1: %

2: %

3: 

4: %

5: 

6: %

7: \documentclass[12pt,preprint,longabstract]{aastex}

8: 

9: %

10: 

11: %

12: 

13: %

14: 

15: %

16: %

17: %

18: %

19: 

20: %

21: %

22: 

23: %

24: 

25: %

26: %

27: 

28: \bibliographystyle{apj}

29: 

30: %

31: %

32: %

33: %

34: %

35: %

36: %

37: %

38: 

39: %

40: 

41: \newcommand{\lastchangedefault}{\texttt{Specify Last Change Date}}

42: \newcommand{\lastchangedate}{\lastchangedefault}

43: \newcommand{\timestamp}{{\small \lastchangedate}}

44: \newcommand{\lastchange}[1]{\renewcommand{\lastchangedate}{#1}}

45: 

46: \newcommand{\ourversion}{Revisions in progress: \today~\time}

47: 

48: %

49: %

50: 

51: %

52: \slugcomment{Resubmitted: March 20, 2007}

53: 

54: %

55: 

56: %

57: %

58: \newcommand{\chase}     {ChASeM33}

59: \newcommand{\oursnr}{M33SNR\,21}

60: \newcommand{\oursnrlong}{M33SNR\,21}

61: %

62: \newcommand{\kms}       {\ensuremath{\;\mathrm{km}\;\mathrm{s}^{-1}}}

63: \newcommand{\percmsq}   {\ensuremath{\;\mathrm{cm}^{-2}}}

64: \newcommand{\cm}        {\ensuremath{\;\mathrm{cm}}}

65: \newcommand{\cmthree}   {\ensuremath{\;\mathrm{cm}^{-3}}}

66: \newcommand{\pc}        {\ensuremath{\;\mathrm{pc}}}

67: \newcommand{\kpc}       {\ensuremath{\;\mathrm{kpc}}}

68: \newcommand{\keV}       {\ensuremath{\;\mathrm{keV}}}

69: \newcommand{\Kelvin}    {\ensuremath{\;\mathrm{K}}}

70: \newcommand{\ks}        {\ensuremath{\;\mathrm{ks}}}

71: \newcommand{\erg}       {\ensuremath{\;\mathrm{erg}}}

72: \newcommand{\second}    {\ensuremath{\;\mathrm{s}}}

73: %

74: \newcommand{\etal}      {{\it et al.\/}}

75: \newcommand{\eg}        {{\it e.g.\/}}

76: %

77: \newcommand{\Chandra}   {\textit{Chandra\/}}

78: \newcommand{\XMM}       {\textit{XMM-Newton\/}}

79: \newcommand{\ROSAT}     {\textit{ROSAT\/}} 

80: \newcommand{\FUSE}      {\textit{FUSE\/}}

81: \newcommand{\Einstein}  {\textit{Einstein\/}}

82: %

83: \newcommand{\Halpha}    {\textrm{H}$\alpha$}

84: \newcommand{\HII}       {\textrm{H}\,\textsc{ii}}

85: \newcommand{\fion}[2]   {[\textrm{#1}\,\textsc{#2}]}

86: 

87: %

88: %

89: %

90: %

91: 

92: \shorttitle{\Chandra{} X-ray Imaging Spectroscopy of \oursnrlong{}}

93: \shortauthors{Gaetz \etal}

94: 

95: %

96: %

97: 

98: \begin{document}

99: 

100: %

101: %

102: %

103: 

104: \title{\Chandra{} ACIS Survey of M33 (ChASeM33): 

105: X-ray Imaging Spectroscopy of \oursnrlong{},

106: the brightest X-ray Supernova Remnant in M33}

107: 

108: %

109: %

110: %

111: %

112: %

113: %

114: 

115: \author{Terrance J. Gaetz\altaffilmark{1}

116: \email{gaetz@cfa.harvard.edu}

117: William P. Blair\altaffilmark{2},

118: John P. Hughes\altaffilmark{3},

119: P. Frank Winkler\altaffilmark{4},

120: Knox S. Long\altaffilmark{5},

121: Thomas G. Pannuti\altaffilmark{6},

122: Benjamin Williams\altaffilmark{7},

123: Richard J. Edgar\altaffilmark{1},

124: Parviz Ghavamian\altaffilmark{2},

125: Paul P. Plucinsky\altaffilmark{1},

126: Manami Sasaki\altaffilmark{1},

127: Robert P. Kirshner\altaffilmark{1},

128: Miguel Avillez\altaffilmark{8,9},

129: and

130: Dieter Breitschwerdt\altaffilmark{8}}

131: 

132: \altaffiltext{1}{Harvard-Smithsonian Center for Astrophysics,

133:     60 Garden Street, Cambridge, MA 02138}

134: 

135: \altaffiltext{2}{Department of Physics and Astronomy, \\

136: Johns Hopkins University, 3400 North Charles Street, Baltimore, MD 21218}

137: 

138: \altaffiltext{3}{Department of Physics and Astronomy, \\

139: Rutgers University, 136 Frelinghuysen Rd., Piscataway, NJ 08854-8019}

140: 

141: \altaffiltext{4}{Department of Physics, \\

142: McCardell Bicentennial Hall 526,

143: Middlebury College, Middlebury, VT 05753}

144: 

145: \altaffiltext{5}{Space Telescope Science Institute, \\

146: 3700 San Martin Drive, Baltimore, MD 21218}

147: 

148: \altaffiltext{6}{Space Science Center, \\

149: 200A Chandler Place, Morehead State University, Morehead, KY 40351}

150: 

151: \altaffiltext{7}{Department of Astronomy, Box 351580, University of 

152: Washington, Seattle, WA  98195}

153: 

154: \altaffiltext{8}{Institut f\"ur Astronomie, Universit\"at Wien,\\

155:      T\"urkenschanzstra{\ss}e 17, A-1180 Wien, Austria}

156: 

157: \altaffiltext{9}{Department of Mathematics, University of \'Evora,\\

158:               R. Rom\~ao Ramalho 59, 7000 \'Evora, Portugal}

159: 

160: %

161: %

162: %

163: %

164: 

165: %

166: \begin{abstract}

167: 

168: We present and interpret new X-ray data for \oursnrlong{}, the brightest X-ray

169: supernova remnant (SNR) in M33. The SNR is in seen projection against (and

170: appears to be interacting with) the bright \HII{} region NGC\,592. Data for this

171: source were obtained as part of the \Chandra{} ACIS Survey of M33 (ChASeM33)

172: Very Large Project. The nearly on-axis Chandra data resolve the SNR into a $\sim5\arcsec$ diameter (20\,pc at our assumed M33 distance of $817\pm58$ kpc) 

173: slightly elliptical shell. The shell is brighter in the east, which 

174: suggests that it is encountering higher density material in that direction.

175: The optical emission is coextensive with the X-ray shell in the north,

176: but extends well beyond the X-ray rim in the southwest.  Modeling with X-ray 

177: spectrum with an absorbed {\tt sedov} model yields a shock

178: temperature of $0.46^{+0.01}_{-0.02}\,\keV$, an ionization timescale of

179: $n_\mathrm{e} t = 2.1^{+0.2}_{-0.3} \times 10^{12}\, \mathrm{cm}^{-3}\, 

180: \mathrm{s}$, and half-solar abundances ($0.45^{+0.12}_{-0.09}$).  Assuming 

181: Sedov dynamics gives an average preshock H density of 

182: $1.7\pm 0.3\, \mathrm{cm}^{-3}$.  The dynamical age estimate is 

183: $6500\pm 600\, \mathrm{yr}$, while the best fit $n_\mathit{e} t$ value 

184: and derived $n_\mathit{e}$ gives $8200\pm 1700\,\mathrm{yr}$; the weighted 

185: mean of the age estimates is $7600\pm 600\,\mathrm{yr}$.  We estimate an 

186: X-ray luminosity (0.25-4.5 keV) of 

187: $(1.2\pm 0.2) \times 10^{37}\,\mathrm{ergs}\, \mathrm{s}^{-1}$ (absorbed), and

188: $(1.7\pm 0.3) \times 10^{37}\,\mathrm{ergs}\, \mathrm{s}^{-1}$ (unabsorbed),

189: in good agreement with the recent \XMM{} determination.  No significant excess 

190: hard emission was detected; the luminosity 

191: $\le 1.2\times 10^{35}\, \mathrm{ergs}\, \mathrm{s}^{-1}$ (2-8 keV)

192: for any hard point source.

193: 

194: \end{abstract}

195: 

196: 

197: %

198: %

199: %

200: %

201: 

202: \keywords{galaxies: individual (M33) --- shock waves --- supernova remnants}

203: 

204: %

205: %

206: %

207: %

208: %

209: %

210: %

211: %

212: 

213: %

214: 

215: %

216: \section{Introduction}

217: \label{sec:introduction}

218: 

219: Multiwavelength studies of nearby galaxies are becoming increasingly

220: effective at providing statistically interesting samples of many

221: classes of objects, including supernova remnants (SNRs).  At

222: a distance of 817$\pm$58~kpc \citep{2001ApJ...553...47F}

223: and with a relatively face-on orientation 

224: \citep[$i=55\arcdeg\pm1\arcdeg$;][]{1989AJ.....97...97Z},

225: the late-type Sc spiral galaxy M33 is a key galaxy for

226: such studies.  At the assumed distance to M33, 1\arcsec\ subtends 4\pc{},

227: allowing the morphology of some objects to be studied and

228: many confused or crowded regions to be at least partially

229: resolved, depending on the wavelength band, instrumentation,

230: and available spatial resolution.

231: 

232: Nearly 100 SNRs have been identified in M33, based on a combination

233: of radio and optical imaging and spectroscopy 

234: \citep{1978A&A....63...63D,

235: 1979A&A....80..212S,

236: 1985ApJ...289..582B,

237: 1986A&AS...64..237V,

238: 1990ApJS...72...61L,

239: 1993ApJ...407..564S,

240: 1998ApJS..117...89G,

241: 1999ApJS..120..247G}.

242: M33 has also been surveyed by each of the imaging X-ray

243: missions, including \Einstein{}

244: \citep{1981ApJ...246L..61L,

245: 1988ApJ...325..531T},

246: \ROSAT{} \citep{1995ApJ...441..568S,

247: 1996ApJ...466..750L,

248: 2001A&A...373..438H},

249: and \XMM{}

250: \citep{2003AN....324...85P,2004A&A...426...11P,2006A&A...448.1247M}.

251: The \ROSAT{} survey of \citet{1996ApJ...466..750L} found 12 of the 98

252: optically identified SNRs in the catalog of 

253: \citet[hereafter GKL98]{1998ApJS..117...89G}, and the \XMM{} survey of

254: \citet{2003AN....324...85P,2004A&A...426...11P} brought the total

255: number of X-ray SNR identifications to 21.

256: \citet{2005AJ....130..539G} used archival {\it Chandra\/}

257: data on M33 in

258: comparison with the optical data sets to detect X-ray counterparts

259: to 22 of the 78 GKL98 SNRs within the \Chandra{}\ fields of view available 

260: at that time.  Using the characteristics of this X-ray sample, X-ray sources 

261: without optical or radio counterparts but with X-ray hardness ratios 

262: similar to those of confirmed SNRs were also identified as candidate SNRs.

263: This earlier work was largely responsible for motivating the \Chandra{} ACIS 

264: Survey of M33 (\chase{}), a Very Large Project to examine the 

265: X-ray point- and extended-source populations and diffuse X-ray emission 

266: in M33 \citep{2007Sasaki..in..prep}.

267: %

268: At this writing, the survey 

269: is in progress and nearly complete.  Here we choose to highlight observations 

270: of a single region in M33 to demonstrate the benefits that this survey 

271: will ultimately provide over much of the galaxy.

272: 

273: \oursnrlong{} (SNR \#21 in the GKL98 catalog; also \#121 in 

274: \citealt{2004A&A...426...11P} and \#108 in \citealt{2006A&A...448.1247M}) 

275: is the brightest X-ray SNR in M33 and is located in the outskirts of the 

276: giant \HII{} region NGC\,592,

277: %

278: which is $\sim$9\arcmin\ due 

279: west of the galaxy's nucleus (see Fig.~\ref{fig:fov.overview}).

280: The relation between the SNR and the \HII{} region is shown in 

281: more detail in Figure~\ref{fig:image.opt.ngc592}, an RGB composite image 

282: constructed from continuum-subtracted Local Group Galaxy Survey data 

283: (LGGS\footnote{{\tt http://www.lowell.edu/users/massey/lgsurvey.html}};

284: \citealt{2006AJ....131.2478M}), 

285: using \Halpha{} (red), \fion{S}{ii} (green), and \fion{O}{iii} (blue) 

286: narrow band images; see \S\ref{sec:imaging.analysis} for further details.  

287: The \HII{} region consists of a pair of bright cores 

288: separated by $\sim25\arcsec$ which are surrounded by extensive 

289: ($\sim 2\arcmin\times2\arcmin$) faint filamentary structure.  As 

290: indicated, the SNR stands out by virtue of its elevated [S~II] emission 

291: relative to H$\alpha$, a signature of shock-heated gas (see, 

292: \eg{}, \citealt{2004ApJS..155..101B}).

293: \clearpage

294: %

295: %

296: %

297: \begin{figure}

298: \epsscale{0.5}

299: \plotone{f1.eps}

300: %

301: %

302: \caption{\label{fig:fov.overview}

303: Large squares indicate the $16\farcm8\times16\farcm8$ 

304: ACIS-I detector footprint for

305: the pointings closest to the SNR (Field 4), superposed

306: on a deep \Halpha{} image taken with 

307: the Burrell Schmidt telescope at Kitt Peak \citep{2006AAS...208.0301M}.  

308: The small rectangle indicates the region covered by 

309: Figure~\ref{fig:image.opt.ngc592}

310: }

311: \end{figure}

312: %

313: 

314: %

315: %

316: \begin{figure}

317: \epsscale{0.5}

318: \plotone{f2.eps}

319: %

320: %

321: \caption{\label{fig:image.opt.ngc592}

322: False color RGB composite of continuum-subtracted images from the 

323: LGGS of the region surrounding \oursnr{}; 

324: \Halpha{} is shown in red, \fion{S}{ii} in green, and \fion{O}{iii} in 

325: blue. The SNR appears in projection against the giant \HII{}

326: region NGC\,592.  The two bright cores of the \HII{} region

327: are seen in magenta, and some of the faint outer emission

328: in the northeast can also be seen.  \oursnr{} appears as the

329: indicated cyan structure.

330: }

331: \end{figure}

332: %

333: %

334: 

335: %

336: %

337: \begin{figure}

338: \centering

339: \epsscale{1.0}

340: %

341: %

342: \plotone{f3.eps}

343: \caption{\label{fig:image.compare.xray.opt}

344: Top left: Exposure-corrected (0.35-4\keV{}) \Chandra{} image, binned by 

345: 0.5 ACIS pixels in each direction and smoothed with a 3-bin radius

346: Gaussian.  The coordinate axes are labeled with J2000 coordinates.

347: The red contours are from a similarly smoothed 

348: image for total counts, ranging from 2--18 cts/bin in steps of 4 ct/bin;

349: the background level is 0.03 ct/bin.  The white ellipse

350: shows the elliptical shell model dimensions.  The central white ``+''

351: is the X-ray center, and the white ``$\times$'' to the west is the centroid

352: position for the nonthermal radio source (see text).

353: Top right: The result of 10 iterations of the Lucy-Richardson

354: deconvolution algorithm.  The slight shift of the bright ridge to larger

355: radii is an artifact.

356: Bottom panels:  False color RGB composite of the optical images, with

357: \Halpha{} (red), \fion{S}{ii} (green), \fion{O}{iii} (blue).

358: Bottom left: Continuum subtracted LGGS images, overlaid with X-ray contours.

359: Bottom right: LGGS images prior to continuum subtraction, overlaid

360: with X-ray contours.  The numbers identify stars from 

361: the LGGS catalog (see text).

362: }

363: \end{figure}

364: %

365: %

366: \clearpage

367: \citet{1981ApJ...246L..61L}

368: first detected \oursnrlong{} in X-rays with \Einstein{}, and suggested

369: that it was an SNR based on its extremely soft spectrum.  

370: \citet{1993ApJ...418..743G} confirmed the SNR identification by combining 

371: radio (VLA and Westerbork), optical (imaging and spectral), and 

372: X-ray (\ROSAT{}) data.  The nonthermal nature of the radio emission was 

373: important for confirming the SNR in the midst of otherwise thermal 

374: emission from the \HII{} region.  \citet{1993ApJ...418..743G} measured 

375: the density-sensitive optical \fion{S}{ii} lines and found 

376: $n_\mathrm{e} = 270 \cm^{-3}$, indicating that a dense preshock 

377: environment is likely responsible for the high X-ray emissivity.  

378: The high luminosity and dense environment are also an indication that the

379: SNR is indeed embedded within the \HII{} region and not just seen in 

380: projection \citep{1993ApJ...418..743G}.  The \ROSAT{} data allowed 

381: estimates of $\sim$500 $\rm km~s^{-1}$ for the shock velocity and 

382: $\sim 4 \times 10^{6}\Kelvin{}$ for the postshock temperature.

383: 

384: More recently, \citet{2004A&A...426...11P} observed \oursnrlong{} as part 

385: of a deep \XMM{} survey of M33 and obtained an absorbed 

386: 0.2--4.5\,keV X-ray flux of 

387: $(1.41\pm0.02)\times 10^{-13}\,\erg{}\,\second^{-2}

388: \cm^{-2}$, which implies an absorbed X-ray luminosity of 

389: $(1.13\pm0.14)\times 10^{37}\,\erg{}\,\second^{-1}$ at our assumed

390: distance to M33.

391: The SNR has been cataloged and called by a variety of names, including

392: M33\,X-3, 2E\,0130.3+3023, 013022+30233, 022+233, GKL\,21, G98-21,

393: GKL98\,21, GDK\,29, RX\,J0133.1+3038, and XMMU\,J013311.6+303841.

394: 

395: We provide below a detailed analysis of the \chase{} data for \oursnr{}.  

396: The data reduction and processing steps are described in 

397: \S\ref{sec:xray.observations.and.data.reduction}.

398: We select the portion of the data with the best imaging 

399: resolution and compare with optical data for the region, 

400: finding very different

401: morphologies for the optical and X-ray emitting components of

402: this SNR (\S\ref{sec:imaging.analysis}).  We then combine the

403: high resolution X-ray data with other portions of the \chase{} data 

404: (in which the SNR was farther off-axis) to improve the statistics and 

405: perform X-ray spectral analyses (\S\ref{sec:spectral.analysis}).  

406: We derive global SNR parameters assuming a Sedov model, and 

407: compare and contrast \oursnr{} with similar SNRs in the LMC --

408: N49 and SNR\,0506-68.0 -- in \S\ref{sec:discussion}.  The last 

409: section, \S\ref{sec:conclusions}, summarizes our results and conclusions.

410: 

411: %

412: \section{X-ray Observations and Data Reduction}

413: \label{sec:xray.observations.and.data.reduction}

414: 

415: The \chase{} survey was designed to cover the inner and most crowded regions

416: of M33 with seven fields, each with a total exposure of 200\ks{} split 

417: into two observing intervals; see \citep{2007Sasaki..in..prep}

418: for further details on the survey.  ACIS-I is the primary detector, 

419: while ACIS-S2 and S3 

420: provide additional (though far off-axis) coverage of portions of the 

421: galaxy adjacent to the primary positions.  At this writing, both 

422: $\sim$100\ks{} Field 4 pointings closest to the \oursnr{} position 

423: have been performed (see Fig.~\ref{fig:fov.overview}).  The first 

424: Field 4 pointing was split into two pieces, ObsIDs 6382 and 7226, 

425: totaling 97.2\ks{} after screening for high background.  In these 

426: data, the SNR is $\sim$1.7\arcmin\ 

427: off-axis on the ACIS-I2 chip and at the same nominal pointing and roll 

428: (differing by less than 0.1\arcsec\ in both); we merged 

429: these datasets for 

430: subsequent analyses.  The second Field 4 pointing, ObsID 6383, 

431: was not split and totaled 91.4\ks{} after screening; the SNR 

432: was $\sim$2.1\arcmin\ off-axis on the ACIS-I3 chip.  The Field 4 

433: observations are all close 

434: enough to on-axis to attain good spatial resolution.  We simulated 

435: the \Chandra{} PSF with 

436: ChaRT\footnote{{\tt http://cxc.harvard.edu/chart/}}

437: \citep{2003ASPC..295..477C} using the fitted \oursnr{} spectrum 

438: (see \S\ref{sec:spectral.analysis}) and obtained half-power diameters

439: of 0.93\arcsec\ and 1.01\arcsec\ for combined ObsIDs 7226+6382 

440: and ObsID 6383, respectively.

441: 

442: \oursnr{} was also observed in five other pointings at off-axis 

443: angles $\sim 8\arcmin$--18\arcmin\ (see Table~\ref{tbl:gkl21.obsids}).  

444: Although the rapid degradation of {\it Chandra's\/} spatial resolution 

445: at large off-axis angles (\Chandra{} 

446: POG\footnote{The Chandra Proposers' Observatory Guide, Version 8.0})

447: %

448: makes these observations unsuitable for spatial studies of this SNR, the 

449: source is bright and isolated enough from other nearby sources to make 

450: these data useful for investigating the spectrum of the SNR as a whole.

451: Moreover, ObsIDs 6380 (Field 3) and 6388 (Field 7) imaged the SNR on 

452: the ACIS-S3 chip which has superior low-energy response and thus 

453: provides improved information about the softest X-rays.

454: 

455: We reprocessed the X-ray data from the Level 1 files to remove pixel 

456: randomization and apply time-dependent gain changes.  We screened for 

457: background flares using the ACIS-I3 light curves except for the 

458: ACIS-S3 data, for which we used the ACIS-S3 light curves.  We used 

459: CIAO version 3.3.0.1 and CALDB version 3.2.0.  

460: In the Field 4 ObsIDs 6282 and 7226, the source dithered across

461: two ACIS I2 ``bad'' columns (162, 196), resulting in the loss of

462: $\sim13$\% of the counts from the SNR.  These columns had been marked

463: bad because of a slight excess background.  However, the SNR spans

464: only a short stretch of the columns, and a careful examination of the

465: rest of the data for these columns showed that the spurious background 

466: would be negligible for the SNR extraction region ($\sim 0.3$ ct for 

467: column 162, effectively 0 for column 196).  Accordingly, we restored 

468: those columns for the analysis to improve the statistics.  All the 

469: Field 4 data were merged for the imaging analysis to improve the 

470: statistics.  We applied a subpixel event repositioning (SER) correction 

471: to the data using software developed by \citet[software available from 

472: the \Chandra{} contributed software 

473: page\footnote{\tt http://cxc.harvard.edu/cont-soft/soft-exchange.html}]

474: {2004ApJ...610.1204L} to improve the spatial resolution.  The SER algorithm

475: uses the distribution of pulse height amplitudes (PHAs) within the

476: $3\times 3$ pixel event islands to improve the event centroiding.

477: 

478: From the processed event list, we construct a 0.35--4\keV{} ``total

479: counts'' image (ct/bin) by binning the events from the merged

480: $\sim200\ks{}$ Field 4 data set by 1/2 ACIS pixel in each

481: direction ($\sim 1/4\arcsec\times 1/4\arcsec$ bins).  We divide this

482: counts image by a corresponding exposure map constructed using the fitted 

483: SNR spectrum (see \S\ref{sec:spectral.analysis}) to generate an exposure

484: corrected image ($\mathrm{phot}\,\mathrm{cm}^{-2}\,\mathrm{s}^{-1}$).  This

485: exposure-corrected image is then smoothed with a 3 bin radius Gaussian

486: kernel.  The image was overlaid by contours from a similarly smoothed

487: version of the original total counts image, with contours ranging from 2 ct/bin

488: to 18 ct/bin at intervals of 4 ct/bin.  The corresponding background level is

489: $\sim 0.03$ ct/bin.

490: The processed X-ray imaging data are shown and compared with optical images 

491: of the region in Figure~\ref{fig:image.compare.xray.opt}; see

492: \S\ref{sec:imaging.analysis} for a discussion of the optical data.

493: The upper left panel of Figure~\ref{fig:image.compare.xray.opt} shows

494: the smoothed, exposure-corrected 0.35--4\keV{} image from the combined 

495: $\sim200\ks{}$ Field 4 data set, binned to 1/2 ACIS pixel in each

496: direction, with contours from the smoothed total counts image as

497: described above.

498: The X-ray imagery reveals a strong asymmetry in the surface brightness 

499: distribution.  The bright clump in the east has a surface 

500: brightness $\sim 5$ times that of the fainter portions of the SNR.  

501: To examine potential finer spatial structure of \oursnr{}, we experimented 

502: with Lucy-Richardson deconvolutions of this image.  The upper right panel 

503: of Figure~\ref{fig:image.compare.xray.opt} shows the result for 10 iterations 

504: with the CIAO tool {\tt arestore} using the combined Field 4 

505: (ObsIDs 6382, 7226, and 6383) total counts image, and merged ChaRT simulations

506: for the PSF.  The deconvolution shows an arc of bright emission in the east, 

507: and possibly some patchy structure within the interior.  The slight shift of 

508: the bright ridge to larger radii is an artifact of the reconstruction.

509: 

510: %

511: %

512: %

513: \input{tab1}

514: %

515: %

516: 

517: %

518: %

519: %

520: \begin{figure}[t]

521: \epsscale{0.5}

522: \plotone{f4.eps}

523: %

524: \caption{\label{fig:elliptical.shell.model}

525: Elliptical shell fit to Field 4 image data.

526: The panels compare the data and projected model for the four quadrants

527: centered on North, South, East, and West.

528: }

529: \end{figure}

530: %

531: %

532: %

533: %

534: \section{Imaging Analysis}

535: \label{sec:imaging.analysis}

536: 

537: For optical images of the \oursnr{} field, we took data from the Local

538: Group Galaxies Survey (LGGS; \citealt{2006AJ....131.2478M}) which used 

539: the KPNO 4m telescope and Mosaic CCD camera to survey most of M33  

540: (3 overlapping fields) through narrow-band \Halpha{}, \fion{O}{iii} 

541: $\lambda$\ 5007, and \fion{S}{ii} $\lambda\lambda 6716, 6731$\ filters, plus 

542: broadband \textit{UBVRI}.  \oursnr{} appears in both the north and central 

543: fields.  From each Mosaic image we clipped out a small section centered 

544: on the SNR and precisely aligned these using several dozen field stars.  

545: We then selected the images with the best seeing in each of the 3 emission 

546: lines (all from the north field) and in the {\it V\/} and {\it R\/} 

547: bands (both from the central field) and matched the point-spread 

548: functions.  Finally we scaled the broadband images appropriately, 

549: and performed continuum subtraction to remove the stars as effectively as

550: possible.  In this case, we subtracted \textit{V} from \fion{O}{iii}, and \textit{R} from 

551: both \Halpha{} and \fion{S}{ii}.  The result is shown 

552: in Figure~\ref{fig:image.opt.ngc592}.  

553: The lower left panel of Figure~\ref{fig:image.compare.xray.opt}

554: is an enlarged version of the same image with colors rescaled to emphasize the 

555: SNR, and the lower right panel is an RGB composite of the narrow-band 

556: images prior to continuum subtraction to show the locations of the bright 

557: stars.  X-ray contours have been plotted on both to facilitate comparison 

558: with the upper panels.  Based on the \textit{V} magnitudes and 

559: \textit{UBVRI} colors obtained by the LGGS, the three stars seen in 

560: projection (labeled ``1'', ``2'', and ``3'') and the star furthest 

561: to the northeast (``5'') are O stars in M33.  

562: Typical O star X-ray luminosities are  $\la 10^{33} \erg{}\,\second^{-1}$

563: \citep{1981ApJ...245..163V} and an O star would be undetectable in the 

564: 500\ks{} of data used in our analysis.  The ``star'' closest to the northeast rim 

565: , ``4'', consists of three stars, two of which appear to be O stars.  The third

566: (much fainter) star has an unusual color, but no X-ray events were detected 

567: in the 190\ks{} of the Field 4 imaging data 

568: (Fig.~\ref{fig:image.compare.xray.opt}).  

569: The X-ray data are not significantly contaminated by stellar emission.

570: %

571: %

572: %

573: \begin{figure}

574: \epsscale{0.5}

575: \plotone{f5.eps}

576: %

577: %

578: \caption{\label{fig:xray.wedge.reg}

579: Unsmoothed total counts image, binned to 1/2 ACIS pixels.

580: The Field 4 Bright Eastern (``BE'') and Faint Western (``FW'') wedge extraction 

581: regions are indicated. The extraction region for the full SNR is 

582: indicated by the ellipse labeled ``SNR''.

583: }

584: \end{figure}

585: %

586: %

587: %

588: 

589: To characterize the overall shape of the X-ray SNR, we fitted the 0.35-4\keV{} 

590: image data with an elliptical shell model (assuming that the symmetry 

591: axis lies in the plane of the sky), with density within the 

592: shell varying azimuthally as 

593: %

594: \begin{equation}

595: \label{eqn:elliptical.shell.model}

596: n(\theta) \propto n_0 (1 + 0.5 A (1 + \cos(\theta-\theta_0))) .

597: \end{equation}

598: %

599: Figure~\ref{fig:elliptical.shell.model} shows four projections of the model 

600: onto the data (including the effect of the simulated \Chandra{} PSF) for 

601: quadrants centered on north, south, east, and west.  The greater brightness 

602: in the eastern quadrant compared to the western quadrant is evident.

603: The model shows the SNR to be slightly elliptical (axis ratio 1.07) with 

604: the position angle of the major axis $\sim 9^{\circ}$ east from north.  

605: The semimajor (semiminor) axis is 2.65\arcsec (2.48\arcsec) with an

606: uncertainty of $\sim \pm 0.05\arcsec$.  At the assumed distance of M33, 

607: this corresponds to a semimajor (semiminor) axis of 10.5\pc{} (9.8\pc{}) with

608: an uncertainty of $\sim\pm0.2\,\mathrm{pc}$, 

609: and an overall size uncertainty of $\sim 0.7\,\mathrm{pc}$ when the

610: distance uncertainty is included.  The surface brightness peaks at 

611: position angle $\sim 104\degr$ east from north.  The fitted value for 

612: the factor $A$ in the elliptical shell model is 1.19; this implies the

613: ambient density on the bright side is about a factor of $1+A \approx 2$ 

614: higher than that on the faint side.  

615: The outline of the elliptical

616: shell model is indicated by an ellipse in the upper panels of

617: Figure~\ref{fig:image.compare.xray.opt}.

618: Compared to the fitted ellipse, the X-ray emission extends slightly further 

619: to the south than other directions (Fig.~\ref{fig:image.compare.xray.opt}, 

620: upper left panel; see also the comparison between the northern vs. southern 

621: quadrants in Figure~\ref{fig:elliptical.shell.model}).  

622: The fractional thickness of the shell,

623: $\Delta R_\mathit{s}/R_\mathit{s}$, is poorly constrained because 

624: the SNR is only barely resolved; we found that a value 

625: of $0.065$ provides an adequate description for the 

626: limb-brightening of the SNR's shell.  

627: The resulting center position is 

628: $(\alpha_\mathrm{J2000},

629: \delta_\mathrm{J2000})$=($01^\mathrm{h}33^\mathrm{m}11\fs73$, 

630: +$30\arcdeg38\arcmin41\farcs9$), indicated 

631: by the central ``+'' in each panel of 

632: Figure~\ref{fig:image.compare.xray.opt}.  This position agrees well

633: with the \XMM{} position determination \citep{2006A&A...448.1247M} 

634: of $(\alpha_\mathrm{J2000}, \delta_\mathrm{J2000})

635:   = (01^\mathrm{h}33^\mathrm{m}11\fs76$,

636: +$30\arcdeg38\arcmin42\farcs1)$, well within the \XMM{} uncertainty of 

637: 0.55\arcsec.

638: 

639: %

640: %

641: %

642: \begin{figure}[t]

643: \epsscale{0.5}

644: %

645: \includegraphics[width=0.4\textwidth,angle=270,clip]

646:   {f6.eps}

647: %

648: %

649: \caption{\label{fig:spectrum.sedov.solarabd}

650: \Chandra{} ACIS-I and ACIS-S3 spectra and best fit {\tt sedov} model 

651: of \oursnr{}; the curves which are higher at low energies are ACIS-S3,

652: and the rest are ACIS-I data.

653: }

654: \end{figure}

655: %

656: %

657: %

658: 

659: Using techniques similar to those in the \citet{2006Book...W} review

660: paper, we also examined the merged Field 4 hard band (3--8\keV{}) image for any 

661: indications of a point source or emission from a plerion.  The extraction 

662: region for the whole SNR includes only 14 counts in the 3--8\keV{} band during 

663: the entire 190\,ks integration.  The best fit {\tt XSPEC sedov} model (see

664: \S\ref{sec:spectral.analysis}) predicts $11\pm5$ counts in this band, 

665: so there is no evidence for a significant excess of hard emission.  

666: The brightest potential point source has a total of 3 counts 

667: (3--8\keV{}) within two adjacent pixels, providing a 3$\sigma$ count rate upper

668: limit of $\la 4.2\times 10^{-5}\,\mathrm{ct}\,\mathrm{s}^{-1}$.  

669: If we assume a Crab-like powerlaw spectrum (powerlaw index $\alpha=2.05$), 

670: the same absorbing column as for the best fit {\tt sedov} model

671: and the response for the ACIS-I2 observation, this count rate limit allows 

672: an unabsorbed flux upper limit for a point source of 

673: $\la 1.5\times 10^{-15}\,\mathrm{erg}\,\mathrm{cm}^{-2}\,\mathrm{s}^{-1}$

674: (2--8\,\keV{}) corresponding to an unabsorbed X-ray luminosity upper 

675: limit of $\la 1.2\times 10^{35}\,\mathrm{erg}\,\mathrm{s}^{-1}$ for

676: the same band.

677: 

678: The centroid of the nonthermal radio source 

679: \citep[precessed to J2000]{1993ApJ...418..743G}

680: is 2.8\arcsec\ west of the center of the X-ray emission, and indicated

681: by a ``$\times$'' toward the right in each panel of 

682: Figure~\ref{fig:image.compare.xray.opt}.

683: Thermal emission from the \HII{} region may confuse the radio picture 

684: \citep{1993ApJ...418..743G}.  The resolution of the radio data is 

685: $\sim 7\arcsec$ so the $\sim 3\arcsec$

686: offset of the radio centroid from the X-ray center is unlikely to be

687: significant.

688: 

689: The relative morphologies of the optical and X-ray emissions from \oursnr{}

690: are interesting and somewhat perplexing. The X-ray SNR appears nearly 

691: circular, although with the strong brightness gradient already described, 

692: but the optical SNR appears somewhat more structured and extended.  The 

693: bright optical emission agrees well with the outer X-ray contour shown in 

694: Figure~\ref{fig:image.compare.xray.opt} along the northern and 

695: northwestern X-ray limbs, apparently validating the relative astrometry 

696: used to align the images.  However, the bright eastern X-ray limb is almost 

697: devoid of optical emission (although this comparison is hampered by the stellar 

698: emission) and faint optical emission extends farther to the southwest

699: from the X-ray contour.  Taken on its own, the optical morphology in the 

700: southwest is reminiscent of a ``blow-out,'' which is consistent with the 

701: idea that the density is lower on the side of the SNR away from the bright 

702: \HII{} cores.  The absence of optical emission on the eastern X-ray peak does 

703: not appear to be due to obscuration, as the $N_\mathrm{H}$ column derived 

704: below is not extreme.  

705: 

706: %

707: %

708: \clearpage

709: \input{tab2}

710: %

711: %

712: 

713: %

714: %

715: %

716: \input{tab3}

717: %

718: %

719: 

720: %

721: \section{Spectral Analyses}

722: \label{sec:spectral.analysis}

723: 

724: The combined Field 4 data (the only X-ray data for which the SNR 

725: is resolved well enough to extract spectra from individual subregions) has 

726: $\sim 3600$ counts.  We extracted spectra from two wedge-shaped regions, 

727: each containing roughly half the total counts:  the bright eastern part 

728: (``BE'', $\sim 1600$ counts), and the fainter western part 

729: (``FW'', $\sim 2000$ counts); see Figure~\ref{fig:xray.wedge.reg}.

730: The BE region has $28$\% of the area.  We also extracted spectra for 

731: the SNR as a whole using an elliptical region (``SNR''; see 

732: Figure~\ref{fig:xray.wedge.reg}).  Because ObsIDs 6382 and 7226 fell on the 

733: ACIS-I2 chip, while ObsID 6383 fell on ACIS-I3, we extracted separate spectra 

734: for the merged 6382+7226 data and the 6383 data.  The data were grouped 

735: to a minimum of 25 counts/bin, and the $\chi^2$ statistic was used for 

736: the fits.  Background spectra were extracted from larger adjacent regions.

737: In order to improve the fit statistics for the spectrum for the SNR as a whole,

738: we also extracted spectra for a number of the far off-axis observations:

739: Field 1 (ACIS-I1), and Fields 3 and 7 (ACIS-S3).  The source and background 

740: extraction regions were enlarged appropriately to account for the expansion 

741: of the \Chandra{} PSF at large off-axis angles.  This provided a total of 

742: $\sim 10000$ counts for the SNR as a whole.  For the ACIS-S3 data 

743: (Fields 3 and 7), the SNR was imaged near the edge of the chip and 

744: $\sim 10\%$ of the flux spilled off the chip.

745: 

746: We fitted the spectra using {\tt XSPEC} version 11.3.2.  In each case, we included 

747: absorption ({\tt phabs}) corresponding to a Galactic column of 

748: $N_\mathrm{H} = 5.5\times 10^{20}\,\cm^{-2}$ toward M33 

749: \citep{1995yCat.8028....0S}; this component was frozen during the fits.  

750: To account for absorption within M33, we included a second absorption 

751: component ({\tt vphabs}) which was allowed to vary:  the abundances of the 

752: {\tt vphabs} component were fixed to be 0.5 solar, where we assume 

753: \citet{1989GeCoA..53..197A} values for the solar abundance set.  The M33 abundance 

754: at the galactocentric radius of \oursnr{} is 

755: $\sim 0.5\,Z_\odot$ \citep{1995ApJ...438..170H}.

756: For the thermal component, we examined collisional ionization equilibrium 

757: models (which provided very poor fits), and a number of nonequilibrium 

758: ionization (NEI) models: {\tt nei}, {\tt pshock}, and {\tt sedov}.  

759: The {\tt nei} and {\tt pshock} models represent impulsive heating to 

760: a constant temperature, $T$; in the former model, the spectrum evaluated 

761: at a single ionization timescale, $\tau \equiv n_\mathit{e} t$, while 

762: in the latter, the spectrum is integrated over ionization timescales from 

763: $\tau = 0$ to $n_\mathit{e} t$.  In the {\tt sedov} model, the 

764: nonequilibrium ionization spectrum is integrated over a Sedov-stage 

765: SNR profile, and the ionization timescale is $\tau = n_\mathit{e,s}\,t_0$ 

766: where $n_\mathit{e,s}$ is the postshock electron density and $t_0$ is 

767: the age of the SNR.  See \citet{2001ApJ...548..820B} for a more detailed 

768: description of the models.  

769: 

770: We performed fits for regions BE, FW, and the whole SNR using the high 

771: spatial resolution Field 4 data, and also for the whole SNR including data 

772: from Fields 1, 3, and 7 in addition to Field 4.  The spectra were fitted 

773: simultaneously with the model parameters (including model normalizations) 

774: linked.  To account for the loss of flux off the chip for the ACIS-S3 

775: data, an additional multiplicative {\tt const} model was applied, allowed 

776: to be free for the ACIS-S3 datasets, and fixed at 1.0 for the other 

777: datasets.  The resulting {\tt const} fit values were $\sim 0.9$ for 

778: the ACIS-S3 data, as expected.

779: 

780: All three models had comparable $\chi^2_\mathit{red} \sim 1$, but the 

781: {\tt sedov} models consistently produced better fits: 1.08 for 261 degrees 

782: of freedom (dof) for {\tt sedov} vs. 1.14 (259 dof) and 1.21 (259 dof)

783: for {\tt pshock} and {\tt nei}, respectively.  The {\tt nei} and 

784: {\tt pshock} models gave lower abundance estimates ($0.15$--$0.19 Z_\odot$ 

785: and $0.22$--$0.34 Z_\odot$, respectively) compared to the {\tt sedov} model

786: ($0.36$--$0.57 Z_\odot$).  The temperatures were 

787: $0.52$--$0.56\keV{}$ ({\tt nei}), $0.56$--$0.6\keV{}$ 

788: ({\tt pshock}), and $0.44$--$0.47\keV{}$ ({\tt sedov}).  The {\tt pshock} 

789: and {\tt sedov} fit parameters are listed in Table~\ref{tbl:gkl21.fits}; 

790: the listed parameter uncertainties are 90\% confidence intervals 

791: ($\Delta \chi^2 \le 2.706$ for one parameter) based on the statistical errors.

792: 

793: In the {\tt sedov} model, the temperature parameter is the postshock (electron) 

794: temperature.  In the Sedov self-similar solution the temperature increases (and 

795: the density decreases) radially inwards so that the emission-weighted 

796: X-ray temperature is $\sim 1.3$ times the shock temperature 

797: \citep{1974ApJ...194..329R}. In the other two models, {\tt nei} 

798: and {\tt pshock}, the temperature parameter is the (constant)

799: postshock temperature.  If the temperature in the SNR is in fact 

800: increasing radially inwards as in the Sedov solution, the fitted 

801: temperature parameters for the {\tt nei} and {\tt pshock} models would 

802: reflect an average X-ray temperature higher than the temperature at the

803: shock front; this would be consistent with the generally higher 

804: temperature parameters in those fits compared to the {\tt sedov} model 

805: fits (Table~\ref{tbl:gkl21.fits}).

806: 

807: Based on the best fitting {\tt sedov} model, we evaluated the X-ray flux and 

808: luminosity (absorbed and unabsorbed) for 0.35--3\,\keV{} (the range best 

809: constrained by the data) and a broader 0.25--4.5\,\keV{} range; the results 

810: are presented in Table~\ref{tbl:flux.lum}.  The quoted errors include the 

811: normalization uncertainty (Table~\ref{tbl:gkl21.fits}) and the distance 

812: uncertainty.  These luminosities may be compared to the luminosities 

813: (scaled to 817\kpc{}) estimated with \Einstein{} IPC 

814: ($1.7\times 10^{37}\,\mathrm{erg}\,\mathrm{s}^{-1}$ absorbed, 0.15--4.5\,keV; 

815: \citealt{1981ApJ...246L..61L}), \ROSAT{} PSPC (flux 

816: $1.4\times 10^{-13}\,\mathrm{erg}\,\mathrm{cm}^{-2}\,\mathrm{s}^{-1}$

817: implying $L_X = 1.1\times 10^{37}\,\mathrm{erg}\,\mathrm{s}^{-1}$

818: absorbed, 0.1--2.4\,keV; \citealt{2001A&A...373..438H})

819: and \XMM{} EPIC (flux $(1.41\pm0.02)\times 

820: 10^{-13}\,\mathrm{erg}\,\mathrm{cm}^{-2}\,\mathrm{s}^{-1}$ 

821: implying 

822: $L_\mathit{X} = (1.13\pm0.16)\times 10^{37}\,\mathrm{erg}\,\mathrm{s}^{-1}$

823: absorbed, 0.2--4.5\,keV; \citealt{2004A&A...426...11P}).  Our luminosity

824: estimate of

825: $L_\mathit{X} = (1.2\pm0.2)\times 10^{37}\,\mathrm{erg}\,\mathrm{s}^{-1}$ 

826: (absorbed, 0.25--4.5\,keV) is in good agreement with the \ROSAT{} 

827: and \XMM{} estimates and in reasonable agreement with the \Einstein{} estimate.

828: It is also worth noting that the SNR X-ray luminosity function for M33

829: does not reach as high as in other nearby galaxies

830: \citep{2005AJ....130..539G}.  \oursnr{}, the brightest SNR in M33, 

831: would be only the fourth or fifth brightest SNR if it were in the

832: Large Magellanic Cloud (LMC), based on the luminosities

833: reported in \citet{1981ApJ...248..925L}.

834: 

835: %

836: \section{Discussion}

837: \label{sec:discussion}

838: 

839: %

840: \subsection{Association with NGC\,592}

841: \label{subsec:assoc.with.ngc.592}

842: 

843: As noted in \S\ref{sec:introduction}, the SNR lies along the line of

844: sight to the bright \HII{} region NGC\,592.  The high X-ray luminosity of 

845: the SNR and the relatively high density interstellar medium (ISM)

846: inferred from the optical \fion{S}{ii} emission imply that

847: the SNR is in fact embedded in the \HII{} region \citep{1993ApJ...418..743G}. 

848: The fact that the X-ray emission from \oursnr{} is brightest on the side toward 

849: the bright \HII{} region core further supports the association with

850: NGC\,592.

851: %

852: This suggests that the \oursnr{} progenitor could have been a massive 

853: star associated with the burst of star formation which produced NGC\,592.

854: 

855: \citet{2006AJ....131..849P} recently reported on \FUSE{} observations of

856: NGC\,592 in which the \FUSE{} 30\arcsec\ aperture covered the bright

857: \Halpha{} cores (but excluded \oursnr{}).  The line profiles show extended 

858: blue absorption wings, indicating the presence of evolved O stars.  Her 

859: fits to synthetic spectra based on stellar population models are consistent 

860: with a $4.0\pm0.5$ Myr population at solar metallicity; the FUV models 

861: are consistent with a range of $\sim 0.4$--1.2 solar.  

862: \citet{2002MNRAS.329..481B} estimated a population age of 

863: $\ga4.5$\,Myr, but \citeauthor{2006AJ....131..849P} argues that this 

864: is inconsistent with the strength of the P-Cygni features.  If we 

865: assume that the \oursnr{} progenitor belongs to the same population, the

866: models of \citet{1990A&AS...84..139M} suggest an initial stellar mass

867: of $\sim 50\,M_\odot$ for half-solar to solar abundances.  Our X-ray 

868: spectroscopic analysis of \oursnr{} with \Chandra{} data also suggests roughly 

869: half-solar abundances.  The lack of an X-ray signature from ejecta is likely 

870: a combination of the age of the SNR ($\sim 5$--10\,kyr) and the limited 

871: \Chandra{} spatial resolution at the distance of M33.  For the

872: integrated spectrum for the SNR as a whole, most of the X-ray 

873: emission arises from the swept-up interstellar medium in M33.  

874: These high spatial resolution X-ray observations 

875: apparently do not provide sufficient resolution and statistics to 

876: separate out high abundance clumps if they are present.

877: 

878: \citet{2006AJ....131..849P} also obtained an intrinsic extinction of 

879: $E(B-V)_i = 0.07 \pm 0.02$ for the \HII{} region.  For this extinction, 

880: the LMC interstellar extinction curve \citep{1986AJ.....92.1068F} implies

881: $N_\mathrm{H} \approx 1.7\times 10^{21}\,\cm^{-2}$, while the Galactic 

882: extinction curve \citep{1978ApJ...224..132B} implies 

883: $N_\mathrm{H} \approx 4\times 10^{20}\,\cm^{-2}$.  The analysis of 

884: the \Chandra{} X-ray spectra provides primarily upper limits 

885: $N_\mathrm{H} \le (2$--10$)\times 10^{20}\,\cm^ {-2}$ for the M33 

886: column, consistent with a solar metallicity extinction curve.  However, 

887: given the complex environment, the position of \oursnr{} along the line of 

888: sight to NGC\,592 cannot be determined at present.  If the SNR is 

889: toward the near side of the \HII{} region, the low column for the 

890: X-ray data could also be consistent with the FUV data and a half-solar 

891: metallicity extinction curve.

892: 

893: %

894: \subsection{Sedov Analysis}

895: \label{subsec:sedov.analysis}

896: 

897: In the following, we examine the global properties of \oursnr{} based on the

898: assumption of a Sedov-stage SNR.  \citet{1999A&A...344..295H} investigated 

899: the properties of nonspherical SNRs using approximate analytic models, 

900: including the cases of an SNR in an exponential density gradient, and 

901: an SNR in an off-center $r^{-2}$ powerlaw distribution (for example, an 

902: explosion in the wind cavity of another star).  The integrated

903: spectral properties of the SNRs were found to be approximately 

904: like those of Sedov models, even though the detailed properties are 

905: expected to vary spatially across 

906: the SNRs.  The surface brightness ($\propto n^2$) varies more strongly 

907: with density around the remnant than the temperature ($\propto n$), and the shock 

908: velocity varies only weakly ($\propto n^{1/2}$).  The properties derived 

909: from the {\tt sedov} model fits should represent values averaged over the SNR.

910: 

911: The Sedov solution depends on three parameters: the explosion energy, 

912: $E_0$, the ambient mass density, $\rho_0$, and the time since the explosion,

913: $t_0$.  The ambient density of hydrogen, $n_0$, is related to the mass

914: density by $n_0 = \rho_0 / (1.43\,\mathrm{amu})$, where we assume

915: a fully ionized gas with half-solar abundances (based on the

916: \citealt{1989GeCoA..53..197A} solar abundance set).

917: The {\tt XSPEC sedov} model fits provide estimates for the postshock 

918: temperature, $T_\mathit{s}$, the ionization timescale, 

919: $\tau \equiv n_\mathit{e,s} t_0$, and also a model normalization 

920: $K = (10^{-14} / (4\pi D^2)) \int\,dV \,n_\mathit{e} \, n_\mathrm{H}$.

921: In the ionization timescale, $n_\mathit{e,s}$ is the postshock electron number

922: density (where we are assuming electron-ion equilibration in the shock),

923: and $n_\mathit{e,s} \approx 4.8\times n_0$ for our assumed abundances, and

924: assuming strong shock jump conditions.

925: For \oursnr{}, we also have an estimate for the physical radius of the SNR, 

926: $R_\mathit{s}$, based on the angular radius, $\theta_\mathit{s}$, and 

927: the distance to M33.  With four observational constraints and three model 

928: parameters, the system is overconstrained.  As noted by 

929: \citet{1998ApJ...505..732H}, the age estimates based on the Sedov dynamical 

930: age and the age based on ionization timescale allow for a consistency check. If 

931: the SNR is well modeled by a Sedov solution, these two age estimates should 

932: be the consistent with each other.  

933: 

934: The X-ray temperature implies a shock velocity,

935: %

936: \begin{equation}

937: \label{eqn:def.v.s}

938: v_\mathit{s} 

939:   = \left( { 16 k T_\mathit{s} \over 3 \mu_\mathit{s} } \right)^{1/2} 

940:   = 904\, T_{\mathit{s},\mathrm{keV}}^{1/2} \,\kms{}

941: \end{equation}

942: %

943: where $kT_{\mathit{s},\mathrm{keV}}$ is the postshock temperature

944: in \keV{} and $\mu_\mathit{s}$ is the postshock mean mass

945: per free particle.  The equation assumes a strong shock into a monatomic

946: gas with three degrees of freedom, and complete equilibration between

947: species \citep{1980ARA&A..18..219M}.  

948: For a fully ionized plasma with half-solar abundances, 

949: $\mu_\mathit{s} = (m_\mathrm{H} + A_\mathrm{He} m_\mathrm{He} + \cdots)/(2 +

950: 3 A_\mathrm{He} + \cdots)

951: \approx 0.61$, where $m_\mathit{H}$ and $m_\mathit{He}$ are the atomic 

952: weights of H and He, respectively, and $A_\mathrm{He}$ is the 

953: fractional abundance of He relative to H.  From the {\tt sedov} 

954: model estimate for the postshock temperature, we have 

955: $v_\mathit{s} = 611\pm 12\kms{}$ where 

956: the error is a purely statistical error, and almost certainly underestimates the 

957: real uncertainty in the velocity.

958: 

959: If the SNR is assumed to be expanding as $R_\mathit{s} \propto

960: t^\delta$, the shock velocity, combined with the current SNR radius, 

961: allows for an estimate of the SNR dynamical age, $t_0$.  The model 

962: normalization, $K$, combined with $v_\mathit{s}$, provides an estimate 

963: for the ambient hydrogen density, $n_0$, and the postshock electron density, 

964: $n_\mathit{e,s} = 4.8\,n_0$, can be estimated by using the strong 

965: shock jump conditions and the (half-solar) abundances.  The postshock 

966: density can be combined with the ionization timescale estimate, 

967: $\tau$, from the {\tt sedov} model fit to provide a second estimate for 

968: the SNR age.  From these basic parameters, the swept up mass, 

969: $M_\mathit{SU}$, and the explosion energy, $E_0$, can be estimated.  We 

970: proceed to the evaluation of the parameters based on the {\tt sedov} fit 

971: parameters from simultaneous fits to the data for the whole SNR (the right 

972: hand column in Table~\ref{tbl:gkl21.fits}).

973: 

974: The mean angular radius of the X-ray SNR is $2.56\arcsec\pm 0.13\arcsec$, 

975: where the uncertainty in the radius includes a 2\% angular size uncertainty 

976: and 3.5\% uncertainty (added in quadrature) for the azimuthal asymmetry.  

977: This angular radius and the assumed distance to M33 

978: imply a radius of $10.1\pm 0.9 \pc{}$ for the SNR, where the distance 

979: uncertainty is now included.  For a Sedov-stage SNR, the deceleration parameter 

980: is $\delta=2/5$, and we estimate the dynamical age to be 

981: $(2/5) R_\mathit{s}/v_\mathit{s} \approx 6500\pm600\,\mathrm{yr}$.

982: Given the radial dependence of the electron and H number densities, the

983: emission integral, $\int\,dV \,n_\mathit{e}\,n_\mathrm{H}$, can be evaluated 

984: from the model normalization.  For the case of the Sedov solution, 

985: %

986: \begin{equation}

987: \int\,dV\,n_\mathit{e}\,n_\mathrm{H} \approx 4\xi

988:       \left(

989:           {n_\mathit{e} \over n_\mathrm{H} }

990:       \right)

991:       n_0^2 

992:       \left(

993:           {4\pi \over 3} R_\mathit{s}^3

994:       \right)

995: \end{equation}

996: %

997: where $n_0$ is the preshock H density and $\xi = 0.517773$ is an

998: integration constant evaluated by numerical integration over the Sedov

999: solution.  The average preshock H density can be estimated as

1000: %

1001: \begin{equation}

1002: n_0 = 1.58 \, K^{1/2}_{-4}\, D_{800} \, R_\mathit{s,10}^{-3/2}

1003:    \, \cm^{-3} ,

1004: \end{equation}

1005: %

1006: where $K_{-4}$ is the {\tt XSPEC sedov} model normalization in units of $10^{-4}

1007: \cm^{-5}$, $D_{800}$ is the distance to the SNR in units of 800\kpc{},

1008: and $R_\mathit{s,10}$ is the radius of the SNR in units of 10\pc{}.  From 

1009: the {\tt sedov} fit parameters, we obtain an average preshock H density of 

1010: $n_0 = 1.7\pm0.3\,\cm^{-3}$.  Approximate pressure equilibrium around the 

1011: rim of the SNR would imply that the temperature should vary as 

1012: $\sim n_0^{-1}$.  Although the temperature derived from the fit to the 

1013: BE region is somewhat higher than that obtained from the FW region fit 

1014: (though not significantly so), even taking the full 90\% confidence ranges 

1015: into account does not provide the $\sim2$:1 ratio implied by the density 

1016: variation from the elliptical shell model fit (see 

1017: \S\ref{sec:imaging.analysis}).  The relative overpressure of the BE 

1018: region compared to the FW region may be an indication that the encounter 

1019: with the density gradient toward the east is relatively recent.  This might 

1020: also explain the lack of optical emission (see \S\ref{sec:imaging.analysis}) 

1021: if there has not been sufficient time for the shocks to become radiative.

1022: In principle, the ionization timescale provides a constraint on the age of the 

1023: encounter; however, 90\% confidence intervals for $\tau$ for the BE and 

1024: FW regions are broad enough to make this a weak constraint.

1025: An additional possibility is that there are sufficient photoionizing photons

1026: toward the east to prevent the postshock gas from recombining.  It is

1027: notable that the edge of the enhanced \fion{S}{ii} in the east coincides

1028: roughly with a bright rim of \Halpha{} emission associated with one of the

1029: bright cores of the \HII{} region.

1030: 

1031: If a Sedov-stage SNR is assumed, the ionization timescale provides for another

1032: estimate for the SNR age: 

1033: $t_0 = \tau / n_\mathit{e,s} \approx 8200\pm 1700\,\mathrm{yr}$.  

1034: The age based on ionization timescale is somewhat larger than the dynamical 

1035: age, but the estimates are reasonably consistent.  The weighted mean of the 

1036: estimates yields an age of $6700\pm600\,\mathrm{yr}$.

1037: 

1038: From the preshock H density and the radius of the SNR, the swept up mass 

1039: is $\sim 260\pm80\,M_\sun$, assuming a constant preshock density (which

1040: is almost certainly not the case, but would provide a reasonable estimate

1041: if the density gradients are not too extreme).  For a Sedov-stage SNR, 

1042: this mass, together with the postshock temperature, implies a total 

1043: explosion energy of $E_0 \approx (1.8\pm0.3)\times 10^{51}\, \erg{}$.

1044: 

1045: \citet{1993ApJ...418..743G} estimated the nonthermal pressure

1046: to be $P_\mathit{nt}/k \sim 2.4\times

1047: 10^{6}\,\mathrm{cm}^{-3}\,\mathrm{K}$

1048: (where $k$ is the Boltzmann constant)

1049: based on the radio flux and assuming a 28\pc\ diameter shell

1050: with thickness $\Delta R/R = 1/12$ for the emitting region.

1051: Scaling their pressure value to our smaller radius and using 

1052: our Sedov model parameters to estimate the emitting volume, we 

1053: obtain $P_\mathit{nt}/k \approx 4.6\times 10^{6}\,\mathrm{cm}^{-3}\,\mathrm{K}$.

1054: As noted by \citealt{1993ApJ...418..743G}, the 

1055: estimated nonthermal pressure is much smaller than the

1056: thermal pressure estimate based on X-ray data 

1057: (see Table~\ref{tbl:gkl21.cmp.snr}).

1058: They note that the nonthermal pressure estimate is likely to be 

1059: a lower limit:  pressure: field/particle nonequilibrium could increase the 

1060: nonthermal pressure, and volume for the radio emitting plasma

1061: (difficult to ascertain, because of the radio spatial resolution) 

1062: could be much smaller than was assumed.

1063: 

1064: %

1065: \subsection{Comparisons with Other SNRs}

1066: \label{subsec:n49.comparison}

1067: 

1068: The asymmetric brightness distribution invites comparison with two 

1069: LMC supernova SNRs which also show large variations in optical

1070: and X-ray surface brightness: N49, and 0506-68.0 (also known as N23).  

1071: Both SNRs show strongly asymmetric X-ray and optical surface 

1072: brightnesses as well as indications of interaction with denser material.

1073: Table~\ref{tbl:gkl21.cmp.snr} summarizes the X-ray properties and 

1074: Sedov model parameters for the SNRs, based on 

1075: \citet{1998ApJ...505..732H, 2006ApJ...645L.117H}.  

1076: 

1077: \citet{1993ApJ...418..743G} compared \oursnr{} to the LMC SNR N49;

1078: in the latter SNR, both the optical and the X-ray emission are strongly 

1079: enhanced in the southeast.  A comprehensive study by 

1080: \citet{1992ApJ...394..158V} examined optical and UV data, and developed a 

1081: self consistent model for N49.  The optical emission is coming from slow 

1082: (40--270\kms{}) shocks into dense (20--940\,cm$^{-3}$) gas as the SNR 

1083: interacts with a molecular cloud to the southeast, while the X-ray 

1084: emission results from faster (730\kms{}) shocks into an intercloud medium 

1085: ($n_0\approx0.9$\,cm$^{-3}$).  A CO cloud to the southeast of the SNR 

1086: \citep{1997ApJ...480..607B} would account for the high densities there.  

1087: \citet{1992ApJ...394..158V} inferred an age of 5400\,yr based on 

1088: Sedov dynamics.  

1089: 

1090: SNR\,0506-68.0 also shows strong asymmetries in its X-ray, optical,

1091: and radio morphologies \citep[and references therein]{2006ApJ...645L.117H}.  

1092: Such asymmetries 

1093: suggest that an interaction is taking place between the SNR and denser gas.  

1094: However, searches for an associated molecular cloud have not been successful; 

1095: moreover, the low absorption (as implied by the low column density) 

1096: argues against the presence of such a cloud \citep{1998ApJ...505..732H}.  

1097: \citet{2006ApJ...645L.117H} noted that the open cluster HS114 

1098: \citep{1966AJ.....71..363H} lies near the high density side of the SNR 

1099: (only 2\,pc in projection, assuming a distance of 50\kpc{} to the LMC).

1100: They found at least a factor of 10 variation in the ambient density

1101: around the rim of the SNR, and also detect regions with enhanced 

1102: abundances in parts of the SNR.  By assuming that the side of the

1103: SNR which is expanding into a region of low density can be modeled 

1104: approximately as a free Sedov expansion, \citet{2006ApJ...645L.117H} 

1105: estimated an age of $\sim 4600\,\mathrm{yr}$.

1106: 

1107: %

1108: %

1109: \clearpage

1110: \input{tab4}

1111: \clearpage 

1112: %

1113: 

1114: \oursnr{} shows similar asymmetries in the X-ray emission, with

1115: the rim toward the \HII{} region core being much brighter than the

1116: rest of the SNR.  Despite the X-ray similarities with the other SNRs, 

1117: the optical emission shows rather different properties.  N49 and 

1118: SNR\,0506-68.0 both show optical brightening on the same side as 

1119: the X-ray brightening.  In contrast, \oursnr{} does not show strong 

1120: optical brightening near the X-ray bright eastern rim (though the situation 

1121: is confused by stellar contamination), but instead shows an optical 

1122: extension toward the southwest well beyond the faint X-ray rim.

1123: The strong X-ray emission ahead of the optical emission in the east is 

1124: puzzling.  As noted in \S\ref{subsec:sedov.analysis}, this may indicate a 

1125: relatively recent encounter with denser gas for which photoionization

1126: associated with the bright \HII{} region core maintains the ionization

1127: at a high enough state that any \fion{S}{ii} emission is weak.  

1128: The extension of the optical emission well beyond the X-ray rim in 

1129: the southwest is suggestive of some sort of breakout into a lower density 

1130: region which has subsequently driven slower (hence X-ray faint) 

1131: radiative shocks into denser material.  In that case it is puzzling

1132: that no X-ray emission is seen within the breakout.  Possibly the 

1133: X-ray emitting plasma driving the optical shocks is tenuous enough 

1134: that the X-ray emission is too faint to be seen even in this 

1135: 190\ks{} exposure.  In any case, the peculiarities of the X-ray and 

1136: optical emission for \oursnr{} suggest an explosion in a complex environment.

1137: 

1138: %

1139: \section{Conclusions}

1140: \label{sec:conclusions}

1141: 

1142: We have presented the results of an analysis of the \chase{} project

1143: data for \oursnr{}, an SNR embedded in the giant \HII{} region NGC\,592 

1144: \citep{1993ApJ...418..743G}.  These data provide the first well-resolved

1145: X-ray imagery of an SNR in M33.  The remnant is slightly elliptical, and

1146: roughly $5\arcsec$ in diameter.  We fitted a three dimensional

1147: elliptical shell model to the X-ray distribution, from which we

1148: measured the X-ray center of the SNR to be at

1149: RA=$01^\mathrm{h}33^\mathit{m}11\fs73$,

1150: Dec=+30$\arcdeg38\arcmin41\farcs9$ (J2000), and 

1151: the dimensions of the SNR shell in projection to be

1152: $\sim 21.0\,\mathrm{pc} \times 19.6\,\mathrm{pc}$ at the assumed distance

1153: to M33.  The \Chandra{} data show the emission to be asymmetric,

1154: with the eastern rim (closest to the bright \HII{} region cores)

1155: roughly $5$ times brighter than the rest of the SNR.  This asymmetry

1156: suggests that the SNR is interacting with the \HII{} region and further

1157: supports the argument that the SNR is actually embedded in,

1158: rather than merely in projection against, the \HII{} region 

1159: \citep{1993ApJ...418..743G}.

1160: The association with NGC\,592 suggests that the progenitor could have

1161: been a high mass star, and that the SNR was the result of a core collapse

1162: supernova.

1163: 

1164: We augmented the high resolution X-ray data with other \chase{}

1165: data taken further off-axis in order to improve the statistics for

1166: X-ray spectral analyses.  By fitting with the {\tt XSPEC sedov} model,

1167: we obtained an estimate for the shock temperature 

1168: $kT = 0.46_{-0.02}^{+0.01}\,\keV{}$, an average preshock 

1169: ISM (hydrogen) density of 

1170: $n_\mathrm{0} \approx 1.7\pm0.3\,\cm^{-3}$, and an ionization

1171: timescale $\tau \equiv n_\mathit{e,s} \,t_0 

1172:  \approx 2.1_{-0.3}^{+0.2}\times 10^{12}\,\mathrm{cm}^{-3}\,\mathrm{s}$.

1173: Although the imaging data show the SNR to be somewhat asymmetric,

1174: the global average properties can be described approximately by a Sedov model

1175: \citep{1999A&A...344..295H}.

1176: The detailed properties would be expected to vary azimuthally around the

1177: SNR because of the variation in the preshock density.  The surface 

1178: brightness will vary most strongly with density ($\propto n^{2}$),

1179: the postshock temperature varies $\propto n$, and the shock velocity

1180: is relatively insensitive to the ambient density ($\propto n^{1/2})$.

1181: However, fits to the high spatial resolution data for sectors 

1182: containing the brightest emission (region BE) and the rest of the 

1183: SNR (region FW) did not show any significant differences 

1184: in spectral properties.

1185: 

1186: From the {\tt sedov} model estimates, we can also derive average 

1187: values for other properties.  The expansion velocity of the SNR 

1188: is $v_\mathit{s} \approx 611\pm12\,\kms{}$

1189: based on the observed size of the remnant and the postshock temperature.

1190: By assuming a Sedov expansion rate, we obtain a dynamical age estimate

1191: of $6500\pm600\,\mathrm{yr}$.  The average postshock electron density,

1192: $n_\mathrm{e, s} = 4.8\,n_0$, where $n_0 = 1.7\pm 0.3$, combined

1193: with the ionization timescale, $\tau$, yields an ionization age

1194: estimate of $8200\pm1700\,\mathrm{yr}$.  The estimates are reasonably 

1195: consistent with each other, and the weighted mean of the ages

1196: is $6700\pm 600\,\mathrm{yr}$.  The large amount of swept up 

1197: mass, $260\pm80\,M_\odot$, also supports the argument that 

1198: \oursnr{} is at least a middle aged SNR.  

1199: 

1200: Our fitted abundance value, $0.45_{-0.09}^{+0.12}$ solar, is 

1201: consistent with the expected half-solar abundances appropriate to the 

1202: galactocentric radius

1203: of \oursnr{}.  Although clumps of enhanced ejecta have been found in 

1204: other SNRs of a similar age (e.g., SNR\,0506-68.0 (N23); 

1205: \citealt{2006ApJ...645L.117H}),

1206: we do not find evidence here for enhanced abundances due to ejecta.

1207: Fits to the SNR as a whole will dilute any ejecta signature in the

1208: much larger mass of swept up ISM, while much smaller extraction regions,

1209: BE and FW, yielded poor counting statistics for the available 190\,ks of

1210: high resolution X-ray imaging data.

1211: 

1212: The total X-ray luminosity (0.25--4.5\,\keV{}) of the source is estimated to 

1213: be $(1.2\pm0.2)\times 10^{37}\,\mathrm{erg}\,\mathrm{s}^{-1}$ (absorbed),

1214: or

1215: $(1.7\pm0.3)\times 10^{37}\,\mathrm{erg}\,\mathrm{s}^{-1}$

1216: (unabsorbed),

1217: in good agreement with the \XMM{} result of 

1218: $(1.13\pm0.14)\times 10^{37}\,\erg{}\,\second^{-1}$ (absorbed),

1219: based on the observations of \citet{2004A&A...426...11P}.

1220: We searched for any emission from a point source or plerion.

1221: We found no evidence for a significant excess of hard emission, and obtained

1222: an upper limit of 

1223: $L_\mathit{X} \le 1.2\times 10^{35}\,\mathrm{erg}\,\mathrm{s}^{-1}$

1224: (2--8 keV) for the unabsorbed luminosity of any hard pointlike source if 

1225: present.

1226: 

1227: We compared the X-ray images with available optical data

1228: and found some puzzling morphological differences.  

1229: In contrast to the slightly elliptical X-ray remnant, the optical remnant

1230: is more oblong.  The bright optical emission agrees with the X-ray

1231: contours in the north and northwest, but falls off more rapidly in the

1232: east, with the region of brightest X-ray emission almost devoid

1233: of optical emission (although this comparison is hampered by the

1234: presence of several bright stars).  To the southwest, the

1235: optical emission extends considerably beyond the X-ray

1236: emission.  The lack of optical emission to accompany the bright X-ray

1237: emission in the east could be the result of a relatively recent interaction 

1238: with denser material, in which the shocked gas has not had time to

1239: become radiative; the apparent overpressure in the BE region compared to 

1240: the FW region would support that interpretation. The edge of the 

1241: \fion{S}{ii} emission in the east coincides with the rim of an

1242: \HII{} loop, and the bright X-ray emission may be the result of an interaction

1243: with a bubble driven by the stars associated with the southern \HII{}

1244: region core.  In that case, the lack of \fion{S}{ii} emission may be

1245: the result of the photoionizing flux from the stars in the 

1246: \HII{} region core preventing the postshock gas from recombining.

1247: On the other hand, the lack of X-ray emission to accompany the 

1248: optical emission in the southwest is also puzzling.  It may be that 

1249: even with 190\,\ks{} of data, the X-ray emission in the southwest is too 

1250: faint to be picked up.  The complex multiwavelength

1251: morphology suggests that the shell is expanding into a non-uniform density

1252: medium, as may be expected for an SNR evolving in an active star

1253: forming region.

1254: 

1255: \acknowledgments

1256: 

1257: This work has made use of SAOImage 

1258: DS9\footnote{\tt http://hea-www.harvard.edu/RD/}, developed 

1259: by the Smithsonian Astrophysical Observatory (Joye \& Mandel 2003), 

1260: the {\tt XSPEC}\footnote{\tt http://xspec.gsfc.nasa.gov/}

1261: spectral fitting package (Arnaud 1996), the 

1262: FUNTOOLS\footnote{\tt http://hea-www.harvard.edu/RD/funtools} 

1263: utilities package,

1264: the HEASARC 

1265: FTOOLS\footnote{\tt 

1266: http://heasarc.gsfc.nasa.gov/docs/software/lheasoft/ftools/}

1267: package, 

1268: and the 

1269: CIAO\footnote{\tt http://cxc.harvard.edu/ciao/}

1270: (\Chandra{} Interactive Analysis of Observations) package.

1271: Support for this work was provided by the National Aeronautics and Space

1272: Administration through \Chandra{} Award Number G06-7073A issued by the 

1273: \Chandra{}

1274: X-ray Observatory Center, which is operated by the Smithsonian

1275: Astrophysical Observatory for and on behalf of the National Aeronautics

1276: Space Administration under contract NAS8-03060.  TJG, PPP, and RJE 

1277: acknowledge support under NASA contract NAS8-03060.

1278: 

1279: {\it Facilities:} \facility{CXO (ACIS)}

1280: 

1281: 

1282: \begin{thebibliography}{}

1283: 

1284: %

1285: 

1286: \bibitem[Anders \& Grevesse (1989)]{1989GeCoA..53..197A} Anders, E, and Grevesse, N., \gca, 53, 197

1287: 

1288: \bibitem[Arnaud (1996)]{1996ASPC..101...17A} Arnaud, K.~A. 1996, in

1289:   ASP Conf. Ser. 101: Astronomical Data Analysis  Software and Systems

1290:   V, ed. G.~H. Jacoby \& J. Barnes

1291: 

1292: \bibitem[Banas et al. (1997)]{1997ApJ...480..607B} Banas, K. R., Hughes, J. P.,

1293: Bronfman, L., and Nyman, L. A. 1997, \apj, 480, 607

1294: 

1295: \bibitem[Blair \& Kirshner (1985)]{1985ApJ...289..582B}

1296:   Blair, W. P., and Kirshner, R. P.  1985, 289, 582

1297: 

1298: \bibitem[Blair \& Long (2004)]{2004ApJS..155..101B} 

1299:   Blair, W. P. \& Long, K. S. 2004, \apjs, 155, 101

1300: 

1301: \bibitem[Bohlin et al. (1978)]{1978ApJ...224..132B}

1302:   Bohlin, R.~C., Savage, B.~D., Drake, J.~F.  1978, \apj, 224, 132

1303: 

1304: \bibitem[Borkowski et al. (2001)]{2001ApJ...548..820B}

1305:   Borkowski, K.~J., Lyerly, W.~J., and Reynolds, S.~P.  2001, \apj, 548, 820

1306: 

1307: \bibitem[Bosch et al. (2002)]{2002MNRAS.329..481B}

1308:   Bosch, G.,Terlevich, E., Terlevich, R.  2002, \mnras, 548, 820

1309: 

1310: \bibitem[Carter et al. (2003)]{2003ASPC..295..477C}

1311:   Carter, C., Karovska, M., Jerius, D., Glotfelty, K., and Beikman, S.  2003, 

1312:   in ASP Conf. Ser. 295: Astronomical Data Analysis  Software and Systems XII, ed.

1313:   H. E. Payne, R. I. Jedrzewski, \& R. N. Hook, 477

1314: 

1315: %

1316: %

1317: %

1318: 

1319: \bibitem[Dodorico et al. (1978)]{1978A&A....63...63D}

1320:   Dodorico, S., Benvenuti, P., Sabbadin, F.  1978, \aap, 63, 63

1321: 

1322: \bibitem[Fitzpatrick (1986)]{1986AJ.....92.1068F} Fitzpatrick, E. L. 1986, \aj, 

1323:    92, 1068

1324: 

1325: \bibitem[Freedman et al. (2001)]{2001ApJ...553...47F}

1326:   Freedman, W. L., Madore, B. F., Gibson, B. K.,   

1327:         Ferrarese, L.,  Kelson, D. D., Sakai, S.,  Mould, J. R.,   

1328:         Kennicutt, R. C.,  Ford, H. C., Graham, J. A.,   

1329:         Huchra, J. P.,  Hughes, S. M. G.,  Illingworth, G. D.,   

1330:         Macri, L. M.,  and Stetson, P. B. 2001, \apj, 553, 47

1331:   

1332: \bibitem[Ghavamian et al. (2005)]{2005AJ....130..539G}

1333:   Ghavamian, P., Blair, W. P., Long, K. S., Sasaki, M., 

1334:         Gaetz, T. J., and Plucinsky, P. P. 2005, \aj, 130, 539

1335: 

1336: \bibitem[Gordon et al. (1993)]{1993ApJ...418..743G}

1337:   Gordon, S. M., Kirshner, R. P., Duric, N., and Long, K. S. 1993, \apj, 418, 743

1338: 

1339: \bibitem[Gordon et al. (1998)]{1998ApJS..117...89G}

1340:   Gordon, S. M., Kirshner, R. P., Long, K. S., Blair, W. P., Duric, N., 

1341:   and Smith, R. C. 1998, \apjs, 117, 89 [GKL98]

1342: 

1343: \bibitem[Gordon et al. (1999)]{1999ApJS..120..247G}

1344:   Gordon, S.~M.,  Duric, N.,  Kirshner, R.~P.,  Goss, W.~M., 

1345:         Viallefond, F.  1999, \apjs, 120, 247

1346: 

1347: \bibitem[Haberl \& Pietsch (2001)]{2001A&A...373..438H}

1348:   Haberl, F., and Pietsch, W. 2001, \aap, 373, 438

1349: 

1350: \bibitem[Henry \& Howard (1995)]{1995ApJ...438..170H}

1351:   Henry, R. B. C., and Howard, J. W.  1995, \apj, 438, 170

1352: 

1353: \bibitem[Hnatyk \& Petruk (1999)]{1999A&A...344..295H}

1354:   Hnatyk, B., and Petruk, O.  1999, \aap, 344, 295

1355: 

1356: \bibitem[Hodge \& Sexton (1966)]{1966AJ.....71..363H}

1357:   Hodge, P. W., and Sexton, J. A.  1966, \aj, 71, 363

1358: 

1359: \bibitem[Hughes et al. (1998)]{1998ApJ...505..732H}

1360:   Hughes, J.~P., Hayashi, I., and Koyama, K.  1998, \apj, 505, 732

1361: 

1362: \bibitem[Hughes et al. (2006)]{2006ApJ...645L.117H}

1363:   Hughes, J.~P.,  Rafelski, M.,  Warren, J.~S.,  Rakowski, C., 

1364:         Slane, P., Burrows, D., and Nousek, J.  2006, \apj, 645, 117

1365: 

1366: \bibitem[Li et al. (2004)]{2004ApJ...610.1204L}

1367:   Li, J., Kastner, J.~H., Prigozhin, G.~Y., Schulz, N.~S., Feigelson, E.~D.,

1368:   Getman, K.~V. 2004, \apj, 610, 1024

1369: 

1370: \bibitem[Long et al. (1990)]{1990ApJS...72...61L}

1371:   Long, K. S., Blair, W. P., Kirshner, R. P., and Winkler, P. F.  1990, \apjs, 72, 61

1372: 

1373: \bibitem[Long et al. (1996)]{1996ApJ...466..750L}

1374:   Long, K. S., Charles, P. A., Blair, W. P., and Gordon, S. M. 1996, \apjl, 466, L750

1375: 

1376: \bibitem[Long et al. (1981a)]{1981ApJ...246L..61L}

1377:   Long, K.~S., Dodorico, S., Charles, P.~A., and Dopita, M.~A. 1981a, \apj, 246, L61

1378: 

1379: \bibitem[Long et al. (1981b)]{1981ApJ...248..925L}

1380:   Long, K. S., Helfand, D. J, and Grabelsky, D.~A.  1981a, \apj, 248, 925

1381: 

1382: \bibitem[Maeder (1990)]{1990A&AS...84..139M}

1383:   Maeder, A.  1990, \aaps, 84, 139

1384: 

1385: \bibitem[Massey et al. (2006)]{2006AJ....131.2478M}

1386:   Massey, P.,  Olsen, K.~A.~G.,  Hodge, P.~W.,  Strong, S.~B., 

1387:         Jacoby, G.~H.,  Schlingman, W.,  and Smith, R.~C. 2006, \aj, 131, 2478

1388: 

1389: \bibitem[McKee \& Hollenbach (1980)]{1980ARA&A..18..219M}

1390:    McKee, C.~F., and Hollenbach, D.~J. 1980,

1391:    \araa, 18, 219

1392: 

1393: \bibitem[McNeil \& Winkler (2006)]{2006AAS...208.0301M}

1394:   McNeil, E.~K., and Winkler, P.~F. 2006, \baas, 39, 80

1395: 

1396: \bibitem[Misanovic et al. (2006)]{2006A&A...448.1247M}

1397:   Misanovic, Z., Pietsch, W., Haberl, F., Ehle, M., Hatzidimitriou, D.,

1398:   and Trinchieri, G. 2006, \aap, 448, 1247

1399: 

1400: \bibitem[Pellerin (2006)]{2006AJ....131..849P}

1401:   Pellerin, A.  2006, \aj, 131, 849

1402: 

1403: \bibitem[Pietsch et al. (2003)]{2003AN....324...85P}

1404:   Pietsch, W.,  Ehle, M.,  Haberl, F.,  Misanovic, Z., Trinchieri, G.  2003,

1405:   Astron. Nach, 324, 85

1406: 

1407: \bibitem[Pietsch et al. (2004)]{2004A&A...426...11P}

1408:   Pietsch, W.,  Misanovic, Z.,  Haberl, F.,  Hatzidimitriou, D., 

1409:         Ehle, M.,  Trinchieri, G.  2004, \aap, 426, 11

1410: 

1411: \bibitem[Rappaport et al. (1974)]{1974ApJ...194..329R}

1412:   Rappaport, S., Doxsey, R., Solinger, A., and Borken, R.  1974, \apj, 194, 329

1413: 

1414: \bibitem[Sabbadin (1979)]{1979A&A....80..212S}

1415:   Sabbadin, F.  1979, \aap, 80, 212

1416: 

1417: \bibitem[Sasaki et al. (2007)]{2007Sasaki..in..prep}

1418:   Sasaki, M., et al. 2007, in preparation

1419: 

1420: \bibitem[Schulman \& Bregman (1995)]{1995ApJ...441..568S}

1421:   Schulman, E., and Bregman, J.~N.  1995, \apj, 441, 568

1422: 

1423: \bibitem[Smith et al. (1993)]{1993ApJ...407..564S}

1424:   Smith, R. C., Kirshner, R. P.,  Blair, W.~P., 

1425:         Long, K.~S.,  and Winkler, P.~F.  1993, \apj, 407, 564

1426: 

1427: \bibitem[Stark et al. (1995)]{1995yCat.8028....0S}

1428:   Stark, A.~A.,  Gammie, C.~F.,  Wilson, R.~W.,  Bally, J., 

1429:         Linke, R.~A.,  Heiles, C.,  and Hurwitz, M.  1995,

1430:         VizieR Online Data Catalog, 8028, 0

1431: 

1432: \bibitem[Trinchieri et al. (1988)]{1988ApJ...325..531T}

1433:   Trinchieri, G., Fabbiano, G., and Peres, G.  1988, \apj, 325, 531

1434: 

1435: \bibitem[Vaiana et al. (1981)]{1981ApJ...245..163V}

1436:   Vaiana, G.~S.,  Cassinelli, J.~P.,  Fabbiano, G.,  

1437:         Giacconi, R.,  Golub, L.,  Gorenstein, P.,  Haisch, B.~M., 

1438:         Harnden, Jr., F.~R.,  Johnson, H.~M.,  Linsky, J.~L., 

1439:         Maxson, C.~W.,  Mewe, R.,  Rosner, R.,  Seward, F., 

1440:         Topka, K.,  Zwaan, C. 1981, \apj, 245, 163

1441: 

1442: \bibitem[Vancura et al. (1992)]{1992ApJ...394..158V}

1443:   Vancura, O., Blair, W. P., Long, K. S., and Raymond, J. C  1992, \apj, 394, 158

1444: 

1445: \bibitem[Viallefond et al. (1986)]{1986A&AS...64..237V}

1446:   Viallefond, F., Goss, W.~M., van der Hulst, J.~M., and Crane, P.~C.  1986,

1447:   \aaps, 64, 237

1448: 

1449: \bibitem[Weisskopf \& Hughes (2006)]{2006Book...W}

1450:   Weisskopf, M.~C., and Hughes, J.~P. 2007, 

1451:   in Astrophysics Updates 2, 

1452:   ed. John W. Mason (Chichester: Praxis Publishing), 55

1453: 

1454: \bibitem[Zaritsky et al. (1989)]{1989AJ.....97...97Z}

1455:   Zaritsky, D., Elston, R., and Hill, J.~M. 1989, \aj, 97, 97

1456: 

1457: 

1458: \end{thebibliography}

1459: 

1460: 

1461: %

1462: 

1463: %

1464: 

1465: 

1466: %

1467: 

1468: %

1469: %

1470: %

1471: %

1472: %

1473: 

1474: \end{document}

1475: