the angle between the observed Hα filament and the projected
1/2
majoraxisofthegalaxy)andhasaradiusatitsbaseofr′=6.5kpc.Thisscenarioassumesacylindrically-symmetricgeometry.Inpanel
(b)itisassumedthattheemissionlieswithintheplaneofthegalaxyandappearsinprojectionaboveandbelowtheopticaldiskofthe
galaxybecauseitextends tomuchlargergalacticradiusthanthestars.Tosavespacenoemissionisshownonthelefthandsideofthis
panelalthoughitisassumedtobethere.
halo of NGC 6810. This halo fraction is not as accurate an vativeapproach and only fitthespectral datain theE=0.5
estimate of the fraction of diffuse X-ray emission from the – 10.0 keV energy band.
halo as might be obtained with Chandra, given the larger InspectionoftheEPICspectrumofNGC6810(Fig.6)
PSFwingsofXMM-Newton,thatwecannotseparatepoint reveals soft thermal emission with line features at ener-
source from diffuse emission, and can not correct for the gies E . 2 keV and a fainter featureless continuum for
probably higher absorption experienced by diffuse emission E & 2 keV. This is typical of CCD spectra of star form-
withinthediskofthegalaxy.Neverthelessitislieswithinthe ing galaxies (see for e.g. Ptak et al. 1997; Dahlem et al.
range of halo diffuse flux fraction found for other starburst 1998; Strickland et al. 2004; Tu¨llmann et al. 2006). There
galaxies with superwinds (7 – 48%, Strickland et al. 2004), is no sign of the scattered AGN continuum or strong Fe-K
andsupportsthehypothesisthatthesoftX-rayemission in fluorescencethatiscommonlyfoundinSeyfert2galaxies,in-
NGC 6810 extendsfrom the disk into its halo. cludingthosewithstarbursts(Levenson et al.2001a,2002).
Astatisticallysatisfactoryfitwasobtainedusingaspec-
tral model comprised of an absorbed collisionally ionized
plasma(weusedtheAPECplasmacode,Smith et al.2001)
to represent the diffuse hot gas, and an additional more-
3.4 Spectral analysis
highly-absorbed power law to represent X-ray binaries and
We created X-ray spectra of NGC 6810 including all emis- any low luminosity AGN component.
sion within a radius of 1′ of the nucleus. No attempt was Using a hot plasma model with Solar abundance ra-
made to create multiple spectra of different regions within tios (Anders& Grevesse 1989) results in clear systematics
NGC 6810 as the ∼ 100′′ angular diameter of the X-ray in the spectral fit residuals. The systematics in the resid-
emissionfromNGC6810isonlyoforderseveralPSFwidths, uals disappeared once we fitted for the abundance of the
and as the X-ray emission from within the galactic disk is α elements (specifically the abundances of O, Ne, Mg, Si
essentially unresolved. and Ar were constrained to be equal) and iron. We ob-
Backgroundspectrawerecreatedfromasource-freecir- tain an enhanced relative abundance of α-elements to iron
cular region of radius 75′′ centered at α =19:43:30.1, δ =- (best fit Zα/ZFe ∼ 2Zα,⊙/ZFe,⊙), again characteristic of
58:37:04.5, directly adjacent to the region the NGC 6810 starburst galaxies (Strickland et al. 2004). Note that the
spectra were extracted from, and also on the same CCD absolute values of the X-ray derived elemental abundances
chips.Thebackground-subtractedcountratesforNGC6810 with respect to hydrogen are ambiguous, given that we are
in each EPIC instrument are given in Table 1. fitting a simple spectral model to emission most probably
Spectral models where fit simultaneously to the PN, coming from many spectrally-distinct regions within NGC
MOS1 and MOS2 spectra, allowing for small cross- 6810 (Dahlem et al. 2000; Weaveret al. 2000) and also be-
calibration uncertainties between the different detectors. cause there are degeneracies in fitting CCD-spectral reso-
Thespectrawere binnedtoensureaminimumof 10 counts lution X-ray spectra of α-enhanced plasmas (Martin et al.
per bin to allow the use of χ2 fitting. As calibration uncer- 2002; Strickland et al. 2004). A two-temperature plasma
taintiesremainsignificantatlowenergiesweadoptaconser- model provides a marginally better fit than a single tem-
8 D. K. Strickland
Table 1. X-ray count rates for NGC 6810, accumulated within a radius of 1′ of the center of NGC 6810. The background has been
subtracted. Thehalocountratefractionisthefractionofthenetemissionaccumulatedatprojectedminoraxisheightsof|z|>2kpc.
Parameter Units Countrateineachenergyband Halocountratefraction
E=0.5–2.0keV E=2.0–10.0keV E=0.5–10.0keV E=0.3–1.1keV
PNcountrate cts/s 0.1193±0.0025 0.0106±0.0012 0.1299±0.0027 0.37±0.02
MOS1countrate cts/s 0.0313±0.0010 0.0039±0.0006 0.0352±0.0011 0.35±0.03
MOS2countrate cts/s 0.0291±0.0009 0.0036±0.0006 0.0327±0.0011 0.35±0.03
Figure 6.(a) TheEPIC PN(filled circles),MOS1(open triangles)andMOS2(open squares) spectraof NGC 6810shown onalinear
flux and energy scale, demonstrating the strength of the soft thermal component at energies below ∼ 1keV in comparison to higher
energyemission.ThePNspectrumisrepeatedonalogarithmicscalein(b),withsolidlinesshowingthebest-fit2-temperaturethermal
plus power law model fit to the spectrum. The predicted total spectrum is shown as a solid cyan line, the two APEC thermal plasma
components as dash red and green lines and the power law component as a dotted blue line. (c) As (b), but showing the MOS1 and
MOS2spectra.
perature model, but we would expect the need for a multi- 4 THE NATURE OF NGC 6810
temperature plasma model to become more apparent with
higher signal-to-noise spectra. The best-fit spectral param- The ∼ 7-kpc-scale extended soft X-ray emission discovered
intheXMM-Newtonobservations,thepreviously-discovered
eters for NGC 6810 are given in Table 2. To ease compar-
ison to other galaxies observed with Chandra we present ∼5 kpc-scale Hα filament (Hameed & Devereux 1999) and
fluxes and luminosities in the E = 0.3 – 8.0 keV energy theminor-axis warm ionized gas kinematics (Coccato et al.
2004)areincombinationconvincingevidenceforagalactic-
band.Thetemperatureofthedominantthermalcomponent,
scale superwind from NGC 6810.
kT ∼ 0.6keV, is typical of values found in fits to the total
X-ray emission from starburst galaxies (Read et al. 1997; Can we identify whether the outflow in NGC 6810
Ptak et al. 1997; Dahlem et al. 1998, 2003). We would ex- is a Seyfert 2-or-supernova-driven wind? It has already
pect that thereis actually a range of temperatures present, beenmentionedthattheeffectivetemperature,apparentα-
withtheextra-planaremissionbeingonaveragecoolerthan elementtoiron abundanceratioand halo-to-totalX-raylu-
thethermal emission from thedisk. minosity ratios of NGC 6810 are consistent with the other
localstarburstswithsuperwinds.Table3isacompilationof
Thepowerlawcomponentbest-fitphotonindexΓ≈1.9 a variety of the pertinent parameters of NGC 6810 derived
is again typical of a purely star forming galaxy. The dom- from theliterature.
inance of the soft thermal component in the spectrum pre-
NGC 6810 should be classified as a starburst galaxy
ventsusfromconstrainingtheabsorptioncolumntowardthe
based on standard IR diagnostics using IRAS fluxes. The
hard X-ray source or sources. We fixed this column density
60µm to 100µm flux ratio of 0.56 is typical of starburst
atN =2×1021cm−2,butequallygoodspectralfitscould
H galaxies (Lehnert & Heckman 1996a,b). Of course a buried
beobtainedwithN =1×1022cm−2.Forthepurposesofil-
H AGN might also elevate the mean dust temperature above
lustratingtheeffectthisuncertaintyhasonthebest-fitspec-
that of a normal spiral galaxy, but we would then expect a
tral parameters the best-fit single temperature hot plasma
model plus heavily absorbed power law (N = 1022cm−2) lowermid-IRspectralslopeαIR∝logS60µm/S25µm.Again
H NGC 6810 has IR properties typical of normal starburst,
is also shown in Table 2.
α = 1.87, significantly higher than the typical value of
IR
The similarity of thebest-fitpower law slope tobinary αIR.1.2ofIRASgalaxieswithAGN(Condon & Broderick
emission from normal galaxies does not alone exclude the 1991).
presence of an AGN, as Seyfert 1 galaxies also have hard With an total observed 4.85 GHz radio flux of 72±8
continuum spectra with Γ ∼ 1.9, but the hard X-ray lumi- mJy (Wright et al. 1994) and a radio to FIR flux ratio of
nosity of NGC 6810 (L ∼{4±2}×1039ergs−1, E =2–8 q = 2.57, NGC 6810 falls on the radio/FIR correlation for
X
keV) is too low for an unobscured Seyfert galaxy. normalandstarburstgalaxies(mean<q>=2.75±0.14,see
XMM-Newton observations of NGC 6810 9
Table 2.BestfitX-rayspectralparametersforNGC6810. Table 3.GeneralpropertiesofNGC6810.
Parameter Units Spectral model Property Value Note
1T+PL 2T+PL
Galaxytype Sab(s):sp 1
NHa 1021cm−2 0.78+−00..2393 0.09+−00..6049 cz 2031kms−1 2
kZZnToαFr1em1 ZZ1k0eαF4Ve,⊙,×⊙K 0020....35185789+−+−+−+−00000000........0109010826753045 0001....55304560+−+−+−+−00000000........0318021136487358 MD1id′i′iasatjoatrnDcaexiDsP.A. 112277378.361.◦19.◦3′,Mp1pc8c.9kpc 3414
knΓNToPHr2Lmb 2 1.1k.00e.42V1×cmK−2 1.1...0.7..3(F+−i00x..76e46d) 2011...039(F310+−+−+−ix000000e......d225118928752) vIMIRRr2o5bAAtaSSrySS2152µµmm 3218.1..524057×kJJm1yy0s1−01M⊙ 7756
normPL 104×L 0.15+−00..1078 0.17+−00..1130 IRASS60µm 18.2Jy 7
χ2 (d.o.f) ... 1500.6(3158) 1492.6(3156) IRASS100µm 32.6Jy 7
fX′,tot 10−14ergs−1 cm−2 22.0+−12..21 22.3+−13..01 S60µm/f100µm 0.56 7
fX,tot 10−14ergs−1 cm−2 34.3 25.8 αIR 1.87 8
ffXX,,APLPEC 1100−−1144eerrggss−−11 ccmm−−22 295.3.0+−+−53..4456..96 196.2.6+−+−53..3356..30 LSFIRRIR 43..72M×⊙10y10r−L1⊙ 79
LX,tot 1040ergs−1 3.0 2.3 S4.85GHz 72±8mJy 10
LLXX,,APLPEC 11004400eerrggss−−11 20..28+−+−0000....4543 10..58+−+−0000....3533 qf[NH=αIlIo]g6(5f8F3IR/H/fα4.85GHz) 260...56973×,01.06−212ergs−1 cm−2 111212,13
FWHM(Hα) 263kms−1 13
Note:– The spectral models used were (in the termi- FWHM([O III]) 304kms−1 13
nology used within xspec) wabs*(vapec+wabs*po) and LHα 6.1×1041ergs−1 12
wabs*(vapec+vapec+wabs*po). Both thermal and power SFRHα 1.88M⊙yr−1 9
lawcomponentsexperienceabsorptionbycomponentNHa,while logfX,APEC/fFIR -3.77–-3.60 14
thepowerlawcomponentexperiencesanadditionalcolumnNHb.
The metal abundances (with respect to the Solar abundances of
Anders&Grevesse(1989))ofboththermalcomponentswereas- 1. FromTully(1988).
sumedtobethesameinthetwo-temperaturehotplasmamodel. 2. RecessionalvelocityfromSandersetal.(2003).
TheAPECmodelnormalizationfactorK=10−14×EI/(4πD2) 3. AssumingH0=75kms−1 Mpc−1.
where the emission integral EI = R nenHdV. The power law 4. The optically-derived inclination is described in § 1. The
model normalization factor L is the photon flux in units of inclination-corrected angular diameter at 25th magnitude di =
photons keV−1s−1 cm−2 measured at E = 1 keV. Fluxes 0.1×10(D25−0.22×R25) (inarcminutes),whereD25andR2525are
and luminosities are quoted in the energy band E = 0.3 – 8.0
giveninTully(1988).
keV. The total observed flux is f′ , i.e. no correction for
X,tot 5. Theinclination-correctedrotationalvelocityisrelatedtothe
absorptionhasbeenmade.Allotherfluxesandluminositieswere
absorption-corrected to provide an estimate of the intrinsic flux W20valuegiveninTully(1988):vrot=0.5×W20/sini.
emitted within NGC 6810. The confidence regions quoted for 6. Baryonic galactic mass Mbary = 109.79 ×
fluxes and luminosities are 68.3% confidence for one interesting (vrot/100kms−1)3.51 (Bell&deJong2001).
parameter, all other confidence regions are 90% confidence for 7. IRAS flux densities are from Sandersetal. (2003), trans-
one interestingparameter. Theerroronthe observed flux f′ formedinto a total 10 –1000µm luminosity usingthe equations
X,tot
was calculated using the Xspec flux Monte Carlo method, all inSanders&Mirabel(1996).
otherfluxandluminosityerrorsaremoreapproximateastheyare 8. AsdefinedinCondon&Broderick(1991)themid-IRspec-
basedonscalingtheuncertainties inthemodelnormalizations. tralslopeαIR=log(S60µm/S25µm)/log(60/25).
9. Star formationrates derived fromthe IR or Hαluminosity
using the equations given in Kennicutt (1998), converted to a
Condon & Broderick 1991). Forbes & Norris (1998) stud-
ied the central few arcseconds of NGC 6810, finding a low Salpeter IMF between mass limits of 1 and 100 M⊙. The SFR
assumingaKroupaIMFis1.7timeslarger.
brightness temperature and a q-value consistent with star-
10. Radiofluxdensityat4.85GHzfromWrightetal.(1994).
burst activity evenin this nuclear region.
11. The Far-IR to radio flux ratio q as defined in
The star formation rate of NGC 6810 is SFR ∼ 2 –
3M⊙yr−1 based on the galaxy’s Hα and IR luminosities Condon&Broderick(1991).
12. Theextinction-correctedHαfluxisbasedontheHα+[Nii]
(these rates assume a Salpeter IMF between mass limits of
flux given in Hameed&Devereux (2005), corrected for [Nii]
1 and 100M⊙). The core-collapse SN rate associated with emission using the Nii/Hα flux ratio measured by Kewleyetal.
this is approximately one-tenth of the star formation rate,
(2001). The Cardellietal. (1989) extinction model was used to
buttheassociatedkineticenergyinputofL ∼(1.2–2.0)×
SN derive A(Hα) from the E(B-V) value found by Kewleyetal.
1042ergs−1 would be sufficient to power a superwind (the (2001).
L value comes from the Leitherer et al. (1999) starburst
SN 13. FromKirhakos&Steiner(1990b).
model for constant star formation rate, evaluated 30 Myr
14. Logarithm of the total thermal X-ray flux to the
after the onset of star formation). FIR flux defined as fFIR = 1.26 × 10−11(2.58S60µm +
In local starbursts with superwinds the soft diffuse X- S100µm)ergs−1 cm−2.
ray emission associated with the wind typically has a lu-
10 D. K. Strickland
minosity that isonly several hundredthsof apercent of the 2-like luminosity of L ∼ (1.8±0.3)×1042ergs−1 if the
X
IRluminosityofthegalaxy.TheXMM-Newtonobservations source lies at the distance of NGC 6810. NGC 6810 also
ofNGC6810lackthespatialresolutionnecessarytocleanly satisfiedtheiropticalspectroscopic criteriaforclassification
separatediffusefrompoint-likeX-rayemission,butareason- as an AGN: [Nii]/Hα > 1.0 and/or FWHM ( [Oiii] or Hα)
ableapproximationistousethetotalestimatedsoftthermal > 300kms−1, although just barely as only the [Oiii] line
X-ray luminosity as a proxy for the truly diffuse emission. width satisfies their criterion (see Table 3).
The logarithm of the ratio of the thermal X-ray flux to the A[Oiii]linewidththislargewouldbeunusualinanor-
FIR flux is ∼−3.8 to -3.6 for NGC 6810 (see Tables 2 and malgalaxy,butnotinastarburstgalaxywithasuperwind,
3).Againthesenumbersareconsistentwithlocalstarbursts where the velocity of warm ionized gas typically reaches
withsuperwinds,wherethelogofthesoftdiffuseX-rayflux 200 – 600kms−1 (Heckmanet al. 1990; Adelberger et al.
to theFIR flux is −3.6±0.2 (Strickland 2004). 2003). Indeed Coccato et al. (2004) specifically discuss the
The soft X-ray luminosity, X-ray/FIR luminosity ratio largelinewidthinNGC6810intermsofapossibleoutflow.
and IRAS 60µm to 100µm flux ratio of NGC 6810 are ex- Kewley et al. (2001) apply a variety of galaxy classification
tremely similar to those of NGC 1511, another starburst methodsbasedonopticallinefluxestoIRASgalaxiesinthe
galaxy with a probable wind observed with XMM-Newton southern hemisphere. Of the eight methods they apply to
(Dahlem et al. 2003). This is despite their baryonic masses NGC6810theyclassifyitasaHiigalaxysixtimes,withone
differing by close to an order of magnitude. It illustrates method giving ambiguous results and one method yielding
how strongly correlated the soft X-ray and FIR properties a classification of extreme starburst / borderline LINER.
of star forming galaxies are, and (contrary to na¨ıve expec- Forbes & Norris(1998)alsosuggesttheNGC6810may
tation) how little effect large galactic mass appears to have have been misclassified as being a Seyfert 2 galaxy, based
on starburst-driven superwinds. Galaxy mass does appears on its starburst-like radio properties and unpublished opti-
toplayaroleindeterminingthecriticalstarformation rate calspectra.HowevertheyalsopointoutthatseveralSeyfert
perunitdiskareaforcreationofradio(andpresumablyhot galaxieswithnuclearstarformationhaveradioandIRprop-
gas) halos around spiral galaxies (Dahlem et al. 2006), but erties dominated by the star formation despite being bona
foragalaxyofthemassandsizeofNGC6810thetransition fideSeyfert 2 galaxies.
between galaxies with and without radio halos occurs at a As we havediscussed thehard X-ray spectrum and lu-
mean star formation rate per unit area approximately two minosityofNGC6810areinconsistentwithaSeyfert2clas-
orders of magnitudelower than that found in NGC 6810. sification. Theidentification of NGC6810 with 1H1930-589
The X-ray emission not associated with the thermal is most probably spurious. That NGC 6810 had a signifi-
components (i.e. the hard X-ray emission and the power cantly higher X-ray luminosity in the early 1980’s is a less
law component, all presumably from compact objects) is likelyalternative,asthereisnootherevidenceofNGC6810
also at the level we would expect from purely stellar everbeingaluminoushardX-raysource(Ward et al.(1978)
processes with no additional AGN contribution required. place a 3σ upper limit on the E = 2 – 10 keV luminosity
Colbert et al. (2004) found an empirical relationship for of NGC 6810 of L < 6×1042ergs−1 based on Ariel 5
X
the X-ray point source luminosity of galaxies from both data),norwouldtheopticalclassificationofthegalaxysug-
young and old stellar populations. Given the stellar mass gest it to be, or have recently been, a Seyfert galaxy. The
of NGC 6810 (∼ 8×1010M⊙, see Table 3) we would ex- possibility remains that either a low luminosity AGN, or a
pect the point source X-ray luminosity from the old stel- more luminous but very heavily obscured AGN, is present
lar population to be L ∼ 1.4 × 1040ergs−1 in the in NGC 6810. Deep observations with the Hard X-ray De-
XP,old
E = 0.3 – 8.0 keV energy band. The expected X-ray emis- tector(HXD)onSuzakuwouldbeneededtorulethesecond
sionfrom pointsourcesassociated withongoingstarforma- possibility out, but at present we conclude that NGC 6810
tion is L ∼ 5.7×1039ergs−1 (we have corrected for does not deserve classification as an active galaxy. Radio
XP,old
the different IMFs used in the the SF rate given in Table 3 continuum observations of NGC 6810 would offer another
and Colbert et al. (2004)). The total predicted X-ray point method of searching for a low luminosity AGN, in addition
source luminosity of ∼ 2×1040ergs−1 is 2.5 times larger toconstrainingthelocationofrecentSNactivitywithinthe
than the luminosity of the power law component in NGC disk.
6810, but this is within the level of scatter expected from There is one respect in which the outflow from NGC
theColbert et al. relationship. 6810 may be less than typical, in that the wind appears to
The above mentioned galactic properties and the lack originatenotinacompact(r.1kpc)nuclearstarburst(as
of any sign of AGN activity in the 0.5 – 10 keV en- seenin “classic” superwind galaxies suchasNGC253, M82
ergy band cast doubt on the classification of NGC 6810 andNGC3079),buttohavealarger∼6.5-kpcradiusbase.
as a Seyfert 2 galaxy. The original classification of NGC Such“disk-wide” superwindsare lesscommon thanthenu-
6810 as a Seyfert 2 galaxy is due to Kirhakos & Steiner clearsuperwinds,althoughthereareahandfulofpreviously
(1990b), based on follow-up optical spectroscopy of IRAS known examples such as NGC 4666 (Dahlem et al. 1997)
sources near to or within the HEAO-1 hard X-ray error and NGC 5775 (Tu¨llmann et al. 2006). The base radius of
boxes of previously unidentifiedX-ray sources (Wood et al. a superwind appears to match the region of active star-
1984; Kirhakos & Steiner 1990a). Kirhakos& Steiner iden- formation (see e.g. Strickland 2004), and as we have men-
tified NGC 6810 as the counterpart to the HEAO-1 source tioned earlier may be related to stellar bars (for unknown
1H1930-589, even though NGC 6810 lies outside the 95% reasons). Indeed, one of the greatest current weaknesses of
confidence error box of this source. This X-ray source had numerical models of superwinds is their failure to produce
a E = 2 – 10 keV X-ray flux of f ∼ (2.0 ± 0.3) × awindbaseradiusequaltotheradiusofthestarformation
X
10−11ergs−1 cm−2, which would correspond to a Seyfert- region without imposing unphysical boundary conditions