X-RAY EMISSION FROM MULTI-PHASE SHOCK IN THE LARGE MAGELLANIC CLOUD ...

22 Pages·2016·0.44 MB·English

ACCEPTEDFORTHEPUBLICATIONINTHEAstrophysicalJournal
PreprinttypesetusingLATEXstyleemulateapjv.25/04/01
X-RAYEMISSIONFROMMULTI-PHASESHOCKINTHELARGEMAGELLANICCLOUDSUPERNOVA
REMNANTN49
SANGWOOKPARK1,DAVIDN.BURROWS,GORDONP.GARMIRE,ANDJOHNA.NOUSEK
DepartmentofAstronomyandAstrophysics,PennsylvaniaStateUniversity,525DaveyLaboratory,UniversityPark,PA.16802
JOHNP.HUGHES
DepartmentofPhysicsandAstronomy,RutgersUniversity,136FrelinghuysenRoad,Piscataway,NJ.08854-8109
AND
2
ROSAMURPHYWILLIAMS
0
DepartmentofAstronomy,B619E-LGRT,UniversityofMassachusetts,Amherst,MA01003
0
AcceptedforthepublicationintheAstrophysicalJournal
2
ABSTRACT
v
o Thesupernovaremnant(SNR)N49intheLargeMagellanicCloudhasbeenobservedwiththeAdvancedCCD
N ImagingSpectrometer(ACIS) on boardthe ChandraX-RayObservatory. The superbangularresolutionof the
Chandra/ACIS images resolves a point source, the likely X-ray counterpart of soft gamma-ray repeater SGR
5
0526−66,andthediffusefilamentsandknotsacrosstheSNR.Thesefilamentaryfeaturesrepresenttheblastwave
2
sweeping through the ambient interstellar medium and nearby dense molecular clouds. We detect metal-rich

ejecta beyondthe main blast wave shock boundaryin the southwestof the SNR, which appear to be explosion
1
v fragments or “bullets” ejected from the progenitor star. The detection of strong H-like Si line emission in the
2 easternside oftheSNRrequiresmulti-phaseshocksin ordertodescribethe observedX-rayspectrum,whereas
5 suchamulti-phaseplasmaisnotevidentinthewesternside. ThiscomplexspectralstructureofN49suggeststhat
5 the postshock regionstoward the east of the SNR mighthave been reheated by the reverse shock off the dense
1 molecular clouds while the blast wave shock front has decelerated as it propagatesinto the dense clouds. The
1 X-rayspectrumofthedetectedpoint-likesourceiscontinuumdominatedandcanbedescribedwithapowerlaw
2 ofG ∼3. Thisprovidesaconfirmationthatthispoint-likeX-raysourceisthecounterpartofSGR0526−66inthe
0
quiescentstate.
/
h
Subjectheadings:ISM:individual(N49)—supernovaremnants—X-rays:ISM—X-rays:individual(SGR
p
0526−66)—stars: neutron—X-rays:stars
-
o
tr 1. INTRODUCTION clouddensityis∼0.9cm−3. ThisdenseISMismostlikelyas-
s sociated with the CO clouds detected in the east of the SNR
a N49isabrightsupernovaremnant(SNR)intheLargeMag-
: ellanic Cloud (LMC) (Dopita1979; Longetal. 1981). The (Banasetal. 1997). The bright optical/UV emission arises
v from “sheets” of emission generated by slow radiative shocks
Xi low metal abundances of N49 (Danziger&Leibowitz1985; ofv≤140kms−1astheblastwaveinteractswiththedenseam-
Russell&Dopita1990) indicated a substantial mixture of
bientmaterial, whiletheX-rayemissionappearsfromthefast
r the supernova (SN) ejecta material with the LMC interstel-
a shock(v∼730kms−1)propagatingintheintercloudmedium.
lar medium (ISM), which is consistent with the relatively
TheX-rayspectrumofN49asobtainedwithASCAwasfit-
old dynamical age as derived to be ∼5000 − 10000 yr
tedwith a non-equilibriumionization(NEI)modelwith kT ∼
(Longetal. 1981; Vancuraetal. 1992). The most remarkable
0.6keV and n t ∼ 1011 cm−3 s (Hughesetal. 1998). Thefit-
featureofN49isstrongenhancementsoftheoptical/X-raysur- e
ted electron temperature and the ionization age are consistent
face brightness in the south-east. This asymmetric intensity
with the Sedov evolution without significant alteration of the
distribution has been attributed to the interaction of the blast
preshock ISM by the stellar winds from the progenitor (the
wave with the nearby molecular clouds located in the east of
model suggested by Shull et al. [1985]), which supports the
the SNR (Vancuraetal. 1992; Banasetal. 1997). The radio
overallpictureproposedbyVancuraetal. (1992). Theoverall
images show similar intensity distribution to the X-ray mor-
X-rayspectrumwasfittedwithlowmetalabundances,whichis
phology, but with a complete shell in the west as well as in
the east (Dickeletal. 1995). The radio spectral index of a = alsoconsistentwiththeoptical/UVspectralanalysis.
The positionof the outburstof the softgamma-rayrepeater
−0.58 (Dickel&Milne1998) is typical for a shell-type SNR.
(SGR) SGR0526−66in1979March5 hasbeenknownto be
Althoughthemeanfractionalpolarizationoverthewholerem-
within the boundary of N49 (Clineetal. 1982) and is consis-
nantis low (<5%), the western side of the SNR is highlypo-
tent with the unresolved X-ray source RX J05260.3−660433
larized. Thehighlyorganizedmagneticfieldinthewestofthe
(Rothschildetal. 1994), suggesting the identification of this
SNR is supportive of the less-disturbed blast wave expanding
sourceas theX-ray counterpartof SGR 0526−66. Thisunre-
intothelowdensityISMinthewest(Dickel&Milne1998).
solvedX-raysourceappearsspatiallyassociatedwithSNRN49
Vancuraet al. (1992)haveperformedextensiveoptical/UV
basedonitspositionalcoincidencewithintheboundaryofthe
spectroscopyandhavefoundthatN49isinteractingwithdense
SNR although an incidental alignment along the line of sight
clouds(n ∼150cm−3 onaverage)while thepreshockinter-
H
[email protected]
1
2 Park
betweenthetwocannotberuledout(Gaensleretal. 2001).Al- ingwastowardthepositionofSGR0526−66(a =05h26m
2000
thoughan8spulsationfromSGR0526−66hasbeendetected 00s.7, d = −66◦ 04′ 35.′′0). In this observation, instead of
2000
during the g -ray burst in 1979 (Mazetsetal. 1979), the X-ray the whole CCD and the standard 3.2 s frame time, a 128-row
timing/spectral analysis of this unresolved X-ray source has subarray with a 0.4 s frame time was used for the temporal
beenlimitedwithpreviousdetectors(e.g.,Marsdenetal.1996). study of the target object SGR 0526−66. Due to the selec-
No counterparts in the infrared and radio bands have been tionofthesubarrayoftheCCDchip,thecoverageoftheSNR
foundforSGR0526−66(Dickeletal. 1995;Smithetal. 1998; was incomplete and only the northern parts of the SNR were
Fenderetal. 1998). These results from SGR 0526−66 share observed. The total exposure was ∼40 ks. Although the an-
typicalcharacteristicsoftheAnomalousX-rayPulsars(AXP), gularcoverageisincomplete,thison-axisobservationprovides
provided that the detection of the X-ray pulsar is confirmed. better angular resolutions than our GTO data, thus we utilize
High resolutionX-ray imagingwith the moderatespectral ca- thesearchivaldataassupplementsintheanalysis. Wehavere-
pability of Chandra observationswould be critical to confirm processedtherawarchivaldatainexactlythesamewayasthe
thedetectionoftheX-raycounterpartofSGR0526−66. GTOdatawereprocessed,includingtheCTIcorrection.
Weherereporttheresultsfromtheimagingandspatiallyre-
solvedspectralanalysisofSNRN49fromtheChandraobser- 3. ANALYSIS &RESULTS
vations. While this work is primarily devoted to the analysis
3.1. X-rayColorImages
of the diffuse X-ray emission from SNR N49, we also report
somepreliminaryresultsfromthedetectedpointsource,theX- The“true-color”X-rayimagesofSNRN49fromtheChan-
raycounterpartofSGR0526−66.Wedescribetheobservation dra/ACISobservationsarepresentedinFigure1. Figure1ais
andthe datareductionin §2. Theimageandspectralanalyses theimagefromtheGTOobservationandFigure1bistheone
arepresentedin§3,andadiscussionin§4. Asummarywillbe fromthearchivalGOdata. Thedifferencesinthepointspread
presentedin§5. function (PSF) between two images due to the angular offset
of∼6.′5fromtheaimpointintheGTOdataareapparent. The
highresolutionon-axisarchivaldatarevealthedetaileddiffuse
2. OBSERVATIONS &DATA REDUCTION
filamentary structure, particularly in the east of the SNR, far
SNR N49 was observed with the Advanced CCD Imaging better than the off-axis GTO data. The detection of the point
Spectrometer(ACIS)onboardtheChandraX-RayObservatory source as the counterpart of SGR 0526−66 is evident in the
on2001September15aspartoftheChandraGuaranteedTime middle of the northern part of the SNR as shown by a bright
Observation(GTO)program(ObsID1041).TheACIS-S3chip whitespot(Figure1b).Thispointsourceisalsodetectedinthe
waschosentoutilizethebestsensitivityandenergyresolution off-axisimage(Figure1a)butappearsextendedbecauseofthe
of the detectorin the softX-raybands. Thepointing(a = PSFdegradation.Despitethesuperiorangularresolutionofthe
2000
05h 25m 26s.04, d = −65◦ 59′ 06.′′9) was roughly toward on-axis data, the selection of the subarray of the CCD allows
2000
theX-raycentroidofthenearbyLMCSNRN49Bwiththeaim only a partial coverage of the SNR: i.e., about a third of the
pointshiftedby245detectorpixels(∼2′)fromthenominalS3 SNRinthesouth-westwasnotcovered(Figure1b).Theimag-
chipaimpointinordertosimultaneouslyobservebothofN49 ing and spectral analysis overthe whole remnantis important
and N49Bon the same CCD chip. This pointingdirectionre- to understand the global nature of the SNR and we primarily
sultsinanangularoffsetof∼6.′5fromtheaimpointatthepo- utilizeourGTOdataforthedataanalysis. Inordertotakead-
sitionofN49. Theresultsfromthedataanalysisofthepointed vantageofthehighangularresolution,weusethearchivaldata
targetSNRN49Barepresentedelsewhere(Parketal. 2002b). forsupplementarypurposes.
We have utilized the data reduction techniques developed InFigure1,thesoftX-rayemission(red-yellow)revealsvar-
at the Pennsylvania State University for correcting the spa- ious filamentary structures across the SNR. The circumferen-
tial and spectral degradation of the ACIS data caused by ra- tialshell-likesoftX-rayfeaturesaroundtheSNRarespatially
diation damage, known as charge transfer inefficiency (CTI) coincident with the radio shells (Figure 2a), representing the
(Townsleyetal. 2000). The expected effects of the CTI blast wave shock front. By comparison, the hard X-ray emis-
correction include an increase in the number of detected sion features show little correlation with the radio intensities
events and improved event energies and energy resolution (Figure2b). Basedonastrophysicalconsiderations,webelieve
(Townsleyetal. 2000; Townsleyetal. 2002a). We screened thatitisreasonabletosupposethatshocksarereflectingoffthe
the data with the flight timeline filter and then applied the densermoleculargasintheeastandmaybepropagatingback
CTI correction before further data screenings by status and intotheinteriorofN49. Thiscouldbeapartialexplanationfor
grade. Weremoved“flaring”pixelsandselectedASCAgrades the highertemperatureswe find in the spectralanalysis of the
(02346). The overall lightcurve was examined for possible easternregion(§3.3.2),althoughadefinitiveconclusionwould
contamination from time-variable background. The overall require a detailed investigation of the hydrodynamicsof N49
lightcurvefromthewholeS3chipshowsrelativelyhighback- that is beyond the scope of this work. Other evidence, albeit
ground(byafactorof>2)forthelast∼10%oftheobservation morespeculative, in supportof reflected shocksin the eastern
∼
time. Thiscontaminationfromthevariablebackgroundishow- region of N49 include (1) the morphologyof the hard X-rays
evernegligiblefortheN49sourceregion(∼1.′4diameter)and whichpeakclosertotheinteriorontheeasternsidethanthesoft
we ignore it in the data analysis. After these data reduction X-raysdo(Figure2),and(2)thecleararc-likefeatureintheSi
process,theeffectiveexposurewas34ksthatcontains∼241k EWmap(§3.2andFigure5f)whichtraceshotgas,thatperhaps
photonsinthe0.3−8keVbandforthewholeremnantwithin delineatestheedgeoftheinwardmovingreflectedshock. The
a1.′4diametercircularregion. western half is, for the most parts, dominated by hard X-ray
DuringChandracycle1(2000January4),N49wasobserved emission(green-blue),whichmayindicatethattheblastwave
withtheACIS-S3asaGuestObserver(GO)program,whichis ispropagatingthroughalowdensityISM,assuggestedbyradio
nowpubliclyavailableinthearchive(ObsID747). Thepoint- data(Dickel&Milne1998).
N49 3
The brightestdiffuseemission is in the southeast, as recog- photonstatisticsneartheedgeoftheSNR,wehavesettheEW
nizedbypreviousX-rayobservations,wheretheSNRisinter- valuesto zerowherethe estimatedcontinuumflux islow. We
acting with dense molecular clouds. The overallbrightX-ray alsosettheEWtozerowheretheintegratedcontinuumfluxis
emissionintheeasternhalfoftheSNRissignificantinallcol- greaterthanthelineflux.WepresenttheEWimagesforO,Ne,
ors and likely has complex spectral structure. While filamen- Mg, Si, andFe (Figure5). TheSEWimageis unreliabledue
tarysoft,red-yellowemissionprevailsintheeasternhalfofthe tothelowphotonstatistics. Thebroadlinecomplexesat∼0.8
SNR, there are apparentblue regions in the east (primarily to −1.1keVarefromvariousionizationstatesofNeKandFeL-
the immediate north of the southeast brightest region; Figure shell line emission, and we present tentative identifications in
1).ThisdiffuseX-rayemissionwithspectralvariationsappears Figure5andTable1. AlthoughMgandSilinesshowHe-like
localizedtoproducespatiallydistinctivelayersofX-raycolors andH-likesubstructuresintheoverallspectrum(Figure4),we
in the eastern half of the SNR. The interactions of the blast presenttheEWimagesfromthecombinedbroadlinesinFig-
wave shock with the clumpy dense molecular clouds in these ure 5 since the H-like lines of these elements are too weak to
regionsappearresponsiblefortheobservedspatio-spectralsub- separatelyproduceusefulEWimagesoverthewholeSNR.
structuresofthediffuseX-rayemission(see §3.3). Forexam- The O EW distribution is generally consistent with the op-
ple,thebrightfilametaryemissionfeaturesfromtheopticalOIII tical OI,III intensity distributionsand shows higherEW in the
imagesarelargelyanti-correlatedwiththeblue,hardX-rayin- east and south of the SNR (Figure 5a; Vancura et al. 1992).
tensity and more correlated with the soft, red X-ray emission The Fe EW (Figure 5d) shows similar distribution to the O
(Figure 3a and Figure 3b). The optical emission from N49 map. Ne/Fe and Mg EW (Figure 5b, Figure 5c, and Figure
is believed to originate primarily from slow radiative shocks 5e) distributions appear to be opposite to O, showing higher
entering dense molecular clouds, while the X-ray emission is intensities in the west of the SNR. The Si EW is relatively
ingeneralfromfastshockspropagatingthroughtheintercloud strong in both east and west (Figure 5f). The O and Fe EW
medium (Vancuraetal. 1992). The high resolution X-ray im- enhancementsare spatially correlated with the location of the
agesfromtheACISdatarevealthatthehardX-raysareindeed dense molecularclouds where the X-ray and optical emission
fromintercloudregionswherelittleopticalemissionoriginates, is generally bright (Figure 5a and Figure 5d). Since the X-
and that the soft X-ray emission is spatially more coincident ray emission in these regionsis presumablydominatedby the
withthebrightopticalemittingregions.Thesefeaturessupport shockeddenseISM,thesehighOandFeEWsareunlikelydue
theinteractionsoftheblastwavewiththecloudsastheorigin to abundance enhancements, but appear to trace shock tem-
oftheX-rayspectralvariationsobservedinthecolorimages. perature variations. Optical spectroscopic studies have indi-
Wedetectaninteresting,previouslyunknownfeatureofN49 catedthatmetalabundancesofN49areclosetotheLMCval-
withourChandraGTOdata: i.e.,theblue,hardX-rayknotin ues (Russell&Dopita1990; Vancuraetal. 1992) and support
the southwest, beyondthe main boundaryof the SNR (Figure temperature variations, rather than abundances, as the origin
1a),fromwhichfaintradioemissionisalsoassociated(Figure of these EW distributions. Based on the plane-parallel shock
2b). Theextendedshapeofthisfeatureappearscausedbythe modelwithamoderateionizationtimescale(n t ∼1012 cm−3
e
off-axisPSFandthetruemorphologicalnatureofthisemission s)andtheLMCabundances,theFeLEW(Figure5d)isinfact
isuncertainwiththecurrentoff-axisdata.Wecannotutilizethe representativeofthebroadlinecomplexatE ∼0.8keV,which
high resolution archivaldata, because this feature lies outside ischaracteristicofa“cool”plasma(kT <0.5keV)asopposed
∼
the field of view (Figure 1b). We find no proper counterparts to the “hot” component (kT ∼ 1 keV) that would produce a
for this feature in other wavelengths, which suggests that it is prominentFeLlinecomplexatE ∼1keV.
unlikelyaprojectionofabackground/foregroundpointsource Indirectcontrast,theH-likeNe(and/orhotFeL)EW(Fig-
(§3.3.3). The X-ray spectrum of this feature indicates metal- ure5c),whichissensitivetohightemperaturesofkT ∼1keV,
richthermalemissionwithstrongatomicemissionlinesrather isenhancedinthewestwheretheblastwaveisexpandingintoa
thanapowerlawdominatedspectrum(§3.3.3). Thesespectral lowdensityISM.TheseEWdistributionsmaythusindicatethe
features supportthe propositionthat this hard X-ray knot is a presenceofanundisturbed(bydenseISM)blastwaveshockin
partoftheSNR,probablyfragment(s)ofSNejecta. thewest,whichmayhaveahighertemperaturethanthatinthe
easternsideoftheSNR.Ifthedominantcontributortotheen-
hancedEWatE∼1.1keVisNe,comparisonsbetweenHe-like
3.2. EquivalentWidthImages
andH-likeNeEWimages(Figure5bandFigure5c)maysug-
TheoverallspectrumofN49showssomebroadatomicemis- gestprogressiveionizationsof Ne by the reverseshockas de-
sion lines (Figure 4) and we generate equivalent width (EW) tected in other SNRs (Gaetzetal. 2000; Flanaganetal. 2001;
images of K-shell line emission from the atomic elements O, Parketal. 2002a). The Si EW enhancementsin both eastand
Ne,Mg,Si,S,andtheL-shelllineemissionofFebyselecting westareintriguingandmaysuggestthepresenceofhotphases
photonsaroundthe broadline complexes(Table 1), following in addition to cool phase shocks in the south-eastparts of the
the methods described by Park et al. (2002a). Images in the SNR(seemorediscussionin§3.3). TheSiEWenhancements
bandpasses of interest for both the lines and continuum were inthesoutheasternquadrantarelargelyshell-like,asindicated
extractedwith2′′ pixelsandsmoothedbyaGaussianwiths = by a dotted line in Figure 5f. This may also be interpretedas
7′′. The underlying continuum was calculated by logarithmi- hot, highly ionized plasma in the southeastern regions of the
cally interpolating between images made from the higher and SNR.
lower energy “shoulders” of each broad line. The estimated The hard X-ray knot in the southwest beyond the SNR
continuumfluxwasintegratedovertheselectedlinewidthand boundary,asrevealedwiththetrue-colorimage,showsstrong
subtracted from the line emission. The continuum-subtracted enhancements in the Si, Mg, and hot Ne/Fe EW images.
lineintensitywasthendividedbytheestimatedcontinuumon The actual spectralanalysis indicates thatthese high EWs are
apixel-by-pixelbasistogeneratetheEWimagesforeachele- caused by metal abundances as well as advanced ionizations
ment. InordertoavoidnoiseintheEWmapscausedbypoor
4 Park
andhightemperature(see§3.3.3). The enhancedEWs forthe insignificantintheselectedenergyrange.Otherelementswere
elementalspeciescausedbyhighmetalabundancessupportthat allowedtovary. Theresultsofthespectralfittingsforthewest
this hard X-ray knot may not be a background extragalactic regionsoftheSNRaresummarizedinTable2andTable3.
source,andthatitislikelyemissionfrommetal-richSNejecta These spectra in the western half of the SNR can be de-
associatedwithN49.TheEWaroundthedetectedpointsource scribed by a thermal plasma with the best-fit kT of 0.52 −
is,ontheotherhand,verylowforallelementalspecies. These 0.65keV.Theimpliedabsorptionsare∼0.6−2.7×1021cm−2.
low EWs are strong indications of the continuum-dominated TheseresultsareconsistentwiththeresultsfromtheASCAdata
X-ray spectrum for the point source. This is in good agree- (Hughesetal. 1998). Thefittedabundancesaregenerallycon-
ments with the presumed identification of the point source as sistentwiththeEWdistributions.Theresultsfromtheregion2
theneutronstar, ortheX-raycounterpartofSGR0526−66in spectrumareperhapsthemostinteresting.ThisiswhereNe/Fe,
thequiescentstate. Mg, and SiEWs are enhancedand the best-fitabundancesfor
thoseelementalspecies(exceptforFe)arealsorelativelyhigh,
3.3. X-raySpectra although the magnitudes of the enhancements are moderate
only at around the solar level. The fitted abundances for this
The high resolution ACIS data, for the first time, allow us
region suggest that the high EW responsible for the emission
toperformspatially-resolvedspectralanalysisinX-raysforthe
lineatE ∼1.1keV(Figure5c)isduetohighlyionizedH-like
study of SNR N49. Despite the PSF degradations due to the
Ne ratherthan hotFe. The best-fit shocktemperatureand the
largeangularoffsetfromtheaim point,theGTOdata provide
ionizationtimescaleforthisregionarehigherthanregion1in
moderateangularresolutionofafewarcsec,whicharestillsuf-
whichHe-likeNe EWis enhanced(Figure5b). Thissupports
ficient to resolve major diffuse structure of N49. Moreover,
theadvancedionizationinregion2assuggestedbytheEWim-
thecompletespatialcoverageoftheSNRwiththeGTOdatais
ages.Theregion3spectrumisontheotherhandfittedwithlow
advantageousoverthearchivaldatafortheefficientandconsis-
LMC-likeabundancesforallelements. ThebroadbandX-ray
tentstudyoftheglobalstructuresoftheSNR.Combiningtwo
surfacebrightnessisalsolow forthisregion,indicatingX-ray
datasetsmaynotbedesirablebecauseofthesignificantdiffer-
emissionfromtheshockedlowdensityLMCISM.
ence in the PSF. We thus make a simple and secure approach
forthespatially-resolvedspectralanalysisbyutilizingtheGTO
3.3.2. SNREast: Multi-PhaseShock
data as the primary data and the archival data as the supple-
mentary data: i.e., we perform all spectral analysis with the We selecttworegionsfromtheeasternhalfoftheSNR(re-
GTO data and thenverifythe resultswith the archivaldata as gions4and5;Figure6). Region4iswithinthebrighteastern
necessary. Weextractthespectrumfromseveralsmallregions rimwhichisevidentlydominatedbythehardX-rayemissionas
overtheSNR.Theregionalselectionswerebasedonthechar- displayedinFigure1. Atotalof∼10000photonsareextracted
acteristicsrevealedbythebroadbandcolorimagesandtheEW fromregion4. Region5isthebrightestknotinthe southeast,
images, and are presented in Figure 6. In all spectral fittings, where we extract ∼17000 counts. We fit these spectra in the
wemadealowenergycutatthephotonenergyofE =0.5keV same way as we did for the western regions: i.e., we use the
becauseofunreliablecalibrationsoftheACISspectruminthe plane-parallelshockmodelwithO,Ne,Mg,Si,S,andFeabun-
softbands.Inmostcases,webinthedatatocontainaminimum dancesvariableinthefittings. Bothofthespectraintheeastern
of20countsforthespectralfitting. Forthespectralanalysisof half of the SNR show significant Si Lya line (E ∼ 2.0 keV)
ourCTIcorrecteddata,wehaveutilizedtheresponsematrices as well asstrong softX-ray emission (E < 1 keV) (Figure7d
appropriatefor the spectralredistributionof the CCD, asgen- and Figure 7e). In orderto adequatelydescribe these spectral
erated by Townsley et al. (2002b). The low energy (E < 1 features, a hotcomponentshockwith kT > 1 keV is required
∼
keV)quantumefficiency(QE)oftheACIShasbeendegraded in addition to the soft component with kT ∼ 0.4 − 0.5 keV.
becauseofthemolecularcontaminationontheopticalblocking The hot components have larger ionization timescales (n t >
e ∼
filter. We have corrected this time-dependentQE degradation 1012 cm−3 s) than the cool components(n t ∼ 1011 cm−3 s).
e
by modifying the ancillary response function (ARF) for each
The fitted ionizationtimescales forthe hotcomponentare not
extracted spectrum by utilizing the IDL version of ACISABS
constrained in the upper bounds and a thermal plasma in the
softwaredevelopedbyChartasetal. (2002).
complete collisionalionization equilibrium(CIE) may also fit
thedatawiththeequivalentstatisticalsignificance.Theimplied
3.3.1. SNRWest: Single-PhaseShock absorbingcolumnsareN ∼3−4×1021 cm−2 forthesere-
H
WeselectthreeregionsfromthewesternhalfoftheSNR(re- gions.
gions1,2,and3;Figure6). Theseregionsarewherethebroad The soft component model for the region 5 spectrum indi-
bandsurfacebrightnessisrelativelylow,whiletheEWimages cateslowmetalabundancesof<0.5solarforfittedelemental
∼
revealsomenoticeablefeatures(Figure5andFigure6).Region species. ThissoftcomponentdominatestheX-rayemissionat
1 is where O and Ne EWs are enhanced and contains ∼3500 E < 1 keV, which consistently supports that the enhanced O
∼
photons. Region 2 is where Ne/Fe, Mg and Si are enhanced andFeEWs(Figure5aandFigure5d)arenotcausedbyhigh
containing∼8100counts.Region3showslowEWformostof abundances. The best-fit model for the hard component im-
the elements and contains∼3200 photons. We fit these spec- pliesoverabundances(typically∼1−2solar)comparedwith
trawiththeplane-parallelshockmodel(Borkowskietal. 2001) the LMC abundancesfor some elemental species. The statis-
as presented in Figure 7a, Figure 7b, and Figure 7c. The ele- tical uncertainties for the fitted abundancesare however large
mental abundanceswere fixed for He (= 0.89), C (= 0.30), N enoughthattherealityofthesehighabundancesareinconclu-
(= 0.12), Ca (= 0.34) and Ni (= 0.62) at values for the LMC sivewiththecurrentdata. Forexample,thisregion5spectrum
(Russell&Dopita1992)(hereafter,allabundancesarewithre- may also be fitted with subsolar abundances, for both of the
specttothesolar;Anders&Grevesse1989)sincethecontribu- softandhardcomponentsofthe shock,withequivalentstatis-
tion fromthese speciesin thespectralfitting isexpectedto be tical significance based on the F-test. The EWs distributions
N49 5
in theeastoftheSNRarethereforemorelikelydominatedby Thesourcepositionswithouroff-axisdatathereforeappearto
the shock temperature and the ionizationstates rather than by have<4′′ accuracyas a conservativelimit. We have searched
metalabundances. The fitted modelparametersforthe region for counterparts around the source position in SIMBAD and
4 spectrum are also generallyconsistent with the results from NED catalogs as well as ∼100 multi-wavelengh on-line cata-
region 5, which also supports the temperature and ionization logs available through HEASARC. We find no candidate for
states for the origins of the observed EW distributions. We an extragalacticcounterpartwithin 1′ radiusof the source po-
note that region 4 EWs appear relatively weak for the atomic sition. The derived X-ray flux is f ∼ 6.0 × 10−14
X(0.5−2keV)
lines at low photonenergies(O and Fe; Figure 5a and Figure ergss−1 cm−2 and f ∼ 1.5× 10−14 ergss−1 cm−2.
X(2−10keV)
5d) and stronger in higher energies (Si; Figure 5f) compared BasedonthelogN−logSrelationsoftheChandraDeepField
withregion5,eventhoughthereisnosignificantdifferencein (Brandtetal. 2001), these X-ray fluxes imply the probability
the fitted shock parametersbetween these two regions. These ofonly<10−4 withina generousareaof5′′ radiusaroundthe
∼
features may imply a difference in the local interstellar envi- sourcepositionforacoincidentaldetectionofanextragalactic
ronmentsdue to the clumpy structure of the interacting dense object.
molecularclouds:i.e.,adensitygradientofthecloudsbetween We find an LMC star, MACS J0525−660#037
thesetworegionsmighthavecauseddeficientsoftX-rayemis- (Tucholkeetal. 1996) which has an angular offset from the
sion in region4 as displayedby the lack of red, softemission positionofthehardX-rayknotof∼6.′′2. Consideringthe po-
there(Figure1). Theanticorrelationsbetweenopticalandhard sitional accuracy of the MACS catalog (∼0.′′5; Tucholke et
X-ray emission are supportiveof this case. The fitted volume al. 1996), the apparentangularoffsetbetween the hard X-ray
emission measure (EM) of the soft componentplasma for the knotand opticalposition of the MACS star is significant. We
region 5 spectrum is indeed significantly higher than that for find another nearby LMC star within the OB association LH
region4,whichalsosupportsthisinterpretation.Theresultsof 53 as cataloged (ID # 1630) by Hill et al. (1995). The rela-
thespectralfittingsfortheeastregionsoftheSNRaresumma- tively small angular offset (∼2′′) of this star from the X-ray
rizedinTable4. hard knot position suggests that this star could be the optical
counterpart of the X-ray knot. The derived X-ray luminosity
3.3.3. ExtendedFeatureBeyondtheShockBoundary of LX ∼ 3.4 × 1034 ergs s−1 (at the distance of 50 kpc) for
theX-rayknotis howeverhigherthanthatoftypicallate-type
The hard X-ray knot in the southwest beyond the main youngstellarobjects(L ∼1028−1030ergss−1;e.g.,Imanish
boundary of the SNR (region 6 in Figure 6) contains ∼800 X
et al. 2001) by a few orders of magnitude. A massive early-
counts and we binned the spectrum to contain a minimum of
typestarshouldbebrighter(L ∼1032 ergss−1; e.g.,Seward
12 counts per bin. The spectrum can be fitted with a single- X
&Chlebowski1982),butmayalsonotmatchthederivedhigh
temperature plane-parallel shock (Figure 7f). The spectrum
L ofthe hardX-rayknot. Moreover,althoughthereare ∼80
featuresstrongatomic emission lines and is mostlikely emis- X
LH 53 stars within 1′ radius of this hard X-ray knot position
sion from metal-rich ejecta. The strong Si Lya line (E ∼ 2.0
(Hilletal. 1995), noneofthemis detectedwith ourACISob-
keV) constrains the plasma temperature to kT > 1 keV. Al-
∼ servationatthefluxlevelofthishardX-rayknot. Itisunlikely
thoughstatisticaluncertaintiesarelargeduetothepoorphoton
foronlyonestartobedetectedwiththeACISatthisparticular
statisticsofthisfaintfeature,thebest-fitmodelindicatesstrong
positionatthisfluxlevel.
enhancementsinmetalabundancesatseveraltimeslargerthan
If this object is a foreground young late-type Galactic star
solar abundances. Thebroadbandline featuresbetween∼0.8
with a typical X-ray luminosity of L ∼ 1028−1030 ergs s−1,
keVand∼1.2keVrepresentFeLlinecomplexaswellasHe- X
this object must be nearby with a distance of ∼30 − 300 pc,
likeand/orH-likeNeKlines. Withthephotonstatisticsofthe
andaVmagnitudeof∼11−14isexpectedfromthemeasured
current data, it is difficult to identify these individual atomic
X-ray flux. We have, however, found no such bright stars in
lines. Wehavethusattemptedfittingthespectrumbyassuming
Galactic star catalogs. We find a faint star in the Guide Star
dominantcontributionsfromNeorFeaswellasfrombothele-
Catalog with an ∼2′′ offset from the source position, but the
ments. Theoverallfitsarestatisticallyindistinguishableamong
cataloged B magnitude is only 19.44. The best-fit absorbing
these cases. We howevernote that the modelfits Fe as domi-
column from the X-ray spectral modeling may also be higher
nantcontributor(Feabundance∼ 0.8,Ne∼0)whenbothNe
(N = 1.1 × 1021 cm−2) than the Galactic absorption toward
andFearesimultaneouslyallowedtovary.EvenwithFeabun- theHposition of this hard X-ray knot (N ∼ 6 × 1020 cm−2;
dancefixedatLMCvalue,thefittedNeabundanceisstilllow H
Dickey&Lockman1990).
(∼0.4). We thereforeobtainthebest-fitmodelbyallowingFe
This hard X-rayknotis thereforemost likely a feature spa-
abundanceto varywith Ne fixed at the LMC abundance. The
tiallyassociatedwithSNRN49.Thisfeaturethenappearstobe
fittedparametersaresummarizedinTable2andTable3.
fragments from the SN explosion ejected outside of the SNR
Althoughthe metal-richspectrumindicatesthatthisfeature
shock boundary or “shrapnel” as discovered in other Galac-
is likely emission from shocked SN ejecta, the isolated mor-
tic SNRs such as Vela and Tycho (Aschenbachetal. 1995;
phologyofthishardX-rayknotoutsideofthemainboundary
Vancuraetal. 1995;Warrenetal. 2002). Thetrueangularsize
of the SNR may suggest the possibility of a projected back-
and shape of this hard X-ray knot is however uncertain be-
groundobject. Thesourcepositionfortheknotisa =05h
2000 causeofthepooroff-axisPSF. We notethatanon-axisChan-
25m 52s.53, d = −66◦ 05′ 13.′′5, andtheACISastrometric
2000 dra/ACISobservationwithafullangularcoverageofN49,tar-
uncertaintiesat∼6′off-axisisgenerally<2′′(e.g.,Feigelsonet
∼ geted on SGR 0526−66has recently been performed(ObsID
al. 2002).TheChandra/ACISpositionofSGR0526−66based
2515). ThedetailedresultsfromthediffuseX-rayemissionof
on the on-axisdata has a conservativeastrometric uncertainty
theSNRhaveyettobereported.
of <2.′′3 (Kaplanetal. 2001andreferencestherein). Thepo-
∼
sitionoftheSGRofouroff-axisdataasdeterminedbywavde-
3.3.4. PointSource
tect algorithm is ∼1.′′8 offset from the on-axis SGR position.
6 Park
Thespectrumfromthepoint-likesourceisextractedfromre- south-eastandhardemissionlines(E >1keV)inthewestof
gion7(Figure6).Duetotheangularoffsetfromtheaimpoint, theSNR.Thesefeaturesareconsistentwiththeoveralldensity
the PSF at the position of the pointsource is significantly de- distributionsoftheambientISM;i.e.,theblastwaveispropa-
graded(>2′′),andweselectarelativelylargearea(acircularre- gatingintoalowdensityISMinthewestandinteractingwith
∼
gionwitharadiusof6′′),whichcontains∼14000photons.Al- dense cloudsin the east. The comparisonsbetweenthe X-ray
thoughthetotalcountrate(∼0.4countss−1)isrelativelyhigh, and optical images also suggest that the hard X-ray emission
pileupisinsignificantduetothelargeoff-axisangle. Sincethe (E >1.6keV)appearstobeproducedinless-denseintercloud
sourcelocationiswithinfairlybrightdiffuseemissionfromthe regionswhilethesoftX-rays(E<0.75keV)originatenearthe
SNR,thedegradedPSFalsomakesthebackgroundestimation denseclouds.
difficult. We thus include a thermal component for the back- Theresultsfromthespatially-resolvedspectralanalysiscon-
ground spectrum in the spectral fitting instead of subtracting firm the speculations inferred from the image analysis by re-
thebackgroundfromthesourcespectrum. vealingthatshocktemperaturesandionizationstatesintheeast
Inadirectcontrasttothediffuseemissionspectra,thepoint oftheSNRaresignificantlydifferentfromthoseinthewestern
source spectrum is dominated by continuum, as perhaps ex- half:i.e.,thewesternportionoftheSNRisdescribedbyasim-
pected from a neutron star (Figure 8a). The source spectrum ple one-temperature plane-parallel shock with relatively large
canbedescribedwithasinglepowerlawmodelofphotonindex ionization timescales while two-temperature model with mul-
G =2.98+0.17. Theembeddedbackgroundthermalcomponent tiple ionizationtimescales is requiredin the eastern half. The
is a kT ∼−00.2.303keV plasmawith LMC-likemetalabundances, global spectral structure of the SNR may then be represented
which is primarily responsible for the soft emission at E < 1 byamulti-phaseshockmodelwiththreecharacteristictemper-
keV.Ablackbodymodelmayalsofitthedata(kT ∼0.52keV) atures, i.e., the “cool” (kT ∼ 0.44+0.19 keV) and “hot”(kT ∼
−0.10
withacceptablestatistics,butappears,qualitatively,inappropri- 1.25+2.58 keV)componentsintheeast,andthe“intermediate”
−0.22
atetodescribethehardtailofthespecrum. (kT ∼ 0.59+0.09 keV) component in the west. The presence
Since the archival data are pointed in the direction of the −0.13
of multiple phases of the shock due to the interactions of the
position of the point source or SGR 0526−66, the on-axis
blastwavewiththedensemolecularcloudshasbeenputforth
highangularresolutionallowsustomanipulatethebackground
by Hester & Cox (1986). Recently, based on the high resolu-
subtraction with less confusion. We extract the source spec-
tion Chandra/ACIS observation of the Galactic SNR Cygnus
trum from a circular region with a radius of 2′′, which con-
Loop, Levenson et al. (2002) reported evidence for the three
tains∼10000photons.Becauseoftheutilizedshorttime-frame
distinct regions characterised as: (i) the forward shock decel-
(0.4s),pileupisnegligiblewiththison-axisobservation. SGR
erated by the clouds, (ii) the postshockSNR interior, and (iii)
0526−66isontopofcomplexfilamentarystructureofthedif-
thereflectedshockthatre-heatedthepostshockregions.Inthis
fuse emissionfromthe SNR andthebackgroundemissionfor
context, we propose that the cool component of X-ray emis-
theSGRisnon-uniform(Figure1b). Wethusextracttheback-
sion from N49 correspondsto the forward shock entering the
ground emission from the immediate annular region centered
denseclouds,theintermediatecomponenttoemissionfromthe
onthepointsourceposition(innerradiusof2.′′5 andouterra-
postshockregion,andthehotcomponenttoemissionfromthe
dius of 4′′), in orderto obtain an “average”spectrumoverthe
re-heated postshock region, respectively. Comparisons of the
variableemissionfeaturesaroundtheSGR.TheSGRspectrum
X-ray emission features with optical images indicate that the
from the on-axis data can be fitted with a single power law
softX-raysarespatiallycoincidentwiththebrightopticalknots
model(Figure8b). The best-fit photonindexis G = 2.72+−00..1110 whilethehardX-rayemissionisanti-correlatedwiththeopti-
which is in good agreement with the result from the off-axis cal intensity distribution. Since the bright filamentary optical
data. The results from the fit with a black bodymodel(kT ∼ emissionfeaturesarebelievedtooriginatefromtheslow,cool
0.50keV)arealsoidenticaltotheoff-axisdataandappearin- radiativeshockenteringdensemolecularclouds, the observed
adequatebythereasonofthesamehard-taildiscrepancy. The correlation/anti-correlationbetween X-ray and optical images
impliedX-rayluminosityisLX =2.1×1035 ergss−1 inthe2 provide good support for the suggested multi-phase shock in-
−10keVbandwiththeoff-axisdata(LX =2.5×1035ergss−1 terpretation.
basedontheon-axisdata),whichisconsistentwiththeASCA Hesteretal. (1994)havedevelopedasimple1-dimensional
upper limit (Hughesetal. 1998) and that of the Galactic SGR modelin orderto describe such a multi-phase shock structure
1806−20(Mereghettietal. 2000). The results from the spec- of the nearby Galactic SNR Cygnus Loop, probably a “mag-
tralmodelfitsofthepointsourcearesummarizedinTable5. nified” version of the distant SNR N49. Our best-fit electron
temperatures indicate that the ratios of temperatures between
4. DISCUSSION
the re-heated postshock and the postshock without re-heating
4.1. Multi-PhaseShockStructure are T /T = ∼1.8 − 2.3, which implies the preshock cloud
rh ps
TheX-raycolorimagesandtheEWimagesindicatecomplex density nc ∼ 30 − 2500 n0 based on the model by Hester et
spectralstructureofthediffuseX-rayemissionacrosstheSNR al. (1994), where n0 is the preshock intercloud density. As-
N49.Thefilamentaryandclumpyemissionfeaturesintheeast- sumingthepreshockinterclouddensityn0 ∼0.9cm−3 asesti-
ernhalfoftheSNRshowspatiallydistinctivebrightstructures matedfromtheopticalspectroscopy(Vancuraetal. 1992),we
indifferentcolors. ThespatialandspectralstructureoftheX- derivenc ∼ 27 −2300cm−3 forthe interactingdense molec-
rayemissioninthewesternhalfisrelativelyuniform,primarily ularclouds. This is in plausibleagreementswith the previous
representedbyintermediate-to-hardemission(green-blue)with estimationsbasedonoptical/UVspectroscopy(nc = 20−940
low surface brightness. The soft (red-yellow)X-ray shell-like cm−3; Vancura et al. 1992). The best-fit EMs in the east-
emission features represent the blast wave shock front as ob- ern regions imply (nc/n0)2(Vc/Vrh) = 2 − 13, where Vc is the
servedintheradioband.TheoverallfeaturesoftheEWimages X-ray emission volume of the cool forward shock component
indicateenhancementsofsoftlineemission(E<1keV)inthe andVrhistheX-rayemissionvolumeofthehotre-heatedcom-
N49 7
ponent(we assume an X-ray emitting volume filling factor of implies n of 2.3 cm−3 to > 68 cm−3 depending on the as-
e ∼
∼1, hereafter). Provided that n /n is in the orderof ∼101 − sumedemissionvolume. Basedonthebest-fitabundances,we
c 0
103, thisimpliesV /V <0.01. ThesoftX-rayemissionfrom may derive the ejecta mass of ∼0.001 M − 0.03 M . Al-
c rh ∼ ⊙ ⊙
the cool forward shock is then from the blast wave entering thoughthisrangeoftheejectamassisinreasonableagreements
small-volume, clumpy clouds while the hot componentorigi- withotherGalacticSN ejectafragments(Tsunemietal. 1999;
natesfromthelarge-volume,re-heatedintercloudpostshockre- Aschenbachetal. 1995; Strometal. 1995), we caution that
gions. Foranadiabaticshock,therelationsbetweentheshock theseestimationsarecrudeduetothelargeuncertaintiesinthe
temperatureandtheshockvelocityisT = 3m¯v2s andcanbewrit- fittedabundancesandassumedemissionvolumes.
s 16k TheangularoffsetofthehardX-rayknotfromthegeomet-
tenas
ricalcenterofN49is∼48′′. Thespatialseparationofthisknot
kTs=0.013(vs/100kms−1)2 keV, from the center of the SNR is then D = 11.6 d (sinq )−1 pc,
50
wmheearneaTtsomisitchemsahsoscmk¯ t≈em0p.7ermatur(em,vs=isptrhoetosnhomcakssv)e.loTchitey,caonodl wq hisertehed5a0ngislethbeetwdiesetanntcheetloinNe4o9f sinighthteanudnitthseomf 5o0vinkgpcd,iraencd-
p p
componentshockvelocityisthenv =582+114kms−1,thehot tion of the knot. Assuming that the SN explosion site, so the
c −71 birthplaceoftheobservedejectafragments,isclosetothege-
componentvelocityisv =981+735kms−1,andtheintermedi-
h −91 ometricalcenterofthecurrentX-rayremnantofN49,themean
atevelocityisvi=674+−4799kms−1.Forcomparisons,Vancuraet transverse velocity is vmean = 1717 D11.6 t6−6100 km s−1, where
al. (1992)havederivedashockvelocityofv=730kms−1 for D isthespatialseparationoftheknotfromthegeometrical
11.6
the X-rayemissionofN49(Note: Forsimplicity, weassumed center of the SNR in the units of 11.6pc, andt is the age
6600
theelectronandiontemperaturesinequalibrationinthesecal- of the SNR in the units of 6600 yr. The analysis on this hard
culations). X-ray knot is however substantially limited by the poor PSF.
The ionization timescale (n t) ratios between the cool and Follow-upon-axisChandraobservationswouldbeessentialto
e
the hot components are (n t) /(n t) < 0.1, which implies t unveilthedetailednatureofthisfeature.
e c e rh ∼ c
< 0.01t , wheret is the ionizationage of the shockentering
∼ rh c
thecloudsandtrhistheionizationageofthere-heatedregions. 4.3. SGR0526−66
AssuminganSNexplosionenergyofE =1051 ergs,theSNR
0
The Chandra/ACIS observations conclusively detect an X-
radiusof r = 10pcbasedonthe X-rayimageofN49, andthe
preshockdensityn =0.9cm−3,theimpliedSedovdynamical ray point source at the position of SGR 0526−66 within the
0
boundaryofSNRN49.TheX-rayspectrumofthispointsource
age (Sedov1959),t is ∼ 6600yr. Assumingt is on the or-
d rh
deroft or∼104yr,themeasuredn t ratiosimplyt <100yr. is dominated by continuum, which can be described with a
d e c ∼
powerlawmodel(see§3.3.4). Thispointsourceisthusconfi-
ThisshortionizationagesuggeststhatthesoftX-rayemission
dentlyidentifiedwiththeX-raycounterpartofquiescentphase
(kT ∼ 0.44 keV) has been produced by a rapidly decelerated
SGR0526−66.
shock entering dense molecular clouds as was observed from
The single power law model indicates a soft spectrum
theCygnusLoopSNR(Hesteretal. 1994).
with G ∼ 3 for SGR 0526−66, which resembles the typ-
ical spectrum of the AXPs (Mereghetti et al. 2002 and
4.2. SupernovaEjectaFragments
references therein), while the X-ray spectra from Galac-
ThehardX-rayknotinthesouthwestbeyondtheSNRblast tic SGRs are described with harder single power laws
wave shock front (region 6; Figure 6) is most likely associ- of G ∼ 2 (Sonobeetal. 1994; Kouveliotouetal. 1998;
atedwiththeSNRratherthanabackground/foregroundobject. Woodsetal. 1999;Mereghettietal. 2000;Kouveliotouetal. 2001).
The X-ray spectrum is dominated by emission lines and can As X-ray spectra from the AXPs and SGRs can be described
bedescribedwithakT ∼1keVplasmawithmetal-richabun- with the power law + black body models, we have made an
dances, which supports the shocked ejecta origin. The fitted attempt to fit the SGR 0526−66 spectrum with such a multi-
plane-parallelshockmodelindicatesahighlyadvancedioniza- componentmodel.Withtheoff-axisGTOdata,thefittedblack
tionstate(n t∼5×1013cm−3s−1)andthespectrummayalso body temperature (kT ∼ 0.55 keV) is consistent with typical
e
be equally described with a thermal plasma in the CIE. The AXP/SGR spectra, and the improvement in the overall fit is
hightemperatureandthelargeionizationtimescaleofthisfea- also statistically significant based on the F-test. The power
ture are also supportive of its origin as the high-speed ejecta law component, with the addition of the black body compo-
fragmentscreatedatthetimeoftheSNexplosion. nent,howeverimpliessubstantiallyharderspectrum(G ∼1.7)
BecauseofthePSFdegradation,theapparentangularexten- thanthoseoftheAXPs(G ∼3−4;Mereghettietal. 2002and
sionofthisfeature(∼5′′ inradius)islikelyanartifactandthe references therein). Nonetheless, the two component model
truemorphologyisuncertainwiththecurrentdata. Inorderto is still in good agreements with Galactic SGRs as they are
estimate the electron density n , we thus assume two extreme fitted with G ∼ 1−2 and black body temperature kT ∼ 0.5
e
cases of a circular shape with a radiusof 5′′ (for the apparent keV (Woodsetal. 1999; Kouveliotouetal. 2001). We how-
angular size and so 1.2 pc radius in physical size at the dis- ever note that this two component model is not required to
tance of d = 50 kpc) and 0.′′5 (for the case of an unresolved describetheon-axisdata. Sincetheembeddedconfusionlevel
point-like source, and so the physical size of < 0.125 pc ra- causedbythebackgroundinthespectralmodelfittingismore
∼
diusatd =50kpc). Theassumedrangeofthephysicalsizeof severe for the off-axisdata than the on-axisdata, we consider
the knot(∼0.1pc − 1.2pc inradius)is reasonablyconsistent thatthepresenceoftheblackbodycomponentisinconclusive,
withtheSNejectafragmentsdetectedinotherGalactcSNRsof althoughtheimprovementinthefitfortheoff-axisdataissta-
Vela and Tycho(Vancuraetal. 1995; Miyataetal. 2001). As- tisticallysignificantwiththeadditionalblackbodycomponent.
sumingthedistanceof50kpctotheSNRandasimplespher- Recently,Kulkarnietal. (2002)havereportedtheresultsfrom
ical geometry for the X-ray emitting volume, the best-fit EM the spectral analysis of SGR 0526−66 utilizing two Chan-
8 Park
dra/ACIS GO observations: one (ObsID 747) of those two enormousenergyreleaseinthe1979Marchg -rayburst.
data sets was identical to that we uitilze in the current work. If SGR 0526−66is indeed the compactremnantof the SN
They found that the observed spectrum of the SGR was best explosionwhichhasalsocreatedSNRN49anditwasbornnear
describedbyablackbody+ powerlawmodel, whichappears thecenterofN49,wemayestimateitstransversevelocity.The
to be inconsistentwith ourresults. Althoughthe originof the angularoffsetofSGR0526−66fromthegeometricalcenterof
inconsistency is unclear, we may attribute the descrepancy to N49 is ∼22′′, which implies that the spatial separation of the
thebackgroundestimationsutilizedinthespectralanalysis. As SGRfromthecenteroftheSNRisD=5.3(sinq )−1 pcatthe
Kulkarnietal. (2002)pointedout,theirSGRsourcespectrum distance of 50 kpc. Assuming that SGR 0526−66is the neu-
apparently include residual emission from the diffuse SNR, tron star born at the time of the SN explosion that produced
whereas such a contaminationin our on-axissource spectrum N49, and that the age of the SGR is t ∼ 6600 yr, the mean
appearstobenegligible.ThesoftX-raysfromthediffuseemis- transversevelocityisv >787kms−1. Thiswouldsuggest
mean ∼
sionoftheSNRmighthavecausedtheslightdifferencesinthe thatSGR0526−66isayoung,highvelocityneutronstarthatis
results of the spectral analysis as presented by Kulkarniet al. inadifferentclassthantypicalAXPsinsteadofbeingthesame
(2002). We havealso noticedthatKulkarnietal. (2002)have classobjectasotherAXPs,simplyinamoreevolvedphase.
not incorperated the CTI and QE corrections in their data re-
duction. AlthoughthesubstantialcontributionsbytheCTIand 5. SUMMARY
theQEdegradationeffectsonapoint-likesourceintheobser-
SNR N49 in the LMC has been observed with Chan-
vationtakeninarelativelyearlyphaseoftheChandramission
dra/ACIS. The high resolution “true-color” ACIS images re-
are unlikely, we speculate that potential contaminations from
vealcomplexX-rayfilamentarystructureswhichrepresentthe
theseartifactsinthespectralanalysisofanyACISdatacannot
blast wave shock sweeping through the ambient ISM and the
becompletelyruledout.Inaddition,wewouldliketocomment
shockinteractingwiththenearbydensemolecularclouds. The
on the spectral analysis of some parts of the diffuse emission
overallfeaturesfromtheEWimagesofthedetectedelemental
fromN49byKulkarnietal.(2002).TheyhaveusedaMEKAL
speciesindicatethatsoftlineemissionisenhancedinthesouth-
modelforsuchaspectralanalysis,whichshouldbeinappropri-
east while hard line emission is enhanced in the west of the
ate forthecomplicateddiffuseSNRemissionaswepresented
SNR.TheseoverallEWfeaturesaremostlikelydominatedby
inthecurrentwork. Morecomplex,physicallyrealisticmodels
thevariationsintheshocktemperatureandtheionizationstates
shouldbeutilizedinorderforan appropriatespectralanalysis
of the plasma across the SNR. The spatially-resolvedspectral
ofsuchdata.
analysisofN49suggeststhattheX-rayemissionfromtheSNR
The results from our simple spectral analysis may sup-
can be described with multiple-phase shocks representing the
port the speculation that SGR 0526−66 is a transition ob-
singly-postshockregion,there-heatedpostshockregionbythe
ject bridging the AXP and the SGR phases of neutron stars
reflectedshocks,andthecoolforwardshockenteringthedense
(Kaplanetal. 2001). Extensive analyses with the on-axis ob-
molecularclouds.
servations, which are beyond the scope of the current work,
WedetectahardX-rayknotinthesouthwestofN49beyond
would however be necessary to understand the detailed astro-
the SNR shock boundary. The X-ray spectrum of this knot is
physical characteristics of this object. For example, the mea-
emission line dominated and can be described with a metal-
surementsoftheperiod(P)andtheperiodderivative(P˙)would
richhotthermalplasma. Thisresultimpliesthatthishardknot
be critical to determine the magnetic field strengh B and the
appearstobeejectafragmentscreatedatthetimeoftheSNex-
spin-downaget ofSGR0526−66. Thisinformationcanpro-
plosion that producedSNR N49. The PSF degradationof the
vide important clues on the nature of SGR 0526−66 such as
off-axisobservationhoweverrestricts the data analysis of this
theidentificationwithahigh-field(B>1015G)neutronstaror
∼ featureandprecluderesolvingtheangularstructureofthisknot.
the “magnetar”(Paczyn´ski1992; Duncan&Thompson1992).
Follow-up on-axis observationswith a decent exposure might
Themeasurementofthespin-downagewouldalsobeusefulto
benecessarytoaddressthedetailednatureofthisfeature.
testthephysicalassociationofSGR0526−66withSNRN49,
The ACIS observationsof N49 also firmly detectthe X-ray
whichshouldprovideimportantimplicationsonthepossibleas-
counterpartof the quiescentphase of SGR 0526−66. The X-
trophysicalconnectionsbetweentheSGRsandtheAXPs(e.g.,
ray spectrum of SGR 0526−66 is continuum dominated and
Gaensleretal. 2001). Kulkarnietal. (2002)haverecentlyre-
canbefittedwitha powerlawmodel. Theimpliedphotonin-
porteda detectionofthe8.04spulsationfromSGR0526−66
dexisG ∼3. Thispowerlawspectrumisgenerallyconsistent
by utilizing∼85ks Chandra/ACISobservations. Thederived
withthosefromtheAXPsalthoughtheobservedg -rayburstof
magnetic dipole field strength is high (B ∼ 7 × 1014 G) and
SGR0526−66isararecharacteristicoftypicalAXPs.
supportsthemagnetarinterpretation. Theestimatedspin-down
age (P/2P˙ ∼ 2000yr) is consistentwith the dynamicalage of
TheauthorsthankJ.DickelforprovidinghisATCAdata.SP
SNRN49,whichsupportthephysicalassociationbetweenN49
andSGR0526−66.ThelocationofSGR0526−66intheLMC alsothankA.BykovforavaluablediscussionontheSNejecta
fragments.ThisworkhasbeeninpartssupportedbyNASAun-
and the likely association with N49 is also supported by the
dercontractNAS8-01128fortheChandraX-RayObservatory.
fitted columns: i.e., the absorbing columns for N49 and SGR
0526−66 are consistent. This SGR location may confirm the JPHacknowledgessupportfromChandragrantsGO1-2052X,
GO2-3068X,andGO2-3080BtoRutgers.
N49 9
REFERENCES
Anders,E.,&Grevesse,N.1989,GeochimicaetCosmochimicaActa,53,197 Kulkarni,S.R.,Kaplan,D.L.,Marshall,H.L.,Frail,D.A.,Murakami,T.,&
Aschenbach,B.,Egger,R.,&Tru¨mper,J.1995,Nature,373,587 Yonetoku,D.2002,ApJ,inpress(astro-ph/0209520)
Banas,K.R.,Hughes,J.P.,Bronfman,L.,&Nyman,L.A.1997,ApJ,480, Levenson,N.A.,Graham,J.R.,&Walters,J.L.2002,ApJ,inpress
607 Long,K.S.,Helfand,D.J.,&Grabelsky,D.A.1981,ApJ,248,925
Borkowski,K.J.,Lyerly,W.J.,&Reynolds,S.P.2001,ApJ,548,820 Marsden,D.,Rothschild,R.E.,Lingenfelter,R.E.,&Puetter,R.C.1996,ApJ,
Brandt, W. N., Alexander, D. M., Hornschemeier, A. E., Garmire, G. P., 470,513
Schneider, D. P., Barger, A. J., Bauer, F. E., Broos, P. S., Cowie, L. L., Mazets,E.P.,Golenetskii,S.V.,Ilinskii,V.N.,Aptekar,R.L.,&Guryan,Y.
Townsley, L.K., Burrows, D.N.,Chartas, G.,Feigelson, E. D.,Griffiths, A.1979,Nature,282,587
R.E.,Nousek,J.A.,&,Sargent,W.L.W.2001,AJ,122,2810 Mereghetti,S.,Cremonesi,D.,Feroci,M.,&Tavani,M.2000,A&A,361,240
Chartas,G.etal.2002,ApJL,Submitted(alsoChandraContributedSoftware: Mereghetti,S.,Chiarlone,L.,Israel,G.L.,&Stella,L.2002,“NeutronStars,
http://asc.harvard.edu/cont-soft/software/ACISABS.1.1.html). Pulsars and Supernova Remnants”, Proceedings of the 270. WE-Heraeus-
Cline,T.L.,Desai,U.D.,Teegarden,B.J.,Evans,W.D.,Klebesadel,R.W., Seminar,BadHonnef,January2002,eds.W.Becker,H.Lesch&J.Tru¨mper,
Laros,J.G.,Barat,C.,Hurley,K.,Niel,M.,Vedrenne,G.,Estulin,I.V.,Kurt, MPE-Reportinpress(astro-ph/0205122)
V.G.,Mersov,G.A.,Zenchenko,V.M.,Weisskopf,M.C.,&Grindlay,J. Miyata,E.,Tsunemi,H.,Aschenbach,B.,&Mori,K.2001,ApJ,559,L48
1982,ApJ,255,L45 Paczyn´ski,B.1992,ActaAstron.,42,145
Danziger,I.J.,&Leibowitz,E.M.1985,MNRAS,216,365 Park,S.,Roming,P.W.A.,Hughes,J.P.,Slane,P.O.,Burrows,D.N.,Garmire,
Dickel,J.R.,Chu,Y.-H.,Gelino,C.,Beyer,R.,Burton,M.G.,Milne,D.K., G.P.,&Nousek,J.A.2002a,ApJ,564,L39
Spyromilio,J.,Green,D.A.,Wilkinson,C.,&Junkes,N.1995.ApJ,448, Park,S.Burrows,D.N.,Garmire,G.P.,Nousek,J.A.,Hughes.J.P.,&Slane,
623 P.O.2002b,ApJ,Submitted
Dickel,J.R.,&Milne,D.K.,1998,AJ,115,1057 Rothschild,R.E.,Kulkarni,S.R.,&Lingenfelter,R.E.1994,Nature,368,432
Dickey,J.M.,&Lockman,F.J.1990,ARA&A,28,215 Russell,S.C.,&Dopita,M.A.,1990,ApJS,74,93
Dopita,M.A.1979,ApJS,40,455 Russell,S.C.,&Dopita,M.A.,1992,ApJ,384,508
Duncan,R.C.,&Thompson,C.1992,ApJ,392,L9 Sedov,L.I.1959,SimilarityandDimensionalMethodsinMechanics,transl.
Feigelson, E. D., Broos, P., Gaffney III, J. A, Garmire, G., Hillenbrand, L. M.Friedman(NewYork:AcademicPress)
A.,Pravdo,S.H.,Townsley,L.,Tsuboi,Y.2002,SubmittedtoApJ(astro- Seward,F.D.,&Chlebowski,T.1982,ApJ,256,530
ph/0203316) Shull,P.,Dyson,J.E.,Kahn,F.D.,&West,K.A.1985,MNRAS,212,799
Fender,R.P.,Southwell,K.,&Tzioumis,A.K.,1998,MNRAS,298,692 Smith,I.A.etal.1998,AdSpS,22,1133
Flanagan, K. A. et al. 2001, in AIP Conference proceedings, 565, Young Sonobe,T.,Murakami,T.,Kulkarni,S.R.,Aoki,T.,&Yoshida,A.1994,ApJ,
SupernovaRemnants,ed.,S.S.Holt&U.Hwang(NewYork:AIP),226 436,L23
Gaensler,B.M.,Slane,P.O.,Gotthelf,E.V.,&Vasisht,G.2001,ApJ,559,963 Strom,R.,Johnston,H.M.,Verbunt,F.,&Aschenbach,B.1995,Nature,373,
Gaetz,T.J.,Butt,Y.M.,Edgar,R.J.,Eriksen,K.A.,Plucinsky,P.P.,Schlegel, 590
E.M.,&Smith,R.K.2000,ApJ,534,L47 Townsley,L.K.,Broos,P.S.,Garmire,G.P.,&Nousek,J.A.2000,ApJ,534,
Hester,J.J.,&Cox,D.P.1986,ApJ,300,675 L139
Hester,J.J.,Raymond,J.C.,&Blair,W.P.1994,ApJ,420,721 Townsley,L.K.,Broos,P.S.,Nousek,J.A.,&Garmire,G.P.2002a,Nuclear
Hill,R.S.,Cheng,K.-P.,Bohlin,R.C.,O’Connel,R.W.,Roberts,M.S.,Smith, Instruments&MethodsinPhysicsResearchSectionA,486,751
A.M.,&Stecher,T.P.1995,ApJ,446,622 Townsley,L.K.,Broos,P.S.,Chartas,G.,Moskalenko,E.,Nousek,J.A.,&
Hughes,J.P.,Hayashi,I.,&Koyama,K.1998,ApJ,505,732 Pavlov,G.G.2002b,NuclearInstruments&MethodsinPhysicsResearch
Imanishi,K.,Koyama,K.,&Tsuboi,Y.2001,ApJ,557,747 SectionA,486,716
Kaplan, D. L., Kulkarni, S. R., van Kerkwijk, M. H., Rothschild, R. E., Tsunemi,H.,Miyata,E.,&Aschenbach,B.1999,PASJ,51,711
Lingenfelter, R.L.,Marsden,D.,Danner, R.,&Murakami, T.2001,ApJ, Tucholke,H.-J.,deBoer,K.S.,&Seitter,W.C.1996,A&AS,119,91
556,399 Vancura,O.,Blare,W.P.,Long,K.S.,&Raymond,J.C.1992,ApJ,394,158
Kouveliotou,C.1998,Nature,393,235 Vancura,O.,Gorenstein,P.,&Hughes,J.P.1995,ApJ,441,680
Kouveliotou, C., Tennant, A., Woods, P. M., Weisskopf, M. C., Hurley, K., Warren,J.S.,Hughes,J.P.,&Slane,P.O.2002,ApJ,inpress
Fender,R.P.,Garrington,S.T.,Patel,S.K.,&Go¨g˘u¨s¸,E.2001,ApJ,558, Woods,P.M.,Kouveliotou,C.,vanParadijs,J.,Finger,M.H.,Thompson,C.
L47 1999,ApJ,518,L103
10 Park
TABLE1
ENERGYBANDSUSEDFORGENERATINGTHEEQUIVALENTWIDTHIMAGES.
Elements Line Lowa Higha
(eV) (eV) (eV)
O 590−740 300−550 1200−1280
FeL 750−870 300−550 1200−1280
Ne(Hea ) 880−950 300−550 1200−1280
Ne(Lya )/FeL 980−1100 300−550 1200−1280
Mg 1290−1550 1200−1280 1630−1710
Si 1750−2100 1630−1710 2140−2280
aThe high andlow energybandsaroundthe selected line ener-
giesusedtoestimatetheunderlyingcontinuum.