0711.4171/ms.tex
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: \makeatletter
9: 
10: \citestyle{apj}
11: %\usepackage{float,longtable,supertabular}
12: %\documentclass[useAMS,usenatbib,18pt]{mn2e}
13: %\documentclass[usenatbib]{mn2e}
14: %\usepackage{natbib}
15: %\usepackage{natbib209}
16: \usepackage{graphics,epsfig,lscape}
17: 
18: 
19: %\bibpunct{(}{)}{;}{a}{}{,}
20: %\bibliographystyle{mn2e}
21: \bibliographystyle{apj}
22: % If your system does not have the AMS fonts version 2.0 installed, then
23: % remove the useAMS option.
24: %
25: % useAMS allows you to obtain upright Greek characters.
26: % e.g. \umu, \upi etc.  See the section on "Upright Greek characters" in
27: % this guide for further information.
28: %
29: % If you are using AMS 2.0 fonts, bold math letters/symbols are available
30: % at a larger range of sizes for NFSS release 1 and 2 (using \boldmath or
31: % preferably \bmath).
32: %
33: % The usenatbib command allows the use of Patrick Daly's natbib.sty for
34: % cross-referencing.
35: %
36: % If you wish to typeset the paper in Times font (if you do not have the
37: % PostScript Type 1 Computer Modern fonts you will need to do this to get
38: % smoother fonts in a PDF file) then uncomment the next line
39: % \usepackage{Times}
40: 
41: %%%%% AUTHORS - PLACE YOUR OWN MACROS HERE %%%%%
42: 
43: 
44: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
45: 
46: \slugcomment{}
47: \shorttitle{{\sl Chandra} observations of PSR B1929+10. }
48: \shortauthors{Misanovic, Pavlov,  \& Garmire}
49: 
50: %\usepackage{babel}
51: \makeatother
52: 
53: \begin{document}
54: 
55: \title{{\sl Chandra} observations of the pulsar 
56: B1929+10 and its
57:  environment}
58: 
59: 
60: \author{ Z.\ Misanovic, G.\ G.\ Pavlov, and G.\ P.\ Garmire}
61: 
62: 
63: 
64: \affil{Dept.\ of Astronomy and Astrophysics,
65: The Pennsylvania State
66: University, 525 Davey Lab., University Park,
67: PA 16802}
68: 
69: 
70: 
71: \begin{abstract}
72: We report on  two {\sl Chandra} observations of the 3-Myr-old pulsar B1929+10, 
73: which reveal a faint compact 
74: ($\sim 9'' \times 
75: 5''$) nebula elongated in the 
76: direction perpendicular to the pulsar's proper motion, two patchy wings, 
77: and a possible short ($\sim 3''$) jet emerging from the pulsar. In addition, 
78: we detect a tail extending up to at least $4'$ in the direction opposite to 
79: the pulsar's proper motion,
80: aligned with the $\sim15'$-long tail  detected in  {\sl ROSAT} and 
81: {\sl XMM-Newton} observations. The overall  morphology of the  nebula  
82: suggests that the shocked pulsar wind is confined by the ram pressure 
83: due to the pulsar's supersonic speed. The shape of the compact nebula 
84: in the immediate vicinity of the pulsar seems to be consistent with the 
85: current MHD models. However, since these models  do not account yet for 
86: the change of the flow velocity at larger distances from the pulsar, 
87: they are not able to constrain the  extent of the long pulsar tail. 
88: The luminosity of the whole nebula as seen by {\sl Chandra} is $L_{\rm PWN}
89: \sim 10^{30}$ ergs s$^{-1}$ in the 0.3--8 keV band, for the distance
90: of 361 pc.
91: Using the {\sl Chandra} and {\sl XMM-Newton} data, we found that
92: the pulsar spectrum is comprised of non-thermal (magnetospheric)
93: and thermal components.
94: The non-thermal 
95:  component can be described by a power-law model 
96: with photon index
97: $\Gamma\approx 1.7$ 
98: and luminosity $L_{\rm PSR}^{\rm nonth}\approx 1.7\times 10^{30}$
99: ergs s$^{-1}$ in the 0.3--10 keV band. The blackbody fit for
100: the thermal component, which presumably emerges from hot polar caps,
101:  gives the temperature $kT\approx 0.3$ keV and projected emitting area $A_{\perp}\sim 3\times
102: 10^3$ m$^2$,
103: corresponding to the bolometric luminosity
104: $L_{\rm bol}\sim (1$--$2)\times 10^{30}$ ergs s$^{-1}$.
105: \end{abstract}
106: 
107: 
108: \keywords{pulsars: individual (PSR B1929+10) --- stars: neutron ---
109: 	 X-rays: stars}
110: 
111: 
112: \section{Introduction}
113: \label{intro}
114: 
115: Pulsar wind nebulae (PWNe), generated 
116: when the magnetized relativistic pulsar winds  
117: shock in the ambient medium,
118:  have been observed 
119: around 
120: $\approx 50$
121: pulsars 
122: \citep[see][for a recent review]{2008AIPC..983..171K}.
123: The shocked pulsar winds produce  
124: synchrotron
125:  radiation  from radio frequencies 
126: through
127:  $\gamma$-rays, 
128:  revealing 
129:  properties of the winds 
130: and the ambient medium.
131: 
132: 
133: The PWN morphology depends on the pulsar's velocity,  pressure,
134:  temperature, and magnetic field of the ambient medium, and the geometry of the wind outflow.
135: For instance,
136: the torus-jet PWN structures observed  around young pulsars 
137: in supernova remnants (SNRs), such as the Crab and Vela 
138: PWNe \citep[]{2000ApJ...536L..81W,2003ApJ...591.1157P},
139: are commonly interpreted as anisotropic outflows 
140: with equatorial and polar components
141: confined by the pressure of the hot gas
142: in the host SNR interiors, where the speed of sound exceeds the pulsar's speed
143:  \citep[]{2003MNRAS.344L..93K,2004A&A...421.1063D}.
144: When the pulsar is moving with a supersonic
145: speed with respect to the ambient medium, bow-shock PWNe are formed
146: \citep{2005A&A...434..189B,2005ApJ...630.1020R},
147: such as the PWNe around PSR J1747$-$2958 
148: \citep[the Mouse PWN;][]{2004ApJ...616..383G}
149:  and PSR B1957--20 \citep{2003Sci...299.1372S}.
150:  The pulsar's supersonic speed causes the ram pressure
151:  to exceed  the  ambient pressure, so that  the wind termination 
152: shock acquires  a bullet-like shape \citep[e.g.][]{2005A&A...434..189B}
153:  and forms a PWN,   sometimes with a long tail   
154:   extending behind the pulsar. The pulsar's speed becomes supersonic
155: as it encounters a relatively cold interstellar medium (ISM), 
156: either after leaving the 
157:  SNR in which it was born, or when the host SNR dissolves in the ISM
158:  background. These last stages of the SNR evolution occur
159:  after $\sim 10^5$ years.
160:     
161: 
162: 
163: The majority of middle-aged 
164: ($\sim 10^5$--$10^6$ yr) and old ($\gtrsim 10^6$ yr) 
165: pulsars are expected to  move
166:  supersonically through the surrounding ISM, and as long as their  winds are
167:  sufficiently strong, they are expected to produce observable
168: bow-shock PWNe. Observations of these objects  provide an  
169:  opportunity to study  bow-shock morphologies and   probe pulsar winds and
170:  their surroundings at later stages of pulsar evolution.
171: 
172: 
173: 
174: 
175: Until now, only a handful of PWNe around middle-aged pulsars have been
176:  detected: e.g.,  PSR\,J0538+2817 \citep{2003ApJ...585L..41R}, 
177: PSR\,B0355+54 \citep{2007Ap&SS.308..309M},
178:  and Geminga
179: \citep{2004MmSAI..75..470C,2006ApJ...643.1146P} in X-rays,  
180:  PSR\,B0906--49  at radio frequencies 
181:  \citep{1998ApJ...499L..69G}, and 
182: PSR\,B1951+32 in X-rays and H$\alpha$
183:  \citep{2004ApJ...610L..33M}.
184:  However, one of the first bow-shock PWN candidates  
185: was detected
186:  serendipitously around a much older pulsar, PSR B1929+10
187: (hereafter B1929).
188: This is one of the nearest pulsars 
189: \citep[$d=361^{+10}_{-8}$ pc;][]{2004ApJ...604..339C},
190:  with a period $P=226.5$ ms and period derivative
191: $\dot{P}=1.16\times 10^{-16}$ s s$^{-1}$. Despite the large
192: spindown age, $\tau\equiv
193: P/2\dot{P} = 3.1$ Myr, its spindown power, $\dot{E}\equiv 4\pi^2 I\dot{P}/P^3
194: =3.9\times 10^{33}$ ergs s$^{-1}$ for the moment of inertia $I=1\times 10^{45}$
195: g cm$^2$, is still relatively high.
196: As the pulsar's transverse velocity in the plane of the sky,
197: $V_{\perp}=177^{+4}_{-5}$\,km\,s$^{-1}$ \citep{2004ApJ...604..339C},
198: substantially exceeds the typical ISM sound speed, one can expect
199: that the pulsar wind outflow forms a bow-shock PWN. An indication of such a
200: PWN was first noticed by \citet{1993Natur.364..127W}, who detected
201: a very long ($\sim1.6$\,pc) tail-like structure 
202: behind this pulsar in  {\sl ROSAT} data. 
203: The tail behind B1929
204:  had been the longest extended structure associated with a compact 
205: Galactic object until the recent
206: discovery of a 
207: 6-pc-long X-ray tail behind the middle-aged pulsar 
208: PSR\,J1509$-$5850, 
209: reported by \citet{2006HEAD....9.0757K} 
210:  and described in detail by 
211: \citet{2008arXiv0802.2963K}
212: Similar  
213: (albeit shorter) tails have been found behind a few other pulsars
214: \citep{2008AIPC..983..171K}
215:  suggesting that such extended PWN morphologies might be ubiquitous.
216: 
217: 
218: 
219: 
220: 
221: The B1929 pulsar  has been extensively observed in X-rays, 
222: in an attempt to study the magnetospheric and thermal components of its
223: emission and compare them with the models for pulsar radiation. 
224: In particular, the thermal emission from 
225: polar caps heated by relativistic particles accelerated in the pulsar
226: magnetosphere is predicted by the current pulsar models
227: \citep[e.g.,][]{2001ApJ...556..987H,2002ApJ...568..862H}, but
228:  a limited sample of the 
229:  observed old pulsars
230: suggests that the nonthermal magnetospheric emission with a power-law spectrum
231: dominates at higher X-ray energies, $E\gtrsim 2$ keV,
232: while the nature of radiation at lower energies has been a matter of debate
233: \citep{2004ApJ...615..908B,2005ApJ...633..367B,2004ApJ...616..452Z,2006ApJ...636..406K}.
234: Observations of B1929 with the
235: {\sl Einstein} \citep{1983IAUS..101..471H}, {\sl ROSAT} \citep{1994ApJ...429..832Y,1997A&A...326..682B}, and {\sl ASCA} \citep{1997ApJ...482L.159W,1998AdSpR..21..213K} observatories
236:  were not able to constrain the nature of the X-ray emission
237:  from the pulsar
238: because its spectrum  could be equally well described  as thermal
239: or  non-thermal. The same result has been obtained by combining the available {\sl ROSAT} and {\sl ASCA}
240:  data by \citet{2005A&A...434.1097S}, who found that the spectrum of B1929 
241: could be fitted  either by a power-law or a  two-temperature blackbody 
242:  model. 
243: 
244: 
245: 
246: 
247: 
248: Analyzing the images obtained in the 40 ks {\sl ROSAT} PSPC observation,
249: \citet{1993Natur.364..127W}  
250: noticed an elongated structure extending approximately 10$'$ behind
251:  B1929, almost aligned with the direction of its proper motion. 
252: \citet{1993Natur.364..127W} have suggested that the  detected tail
253: is synchrotron radiation from the  wind of 
254:  B1929 confined by the ram pressure.
255: The detection of this diffuse emission  has also 
256:  been reported by \citet{1994ApJ...429..832Y} in the analysis of
257: the same  {\sl ROSAT} PSPC data.  
258: \citet{1998AdSpR..21..213K} 
259: also claim detection of some
260:  diffuse emission in the vicinity of B1929 in the {\sl ASCA} images,
261: but the poor angular resolution 
262: of {\sl ASCA} prevented a detailed analysis and did not allow to separate
263: the contribution from background sources in the alleged PWN emission. 
264: 
265: 
266: 
267: In a multi-wavelength study of B1929 \citep{2006ApJ...645.1421B},
268:    the tail-like structure extending up to more than 1\,pc behind the pulsar
269:  has been observed by
270:  {\sl XMM-Newton}, confirming the {\sl ROSAT} finding.
271: The spectral analysis suggests that this   
272: extended emission is nonthermal, likely produced by synchrotron radiation
273: of the shocked pulsar wind.  
274: \citet{2006ApJ...645.1421B} also report on an elongated faint diffuse radio
275: emission found in the Effelsberg radio continuum survey data, whose brightness
276: distribution roughly coincides with the X-ray tail.
277: These authors also conclude that the radiation of the B1929 pulsar is 
278: predominantly nonthermal (i.e., magnetospheric), with a rather soft 
279: power-law spectrum (photon index $\Gamma\approx 2.7$). 
280: 
281: 
282: 
283: To understand the nature of the extended emission behind B1929
284: and prove that it is indeed a tail of a bow-shock PWN,
285: the nebula in the immediate vicinity of the pulsar should be observed with 
286: an angular resolution  much better than those of {\sl ROSAT} PSPC
287: ($25''$) or {\sl XMM-Newton} ($15''$ 
288: half-energy width).
289: Therefore, we conducted two {\sl Chandra}
290:  observations of the B1929 pulsar and its surroundings. 
291: The first results, including the detection of a faint nebula in the 
292: immediate vicinity of
293:  B1929, have already been reported by \citet{2006HEAD....9.0758M}. 
294: In this paper we present a more detailed analysis
295: of these observations.  
296: 
297: 
298: 
299: 
300: \section{Observations and results}
301: \label{observations-and-results}
302: 
303: 
304: 
305: We observed the field around B1929 with the Advanced CCD Imaging
306:  Spectrometer (ACIS) aboard 
307: {\sl Chandra} (Table~\ref{table-observations}).
308: Both observations were carried
309:  out in  very faint mode on the ACIS-S3 chip, with the target imaged
310: about $8''$ from the optical axis.
311: The data were analyzed
312: using the {\sl Chandra} Interactive Analysis of Observations (CIAO) software
313:  (ver.\,3.3.0.1; CALDB ver.\,3.2.0 for observation 6657, and CALDB ver.\,3.2.2 for observation
314:  7230).
315: 
316: Table~\ref{table-observations} also includes  three archived {\sl
317:   XMM-Newton} observations of B1929, which we analyzed in addition to the {\sl
318:   Chandra} data, using SAS (ver. 7.1.0).
319: 
320: 
321: 
322: \clearpage
323: \begin{figure}
324: \begin{center}
325: \includegraphics[height=5cm,angle=0]{fig1a.ps}\\
326: \includegraphics[height=5cm,angle=0]{fig1b.ps}\\
327: \includegraphics[height=5cm,angle=0]{fig1c.ps}
328: \end{center}
329: \caption{ 
330: Images of the  field around 
331: B1929 in the 0.3--8 keV band, combining the two ACIS-S3 observations. 
332: {\em Top}: 
333: Unsmoothed image (pixel size is $0.492''$).
334: The circular and  elliptical regions mark the smallest structures detected 
335: around 
336: the pulsar at 
337: $\geq 3 \sigma$ levels
338:  in at least one of the observations (see Table~\ref{table-regions}).
339: {\em  Middle}: The same image smoothed using a Gaussian of FWHM $1''$. 
340: The blue arrow
341:  shows 
342:  the direction of the measured pulsar's proper motion
343:  \citep{2004ApJ...604..339C}, while the black arrow 
344: indicates the proper motion 
345: corrected for the Galactic rotation and solar peculiar velocity 
346: (see text for details).  
347: {\em Bottom}: 
348: In this image the pulsar contribution is subtracted (see text for 
349: details) and smoothing with a Gaussian of FWHM $0.5''$ is applied. 
350: The brightness scales
351: in the middle and bottom panels
352:  are selected to emphasize different components of the extended emission in
353:  the vicinity of B1929.  
354: The intensity scale is linear, with a range of 0 to 1.8 
355: cts/pixel in the top
356: panel, and 0 to 0.9 
357: cts/pixel (middle and bottom).
358: The position of the pulsar  is indicated by
359:  a 1.5$\arcsec$ circle in all panels.
360: \label{fig-unsmoothed-images}}
361: \end{figure} 
362: 
363: 
364: \clearpage
365: \begin{figure}
366: \begin{center}
367: \includegraphics[height=6.5cm,angle=0]{fig2.ps}
368: \caption{ 
369:   Radial profiles of the combined 
370: pipeline-processed and subpixelized data (solid lines),
371: and
372: MARX-simulated data (dotted lines) for various blurring parameters (see text).
373:   The counts are measured in annular regions centered at the pulsar's position.
374: \label{fig-radial-profiles}}
375: \end{center}
376: \end{figure} 
377: \clearpage
378: 
379: \subsection{X-ray imaging}
380: \label{x-ray-imaging}
381: 
382: 
383: We first produced new level-2 event files with the pipeline 
384: pixel randomization disabled and then applied the
385: method  by 
386: \citet{2001ASPC..251..576M} and \citet{2001ApJ...554..496T},
387: which allows one to reach a subpixel spatial resolution. 
388:  To minimize the background contribution,
389: we chose the energy band of 0.3--8 keV.
390: 
391: 
392: 
393: The only relatively bright source on the S3 chip
394: is the B1929 pulsar, centered at
395: R.A. = $19^{\rm h} 32^{\rm m} 13.999^{\rm s}$, 
396: decl.\ = $10^\circ 59' 32.64''$, and
397: R.A. = $19^{\rm h} 32^{\rm m} 14.000^{\rm s}$,
398: decl. = $10^\circ 59' 32.85''$ (J2000)
399: in the first and second observation, respectively (the centroiding
400: $1\, \sigma$ uncertainty is
401: $0.02''$ for each coordinate, 
402: as determined by the CIAO procedure {\sc celldetect}).
403: The offsets of these positions
404: from the radio positions of the pulsar extrapolated to the epochs 
405: of the two  observations
406: ($0.17''$ and $0.14''$ in R.A., $-0.04''$ and $0.16''$ in decl., 
407: respectively)
408: are smaller than the error in the absolute {\sl Chandra} astrometry,
409: $0.45''$ at the 90\% confidence level for on-axis observations on the S3 
410: chip\footnote{See \S\,5.4 and
411:  Fig.\ 5.4 in the Chandra Proposers' Observatory
412: Guide, ver.\ 10, at http://asc.harvard.edu/proposer/POG.}.
413: 
414: 
415: 
416: \subsubsection{Combined ACIS image.}
417: 
418: Although both images show
419: some extended emission in the immediate vicinity of the pulsar,
420: its surface brightness is very low.
421: Therefore, to increase the signal-to-noise ratio, 
422: we have to combine the two ACIS observations.
423: To align the images, we reprojected the event files to a common position,
424: corrected the aspect solution of the second observation for the 
425: pulsar's positional difference of $0.02''$ in R.A. and $0.21''$ in decl., 
426: and  added the two images.  
427: 
428: 
429:   
430: A nebula surrounding the pulsar 
431: is clearly seen in the unsmoothed and smoothed images shown in the top and
432:  middle panels of
433:  Figure~\ref{fig-unsmoothed-images}. 
434: To separate this extended
435:    emission from the pulsar, we simulated the
436:  ACIS-S3 point source 
437: observations using the {\sl Chandra} data simulator MARX\footnote{See
438: http://space.mit.edu/CXC/MARX/}.
439: We produced simulated images of a point
440: source at the same position on the detector,
441:  and with the same X-ray flux and spectral shape
442:  as those of the observed pulsar,
443:  running MARX for an effective exposure 100 times longer than the actual
444:  observations
445:  to reduce statistical errors.
446: The simulated images were then  scaled 
447: to the actual exposure times, combined,
448: and compared with the data, both pipeline-processed and those
449: with subpixel resolution.
450: 
451: 
452: 
453: The width of the simulated point spread function 
454: (PSF) depends on the value of the MARX parameter
455: Dither Blur, which is a combination of the aspect reconstruction error, 
456: ACIS pixelization, and pipeline pixel randomization.
457: According to the MARX manual\footnote{See also http://www.astro.psu.edu/xray/docs/TARA/ae\_users\_guide/node16.html for a more detailed discussion.},
458: appropriate values of this parameter are about $0.35''$ if the pixel
459: randomization is applied and $0.20''$--$0.25''$ if it is switched off, but the
460:  best values of this parameter may vary from observation to observation.
461: 
462: 
463: 
464: We found that for the default Dither Blur = $0.35''$,
465:  the simulated PSF is significantly {\em broader} than
466:  the core of the observed 
467: B1929 image, not only for the sharper image with subpixel resolution
468:  but also for
469: the pipeline-processed one (see Fig.~\ref{fig-radial-profiles}).
470: We repeated the simulation for a number of smaller Dither Blur values and found
471: that  the PSF core of the simulated data with Dither Blur reduced 
472: to 0\farcs15 matched well to the data with subpixel resolution, 
473: while a simulation  with Dither Blur = $0.25''$ is close to the 
474: pipeline-processed
475: data. 
476: In both  cases, 
477: the shape 
478:  of the simulated  PSF  matched well to the observed one only up to
479: $\approx 0.6''$ from the center,  
480: indicating the 
481: contribution of extended emission at larger distances from the pulsar.
482: Using the excess of the data counts with respect to the simulation in the
483: $0.5''$--$1.5''$ annulus, we estimate the signal-to-noise ratio of the PWN
484: to be $\gtrsim 20$.
485: For comparison, we also simulated the ACIS-S observation of the central
486: compact source of the Cas A SNR
487: (ObsID 6690) 
488: and found no indication of an extended emission around this point source.
489: 
490: 
491: 
492: The PWN image in the vicinity of B1929, obtained by
493: subtracting the simulated data with Dither Blur = $0.15''$ from the 
494: subpixel resolution image and slight smoothing with a Gaussian of 
495: $0.5''$ FWHM, 
496: is shown in the bottom panel of Figure~\ref{fig-unsmoothed-images}.
497: The image reveals a compact, $\sim 9''\times 5''$, emission 
498: elongated in the direction perpendicular to the pulsar's proper motion,
499: two faint wings extending in the direction opposite to the pulsar's motion
500: and seen up to $\sim 11''$ from the pulsar,
501: and a hint of a short, $\sim 3''$, linear structure immediately behind the 
502: pulsar.
503: 
504: \clearpage
505: \begin{figure*}
506: \begin{center}
507: \includegraphics[height=4.7cm,angle=0]{fig3a.ps}
508: \includegraphics[height=4.7cm,angle=0]{fig3b.ps}
509: \includegraphics[height=4.7cm,angle=0]{fig3c.ps}
510: \end{center}
511: \caption{
512: {\em Left:}   
513: ACIS-S3 image of  
514:  B1929 and its environment produced by combining observations 6657 and 7230,
515: in the 0.3--8 keV band. The unbinned  image is smoothed using a Gaussian of FWHM 2\arcsec.
516: The elliptical regions mark the smallest structures  detected around B1929 
517: (see Table~\ref{table-regions} and Fig.\ \ref{fig-unsmoothed-images}). 
518: The pulsar's position is marked by
519:  a  $1.5''$ circle, while the $40'' \times 20''$ box shows the region 
520: used for the background evaluation. 
521: The arrows indicate the direction of the pulsar's proper motion 
522: (as in Fig.~\ref{fig-unsmoothed-images}).
523: The green circles enclose the optical/infrared sources in the field of view. The X
524: marks the X-ray source found with {\sc celldetect} in the combined image,
525: while Y and Z mark other possible X-ray point sources that might be just below
526: the decection threshold.
527:  {\em Middle:} Unsmoothed 
528: combined image showing the same region. The image is binned by a factor of two. The ellipse
529: shows the 
530:  region of enhanced emission, in which the diffuse emission is detected at
531:  a  $\sim$10$\sigma$ level in the combined image.  The red box and the line
532:  mark the 
533: size of the rectangular region and the direction in which the region is moved
534:  to produce the linear profile shown in Fig.~\ref{fig-profile}. 
535:  {\em Right:} The same unbinned image smoothed using an adaptive Gaussian kernel with the size self-adjusting
536: to show the structures with the signal-to-noise ratio in the range 2.2 to 4.
537: The overlaid contours correspond to the combined {\sl XMM-Newton} observations. The contour levels are the same as in Fig.\ \ref{fig-large}.
538:  The intensity scale is linear in all  images, with a range of 0 to 0.2
539: cts/pixel in the
540:  left and right panels and 0 to 1.8 
541: cts/pixel (middle). 
542: \label{fig-combined-conv-smooth-image}}
543: \end{figure*} 
544: \clearpage
545: 
546: 
547: After confirming that there is a PWN in the immediate vicinity of B1929, 
548: we proceeded by searching for the nebular emission farther from the pulsar.  
549: Figure~\ref{fig-combined-conv-smooth-image} shows a $2.5' \times 2.5'$ image
550:  combining the two {\sl Chandra} observations. The image on the
551:  left is
552:  smoothed using a fixed-size Gaussian of 2$\arcsec$ FWHM, while the right panel
553:  shows a heavily
554: smoothed image using an adaptive kernel with an adjustable size to show the
555:  structures with the signal-to-noise ratio in the range 2.2 to 4.
556: The middle panel shows the combined unsmoothed data binned by a factor of two.
557: The images
558: show several regions of extended emission:
559:  a bow-shock-shaped
560:  extended emission
561:  with the Southern Wing (region 1) and Northern Wing (region 2),
562:  some faint emission in front of the pulsar
563:  (region 3 or the Front),
564: faint diffuse emission elongated in the direction
565:  opposite to the direction of the pulsar's projected velocity, with an enhancement at the
566:  distance of
567:  $\approx$\,$30\arcsec$ (region 4 or the Inner Blob).
568: A more 
569: extended but fainter region 5 (Outer Blob)
570: is seen farther behind the pulsar, at a distance of 
571:  approximately $1.5\arcmin$--$2\arcmin$. 
572: Regions 1 through 3 are also shown in the top panel of Figure~\ref{fig-unsmoothed-images}, while the numbers of counts in the regions are given in
573: Table~\ref{table-regions}.
574:  
575: 
576: 
577: Although these
578: structures (regions 1 to 5)  
579: were detected at a level of at least 3$\sigma$ 
580:  in one or both individual observations, the small number of detected 
581: counts prevented the detailed spectral analysis.  
582: Therefore, we selected a larger region
583: of the extended emission to study the spectral properties of the PWN
584:    (\S\ref{spectrum-nebula}).
585: The region is marked by the 
586:  ellipse ( semi-axes 12$''$ and 25$''$) in the
587:  middle panel of Figure~\ref{fig-combined-conv-smooth-image}, detected at a level of at least 5$\sigma$ in each
588:  observation.
589: 
590: 
591: 
592: The blue arrows in Figures~\ref{fig-unsmoothed-images} 
593:   and \ref{fig-combined-conv-smooth-image}
594:  indicate  the direction of the pulsar's proper motion 
595:  ($\mu _{\alpha}=94.09\pm0.11$\,mas\,yr$^{-1}$, 
596: $\mu _{\delta}=42.99\pm0.16$\,mas\,yr$^{-1}$)
597: as measured by  \citet{2004ApJ...604..339C} with respect to extragalactic 
598: reference radio sources.  
599: To determine  
600: the proper motion 
601: with respect to the local ISM,  
602: we corrected the
603: measured proper motion 
604:  for the 
605: effects of Galaxy rotation 
606: and the Sun's peculiar motion with respect to the local standard of rest
607:  (LSR). 
608: Following the procedure by \citet{1987AJ.....93..864J}, 
609: and adopting the value for the Sun's peculiar velocity 
610: from \citet{1986ApJ...303..724B} and Oort constants from 
611: \citet{1987gady.book.....B}, 
612: we obtained the corrected proper motion
613: $\mu_\alpha=91.6\pm 0.2$ mas yr$^{-1}$,
614: $\mu_\delta=49.1\pm 0.2$ mas yr$^{-1}$, shown by 
615: the black arrows in Figures~\ref{fig-unsmoothed-images} and 
616: \ref{fig-combined-conv-smooth-image}.
617: The total proper motion, $\mu=104.0\pm 0.3$ mas yr$^{-1}$,
618: corresponds to the transverse velocity $v_\perp = 178\pm 8$ km s$^{-1}$.
619: Its direction (position angle $61.7^\circ \pm 0.1^\circ$, 
620: counted East of North)
621:  seems to be  
622: better aligned with the extended tail of diffuse emission,
623: which further 
624: supports the PWN interpretation of the detected X-ray emission
625:  around this pulsar.
626: 
627: 
628: 
629: Since the observed diffuse emission might be partly due to unresolved 
630: point X-ray sources in the field of view (e.g., stars or background AGNs), 
631: we examined
632:  available
633:  optical, near-infrared, and radio data of the region around B1929. 
634: We found several sources (marked with the green circles in Fig.\ \ref{fig-combined-conv-smooth-image}, {\em left}) in the 
635: USNO-B2 \citep{2003AJ....125..984M} and 2MASS \citep{2003tmc..book.....C}
636:  catalogs,
637:  which coincide with the regions
638:  resembling faint point-like objects in our smoothed images. Although 
639:   these possible sources are 
640: very faint in the {\sl Chandra}
641:  images
642: (only one 
643:  of these sources, marked with an X, was detected by the 
644: CIAO tool {\sc celldetect} above the source threshold of 3, while Y and Z mark
645: other two possible X-ray point sources that might be just below the detection threshold), 
646: they might contribute to the observed extended emission.
647: It is clear, however, that most of the detected tail-like structure 
648: is indeed diffuse emission associated with the pulsar.
649:  
650: 
651:  
652: 
653: Figure~\ref{fig-profile} shows the brightness distribution along the tail, 
654: which further demonstrates 
655: the excess emission over the background in front of the pulsar, 
656: and also in the direction approximately opposite to the projected pulsar's
657:  velocity. The photons were collected  from  $1\arcsec \times 35\arcsec$
658:  rectangular regions (see Fig.~\ref{fig-combined-conv-smooth-image}) along the
659:  tail, and also from the background regions of the same area on both sides of
660:  the tail in the unsmoothed combined {\sl Chandra} image.
661:  We show only the brightest
662: part of the tail, up to $\sim 40''$ from the pulsar, because the extended
663:  emission farther away is too faint and
664: cannot be distinguished from the background using such small collecting
665:  regions, but it is much better seen in the smoothed image.
666: 
667: 
668: \clearpage
669: \begin{figure}
670: \begin{center}
671: \includegraphics[height=8cm,angle=270]{fig4.ps}
672: \end{center}
673: \caption{X-ray linear brightness profile along the direction of the pulsar's 
674: tail.
675: The black data points were obtained
676: by counting the photons in the $1\arcsec \times 35\arcsec$ 
677: rectangular regions (see Fig.~\ref{fig-combined-conv-smooth-image}).
678: The pulsar moves towards the negative values on the X-axis. 
679: The red  
680: data points show the X-ray profile of the background emission, 
681: evaluated in the rectangular regions of the same size on both sides of the
682:  tail emission.
683:  \label{fig-profile}}
684: \end{figure} 
685: \clearpage
686: 
687: To look for extended emission at even larger distances from the
688: pulsar and compare it with the $10'-15'$ tail detected by
689: \citet{2006ApJ...645.1421B}, we
690: produced heavily binned 
691: images shown in 
692: Figure~\ref{fig-large}. 
693: Although 
694: the smaller ACIS-S field of view and
695:  the relatively short {\sl Chandra} exposure
696: do not allow us to detect the full extent of the faint diffuse 
697: emission observed by {\sl XMM-Newton}, 
698: at least some parts of the B1929 long tail 
699: are 
700: seen up to $\approx 4'$ from the pulsar in  
701: the combined ACIS-S3 image.
702:  The long tail extends  from the 
703: compact PWN resolved in the {\sl Chandra} images in the same direction 
704: as the emission detected in {\sl XMM-Newton} data. 
705: 
706: 
707: 
708: 
709: \clearpage
710: \begin{figure*}
711: \begin{center}
712: \includegraphics[height=5cm,angle=0]{fig5a.ps}
713: \includegraphics[height=5cm,angle=0]{fig5b.ps}
714: \end{center}
715: \caption{
716: 10\arcmin $\times$ 8\arcmin\,  ACIS-S3 field around B1929. 
717: The image on the left is produced from the combined event file, binned
718: by a factor of 4 and then smoothed by a Gaussian of FWHM 6$''$.
719: The contours, at levels 
720: 0.08, 0.1, 0.12, 0.13, 0.14, 0.15, 0.2, 0.3, 0.5, 0.9, 2 and 5 
721: counts\,pixel$^{-1}$, correspond to the three combined 
722: {\sl XMM-Newton} observations (MOS1 and MOS2 data, pixel size 1.1$''$).
723:    The right panel shows the same image, which is binned
724:  by a factor of 8 but not smoothed.
725: To minimize the background, both images  include only
726: the 0.5--7 keV 
727:  energy range. 
728: The intensity scale is linear, with ranges of 0 to 0.8 
729: counts\,pixel$^{-1}$ (left) and
730: 0 to 6 counts\,pixel$^{-1}$ (right).
731:  \label{fig-large}}
732: \end{figure*} 
733: \clearpage 
734: 
735: 
736: \subsubsection{Individual observations and variability}
737: \label{variability}
738: 
739: Since some PWNe show significant temporal variations on time-scales of months 
740: \citep[e.g., the Crab and Vela PWNe,][]{2002ApJ...577L..49H,2003ApJ...591.1157P}, we have examined the individual ACIS-S observations, which are separated 
741: by $\simeq$6 months, in search for  variability of the  PWN of B1929.
742: Figure~\ref{fig-individual-conv-smooth-image} shows the 
743: images of
744:  the individual
745: {\sl Chandra} observations of B1929 and its PWN.
746: Similarly to the combined data, the images on the left were smoothed 
747: using a Gaussian of FWHM 2\arcsec, while on the right we show  the same 
748: images adaptively smoothed to reveal structures with the 
749: signal-to-noise ratio in the range 2.2 to 4. The bright features seen
750: in the combined image (Fig.~\ref{fig-combined-conv-smooth-image}) are also
751:  visible in these individual images, but their apparent brightness seem to 
752: show a substantial difference between the observations.
753: For example, the Inner Blob (region 4), clearly detected in the
754:  observation 6657
755: (December 2005), is not visible in the 
756: image of the observation 7230 (May 2006), while the Outer Blob 
757: (region 5) appears fainter and
758: slightly smaller in the second observation. 
759:  In addition,
760: the Southern Wing and the emission in front of the pulsar (region 3) are
761:  slightly brighter in the observation 6657, while 
762:  the Northern Wing is brighter in the second observation. 
763: 
764: 
765: 
766: To quantify this apparent variability,
767:  we measured the total background-subtracted 
768: counts detected in each 
769: of the extended regions 
770: in the unsmoothed images of both data sets,
771:  and  calculated the corresponding 
772: surface brightness 
773: (listed in Table~\ref{table-regions}). The background
774: emission was determined using the rectangular region with an area of
775:  784 arcsec$^2$  shown in
776:   Figure~\ref{fig-combined-conv-smooth-image} (left), from which we measured 
777:   $(3.3\pm0.4)\times 10^{-6}$\,cts\,s$^{-1}$\,arcsec$^{-2}$ and $(2.2\pm0.4)\times 10^{-6}$\,cts\,s$^{-1}$\,arcsec$^{-2}$  in the observations 6657 and 
778: 7230, respectively.
779: 
780: 
781: 
782: 
783: 
784: As shown in Table~\ref{table-regions}, the
785:  surface brightness of the five extended regions is found to change 
786: by a factor of two (e.g. region 1) or  even 
787: four (region 3). However, due to the large statistical errors,
788:  the variability 
789: could only be detected at a level of 
790: up to $\approx2.5\sigma$ (region 5).
791: Deeper observations would be needed to further  examine possible variability
792:  of the nebula associated with this old pulsar.
793:  We also note that the variation of the pulsar count rates
794: between the two {\sl Chandra}
795: observations was found to be within the statistical uncertainty
796:   (see \S~\ref{spectral-analysis-pulsar}).
797: 
798: 
799: 
800: 
801: \clearpage
802: \begin{figure}
803: \begin{center}
804: \vspace{3mm}
805: \includegraphics[height=5.2cm,angle=0]{fig6a.ps}
806: \includegraphics[height=5.2cm,angle=0]{fig6b.ps}
807: \includegraphics[height=5.2cm,angle=0]{fig6c.ps}
808: \includegraphics[height=5.2cm,angle=0]{fig6d.ps}
809: \end{center}
810: \caption{ {\em Left}: 
811: 0.3-8\,keV  images of 
812: the first ({\em top}) and 
813: second ({\em bottom}) ACIS-S3 observations, 
814:  smoothed using a Gaussian of FWHM 2\arcsec.
815: {\em  Right}: The same images adaptively smoothed to show the structures 
816: with the signal-to-noise ratio in the range 2.2 to 4.
817: The intensity scale is linear with a range of 0 to 0.2 cts/pixel 
818: in all panels.
819:  \label{fig-individual-conv-smooth-image}}
820: \end{figure} 
821: \clearpage
822: 
823: 
824: 
825: %%%%%%%%%%%%%
826: \subsection{Connection with the {\sl XMM-Newton} data}
827: \label{xmm}
828: 
829: 
830: To examine the connection between the
831: {\sl Chandra} and {\sl XMM-Newton} data on the PWN, we produced
832:  combined
833: MOS1 and MOS2\footnote{The PN observations were taken in the small-window mode and have high background.}
834:   broad-band images for each of the three available
835:  {\sl XMM-Newton}
836:  observations of the B1929 field
837:  listed in Table~\ref{table-observations}
838:  \citep[see][for more detail about the {\sl XMM-Newton} data
839: and for the combined image of all three observations]{2006ApJ...645.1421B}.
840:  
841: 
842:  The smoothed images are shown in Figure~\ref{fig-xmm},
843:  overlayed with the
844: structures detected in the
845:  ACIS-S observations. 
846:  Since the
847:  MOS PSF is much broader than that of ACIS,
848: the compact PWN (regions 1--3 in Figs.~\ref{fig-unsmoothed-images} and \ref{fig-combined-conv-smooth-image}) and the beginning of the
849: PWN tail (region 4) cannot be resolved from the pulsar.
850: However, at least in the first two {\sl XMM-Newton} observations
851:  we clearly see the local brightening in the PWN tail $\approx 2'$
852:  from the pulsar, approximately coinciding 
853: with the Outer Blob (region 5) detected
854:  in the {\sl Chandra} data.
855: From  the three unsmoothed MOS1 images, we calculated an average background-subtracted count rate of
856: $1.5 \pm 0.3$\,counts\,ks$^{-1}$ in the Outer Blob.
857: With account for different ACIS-S and MOS1 responses,
858: this count rate is consistent
859: with
860: that measured in the combined {\sl Chandra} data in the same region
861: (see Table~\ref{table-regions}).
862: 
863: 
864: \clearpage
865: \begin{figure*}
866: \begin{center}
867: \includegraphics[height=5cm,angle=0]{fig7a.ps}
868: \includegraphics[height=5cm,angle=0]{fig7b.ps}
869: \includegraphics[height=5cm,angle=0]{fig7c.ps}
870: \end{center}
871: \caption{Combined MOS1+MOS2 images of B1929 and its surroundings in the three {\sl XMM-Newton} observations.
872: Broad-band (0.3--10\,keV) images were smoothed using a Gaussian of FWMH $6''$. 
873:  The five regions of enhanced emission and possible pointlike
874:  sources X, Y and Z detected in the  ACIS-S observations are
875: shown (see
876: left panel of
877: Fig.\ \ref{fig-combined-conv-smooth-image}
878: and Table~\ref{table-regions}).
879: The bright source (Var) $\sim2.5'$
880:  south from the pulsar
881: exhibits strong variability, also clearly visible in the unsmoothed MOS
882:  images.
883: The intensity scale is linear, with a range of 0 to 6 counts per pixel in all panels.
884: \label{fig-xmm}}
885: \end{figure*}
886: \clearpage
887: 
888: Since our {\sl Chandra} data  indicate  possible variability of
889:  the  PWN, we examined the {\sl XMM-Newton} data
890: for brightness variations in the extended
891:  emission.
892: From the number of the background-subtracted counts in the Outer Blob,
893: we determined the averaged MOS1 count rates of
894: $1.7 \pm 0.5$, $1.6 \pm 0.5$, and $1.3 \pm 0.5$ counts ks$^{-1}$
895: in observations 0113051301, 0113051401 and 0113051501, respectively.
896: Thus, although the Outer Blob changes its surface brightness by a factor of 2 
897: at a $2.5 \sigma$ level in our
898: ACIS-S data, the surface brightness
899: of the same region in the MOS1
900: data,
901:  separated  by approximately the same period of $\sim$5 months as the two
902:  {\sl Chandra} pointings, does not show a significant brightness change.
903: 
904: 
905: Figure~\ref{fig-xmm} suggests that
906:  the  region
907: southwest of the Outer Blob
908: might be variable. 
909: However, in the ACIS image we see no
910: diffuse emission in this area, 
911: but there are two faint pointlike sources, X and Y
912: (see Figs.\ \ref{fig-combined-conv-smooth-image} and \ref{fig-xmm}).
913: Hence, the apparent variability of this region could be due to the variability 
914: of these two 
915: sources, which could not be resolved in the {\sl XMM-Newton} observations.
916: Thus, although the  {\sl Chandra}  observations
917: hint that the tail of B1929 might be variable on
918:   time-scales of several months,    
919:  longer high-reslolution observations are required
920: to firmly establish this variability.
921: 
922: 
923: 
924: Interestingly, the {\sl XMM-Newton} images show
925: a highly variable
926: source  (Var) $\sim2.5\arcmin$
927: south of the pulsar.
928: It was almost as bright as the pulsar in the 2004 April 27 observation,
929:  while it is
930: barely visible in the 2003 November 10 image.
931: The  source was at the edge of the S3 chip in the first {\sl Chandra}
932:  observation and barely detectable. However, it completely disappeared in the second observation, although its position was within the field of view.  
933: As it is apparently a point source, strongly detached from the other
934: PWN structures, we conclude that it is unrelated to the B1929 PWN.
935: 
936: 
937: %%%%%%%
938: 
939: 
940: 
941: 
942: 
943: \clearpage
944: \begin{figure}
945: \begin{center}
946: \includegraphics[height=8.5cm,angle=270]{fig8a.ps}
947: \includegraphics[height=8.5cm,angle=270]{fig8b.ps}
948: \includegraphics[height=8.5cm,angle=270]{fig8c.ps}
949: \end{center}
950: \caption{{\em Top}: Spectra of B1929 extracted from the 
951: {\sl Chandra} observations 
952: 6657 (black) and 7230 (red)
953: and the best-fit absorbed PL model. 
954: The model parameters are given in Table~\ref{table-spectrum}. 
955: {\em Middle}: PN, MOS1 and MOS2 spectra of B1929 extracted from the 
956: three {\sl XMM-Newton} 
957: observations (PN: black, red, green; MOS1: blue, cyan, magenta; MOS2: yellow, orange, lime-green; in the observations  0113051301, 0113051401 and 0113051501, respectively),
958: and fitted simultaneously with the {\sl Chandra}
959:  data (Obs.\ 6657: dark-green; Obs.\ 7230: dark-blue)  using the  
960: absorbed PL+BB model. The model parameters are given in
961: Table~\ref{table-spectrum}.
962: {\em Bottom:} Photon spectrum for the best-fit 
963: absorbed BB+PL model and its components.
964: The unabsorbed luminosity of the BB component,
965:   $\approx$ $1.2 \times 10^{30}$ ergs s$^{-1}$,  is estimated to be
966:  $\sim$45\% of the total luminosity,
967:  $\approx$ $2.6  \times 10^{30}$ ergs s$^{-1}$,
968:  in the 0.3--8\,keV band.
969: \label{fig-combined-spectra-pl}}
970: \end{figure}    
971: \clearpage
972: 
973: 
974: 
975: \clearpage
976: \begin{figure}
977: \begin{center}
978: \includegraphics[height=8cm,angle=90]{fig9a.ps}\\
979: \includegraphics[height=8cm,angle=90]{fig9b.ps}
980: \end{center}
981: \caption{{\em Top:} 68\%, 90\%, and 99\% confidence contours
982: for the {\sl Chandra} ACIS and  {\sl XMM-Newton} 
983: EPIC spectra of B1929 (dashed and solid lines, respectively), 
984: computed for two interesting parameters for the PL model.
985:  {\em Bottom:} 68\%, 90\%, and 99\% confidence contours for the
986: {\sl Chandra} and {\sl XMM-Newton} observations fitted simultaneously with
987:  the PL model.
988: \label{fig-individual-contours-pl} }
989: \end{figure}   
990: \clearpage
991: 
992: 
993: 
994: \clearpage
995: \begin{figure}
996: \begin{center}
997: \includegraphics[height=8cm,angle=90]{fig10a.ps}\\
998: \includegraphics[height=8cm,angle=90]{fig10b.ps}\\
999: \includegraphics[height=8cm,angle=90]{fig10c.ps}
1000: \end{center}
1001: \caption{68\%, 90\%, and 99\% confidence contours for various 
1002: parameter pairs for the 
1003: PL+BB fit to the spectrum of B1929. 
1004:  The confidence levels correspond to two interesting parameters. 
1005: The {\sl Chandra} ACIS-S and  {\sl XMM-Newton} spectra are fitted
1006:  simultaneously 
1007: with all parameters 
1008: allowed to vary. The top panel shows the lines of constant bolometric
1009:  luminosity
1010: of an equivalent sphere (see \S\,3.2) in units of 
1011: $10^{30}$\,ergs\,s$^{-1}$ for the BB component, 
1012: while the middle panel shows the lines of constant unabsorbed flux 
1013: in units of $10^{-13}$\,ergs\,cm$^{-2}$\,s$^{-1}$ for the PL
1014: component. 
1015: \label{cont-T-rad} }
1016: \end{figure} 
1017: \clearpage  
1018: 
1019: \subsection{Spectral analysis}
1020: \label{spectral-analysis}
1021: 
1022: \subsubsection{The pulsar}
1023: \label{spectral-analysis-pulsar}
1024: 
1025: 
1026: 
1027: The spectral {\sl Chandra} data reduction was done by applying the standard
1028:  procedures
1029:  in CIAO.
1030: From a circular region with a radius
1031: of $1.6''$ (or 3.3 pixels) centered on the pulsar, 
1032: we extracted 645 and 690 counts 
1033: from the first and second observation, respectively.
1034: Four circular background regions, each with a radius of 5$\arcsec$\,,
1035:   were selected  in the vicinity of
1036:  the pulsar and then combined
1037: and used for the background subtraction. 
1038: The background contribution in the source aperture is negligibly small
1039: ($\sim 1$ count in each of the observations), but the aperture
1040: may include a small number of 
1041: photons emerging from the compact
1042:  nebula. 
1043: Using the radial profiles in
1044:  Figure~\ref{fig-radial-profiles},
1045:  we estimated the PWN contribution to be around 2\%--3\%;
1046: hence it does not influence the results of the spectral fitting
1047: significantly.
1048:  We grouped the extracted counts in a minimum of 15 and 17 counts per energy
1049:  bin for the spectral analysis.
1050: The response matrices (rmf) were calculated
1051: using the CIAO task {\sc mkacisrmf}, and the spectral analysis was
1052: performed using XSPEC (ver.\ 11.3.2). 
1053: 
1054: 
1055: Since we did not find statistically significant changes between the 
1056: two {\sl Chandra} observations, 
1057: we fit both data sets simultaneously.
1058: The absorbed power-law (PL) model yielded 
1059: a very good
1060:  fit  (Table~\ref{table-spectrum};
1061:  Fig.\ \ref{fig-combined-spectra-pl}, top),
1062: albeit with surprisingly large best-fit hydrogen column density,
1063: $N_{\rm H}\approx2.5\times 10^{21}$ cm$^{-2}$, and photon index, 
1064: $\Gamma\approx2.9$.
1065: On the contrary, we were not able to produce a statistically
1066: acceptable fit with the absorbed blackbody (BB) model  
1067: as the discrepancy between the best-fit model and the data was large
1068: below 0.7 keV and particularly above 2 keV, resulting
1069: in $\chi_\nu^2 =1.86$ (vs.\ $\chi_\nu^2 =0.91$ for the PL fit)
1070: for 71 d.o.f.
1071: 
1072: 
1073: 
1074: The pulsar models predict that a 
1075:  thermal component from heated polar caps, and perhaps
1076: from the whole NS surface, may be detectable,  making the main contribution
1077:  at lower photon energies.  
1078:  Therefore, we attempted to
1079: fit the {\sl Chandra} spectra with the PL+BB model. 
1080: The addition of the BB component, with the best-fit temperature
1081: $kT\approx 0.3$ keV and emitting area $A\sim 4\times 10^3$ m$^2$,
1082: resulted in a substantially smaller hydrogen column density,
1083:  $N_{\rm H}<1.4\times 10^{21}$ 
1084: cm$^{-2}$, and a softer PL spectrum, $\Gamma\approx 1.8$
1085: (see Table~\ref{table-spectrum}),
1086: but the probability that the BB component 
1087: is required by the data was only 69.5\%, according to the F-test.
1088: We also tested the  BB+BB model, but it yielded either  
1089: statistically  unacceptable fits or  unphysical fitting parameters
1090: (e.g., $kT\sim 100$\,keV, or an extremely low absorption column).
1091: 
1092: To compare  the {\sl Chandra} ACIS spectra with those
1093: obtained with the
1094: {\sl XMM-Newton} EPIC detector, we
1095: extracted and fitted the pulsar spectra
1096: from the three {\sl XMM-Newton}
1097: observations. The source counts were extracted from a circular region with a 
1098: radius of 13$''$ centered on the pulsar, 
1099: while a 38$''$-radius circular region northwest
1100: of the pulsar was used for the background subtraction.
1101:  We collected a total of 450 (with 96.8\% of the source
1102: contribution), 658 (94.8\%) and 805 (75.3\%) counts in the PN 
1103: (small-window mode)
1104: observations 0113051301, 0113051401, and 0113051501, respectively.
1105:  In addition, we extracted
1106: 172 (98.0\%), 282 (97.3\%), and 110 counts from the MOS1, and 153 (97.3\%), 276
1107: (98.0\%), and
1108:  121 (96.8\%) from the MOS2 observations, with MOS cameras
1109: operated in the large window mode. The EPIC response matrices and 
1110: ancillary files
1111:  were calculated using the SAS  tasks {\sc rmfgen} and
1112:  {\sc arfgen}.
1113: From the number of the PWN counts detected with ACIS in an aperture of $13''$,
1114: we estimate that the contribution from the nebula in the EPIC spectra is 
1115: about 5\%. 
1116: 
1117: 
1118: Fitting all the 9 {\sl XMM-Newton} data sets (PN, MOS1 and MOS2 in the three 
1119: observations) simultaneously, 
1120: we found that the PL model produced a statistically
1121:  acceptable fit, with the model parameters consistent with those obtained 
1122: from the {\sl Chandra} data 
1123: (Fig.~\ref{fig-individual-contours-pl})\footnote{We should mention that 
1124: our PL fitting parameters differ significantly from
1125: those obtained by Becker et al.\ (2006) from fitting the
1126: combined {\sl XMM-Newton} EPIC
1127: and {\sl ROSAT}\, PSPC data (they do not provide model parameters from fitting
1128: the {\sl XMM-Newton} data separately). 
1129: For instance, these authors obtained a lower hydrogen column density,
1130:  $N_{\rm H}=(1.6\pm 0.2)\times 10^{21}$ cm$^{-2}$, and 
1131: a smaller photon index, $\Gamma = 2.72^{+0.12}_{-0.09}$. We believe that
1132: the discrepancy is caused by systematic errors due to poor cross-calibration
1133: of the PSPC and EPIC detectors.}. 
1134: Adding the BB component to the spectral model resulted in about the same
1135: fitting parameters as for the {\sl Chandra} data alone
1136: (Table~\ref{table-spectrum}), but improved the fit
1137: more significantly (94.8\%, according to the F-test).
1138: 
1139:  
1140: 
1141: 
1142: Finally, we fit the {\sl Chandra} and {\sl XMM-Newton} data simultaneously,
1143: first with the PL model.
1144: Because of the 
1145: larger PWN contribution 
1146: in the {\sl XMM-Newton} data,
1147: we first linked  all the model 
1148: parameters except for
1149: the PL normalization. However, since we found that the difference between the
1150: PL normalizations 
1151: in the {\sl Chandra} and
1152: {\sl XMM-Newton} spectra was within the statistical errors,
1153: we finally linked all the model parameters.
1154: As expected, we obtained the PL parameters similar to those obtained from
1155: the separate fits to the {\sl Chandra} and {\sl XMM-Newton} data, but
1156: with smaller uncertainties (Fig.~\ref{fig-individual-contours-pl}, bottom).
1157: 
1158: 
1159: 
1160: 
1161: 
1162: Fitting the combined data with the PL+BB model, we found that the
1163: addition of the BB component provided a very significant
1164: improvement of the fit (99.98\%, according to the F-test).
1165: The best PL+BB fit 
1166: is shown in Figure~\ref{fig-combined-spectra-pl},
1167: the fitting parameters are given in Table~\ref{table-spectrum},
1168: and
1169: the confindence contours for various model parameters are presented in 
1170: Figure~\ref{cont-T-rad}.  We note that the hydrogen column density,
1171:  $N_{\rm H}<4\times 10^{20}\,{\rm cm}^{-2}$, 
1172: is significantly lower in the PL+BB fit and consistent with the value 
1173: estimated from the pulsar's dispersion measure, 
1174: while the slope of the PL components, $\Gamma\approx 1.7$, is similar to
1175: those found in many other pulsars
1176: (see \S\,\ref{spectral-discussion} for further discussion). 
1177: 
1178: 
1179: 
1180: 
1181: 
1182: 
1183: \clearpage
1184: \begin{figure}
1185: \begin{center}
1186: \includegraphics[height=8cm,angle=90]{fig11a.ps}\\
1187: \includegraphics[height=8cm,angle=90]{fig11b.ps}
1188: \end{center}
1189: \caption{{\em Top:} 
1190:  68\%, 90\%, and 99\% confidence contours in 
1191: the $N_{\rm H}-\Gamma$ plane
1192:  of the  PL model  for the PWN spectral data 
1193: (the elliptical region in Fig.~\ref{fig-combined-conv-smooth-image}).
1194: {\em Bottom}: 68\%, 90\%, and 99\% confidence contours in 
1195: the 
1196:  $\Gamma$--Normalization plane
1197:  of the PL model with the absorption
1198: column fixed to the best-fit value obtained from the spectrum of the pulsar.
1199: \label{cont-PWN}} 
1200: \end{figure}   
1201: \clearpage
1202: 
1203: \subsubsection{The nebula}
1204: \label{spectrum-nebula} 
1205: 
1206: As the detected PWN is very faint, we have to choose a sufficiently large extraction region
1207: to measure its spectrum.
1208: The large-scale PWN
1209:  appears to consist of  
1210: a smaller, slightly brighter part up to 30$''$ behind the pulsar (marked by
1211: the ellipse in the middle panel of 
1212: Figure~\ref{fig-combined-conv-smooth-image}), 
1213: and a fainter emission 
1214: farther down, up to a distance of
1215:  $\sim$2$'$.
1216:  From the brighter part 
1217: of the nebula region (the ellipse) we collected 138
1218: and 136 counts (64\% and 63\% of which are from the nebula)
1219:  in the first and second
1220:  observation, respectively. 
1221: 
1222: 
1223: To constrain the model parameters, we fit  the spectra from the two observations simultaneously, 
1224: with the absorbed PL model. 
1225: The fit yielded a PL slope 
1226: $\Gamma$=1.0--1.7  at a 90\% confidence level (Fig.\ \ref{cont-PWN}, top) 
1227: and an absorption column 
1228:  $N_{\rm H} = 0.0-0.9 \times  10^{21}$ cm$^{-2}$ (90\% confidence).
1229: We also fitted the PWN spectra with the absorption column fixed to the
1230: best-fit value of $N_{\rm H} = 1.7 \times  10^{21}$ cm$^{-2}$ obtained from
1231: the spectrum of the pulsar, but the photon index range remains the same
1232: (Fig.\ \ref{cont-PWN}, bottom).
1233: We also attempted to fit the spectra from the PWN region farther down the
1234: tail, to
1235: see if there is any indication of spectral softening, but the number of
1236: photons from this faint part of the nebula is too small for a miningfull
1237: spectral analysis.   
1238: 
1239: The PWN flux, $F_{\rm PWN}=2.1$--$3.6\times10^{-14}$,
1240:  determined from the 
1241: elliptical region, is poorly constrained because of the low statistics.
1242: We also measured a  flux of 
1243:  $3.4-5.8\times10^{-14}$ ergs~cm$^{-2}$~s$^{-1}$ from the whole visible PWN.
1244: The unabsorbed luminosity of the whole nebula estimated from this 
1245: observed flux 
1246: is 
1247: $L_{\rm PWN} = 5.3$--$8.6\times 10^{29}$\,ergs\,s$^{-1}$, in the 0.3$-$8\,keV band.
1248: 
1249: 
1250: 
1251: 
1252: The absorption column determined from the spectral fit of the nebula is
1253:  significantly lower
1254: than that determined from the 
1255: PL fit of the pulsar spectrum, but it is in a good agreement with the value
1256: derived from the PL+BB model. This further supports our detection of the thermal component in the spectrum of B1929.
1257: 
1258:  
1259: 
1260: 
1261: 
1262:  
1263: 
1264: 
1265: \section{Discussion and conclusions}
1266: \label{discussion}
1267: 
1268: 
1269: \subsection{The nebula around PSR\,B1929+10}
1270: 
1271: \subsubsection{PWN morphology}
1272: 
1273: 
1274: In the analysis of the two ACIS-S observations of B1929, we detected a faint 
1275: PWN surrounding this old pulsar.
1276: We found that the pulsar is
1277:  immersed in a very compact, $\approx 9''\times 5''$ 
1278: nebula
1279: elongated in the
1280: direction perpendicular to the pulsar's proper motion
1281: (Fig.\ 1, bottom). 
1282: Together with the two patchy ``wings'' (best seen in slightly smoothed
1283: images), the overall compact structure 
1284: looks like a bow shock PWN, with an apex at $\approx 3''$
1285: ($\approx 1.6\times 10^{16}$ cm at $d=361$ pc)
1286:  ahead of 
1287: the pulsar. 
1288: In addition, there is marginal evidence of
1289: a short jet-like feature emerging 
1290: from the pulsar between the wings, in the direction opposite to 
1291: the pulsar's proper motion  
1292: (seen up to $3''$--$4''$ from the pulsar in the unsmoothed combined image 
1293: and in the pulsar-subtracted image in Fig.\ \ref{fig-unsmoothed-images}).
1294:  There is also some very
1295: faint emission  in front of the 
1296: apparent bow-shock apex, seen up to $6''$--$7''$ from the pulsar
1297: (see Fig.\ \ref{fig-profile}). 
1298: 
1299: 
1300: In addition to the compact PWN, 
1301: a faint, inhomogeneous
1302:  tail-like structure was detected, extending
1303:  in the direction approximately opposite to the pulsar's proper motion up to 
1304: at least $\sim1.5'$--2$\arcmin$. This  
1305: structure is well aligned with 
1306: the much longer tail (up to 10\arcmin--15\arcmin)
1307: detected in the {\sl ROSAT} and {\sl XMM-Newton} data 
1308: \citep{1993Natur.364..127W,2006ApJ...645.1421B}.
1309: 
1310: 
1311: The surface brightness of the detected 
1312: PWN is not uniform: e.g., the two wings are
1313: brighter than the other PWN structures
1314: (see Table~\ref{table-regions}). 
1315: Furthermore, the {\sl Chandra\/} and {\sl XMM-Newton\/} data suggest 
1316: a possible variability of the detected PWN, although
1317: the number of detected photons was too small to establish it unambiguously.  
1318: 
1319: 
1320: 
1321: The approximate axial symmetry with respect to the direction
1322: of the pulsar's proper motion strongly suggests that the PWN 
1323: is shaped by the oncoming flow of the ambient matter in the pulsar's
1324: reference frame. 
1325: The corrected pulsar's proper motion and distance correspond to
1326: the pulsar's speed $v_{\rm PSR} = 178\, (\sin i)^{-1}$
1327: km s$^{-1}$, where $i$ is the 
1328: angle between the velocity
1329: vector and the line of sight.
1330: It exceeds the ISM sound speed,
1331: $c_s = (5kT/3\mu m_H)^{1/2} = 37 \mu^{-1/2} T_5^{1/2}$ km s$^{-1}$
1332: (where $\mu$ is the molecular weight, $T_5=T/10^5$ K), for
1333: $T< 2.3\times 10^6 \mu\, (\sin i)^{-2}$ K, i.e., for virtually any
1334: plausible ISM temperature, which supports the bow-shock interpretation
1335: of the detected nebular emission. 
1336: 
1337: 
1338: 
1339: Our current understanding of the pulsar wind dynamics in a bow-shock
1340: PWN mostly relies upon numerical models. A recent work in this
1341: direction has been done by \citet{2003A&A...404..939V}, \citet{2005A&A...434..189B}, and \citet{2007MNRAS.374..793V}.
1342: These models show that at supersonic pulsar speeds
1343: the  termination shock (TS) of an {\em isotropic} pulsar wind
1344:  acquires a bullet-like shape,
1345:  with a distance 
1346: \begin{equation}
1347: R_h \simeq
1348: (\dot{E}/4\pi c p_{\rm ram})^{1/2}\, 
1349: \end{equation}
1350:  between the pulsar and the
1351: bullet head, where 
1352: \begin{equation}
1353: p_{\rm ram} = \rho_{\rm amb}\,v_{\rm PSR}^2
1354: \end{equation}
1355:  is
1356: the ram pressure, 
1357: and $\rho_{\rm amb}$ is the ambient density.
1358: %
1359:  The shocked pulsar wind is confined between the TS and
1360: the contact discontinuity
1361:  (CD) surface, while the forward bow
1362: shock (FBS) separates the shocked ambient medium (between the CD and the
1363: FBS) from
1364: the unshocked one.
1365:  For large Mach numbers, $\mathcal{M}\equiv v_{\rm \sc PSR}/c_{s}
1366: =(3p_{\rm ram}/5p_{\rm amb})^{1/2} \gg 1$, and
1367: small values of the magnetization parameter\footnote{The magnetization
1368: parameter in the pre-shock wind is defined  
1369: as the ratio of the Poynting flux to the kinetic energy flux.} 
1370: of the
1371: pre-shock pulsar wind, $\sigma < 0.1$,  the bullet's
1372: cylindrical radius is $r_{\rm TS}\sim R_{h}$ 
1373:  and the distance of
1374: its back surface from the pulsar is $R_{b}\sim 6R_{h}$ 
1375: \citep[see Figs.\ 1--3 in][]{2005A&A...434..189B}.
1376: The shape of the CD surface ahead of the pulsar is similar to that of the
1377: TS but with the apex
1378: at $R_{\rm CD}\approx 1.3 R_h$, while  it acquires
1379: a cylindrical shape with a radius $r_{\rm CD}\approx 4 R_h$ behind the
1380:  TS bullet.
1381: The shocked wind flows away from the pulsar within this cylinder,
1382:  forming a PWN tail. The flow
1383: velocity in the central part of the tail (the inner channel, 
1384: $r\lesssim r_{\rm TS}$) is $\sim 0.1c$--$0.2c$,
1385: while it is as high as $v_{\rm tail}=0.8c$--$0.9c$ in the bulk
1386:  of the tail's volume,
1387: at $r_{\rm TS} \lesssim r < r_{\rm CD}$ (where $r$ is the cylindrical
1388: radius). The magnetic field, purely
1389: toroidal in these models, also depends on $r$, being enhanced toward
1390: the tail axis and the CD surface. Its typical value in the tail is
1391: \begin{equation}
1392: B_{\rm tail}\sim 2.8\, (c/v_{\rm tail})\, (\sigma p_{\rm ram})^{1/2} {\rm G},
1393: \end{equation} 
1394: where the $p_{\rm ram}$ is in units of dyn cm$^{-2}$
1395: \citep[see eq.\ (13) of][]{2005A&A...434..189B}.
1396: %
1397:  In the numerical simulations, the flow velocity and the magnetic field
1398: do not show substantial changes along the tail,
1399: but these simulations are limited to relatively short
1400: distances from the pulsar 
1401: \citep[e.g., $<13 R_h$ in][]{2005A&A...434..189B}.
1402: 
1403: 
1404: 
1405: 
1406: 
1407: To interpret the B1929 PWN 
1408: in the framework of 
1409: the current MHD models, we attempted to match the observed
1410:  morphological structures with those
1411: predicted by the simulations. 
1412: As the brightest X-ray emission is expected from the shocked
1413: wind between the TS and CD in the front part of the PWN,
1414: we assume that the very compact structure in the immediate vicinity
1415: of the pulsar (Fig.~\ref{fig-unsmoothed-images}, bottom)
1416:  is the PWN head and estimate the distance from the pulsar
1417: to the CD apex to be $\sim 3''$.
1418: This corresponds to
1419: $R_{\rm CD}\sim 1.6\times 10^{16}$ cm,
1420: $R_h \sim r_{\rm TS}\sim 1.2\times 10^{16}$ cm (or $2.3''$),
1421: $R_b\sim 7.5\times 10^{16}$ cm (or $14''$), 
1422: and $r_{\rm CD}\sim 4.8\times 10^{16}$ cm (or $9''$). 
1423: %
1424: In this picture, the patchy wings in the PWN image are produced by
1425: the shocked wind between the TS and CD,  while the ``blobs'' farther
1426: behind the pulsar are in the pulsar tail.
1427: However, the faint emission ahead
1428: of the pulsar (region 3, or the Front, in Fig.~\ref{fig-unsmoothed-images}, top, and Table~\ref{table-regions})
1429:  is not explained in this model. We can only speculate that this emission might
1430: be produced by a polar outflow along the pulsar's spin axis
1431: (not included in the models), possibly scattered by the head wind,
1432:  or it might suggest that, because of the 
1433: Rayleigh-Taylor instability, the CD surface
1434: acquires an irregular shape ahead of the pulsar, with ``fingers'' of the
1435: shocked wind penetrating into the shocked ISM.
1436: %
1437: As the above estimates of sizes of the PWN elements are based on simplified
1438: models and shallow images, one should consider them as order-of-magnitude
1439: estimates rather than accurately established values. Therefore,
1440: in the following estimates, we will
1441: retain the explicit dependence on $R_h$, scaling it as 
1442: $R_h=R_{h,16}\times 10^{16}$ cm.
1443: 
1444: 
1445: Using this estimate for $R_h$ and equation (1), we can calculate the 
1446: ram pressure:
1447: \begin{equation}
1448: p_{\rm ram}=
1449: \dot{E}f_\Omega (4\pi cR_h^2)^{-1} = 
1450: 1.0\times 10^{-10} 
1451: f_\Omega R_{h,16}^{-2}\,\,
1452: {\rm dyn}\,\,{\rm cm}^{-2},
1453: \end{equation}
1454: where 
1455: we introduced the factor $f_\Omega$ 
1456: to account for possible anisotropy of the pulsar outflow
1457: (e.g., $f_\Omega <1$ for a mostly equatorial outflow if the
1458: equatorial plane is perpendicular
1459: to the pulsar's velocity). Since, on the other hand,
1460: \begin{equation}
1461: p_{\rm ram}=\rho_{\rm amb}V_{\rm PSR}^2 =5.3\times 10^{-10} n_{\rm amb}\, (\sin i)^{-2}\,\,{\rm dyn}\, {\rm cm}^{-2}\,,
1462: \end{equation}
1463: where $n_{\rm amb}\equiv \rho_{\rm amb}/m_H$, we obtain an estimate
1464: for the ambient number density: 
1465: \begin{equation}
1466: n_{\rm amb}= 0.19 f_\Omega R_{h,16}^{-2} \sin^2i\,\,\,  {\rm cm}^{-3}\,.
1467: \end{equation}
1468: %
1469: 
1470: 
1471: 
1472: Using equation (4), we can estimate the Mach number:
1473: \begin{equation}
1474: \mathcal{M}= 
1475: 7.7 f_\Omega^{1/2}R_{h,16}^{-1}p_{\rm amb,-12}^{-1/2}\,,
1476: \end{equation}
1477: where we scaled $p_{\rm amb}$ to $10^{-12}$ dyn cm$^{-2}$, a typical value
1478: for the ISM.
1479: From the corresponding sound speed, 
1480: $c_s = 23 f_\Omega^{-1/2} R_{h,16} p_{\rm amb,-12}^{1/2} (\sin i)^{-1}$
1481: km s$^{-1}$, we estimate the temperature:
1482: \begin{equation}
1483: T=3.9\times 10^4 \mu f_\Omega^{-1} R_{h,16}^2 p_{\rm amb,-12}(\sin i)^{-2}\, {\rm K}\,.
1484: \end{equation}
1485: The ambient pressure in the unperturbed ISM is in the range
1486: $p_{\rm amb,-12}\approx 0.2$--2 
1487: \citep{2001ApJ...551L.105H,2001RvMP...73.1031F}.
1488: The pressure in the pulsar's neighborhood might be somewhat higher if
1489: the pulsar's UV and soft X-ray emission heats and ionizes the surrounding
1490: medium, which is supported by the lack of an H$_\alpha$
1491: bow-shock PWN in the observations reported by \citet{2006ApJ...645.1421B}
1492: \citep[see][for a detailed discussion on conditions
1493: of observability of H$_\alpha$ bow-shock PWNe]{2001A&A...375.1032B}.
1494: As the factor $\mu f_\Omega^{-1} R_{h,16}^2 (\sin i)^{-2}$ is also very likely
1495: greater than unity, we expect $T\sim 10^5$ K to be a realistic estimate 
1496: for the ambient temperature.
1497: 
1498: 
1499: We should note that although the comparison of our data with the numerical
1500: models leads to reasonable estimates for the parameters of the ambient
1501: medium, these two-dimensional
1502:  models are based on a number of assumptions that are not
1503: necessarily realistic.  
1504: In particular, the models assume an isotropic
1505: pulsar outflow, whereas we know that it is predominantly equatorial, at
1506: least in the case of young pulsars.
1507: Obviously, a substantial anisotropy would distort the shape of the 
1508: bow shock in the pulsar vicinity (e.g., decreasing $R_h$ in the case
1509: of an equatorial outflow in the plane perpendicular to the pulsar
1510: velocity).
1511: For B1929, a hint of an equatorial outflow might be suggested by
1512: the elongation of the most compact PWN structure 
1513: perpendicular to the proper motion (Fig.~\ref{fig-unsmoothed-images}, bottom).
1514:  Such a
1515: structure, together with the ``wings'',
1516:  might be interpreted as a torus of the shocked equatorial wind,
1517: compressed and possibly bent by the oncoming wind.
1518: Such an interpretation might also be supported by the marginal detection 
1519: of the jet-like structure immediately behind the pulsar and the faint
1520: emission ahead of the pulsar (a crushed counter-jet?).
1521: 
1522: 
1523: 
1524: 
1525: 
1526: 
1527: 
1528: 
1529: Application of the PWN simulations to the observed tail
1530: leads to some discrepancies. For instance, although the tail 
1531: may look like 
1532: a cylinder with the predicted diameter of $\sim 18''$ at the relatively
1533: small distances, $\lesssim 2'$, from the pulsar in the ACIS images, 
1534: its transverse
1535: size apparently increases 
1536: at larger distances, possibly reaching
1537: $\sim 5'$ at the edge of the MOS field of view ($15'$ from the pulsar).
1538: The tail's surface brightness looks very patchy, which might suggest
1539: some interaction of the shocked wind with the surrounding medium,
1540: such as the Kelvin-Helmholtz instability caused by the shear at the CD
1541: surface. This instability
1542: may produce clumps of shocked ISM embedded in the
1543: shocked wind, which would distort the flow pattern and compress the
1544: wind material \citep{2005A&A...434..189B}. 
1545: The same instability may result in temporal variations in the X-ray
1546: emission of the tail, with a time scale of 
1547: $r_{\rm CD}/c \sim 0.5$ months, which might explain the alleged
1548: variability of the PWN (\S~\ref{variability}).
1549: 
1550: 
1551: Using equations (3) and (4), the magnetic field in the tail can be estimated
1552: as
1553: \begin{equation}
1554: B_{\rm tail} = 
1555: 56 \sigma^{1/2} (0.5c/v_{\rm tail}) f_\Omega^{1/2}
1556: R_{h,16}^{-1}\,\,\,\mu{\rm G}\,.
1557: \end{equation}
1558: We should note that this field becomes rather small, comparable to the interstellar magnetic field, at the  usually assumed small values of
1559: the magnetization parameter, $\sigma\lesssim 10^{-2}$.
1560: 
1561: 
1562: 
1563: Using this estimate of $B_{\rm tail}$, we can estimate the synchrotron
1564: cooling time for relativistic electrons
1565: in the tail, $\tau_{\rm syn}=5.1\times 10^{10} \gamma_8^{-1}B_{-5}^{-2}$ s,
1566: where $\gamma=10^8 \gamma_8$ is the electron Lorentz factor
1567: and $B_{-5}=B/10^{-5}\,{\rm G}$. As the characteristic energy of synchrotron
1568: photons is
1569: \begin{equation}
1570: E \sim 0.5 B_{-5} \gamma_8^2\,\,\, {\rm keV},
1571: \end{equation}
1572: photons with energy $E$
1573: are emitted by electrons with $\gamma_8\sim 2 (E/1\,{\rm keV})^{1/2}
1574: B_{-5}^{-1/2}$, 
1575: and the synchrotron cooling
1576: time corresponding to $E=8$ keV photons (upper energy of our band) is
1577: \begin{equation}
1578: \tau_{\rm syn}= 6.8\times 10^8 (\sigma f_\Omega)^{-3/4} 
1579: (v_{\rm flow}/0.5c)^{3/2}
1580: (E/8\,{\rm keV})^{-1/2} 
1581: R_{h,16}^{3/2}\,\, {\rm s}.
1582: \end{equation}
1583: It follows from this equation that the projected tail length,
1584: $l_{\rm tail}\sim v_{\rm flow} \tau_{\rm syn}\,\sin i$, as observed
1585: at energy $E$, can be estimated as
1586: \begin{equation}
1587: l_{\rm tail} \sim 
1588: 1\times 10^{19}
1589: \left(\frac{v_{\rm flow}}{0.5c}\right)^{5/2} 
1590: \left(\frac{8\,{\rm keV}}{E}\right)^{1/2} 
1591: \left(\frac{R_{h,16}^2}{\sigma f_\Omega}\right)^{3/4}\sin i\,\,\, {\rm cm}.
1592: \end{equation}
1593: For instance, for $\sigma=0.01$, $R_{h,16}f_\Omega^{-1/2}=2$, 
1594: $v_{\rm flow}=0.5c$, $\sin i=0.5$, and $E<8$ keV
1595: we obtain $l_{\rm tail}\gtrsim 150$ pc (i.e., $\sim 24^\circ$), 
1596: two orders of magnitude larger than observed by
1597: {\sl XMM-Newton} and {\sl ROSAT}. The main reason for this discrepancy
1598: is the assumption that the flow speed remains very high along the 
1599: entire tail in the
1600: ideal MHD model. \citet{2005A&A...434..189B}
1601: have noticed that a similar discrepancy arises
1602: when the same model is applied to the 
1603: X-ray tail of the Mouse PWN: the observed
1604: tail length is much shorter than that estimated from the model at
1605: any $\sigma <1$. These authors suggest that the 
1606: flow can be slowed down by the interaction with the ambient
1607: medium,
1608: which could also explain the 
1609: divergence 
1610:  of the Mouse tail in the radio.
1611: As these factors have not been included in the current models,
1612: we can only empirically estimate an average flow velocity that would
1613: be consistent with the observed length of 1.5 pc:
1614: $v_{\rm flow} \sim 0.1 c\, \sigma_{-2}^{3/10} 
1615: (f_\Omega/R_{h,16}^2)^{3/10} (\sin i)^{-2/5}$,
1616: assuming that equation (3) is still applicable
1617: (here $\sigma_{-2}=\sigma/10^{-2}$).
1618: We note that this velocity is still much larger than the pulsar velocity.
1619: This means that the 
1620: equation 
1621: \begin{equation}
1622: l_{\rm tail}= v_{\rm PSR,\perp} \tau_{\rm syn}
1623: \approx 0.3 (v_{\rm PSR,\perp}/178\,\,{\rm km}\,\,{\rm s}^{-1})
1624:  \gamma_8^{-1} B_{-5}^{-2}\,\,\, {\rm pc},
1625: \end{equation}
1626:  often
1627: used for estimating the magnetic field \citep[e.g.,][]{2006ApJ...645.1421B},
1628: is inapplicable in this case\footnote{Equation (13) implies that the
1629: shocked wind flow acquires the speed of the ambient ISM matter
1630: in the immediate vicinity of the pulsar, so that the observed elongated PWN
1631: is actually a ``trail'' of the decelerated wind left behind the moving pulsar.
1632: Such an assumption strongly contradicts all the models of bow-shock PWNe
1633: \citep[e.g.,][]{2005ApJ...630.1020R,2005A&A...434..189B}.
1634:  \citet{2006ApJ...645.1421B} obtained
1635: a reasonable estimate for the magnetic field, $B\lesssim 12$ $\mu$G,
1636:  using eq.\ (13) because they assumed $\gamma=10^6$. However, at such
1637: $B$ and $\gamma$,
1638: the energy of a synchrotron photon is $E=6$ eV (see eq.\ [10]), well
1639: below the X-ray band.}.
1640: 
1641: 
1642: 
1643: It is worthwhile to mention that 
1644: the length of the tail of B1929 measured so far might be limited by the
1645:   field of view of the available {\sl XMM-Newton} observations,
1646:  and further deep observations along the tail are needed to determine its full extent.
1647: Recent observations with {\sl Chandra} and {\sl XMM-Newton} have discovered
1648: a number of other PWNe with very long tails \citep{2008AIPC..983..171K},
1649: including the 6\,pc tail ($l_{\rm tail}/R_{h}\sim 600$) behind 
1650: PSR J1509$-$5850, the longest pulsar tail known to date \citep{2008arXiv0802.2963K}.
1651: Detection of more such objects
1652:  would help to facilitate further modeling
1653:   that would account for cooling in the varying magnetic field  at larger distances from the pulsar.
1654: 
1655: 
1656: \subsubsection{X-ray luminosity and spectrum of the PWN}
1657: 
1658: 
1659: From our  estimate of the total unabsorbed flux of the 
1660: B1929 PWN detected with {\sl Chandra},
1661: we 
1662: estimated the X-ray PWN luminosity, $L_{\rm PWN}=(5.3$--$8.6)\times 10^{29}$
1663: ergs s$^{-1}$, which corresponds to the PWN efficiency
1664:  $\eta_{\rm PWN} \equiv L_{\rm PWN}/\dot E = 
1665: (1.4$--$2.2)\times 10^{-4}$ in the 0.3--8\,keV band.
1666: Similar values of the luminosity and efficiency have been measured
1667: in the {\sl XMM-Newton} data
1668: by \citet{2006ApJ...645.1421B} from a larger area farther
1669: southwest
1670: from the pulsar.
1671: As both measurements were taken for small fractions of the PWN,
1672: the luminosity and efficiency of the entire PWN can be higher than these
1673: estimates.
1674: 
1675: 
1676: The estimated efficiency of the B1929 PWN is within the range
1677: $\eta_{\rm PWN}\sim 10^{-4.5}$--$10^{-2}$ found by \citet{2007ApJ...660.1413K}
1678: in a recent study of several energetic middle-aged Vela-like pulsars and
1679:  their PWNe. 
1680: Compared to a few known PWNe with long tails, which, on average, show higher
1681: efficiencies, $\eta_{\rm PWN}\sim 10^{-3.8}$--$10^{-1.7}$,
1682:  than more compact PWNe \citep[see][]{2008arXiv0802.2963K}, 
1683: the B1929 PWN is among the least efficient ones;
1684: however, we stress that a deeper observation of B1929 is needed to 
1685: measure the efficiency more accurately.
1686: 
1687: The spectral slope of the PWN spectrum is also consistent with the values  
1688: measured for other PWNe \citep[e.g., see][]{2007ApJ...660.1413K,2008AIPC..983..171K}. 
1689: 
1690:  
1691: \subsection{Spectral properties of PSR\,B1929+10}
1692: \label{spectral-discussion}
1693: 
1694: As we mention in \S\ref{intro}, it has been a matter of debate
1695:  whether the X-ray 
1696: emission from B1929 is predominantly magnetospheric or it has a significant
1697: thermal component emitted from the neutron star surface.
1698: Most recently, \citet{2006ApJ...645.1421B} have concluded, 
1699: based on the analysis of
1700: the {\sl XMM-Newton} data, that the spectrum of B1929 is best
1701: described by a PL model (i.e.\  the emission is
1702: predominantly magnetospheric),
1703: while the contribution of the thermal component, modeled as BB radiation,
1704: does not exceed $\sim$7\%. Our analysis of the {\sl Chandra} and
1705:  {\sl XMM-Newton}
1706: spectra has shown, however, that 
1707: adding the BB component to the PL model significantly improves the fit to the
1708: combined {\sl Chandra} plus {\sl XMM-Newton} spectrum, with the 
1709: the best-fit BB component
1710: providing $\sim$45\% of the luminosity in the 0.3--10 keV band.
1711: 
1712: 
1713: In addition to the better fit, there are other serious arguments in favor of
1714:  the
1715: PL+BB model. First of all,  
1716: the hydrogen column density, $N_{\rm H}=(2.23\pm0.27)\times 10^{21}$ cm$^{-2}$,
1717: obtained from the PL fit, looks unreasonably large. In particular,
1718: it is 
1719: much larger than the standard estimate, 
1720: $N_{\rm H}\sim 10 N_e \approx 1\times 10^{20}$ cm$^{-2}$, obtained
1721: from the pulsar's dispersion measure, DM = 3.178 pc cm$^{-3}$, under
1722: the usual assumption of a 10\% ISM ionization (where
1723: $N_e$ is the electron column density). In other words, the PL model
1724: requires a very low ISM ionization, $\sim$0.4\%, (i.e., a very large
1725: ratio $N_{\rm H}/N_e \sim 230$, much larger than for any other
1726: radio pulsar detected in X-rays). Also, taking into account that B1929 is
1727: a nearby pulsar ($d=361$ pc),
1728: the $N_{\rm H}$ obtained from the PL fit
1729: is uncomfortably close to the HI column density, 
1730: $N_{\rm HI}=3.5\times 10^{21}$, throughout
1731: the entire Galaxy in the direction of B1929 \citep{2005A&A...440..775K}.
1732: On the contrary,
1733: $N_{\rm H}= 0.17^{+0.23}_{-0.17}\times 10^{21}$ cm$^{-2}$ inferred from the
1734: PL+BB fit is consistent with the usual estimate based on the dispersion
1735: measure as well as with the $N_{\rm H}$ estimated from the PWN spectrum
1736: (see Fig.\ \ref{cont-PWN}),
1737:  and it is much lower than the total Galactic $N_{\rm HI}$.
1738: 
1739: 
1740: Also, the PL+BB model looks more attractive than the PL model
1741:  because it gives the slope of the PL component,
1742: $\Gamma\approx 1.7$, within the range of spectral slopes found 
1743: for a large sample of young
1744: pulsars \citep[$\Gamma \approx 1-2$;][]{2003ApJ...591..361G},
1745: while the PL fit to the B1929 spectrum results in
1746: a considerably steeper slope, $\Gamma \approx 3$.
1747: One could argue that 
1748: the PL fits to the spectra of other old pulsars \citep[e.g., 
1749: B0950+08, B2224+65, B0823+26, B0628--28, 
1750: B1133+16, B0943+10;][]{2004ApJ...616..452Z,2004ApJ...615..908B,2005ApJ...630L..57T,2006ApJ...636..406K,2005ApJ...624L.109Z}
1751: also show rather steep slopes, $\Gamma\approx 2-3$, suggesting that
1752: pulsar spectra might soften with increasing age or decreasing spin-down
1753: power \citep{2006ApJ...636..406K}. However, 
1754: a plausible alternative 
1755: interpretation
1756: of the softer spectra of old pulsars is
1757: that they are, in fact, comprised of
1758: a (soft) thermal component and a PL component with a more gradual 
1759: slope\footnote{Note
1760: that 
1761: the observed spectra of these pulsars can be satisfactorily fitted
1762: with quite different models
1763: because 
1764: these objects are very faint in the X-ray range.},
1765: similar to those found in young pulsars \citep{2004ApJ...616..452Z}.
1766: 
1767: 
1768: Thus, based on the goodness of fit and the astrophysical arguments,
1769: we conclude that the X-ray emission from B1929 most likely includes both
1770:  thermal 
1771: and magnetospheric components.
1772: The BB model for the thermal component gives the 
1773: apparent temperature $kT_a\approx0.3$\,keV
1774: and projected emitting area $A_{\perp,a}\sim 3000$ m$^2$, 
1775: at the distance of 361 pc,
1776: suggesting that the thermal emission originates from small heated spots
1777: (e.g., polar caps). These temperature and area correspond to the
1778: apparent radius $R_a^{\rm es}=(A_{\perp,a}/\pi)^{1/2}\approx 33$ m 
1779: and bolometric luminosity $L_{{\rm bol},a}^{\rm es}=4 A_{\perp,a}\sigma T_a^4
1780: \approx 1.1\times 10^{30}$ ergs s$^{-1}$ of an equivalent sphere.
1781: The true size and luminosity of the polar caps depend on the geometry
1782: and the gravitational redshift factor, $g_r=(1-R_g/R_{\rm NS})^{1/2}$,
1783: where $R_g=2.95 M_{\rm NS}/M_\odot$ km, $R_{\rm NS}$ and $M_{\rm NS}$ are
1784: the mass and radius of the neutron star. 
1785:  If there are two identical hot spots at the poles of a centered magnetic
1786: dipole, the polar cap radius 
1787: and the luminosity of two
1788: polar caps, as measured at the neutron star surface, are
1789: $R_{\rm pc}=
1790: R_{a}^{\rm es}f^{-1/2}$ and $L_{\rm bol,pc}=
1791: L_{{\rm bol},a}^{\rm es}/(2fg_r^4)$, where $f\leq 1$ is a geometrical factor
1792: depending on the angles $\zeta$ beween the line of sight and the
1793: spin axis and $\alpha$ between the spin and magnetic axes, as well as
1794: on $g_r$ \citep{2007ApJ...664.1072P}.
1795: For instance, assuming $M_{\rm NS}=1.4 M_\odot$, $R_{\rm NS}=10$ km 
1796: (i.e.\ $g_r=0.766$),
1797: $\alpha=36.0^\circ$, and $\zeta=61.5^\circ$ \citep{2001ApJ...553..341E} and
1798: using the approach described by \citet{1995AstL...21..149Z} and
1799:   \citet{2002ApJ...566L..85B},
1800: we find 
1801: $f=0.897$,
1802: which gives
1803:  $R_{\rm pc}= 
1804: 1.06 R_{a}^{\rm es}\approx 35$ m and 
1805: $L_{\rm bol,pc}=1.62 L_{{\rm bol},a}^{\rm es}
1806: \approx 1.8\times 10^{30}\, {\rm ergs\, s}^{-1}=4.7\times 10^{-4}\dot{E}$.
1807: 
1808: 
1809: 
1810: 
1811: 
1812: 
1813: 
1814: 
1815: 
1816: The estimated polar cap radius is a factor of $\sim$9 smaller than 
1817: the conventional polar cap radius of a radio pulsar,
1818: $\tilde{R}_{\rm pc} = (2 \pi R_{\rm NS}^3/cP)^{1/2}$
1819: \citep[see e.g.,][]{1991tnsm.book.....M}, which is
1820: $\approx 300$ m for B1929, assuming $R_{\rm NS} =10$ km.
1821: However, the observed 
1822: polar cap radii can be substantially smaller
1823: or larger than the conventional value.
1824: In particular,
1825: BB fits of several old pulsar spectra show $R_{\rm pc}\sim (0.1$--$0.2)\,
1826: \tilde{R}_{\rm pc}$ 
1827: \citep[e.g.,][]{2004ApJ...616..452Z,2005ApJ...624L.109Z,2006ApJ...636..406K}.
1828: For some pulsars, 
1829: such a discrepancy can be alleviated assuming that the polar cap
1830: is covered by a hydrogen or helium atmosphere, in which case the effective
1831: temperature would be a factor of 2 lower, and the radius a factor of 3--10
1832: larger 
1833:  while the bolometric luminosity would not change significantly 
1834: \citep{1995lns..conf...71P,2004ApJ...616..452Z}. 
1835: Another explanation for such a discrepancy
1836: was proposed by 
1837: \citet{2005ApJ...624L.109Z}, who suggested that only a small fraction of the polar cap surface,
1838: associated with footprints of sparks produced by intermittent breakdowns of an ``inner gap'' above the polar cap, is hot enough to emit X-rays.
1839: 
1840: 
1841: 
1842: The luminosity of the detected thermal component, which is less dependent on
1843: the assumed spectral model than the radius and temperature,
1844: can be compared with the models for 
1845: pulsar polar cap heating \citep{2002ApJ...568..862H,2001ApJ...556..987H}. 
1846: For B1929, these models predict
1847: the polar cap luminosity 
1848: $L_{\rm bol,pc}\sim 
1849: 10^{31}$\,ergs\,s$^{-1}$ if the polar cap
1850:  is heated by positrons produced through curvature radiation 
1851: of electrons accelerated in the neutron star magnetosphere.
1852: Although the polar cap thermal luminosity estimated from the
1853: PL+BB spectral model is a factor
1854: of a few lower than that predicted by \citet{2001ApJ...556..987H},
1855: this can be considered as a reasonably good agreement, with account
1856: of the uncertainties of both the theoretical model
1857:  and the observational results.
1858: We expect that deeper observations of other old pulsars would also detect
1859: thermal components from their polar caps and help us better understand
1860: the polar cap heating mechanisms.
1861: 
1862: 
1863: 
1864: As Figure~\ref{fig-combined-spectra-pl}  shows, the thermal component
1865:  dominates at 
1866: $0.5\,{\rm keV}\lesssim E \lesssim 2\,{\rm keV}$. This means that we can
1867: expect different pulse shapes within and outside this energy range,
1868: because the pulsations of thermal radiation should be smoother and
1869: shallower than those of the magnetospheric radiation. 
1870: The study of the energy dependence of pulsations not only can
1871: confirm the presence of the thermal component, but it would also allow
1872: one to infer the polar cap geometry and emission mechanism
1873: (e.g., BB vs.\ a light-element atmosphere), and even measure the mass-to-radius
1874: ratio for the neutron star \citep{1995AstL...21..149Z,1997ApJ...490L..91P,2004ApJ...616..452Z}.
1875:  Since the current {\sl XMM-Newton} data with
1876: sufficient time resolution do not
1877: have enough counts for such an analysis, a new deep observation of
1878: B1929 is needed to perform this important test. 
1879: 
1880: 
1881: The spectral slope
1882:  of the PL (magnetospheric) component inferred from the PL+BB fit 
1883: is $\Gamma\approx 1.7$,
1884: similar to those of younger pulsars. The luminosity of this component,
1885:  $L_X=
1886: 4\pi d^2 F_X^{\rm unabs} \approx 1.7\times 10^{30}$ ergs s$^{-1}$ in
1887: the 0.3--10 keV band,
1888: correponds to the nonthermal X-ray efficiency
1889:  $\eta_{\rm nonth}=L_X/\dot{E}\approx 4.4\times 10^{-4}$, typical for
1890: the whole population of radio pulsars detected in the X-rays, and comparable
1891: to the thermal (polar cap) efficiency. Notice that the one-component PL fit
1892: results not only in a much softer spectrum ($\Gamma\approx 3.0$) but also
1893: in a higher luminosity, $L_X\approx 6.6\times 10^{30}$ ergs s$^{-1}$,
1894: and efficiency, $\eta_{\rm nonth}\approx 1.7\times 10^{-3}$, above the
1895: typical value for X-ray emitting radio pulsars. A similar trend
1896: has been observed in other old pulsars: PL fits of their spectra yield
1897: X-ray efficencies noticeably higher that those of young and middle-aged
1898: pulsars, which might suggest that the X-ray efficiency in this energy
1899: range grows with increasing age and decreasing spin-down power, perhaps
1900: at the expence of the efficiency at higher photon energies.
1901: However, this conclusion is based on the assumption that the contribution
1902: of the thermal component is negligible in the soft X-ray range, which
1903: we believe is not valid at least for the best-studied old pulsars,
1904: B1929 and B0950+08. To understand the evolution of X-ray properties
1905: of pulsars, deep observations of a larger sample of old pulsars are
1906: warranted.
1907: 
1908: 
1909: 
1910: 
1911: 
1912: 
1913: 
1914: \subsection{Summary}
1915: 
1916: Two ACIS-S observations of B1929 revealed a faint 
1917: PWN surrounding
1918:  this old, nearby pulsar.
1919:  The observed morphology includes 
1920:  a compact nebula with two patchy wings in the immediate vicinity of B1929, and a tail
1921:  extending in the direction opposite to the pulsar's proper motion, 
1922: aligned with the much longer tail detected in the previous {\sl ROSAT} and
1923: {\sl XMM-Newton}
1924:  observations.  
1925: The shape of the nebula and its spectral properties
1926:  are  consistent with the proposed
1927: bow-shock classification.
1928: The properties of the compact nebula suggest that the pulsar wind outflow
1929: is anisotropic, possibly concentrated toward the equatorial plane
1930: perpendicular to the pulsar velocity. The size of the observed tail
1931: implies an average flow velocity of $\sim 0.1 c$. The whole observed
1932: PWN radiates about $2\times 10^{-4}$ of the pulsar spin-down power
1933: in X-rays.
1934:  By comparing the observed PWN properties with the bow-shock MHD models,
1935: we estimated the temperature  of the local ISM to be $\sim10^5$\,K,
1936:  which is consistent with the lack of
1937:  the H$\alpha$ bow-shock emission around this pulsar. 
1938: 
1939: 
1940: In the spectral analysis of the combined {\sl Chandra} 
1941: and {\sl XMM-Newton} data, we detected  a thermal component
1942: in the pulsar radiation,
1943: whose luminosity, $\sim (1$--$2) \times 10^{30}$\,ergs\,s$^{-1}$,
1944:  is comparable with the magnetospheric  
1945: luminosity in the X-ray band. 
1946: The thermal radiation is likely emanating from polar caps
1947: heated by positrons created by the 
1948: curvature radiation of ultrarelativistic electrons accelerated in the
1949: pulsar magnetosphere.
1950: The spectrum and X-ray efficiency of the magnetospheric component are
1951: similar to those found in young and middle-aged pulsars.
1952: Further X-ray observations of the pulsar will allow one to
1953: better characterize its spectral and timing properties and understand
1954: the nature of the magnetospheric radiation and the mechanisms of polar
1955: cap heating.
1956: 
1957: 
1958: 
1959: 
1960: We thank Patric Broos for the useful discussion about the MARX  
1961: simulations, Oleg Kargaltsev and Leisa Townsley for  the help
1962:  with the high-resolution imaging techniques, and Slava Zavlin
1963: for the discussion about the {\sl XMM-Newton} spectral results.
1964: We wouls also like to thank the referee for very helpful comments
1965: and suggestions.
1966: This work was partially supported by NASA grant NAG5-10865
1967: and {\sl Chandra} award SV4-74018.
1968: 
1969: 
1970:  
1971: \bibliographystyle{apj}
1972: \bibliography{./paper}
1973: 
1974: 
1975: 
1976: \clearpage
1977: 
1978: 
1979: \rotate{
1980: \begin{table}
1981: \scriptsize
1982: \caption[]{{\sl Chandra} and {\sl XMM-Newton} observations of B1929 }
1983: \begin{center}
1984: \begin{tabular}{cccc}
1985: \hline\hline\\
1986: Instrument & Obs ID & Date & Exposure  \\
1987: \hline\\
1988: EPIC & 0113051301 & 10 Nov 2003 & 10.5 / 10.5 / 7.3 \\
1989: EPIC & 0113051401 & 27 Apr 2004 & 15.2 / 16.4 / 11.0 \\
1990: EPIC & 0113051501 & 29 Apr 2004 &  7.3 / 7.6 / 10.3 \\
1991: ACIS & 6657 & 4 Dec 2005 & 20.9 \\
1992: ACIS & 7230 & 28 May 2006 & 24.6 \\ 
1993: \hline\\
1994: \end{tabular}
1995: \end{center}
1996: \tablecomments{
1997: The good-time exposures are given in ks. The first, second, and third
1998:  EPIC exposures are for MOS1, MOS2,
1999: and PN, respectively. The effective PN exposure in each observation is
2000:  $\sim$70\% of the total PN 
2001: good-time exposure because of the reduced efficiency of the small window mode.
2002: }
2003: \label{table-observations}
2004: \end{table}
2005: }
2006: 
2007: 
2008: \clearpage
2009: 
2010: 
2011: \rotate{
2012: \begin{table}
2013: \scriptsize
2014: \caption[]{Background-subtracted counts and surface brightness of 
2015: PWN regions in {\sl Chandra} observations 6657, 7230, and combined data}
2016: \begin{center}
2017: \begin{tabular}{ccccccc}
2018: \hline\hline\\
2019: Region &   &  1 (Southern Wing)  & 2 (Northern Wing) & 3 (Front) & 4 (Inner Blob) & 5 (Outer Blob)  \\
2020: \hline \\
2021:  Extraction area  &  & 31.5  & 46.7  &  77.9 & 157 & 1884 \\
2022: arcsec$^2$ &  &  &   &   &  &  \\
2023: \hline \\
2024: Counts & 6657 & 18.9$\pm$4.6  & 6.9$\pm$3.2 & 12.7$\pm$4.2 & 23.5$\pm$5.8 & 
2025: 57$\pm$14  \\
2026: (bkg-subtracted)    & 7230 & 15.3$\pm$4.1  & 17.3$\pm$4.5 & 3.7$\pm$2.8 & 14.4$\pm$4.8 & 
2027: 116$\pm$15  \\
2028:  & comb. & 34.1$\pm$6.1   & 24.1$\pm$5.5  &  16.4$\pm$5.0 &
2029: 37.9$\pm$7.7 & 173$\pm$20  \\
2030: &&&&&& \\
2031: \hline \\
2032: Bkg-subtracted & 6657 & 2.9$\pm$0.6  & 0.7$\pm$0.3 & 0.8$\pm$0.2 & 0.7$\pm$0.2 & 0.14$\pm$0.03   \\
2033:  surface brightness  & 7230 & 2.0$\pm0.5$  & 1.5$\pm$0.4 & 0.2$\pm$0.1 & 0.4$\pm$0.1 & 0.25$\pm$0.03  \\
2034: ($10^{-5}$ cts\,s$^{-1}$\,arcsec$^{-2}$)  & comb. & 2.4$\pm$0.4  & 1.4$\pm$0.2  &  0.5$\pm$0.1 & 0.6$\pm$0.1 & 0.20$\pm$0.02  \\
2035:  & & & & & & \\
2036: \hline\\
2037: \end{tabular}
2038: \end{center}
2039: \tablecomments{
2040: The regions are shown in Fig.\ \ref{fig-combined-conv-smooth-image}.
2041:  The errors represent statistical uncertainties at the 68\% confidence
2042: level. 
2043:  }
2044: \label{table-regions}
2045: \end{table}
2046: }
2047: 
2048: 
2049: \rotate{
2050: \begin{table}
2051: \scriptsize
2052: \caption{
2053: Fits to the spectrum of B1929
2054: for the  {\sl Chandra} and {\sl XMM-Newton} observations.}
2055: \begin{center}
2056: \begin{tabular}{cccccccccc}
2057: \hline\hline\\
2058: Model & $N_{\rm H}$ & $\Gamma$ & PL Norm. & $kT$  & Radius\tablenotemark{a}  & $\chi_{\nu}^2$/d.o.f & Absorbed Flux  & Luminosity  \\
2059:  &$10^{21}$ cm$^{-2}$ & & $10^{-5}$\,cm$^{-2}$\,s$^{-1}$\,keV$^{-1}$ & keV  & m &   & $10^{-13}$\,ergs\,cm$^{-2}$\,s$^{-1}$ & $10^{30}$\,ergs\,s$^{-1}$ \\
2060: \hline \\
2061: Chandra: &&&&&&&&\\
2062: PL & $2.45^{+0.52}_{-0.48}$ & $2.94^{+0.25}_{-0.22}$ &  $7.99^{+1.76}_{-1.17}$ & $\cdots$ & $\cdots$ &
2063: 0.91/71 & $1.37^{+0.14}_{-0.27}$ & $6.26^{+2.66}_{-0.93}$ \\
2064: PL+BB & $0.49^{+0.93}_{-0.49}$& $1.82^{+1.04}_{-0.56}$ & $1.91^{+3.86}_{-1.33}$ & $0.29^{+0.05}_{-0.06}$  & $33.8^{+46.3}_{-6.3}$ & 0.88/69 & $1.52^{+0.23}_{-0.37}$ & $4.66^{+0.86}_{-1.12}$ \\
2065: XMM-Newton: & & & & & & & & \\
2066:  PL & $2.22^{+0.37}_{-0.30}$ & $2.99^{+0.20}_{-0.15}$ & $8.96^{+1.32}_{-0.96}$ & $\cdots$ & $\cdots$ & 0.97/190 & $1.59^{+0.12}_{-0.25}$ & $7.08^{+2.18}_{-0.84}$ \\
2067:  PL+BB & $0.05^{+0.11}_{-0.05}$& $1.63^{+0.86}_{-0.29}$ & $1.43^{+3.73}_{-0.42}$ & $0.30^{+0.02}_{-0.05}$ & $34.9^{+11.0}_{-4.9}$ & 0.94/188 & $1.82^{+0.13}_{-0.26}$ & $2.87^{+0.19}_{-0.39}$ \\
2068:  combined: & & & & & & & & \\
2069:  PL & $2.23^{+0.27}_{-0.27}$& $2.95^{+0.14}_{-0.13}$ &$8.43^{+0.91}_{-0.79}$ & $\cdots$ & $\cdots$ &  1.03/264 & $1.51^{+0.12}_{-0.24}$ & $6.62^{+1.98}_{-0.82}$ \\
2070:  PL+BB & $0.17^{+0.23}_{-0.17}$& $1.73^{+0.46}_{-0.66}$ &$1.64^{+1.75}_{-0.28}$ &$0.30^{+0.02}_{-0.03}$ & $33.1^{+5.9}_{-4.6}$  &  0.98/262 & $1.75^{+0.11}_{-0.22}$ & $2.84^{+0.15}_{-0.22}$ \\
2071: \hline \\
2072: \end{tabular}
2073: \end{center}
2074: \tablecomments{The observed flux and unabsorbed luminosity, 
2075: $L_X=4\pi d^2 F_X^{\rm unabs}$,
2076:  are calculated for the 0.3--8\,keV energy band for {\sl Chandra}, and 0.3--10\,keV for {\sl XMM-Newton} and combined data. The listed uncertainties are at a 90\% confidence level determined for 2 interesting parameters.
2077: }
2078: \tablenotetext{a}{Radius of equivalent sphere for the BB component (see \$\,3.2).}
2079: \label{table-spectrum}
2080: \end{table}
2081: }
2082: 
2083: 
2084: 
2085: \end{document}
2086: