1: %\documentclass[12pt,preprint]{aastex}

2: \documentclass{emulateapj}

3: %\usepackage{emulateapj}

4: \usepackage{apjfonts}

5: %\usepackage[T1]{fontenc}

6: %\usepackage[latin1]{inputenc}

7: %\setcounter{tocdepth}{3}

8: 

9: \makeatletter

10: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%% New commands.

11: \newcommand{\psr}{PSR B1823--13}

12: \newcommand{\pwn}{G18.0--0.7}

13: \newcommand{\ros}{{\sl ROSAT}}

14: \newcommand{\chan}{{\sl Chandra}}

15: \newcommand{\cxo}{{\sl CXO}}

16: \newcommand{\einst}{{\sl Einstein}}

17: \newcommand{\xmm}{{\sl XMM-Newton}}

18: \newcommand{\edot}{\dot{E}}

19: \newcommand{\fdeg}{\hbox{$.\!\!^{\circ}$}}

20: \newcommand{\fars}{\hbox{$.\!\!^s$}}

21: \def\aciss3{{ACIS-S3}}

22: \def\asca{{\sl ASCA}}

23: \def\pdot{\dot{P}}

24: \def\edot{\dot{E}}

25: \def\fluxu{ergs cm$^{-2}$ s$^{-1}$}

26: \def\lumu{ergs s$^{-1}$}

27: \def\ssbu{ergs cm$^{-2}$ s$^{-1}$ arcsec$^{-2}$}

28: \def\sbu{ photons cm$^{-2}$ s$^{-1}$ arcsec$^{-2}$}

29: \def\scbu{counts s$^{-1}$ arcsec$^{-2}$}

30: \def\msol{{$M_\odot$}}

31: \def\ls{\lower 2pt \hbox{$\;\scriptscriptstyle \buildrel<\over\sim\;$}}

32: 

33: 

34: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%% User specified LaTeX commands.

35: 

36: %\slugcomment{}

37: \shorttitle{{\sl Chandra}  Observation of PSR B1823--13 }

38: \shortauthors{Pavlov, Kargaltsev, \& Brisken}

39: 

40: %\usepackage{babel}

41: \makeatother

42: \begin{document}

43: 

44: \title{Chandra Observation of

45: PSR B1823--13

46:  and its Pulsar Wind Nebula}

47: 

48: 

49: \author{G.\ G.\ Pavlov\altaffilmark{1}, O.\ Kargaltsev\altaffilmark{1},

50: and W.\ F.\ Brisken\altaffilmark{2}}

51: 

52: 

53: %\affil

54: \altaffiltext{1}{Dept. of Astronomy and Astrophysics,

55: Pennsylvania State

56: University, 525 Davey Lab., University Park,

57: PA 16802; pavlov@astro.psu.edu, oyk1@psu.edu}

58: \altaffiltext{2}{National Radio Astronomy Observatory, P.O. Box O, Socorro, NM 87801; wbrisken@nrao.edu}

59: 

60: 

61: %\today

62: 

63: \begin{abstract}

64: We

65: report on an observation of the Vela-like pulsar B1823$-$13 and

66: its synchrotron nebula with the {\sl Chandra X-ray Observatory}.

67: The pulsar's

68: spectrum fits a power-law model with a photon index $\Gamma_{\rm PSR}

69: \approx 2.4$ for the plausible hydrogen column density $n_{\rm H} =

70: 1\times 10^{22}$ cm$^{-2}$, corresponding to the luminosity $L_{\rm

71: PSR}\approx  8\times 10^{31}$ ergs s$^{-1}$ in the 0.5--8 keV band, at a

72: distance of 4 kpc. The pulsar

73: radiation likely includes magnetospheric and thermal

74: components, but they cannot be reliably separated because of the small

75: number of counts detected and strong interstellar absorption.

76: The pulsar is

77: surrounded by a compact, $\sim 25''\times 10''$,

78:  pulsar wind nebula (PWN) elongated in the east-west direction, which

79: includes a brighter inner component

80: $\approx 7''\times 3''$,

81: elongated in the northeast-southwest direction.

82: The slope of the compact PWN spectrum is $\Gamma_{\rm comp}\approx 1.3$,

83: and the 0.5--8 keV luminosity is $L_{\rm comp}

84: \sim 3\times 10^{32}$

85: ergs s$^{-1}$.

86: The compact PWN is surrounded by asymmetric

87: diffuse emission (extended PWN) seen

88: up to at least $2.4'$ south of the pulsar,

89: with a softer spectrum

90: ($\Gamma_{\rm ext} \approx 1.9$ for

91: $n_{\rm H}=1\times10^{22}$ cm$^{-2}$),

92: and the 0.5--8 keV luminosity

93: $L_{\rm ext}\sim 10^{33}$--$10^{34}$ ergs s$^{-1}$.

94: We also measured the pulsar's proper

95: motion using archival VLA data: $\mu_\alpha=23.0\pm 2.5$ mas yr$^{-1}$,

96: $\mu_{\delta}=-3.9\pm 3.3$ mas yr$^{-1}$, which corresponds to

97: the transverse velocity $v_\perp

98: \approx 440$ km s$^{-1}$.

99: The direction of the proper motion is approximately parallel to

100: the elongation of the compact PWN, but

101: it is nearly perpendicular to

102: that of the extended PWN and to the direction towards the center

103: of the bright VHE $\gamma$-ray source HESS J1825--137,

104: which is likely powered by PSR B1823$-$13.

105: \end{abstract}

106: 

107: \keywords{

108: 	ISM: individual (G18.0$-$0.7, HESS J1825$-$137) ---

109:         pulsars: individual (PSR B1823--13 = J1826--1334) ---

110: 	stars: neutron ---

111: 	 X-rays: ISM}

112: 

113: 

114: \section{Introduction}

115: 

116: About a dozen of

117: $\sim 80$ radio pulsars observed in X-rays belong to the group of the so-called

118: Vela-like pulsars, with spindown ages $\tau\equiv P/2\dot{P} = 10$--30 kyr

119: and spindown powers $\dot{E}$ of a few $\times 10^{36}$ ergs s$^{-1}$.

120: Their radiation usually show

121: a thermal component,

122:  possibly emitted from the cooling neutron star (NS) surface,

123: and a nonthermal component with a power-law (PL) X-ray spectrum,

124: presumably originating in the pulsar magnetosphere (see Kaspi et al.\ 2006

125: for a review).

126: These pulsars are enveloped by pulsar wind nebulae

127: (PWNe) whose X-ray emission is due to the wind of ultrarelativistic

128: particles shocked in the ambient medium (Kargaltsev et al.\ 2007, and

129: references therein). Therefore, in addition to studying pulsar

130: magnetospheres and NS

131: cooling, X-ray observations of Vela-like pulsars

132: allow one to study pulsar winds, the ambient medium, and their interaction.

133: 

134: PSR B1823--13 (also known as PSR J1826--1334,

135: hereafter referred to as B1823) is a typical Vela-like

136: pulsar.

137: It was detected in the

138: Jodrell Bank radio survey (Clifton et al.\ 1992).

139: Its period, $P=104$ ms, and period derivative, $\dot{P}=7.5\times 10^{-14}$

140: s s$^{-1}$, imply the spindown age $\tau = 21$ kyr,

141: spindown power $\dot{E}=2.8\times

142: 10^{36}$ ergs s$^{-1}$, and surface magnetic field

143: $B=2.8\times 10^{12}$ G.

144: The distance to the pulsar,

145: $d\approx 4$ kpc, was estimated from the

146: pulsar's dispersion measure (${\rm DM} = 231$ pc cm$^{-3}$)

147: using the models of Galactic electron distribution (Taylor \& Cordes 1993;

148: Cordes \& Lazio 2002).

149: Although such a young pulsar is expected to reside in a supernova remnant

150: (SNR), observations of this field with the Very Large Array (VLA)

151: have not found an SNR, nor a radio PWN

152: (Braun, Goss, \& Lyne 1989;

153: Frail \& Scharringhausen 1997;

154: Gaensler et al.\ 2000).

155: 

156: 

157: This pulsar is particularly interesting because it possibly powers

158: an extended ($\sim 1^\circ$, i.e. $\sim 70$ pc, in size)

159: Very High Energy (VHE) $\gamma$-ray source HESS J1825--137

160: (Aharonian et al.\ 2005,2006).

161: The $\gamma$-ray emission extends asymmetrically

162: south-southwest of B1823, with

163: the peak surface brightness

164: $\approx 10'$ from the pulsar.

165: Aharonian et al.\ (2006, hereafter A06) found a softening

166: of the

167: 0.25--35 TeV spectrum with increasing distance from the pulsar

168: (hence an energy-dependent morphology), with a photon index

169: $\Gamma_\gamma$ increasing from $\approx 1.9$ to $\approx 2.5$.

170: These authors discuss a scenario in which the $\gamma$-rays are produced

171: by Compton upscattering of

172: the cosmic microwave background (CMB) photons off

173: the ultrarelativistic electrons supplied by the pulsar.

174: As the $\gamma$-ray luminosity of the source, $L_\gamma \sim 3\times 10^{35}\,

175: d_4^2\,\,\,

176: {\rm ergs\,\, s}^{-1}$ above 200 GeV [$d_4\equiv d/(4\,{\rm kpc})$],

177: is only a factor of 10 lower than the

178: pulsar's spindown power,

179: A06 propose that

180: the spindown power was significantly higher in the past,

181: when the electrons responsible for the currently observed

182:  TeV emission were produced

183: by the pulsar.

184: 

185: Although the field of B1823 has been

186: observed with \ros, \asca, and \xmm,

187: the X-ray spectrum of the pulsar and the structure of its surroundings

188: remained unclear because of the relatively low angular resolution of

189: those missions.

190: Finley

191: et al.\ (1996, F96 hereafter) observed B1823

192: with the \ros\/ observatory and,

193: despite the small number of source counts detected

194: (e.g., 127 counts from a $5'$ radius circle around the pulsar position

195: in the 14.7 ks PSPC exposure)

196: found evidence of a point source surrounded by a compact PWN

197: of $\sim 20''$ radius,

198: and a  dimmer

199: component

200: extending up to $\sim4'$ south-southwest of the pulsar.

201: Because of the

202: poor statistics, it was impossible

203: to measure the spectra.

204: For an upper limit on the hydrogen column

205: density, $n_{\rm H,22}\equiv n_{\rm H}/(10^{22}\,{\rm cm}^{-2}) < 4$,

206: F96 estimated the upper limit

207: fluxes at a few $\times 10^{-12}$ \fluxu\ level for the emission

208: components, corresponding to luminosities

209: $\lesssim (4$--$8)\times 10^{33} d_4^2$ erg s$^{-1}$ in the 0.5--2.4 keV range.

210: Sakurai et al.\ (2001) were able to detect the extended structure

211: around the B1823 position with {\sl ASCA},

212: but that observation lacked the spatial resolution to separate the components.

213: 

214: Gaensler et al.\ (2003, G03 hereafter)

215: observed B1823 with the \xmm\ EPIC detectors.  The \xmm\ resolution

216: turned out to be too low to

217: resolve the pulsar,

218: but G03 detected two components of

219: the PWN

220: (designated as G18.0--0.7 based on its Galactic coordinates)

221: associated with this pulsar: a ``core'' of a $30''$ extent, elongated

222: in the east-west direction, and

223: a large-scale diffuse component of lower surface brightness,

224:  $\sim 5'$ in extent,

225: seen mostly

226: to the south of the pulsar.

227: Fitting the core and diffuse component spectra with a PL model, G03 found

228: the photon indices $\Gamma_{\rm core} \approx 1.6$ and

229: $\Gamma_{\rm diff}\approx 2.3$, and 0.5--10 keV luminosities

230: $L_{\rm core}\approx 9\times10^{32}d_4^2$ \lumu,

231: $L_{\rm diff} \approx 3\times10^{33} d_4^2$ \lumu,

232: for a foreground hydrogen column density $n_{\rm H,22} = 1.2$.

233: They

234: suggested that the core

235: represents the pulsar

236: wind termination shock at a distance $R_s\approx 15''$ from the

237: the pulsar  while the diffuse component indicates the

238: shocked downstream wind. G03 propose that the asymmetric morphology of

239: the diffuse component with respect to the pulsar is the result of a reverse

240: shock from an associated SNR, which has compressed and distorted the PWN.

241:  They place limits on the pulsar's

242: thermal emission:

243: $kT < 0.147$ keV and

244: bolometric luminosity

245: $L_{\rm bol} < 8.7\times10^{33} d_4^2$ \lumu\

246: (as seen by a distant observer),

247: assuming that the NS emits as a blackbody of a radius $R=12$ km,

248: for

249: $n_{\rm H,22} < 2$.

250: 

251: Because of its superior

252: angular resolution,

253:  \chan\

254: is suited much better than any other X-ray observatory

255:  for studying small-scale

256: extended structures

257: (such as the core of the B1823 PWN)

258: and resolving

259: point sources

260: possibly projected onto such structures.

261: Therefore,

262: our group proposed a {\sl Chandra} observation of

263: B1823,

264: with the goal to separate the pulsar from the PWN,

265: measure their spectra, investigate the PWN morphology with the best

266: possible resolution, and compare the pulsar and PWN properties with

267: those of other Vela-like pulsars.

268: The data from that observation immediately showed that the pulsar

269: can indeed be separated from the PWN,

270: the ``core'' PWN is highly nonuniform in surface brightness,

271: and the overall PWN morphology is more

272: complicated than it had been presumed based on the previous low-resolution

273: X-ray observations.

274: These results

275: have been

276: reported at several conferences (Teter et al.\ 2003a,b; Gonzalez \& Kaspi

277: 2006; Kargaltsev et al.\ 2006), but they have not been published in

278: a refereed journal.

279: In this paper, we decribe the analysis and results of

280: the \chan\ observations

281: and present infrared (IR) and radio images of the field (\S2).

282: As the interpretation of the PWN morphology strongly depends on the

283: direction and magnitude of the pulsar's proper motion, we also

284: report the proper motion measurement in \S3. Implications

285: of our findings for the pulsar and PWN physics, including the

286: VHE $\gamma$-radiation,

287: are discussed in \S4, while \S5 summarizes the results

288: of this work.

289: 

290: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

291: \section{Observations and Data Analysis}

292: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

293: \subsection{{\sl Chandra} ACIS observation}

294:  B1823 was observed with the Advanced CCD Imaging Spectrometer (ACIS)

295: aboard {\sl Chandra} on 2002 October 24

296: (ObsID 2830).

297: Although last 13 ks of the 40,836 s exposure suffered from flaring particle

298: background ($9.7\times 10^{-7}$ counts s$^{-1}$ per pixel, on average),

299: we have checked that filtering out the periods of elevated background

300: at the expense of shortening the exposure time would actually reduce

301: the signal-to-noise ratio for the pulsar and the compact PWN and only

302: very slightly increase it for the large-scale extended PWN;

303: therefore, we chose to use the whole exposure for our analysis.

304: 

305:  The

306: observation was carried out in Faint mode, and the target was

307: imaged on S3 chip, with the radio pulsar position

308:  $\approx35''$ from the optical axis, where the angular resolution

309: is still not degraded.

310:  The other ACIS chips

311: activated during this observation were S1, S2, S4, I2 , and I3.

312: The detector was operated in Full Frame mode, which provides time

313: resolution of 3.24 seconds.

314: We analyzed the data

315: reprocessed by the Chandra X-ray Center (CXC)

316: on 2006 October 11 (rev.\ 2; ASCDSVER 7.6.9).

317: The data were reduced using the Chandra Interactive Analysis of

318: Observations (CIAO) software (ver.\ 3.4; CALDB ver.\ 3.3.0.1).

319: We used the software package XSPEC (ver.\ 11.3.2) for the spectral

320: analysis.

321: 

322: 

323: \begin{figure*}[t]

324: %\epsscale{0.75}

325:  \centering

326: \includegraphics[width=7.2in,angle=0]{f1.eps}

327: \caption{ {\em Top left:} $77''\times 48''$ ACIS-S3 image

328:  of B1823 and its PWN (0.7--7 keV; pixel size $0\farcs49$).

329: {\em Top right:}

330: Adaptively smoothed

331: image (0.7--7 keV; pixel size $0\farcs49$)

332:  of the same region obtained by removing the

333: pipeline pixel-randomization and

334: applying the sub-pixel

335: resolution tool

336: (Tsunemi et al.\ 2001; Mori et al.\ 2001).

337: {\em Bottom left:}

338:  The same image as in the top left panel showing the extraction

339: regions used for the spectral analysis of the PWN and the pulsar (see

340: \S2). `b' and `X' mark the ``blob'' in the compact PWN and a point

341: source unrelated to the PWN, respectively.  {\em Bottom right:}

342:  Heavily binned (pixels size $3\farcs94$; $0.5-8.0$ keV) and

343:  smoothed (with a gaussian kernel of $r=27''$)

344:  ACIS-S3 image of B1823 and its surroundings.

345: The brightness and smoothing scales are chosen

346:  to stress the fainter, more extended emission south of the compact PWN.

347: The polygon shows the region from which we extract counts for the

348: spectral analysis of the large-scale extended emission, with the

349: background estimated from the $r=1.28'$ circle northwest of the

350: PWN. The annulus around the compact PWN is used for measuring

351: the compact PWN background.

352: }

353: \end{figure*}

354: 

355: \begin{figure}

356: %\epsscale{0.75}

357:  \centering

358: \includegraphics[width=3.2in,angle=0]{f2.eps}

359: \caption{

360: {\em Top:}

361: $62''\times  48''$

362: ACIS-S3 image ($0.7-7$ keV; pixel size $0.49''$; smoothed with $1.5''$

363: gaussian kernel)

364: of the immediate surroundings of B1823, including the compact PWN

365: and Source X  (likely a backlground AGN).

366: The

367:   white arrow shows the direction of the pulsar's proper motion

368: (see \S3).

369: The length of the arrow, $7''$,

370:   corresponds to the distance that the pulsar travels during

371: 300 years.

372:  The {\em middle} and {\em bottom} panels show  the  linear profiles

373: extracted by moving the rectangular box

374: ($0.49''\times17''$ and $0.98''\times17.1''$, respectively)

375:  along the horizontal line

376:   shown in the top panel. The one-dimensional point spread function

377: centered on the pulsar position is shown in the middle panel.

378:   }

379: \end{figure}

380: 

381: \subsection{X-ray images of B1823 and its vicinity }

382: 

383: Figure 1 shows the ACIS-S3 images of

384: the B1823 field, with

385: an extended source around the radio pulsar position.

386: The coordinates of the highest peak in the X-ray brightness distribution

387: (which is at the center of a $2\times 2$ pixel island, with

388: equal numbers of counts in each of the 4 pixels)

389: are

390: $\alpha=18^{\rm h}26^{\rm m}13.187^{\rm s}$,

391: $\delta =-13^{\circ}34' 46.63''$ (J2000),

392: with an estimated $1 \sigma$

393: centroiding error of $0.2''$ for each of the coordinates.

394: This position differs by $0.08''$ in right ascension and $0.17''$ in

395: declination from

396: the radio pulsar position,

397: $\alpha = 18^{\rm h}26^{\rm s}13.182^{\rm s}\pm 0.04''$,

398: $\delta =-13^{\circ}34'46.8'' \pm 0.2''$,

399: calculated from the original position in Hobbs et al.\ (2004)

400: for the epoch of the {\sl

401: Chandra} observation (MJD 52,571), using  the  proper motion

402: measurement described in \S3.

403: These differences are smaller than the combined errors

404: ($0.3''$ in right ascension and $0.4''$ in declination, at the 68\% level)

405: of the

406: peak position measurement, radio position, and {\sl Chandra} absolute

407: astrometry.

408: Therefore, there are no doubts that the highest brightness peak corresponds

409: to the pulsar B1823 while the surrounding emission represents its PWN.

410: 

411: We see from Figures 1 and 2 that the PWN morphology is asymmetric

412: and nonuniform.

413: The brightest {\em inner} PWN component

414: (surface brightness in the range of 4--23 counts arcsec$^{-2}$ in 0.7--7 keV),

415: is $\sim 7''\times 3''$ (i.e.\ $0.14\times 0.06$ pc$^2$

416: at $d=4$ kpc) in size.

417: It is approximately centered on

418: the pulsar and

419:  elongated along the northeast-southwest direction

420: (position angle

421: ${\rm P.A.}\sim 50^{\circ}$, measured east of north).

422: 

423: The inner component is

424: embedded in

425: a dimmer (average surface brightness $\approx 1.2$ counts

426: arcsec$^{-2}$ in 0.7--7 keV)

427:  {\em outer} component of the {\em compact PWN}, within a

428: $25''\times 10''$

429: area.

430: It is stretched along the east-west direction,

431: with the pulsar

432: being closer to its east end (i.e. the outer component is more

433: extended in the direction approximately opposite to the direction

434: of the pulsar's proper motion; see \S3).

435: Its surface brightness distribution is rather nonuniform,

436: with a prominent enhancement

437: $\approx8''$ west of the pulsar (we designate it as the ``blob''

438: and mark `b' in the figures).

439:  The smoothed subpixel-resolution image shown in Figure 1

440: ({\em top right}) suggests that the

441: blob is connected to the

442: inner PWN component  by a faint ``bridge''.

443: We have checked that there are no optical-NIR counterparts to the blob in

444: the catalogs available (e.g.,

445: the nearest 2MASS\footnote{Two Micron All Sky Survey

446: (Skrutzkie et al.\ 2006).}

447:  counterpart is $6.5''$ away).

448: 

449: Another compact

450: source is seen in the images $\approx 18''$ east

451: of the pulsar. This object, CXOU 182614.4$-$133448

452: (hereafter referred to as Source X) is point-like. The analysis of

453: its spectrum (\S2.5) shows that it is

454: likely a background AGN,

455: unrelated to B1823.

456: 

457: The smoothed image in Figure 2 ({\em top})

458:  shows

459: marginal evidence of two $\sim 20''$-long ``tails'', apparently emanating

460: from the pulsar in the west-northwest and west-southwest directions.

461: The linear profiles in the middle and bottom panels of Figure 2

462:     show that the outer PWN structure is

463: discernible

464: above the local background level of

465: $\sim 0.2$ counts arcsec$^{-2}$

466:  up to

467: $\approx 7''$ and $\approx 23''$

468: east and west of the pulsar,

469: respectively.

470: 

471:  The compact PWN (comprised of the above-described inner and outer

472: components) is, in turn, immersed in an even

473: dimmer large-scale diffuse

474:    emission

475: apparent in heavily binned

476: images.

477: As it is natural to assume that this diffuse emission is also produced

478: by the wind from B1823,

479: we will call it the {\em extended} PWN.

480: It is mostly concentrated southward

481:    of the compact PWN, and it is discernible in the ACIS images

482: up to

483: the edge of the S3 chip (i.e.\ $\approx2.4'$ from the

484:    pulsar; Fig.\ 1, {\em bottom right}).

485: (It likely extends further into the S2 chip, but

486: the lower sensitivity of this chip

487:  does not allow a reliable analysis of the faint

488: diffuse

489:  emission.)

490: The extended

491: PWN is somewhat better seen in the {\sl XMM-Newton}

492: EPIC MOS1+MOS2 image shown in Figure 3{\em c}

493: (see also G03).

494: However,

495: the true extent of this

496: dim diffuse component is difficult to determine

497: even in the large field-of-view of the MOS detectors

498: because of the strong vignetting at large  off-axis angles and

499:    high EPIC background.

500: 

501: \begin{figure*}[t]

502: %\epsscale{0.75}

503:  \centering

504: \includegraphics[width=6.5in,angle=0]{f3.eps}

505:  \caption{

506: Large-scale images of the B1823 vicinity and the central part

507: of HESS\,J1825

508: observed with different telescopes and at different wavelengths.

509:   All eight panels show the same area on the sky.   The TeV

510: brightness contours

511:  (adopted from

512: A06) are overlayed on top of

513: the images in panels {\em b--f}.

514: {\em a:} ACIS

515: S3, S2, S1, I2 and I3

516: image (0.7$-$7 keV; pixel size $3\farcs94$),

517: corrected for the exposure map nonuniformity.

518: {\em b:}

519:  The same image with point sources removed

520: (except for B1823)

521: and gaussian ($15.7''$ kernel) smoothing applied.

522:  {\em c:}   {\sl XMM-Newton} EPIC MOS1+MOS2  combined

523: image (0.5--8 keV; pixel size $10''$; smoothed with the $40''$ gaussian

524:  kernel).

525: {\em d:}

526:  {\sl Spitzer} IRAC 8\,$\mu$m image from the GLIMPSE survey.

527: {\em e:}  {\sl Spitzer} MIPS 24\,$\mu$m image from the GLIMPSE survey.

528: {\em f:} NRAO VLA Sky Survey (NVSS)

529: image at 1.4 GHz.

530: The star in panels {\em b} and {\em c} marks the location of

531: the brightest part of the extended

532:    IR and radio source A

533:  seen in the {\sl Spitzer} and NVSS images.

534: The diamond in {\em panels d--f} marks the B1823 position.

535: The small white circle

536:  marks the position of the brightness peak in the distribution

537:    of the extended TeV emission from HESS\,J1825.

538:  The projected place of birth of the pulsar estimated from the measured proper motion (see \S3)

539:  and pulsar's characteristic age ($\tau=21$ kyr) is shown by the dashed circle.}

540: \end{figure*}

541: 

542: \subsection{PWN spectrum}

543: 

544: We extracted the compact PWN spectrum\footnote{As we found no statistically

545: significant differences between the spectra

546: of the inner and outer components of the compact PWN, we present here

547: only the spectrum of the entire compact PWN.} (Fig.\ 4, {\em top})

548: from the

549: ellipse  shown in Figure 1 ({\em bottom left}),

550: excluding the counts

551: within the $r=3''$ circle centered on the X-ray position of the pulsar (see

552: \S2.1) and restricting the energy range to 0.7--7 keV.

553: This region of

554: 225 arcsec$^2$ area

555:   contains

556:  415 counts.

557: We estimated the background from the annulus of

558: 2916 arcsec$^2$ area

559:  shown in Fig.\ 1 ({\em bottom right}),

560: which contains

561:   511 counts. Thus, there are $375\pm 20$ compact PWN counts in the aperture

562: chosen.

563: 

564: To analyze the spectrum of the

565: extended PWN

566: (Fig.\ 4, {\em bottom}), we

567: chose the polygon encompassing the region of brightest

568: emission south of the pulsar,

569: with the area of  3.4 arcmin$^{2}$  (see Fig.\ 1, {\em bottom right}),

570: which contains

571: 2329 counts in the 0.7--7 keV band.

572: The background was

573: extracted from the $r=1.28'$ circular region

574: shown in the same panel, which contains

575:  2004 counts. Subtracting the background contribution, we obtain

576: $1008\pm 57$ extended PWN counts in the polygon aperture.

577: (We have also tried several

578: other background regions but this

579: did not lead to significant changes in the

580: spectral fits.)

581: 

582: First, we fit the  compact and extended PWN spectra with the absorbed power-law

583: (PL) model, allowing the hydrogen column density

584: to vary. The confidence contours obtained from these fits are shown in

585: Figure 5.

586: We see that the 90\% confidence contours for the compact and extended PWN

587: do not overlap, which suggests that the two spectra are different.

588: Since we do not expect the hydrogen column density,

589: $n_{\rm H,22}\sim 1$, to differ significantly

590: on such a small, $\sim 2'$, scale,

591: we ascribe the difference to different spectral slopes,

592: $\Gamma_{\rm ext}- \Gamma_{\rm comp} \approx 0.5$.

593: 

594: To estimate the spectral slopes more accurately,

595: it would be useful to restrict

596: the range of $n_{\rm H}$

597: (which would also be helpful for the spectral analysis of the

598: pulsar -- see \S2.4). The fits with free $n_{\rm H}$ give

599: $n_{\rm H,22}

600: = 1.24\pm0.26$ and $n_{\rm H,22}=0.72\pm0.15$ for the compact

601: and extended PWNe, respectively (the uncertainties

602: correspond to the 68\%

603: confidence level for a single interesting parameter).

604: The  values

605: of $n_{\rm H,22}$

606: obtained by G03 using the {\sl XMM-Newton} data

607: are in the range

608: of 0.8--1.7, depending upon the

609: spectral model and extraction region chosen.

610: The estimate based on

611: the pulsar's dispersion measure (DM$=231$ cm$^{-3}$)

612: and the usual assumption of 10\% degree of ionization of

613: the interstellar medium

614: (ISM)

615: gives

616: $n_{\rm H,22}=0.7$.

617: This is close to the lower limit,

618: $n_{\rm H, 22}\gtrsim 0.7$, estimated from the observed colors of

619: the foreground ($d=2.9$ kpc) O6 star HD 169727,

620: projected $\approx 8'$ southeast

621: of the pulsar (F96).

622: Finally,

623: the total Galactic HI column in that direction is $n_{\rm HI,22}

624: =1.4$--1.5

625: (Dickey \& Lockman 1990).

626: Since the $n_{\rm H}$ values deduced from an X-ray spectrum

627: under the assumption of standard element abundances usually

628: exceed the $n_{\rm HI}$ measured from 21 cm observations by a

629: factor of 1.5--3 (e.g., Baumgartner \& Mushotzky 2005),

630: this $n_{\rm HI}$ value does not contradict to those estimated

631: from the X-ray fits and the dispersion measure.

632:  Thus,

633: with allowance for

634: different kinds of uncertainties, it seems reasonable to assume

635: $0.7 \lesssim n_{\rm H,22}\lesssim 1.3$,

636: with $n_{\rm H,22}\approx 1$ being the most plausible estimate.

637: 

638: 

639:  With the hydrogen column density fixed at $n_{\rm H,22}=1.0$,

640: the absorbed PL fits to the compact and extended PWN

641: yield photon indices

642: $\Gamma_{\rm comp}\simeq 1.3$ and $\Gamma_{\rm ext}\simeq 1.9$,

643: and 0.5--8 keV luminosities $L_{\rm comp}\simeq 3\times 10^{32}$

644: and $L_{\rm ext}\sim 9\times 10^{32}$ ergs s$^{-1}$

645: (see also Fig.\ 6 and

646:    Table 1; the extended PWN luminosity is for the polygon region).

647: 

648: \subsection{Pulsar spectrum}

649: 

650: Although the peak of the surface brightness distribution is

651: clearly dominated by the pulsar,

652: the surface brightness of the inner PWN

653: grows toward the pulsar's position, which introduces some ambiguity

654: in separating the pulsar contrubution from that of the PWN

655: and complicates the

656: measurement of the pulsar spectrum.

657: However, based on the surface brightness distribution near the peak

658: (see, e.g., the one-dimensional profiles in Fig.\ 2),

659: it seems reasonable to assume that the PWN

660: surface brightness reaches a plateau of about $3''$ in size,

661: centered on the pulsar, which can be used to estimate the

662: local background at the pulsar position.

663: Therefore, we extracted the pulsar

664: spectrum from a small circular aperture (small black circle in Fig.\ 1,

665: {\em bottom left}) with the radius of 1 ACIS pixel ($\simeq 0.49''$,

666: 55\% encircled energy radius)

667: and took the background

668: from

669: the $0.74''<r<1.5''$ circular annulus

670: centered on the pulsar

671: (shown by the

672: white circles in Fig.\ 1, {\em bottom left}).

673:  The background

674: contributes $\approx 22$ counts

675: to 62 counts extracted from the source region.

676:  Given the

677: small number of source counts

678:  and the appreciable background contribution,

679: we chose not to subtract the background channel by channel but

680: rather fit the background spectrum (modeled as an absorbed PL model

681: with the same $n_{\rm H}$ as for the source) simultaneously with the

682: source spectrum.

683: 

684: To obtain better constrained fits with the small number of counts

685: available, we are forced to freeze the hydrogen column density.

686: Since the interpretation of the pulsar spectrum (e.g., thermal vs.\

687: nonthermal) may depend on the choice of the $n_{\rm H}$ value,

688:  we  fit the

689: pulsar spectrum with the absorbed PL model for three

690: plausible

691: values of the hydrogen column density:

692:  $n_{\rm H,22}=0.7$, 1.0, and 1.3 (see Table 2 and Fig.\ 7).

693: As expected, the inferred photon indices depend on the

694: choice of $n_{\rm H}$, changing from $\Gamma_{\rm

695: PSR}=2.0\pm0.4$ to $3.0\pm0.6$ with increasing

696: $n_{\rm H}$ (see

697: Table 2 for details).

698:  The pulsar's absorbed flux

699: is $F_{\rm PSR}=(1.6\pm0.4) \times10^{-14}$ ergs cm$^{-2}$ s$^{-1}$

700: in the 0.7--7 keV band (corrected for the background and finite

701: aperture).

702:  The unabsorbed pulsar luminosity in the 0.5--8

703: keV band increases from $L_{\rm PSR} \approx 0.6\times 10^{32}$ to $1.1

704: \times 10^{32}$ ergs cm$^{-2}$ s$^{-1}$, respectively.

705: 

706: The steepness of the best-fit PL spectra, especially in the case of

707: larger $n_{\rm H}$, indicates the presence of thermal emission from

708: the NS surface, such as seen in other Vela-like pulsars (e.g.,

709: Pavlov et al.\ 2001).

710: However, adding the

711: second blackbody (BB) component is not justified statistically (i.e.\ the

712: contribution of the BB component is negligible, and the fitting

713: parameters of the BB component are unconstrained) at $n_{\rm

714: H,22}=0.7$ and 1.0, while in the case of

715:  $n_{\rm H,22}=1.3$ the

716: fit

717:  converges

718: to a reasonable (although still poorly constrained)

719:  values of the surface temperature $kT\simeq 100$ eV,  projected

720:  emitting area $\mathcal{A}\sim 500d_4^2$ km$^{2}$

721: (corresponding to the bolometric luminosity

722: $L_{\rm bol} \equiv 4{\mathcal{A}}\sigma T^4

723: \sim 2\times 10^{33}

724: d_4^2$ ergs s$^{-1}$),

725:  and the slope of the PL component $\Gamma_{\rm PSR}

726: \simeq 1.9$.

727: The fitting parameters become better constrained if we fix the

728: area, but

729: the best-fit temperature and luminosity depend on the value of

730: $\mathcal{A}$

731: because of the strong $\mathcal{A}$-$T$ correlation.

732: For instance,

733: $kT$ increases from 97 to 139 eV, and $L_{\rm PSR}^{\rm bol}$

734: decreases from $1.8\times 10^{33}$ to $3.0\times 10^{32}$ ergs s$^{-1}$,

735:   when $\mathcal{A}$ decreases

736: from 500 km$^{2}$ (approximate projected area of the NS surface)

737: to 20 km$^{2}$ (projected area in the PL+BB fit for the Vela pulsar

738: spectrum).

739: Given the errors of these fits and the uncertainty of ISM

740: absorption,

741:  the obtained BB parameters

742: are just crude estimates.

743: We can, however, treat

744: $L_{\rm bol}\sim 2\times 10^{33}

745: d_4^2$ ergs s$^{-1}$

746:   as

747: an upper limit for the bolometric luminosity.

748: 

749: 

750: \begin{figure}

751: %\epsscale{0.75}

752:  \centering

753: \includegraphics[width=2.9in,angle=0]{f4.eps}

754:  \caption{ Spectrum of the compact PWN ({\em top})

755: and the

756: extended PWN ({\em bottom})

757:  fitted with the PL model at fixed $n_{\rm H,22}=1.0$.

758:  }

759: \end{figure}

760: 

761: \begin{figure}

762: %\epsscale{0.75}

763:  \centering

764: \includegraphics[width=2.5in,angle=90]{f5.eps}

765: \caption{Confidence contours (68\% and 90\%) in the $n_{\rm

766: H}$--$\Gamma$ plane for the PL fit to the

767:  compact and extended PWN spectra.

768:  The contours

769:  are obtained with the PL normalization fitted at each point

770: of the grid.

771: }

772: \end{figure}

773: 

774: 

775: \begin{figure}

776: %\epsscale{0.75}

777:  \centering

778: \includegraphics[width=2.5in,angle=90]{f6.eps}

779: \caption{  Confidence contours (68\% and 90\%) for the PL fit to the

780:  compact PWN and extended PWN

781: for fixed $n_{\rm H,22}=1.0$.

782:  The PL normalization

783: is in units of $10^{-5}$ photons cm$^{-2}$ s$^{-1}$ keV$^{-1}$ at 1

784: keV. The dashed curves are the loci

785:  of constant unabsorbed flux in the 0.5--8 keV band;

786: the flux values near the curves are in units of 10$^{-13}$ ergs

787: cm$^{-2}$ s$^{-1}$.

788: }

789: \end{figure}

790: 

791: \begin{figure}

792: %\epsscale{0.75}

793:   \centering

794:  \vbox{

795: \includegraphics[width=2.5in,angle=90]{f7.eps}

796: } \caption{ Confidence contours (68\% and 90\%) for the PL fit to the pulsars's spectrum.

797:  The PL normalization

798: is in units of $10^{-5}$ photons cm$^{-2}$ s$^{-1}$ keV$^{-1}$ at 1

799: keV. The dashed curves are the lines of constant unabsorbed flux

800: in the 0.5--8 keV band (the flux values are in units of $10^{-14}$

801: ergs cm$^{-2}$ s$^{-1}$).

802: }

803: \end{figure}

804: 

805: \subsection{Other X-ray sources}

806: 

807: Among X-ray

808: sources serendipitously detected in this ACIS observation, most interestng are

809: CXOU\,182614.4$-$133448 (Source X, see Figs.\ 1 and 2) $\approx 18''$ east of

810: the B1823 pulsar,

811: %which might be related to the PWN,

812: and a few sources

813:  on the S2 and I2 chips projected within the brightest part of

814: HESS\,J1825 (see Fig.\ 8),

815: which might supply additional ultrarelativistic

816: electrons to power HESS\,J1825. The positions

817: and spectral characteristics of these sources

818: are presented in Table 3.

819: 

820: The strongly absorbed ($n_{\rm H,22}\sim 3$) point-like Source X

821: shows a hard spectrum ($\Gamma\sim 0.2$) and the

822: 1--8 keV flux

823: of $\sim 3\times 10^{-14}$ ergs cm$^{-2}$ s$^{-1}$

824: (see Table 3 for other properties). There is no

825: an obvious optical-NIR counterpart for this source in stellar catalogs.

826: 2MASS images show a point source with $J=13.4$, $H=12.3$ and

827: $K_s=11.85$ close to Source X

828: ($\alpha=18^{\rm h}26^{\rm m}14.32^{\rm s}$,

829: $\delta=-13^\circ 34'48.2''$),

830: but, since the offset of $1.6''$ from the pulsar position exceeds the

831: combined centroiding and absolute astrometry error

832: ($0.7''$ at the 90\% confidence level), the 2MASS source is unlikely

833: to be a counterpart of Source X.

834: Based on the X-ray spectrum, we conclude that Source X

835: is, most likely,

836: a background AGN, not related to either B1823 or HESS\,J1825.

837: 

838: The other X-ray sources listed in Table 3 also look point-like

839: (the apparent extension of the sources s4 and i1 can be

840: explained by the asymmetric PSF broadening at the large off-axis distances).

841: 2MASS and Digital Sky Survey (DSS\footnote{see http://archive.eso.org/dss/dss})

842:  images do not show NIR-optical objects within most

843: conservative error circles around

844: the sources s1, s5, and i1 (the nearest objects are offset by

845: $4.1''$, $4.0''$, and $3.4''$, respectively).

846: The strongly absorbed s1 and s5 are likely extragalactic sources

847: (possibly AGNs),

848: while the nature of i1 is unclear. The source s4 shows a possible

849:  NIR-optical

850: counterpart DENIS\,J182620.9$-$134425

851:  (offsets are in the range of $0.36''$--$0.86''$ in different catalogs;

852:  magnitudes $K_s=11.75$, $H=12.2$, $J=13.3$

853: in the 2MASS catalog; $I=15.0$, $R=15.8$--16.4, $V=16.5$--17.1, $B=17.5$--19.1

854: in other catalogs). The source looks slightly extended in the NIR-optical

855: images, likely a blend of two field stars.

856: Close to the sources s2 and s3, there are

857: point-like

858: NIR-optical objects (DENIS J182617.1$-$134111

859: and J182618.0$-$134225, respectively),

860: but their relatively large

861: offsets from the X-ray positions

862:  ($1.4''$--$1.9''$ and $1.8''$--$2.3''$,

863: respectively, somewhat different in different catalogs)

864: make their association with the X-ray sources questionable.

865: Overall, the X-ray sources in the brightest part of HESS\,J1825

866: are rather faint and unremarkable,

867: and there is no evidence that any of the sources is related

868: to the VHE $\gamma$-ray source HESS\,J1825.

869: 

870: \begin{figure}[]

871: %\epsscale{0.75}

872:  \centering

873: \includegraphics[width=3.5in,angle=0]{f8.eps}

874: \caption{ACIS field sources in the vicinity of the HESS\,J1825 center (shown by the circle).

875: }

876: \end{figure}

877: 

878: \subsection{Vicinity of PSR B1823-13 and HESS J1825--137 at other wavelengths}

879: To check whether or not there are luminous sources within the central part

880: of HESS\,J1825, which could enhance the VHE $\gamma$-radiation by providing

881: additional photons for the inverse Compton (IC) scattering

882: off ultrarelativistic electrons,

883: we have examined optical-NIR, IR, and radio images of this field.

884: Three examples, {\sl Spitzer} IRAC 8\,$\mu$m and MIPS 24\,$\mu$m

885: images from the Galactic Legacy Infrared Mid-Plane Survey

886: (GLIMPSE\footnote{http://www.astro.wisc.edu/sirtf/}) and a NRAO

887: VLA Sky Survey (NVSS; Condon et al.\ 1998) 1.4 GHz image are shown

888: in Figure 3. The images do not show anything remarkable in the

889: immediate vicinity, within $\sim 2'$, of B1823. The most interesting

890: object within the brightest part of HESS\,J1825 is the extended

891: ($\approx 4'$ in size) source, marked A in Figure 3,

892: prominent in both the IR and radio images.

893: The source with a two-sided shell morphology is centered at

894: $\alpha \approx 18^{\rm h}26^{\rm m}02^{\rm s}$,

895: $\delta \approx -13^{\circ}38'08''$

896: (measured in the IRAC 8\,$\mu$m image), i.e.\

897:   between the pulsar and the brightest part of HESS\,J1825.

898: It is listed in the IRAS catalog of point sources\footnote{Infrared

899: Astronomical Satellite Catalog, 1988, The Point Source Catalog, ver.\ 2.0,

900: NASA RP-1190.} as IRAS\,18231--1340, with fluxes of

901: 4.25, 13.4, 161, and 393 Jy at 12, 25, 60, and 100 $\mu$m, respectively.

902: Its radio fluxes are 0.13 and 0.03 Jy at 4.85 and 1.4 GHz, respectively

903: (Griffith et al.\ 1994; Condon et al.\ 1998).

904: This object is barely visible in the IRAC 4.5\,$\mu$m image, and it is

905: not seen in the

906: 2MASS images of the field. The IR-radio spectrum and

907:    the lack of an extended optical or NIR counterpart

908: suggest that

909: the IR emission is thermal

910: emission of

911: dust with $T\sim 30$--50 K, possibly mixed with the

912: molecular/atomic cloud(s) that are heated and partly ionized by

913: radiation from embedded young stars.

914: Its luminosity can be crudely estimated as

915: $L_A\sim 3\times 10^{37}d_4^2$

916: ergs s$^{-1}$.

917: 

918: The nature of the other two bright radio sources within the brightest part

919: of HESS\,J1825 (B and C in Fig.\ 3{\em f}), centered at

920: $\alpha\approx 18^{\rm h}25^{\rm m}57^{\rm s}$,

921: $\delta\approx -13^\circ 47'56''$ and

922: $\alpha\approx 18^{\rm h}25^{\rm m}38^{\rm s}$,

923: $\delta\approx -13^\circ 36'37''$, respectively, is unknown.

924: However, as they are not seen in IR, their bolometric luminosity

925: is much lower than that of Source A.

926: 

927: 

928: \section{Proper motion of PSR B1823--13}

929: 

930: We measured the proper motion

931: using two unrelated archival

932: VLA observations.  Both observations were made in the 20~cm band in

933: the A-configuration providing about $1.5''\times 2.1''$ resolution.

934: The reduction technique

935: is identical to that used by

936: Brisken et al.\ (2006)

937: in the determination of the proper motion of PSR B1800--21 using

938: 1.4 GHz data.

939: All synthesis calibration and imaging were

940: performed using the AIPS software package.

941: Project AF118 was observed on MJD 46573

942: in right circular polarization in spectral line mode providing

943: 15 channels spanning 50~MHz bandwidth at 1385~MHz.

944: The upper 4 channels were discarded due to severe interference.

945: Project AC761 was observed on MJD 53265 in dual circular polarization

946: in spectral line mode providing 15 channels across 25~MHz at 1665~MHz.

947: After standard calibration and editing, AIPS task UVFIX was used to

948: correct the baseline vectors of the pulsar field for the half-integration time

949: time-stamp error

950:  common to all VLA observations as well as including

951: relativistic aberration correction.  UVFIX was also used to convert the

952: B1950 coordinates used in AF118 to the J2000.0 coordinates.

953: 

954: The temporal and spectral resolution of both data sets allowed imaging of

955: the entire 25\arcmin\ primary beam.  In addition to the pulsar, eleven other

956: sources were detected in both epochs, five of which were bright and

957: compact enough to be used as reference sources.  Accurate positions for

958: the pulsar and these five objects were measured for each epoch.

959: The motion of the pulsar was then measured in a frame defined by

960: these sources using a fitting prodedure described in

961: McGary et al.\ (2001), yielding a proper motion

962: $\mu_{\alpha} = 22.7 \pm 2.2$~mas~yr$^{-1}$,

963: $\mu_{\delta} = -5.3 \pm 2.7$~mas~yr$^{-1}$

964: (68\% confidence intervals). The global linear fit used to determine

965: the proper motion and a possible translation and rotation of the coorinates

966: between the two epochs yielded a reduced $\chi^2$ of 1.3,

967: indicating that the formal uncertainties may be underestimates at the

968: 15\% level.  Assuming a distance of $4 \pm 1$~kpc,

969:  this proper motion implies a local standard of rest (LSR) proper

970: motion of $\mu_{\alpha} = 23.0 \pm 2.5$~mas~yr$^{-1}$,

971: $\mu_{\delta} = -3.9 \pm 3.3$~mas~yr$^{-1}$, which corresponds

972: to the velocity $v_\perp = (443\pm 48) d_4$ km s$^{-1}$ at a position angle

973: of $99\fdeg6\pm 8\fdeg1$ (east of north).

974: At this proper motion, the pulsar has traveled about $8.2'$ in the plane

975: of the sky during its characteristic age of 21 kyr. The projected

976: birthplace of the pulsar (hence the SN location) is marked in Figure 3{\em b-f}.

977: 

978: \section{Discussion}

979: 

980: \subsection{Pular Wind Nebula}

981: From their analysis of the {\sl XMM-Newton} data,

982: G03 have concluded that the B1823 PWN

983: (which they designated as G18.0--0.7)

984: is comprised

985: of two components (``core'' and ``diffuse'') of different brightness and

986:  size and perhaps different

987: spectra. Our {\sl Chandra} ACIS observation has confirmed the presence

988: of the two components, which we designate as the ``compact PWN''

989: and ``extended PWN'',

990: and allowed us to investigate the properties of the compact PWN in much more

991: detail.

992: The better resolved images and the measurement of

993: the pulsar's proper motion not only allow a more definitive

994: interpretation of the compact PWN properties,

995: but they also help better constrain the models

996: the puzzling extended PWN and its connection to HESS\, J1825.

997: 

998: \subsubsection{Compact PWN}

999: 

1000: The east-west elongation of the compact PWN, approximately coinciding

1001: with the direction of the pulsar's proper motion (see Fig.\ 2), strongly

1002: suggests that the compact PWN

1003: is shaped by the oncoming flow of the ambient matter in the pulsar's

1004: reference frame.

1005: The transverse velocity

1006: of the pulsar, $v_\perp \approx 440 d_4$ km s$^{-1}$, is high enough to

1007: assume that the pulsar's motion is supersonic,

1008: i.e., ${\mathcal M}\equiv v/c_s >1$, or, equivalently,

1009: $T< 1.4\times 10^7 \mu (d_4/\sin i)^{2}$ K, where

1010: ${\mathcal M}$ is the Mach number, $v=v_\perp/\sin i$ is the pulsar's

1011: speed, $i$ is the angle between the velocity vector and

1012: the line of sight, $c_s=(5kT/3\mu m_{\rm H})^{1/2}=

1013: 117 \mu^{-1/2} T_6^{1/2}$ km s$^{-1}$ is the speed

1014: of sound, $\mu$ is the molecular weight,

1015: and $T=10^6 T_6\,{\rm K}$ is the temperature of the ambient medium.

1016: %The wind from a supersonically moving pulsar

1017: When a pulsar moves supersonically, the pulsar wind (PW) termination

1018: shock (TS) acquires a bow-like shape ahead of the pulsar,

1019: with the apex at a distance $R_h=(\dot{E} f_\Omega/4\pi c p_{\rm ram})^{1/2}$,

1020: determined by the balance between the PW pressure and the

1021: ram pressure, $p_{\rm ram}=\rho v^2$, where $\rho$ is the density of

1022: the ambient medium, and $f_\Omega$ takes into account anisotropy of

1023: the pulsar wind.

1024: For B1823, these equations yield

1025: $p_{\rm ram}=3.3\times 10^{-9} n (d_4/\sin i)^{2}$ ergs cm$^{-3}$ and

1026: $R_h=4.8\times 10^{16} n^{-1/2} f_\Omega^{1/2} (d_4/\sin i)^{-1}$ cm,

1027: which translates into the projected angular distance

1028: ${\mathcal R}_h \approx 0.8'' f_\Omega^{1/2} n^{-1/2}(d_4/\sin i)^{-2}$,

1029: where $n=\rho/m_{\rm H}$.

1030: 

1031: In an idealized picture,

1032: the shocked PW,

1033: responsible for the PWN X-ray emission,

1034: is confined between the TS and

1035: the contact discontinuity (CD) surface, which separates the shocked

1036: PW from the shocked ambient medium.

1037: For ${\mathcal M}\gg 1$

1038: and a nearly isotropic

1039: preshock PW with a low magnetization parameter $\sigma$,

1040: the distance between the pulsar and the head of the CD surface is

1041: $\approx 1.3 R_h$ (Bucciantini et al.\ 2005, hereafter B05).

1042: In a more realistic picture,

1043: one should not expect a smooth CD surface because it is unstable.

1044: As a result, the shocked PW can penetrate into the shocked ambient medium,

1045: and vice versa.

1046: 

1047: The images of the compact PWN (Figs.\ 1 and 2)

1048: do not show a sharp PWN boundary ahead of the pulsar that could be

1049: unambiguosly identified as a CD surface head.

1050: Instead, we see a two-component structure, with the bright inner

1051: component up to $3''$ east of the pulsar and some fuzzy emission

1052: discernible up to $7''$.

1053: Although

1054: the quality of the image does not

1055: allow one to make a definitive conclusion, we can assume that the

1056: fuzzy emission represents the shocked PW that has penetrated into the

1057: shocked ambient medium because of CD instabilties, while

1058: the inner component is emitted by the shocked PW immediately

1059: beyond the TS, so that ${\mathcal R}_h\sim 2''$--$3''$ seems to be

1060: a reasonable estimate for the angular distance to the TS head.

1061: Scaling ${\mathcal R}_h$ to $3''$, we can estimate the ambient

1062: number density,

1063: $n\sim 0.07 f_\Omega (d_4/\sin i)^{-4}({\mathcal R}_h/3'')^{-2}$ cm$^{-3}$,

1064: and the ram pressure,

1065: $p_{\rm ram}\sim 2.3\times 10^{-10} f_\Omega (d_4/\sin i)^{-2}

1066: ({\mathcal R}_h/3'')^{-2}$ ergs cm$^{-3}$.

1067: Such a ram pressure is typical for Vela-like pulsars (Kargaltsev

1068: et al.\ 2007), while

1069: such a number density can be found in a warm phase of the ISM or

1070: within an evolved SNR. As no SNR emission is seen in the vicinity

1071: of B1823, we cannot measure the temperature and pressure of the

1072: ambient medium, but it seems reasonable to assume $T\sim 10^5$ K,

1073: which could explain the lack of X-ray SNR emission. Scaling the

1074: temperature to this value, we obtain $p_{\rm amb} = nkT/\mu \sim

1075: 1\times 10^{-12} f_\Omega (d_4/\sin i)^{-4}({\mathcal R}_h/3'')^{-2}\mu^{-1}

1076:  T_5 \ll p_{\rm ram}$,

1077: $c_s \sim 40 \mu^{-1/2} T_5^{1/2}$ km s$^{-1}$,

1078: and ${\mathcal M} \sim 10 (d/\sin i) \mu^{1/2} T_5^{-1/2}$.

1079: 

1080: The existing MHD models of bow-shock PWNe associated with supersonically

1081: moving pulsars assume a nearly isotropic preshock PW.

1082:  However, the high-resolution {\sl Chandra} images of the whole

1083: compact PWN can hardly be reconciled with this assumption. According

1084: to B05, the TS in an isotropic PW of a supersonically moving pulsar

1085:  acquires a bullet-like shape,

1086: with the bullet's cylindrical radius $r_{\rm TS} \sim R_h$ and

1087: the distance of its back surface from the pulsar $R_b\sim 6 R_h$.

1088: The shocked PW outside the TS is confined inside the CD surface,

1089: which has a cylindrical shape behind the TS bullet, with a

1090: cylindrical radius $r_{\rm CD}

1091: \sim 4 R_h$. The collimated PW flows with subrelativistic velocities:

1092: $0.1c$--$0.2c$ in the inner channel, $r\lesssim r_{\rm TS}$, and

1093: up to $0.8c$--$0.9c$ in the outer channel, $r_{\rm TS}\lesssim r \lesssim

1094: r_{\rm CD}$ (see Figs.\ 1--3 in B05). The images of the compact PWN

1095: are not quite consistent with this simple picture. First of all, the

1096: bright inner component, which could be naturally associated with

1097: the PW immediately outside the TS, is inclined to the direction

1098: of the proper motion by an angle $\sim 50^\circ$, while it should be

1099: coaligned with the proper motion if the wind is isotropic. Second,

1100: the outer PWN, which might represent the shocked PW confined by

1101: the CD surface, not only has an irregular boundary (which could be

1102: explained by instabilities), but it is also rather nonuniform.

1103: (Particularly puzzling is the blob $\approx 8''$ west of the pulsar;

1104: see Figs.\ 1 and 2, and \S2.2). Finally, we do not see an extended

1105: tail behind the TS aligned with the proper motion direction.

1106: These differences between the model and the observed image suggest

1107: that the preshock PW is, in fact, substantially anisotropic.

1108: To understand the PW geometry and interpret the PWN morphology more

1109: or less unambiguously, deeper high-resolution X-ray observations

1110: are needed, as well as modeling of intrinsically anisotropic,

1111: magnetized

1112: PWs from fast-moving pulsars. Here we can only speculate that

1113: the inner PWN likely represents a shocked equatorial outflow (similar to

1114: the famous Crab torus; Weisskopf et al.\ 2000), seen almost

1115: edge-on. If this is the case,

1116: then the spin axis of the B1823 pulsar is strongly misaligned

1117: (by $\sim 40^\circ$--$50^\circ$ in the plane of the sky)

1118: with the direction of its motion, contrary to many other pulsars.

1119: The absense of jets along the spin axis in the observed images

1120: can be explained by their faintness (e.g., the jets of the Vela

1121: PWN would be unobservable at $d\sim 4$ kpc with such an exposure --

1122: see Fig.\ 8 in Kargaltsev et al.\ 2007), or possibly the jets

1123: are deflected (or even crushed) by the oncoming flow of the ambient

1124: medium. The blob and the apparent wiggles in the west part

1125: of the outer component could be ascribed to instabilities

1126: in the tail formed by the supersonic motion of the pulsar.

1127: 

1128: The luminosity, $L_{\rm comp}\sim 3\times 10^{32} d_4^2$ ergs s$^{-1}$

1129: in the 0.5--8 keV band\footnote{We should note that our estimate for the

1130: compact PWN luminosity is a factor of 3 lower than that obtained

1131: by G03 for the ``core'' component. The discrepancy is possibly

1132: caused by the much larger area from wich G03 extracted the spectrum

1133: (1018 vs.\ 225 arcsec$^2$), which included some contribution from

1134: the extended (``diffuse'') PWN, the pulsar, and field sources, such as Source X

1135: in our Figures 1 and 2. Also, G03 extracted the background spectrum

1136: from a region outside the extended PWN, while we used an annulus

1137: filled with the extended PWN emission (see \S2.3).

1138:  },

1139:  the X-ray efficiency,

1140: $\eta_{\rm comp}\equiv L_{\rm comp}/\dot{E}

1141: \approx 1.2\times 10^{-4} d_4^2$,

1142:   and the spectral slope, $\Gamma_{\rm comp}\approx

1143: 1.3$ (at the assumed $n_{\rm H,22}=1.0$) of the compact PWN

1144: are not unusual for Vela-like PWNe

1145: (e.g., Kargaltsev et al.\ 2007; Kargaltsev \& Pavlov 2007).

1146: Also, similar to other Vela-like PWNe, the PWN luminosity

1147: exceeds by a factor of a few the pulsar luminosity in the

1148: same X-ray energy range,  $L_{\rm comp}\sim 4 L_{\rm PSR}$.

1149: 

1150: \subsubsection{Extended PWN}

1151: The morphology of the extended PWN is even more puzzling than that

1152: of the compact one. In the conventional picture of a bow-shock PWN,

1153: we would expect to see an extended component

1154: {\em behind} the compact PWN of

1155: the moving pulsar,

1156: which would look like a ``tail''. In B1823, however,

1157: the PWN extends southward from the pulsar, almost

1158: perpendicular to the proper motion direction,

1159: similar to the extended diffuse PWN of the Vela pulsar

1160: (Pavlov et al.\ 2003). This means that something

1161: is deflecting (or has deflected) the shocked PW in this direction.

1162: One could imagine a local north wind in the turbulent ambient medium,

1163: but such a wind would have to blow with a

1164: velocity of several hundred

1165: km s$^{-1}$ (comparable to the pulsar velocity) to displace the PWN bubble.

1166: If such a wind were blowing now, we would see the compact PWN elongated

1167: southward from the pulsar, and, more importantly,

1168: it would be not a usual wind but a shock, since the required velocity

1169: almost certainly exceeds the sound speed. Thus, it seems we have

1170: to assume that the PWN bubble was crushed/displaced by a passing

1171: shock some time ago. A natural candidate for such a shock is the

1172: reverse shock from the surrounding SNR that has propagated back

1173: inward and collided with the PWN, displacing the PWN bubble from

1174: the pulsar, as has been suggested by Blondin et al.\ (2001) for the

1175: strongly dispaced Vela X component of the Vela PWN.

1176: Such an explanation for the

1177: one-sided extended PWN of B1823 has been proposed by G03, and we

1178: currently do not see an alternative explanation. If this hypothesis

1179: is correct, we may expect the extended PWN to consist of two components:

1180: a strongly displaced one (similar to Vela X),

1181: which represents a relic population of

1182: relativistic electrons that have lost a substantial fraction of

1183: their energy to synchrotron, inverse Comption, and adiabatic cooling,

1184: and a younger one, which represents electrons that

1185: entered the PWN after the inverse

1186: shock passage.

1187: 

1188: Although the crushed/dragged PWN hypothesis looks attractive

1189: and provides a plausible explanation for the extended X-ray PWN,

1190: we should mention that it does not so readily expalin the huge,

1191: about 70 pc at $d=4$ kpc, size of the

1192: TeV PWN. Traveling 70 pc in time $t$ requires the average

1193: velocity $v\approx 6,900\, (t/10\,{\rm kyr})^{-1}$ km s$^{-1}$, too high

1194: for a reverse shock. Even if we assume $t=21$ kyr

1195: (i.e., equal to the pulsar's spindown age), we still need

1196: $v\approx 3,300$ km s$^{-1}$. Therefore, it is hard to imagine

1197: how the reverse shock could have dragged the relic PWN so far,

1198: even if it passed the pulsar very long time ago.

1199: Perhaps, one should explore

1200: another possibility, that during a substantial fraction of the pulsar's

1201: lifetime the PWN electrons have been

1202: transported southward from the pulsar

1203: along an anisotropic large-scale

1204: magnetic field.

1205: Alternatively, one might assume a strongly anisotropic supply

1206: of electrons into the PWN, e.g., through a one-sided jet

1207: (which, however, is not seen in the images available).

1208: At this point, however, these possibilities remain highly

1209: speculative.

1210: 

1211: As has been noticed by G03, the extended PWN spectrum is softer

1212: than the compact PWN spectrum, which is also confirmed by our

1213: results (see Fig.\ 5). The difference of the photon indices,

1214: $\Gamma_{\rm ext} - \Gamma_{\rm comp} \approx 0.5$, supports the hypothesis

1215: that the extended X-ray PWN electrons have lost their energy to synchrotron

1216: cooling after traveling a distance of a few parsecs.

1217: The slope $\Gamma_{\rm ext}\approx

1218: 1.9\pm 0.1$ we found for the X-ray spectrum is very close to

1219: the slope of the VHE $\gamma$-ray spectrum within $\approx 6'$ from the pulsar,

1220: $\Gamma_{\gamma}=1.8$--2.0 (A06),

1221: which confirms that both

1222: the X-ray and $\gamma$-ray emission are due to the same electron population

1223: with a power-law energy spectrum $dN_e/dE_e\propto E_e^{-p}$,

1224: with $p=2\Gamma_{\rm ext}-1\approx 2.5$--3.

1225: Interestingly, the TeV spectrum

1226: at larger distances from the pulsar, where we do not see the

1227: X-ray PWN, becomes even softer ($\Gamma_\gamma$ grows up to

1228: $\approx 2.6$ at $60'$ from the pulsar).

1229: 

1230: Because we can see only part of the extended PWN in our ACIS

1231: image (see \S2.2),

1232: we can measure only a fraction of its X-ray luminosity

1233: ($L_{\rm ext}\sim 1\times 10^{33} d_4^2$ ergs s$^{-1}$ in the

1234: 3.4 arcmin$^2$ region shown in Fig.\ 1, {\em bottom right}).

1235: The total extended PWN luminosity should be a factor of a few

1236: higher. For instance, G03 found $L_{\rm ext} \sim 3\times 10^{33} d_4^2$

1237: ergs s$^{-1}$ in a 33.4 arcmin$^2$ area, which corresponds to

1238: the extended PWN efficiency $\eta_{\rm ext} \sim 10^{-3} d_4^2$,

1239: a factor of $\sim 10$ higher than that of the compact PWN.

1240: 

1241: \subsection{Pulsar}

1242: Although we have resolved the pulsar from the PWN, the small

1243: number of pulsar counts detected and the strong

1244: and not well determined  background

1245: from the inner component of the compact PWN preclude a precise

1246: spectral analysis. We can only estimate the pulsar

1247: luminosity, $L_{\rm PSR} \sim 1\times 10^{32} d_4^2$ ergs s$^{-1}$

1248: in the 0.5--8 keV band, which is a small fraction,

1249: $\sim 3\times 10^{-5}$, of the pulsar's spindown power $\dot{E}$.

1250: The slope of the pulsar spectrum, $\Gamma_{\rm PSR}\sim2$--3,

1251: is strongly correlated with the ISM column density (see Fig.\ 7). It is

1252: likely that the spectrum consists of two components, magnetospheric

1253: and thermal, but the photon statistics does not allow one to separate

1254: these components unambiguously. For $n_{\rm H,22}=1.3$ (at the

1255: higher end of the plausible $n_{\rm H}$ range) and projected emitting areas

1256: in the range of $(20$--$500)d_4^2$ km$^2$, we found $kT\sim

1257: 90$--150 eV and $L_{\rm bol} \sim (3$--$20)\times 10^{32} d_4^2$ ergs s$^{-1}$

1258: for a possible BB component (higher luminosities correspond to lower

1259: temperates and larger emitting areas).

1260: These temperatures and luminosities are generally similar

1261: to those of other Vela-like pulsars

1262: (e.g., $kT\approx 130$ eV, $\mathcal{A}\approx 20 (d/300\,{\rm pc})^2$ km$^2$,

1263:  $L_{\rm bol}\sim 2\times 10^{32}

1264: (d/300\,{\rm pc})^2$ ergs s$^{-1}$ for the Vela pulsar; Pavlov et al.\ 2001).

1265: However, because of the large uncertainties and  poorly constrained

1266: $n_{\rm H}$, we cannot be sure even in the

1267: detection of the thermal component,

1268: not to mention measuring its parameters. Therefore, we should treat

1269: the estimated bolometric luminosities of B1823 as upper limits

1270: rather than actually measured values.

1271: The most conservative of the above limits,

1272: $L_{\rm bol}\sim 2\times 10^{33} d_4^2$ ergs s$^{-1}$,

1273: corresponding to

1274: $\mathcal{A}=500 d_4^2$ km$^2$ and $T\approx 100$ eV,

1275: is close to the bolometric luminosities predicted by the so-called

1276: ``standard'' NS cooling model for the NS age of 20 kyr

1277: (see, e.g., Fig.\ 2 in Tsuruta 1998).

1278: Overall, we can conclude that B1823 looks like a typical Vela-like

1279: pulsar, but a deeper high-resolution observation is needed to infer its

1280: properties in more detail.

1281: 

1282: \subsection{The lack of a host SNR?}

1283: 

1284: We see no emission in our ACIS images that could be ascribed

1285: to an SNR associated with B1823, and no such emission has been

1286: reported from the previous X-ray (F96; G03) or radio

1287: (Braun et al.\ 1989; Brogan et al.\ 2006) observations.

1288: Using our proper motion measurement, we extrapolated

1289: the pulsar trajectory back in time to check whether or not

1290: it crosses a known SNR. The nearest SNR to the pulsar trajectory

1291: is the $8'\times 8'$ shell SNR G18.16--0.16

1292:  (Brogan et al.\ 2006) at a distance of $35'$ from the current

1293: pulsar position.

1294: Since

1295: the extrapolated trajectory

1296: misses the SNR center by $19'\pm 4'$,

1297: we conclude that G18.16--0.16 is not related to B1823.

1298: 

1299: The failure to detect the host SNR in X-rays can be attributed

1300: to the strong ISM absorption of the soft thermal emission from the

1301: SNR interiors (see Kargaltsev et al.\ 2007, where a similar situation

1302: is discussed regarding the Vela-like PSR B1800--21, which also lacks

1303: an associated SNR).

1304: Therefore, it is possible that B1823 is still within the interiors

1305: of an evolved SNR, which is undetectable in X-ray or radio, but whose

1306: pressure can contribute to the PWN confinement.

1307: 

1308: \subsection{Connection between B1823 and HESS\,J1825}

1309: 

1310: The important discovery of the VHE $\gamma$-ray spectral softening

1311: with increasing distance from B1823

1312: (A06) has provided further evidence that

1313: this pulsar/PWN is the main source of the

1314: ultrarelativistic electrons powering the TeV plerion.

1315: Qualitative models for this VHE $\gamma$-ray

1316:  source have been discussed in a number

1317: of works

1318: (A06; de Jager et al.\ 2005; de Jager 2007).

1319: Generally, they interpret the $\gamma$-ray

1320:  emission as produced by the Compton

1321:  upscattering

1322: of the CMB photons by electrons with energies $E_e \sim 25

1323: (E_\gamma/1\,{\rm TeV})^{1/2}\,\, {\rm TeV} \sim 10$--100 TeV

1324: [electron Lorentz factor $\gamma_e\sim (2$--$20)\times 10^7$],

1325: supplied by the B1823 pulsar. The fact that the pulsar is located at the

1326: northern boundary of the TeV plerion is explained by the effect of the

1327: SNR reverse shock that has dragged the relic PWN southward.

1328: (An alternative explanation,

1329: that the TeV plerion represents electrons left behind the pulsar

1330: moving northward, is invalidated by our measurement of the proper

1331: motion.) The softening of the TeV spectrum with the distance from

1332: the pulsar can be ascribed to the radiative losses (mostly to the

1333: IC scattering losses) or some energy-dependent ``convection'', or  variation

1334: of the particle injection spectrum with the pulsar's age.

1335: As A06 note, even at very low ambient magnetic fields, $B\lesssim 3\, \mu$G,

1336: when the IC energy losses dominate the synchrotron losses, the

1337:  $\gamma$-ray

1338: efficiency, $\eta_\gamma =L_\gamma/\dot{E}\approx 0.1$,

1339: is unrealistically high if we use

1340: the current value of $\dot{E}$.

1341: This supports the hypothesis that the

1342: VHE $\gamma$-ray emission is produced by relic electrons

1343: emitted by the pulsar when it was very young and much more

1344: powerful.

1345: 

1346: Modelling of the TeV plerion is beyond the scope of

1347: this work. We only note some open problems with the current

1348: interpretation of this remarkable object. First, as we have mentioned

1349: in \S4.1, it is difficult to reconcile the very large size of the

1350: TeV plerion with the hypothesis that the relic PWN was dragged by

1351: the SNR reverse shock to its current position. In principle,

1352: other sources of ultrarelativistic electrons

1353: within the  $1^\circ$ size field could contribute to the observed

1354: TeV emission and increase the apparent size of the TeV source.

1355: Our analysis of several X-ray sources in the brightest

1356: part of the TeV plerion has not shown any obvious candidates

1357: (see \S2.5), but one cannot rule out the possibility that there

1358: are such sources outside the small ACIS field-of-view.

1359: 

1360: Another possible problem with the current interpretation of

1361: the TeV source is the spatial distribution of its brightness and spectrum.

1362: The brightest part of the TeV plerion is offset by $\sim 10'$

1363: ($\sim 12$ pc in the plane of the sky)

1364: south-southwest of the pulsar (see Fig.\ 1 of A06),

1365: while

1366: the plerion extends $\sim 50'$ further in the same direcion

1367: (almost perpendicular to the direction of the pulsar's proper motion).

1368: One could speculate

1369: that the relic PWN was displaced by $\sim 10'$

1370: some time ago [which would require a reverse shock speed of $\sim

1371: 1000 (t/10\,{\rm kyr})^{-1}$ km s$^{-1}$], and its relativistic electrons

1372: have been diffusing from that position. It is however, unclear as to

1373: why they diffused not radially but preferentially in the

1374: south-southwest direction.

1375: If the passing shock not just displaced

1376: the relic PWN to a certain distance but smeared it all over the

1377: whole extent of the currently seen TeV plerion, then a very high

1378: shock speed would be required (see \S4.1).

1379: %

1380: Furthermore, the spectral map of HESS\,J1825 (see Figs.\ 5 and 6 of

1381: A06) shows that the VHE $\gamma$-ray spectrum softens radially not from the

1382: maximum of the brightness distribution but from the pulsar position.

1383: This hints that at least some contribution to the TeV photons

1384: comes from the same young electrons that are responsible for the

1385: extended X-ray PWN of a much smaller size. This conjecture is

1386: further supported by the fact that the spectral slopes of the

1387: X-ray (synchrotron) and $\gamma$-ray (IC) spectra are

1388: about the same at a few arcminute distance from the pulsar, which

1389: suggests that they are generated by the same population of

1390: electrons. On the other hand, if the TeV $\gamma$-ray emission

1391: is due to the IC scattering of the CMB photons, and the relativistic

1392: electrons are mostly supplied by the pulsar, one should expect the

1393: TeV brightness to peak at the pulsar position, not $10'$ from the pulsar.

1394: This inconsistency might be alleviated assuming that

1395: in the immediate vicinity of the pulsar,

1396: where the electrons are more energetic, they upscatter the CMB

1397: photons in the Klein-Nishina regime, which results in a lower

1398: brightness of the IC scattered radiation, but it remains to be

1399: seen from detailed models whether or not this assumption is realistic.

1400: Overall, it seems plausible at this point that, at least in the

1401: extended X-ray PWN region, the TeV emission can be produced by two

1402: electron populations of different ages. To infer the properties

1403: of these populations, deep observations with {\sl XMM-Newton}

1404: and

1405: {\sl GLAST} of a region from which {\em both} X-ray

1406: and VHE $\gamma$-ray emission is seen would be very helpful.

1407: 

1408: Another issue related to the nature of the TeV plerion and its

1409: connection to the B1823 pulsar/PWN is whether the TeV photons

1410: are produced by upscattering of the CMB or some other radiation.

1411: Close to the Galactic plane, where HESS\,J1825 is situated,

1412: a factor of a few higher radiation energy density can be provided

1413: by IR emission from interstellar dust and Galactic starlight.

1414: For instance,

1415: at the galactocentric distance

1416: of 4 kpc, the models of interstellar radiation by Strong

1417: et al.\ (2000) give the energy densities of 0.6 and 2.7 eV cm$^{-3}$ for

1418: these two components, respectively,

1419: versus 0.26 eV cm$^{-3}$ for

1420: the CMB. Moreover, if there are very luminous sources of radiation

1421: (such as

1422: molecular/dust complexes

1423: in star-forming regions) in the vicinity of the source of

1424: ultrarelativistic electrons, the high radiation energy density

1425: around these sources can strongly enhance the TeV luminosity

1426: (see, e.g., Kargaltsev \& Pavlov 2007).

1427: In this regard, the bright radio-IR Source A (see Fig.\ 3),

1428: whose spectrum peaks at $\lambda \gtrsim 100\,\mu$m, might be

1429: of some interest.

1430: With the

1431: bolometric flux $F_A \sim 2\times 10^{-8}$ ergs cm$^{-2}$ s$^{-1}$,

1432: its radiation energy density exceeds the ambient radiation energy

1433: density $U_{\rm amb}$ at angular distances $\theta < (F_A/cU_{\rm

1434: abm})^{1/2} \sim 2.2' (U_{\rm amb}/1\,{\rm eV\, cm}^{-3})^{-1/2}$. If

1435: the distance to Source A is about the same as to B1823, and the TeV

1436: emission is produced by the IC scattering of IR photons with energies

1437: around 0.03 eV (which would require electron energies $E_e \sim

1438: 1$--20 TeV), then Source A could create a few arcminute size

1439: brightening. Moreover, if there are hot and massive stars hidden

1440: within this dust/molecular complex of $\sim 4 d_4$ pc size, their

1441: shocked winds could be additional sources of TeV electrons, further

1442: enhancing the VHE $\gamma$-ray emission from this region. (VHE

1443: $\gamma$-rays apparently produced by stellar winds in the open

1444: stellar cluster Westerlund 2 have recently been detected by

1445: Aharonian et al.\ 2007.)

1446:   Although this hypothesis looks rather

1447: speculative at this point, and Source A can not be responsible for the

1448: bulk of the detected TeV radiation, dedicated observations of this

1449: region in the IR and radio would help evaluate the Source A contribution

1450: to HESS\,J1825.

1451: 

1452: \section{Conclusions}

1453: 

1454: The high-resolution {\sl Chandra} ACIS observation has allowed us to

1455: resolve the B1823 pulsar from its PWN, image the PWN structure, and measure

1456: the spectra of the pulsar and the PWN components. We found that the

1457: pulsar's X-ray luminosity,

1458: $L_{\rm PSR}\sim 0.8\times 10^{32}$ ergs s$^{-1}$ in the 0.5--8 keV band,

1459: and spectrum

1460:  are similar to those of other Vela-like

1461: pulsars, but the small number of pulsar counts and the contamination

1462: from the inner PWN component do not allow an unambiguous separation

1463: of the thermal and nonthermal components of the pulsar radiation.

1464: 

1465: We have measured the pulsar's proper motion and found its transverse

1466: velocity, $v_\perp = 443\pm 48$ km s$^{-1}$ at $d=4$ kpc, close to

1467: an average velocity of young radio pulsars. We found no indication

1468: of an SNR, in neither X-rays nor radio, between the current

1469: position of the pulsar and its birthplace, $\sim 8'$ westward.

1470: 

1471: We have confirmed

1472: the presence of the compact and extended PWNe of different shapes,

1473: surface brightnesses, and spectra.

1474: The compact PWN, of $\sim 0.5\times 0.2$ pc$^2$ size in the plane

1475: of the sky, is elongated

1476: approximately along the proper motion direcion, with the pulsar

1477: close to its leading edge. This suggests that the postshock pulsar

1478: wind in the compact PWN, responsible for its X-ray emission, is

1479: confined by the ram pressure due to the supersonic motion of the pulsar.

1480: The X-ray spectrum and luminosity,

1481: $L_{\rm comp}\sim 3\times 10^{32}$

1482: ergs s$^{-1}$, of the compact PWN are similar to those of other

1483: Vela-like PWNe.

1484: 

1485: The X-ray surface brightness distribution within the compact PWN

1486: is very nonuniform. In particular, we see a much brighter inner

1487: PWN component, $\sim 0.14\times 0.06$ pc$^2$, around the pulsar,

1488: inclined by $\sim 50^\circ$ with respect to the direction of the proper

1489: motion. Such a structure indicates that the pulsar wind is intrincically

1490: anisotropic. If the inner PWN component represents the shocked

1491: equatorial outflow

1492: from the pulsar magnetosphere, the pulsar's spin axis is strongly

1493: inclined to the proper motion direction, in contrast to many other

1494: pulsars.

1495: 

1496: The extended PWN is mostly seen south-southwest of the pulsar

1497: (up to $2.4'$ from the pulsar, in the ACIS image), almost

1498: perpendicular to the proper motion direction. Its softer spectrum

1499: indicates that its X-ray emission is generated by older electrons

1500: that have lost part of their energy because of the radiative cooling.

1501: The southward displacement of the extended X-ray PWN with respect to the

1502: pulsar might be

1503: caused by a strong pressure gradient in the

1504: ambient medium and/or by

1505: the reverse SNR shock, which

1506: has swept the PWN material southward.

1507: 

1508: The extended X-ray PWN is apparently connected to the

1509: VHE $\gamma$-ray PWN that extends in the same direction but to

1510: a much larger distance from the pulsar, up to 70 pc in the plane

1511: of the sky. The $\gamma$-ray emission is likely due to the IC

1512: scattering of background radiation off the relativistic electrons

1513: with energies up to 100 TeV supplied by the B1823 pulsar.

1514: However, the

1515: very high

1516: luminosity, $L_\gamma \sim 3\times

1517: 10^{35}$ ergs s$^{-1}$, can only be explained if the

1518: bulk of $\gamma$-radiation is generated by relic electrons,

1519: produced long ago, when the pulsar was much more powerful.

1520: The relic electrons could be displaced from the pulsar by a reverse

1521: SNR shock, but

1522: it remains unclear as to why the $\gamma$-ray PWN is so large.

1523: We have not found X-ray sources in the brightest part of the $\gamma$-ray

1524: PWN which could significantly contribute to the $\gamma$-radiation.

1525: Although the IR-radio images of the field show a star-forming region

1526: in the brightest part of the $\gamma$-ray PWN, which could enhance

1527: the $\gamma$-ray emission by supplying additional photons for the

1528: IC scattering, and perhaps even additional relativistic electrons,

1529: it can be responsible only for a small fraction of the observed

1530: VHE $\gamma$-radiation.

1531: To better understand the multiwavelength nature of the complex

1532: B1823 PWN, deep X-ray and IR observations of the field would

1533: be particularly useful.

1534: 

1535: \acknowledgements

1536: Our thanks are due to Sachiko Tsuruta, PI of the {\sl Chandra} observation,

1537: Marcus Teter, who performed the initial analysis of

1538: the X-ray data, and Vlad Kondratiev, who inspected the archival radio data.

1539: Support for this work was provided by the National Aeronautics and

1540: Space Administration through Chandra Award Number AR5-606X

1541:  issued by the Chandra X-ray Observatory Center,

1542: which is operated by the Smithsonian Astrophysical Observatory for and

1543: on behalf of the National Aeronautics Space Administration under

1544: contract NAS8-03060.

1545: This work was also partially supported by NASA grant NAG5-10865.

1546: 

1547: \clearpage

1548: 

1549: \begin{thebibliography}

1550: 

1551: \bibitem[]{}

1552: Aharonian, F.\ A., et al. 2005, A\&A, 442, L25

1553: 

1554: \bibitem[]{}

1555: Aharonian, F.\ A., et al. 2006, A\&A, 460, 365 (A06)

1556: 

1557: \bibitem[]{}

1558: Aharonian, F.\ A., et al. 2007, A\&A, 467, 1075

1559: 

1560: \bibitem[]{}

1561: Baumgartner, W.\ H., \& Mushotzky, R.\ F. 2006, ApJ, 639, 929

1562: 

1563: \bibitem[]{}

1564: Blondin, J., Chevalier, R., \& Frierson, D. 2001, ApJ, 563, 806

1565: 

1566: \bibitem[]{}

1567: Braun, R., Goss, W.\ M. \& Lyne, A.\ G. 1989, ApJ, 340,355

1568: 

1569: \bibitem[]{}

1570: Brisken, W.\ F., Carillo-Barrag\'an, M., Kurtz, S., \& Finley, J.\ P. 2006, ApJ, 652, 554

1571: 

1572: \bibitem[]{} Brogan, C.\ L., Gelfand, J.\ D., Gaensler, B.\ M.,

1573: Kassim, N.\ E., \& Lazio, T.\ J.\ W. 2006, 639, L25

1574: 

1575: \bibitem[]{} Bucciantini, N., Amato, E., \& Del Zanna, L. 2005, A\&A,

1576:     434, 189 (B05)

1577: 

1578: \bibitem[]{} Cash, W. 1979, ApJ, 228, 939

1579: 

1580: \bibitem[]{}

1581: Clifton, T.\ N., Lyne, A.\ G., Jones, A.\ W., McKenna, J., \& Ashworth, M.

1582: 1992, MNRAS, 254, 177

1583: 

1584: \bibitem[]{}

1585: Condon, J.\ J., et al.\  1998, AJ, 115, 169

1586: 

1587: \bibitem[]{} Cordes, J.\ M., \& Lazio, T.\ J.  2002, astro-ph/0207156

1588: 

1589: 

1590: \bibitem[]{}

1591: de Jager, O.\ C., et al. 2005, in Proc.\ 29th International Cosmic Ray Conf.,

1592: vol.\ 4, eds.\ B. Sripathi Acharya, S.\ Gupta, P.\ Jagadeesan, A.\ Jain,

1593: S.\ Karthikeyan, S.\ Morris, \& S.\ Tonwar (Mumbay: TIFR), p.239

1594: 

1595: \bibitem[]{}

1596: de Jager, O.\ C. 2007, ApJ, 658, 1177

1597: 

1598: \bibitem[]{}

1599: Finley, J.\ P., Srinivasan, R., \& Park, S. 1996, ApJ, 466, 938 (F96)

1600: 

1601: \bibitem[]{}

1602: Frail, D.\ A., \& Scharringhausen, B.\ R. 1997, 480, 364

1603: 

1604: \bibitem[]{}

1605: Gaensler, B.\ M., Stappers, B.\ W., Frail, D.\ A., Moffett, D.\ A.,

1606: Johnston, S., \& Chatterjee, S. 2000, MNRAS, 318, 58

1607: 

1608: \bibitem[]{}

1609: Gaensler, B.\ M., Schulz, N.\ S., Kaspi, V.\ M., Pivovaroff, M.\ J.,

1610: \& Becker, W. 2003, ApJ, 588, 441 (G03)

1611: 

1612: \bibitem[]{}

1613: Griffith, M.\ R., Wright, A.\ E., Burke, B.\ F., \& Ekers, R.\ D. 1994,

1614: ApJS, 90, 179

1615: 

1616: \bibitem[]{}

1617: Hobbs, G., Lyne, A.~G., Kramer, M., Martin, C.~E. \& Martin C.,

1618: J., 2004, MNRAS, 353, 1311

1619: 

1620: \bibitem[]{}

1621: Kargaltsev, O., \& Pavlov, G.\ G. 2007, ApJ, submitted

1622: (arXiv:astro-ph/0705.2378)

1623: 

1624: \bibitem[]{}

1625: Kargaltsev, O., Pavlov, G.\ G., \& Garmire, G.\ P. 2007, ApJ, 660, 1413

1626: 

1627: \bibitem[]{}

1628: Kaspi, V.\ M., Roberts, M.\ S.\ E., \& Harding, A.\ K. 2006, in Compact Stellar X-ray Sources, ed.\ W.\ H.\ G.\ Lewin \& M.\ van der Klis (Cambridge: Cambridge Univ.\ Press), 279

1629: 

1630: \bibitem[]{}

1631: McGary, R.\ S., R.\ S., Brisken, W.\ F., Fruchter, A.\ S., Goss, W.\ M., \& Thorsett, S.\ E. 2001, ApJ, 1192

1632: 

1633: \bibitem[]{}

1634: Mori, K., Tsunemi, H., Miyata, E., Baluta, C. J., Burrows, D.

1635:     N., Garmire, G. P., \& Chartas, G. 2001, in ASP Conf.\ Ser.\ 251, New

1636:     Century of X-Ray Astronomy, eds. H.\ Inoue \& H. Kunieda (San

1637:     Francisco: APS), 576

1638: 

1639: \bibitem[]{}

1640: Pavlov, G.\ G., Zavlin, V.\ E., Sanwal, D.,

1641:  Burwitz, V., \& Garmire, G.\ P. 2001, ApJ, 522, 129

1642: 

1643: \bibitem[]{}

1644: Pavlov, G.\ G., Teter, M.\ A., Kargaltsev, O., \& Sanwal, D. 2003,

1645: ApJ, 591, 1157

1646: 

1647: \bibitem[]{}

1648: Skrutskie, R.\ M., et al. 2006, AJ, 131, 1163

1649: 

1650: \bibitem[]{}

1651: Strong, A.\ W., Moskalenko, I.\ V., \& Reimer, O. 2000, ApJ, 537, 763

1652: 

1653: \bibitem[]{}

1654: Taylor, J.\ H., \& Cordes, J.\ M. 1993, ApJ, 411, 674

1655: 

1656: \bibitem[]{}

1657: Teter, M.\ A., Sanwal, D., Pavlov, G.\ G., Fukumura, K., Tsuruta, S. 2003a,

1658: BAAS, 36, 634

1659: 

1660: \bibitem[]{}

1661: Teter, M.\ A., Sanwal, D., Pavlov, G.\ G., \& Tsuruta, S. 2003b,

1662: BAAS, 35, 706

1663: 

1664: \bibitem[]{}

1665: Tsunemi, H., Mori, K., Miyata, E., Baluta, C. J., Burrows, D.

1666:     N., Garmire, G. P., \& Chartas, G. 2001, ApJ, 554, 496

1667: 

1668: \bibitem[]{}

1669: Tsuruta, S. 1998, Phys.\ Rep., 292, 1

1670: 

1671: \bibitem[]{} Yakovlev, D.\ G., \& Pethick, C.\ J. 2004, ARA\&A, 42, 169

1672: 

1673: \bibitem[]{}

1674: Weisskopf, M.\ C. et al.\ 2000, ApJ, 536, L81

1675: 

1676: \end{thebibliography}

1677: 

1678: \clearpage

1679: 

1680: \begin{table}[]

1681: \caption[]{PL fits to the PWN spectrum} \vspace{-0.5cm}

1682: \begin{center}

1683: \begin{tabular}{ccccccc}

1684: \tableline\tableline Region &

1685: $\mathcal{N}$\tablenotemark{a}  &

1686: $\Gamma$ & $F$\tablenotemark{b} & $F_{\rm unabs}$\tablenotemark{c} & $\chi^2$/do

1687: f  \\

1688: \tableline

1689:   Compact PWN          &

1690: $2.14\pm 0.25$      &

1691: $1.33^{-0.13}_{+0.11}$ & $1.1\pm0.1$ & $1.7\pm 0.1$ &  0.93/18 \\

1692:  Extended PWN      &

1693: $9.0\pm 1.0$ &  $1.87\pm0.14$ & $2.5\pm 0.2$ & $4.5^{+0.3}_{-0.2}$ & 1.28/13 \\

1694:       \tableline

1695: \end{tabular}

1696: \end{center}

1697: \tablecomments{The fits are for fixed $n_{\rm H,22}\equiv n_{\rm

1698: H}/(10^{22}\,{\rm cm}^{-2}) =1.0$. The uncertainties are given at 68\%

1699: confidence level for a single interesting parameter.}

1700: \tablenotetext{a}{Spectral flux in units of $10^{-5}$ photons

1701: cm$^{-2}$ s$^{-1}$ keV$^{-1}$ at 1 keV.}

1702: \tablenotetext{b}{Observed energy flux in units of $10^{-13}$ ergs cm$^{-2}$ s$^

1703: {-1}$, in the 0.7--7 keV band.}

1704: \tablenotetext{c}{Unabsorbed energy flux in units of $10^{-13}$ ergs cm$^{-2}$ s

1705: $^{-1}$, in the 0.5--8 keV band.}

1706: \end{table}

1707: 

1708: \begin{table}[]

1709: \caption[]{Fits to the pulsar spectrum} \vspace{-0.5cm}

1710: \begin{center}

1711: \begin{tabular}{ccccccc}

1712: \tableline\tableline Model & $n_{\rm H,22}$   &

1713: $\mathcal{N}$\tablenotemark{a} or $\mathcal{A}$\tablenotemark{b} &

1714: $\Gamma$ or $kT$\tablenotemark{c} & $C$  & $L_{\rm X}$ or

1715: $L_{\rm bol}$\tablenotemark{d} \\

1716: \tableline

1717: PL & $0.7$ &

1718: $6.8_{-1.6}^{+2.1}$ & $1.98_{-0.36}^{+0.40}$ & $639$ &

1719: 0.6 \\

1720: PL & $1.0$ & $11.5_{-3.2}^{+4.2}$ & $2.43_{-0.40}^{+0.53}$   &  $646$    &

1721: 0.8 \\

1722: PL & $1.3$ & $19.6^{+8.0}_{-6.2}$ & $2.92_{-0.54}^{+0.62}$   &  $657$    &

1723: 1.1 \\

1724: PL+BB(PL)    &     $1.3$ & $7.5^{+3.0}_{-1.9}$ & $1.9\pm 0.6$  &  $650$   &

1725: 0.7 \\

1726:  PL+BB(BB)  &  $1.3$ & $500$ &

1727: $97^{+4}_{-5}$   &  $650$   & $18$   \\

1728: PL+BB(PL)    & 1.3 &    $5.0^{+5.1}_{-3.1}$ & $1.7\pm 0.7$ & 651 & 0.6 \\

1729:  PL+BB(BB)  &  $1.3$ & $20$ & $139^{+9}_{-6}$ & 651 & $3.0$  \\

1730:   \tableline

1731: \end{tabular}

1732: \end{center}

1733: \tablecomments{The fits are for fixed $n_{\rm H,22}\equiv n_{\rm

1734: H}/10^{22}$ cm$^{-2}$. The uncertainties are given at 68\%

1735: confidence level for a single interesting parameter. In each case the

1736: fits were done using the C-statistics (Cash 1979) and the unbinned

1737: source and background spectra with the total of 1052 channels.

1738:  }

1739:  \tablenotetext{a}{Spectral flux in units of $10^{-6}$

1740: photons cm$^{-2}$ s$^{-1}$ keV$^{-1}$ at 1 keV.}

1741: \tablenotetext{b}{Projected area of the emitting region for the BB

1742: model, in units of km$^2$, fixed in the fits.}

1743: \tablenotetext{c}{BB temperature in eV.}

1744: \tablenotetext{d}{Unabsorbed PL luminosity in the 0.5--8 keV band or

1745: bolometric BB luminosity, in units of $10^{32}$ ergs s$^{-1}$.}

1746: \end{table}

1747: 

1748: 

1749: \begin{table}[]

1750: \caption[]{Positions and spectral properties of the field X-ray sources}

1751: \vspace{-0.5cm}

1752: \begin{center}

1753: \begin{tabular}{cccccccccc}

1754: \tableline\tableline Source\tablenotemark{a} & R.A. & Decl. &

1755: $\delta r$\tablenotemark{b} & $n_{\rm H,22}$   &

1756: $\mathcal{N}$\tablenotemark{c}  &

1757: $\Gamma$

1758: & $F_{\rm obs}$\tablenotemark{d}  & $F_{\rm unabs}$\tablenotemark{e} \\

1759: \tableline

1760: X & 18:26:14.43 & $-$13:34:48.0 &

1761: $0.4''$ & $2.9^{+2.9}_{-1.6}$ & $0.10_{-0.05}^{+0.08}$ & $0.2\pm0.5$              &

1762: $2.9\pm0.5$ & $3.7$ \\

1763: s1 & 18:26:14.06 & $-$13:40:35.8 &

1764: $0.5''$ & $4.3^{+1.1}_{-0.8}$ & $2.2_{-1.0}^{+1.7}$ & $2.1\pm0.5$

1765: & $3.2\pm0.4$ & $11.9$ \\

1766: s2 & 18:26:17.18 & $-$13:41:12.7 &

1767: $0.5''$ & $0.9^{+0.7}_{-0.3}$ & $1.8_{-0.4}^{+0.6}$ & $3.1^{-0.3}_{+0.5}$

1768:  & $1.5\pm0.3$  & $7.0$ \\

1769: s3 & 18:26:17.99 & $-$13:42:27.5 &

1770: $0.5''$ & $0.5^{+0.8}_{-0.4}$ & $0.5\pm0.2$   & $1.9\pm0.4$   &

1771: $1.7\pm0.4$  & $2.7$ \\

1772: s4 & 18:26:20.88 & $-$13:44:25.7 &

1773: $0.8''$ & $1.2^{+0.6}_{-0.4}$ & $3.5_{-0.6}^{+0.7}$ & $3.1\pm0.3$

1774:   & $2.8\pm0.3$  & $14.2$ \\

1775: s5 & 18:26:24.76 & $-$13:37:06.2 &

1776: $0.5''$ & $3.0^{+2.6}_{-2.3}$ & $0.8_{-0.4}^{+0.6}$ & $1.8\pm0.5$

1777: & $1.9\pm0.3$ & $4.6$ \\

1778: i1 & 18:25:40.55 & $-$13:40:11.2 &

1779: $0.6''$ & $1.1^{+1.2}_{-1.0}$ & $1.3_{-0.3}^{+0.4}$ & $1.9\pm0.3$

1780:   & $3.2\pm0.6$ & $6.4$ \\

1781:  \tableline

1782: \end{tabular}

1783: \end{center}

1784: \tablecomments{ The uncertainties of the fitting parameters

1785:  are given at 68\% confidence level

1786: for a single interesting parameter.

1787:  }

1788: \tablenotetext{a}{Source designation as shown in

1789: Figs.\ 1 and 8.}

1790: \tablenotetext{b}{Radius of the position error circle at the 68\% confidence level.}

1791: \tablenotetext{c}{Spectral flux in units of $10^{-5}$

1792: photons cm$^{-2}$ s$^{-1}$ keV$^{-1}$ at 1 keV.}

1793: \tablenotetext{d}{Observed flux in the 1--8 keV band, in units of

1794: $10^{-14}$ ergs  cm$^{-2}$ s$^{-1}$.}

1795: \tablenotetext{e}{Best-fit value of the unabsorbed

1796: PL flux in the 0.5--8 keV band, in units of $10^{-14}$ ergs  cm$^{-2}$

1797: s$^{-1}$.}

1798: %\tablenotetext{f}{Offset between the X-ray and optical positions.}

1799: \end{table}

1800: 

1801: 

1802: 

1803: \end{document}

1804: