Table Of ContentMon.Not.R.Astron.Soc.000,1–6(2008) Printed20January2009 (MNLATEXstylefilev2.2)
The nature of the ASCA/INTEGRAL source
AX J183039–1002: a new Compton-thick AGN?
9
0 L. Bassani,1⋆ R. Landi,1 R. Campana,2 V. A. McBride,3 A. J. Dean,3 A. J. Bird,3
0
D. A. Green,4 P. Ubertini,5 A. De Rosa5
2
1 INAF/IASF-Bologna, Via P. Gobetti 101, I-40129 Bologna, Italy
n
2 Department of Physics, Universityof Rome “La Sapienza”, Piazzale A. Moro 2, I-00185, Rome, Italy
a
3 School of Physicsand Astronomy, University of Southampton, Highfield, SO17 1BJ, United Kingdom
J
4 Astrophysics Group, Cavendish Laboratory, 19 J. J. Thomson Ave., Cambridge CB3 0HE, United Kingdom
6 5 INAF/IASF-Roma, Via Fosso del Cavaliere 100, I-00133, Roma, Italy
1
]
A
Accepted 2009January16.Received2009January15;inoriginalform2008October10.
G
.
h
ABSTRACT
p
- We reportonthe identification of the X/softγ-raysourceAX J183039–1002detected
o
with ASCA and INTEGRAL/IBIS. The source, which has an observed 20–100 keV
tr flux of∼8.6×10−11 ergcm−2 s−1,isinside a diffuse radiosupernovaremnant(SNR)
s
andisspatiallycoincidentwithacompactradiosource.WeanalyzedarchivalChandra
a
[ and XMM-Newton observations in order to identify the ASCA/INTEGRAL source.
A point-like Chandra X-ray object was found to be positionally coincident with the
1 compactradiosourceandwithintheerrorcircleoftheASCAandINTEGRALsources.
v
Although the association of a compact radio/X-ray source with a radio supernova
9
remnant could be indicative of a pulsar wind nebula (PWN), the XMM-Newton X-
9
ray spectrum is compatible with an absorbed, Seyfert-2 like AGN, since it provides
4
evidence foranironemissionline of∼ 1keVequivalentwidth; furthermorethe X-ray
2
. source spectrum is similar to that of other Compton thick AGN where the .2 keV
1
data are associated to a warm reflector and the &10 keV one to a cold reflector.
0
9 Key words: X-rays: general, X-rays: individuals: AX J183039–1002
0
:
v
i
X
1 INTRODUCTION courseaprimaryobjectiveofthesurveyworkbutitismade
r
a verydifficultbytherelativelylargeINTEGRALerrorboxes.
A key strategic objective of the INTEGRAL (International
Cross correlations with catalogues in other wavebands can
Gamma Ray Astrophysics Laboratory) mission isahigh en-
beusedasausefultoolwithwhichtorestrictthepositional
ergysurveyoftheGalacticplaneandCentre(Winkleretal.
uncertaintyoftheobjectsdetectedbyIBISandthereforeto
2003)inprimisandofalltheskyasaby-productofpointing
facilitate the identification process. Observations at softer
observations.Thismakesuseoftheuniqueimagingcapabil-
X-ray energies, where the positional accuracy is much bet-
ity of IBIS (Imager on Board Integral Satellite; Ubertini et
ter,hasalreadyprovedaninvaluableaidintheidentification
al.2003),whichallowsthedetectionofsourcesatthemCrab
andclassificationprocess(Stephenetal.2006;Masettietal.
level, with an angular resolution of 12′ and a point source
2008 and references therein), but the use of data at other
location accuracy (PSLA) of typically 1′–3′ within a large
wavebands,forexampletheradio,isalsousefultohelpiden-
(29◦ ×29◦) field of view. So far, several surveys produced
tify peculiar and interesting objects.
from the IBIS/ISGRI data have been reported, the most
complete being those of Bird et al. (2007) and Krivonos et
In this Letter we report on the unusual nature of
al.(2007);bothreportedaround400sourcesdowntoaflux
level of a few mCrab above ∼20 keV. A significant fraction the ASCA/INTEGRAL source AX J183039–1002, located
withindiffuseradioemissionandlikelyassociatedtoacom-
(25–30%) of the objects in these surveys have no obvious
pactradiosource.WeuseunpublishedChandra andXMM-
counterpart at other wavelengths and therefore cannot yet
Newton observationstolocalizeandidentifytheX-raycoun-
beassociated with anyknown class of high energy emitting
terpart of this high energy emitter and to characterize its
objects.Searchingforcounterpartsofthesenewsourcesisof
broad-bandspectrum.Weprovideargumentsfortheidenti-
ficationofthisX/gamma-raysourcewitheitheranewcom-
⋆ E-mailaddress:[email protected] posite Supernova/PWNsystem ora backgroundAGN;this
2 L. Bassani et al.
second option is however the more convincing on the basis
of the source spectral characteristics.
2 THE ASCA/INTEGRAL SOURCE
AX J183039–1002
AX J183039–1002 was first reported by Sugizaki et al.
(2001) as an unidentified and faint X-ray source detected
by ASCA during the Galactic Plane Survey.It was located
at RA(J2000) = 18h30m39s and Dec(J2000) = –10◦02′42′′
withapositionaluncertaintyof∼3′.ItssoftX-rayspectrum
iswelldescribed(χ2/ν =4.9/10) byanabsorbedflatpower
lawwithΓ=0.04+−00..7679,NH =(cid:0)3.1+−32..46(cid:1)×1022 cm−2 andan
unabsorbed 0.7–10 keV fluxof ∼2.3×10−12 erg cm−2 s−1.
AXJ183039–1002 was thenreported as anIBIS source
both by Bird et al. (2007) and Krivonos et al. (2007); here
we use data collected in the 3rd IBIS survey (Bird et al.
2007), whichconsists ofallexposuresfrom thebeginningof
the mission (November 2002) up to April 2006. The total
exposureonthisregion is∼1.78Ms.ISGRIimages foreach
available pointing were generated in various energy bands Figure 1. IBIS/ISGRI 20–40 keV image of AX J183039–1002.
usingtheISDCofflinescientificanalysissoftwareversion5.1 ThegreenandwhitecirclescorrespondtotheINTEGRAL/IBIS
(OSA 5.1; Goldwurm et al. 2003). The individual images andASCA/GISerrorboxes.
were then combined to produce mosaics of the sky region
of interest here to enhance the detection significance using
thesystemdescribedindetailbyBirdetal.(2004,2007).A 0.26 mJy, and is also detected at 6 cm with an integrated
clearexcessisobservedintheIBISmapwithasignificanceof flux density of 1.89±0.18 mJy (White, Becker & Helfand,
∼8σatapositioncorrespondingtoRA(J2000)=18h30m39s.8 2005). The diffuseemission (G21.6417+0.0000) around this
andDec(J2000)=–10◦00′25′.′2andwithanassociateduncer- discrete source has an integrated flux of 1.54 Jy at 20 cm
tainty of 4.8′ (90% confidence level). Our position is com- and a diameter of 2.8′.
patible with that reported by Krivonos et al. (2007) (note Comparison with mid-infrared images obtained with
that their 2.1′ error radius corresponds to 68% confidence MSX (Price et al., 2001) allows us to discriminate between
level). Figure 1 shows the 20–40 keV image of the region thermal and non-thermalsources: indeed high valuesof the
surroundingAXJ183039–1002 withsuperimposed boththe radio-to-IRfluxratio(20cmversus20µm,wherethesource
INTEGRALandASCAerrorcircles.TheIBIS/ISGRIspec- isgenerallynotdetectedintheMSX)generallyindicatenon
trum was obtained following usual procedures: fluxes were thermal radio emission such as is produced in a supernova
extractedfromthelocationofthesourcein10narrowenergy remnant(SNR,see,e.g.,Fu¨rstetal.1987,Green1989).This
bandsoverthe17–100 keVrangeforallavailablepointings; evidence together with the morphology of the diffuse radio
a spectral pha file was then made by taking the weighted emission(inthiscasealmostacompleteshellstructure)and
meanofthelightcurveobtainedineachband.Anappropri- thedetectionofthesourceinthe90cmimages,ledHelfand
atelyre-binnedrmffilewasalsoproducedfromthestandard andco-workerstoproposeG21.6417+0.0000 asahighprob-
IBISspectralresponsefiletomatchthephafileenergybins. ability SNRcandidate.
Here and in the following, spectral analysis was performed The association of a hard X-raysource coincident with
withXSPECv.11.3.2packageanderrorsarequotedat90% a radio point source and surrounded by a very low surface
confidencelevelforoneinterestingparameter(∆χ2 =2.71). brightness radio shell is suggestive of a situation where a
A simple power law provides a good fit to the IBIS data SNR harbors a classic young radio pulsar possibly together
(χ2/ν = 7.6/8) and a photon index Γ = 3.1+−00..97 (much withitswindnebula(PWN).Adisplacementofthecompact
steeper than the ASCA one) combined to an observed 20– sourcewithrespecttothecentreoftheshellstructureisnot
100 keVflux of 8.6 × 10−11 erg cm−2 s−1. unusual, since many pulsars are born with high kick veloc-
ities which allow them to propagate through the shocked
ejecta in the SNR interior (Gaensler and Slane 2006). Al-
3 RADIO COUNTERPART thoughthedistanceandageofG21.6417+0.000 arenotwell
known, it is clear from its small angular size that it’s rela-
The Multi-Array Galactic Plane Imaging Survey (MAG- tively small and young. If the compact source is a PWN,
PIS, Helfand et al., 2006) has mapped the region of in- therefore, its offset from the centre of the shell would re-
terest here with a 20 cm sensitivity of 1–2 mJy and de- quireanunusuallylarge kickvelocity.Away toknowifour
tected a compact source surrounded by diffuse extended discreteradiosourceisaPWNisbymeansofitsradiospec-
emission (see Figure 2 which is a MAGPIS extract). The trum, which is expected to have a flat energy index 0.3 6
compact source (G21.63201–0.00682) which is located at α60.1 (definedbySν ∝ν−α,Chevalier 1998): comparison
RA(J2000)=18h30m38s.29andDec(J2000)=–10◦02′46′.′4(5′′ between 20 and 6 cm data provides an α value of 0.9, con-
uncertainty) has an integrated 20 cm flux density of 5.13± siderably steeper thanwhat generally found in PWN.Also,
The ASCA/INTEGRAL source AX J183039–1002 3
Table 1.Chandra X-raydetections.
Source RA(J2000)† Dec(J2000)† S/N CountRate‡
N1 183038.3 -100247.1 27 38.90±1.50
N2 183037.1 -100245.4 4.7 1.25±0.28
† 1σ uncertaintyonChandra positionsis0.6′′.
‡ Inunitsof10−3 counts s−1 inthe3–10keVenergyband.
2objectsabovea3σdetectionthreshold;theirpositions,de-
tectionsignificanceandcountratesarereportedinTable1.
Figure 3 is a cut-out of the 3–10 keV Chandra image,
showing the2objects detectedwithin theASCAerrorbox,
together with the location of the compact MAGPIS radio
sourceG21.63201–0.00682. Itisevidentfromthefigurethat
the Chandra observation further confirms the association
of source N1 with the compact MAGPIS radio object and,
Figure2.MAGPIS(20cm)imageofAXJ183039–1002,showing
purely for its brightness, also with the ASCA/IBIS source.
boththediffuseemissionandthecompactradiosource.
The reduced positional uncertainty does not contain any
optical/infrared counterpart for this Chandra source in the
theratiobetweenradioandX-raycoreluminositiesishigher HEASARCarchiveandevenintheDSSinfraredimage(see
(∼2.3×104)thanusuallyfoundinPWNe(Becker&Helfand, Sect.6 for a discussion on the expected optical extinction).
1987), indicating either a highly unusual PWN or an alter- To look for the possible presence of a PWN around
native nature for AX J183039–1002, for example that of a sourceN1,wehavealsocomparedtheChandra PSFagainst
backgroundAGN.Indeedtheradiospectrumof0.9istypical that of the source but havefound no evidencefor it: object
of an activegalaxy, although thisscenario is disregarded at N1 is seen as point-like by Chandra implying that a PWN
firstglanceforthelowprobabilityofachancealignmentbe- is not present or, if present,is extremely compact.
tweenanAGNandtheSNR.Howeveramoreindepthanal- TheCIAO script psextract was thenused togenerate
ysis indicates that the probability of such an association is the spectrum, with appropriate background and response
smallbutnotinsignificant.Theunabsorbed2–10keVX-ray files,ofbothsourceN1andN2;spectrawereextractedusing
flux of the likely counterpart of AX J183039–1002 is about aradiusof5′.′6,whilebackgroundfilesweregeneratedusing
2×10−12 erg cm−2 s−1, and from the LogN–LogS relation a circular region of radius 100′′. With a count rate of 0.119
(LaFrancaetal.,2005)weestimateadensityforsuchbright counts/frame,thepile-upfractionisinsignificantatlessthan
AGNs of about 0.1 sources per square degree. The SNR 2%.InallourfittingproceduresweusedaGalactic column
has a radius of about 2.8′, i.e. an area of 6.15 arcmin2. We density, which in the direction of AX J183039–1002 is 1.52
have then a mean value of about 2×10−4 AGNs inside the × 1022 cm−2 (Dickey& Lockman, 1990).
SNR. However, we have observed in all the sky about 200– The spectrum of source N2 is very soft with very few
300SNRs(forexampleGreen’scatalog reports265sources, counts above 3 keV; when compared to the IBIS spectrum
see http://www.mrao.cam.ac.uk/surveys/snrs/). Assum- wehavetoassumeacross-calibrationconstant>4×104,im-
ingthatallofthesehavethesameareaofG21.6517+0.000, plying that the extrapolation of the Chandra data predicts
we can next estimate the probability of finding at least an IBIS flux far below that observed, a further indication
one chance association with an AGN to be about 5%. This that this source is not the counterpart of the ASCA/IBIS
probability is small but not insignificant. Moreover, many object.
SNRs have a radius much larger than 2.8′, therefore the ThespectrumofsourceN1isinsteadveryhardasasim-
“true”chanceassociationprobabilityshouldbeconsiderably ple power law model provides a flat spectrum (Γ = −0.86)
greater than ∼5%. andanunacceptablefit(χ2/ν =87.8/45). Residualstothis
model show some curvature possibly due to intrinsic ab-
sorption. Addition of this extra component provides a fit
4 CHANDRA OBSERVATION improvement (χ2/ν = 40/45), which is significant at the
99.99%confidencelevel,asteeperspectrum(Γ=1.01+0.57),
OnFebruary19,2006Chandra carriedoutanobservationof −0.38
theskyregioncontainingAXJ183039–1002. Datareduction a column density of NH = (cid:0)11.4+−32..89(cid:1)×1022 cm−2, and an
was performed with CIAO v.3.4 and CALDBv.3.2.4 to ap- unabsorbed 0.7–10 keV flux of (cid:0)3.2+−00..49(cid:1)×10−12 erg cm−2
plythelatestgaincorrections.Subsequentfilteringonevent s−1.TheChandra spectrumiscompatiblewithinerrorswith
grade and exclusion of high background times resulted in a the ASCA one. The spectrum is quite flat (as found previ-
total exposure of 19.2 ks. In order to locate the source of ously) and the column density exceeds largely the Galactic
hard X/gamma-ray emission associated to the ASCA/IBIS NH value:it pointstoa sourceable toemit from afew keV
object,weconcentratetheChandra imageanalysistothe3– upto∼100keV,constantintimeand,asunderlinedbefore,
10 keVband:within thesmaller ASCAerror boxwe detect able to produceemission also at radio frequencies.
4 L. Bassani et al.
AX J183030−1002 − XMM−Newton absorbed powerlaw
2
nts/sec/keV 0.010.0
alized cou 5×10−3
norm ×10−3
2
4
2
χ
0
2
−
1 2 5 10
channel energy (keV)
Figure3.Cut-outoftheChandra 3–10keVbandimageshowing
the2objectsdetectedwithintheASCAerrorbox,togetherwith Figure 4. XMM-Newton spectrum of AX J183039–1002 fitted
theXMM-Newton 1.06′′ 1σuncertaintyaswellasthelocationof withanabsorbedpowerlaw:notethesystematicresiduals.
thecompact MAGPISradioobject.
INTEGRALone to obtain a broad-band spectrum over the
5 XMM-NEWTON OBSERVATION 1–200 keV energy band. The model used to fit the XMM-
Newton spectrum gives residuals in the IBIS data suggest-
The Chandra N1 source has also been detected during the
ing the presence of some curvature or a cut-off above 10
XMMGalacticPlanesurvey(Handsetal.,2004)asXGPS–
keV, which is immediately obvious also from the compar-
I J183038–100248. It is also reported in the 2nd Serendip-
ison between the X and γ-ray photon indices. To describe
itous Source Catalogue as 2XMM J183038.2–100246 (Wat-
thebroad-bandsourcespectrumandinviewofthelowsta-
son et al., 2007, http://xmmssc-www.star.le.ac.uk); the
tistical quality of the data we have therefore replaced the
source 0.2–12 keV flux is ∼2×10−12 erg cm−2 s−1 (i.e.
absorbed power law component with other simple models,
compatible with the ASCA/Chandra flux), with almost all
like an absorbed broken power law or an absorbed cut-off
counts coming from the hardest X-ray band (4.5–12 keV),
power law; we have further fixed the cross-calibration con-
again an indication of a flat 2–10 keV spectrum.
stantC betweenXMM-Newton andIBISto1inviewofthe
We have then analyzed the spectrum of
non-variability of the source flux in the keV band given by
2XMM J183038.2–100246, using its pipelined and re-
comparing the ASCA, Chandra and XMM-Newton fluxes.
duced 2XMM catalogue data products available at the
Weemphasizethatthefitdoesnotchangeconsiderablyleav-
HEADAS archive1. This observation was performed on
ing C free to vary and in any case the large uncertainty on
September13, 2003, for a total of 4335 secondsof exposure
thisparameterreachesasalowerboundaryC =1.Thefirst
time with theEPIC-pn instrument.
model gives the same parameters found when using XMM-
The source was constant during the entire observation
Newton data alone, a break at 26±5 keV where the hard
and its spectrum is poorly described by a simple absorbed
power law steepens to Γ > 2.8 (χ2/ν = 12/21). This fit is
powerlaw(χ2/ν =62/19),atvariancewiththatobtainedby
shown in Figure5.Thesecond model,apart from theother
ASCAandChandra.Excesscountsarevisibleatlowenergies
parametersthatdon’tchangesignificantly,hasΓ=0.0±0.5
andataround6.4keV(Figure4).Agooddescriptionofthe
with a cut-off energy at 14±5 keV(χ2/ν =18/22).
data (χ2/ν = 12/14) is achieved with a model consisting
The two fits are equally acceptable and can be used
of a blackbody (kT =0.20±0.02 keV),an absorbed power
to describe in a phenomenological way the source spec-
law (NH = (8.8±4.0)×1022 cm−2; Γ = 0.6±0.5) and an trum,i.ealow-energyblackbodytogetherwithanabsorbed
iron Kα line. Both these two extra features are required by
flat power law steepening at around 10–30 keV and a low-
XMM-Newton data, with a significance of 99.99% for the
redshift iron line.
blackbody component and 99.6% for the line. This line is
verylittleredshifted(z60.08)andhasanequivalentwidth
of ∼1 keV (σ = 0.4±0.2 keV). Considering the errors on
thelinewidthandnormalization givenbythelow statistics 6 WHAT TYPE OF SOURCE IS
ofthedata,theEWcouldbesignificantlylowerthan1keV. AX J183039–1002?
If present, the broad line may be due to a blend of narrow
AXJ183039–1002iscertainlyabroad-bandemitteroverthe
lines(asoftenseeninCompton-thickAGNs,e.g.Mattetal.
2004). Note that Chandra does not clearly detect the line. 1–100keVband;itisalsoaweakandcompactradiosource
sitting inside or behind a possible supernova remnant. No
Thisisnotsurprising,sincetheeffectiveareaofACISinthe
opticalsourceispresentwithintherestrictedChandra error
Fe-K band is so small.
boxdowntoalimitingBmagnitudeof∼20–21;thissuggests
Finally, we combined the XMM-Newton data with the
either that the source is weak or heavily obscured. Outside
the Chandra positional uncertainty we find a DENIS and
1 http://heasarc.gsfc.nasa.gov/W3Browse/xmm-newton/ a 2MASS object at 1.1′′ and 2.6′′ angular distance, respec-
xmmssc.html tively.Bothhavesimilarnearinfraredmagnitudes(J =15.8
The ASCA/INTEGRAL source AX J183039–1002 5
AX J183039−1002 − XMM−Newton and INTEGRAL−IBIS/ISGRI Wehavephenomenologicallyfittedthesourcespectrumwith
simplemodels,likeablackbodyandbrokenpowerlaw.From
acomparisonwithothersimilarsourcesintheliterature(e.g.
ec/keV 0.01 NcaGnCas1s0o6c8ia,tMetahtteelotwa-l.en1e9r9g7y;sbeleacaklsbooMdya-iloikliencooemtpaol.n1e9n9t8w)iwthe
nts/s a warm reflector, and the high-energy component as a cold
u
alized co 10−3 rteofltehcitsorpcuozmzliengwiflrlomonltyhecoinmneerfrtoomrusd.eTtahieleudltrimadaitoeosbolsuetrivoan-
m tions:hopefullyitwill notbelongbeforeAXJ183039–1002
or
n 0−4 is definitely properly classified. If the heavily absorbed na-
1 tureofthissourceisconfirmed,thisbringsto5thenumber
2
of candidate Compton-thick AGN in the third IBIS cata-
1
χ 0 logue,besidesthe7confirmedsources,i.e.about20%ofthe
1 total sample of type-2 active galaxies so far detected (Bird
−
−2 et al. 2007).
1 2 5 10 20 50
channel energy (keV)
Figure5.Broad-bandspectrumofAXJ183039–1002fittedwith
a low energy blackbody, a redshifted iron line and an absorbed
ACKNOWLEDGMENTS
brokenpower law:XMM-Newton datato theleftandIBISdata
totheright.
We acknowledge ASI financial and programmatic support
viacontractsI/008/07/0 andI/088/06/0. Wealsoacknowl-
edge thefundingby PPARCgrant PP/C000714/1.
andK =13)andcompatibleRAandDecwithintherespec-
tivepositional errorssothattheycanbeassumedtobethe
samesource.WhiletheDENISpositioniscompatiblewithin
errorswiththeChandra one,the2MASSlocationistoooff- REFERENCES
set from the X-ray one to claim a match; the positional
accuracy of the 2MASS catalogue is 61′′, which combined Becker, R. H.,& Helfand, D. J. 1987, AJ,94, 1629
withtheChandrauncertaintiesprovidesatmostanangular Becker, R. H.,& Helfand, D. J. 1987, ApJ,316, 660
distance of 1.6′′, smaller than the observed one. The mis- Bird, A. J., Barlow, E. J., Bassani, L., et al. 2004, ApJ,
match is however not too big and could be explained by a 607, L33
differenceinobjectcentroids,dueforexampletothesource Bird, A. J., Malizia, A., Bazzano, A., et al. 2007, ApJS,
shape (as expected in an extended source like a galaxy) or 170, 175
tothesourcepropermotion(astypicalofagalacticsource). Chevalier, R.A., 1998, Mem. S. A.It., 69, 977
We therefore assume that the DENIS/2MASS source could Dean, A.J., et al., 2008, MNRAS,submitted
bethe near-infrared counterpart of AX J183039–1002. Fu¨rst, E., Reich,W., Sofue, Y.,1987, A&ASS,71, 63
Theoverallmorphologyofthesourcesuggeststwopos- Gaensler, B. M. & Slane, P. O., 2007, ARA&A,44, 17
sible scenarios: a PWN associated to the SNR or a back- Green, D. A., 1989, AJ, 98, 1358
ground AGN. In the first case the source would be a very Goldwurm, A., David, P., Foschini, L., et al. 2003, A&A,
unusual pulsar/PWN system from the radio and X/soft 411, L223
gamma-ray spectral characteristics. For example, the soft Hands,A.D.P.,Warwick,R.S.,Watson,M.G.,&Helfand,
X-rayspectralindexofaPWNisusuallyaround1.5–2(Kar- D. J., 2004, MNRAS,351, 31
galtsev&Pavlov2008),softerthantheASCAandChandra Helfand, D. J., Becker, R. H., White, R. L., Fallon, A., &
spectra. Furthermore, the photon index Γ > 2.8 observed Tuttle, S., 2006, AJ, 131, 2525
in the IBIS band is significantly higher than the usual val- Kargaltsev,O.&Pavlov,G.G.,2008,AIPConferencePro-
uesfortheINTEGRALPWNe,thatclusteraroundΓ∼2.2 ceedings, 983, 171
(Dean et al. 2008). The alternative possibility that we are Krivonos, R., Revnivtsev, M., Lutovinov, A., et al. 2007,
dealing with a magnetar is disregarded on the basis of the A&A,475, 775
stable 2–10 keV flux over the years and the shape of the Kuiper, L., Hermsen, W., den Hartog, P. R., Collmar, W.
high energy emission (typically IBIS sees a hard tail in this 2006, ApJ,645, 556
typeof sources, Kuiperet al. 2006). La Franca, F., Fiore, F., Comastri, A., et al., 2005, ApJ,
In the second, more likely, case we would deal with 635, 864
a heavily absorbed AGN (a type-2 Seyfert?) with an ex- Maiolino, R., Salvati, M., Bassani, L., et al. 1998, A&A,
tremely low exponential cut-off energy. The radio flux and 338, 781
spectral index would be compatible with the range of val- Masetti, N.,Mason,E.,Morelli, L.,etal. 2008,A&A,482,
ues found in other INTEGRAL selected AGN, which are 113
almost invariablyassociated withradiosources(Mushotzky Matt, G., Guainazzi, M., Frontera, F., et al. 1997, A&A,
2004). In this case, the spatial coincidence between a back- 325, L13
ground AGN and a SNR is fortuitous. But the more con- Matt, G., Bianchi, S., Guainazzi, M., Molendi, S., 2004,
clusiveevidencefor theassociation with aheavilyabsorbed A&A,414, 155
AGNcomesfromtheXMM-Newton spectrum,thatshowsa Mushotzky,R., ASSL,308, 53
strongironlineandisreminescentofaCompton-thickAGN. Predehl, P. & Schmitt, J. H. M. M. 1995, A&A,293, 889
6 L. Bassani et al.
Price,S.D.,Egan,M.P.,Carey,S.J.,etal.2001,AJ,121,
2819
Stephen,J. B., Bassani, L., Malizia, A., et al. 2006, A&A,
445, 869
Sugizaki, M., Mitsuda, K., Kaneda, H., et al. 2001, ApJS,
134, 77
Ubertini, P., Lebrun, F., Di Cocco, G., et al. 2003, A&A,
411, L131
Watson,M.G.,Schr¨oderA.C.,Fyfe,D.,etal.2008,astro–
ph(0807.1067)
White, R. L., Becker, R. H., & Helfand, D. J. 2005, AJ,
130, 586
Winkler,C.,Gehrels,N.,Sch¨onfelder,V.,etal.2003,A&A,
411, L355
