1: \documentclass{emulateapj}
2: \usepackage{apjfonts}
3: \usepackage{graphicx}
4: \usepackage{epsfig}
5: \usepackage{rotating}
6:
7: \newcommand{\bc}{\begin{center}}
8: \newcommand{\ec}{\end{center}}
9: \newcommand{\be}{\begin{equation}}
10: \newcommand{\ee}{\end{equation}}
11: \newcommand{\ba}{\begin{eqnarray}}
12: \newcommand{\ea}{\end{eqnarray}}
13: \newcommand{\bt}{\begin{tabular}}
14: \newcommand{\et}{\end{tabular}}
15: \newcommand{\un}{\underline}
16: \newcommand{\ov}{\overline}
17: \newcommand{\ros}{{\sl ROSAT}}
18: \newcommand{\chan}{{\sl Chandra}}
19: \newcommand{\einst}{{\sl Einstein}}
20: \newcommand{\nh}{n_{\rm H}}
21: \newcommand{\dd}{{\rm d}}
22: \newcommand{\edot}{\dot{E}}
23: \def\farcs{\hbox{$.\!\!^{\prime\prime}$}}
24: \def\farcm{\hbox{$.\!\!^{\prime}$}}
25:
26:
27: %\documentstyle[emulateapj,psfig,graphicx]{article}
28:
29: \def\chan{{\sl Chandra}}
30: \def\xmm{{\sl XMM}-Newton}
31: \def\ed{\dot{E}}
32: \def\lnon{L^{\rm nonth}}
33: \def\nh{n_{\rm H,20}}
34: \def\tbb{T^\infty_{\rm bb}}
35: \def\rbb{R^\infty_{\rm bb}}
36: \def\tpc{T_{\rm pc}}
37: \def\rpc{R_{\rm pc}}
38: \def\tef{T_{\rm eff}}
39: \def\lbol{L_{\rm bol}}
40: \def\lbolpc{L_{\rm bol}^{\rm pc}}
41: \def\epc{\eta^{\rm pc}}
42:
43: \newcommand{\gapr}{\raisebox{-.6ex}{\mbox{
44: $\stackrel{>}{\mbox{\scriptsize$\sim$}}\:$}}}
45: \newcommand{\lapr}{\raisebox{-.6ex}{\mbox{
46: $\stackrel{<}{\mbox{\scriptsize$\sim$}}\:$}}}
47: \def\farcs{\hbox{$.\!\!^{\prime\prime}$}}
48: \def\farcm{\hbox{$.\!\!^{\prime}$}}
49:
50:
51: \begin{document}
52: \submitted{Submitted to ApJ 2007 May 12}
53:
54:
55:
56: \title{X-ray emission from PSR J1809--1917 and its pulsar wind nebula,
57: possibly associated with the TeV gamma-ray source HESS J1809--193}
58:
59: \author{
60: O.\ Kargaltsev and G.\ G.\ Pavlov}
61: \affil{The Pennsylvania State University, 525 Davey Lab, University
62: Park, PA 16802, USA} \email{oyk100@psu.edu,pavlov@astro.psu.edu}
63:
64:
65: \begin{abstract}
66:
67: We detected X-ray emission from the 50-kyr-old pulsar J1809--1917
68: and resolved its pulsar wind nebula (PWN) with the {\sl Chandra X-ray
69: Observatory}. The pulsar's observed flux is $F_{\rm psr} = (1.8\pm
70: 0.2)\times 10^{-14}$ ergs cm$^{-2}$ s$^{-1}$ in the 1--6 keV band. A
71: two-component blackbody+power-law (BB+PL) fit of the pulsar's
72: spectrum yields the photon index $\Gamma_{\rm psr}=1.2\pm 0.6$
73: and luminosity $L_{\rm psr}=(4\pm 1)\times 10^{31}$ ergs s$^{-1}$ of
74: the PL component, in the 0.5--8 keV band, for a plausible distance
75: $d=3.5$ kpc and $n_{\rm H}=0.7\times 10^{22}$ cm$^{-2}$.
76: The BB component corresponds to the temperature $T\approx 2$ MK,
77: and bolometric luminosity
78: $L_{\rm bol}
79: \sim 1\times 10^{32}$ ergs s$^{-1}$. The bright inner PWN
80: component of a $3''\times 12''$ size is elongated in the north-south
81: direction,
82: with the
83: pulsar
84: close to its south
85: end. This component
86: is immersed in a larger ($\approx 20''\times 40''$), similarly elongated
87: outer PWN component of lower surface brightness. The elongated
88: shape of the compact PWN can be explained by the ram pressure
89: confinement of the pulsar wind due to the supersonic motion of the
90: pulsar. The observed flux of the compact PWN, including both
91: components, is $F_{\rm pwn}\simeq (1.5\pm 0.1)\times10^{-13}$
92: ergs cm$^{-2}$ s$^{-1}$ in the 1--6 keV band. The PWN spectrum can
93: be fitted with a PL model with $n_{\rm H}\approx 0.7 \times 10^{22}$
94: cm$^{-2}$ and photon index $\Gamma_{\rm pwn}=1.4\pm0.1$,
95: corresponding to the 0.5--8 keV luminosity $L_{\rm pwn} \approx
96: 4\times 10^{32}$
97: ergs s$^{-1}$. The compact PWN appears to be inside a more
98: extended ($\approx4'\times4'$) emission with the total observed flux
99: $F_{\rm ext}\sim 5\times10^{-13}$ ergs s$^{-1}$ in the 0.8--7 keV
100: band. This large-scale emission is more extended to the south of the
101: pulsar, i.e.\ in the direction of the alleged pulsar motion. To explain
102: the extended X-ray emission
103: ahead of the moving pulsar, one has to
104: invoke strong intrinsic anisotropy of the pulsar wind or assume that
105: this emission comes from a relic PWN swept by the asymmetrical
106: reverse SNR shock.
107: The pulsar
108: and its PWN are located within the extent of the unidentified TeV
109: source HESS~J1809--193. The
110: %center
111: brightest part of the TeV source is offset by
112: $\sim 8'$
113: to the south of the pulsar,
114: i.e.\ in the same direction as the large-scale X-ray emission. Although
115: the
116: association between the PSR~J1809--1917 and HESS~J1809--193
117: is plausible, an alternative source of relativistic electrons powering
118: HESS~J1809--193 might be the serendipitously discovered X-ray
119: source CXOU J180940.7$-$192544.
120: In addition to the
121: CMBR or Galactic starlight background, the low-frequency seed
122: photons for Compton upscattering to TeV energies might be supplied
123: by bright infrared emission from dust-molecular clouds seen within
124: HESS~J1809--193.
125:
126:
127: \end{abstract}
128: %%%%%%%%%%%%%%%%%%%%%%%%%%%
129: \keywords{pulsars: individual (PSR J1809--1917)
130: --- X-rays: individual (CXOU~J180940.7$-$192544, CXOU~180933.3$-$192959) ---
131: gamma-rays: individual (HESS J1809--193) ---
132: ISM: individual (IRAS 18067--1927, IRAS 18067--1921)}
133: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%
134: \section{Introduction}
135: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%
136:
137:
138:
139: {\sl Chandra} and {\sl XMM-Newton} observations
140: have
141: established
142: the ubiquity of
143: X-ray pulsar wind nebulae (PWNe) around young rotation-powered
144: pulsars (see the reviews by Kaspi et al.\ 2006 and Gaensler \& Slane
145: 2006). The X-ray PWN emission is produced by relativistic particles
146: gyrating in the magnetic field downstream of the termination shock
147: in the pulsar
148: wind (Kennel \& Coroniti 1994; Arons 2004).
149: Most of the PWNe have been discovered around young ($\tau\lesssim 30$ kyr)
150: pulsars.
151: The innermost parts of the young PWNe often show
152: axisymmetric morphologies, including toroidal structures and jets
153: along the pulsar's spin axis. Recently, it has become apparent that
154: PWNe
155: accompanying older pulsars
156: can also be quite luminous (e.g., McGowan et al.\ 2006). Many of
157: these older PWNe exhibit cometary morphologies
158: indicating that the pulsar wind is confined by the ram pressure
159: caused by the supersonic motion of the pulsar in the ambient
160: medium. Studying X-ray PWNe of various ages helps understand the
161: nature and evolution of the ultrarelativistic pulsar winds and their
162: interaction with the ambient medium.
163:
164:
165: An interesting object for such investigations is PSR J809--1917
166: (hereafter J1809). The discovery of this radio pulsar ($P=82.7$ ms) in
167: the Parkes Multibeam Pulsar Survey\footnote{
168: http://www.atnf.csiro.au/research/pulsar/pmsurv} was reported by
169: Morris et al.\ (2002). The pulsar's dispersion measure, ${\rm
170: DM}=197\, {\rm cm}^{-3}\, {\rm pc}$, and the Galactic electron
171: distribution models by Taylor \& Cordes (1993) and Cordes \& Lazio
172: (2002) give the distance to the pulsar of 3.7 and 3.5 kpc, respectively.
173: Having the spin-down age $\tau\equiv P/2\dot{P} =51$ kyr and
174: spin-down power $\dot{E}\equiv 4\pi^2 I \dot{P} P^{-3} \simeq
175: 1.8\times 10^{36} $ ergs s$^{-1}$, J1809 is somewhat older and less
176: energetic than the famous Vela pulsar ($\tau = 11$ kyr,
177: $\dot{E}=6.9\times 10^{36}$ ergs s$^{-1}$), which is accompanied by
178: a remarkable PWN resolved in radio (Dodson et al.\ 2003b) and
179: X-rays
180: (Pavlov
181: et al.\ 2003, and references therein). However, it is much more
182: energetic than typical ``middle-aged'' pulsars, such as B0656+14
183: ($\tau=110$ kyr, $\dot{E}=3.8\times 10^{34}$ ergs s$^{-1}$) and
184: Geminga ($\tau =340$ kyr, $\dot{E} =3.2\times 10^{34}$ ergs
185: s$^{-1}$), whose PWNe are very faint (e.g., Pavlov et al.\ 2006).
186:
187:
188: J1809 is young enough to look for a remnant of the supernova that
189: created the pulsar. Deep radio observations by Brogan et al.\ (2004)
190: %
191: have
192: revealed two compact HII regions and two SNRs, G11.03$-$0.05 and
193: G11.18+0.11, projected near the pulsar (the pulsar's Galactic
194: coordinates are $l=11.094^\circ$, $b=+0.080^\circ$). The distances to
195: the SNRs, estimated from the radio surface brightness-diameter
196: ($\Sigma-D$) relation (Case \& Bhattacharya 1998), are $\sim 16$ and
197: $\sim 17$ kpc, respectively. However, given the very large uncertainty
198: of the $\Sigma-D$ relation for faint SNRs, the association of J1809
199: with one
200: of the SNRs cannot be ruled out despite the discrepant distance estimates.
201: Brogan et al.\ (2004)
202: estimated that the pulsar would need to have the transverse speed of
203: about 200 or
204: 140 km s$^{-1}$ to originate from the geometrical
205: center of G11.03$-$0.05 or G11.18+0.11, respectively,
206: assuming that all three are at the same distance of
207: 4~kpc and the pulsar's true age is 50 kyrs.
208: The estimated speeds are close to the average speeds recently measured
209: for a large sample
210: of radio pulsars by Hobbs et al.\ (2005).
211:
212:
213: Before the detailed radio studies were carried out, the region
214: had been observed by {\sl ASCA}, first as part of the
215: Galactic plane survey (Sugizaki et al.\ 2001),
216: and then with a deeper follow-up exposure (Bamba et al.\ 2003;
217: Ueno et al.\ 2005).
218: These observations revealed an amorphous diffuse emission that encompassed the pulsar position.
219: No pulsar has been detected in these X-ray observations.
220: The observed large-scale emission was attributed to a new SNR,
221: dubbed G11.0+0.0. Bamba et al.\ (2003) have
222: found that the X-ray spectrum
223: of the putative SNR fits
224: the absorbed power-law (PL) model with photon index
225: $\Gamma=1.6^{+0.3}_{-0.2}$,
226: hydrogen column density $n_{\rm H,22}\equiv n_{\rm
227: H}/(10^{22}~{\rm cm}^{-2})=0.8\pm0.3$,
228: and
229: 0.7--10 keV flux
230: of $3.8\times 10^{-12}$ ergs cm$^{-2}$ s$^{-1}$.
231: Using the best-fit $n_{\rm H}$ and assuming that the mean density in
232: the Galactic plane is 1 H cm$^{-3}$, Bamba et al.\ (2003)
233: estimated the distance of 2.6 kpc,
234: which gives the X-ray luminosity of $\approx3.7\times 10^{33}$ ergs s$^{-1}$ (in 0.7--10 keV).
235: These authors suggested that G11.0+0.0 could be
236: a Crab-like (plerionic) SNR.
237:
238:
239: Studying of the J1809 field has become particularly interesting after
240: the recent discovery of the TeV $\gamma$-ray source
241: HESS~J1809--193 (hereafter HESS\,J1809; Aharonian et al.\ 2007). The
242: brightest, firmly detected part of this extended source (radius
243: $\sim12'$--$15'$) is centered at R.A.$ = 18^{\rm h}09.8^{\rm m}$,
244: decl.$=-19^{\circ}25'$, about $8'$ from J1809; its $\gamma$-ray flux
245: is $F_\gamma\approx 1.4\times 10^{-12}$ ergs cm$^{-2}$ s$^{-1}$ in
246: the 1--10 TeV band. Given the small angular separation, it seems
247: plausible that the HESS source could be powered by this pulsar, in
248: which case its $\gamma$-ray luminosity, $L_\gamma\approx 2\times
249: 10^{34} d^{2}_{3.5}$ ergs s$^{-1}$, would be about 1\% of the pulsar's
250: spindown power.
251:
252: In this paper, we describe the results of a {\sl Chandra} observation of
253: PSR J1809--1917 and its compact synchrotron nebula\footnote{
254: Preliminary results of this observation have been presented by
255: Sanwal et al.\ (2005).} and discuss its possible connection to
256: HESS~J1809. We also
257: describe the multiwavelength properties of objects located within
258: the central part of HESS~J1809,
259: including two newly discovered X-ray sources, and discuss their
260: relation to HESS~J1809. The details of the observation and the data
261: analysis are presented in \S2. In \S3 we discuss possible
262: interpretations of the PWN morphology, describe inferences from the
263: pulsar spectrum, and speculate on the nature of HESS~J1809 and its
264: relation
265: to the other sources in the field.
266: Our main results are summarized
267: in \S4.
268:
269: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
270: \section{Observations and Data Analysis}
271: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
272:
273: J1809 was observed with the Advanced CCD Imaging Spectrometer (ACIS)
274: on board {\sl Chandra} on 2004 July 21 (ObsID 3853).
275: The useful scientific exposure time was 19,955 s. The
276: observation was carried out in Faint mode, and the pulsar was
277: imaged on S3 chip, $\approx0.72'$ off-axis. The other ACIS chips
278: activated during this observation were S1, S2, S4, I2, and I3. The
279: detector was operated in Full Frame mode, which provides time
280: resolution of 3.24 seconds. The data were reduced using the Chandra
281: Interactive Analysis of Observations (CIAO) software (ver.\ 3.2.1;
282: CALDB ver.\ 3.0.3).
283:
284: \begin{figure*}[t]
285: \centering
286: \includegraphics[width=6.5in,angle=0]{f1.eps}
287: \caption{ {\em Top left:} ACIS-S3 image
288: of J1809 and its PWN (0.8--7 keV; pixel size $0.49''$). {\em Top right:}
289: Extraction regions used for the analysis of the PWN components
290: (0.8--7 keV; pixel size $0.98''$).
291: {\em Bottom left:} Adaptively smoothed sub-pixel resolution image
292: (0.8--7 keV; pixel size $0.25''$)
293: obtained by removing the pipeline pixel-randomization and applying
294: the sub-pixel resolution tool (based on analyzing the charge
295: distribution produced by an X-ray event; Tsunemi et al.\ 2001; Mori et
296: al.\ 2001). The straight line shows the approximate symmetry axis of
297: the X-ray PWN. {\em Bottom right:}
298: Heavily binned (pixels size $3.94''$) and
299: smoothed (with a gaussian kernel of $r=25''$)
300: ACIS-S3 image of J1809 and its PWN. The brightness and smoothing scales are chosen
301: to show the fainter, more extended emission.
302: }
303: \end{figure*}
304:
305:
306: \begin{figure}
307: \centering
308: \includegraphics[width=2.5in,angle=0]{f2.eps}
309: \caption{{\em Top:}
310: One-dimensional
311: surface brightness distribution along the symmetry axis of the J1809 PWN.
312: {\em Bottom:}
313: Extraction regions used to measure the pulsar's spectrum
314: (see \S2.2.2) and the sliding box used for measuring the one-dimensional
315: surface brightness distribution (see \S2.1 for details).
316: }
317: \end{figure}
318:
319: \subsection{Images}
320:
321: Figure 1 shows the ACIS-S3 image of the region around J1809. An
322: extended X-ray source is clearly seen in the image around
323: R.A.$=18^{\rm h}09^{\rm m}43.123^{\rm s}$, decl.$=-19^{\circ}17'
324: 38.17''$ (these are the coordinates of the center of the brightest
325: pixel). The difference of $0.5''$ between this position and the radio
326: position from Morris et al.\ (2006) is within the uncertainty of
327: absolute {\sl Chandra} astrometry ($0.6''$ at the 90\% confidence
328: level).
329: The close match between the X-ray and radio positions and
330: the extended morphology of the observed
331: X-ray emission allow us to conclude that we detected the X-ray
332: emission from J1809 and its PWN.
333:
334: The brighter {\em inner PWN} component of an $\approx
335: 12''\times3''$ size (i.e. $0.2\times 0.05$ pc$^2$ at $d=3.5$ kpc) is
336: elongated
337: along the approximate symmetry axis
338: (position angle $\approx14^{\circ}$ east of north). The linear profile
339: of the surface brightness distribution along the symmetry axis,
340: extracted with the $4.9''\times0.49''$ (i.e.\ $10\times1$ pixels) sliding
341: box, is shown in Figure 2. The inner PWN is surrounded by a similarly
342: elongated $\sim20''\times40''$ ``halo'' of lower surface brightness,
343: which we will call the {\em outer PWN} component (see Fig.\ 1, {\em
344: top}).
345: In addition to these two relatively compact
346: components,
347: the heavily binned and smoothed image (see Fig.\ 1, {\em bottom
348: right}, and Fig.\ 3{\em a}) reveals
349: even fainter large-scale {\em extended emission},
350: concentrated at the lower half
351: of the S3 chip (south of the pulsar).
352: The morphology of this emission (see also the {\sl ASCA} image in
353: Fig.\ 3{\em c}) indicates that it possibly extends further
354: south but may not be discernible on the S2 chip because of its lower sensitivity.
355:
356: \begin{figure*}[t]
357: \centering
358: \includegraphics[width=6.3in,angle=0]{f3.eps}
359: \caption{ J1809 and its vicinity at different wavelengths.
360: All the eight panels show the same area on the sky. {\em a:} ACIS-S3,
361: S2 and S1 images (0.8$-$7 keV; smoothed with $23.6''$ gaussian
362: kernel). The white ellipse shows the region used to estimate the flux
363: of
364: the large-scale diffuse emission (see \S2.2.1).
365: {\em b:}
366: The same image divided by the exposure map with point sources
367: removed (except for J1809). The contours show the radio emission
368: from two SNRs and two compact (possibly HII) regions (labeled A and
369: B; adopted from Brogan et al.\ 2004).
370: {\em c:} {\sl ASCA} GIS2 and GIS3 combined image (39 ks
371: total exposure;
372: 0.5$-$10 keV) of the same region (the image has been divided by the
373: exposure map; astrometry
374: has been corrected, courtesy of
375: %Eric
376: E.\ Gotthelf). {\em d:} Combined 9 ks EPIC-PN and MOS1+2 image
377: (0.5--10 keV) obtained in the Galactic plane survey (PI: R.\ Warwick).
378: {\em e:} The same as in the {\em panel
379: a} with no smoothing applied (pixel size is $7.4''$).
380: {\em f:} NRAO VLA Sky Survey (NVSS; Condon et al.\
381: 1998) image at 1.4 GHz.
382: {\em g:}
383: {\sl Spitzer} IRAC 8~$\mu$m image from the GLIMPSE survey.
384: {\em h:} {\sl Spitzer} MIPS 24~$\mu$m image.
385: The white circle in {\em panels c, f, g,} and {\em h} marks
386: the position of the pulsar, the diamond and the star mark the Ch1 and
387: Ch2 positions, respectively, and
388: the
389: cross shows the
390: %center of
391: %
392: %the brightest part
393: position of the peak of the TeV brightness distribution of HESS\,J1809.
394: % source.
395: }
396: \end{figure*}
397:
398:
399:
400: The only
401: source on the S2 chip detected above the $3 \sigma$ level is located
402: at ${\rm R.A.}=18^{\rm h}09^{\rm m}40.725^{\rm s}$, ${\rm
403: decl.}=-19^{\circ}25'44.10''$ (the $1\sigma$ centroid uncertainty is
404: $0.28''$ and $0.29''$ in right ascension
405: and declination,
406: respectively), well within the brightest central part of HESS~J1809
407: (about $2'$ west of the peak of the TeV brightness distribution; see
408: Fig.\ 3). We designate this source CXOU~J180940.7$-$192544 and call
409: it Ch1 hereafter. Although Ch1 appears to be extended in the ACIS
410: image, the point spread function (PSF) simulation shows that this is
411: likely
412: the result of the off-axis location
413: (off-axis angle $\theta=7\farcm4$). We find no significant large-scale
414: non-uniformities in the X-ray background on the S2 chip (see Fig.\ 3).
415:
416: We have also examined the S1 chip image and found only one source
417: detected above
418: the $3\sigma$ level.
419: The source, CXOU~180933.3$-$192959 (hereafter Ch2), is located at
420: ${\rm R.A.}=18^{\rm h}09^{\rm m}33.336^{\rm s}$, ${\rm
421: decl.}=-19^{\circ}29'59.89''$ (the $1\sigma$ centroid uncertainty is
422: $0.70''$ in R.A. and $0.74''$ in decl.), which is
423: about $6'$ from
424: the center of HESS~J1809. The X-ray source is consistent with being
425: point-like; however, due to the broadened PSF (FWHM $\approx10''$)
426: at the large off-axis angle ($11.8'$), a compact extended source
427: cannot be ruled out. We have also searched for diffuse emission
428: features on the S1 chip and produced an image corrected for the
429: exposure map non-uniformities (see Fig.\ 3{\em b}). In this image the
430: X-ray emission is systematically brighter toward the western edge of
431: the S1 chip, likely due to the imperfection of the mono-energetic (2
432: keV) exposure map correction at large off-axis angles (see, however,
433: \S2.5). We found no traces of the radio SNRs G11.03--0.05 and
434: G11.18+0.11 in the ACIS images (see Fig.\ 3).
435:
436:
437: \subsection{Spectral analysis}
438:
439: \subsubsection{PWN spectrum}
440:
441: We extracted the PWN spectra from two elliptical regions shown in
442: Figure 1 ({\em top right}). The smaller elliptical region (of 25.1
443: arcsec$^2$ area) encompasses the brighter inner PWN, while the
444: larger elliptical region
445: (of 399.4 arcsec$^2$ area) includes
446: the outer PWN component of a lower surface brightness. To avoid the
447: contamination of the PWN spectrum by the pulsar, we excluded from
448: these regions the circular region of $1.46''$ radius centered on the
449: brightest pixel. The background (367 counts in 5202 arcsec$^2$ area,
450: 0.3--8 keV band) was
451: measured from the
452: $37''<r<55''$ annulus centered on the pulsar.
453: The total numbers of counts
454: extracted from the smaller and larger elliptical regions (excluding the
455: $1.46''$ radius circle)
456: are 153 and
457: 365, of which 99.2\% and 92.4 \% are expected to come from the
458: source,
459: which gives
460: $151\pm 12$ and $337\pm 19$
461: PWN counts in the two regions.
462: The observed PWN fluxes (in the 1--6 keV band) are $F_{\rm inner} =
463: (8.1\pm0.7)\times 10^{-14}$ and $F_{\rm pwn}=(14.7\pm0.8) \times
464: 10^{-14}$ ergs s$^{-1}$ cm$^{-2}$ for the inner and the entire
465: (inner+outer) PWN, respectively. The corresponding average
466: intensities are $I_{\rm inner}=(4.4\pm 0.4)\times 10^{-15}$ and
467: $I_{\rm pwn}= (3.7\pm 0.2)\times 10^{-16}$ ergs cm$^{-2}$ s$^{-1}$
468: arcsec$^{-2}$.
469:
470:
471: To investigate the PWN spectral properties, we first fit the spectra for
472: each of the two PWN regions with the absorbed PL model, allowing
473: the hydrogen column density, $n_{\rm H}$, to vary.
474: These fits
475: result in
476: spectral slopes $\Gamma_{\rm pwn}\approx1.2$--2.0
477: %
478: and
479: $n_{\rm H,22} \sim0.5$--1.1 (the ranges correspond to the 68\%
480: confidence level for a single interesting parameter). The difference in
481: the best-fit parameters for the inner and the entire PWN (see
482: Fig.\ 4)
483: is statistically insignificant.
484: In particular, we see no
485: spectral softening (expected due to synchrotron cooling)
486: in the spectrum extracted from the larger region
487: (entire PWN)
488: compared
489: to the spectrum of the inner PWN\footnote{We have also measured
490: the spectrum of the outer PWN separately and found that the best-fit
491: PL parameters are consistent with those obtained for the inner PWN
492: and the entire PWN spectra, but their uncertainties are larger because
493: of the larger background contribution.}. Therefore, below we will use
494: the better constrained best-fit parameters
495: for the entire PWN
496: (see Figs. 5 and 6) whenever we refer to the PWN spectral properties.
497:
498: Given the J1809's dispersion measure, ${\rm DM}=197$ cm$^{-3}$ pc
499: (i.e., electron column density $n_e = 6.08\times 10^{20}$
500: cm$^{-2}$), the $n_{\rm H,22}$ value of 0.72 (obtained from the fit to the spectrum of the
501: entire PWN) corresponds to the ISM
502: ionization degree $n_e/n_{\rm H} \approx 8.4\%$, only slightly below
503: the commonly used value of 10\%, which corresponds to $n_{\rm H,22}=
504: 0.61$. The total Galactic HI column density in this direction is
505: $1.8\times 10^{22}$ cm$^{-2}$ (Dickey \& Lockman, 1990).
506:
507:
508:
509: \begin{figure}
510: \centering
511: \includegraphics[width=2.5in,angle=90]{f4.eps}
512: \caption{Confidence contours (68\% and 90\%) in the $n_{\rm
513: H}$--$\Gamma$ plane for the PL fit to the inner ({\em dashed})
514: and entire ({\em solid}) PWN spectra. The contours
515: are obtained with the PL normalization fitted at each point
516: of the grid.}
517: \end{figure}
518:
519:
520: The isotropic luminosity of the entire (inner+outer) X-ray PWN is
521: $L_{\rm pwn}\equiv 4\pi d^2 F_{\rm pwn}^{\rm unabs}\approx
522: (3.9\pm0.3)\times 10^{32}d_{3.5}^{2}$ ergs s$^{-1}$ in the 0.5--8 keV
523: band ($d_{3.5} = d/3.5\, {\rm kpc}$), with approximately equal
524: contributions from the outer and the inner PWN components (see
525: Table 1).
526:
527:
528: We have also attempted to fit the spectrum
529: of the faint large-scale emission (within the ellipse shown in Fig.\
530: 3{\em a}, excluding the compact PWN; $13.2$ arcmin$^{2}$ area)
531: surrounding J1809 and its compact
532: PWN. Because of the low surface brightness of this emission and large
533: background contribution ($\sim65$\%), the spectral fits yield
534: inconclusive results. With the absorption being fixed at $n_{\rm
535: H,22}=0.72$, the fits with single PL and thermal plasma\footnote{The
536: model ``mekal'' in XSPEC, with standard abundance} models give
537: $\Gamma = 2.3\pm 0.3$ ($\chi_\nu^2 \approx 2$) and $kT=0.67\pm
538: 0.08$ keV, ($\chi_\nu^2 \approx 3$). The main contribution to the
539: large values of $\chi_\nu^2$ comes from high-energy channels
540: ($\gtrsim 4$ keV and $\gtrsim 2$ keV for PL and mekal, respectively),
541: which suggests either a mixture of thermal and hard non-thermal
542: emission or varying $\Gamma$ or $kT$ within the extraction region.
543: The measured flux, independent of the
544: model, is
545: $F_{\rm ext}\sim5\times10^{-13}$ ergs s$^{-1}$ cm$^{-2}$ in the
546: 0.8--7 keV band, corresponding to the average surface brightness
547: $I_{\rm ext}\sim 1\times 10^{-17}$ ergs cm$^{-2}$ s$^{-1}$
548: arcsec$^{-2}$.
549: The unabsorbed
550: flux, obtained from the PL model, is $F_{\rm ext}^{\rm unabs}\sim
551: 8\times10^{-13}$ ergs s$^{-1}$ cm$^{-2}$ in the 0.5--8 keV band.
552:
553: \subsubsection{Pulsar spectrum}
554:
555: To minimize the contamination by the PWN, the pulsar spectrum was
556: extracted from a small circular aperture (green circle in the bottom
557: panel of Fig.\ 2) with the radius of 1.5 ACIS pixels ($\simeq 0.74''$,
558: 85\% encircled energy radius), while the background was taken from
559: the 10 arcsec$^{2}$ region between the white circle and white ellipse
560: in the bottom panel of Fig.\ 2. The background region includes the
561: bright part of the PWN; it contributes $\approx
562: 8$ counts to the total of 67 counts extracted from the source
563: aperture.
564: Given the small number of counts
565: and the large background contribution, we chose not to subtract the
566: background but rather to fit it simultaneously with the source
567: spectrum, using an additional absorbed PL model with the same
568: %
569: $n_{\rm H}$ as for the source.
570: The pulsar's absorbed flux
571: is $F_{\rm psr}=(1.8\pm0.2) \times10^{-14}$ ergs cm$^{-2}$ s$^{-1}$ in the
572: 1--6 keV band (aperture corrected).
573:
574: Although the
575: absorbed PL model formally fits the pulsar spectrum indicating a soft
576: PL ($\Gamma_{\rm psr}\approx2.6$; see Table 2 and Figs.\ 7 and 8),
577: the fit yields $n_{\rm H,22}\approx 0.4$, smaller than that for the PL
578: fit to the entire PWN spectrum.
579: To obtain a better constrained fit, we
580: fixed the hydrogen column density at $n_{\rm
581: H,22}=0.72$, obtained above from the PL fit to the PWN spectrum.
582: With this $n_{\rm H}$, the single PL fit is still acceptable,
583: but it yields a
584: large photon index, $\Gamma=3.2\pm0.4$, suggesting a thermal
585: emission contribution. A two-component, BB+PL, fit yields reasonable
586: values of fitting parameters (see Table 2 and Figs.\ 7--9), which,
587: however, are poorly constrained because of the small number of
588: photons detected. The slope of the PL component is $\Gamma_{\rm
589: psr}= 1.2\pm 0.6$, and its unabsorbed luminosity is $L_{\rm
590: psr}\sim4 \times 10^{31}d_{3.5}^2$ ergs s$^{-1}$ in the 0.5$-$8 keV
591: band. The temperature and the projected area
592: of the BB component are
593: strongly correlated (see Fig.\ 9), which results in large
594: uncertainties for these parameters. The
595: best-fit temperature for the BB component
596: is $T\approx2$ MK,
597: while the projected emitting
598: area,
599: $\mathcal{A}\sim 3 \times 10^{6}d_{3.5}^2$ m$^2$,
600: is smaller than that of the surface of a neutron star ($\pi
601: R^{2}\sim3\times 10^{8}$ m$^{2}$), but
602: larger than the
603: conventional polar cap area $A_{\rm pc}=2\pi^2R^3/cP\approx
604: 2.5\times10^5$
605: m$^{2}$. The corresponding bolometric luminosity,
606: $L_{\rm bol} \equiv 4{\mathcal{A}}\sigma T^4 \sim 1\times 10^{32}d_{3.5}^2$
607: ergs s$^{-1}$.
608:
609:
610: \begin{figure}
611: \centering
612: \includegraphics[width=3.2in,angle=0]{f5.eps}
613: \caption{ Entire (outer+inner)
614: PWN spectrum fitted with the PL model (see Table 1 for details).
615: }
616: \end{figure}
617:
618:
619:
620: \begin{figure}
621: \centering
622: \includegraphics[width=2.5in,angle=90]{f6.eps}
623: \caption{ Confidence contours (68\% and 90\%) for the PL fit to the
624: entire PWN spectrum (for a fixed $n_{\rm H,22}=0.72$).
625: The PL normalization
626: is in units of $10^{-6}$ photons cm$^{-2}$ s$^{-1}$ keV$^{-1}$ at 1
627: keV. The dashed curves are the loci
628: of constant unabsorbed flux in the 0.5--8 keV band;
629: the flux values near the curves are in units of 10$^{-14}$ ergs
630: cm$^{-2}$ s$^{-1}$.
631: }
632: \end{figure}
633:
634:
635:
636:
637: \begin{figure}
638: \centering
639: \vbox{
640: \includegraphics[width=3.2in,angle=0]{f7.eps}
641: } \caption{ {\em Top:} Pulsar spectrum fitted with the absorbed PL
642: model. {\em Middle:} Pulsar spectrum fitted with the absorbed PL model
643: where the $n_{\rm H}$ was fixed at the best-fit value for the entire
644: PWN
645: ($n_{\rm H,22}=0.72$). {\em Bottom:} The pulsar spectrum fitted with the
646: PL+BB model with
647: fixed $n_{\rm H,22}=0.72$. The dashed and dash-dotted histograms
648: correspond to
649: the BB and PL components, respectively.
650: The residual panels show the contributions of the energy bins into
651: the best-fit C-statistic (multiplied by $-1$ when the number of data
652: counts is smaller than the number of model counts).
653: }
654: \end{figure}
655:
656: \begin{figure}
657: \centering
658: \vbox{
659: \includegraphics[width=3.2in,angle=0]{f8.eps}
660: } \caption{ Confidence contours (68\% and 90\%) for the PL fit to the pulsars's spectrum with
661: $n_{\rm H,22}=0.72$ (green), PL+BB fit with $n_{\rm H,22}=0.72$
662: (red)
663: and PL fit with
664: $n_{\rm H,22}=0.4$ (black) which is the best-fit value for the PL
665: model.
666: The PL normalization
667: is in units of $10^{-6}$ photons cm$^{-2}$ s$^{-1}$ keV$^{-1}$ at 1
668: keV. The dashed curves are the lines of constant unabsorbed flux
669: in the 0.5--8 keV band (the flux values are in units of $10^{-14}$
670: ergs cm$^{-2}$ s$^{-1}$).
671: }
672: \end{figure}
673:
674:
675:
676: \begin{figure}[]
677: \centering
678: \includegraphics[width=3.2in,angle=0]{f9.eps}
679: \caption{ Confidence contours (68\% and 90\%) for the BB component
680: of the BB+PL fit to the pulsar's spectrum, for $n_{\rm H,22}=0.72$.
681: The BB normalization (vertical axis) is the
682: projected emitting area in units of m$^2$, assuming the distance of
683: 3.5 kpc. The lines of constant bolometric flux (in units of
684: 10$^{-14}$ ergs cm$^{-2}$ s$^{-1}$),
685: are plotted as dashed lines. }
686: \end{figure}
687:
688: \begin{table}[]
689: \caption[]{PL fits to the PWN spectrum} \vspace{-0.5cm}
690: \begin{center}
691: \begin{tabular}{ccccccc}
692: \tableline\tableline Model & $n_{\rm H,22}$ &
693: $\mathcal{N}$\tablenotemark{a} &
694: $\Gamma$ & ($C$ or $\chi^2$)\tablenotemark{b}/dof & $L_{\rm X}$\tablenotemark{c} \\
695: \tableline
696:
697: Entire PWN & $0.72$ &
698: $35.6^{+4.6}_{-3.8}$ &
699: $1.41^{+0.12}_{-0.11}$ & $0.82/34$ & $3.89_{-0.27}^{+0.24} $ \\
700: Inner PWN & $0.98$ &
701: $30.6^{+5.3}_{-4.6}$ & $1.85_{-0.17}^{+0.17}$ & $335/526$ & $2.21_{-0.17}^{+0.19} $ \\
702: \tableline
703: \end{tabular}
704: \end{center}
705: \tablecomments{The fits are for fixed $n_{\rm H,22}\equiv n_{\rm
706: H}/10^{22}$ cm$^{-2}$. The uncertainties are given at 68\%
707: confidence level for a single interesting parameter. }
708: \tablenotetext{a}{Spectral flux in units of $10^{-6}$ photons
709: cm$^{-2}$ s$^{-1}$ keV$^{-1}$ at 1 keV.} \tablenotetext{b}{We use the
710: C statistic (Cash 1979) for the inner PWN (which has less counts) and
711: the $\chi^2$ statistic for the entire PWN.}
712: \tablenotetext{c}{Unabsorbed isotropic luminosity in the 0.5--8 keV
713: band, in units of $10^{32}$ ergs s$^{-1}$. }
714: \end{table}
715:
716:
717:
718: \begin{table}[]
719: \caption[]{Fits to the pulsar spectrum} \vspace{-0.5cm}
720: \begin{center}
721: \begin{tabular}{ccccccc}
722: \tableline\tableline Model & $n_{\rm H,22}$ &
723: $\mathcal{N}$\tablenotemark{a} or $\mathcal{A}$\tablenotemark{b} &
724: $\Gamma$ or $kT$\tablenotemark{c} & $C$ & $L_{\rm X}$\tablenotemark{d} or $L_{\rm bol}$\tablenotemark{e} \\
725: \tableline
726:
727: PL & $0.72$ & $22.6_{-4.2}^{+4.9}$ & $3.22_{-0.38}^{+0.42}$ & $412$ & $0.98_{-0.17}^{+0.22}$ \\
728: PL & $0.4$ & $13.6^{+3.1}_{-2.4}$ & $2.59_{-0.28}^{+0.34}$ & $411$ & $0.65_{-0.09}^{+0.10}$ \\
729: PL+BB(PL) & $0.72$ & $2.84^{+2.66}_{-1.51}$ &
730: $1.23\pm0.62$ & $407$ & $0.37^{+0.12}_{-0.10}$ \\
731: PL+BB(BB) & $0.72$ & $2.8^{+5.6}_{-1.8}$ &
732: $0.17_{-0.03}^{+0.03}$ & $407$ &
733: $1.0^{+0.6}_{-0.4}$ \\
734: \tableline
735: \end{tabular}
736: \end{center}
737: \tablecomments{The fits are for fixed $n_{\rm H,22}\equiv n_{\rm
738: H}/10^{22}$ cm$^{-2}$. The uncertainties are given at 68\%
739: confidence level for a single interesting parameter. In each case the
740: fits were done using the C statistics (Cash 1979) and the unbinned
741: source and background spectra, with the total of 1052 channels. }
742: \tablenotetext{a}{Spectral flux in units of $10^{-6}$
743: photons cm$^{-2}$ s$^{-1}$ keV$^{-1}$ at 1 keV.}
744: \tablenotetext{b}{Projected area of the emitting region for the BB
745: model, in units of $10^{6}$ m$^2$.} \tablenotetext{c}{BB temperature
746: in keV.} \tablenotetext{d}{Unabsorbed PL luminosity in the 0.5--8 keV
747: band, in units of $10^{32}$ ergs s$^{-1}$.}
748: \tablenotetext{e}{Bolometric BB luminosity, in units of $10^{32}$ ergs
749: s$^{-1}$.}
750:
751: \end{table}
752:
753:
754: \subsubsection{ Ch1 spectrum}
755:
756: We extracted 99 events from the elliptical
757: region (with the semi-major and semi-minor axes of $10.3''$ and $6.3''$,
758: respectively)
759: centered on the best-fit
760: position of Ch1 and
761: conforming to the shape of the off-axis PSF.
762: Based on the PSF simulation, the extraction region contains
763: $\approx95\%$ of point source counts. The expected background contribution, 4.4 counts,
764: is negligible ($\approx4\%$ of the total number of counts).
765: The measured source flux
766: in
767: 0.8--7 keV is $F_{\rm Ch1}=(8.6\pm0.9)\times10^{-14}$
768: ergs cm$^{-2}$ s$^{-1}$, after correcting
769: for vignetting and finite extraction aperture size.
770: The spectrum fits well by an absorbed PL model (see Table 3) with
771: $n_{\rm H,22}=1.2\pm0.4$ and $\Gamma=1.4\pm0.4$.
772: The unabsorbed flux in 0.5--8 keV is
773: $F_{\rm Ch1}^{\rm unabs}\approx1.4\times10^{-13}$ ergs cm$^{-2}$ s$^{-1}$.
774:
775: \subsubsection{ Ch2 spectrum}
776:
777: We extracted 90 events within a $30''$ radius around
778: the best-fit
779: position of Ch2. Based on the PSF simulation, the extraction region contains
780: $\sim85\%$ of the source counts if Ch2 is a point source.
781: The background contributes $\approx32$ counts,
782: (i.e.\ $\approx36\%$ of the total number of counts). The measured
783: 0.8--7 keV source flux is $F_{\rm Ch2}\approx 9\times10^{-14}$
784: ergs cm$^{-2}$ s$^{-1}$, after correcting
785: for vignetting and finite extraction aperture size
786: (assuming a point source). The thermal plasma (mekal)
787: model fits the Ch2 spectrum best, although fits with absorbed
788: PL and BB models are also acceptable (see Table 3 for details).
789:
790: \begin{table}[]
791: \caption[]{Fits to the Ch1 and Ch2 spectra} \vspace{-0.5cm}
792: \begin{center}
793: \setlength{\tabcolsep}{1pt}
794: \begin{tabular}{cccccccc}
795: \tableline\tableline Source & Model & $n_{\rm H,22}$ &
796: $\mathcal{N}$\tablenotemark{a} or $\mathcal{A}$\tablenotemark{b} &
797: $\Gamma$ or $kT$\tablenotemark{c} & ($C$ or $\chi^2$)\tablenotemark{d}/dof
798: & $F_{\rm X}^{\rm un}$ or $F_{\rm bol}$\tablenotemark{e} \\
799: \tableline
800:
801: Ch1 & PL & $1.17^{+0.43}_{-0.39}$ & $17.8_{-7.3}^{+12.8}$ & $1.42_{-0.38}^{+0.40}$ & $291/523$ & $1.4\pm0.3$ \\
802: Ch2 & PL & $0.66 (<1.4)$ & $\sim45$ & $\sim6$ & $2.3/3$ & $\approx4.7$ \\
803: Ch2 & BB & $0.3 (<1.1)$ & $\sim0.28d_1^2$ & $\approx0.16$ &
804: $2.4/3$ & $\approx0.90$ \\
805: Ch2 & mekal &
806: $0.6_{-0.2}^{+0.6}$ & $\sim2.6$ & $0.7\pm0.2$ & $1.1/3$ &
807: $\approx0.54$
808:
809: \\
810: \tableline
811: \end{tabular}
812: \end{center}
813: \tablecomments{ The uncertainties are given at the 68\% confidence
814: level for a single interesting parameter. For Ch2, the upper limits on
815: $n_{\rm H}$ value (at 68\% confidence) are given in brackets, the
816: lower limits are not constrained at the same confidence level. }
817: \tablenotetext{a}{Normalization for the PL model is the spectral flux
818: at 1 keV in $10^{-6}$ photons cm$^{-2}$ s$^{-1}$ keV$^{-1}$.
819: Normalization for the mekal model is the Emission Measure (EM) in
820: $10^{53}$ cm$^{-3}$, scaled to a distance of 1 kpc. }
821: \tablenotetext{b}{Projected area of the emitting region for the BB
822: model, in units of $10^{6}$ m$^2$, normalized to $d=1$ kpc.}
823: \tablenotetext{c}{BB temperature in keV.} \tablenotetext{b}{We use
824: the C statistic for Ch1 and the $\chi^2$ statistic for Ch2.}
825: \tablenotetext{e}{Unabsorbed PL flux in the 0.5--8 keV band or
826: bolometric BB flux, in units of $10^{-13}$ ergs cm$^{-2}$ s$^{-1}$
827: (corrected for vignetting and finite extraction aperture size).}
828: \end{table}
829:
830:
831:
832:
833: \subsection{Timing}
834:
835: The 3.24 s time-resolution of this observation is insufficient to
836: observe the 82.7 ms pulsations from the pulsar. We found no
837: significant variability in the Ch1 and Ch2 lightcurves.
838:
839: \subsection{Archival X-ray data}
840:
841: The region of interest has been previously observed by {\sl ASCA} for
842: 39 ks
843: (Bamba et al.\ 2003). The {\sl ASCA} GIS image in Figure 3{\em c} shows
844: a region of enhanced X-ray brightness
845: that encompasses J1809 and
846: its compact PWN; however, the diffuse X-ray emission also extends at
847: least $\sim10'$ southward,
848: covering the
849: central region of
850: %
851: the brightest part of HESS~J1809.
852:
853: The field was also partly observed
854: by the {\sl XMM-Newton} as a part of the Galactic plane survey (PI: R.\ Warwick). A point-like object is clearly
855: seen at the pulsar position in the combined EPIC
856: (MOS1+MOS2+PN) image shown in Figure 3.
857: However, the short
858: 8 ks exposure, the off-axis location, and the high EPIC background
859: do not allow one to detect the faint
860: extended component, while the compact PWN cannot be resolved from the pulsar because of the broad
861: PSF of {\sl XMM-Newton}.
862:
863: \subsection{Optical-IR-radio data}
864:
865: To understand the nature of the X-ray sources Ch1 and Ch2, projected
866: within the HESS source image and look for other sources that could be
867: related to the TeV emission, we have examined the field at other
868: wavelengths. We found no counterparts to Ch1
869: in the Two Micron All Sky Survey
870: (2MASS; Skrutskie et al.\ 2006) or Digital Sky Survey (DSS2)\footnote{
871: see http://archive.eso.org/dss/dss}
872: catalogs,
873: up to the limiting magnitudes $K_s=15.4$, $H=16$, $J=17.5$, $R=19$,
874: and $B=21$. The nearest optical/NIR
875: source is a 2MASS point source
876: ($J=15.45\pm0.05$, $H=13.10\pm0.05$,
877: $K=12.06\pm0.04$) located at ${\rm R.A.}=18^{\rm h}09^{\rm
878: m}40.92^{\rm s}$, ${\rm decl.}=-19^{\circ}25'45.7''$. The $3.2''$ offset
879: from the best-fit Ch1 position substantially exceeds the position
880: uncertainty of $\approx 0.4''$. Figure 3 shows that Ch1 is projected
881: very close to the extended Source A seen in the radio and IR images.
882:
883:
884: The only NIR/optical source within the $5''$ radius of the Ch2 position
885: is a star, NOMAD1~0705$-$0568334 in the NOMAD catalog (Zacharias
886: et al.\ 2005), offset by only $0.9''$ from Ch2. Given the relatively large
887: uncertainty of the Ch2 position (see \S2.1), the positions of the star
888: and Ch2 can be considered coincident. The X-ray-to-optical flux ratio,
889: $F_X/F_V\sim 2\times 10^{-3}$, is typical for a K star with coronal
890: X-ray emission (Maccacaro et al.\
891: 1988).
892: The soft X-ray spectrum of Ch2 and the optical-NIR magnitudes of the
893: star ($B=14.96$, $V=14.66$, $R=14.19$, $J=13.35$, $H=12.84$, and
894: $K=12.72$) support such an interpretation. No Ch2 counterpart is
895: seen in the NVSS 20 cm image or in the {\sl Spitzer} 8 and 24 $\mu$m
896: images.
897:
898:
899: The NVSS 20 cm image ($45''$ restoring beam size) and the $25''$
900: resolution images by Brogan et al.\ (2004)
901: show two bright
902: compact radio sources (marked A and B in Fig.\ 3) projected near the
903: HESS source center,
904: with the 1.4 GHz spectral fluxes
905: of 0.3 Jy and 0.15 Jy for Source A and Source B, respectively. Within
906: these sources, 6.7 GHz methanol masers have been detected
907: (Pestalozzi et al.\ 2005):
908: G10.95+0.02 in Source A
909: (peak flux 15 Jy at 6.7 GHz, $V_{\rm LSR}=24\pm 1$ km s$^{-1}$,
910: near-distance 3.2 kpc) and G11.03+0.06 in Source B (peak flux 0.7 Jy
911: at 6.7 GHz, $V_{\rm LSR}=20\pm 1$ km s$^{-1}$, near-distance 2.9
912: kpc).
913:
914: Both A and B are also very bright far-IR sources, IRAS\,18067--1927
915: and IRAS\,18067--1921 in the IRAS Point Source Catalog v2.1 (IPAC
916: Infrared Science Archive\footnote{http://irsa.ipac.caltech.edu/}), with
917: peak fluxes of 1292 and 513 Jy at 60 $\mu$m, respectively. The {\sl
918: Spitzer\/} IRAC 8\,$\mu$m and MIPS 24\,$\mu$m images (Fig.\ 3{\em
919: g,h}; taken from the GLIMPSE
920: survey\footnote{http://www.astro.wisc.edu/sirtf/}) show extended
921: sources of irregular shape at these positions, with characteristic sizes
922: of $\sim 1'$. They are not seen in the DSS and 2MASS images,
923: which suggests that they are intrinsically very cold
924: ($T\sim 70$ K) and/or strongly absorbed molecular/dust complexes.
925:
926:
927:
928:
929:
930: The {\sl Spitzer} images also reveal an extended
931: source with
932: an interesting morphology (shell-like in the 24 $\mu$m image),
933: located $\sim10'$ southwest
934: of the
935: HESS~J1809 center (see Fig.\ 3{\em g,h}). The source has an extended
936: radio counterpart (NVSS~180919--192904)
937: clearly seen in the 20 cm image shown in the same figure.
938: It has been detected by Helfand et al.\ (2006) in the Multi-Array
939: Galactic Plane Survey (MAGPIS) and proposed to be an SNR candidate
940: G10.8750+0.0875. It is possible that this extended source contributes
941: to the non-uniform X-ray background on the S1 chip (see \S2.1 and
942: Fig.\ 3{\em b}). The SNR candidates G11.03$-$0.05 and G11.18+0.11
943: (Brogan et al.\
944: 2004; see Fig.\ 3{\em b})
945: are too faint to be seen in the shallow NVSS images.
946:
947:
948:
949: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
950: \section{Discussion.}
951:
952:
953: The sub-arcsecond angular resolution of {\sl Chandra} has allowed us
954: to resolve the compact J1809 PWN, disentangle the pulsar and the
955: extended emission components, and measure their properties
956: separately. Thanks to the very low ACIS background, we were also
957: able to detect the faint large-scale emission surrounding the compact
958: PWN. The observation has also provided serendipitous coverage of
959: the central part of the extended TeV source
960: HESS\,J1809 and revealed
961: two X-ray sources bright enough to permit spectral measurements and
962: accurate determination of their positions.
963: In this section we discuss some implementations of our findings.
964:
965:
966: \subsection{The J1809 PWN}
967:
968:
969: \subsubsection{Luminosity and spectrum}
970:
971: At the plausible distance of 3.5 kpc, the
972: unabsorbed
973: X-ray luminosity
974: of the
975: compact PWN, $L_{\rm pwn}\sim 4 \times 10^{32}$ ergs s$^{-1}$,
976: corresponds to the X-ray efficiency, $\eta_{\rm pwn}\equiv L_{\rm
977: pwn}/\dot{E} \sim 2.2\times10^{-4}$, similar to
978: those of PWNe around younger and more powerful Vela-like pulsars
979: (see Fig.\ 10). Some of the detected
980: X-ray PWNe associated with older and less powerful pulsars
981: have similar or higher X-ray efficiencies (e.g., $\eta_{\rm pwn}\sim 2\times 10^{-3}$
982: and $4\times 10^{-4}$ for J1509--5850 [$\tau=150$ kyr] and
983: B0355+54 [$\tau=560$ kyr], respectively; Kargaltsev et al., in
984: preparation), while others show much lower efficiencies (e.g.,
985: $\eta_{\rm pwn} \sim 5\times 10^{-5}$ and $7\times 10^{-6}$ for PRS
986: J1740+1000 [$\tau=100$ kyr] and Geminga [$\tau=340$ kyr],
987: respectively; Kargaltsev et al., in preparation; Pavlov et al.\ 2006).
988: This indicates that
989: $\eta_{\rm pwn}$ is not significantly correlated with $\tau$ or
990: $\dot{E}$, at least for young and middle-aged pulsars ($\tau \lesssim
991: 1$ Myr), perhaps because, in addition to $\dot{E}$ and $\tau$, it
992: depends on other factors (e.g., the pulsar's speed and the angle
993: between the spin and magnetic axes).
994:
995: The spectral slope of the J1809 PWN, $\Gamma_{\rm pwn} =1.4\pm
996: 0.1$,
997: is similar to those of the PWNe
998: around Vela-like pulsars (listed in
999: Table 2 of Kargaltsev et al.\
1000: 2007a, hereafter KPG07a), except for
1001: two bow-shock PWNe with prominent tails,
1002: J1747--2958 (the Mouse; Gaensler
1003: et al.\ 2004) and B1757--24 (the Duck; Kaspi et al.\ 2001), which show
1004: softer spectra ($\Gamma=2.0\pm0.2$ and $2.5\pm0.3$,
1005: respectively). Comparing the J1809 spectrum with those of older
1006: PWNe, we see that it is similar to the spectra of relatively compact
1007: PWNe (e.g., $\Gamma=1.4\pm0.3$ for the bright part of the
1008: B0355+54 PWN; McGowan et al.\ 2006), but it is harder than the
1009: spectra of the extended tails in
1010: J1740+1000
1011: and J1509--5850 PWNe ($\Gamma=1.8\pm0.4$ and $2.2\pm0.3$,
1012: respectively; Kargaltsev et al., in preparation).
1013: It hints that PWN
1014: spectra are correlated with PWN morphology rather than with the
1015: pulsar age and spindown power.
1016:
1017: \begin{figure}[]
1018: \centering
1019: \includegraphics[width=2.7in,angle=90]{f10.eps}
1020: \caption{ PWN luminosity versus pulsar spin-down power showing
1021: J1809 and 10 Vela-like pulsars observed with {\sl Chandra}. The
1022: luminosities are estimated for the 0.5--8 keV band. The dash-dot lines
1023: are the lines of constant PWN efficiency, $\eta_{\rm pwn}$. The error
1024: bars include the statistical uncertainties and the nominal 30\%
1025: distance uncertainties, except for the Vela pulsar whose parallax has
1026: been measured (Dodson et al.\ 2003a).
1027: }
1028: \end{figure}
1029:
1030: \subsubsection{PWN morphology}
1031:
1032:
1033:
1034: The unknown proper motion of J1809 and the faintness
1035: of the surrounding large-scale emission
1036: complicate the interpretation of the observed PWN,
1037: but its ``cometary'' appearance strongly suggests that the pulsar's
1038: motion plays a major role.
1039:
1040:
1041: \medskip
1042: \noindent
1043: {\sl 3.1.2.1. Bow shock in a nearly isotropic pulsar wind? ---}
1044: %
1045: The compact J1809 PWN is elongated approximately along the
1046: north-south direction, with the pulsar located much closer to its
1047: southern end (see Fig.\ 1).
1048: Such a
1049: cometary morphology can be attributed to a bow shock created by
1050: the pulsar moving supersonically in the southern direction (${\rm
1051: P.A.}\approx 194^\circ$). The apex of the termination shock (TS) of
1052: the pulsar wind is located at the
1053: distance
1054: %
1055: \be R_h \approx \left[\frac{\dot{E} f_\Omega}{4\pi c (p_{\rm
1056: amb}+p_{\rm ram})}\right]^{1/2} \ee
1057: %
1058: ahead of the pulsar. At this
1059: distance, the pulsar wind pressure, $p_{w} = \dot{E} f_\Omega (4\pi c
1060: r_s^2)^{-1}$ ($f_\Omega$ takes into account anisotropy of the pulsar
1061: wind), is balanced by the sum of the ambient pressure, $p_{\rm
1062: amb}= \rho kT (\mu m_{\rm H})^{-1}= 1.38\times 10^{-12} n \mu^{-1}
1063: T_4$ ergs cm$^{-3}$, and the ram pressure, $p_{\rm ram}= \rho v^2 =
1064: 1.67\times 10^{-10} n v_7^2$ ergs cm$^{-3}$ ($T_4 = T/10^4\,{\rm
1065: K}$, $v_7 = v/10^7\,{\rm cm\, s}^{-1}$, $\mu$ is the molecular weight,
1066: and $n = \rho/m_{\rm H}$ is in units of cm$^{-3}$). Assuming
1067: $p_{\rm ram}\gg p_{\rm amb}$ (or ${\mathcal M} \gg 1$, where
1068: ${\mathcal M}=v/c_s$ is the Mach number,
1069: $c_s=(5 kT/3 \mu m_{\rm H})^{1/2}=
1070: 12 \mu^{-1/2} T_4^{1/2}$ km s$^{-1}$
1071: is the sound speed in the ambient medium),
1072: we obtain $R_{h}= 1.7\times 10^{17}n^{-1/2} f_\Omega^{1/2}
1073: v_7^{-1}$ cm.
1074:
1075:
1076: The shocked pulsar wind, observed in X-rays as a bow-shock PWN,
1077: is confined between the TS and the contact discontinuity
1078: (CD) surface. For ${\mathcal M}\gg 1$ and a {\em nearly isotropic}
1079: preshock wind with a low magnetization parameter $\sigma$, the TS
1080: acquires a bullet-like shape (Bucciantini et al.\ 2005, hereafter B05).
1081: The distance $R_h$ between the pulsar and the bullet head is given
1082: by equation (1), while the CD surface head is at a distance $\approx
1083: 1.3 R_h$ from the pulsar. Assuming that the sharp rise in the
1084: brightness profile (Fig.\ 2) at about $3''$ from the
1085: pulsar (i.e.\
1086: $1.6\times 10^{17} d_{3.5}$ cm in the plane of the sky) corresponds
1087: to the CD head,
1088: we can estimate
1089: the pulsar velocity,
1090: %
1091: \be v \sim
1092: 140\, n^{-1/2} d_{3.5}^{-1} f_\Omega^{1/2} \sin i\,\,\, {\rm km\,\, s}^{-1},
1093: \ee
1094: %
1095: and the Mach number,
1096: ${\mathcal M}\sim 12 n^{-1/2} T_4^{-1/2} \mu^{1/2} f_\Omega^{1/2} d_{3.5}^{-1}\sin i$,
1097: where $i$ is the angle between the velocity vector and the line of sight.
1098:
1099:
1100: The observed projected length, $\sim 12''$, and width, $\sim 3''$, of
1101: the inner PWN are close to the length $\sim 7 R_h\sim 15''$, and
1102: width, $\sim R_h\sim 2''$, of the TS bullet,
1103: predicted by the
1104: B05 models for $\sigma\sim {\rm a\,\,\, few}\times 10^{-3}$. The
1105: width of the outer PWN, $\sim 20''$, can be interpreted as the
1106: diameter of the CD shell (approximately cylindrical behind the pulsar,
1107: with a diameter $\sim 8 R_h \sim 18''$, according to B05).
1108:
1109: The B05 models also predict a tail behind the back surface of the TS
1110: bullet, where the collimated shocked pulsar wind flows with
1111: subrelativistic velocities: $0.1c$--$0.3c$ in the inner channel with a
1112: cylindrical radius $\sim R_h$ ($\sim 2''$ in our case), and up to
1113: $0.8c$--$0.9c$ in the outer channel, confined between the cylindrical
1114: surfaces with radii $\sim R_h$ and $\sim 4 R_h$. The lack of an
1115: extended tail in the ACIS images is puzzling; it could be due to a
1116: relatively low Mach number of the pulsar and insufficient sensitivity
1117: of the short ACIS exposure.
1118: %our images could be explained assuming
1119: %that this collimated flow is receding from the observer, which would
1120: %lower its surface brightness.
1121: On the other hand, the shape of the faint
1122: thin feature northeast of inner PWN hints that it may be a bent
1123: extension of the inner PWN, which is difficult to explain if this is a
1124: strongly collimated, mildly relativistic flow just behind the back
1125: surface of the TS bullet (unless the flow is subject to kink instabilities).
1126: However, statistical significance of this feature is too low to draw any
1127: definitive conclusions.
1128:
1129:
1130: \medskip\noindent
1131: {\sl 3.1.2.2. Effects of pulsar wind anisotropy.} --- We should note that
1132: the B05 models assume an isotropic pulsar wind ($f_\Omega =1$).
1133: We know from observations of young pulsars, such as Crab and Vela,
1134: that the wind is not isotropic, but it is mostly confined
1135: %around
1136: %[{\bf in} OK]
1137: to the equatorial plane. In addition, in young PWNe we
1138: often see jets along the pulsar's spin axis. These jets can be formed by
1139: polar outflows originating in the pulsar magnetosphere or they can be
1140: created by tangential inflows just outside the equatorial TS surface
1141: converging toward the spin axis
1142: (Komissarov \& Lyubarsky 2004). We are unaware of PWN models
1143: that include both the pulsar wind anisotropy and the ram pressure
1144: effects. We can, however, expect that, in the case of an anisotropic
1145: wind, the PWN morphology, at least close to the pulsar, would
1146: depend on the orientation of the spin axis with respect to the pulsar's
1147: velocity.
1148:
1149:
1150: As {\sl Chandra} observations have shown, the spin axis is oriented
1151: along the direction of pulsar's motion in a number of
1152: young pulsars (e.g., Ng \& Romani 2004). In this case, we expect that
1153: the equatorial
1154: outflow would form a shell between the TS and the CD surface behind
1155: the pulsar, filled by a relativistic plasma with a subrelativistic bulk flow
1156: velocity. The morphology of the X-ray emission from such a PWN
1157: would be generally similar to that in the case of isotropic wind,
1158: although we may expect to see a shell-like (rather than filled) PWN
1159: appearance at sufficiently high resolution. Such a structure (``outer
1160: tails'') is possibly seen in the Geminga PWN (Caraveo et al.\ 2003;
1161: Pavlov et al.\ 2006). As for the jets coaligned with the pulsar velocity
1162: vector, the rear jet is expected to be seen along the axis of the shell
1163: filled by the shocked equatorial wind
1164: behind the pulsar (at least if the jet originates
1165: from the pulsar magnetosphere), as observed in the Geminga PWN
1166: (Pavlov et al.\ 2006). The front jet is expected to
1167: pierce the head of the shell (as $f_\Omega \gg 1$ in the jet),
1168: and it can be seen ahead of the pulsar unless
1169: it is crushed by the ram pressure.
1170: The jets could be of quite different brightness and length because
1171: of the Doppler
1172: boosting and different effective pressures in front and behind the pulsar.
1173:
1174:
1175: In the case of J1809, the high-resolution images shown in Figure 1
1176: indicate the presence of a narrow structure
1177: (with a linear extent of $\simeq 3''$)
1178: just
1179: south of the pulsar, which might be interpreted as a front jet.
1180: Furthermore, the heavily binned low-resolution image (Fig.\ 3{\em b})
1181: shows a faint structure extending in the direction of the presumed
1182: pulsar's proper motion. Although tentative, the structure is
1183: consistent
1184: with the X-ray morphology
1185: seen in the {\sl ASCA} GIS image (Fig.\ 3{\em c}).
1186: Therefore, one could speculate that
1187: these structures are connected to each other and could be parts of the front
1188: jet.
1189: In this interpretation, the rear jet may contribute to the inner PWN
1190: emission, and the above-mentioned bent extension of the inner PWN
1191: (\S3.1.2.1) could represent the outer part of the rear jet.
1192: Although barely supported by the existing data, the jet hypothesis
1193: offers a way to explain the offset large-scale X-ray (and TeV) emission
1194: south of the pulsar, which cannot be associated with the shocked
1195: pulsar wind under the assumption of wind isotropy. We speculate
1196: that this emission
1197: could be produced by particles supplied through
1198: the front jet (similar to the Vela PWN; Pavlov et al.\ 2003; Kargaltsev
1199: \& Pavlov 2004).
1200: However, to account for the
1201: luminous TeV emission ($L_{\gamma} \sim 2\times10^{-2}d_{3.5}^2
1202: \dot{E}$), the front
1203: jet would have to carry
1204: a substantial fraction of the pulsar's spindown power,
1205: unlike the northwest
1206: jet of the Vela PWN
1207: whose energy injection rate
1208: is only $\sim 10^{-3}\dot{E}$.
1209:
1210: We cannot also exclude a possibility that the spin axis of the J1809
1211: pulsar, and hence the PWN jets,
1212: are substantially misaligned with the direction of pulsar motion.
1213: For instance, one could assume that the jets are nearly perpendicular
1214: to the velocity vector (i.e. this vector lies in the equatorial plane; a
1215: possible example is PSR B1706--44 and its PWN; Romani et al.\ 2005).
1216: In this case, the jets would be bent backwards (or even destroyed)
1217: by the ram pressure of the oncoming ambient medium,
1218: while the TS in the equatorial outflow (ring-like without ram pressure
1219: applied) would turn into an ellipse-like structure elongated in the
1220: direction of pulsar motion, with the pulsar displaced along the major
1221: axis toward the TS head. The observational appearance of the
1222: shocked pulsar wind would then depend on the inclination of the
1223: orbital plane to the line of sight, but at most inclinations we would see
1224: an elongated structure with the pulsar shifted from the center of the
1225: PWN in the direction of the velocity vector, in a qualitative agreement
1226: with the observed shape of the inner PWN. Of course, such an
1227: interpretation would imply a higher pulsar velocity than that
1228: estimated above for a nearly isotropic pulsar wind (see eq.\ [2]). On
1229: the other hand, in such a geometry the wind could propagate to
1230: larger distances ahead of the pulsar
1231: because the CD could be easier
1232: destroyed by various instabilities, and the pulsar wind would be
1233: mixed with the shocked ambient medium. We note, however, that
1234: this picture can hardly explain the large-scale X-ray emission
1235: south-southwest of the pulsar, which, in this case, could be attributed
1236: to the host SNR or the crushed relic PWN (see \S3.3 and \S3.4.2,
1237: respectively).
1238:
1239: Overall, we can conclude that
1240: the observed PWN morphology is generally
1241: consistent with the assumption of the supersonic pulsar motion,
1242: which implies a low ambient pressure (e.g., $p_{\rm amb}\ll 2\times
1243: 10^{-10} f_\Omega d_{3.5}^{-2} \sin^2i$ ergs cm$^{-3}$ for a nearly
1244: isotropic pulsar wind). On the other hand, modeling of magnetized
1245: anisotropic winds from fast-moving pulsars, deeper X-ray
1246: observations, and proper motion measurements
1247: are needed to firmly establish the nature of the J1809 PWN
1248: and infer its properties quantitatively.
1249:
1250:
1251: \subsection{The pulsar}
1252:
1253: The spectrum of the J1809 pulsar is less certain
1254: than that of the PWN.
1255: Although the one-component PL model formally fits the spectrum,
1256: the fit yields a rather large $\Gamma\simeq2.6-3.2$ and suggests a
1257: smaller $n_{\rm H}$ than the one obtained from the PL fit to a better quality
1258: PWN spectrum.
1259: An alternative description of the pulsar spectrum is
1260: provided by the two-component PL+BB model that is often used to fit the spectra of young and middle-aged
1261: pulsars (see, e.g., KPG07a
1262: and references therein).
1263: The PL component of the PL+BB fit gives $\Gamma_{\rm psr} \simeq
1264: 1.2\pm 0.6$, similar to $\Gamma_{\rm pwn}$
1265: and to the spectral slopes of
1266: other pulsars of similar ages
1267: (see e.g., KPG07a).
1268: In any case, the slope is much softer than $\Gamma_{\rm psr} =
1269: 2.1 - 2.9 \dot{E}_{36}^{-1/2}\approx 0.1$ predicted by the Gotthelf's (2003)
1270: correlation.
1271: The X-ray efficiency of the pulsar, $\eta_{\rm psr} \equiv L_{\rm psr}/\dot{E}
1272: \sim 2\times 10^{-5} d_{3.5}^2$ in the 0.5--8 keV band, is also not unusual
1273: for Vela-like pulsars, as well as the ratio
1274: $L_{\rm pwn}/L_{\rm psr} \sim 10$ (see KPG07a).
1275:
1276: The possible thermal component of the PL+BB fit
1277: is
1278: poorly constrained, not only because of the scarce
1279: statistics but also because the soft thermal radiation
1280: is strongly absorbed by the ISM. The BB temperature,
1281: $T \sim 1.7$--2.3 MK, emitting area $\mathcal{A} \sim 10^6$--$10^7$
1282: m$^2$,
1283: and bolometric luminosity, $L_{\rm psr}^{\rm bol}
1284: \sim (0.6$-$1.6) \times 10^{32}$ ergs s$^{-1}$, are similar to those
1285: found from the PL+BB fits for Vela-like pulsars. The temperature is a
1286: factor of two higher than the NS surface temperatures
1287: predicted by standard NS cooling
1288: models for the 50 kyr age (e.g., Yakovlev \& Pethick 2004), and the
1289: corresponding emitting area is smaller than the NS surface area.
1290: However, the actual spectrum of the NS thermal radiation can differ
1291: substantially from the BB model. In particular, fitting the spectra with
1292: hydrogen atmosphere models (Pavlov et al.\ 1995) yields lower
1293: effective temperatures and larger emitting areas (see Pavlov et al.\
1294: 2001a for the specific example of the Vela pulsar). Unfortunately, the
1295: quality of the data does not warrant fits with more complicated
1296: atmosphere models.
1297: Unlike the temperature and the area,
1298: the bolometric luminosity is not so sensitive to the
1299: presence and
1300: properties of the NS atmosphere. The comparison with the NS cooling
1301: models shows that the J1809's bolometric luminosity is a factor of
1302: $\sim 10$ lower than predicted by the ``basic'' theoretical cooling
1303: curve ($M_{\rm NS} = 1.3 M_\odot$, no superfluidity); it is consistent
1304: with the cooling curves for heavier NSs (e.g., $M_{\rm
1305: NS}=1.5$--$1.6\, M_\odot$) for various superfluidity models
1306: (Yakovlev \& Pethick 2004).
1307:
1308:
1309: \subsection{Host SNR}
1310:
1311: With the velocity given by Equation (2), the pulsar would have
1312: traveled a distance of $\sim 7 n^{-1/2} d_{3.5}^{-1} f_\Omega^{1/2}
1313: \sin i$ pc during the time equal to its spindown age, $\tau=51$ kyr.
1314: This corresponds to a displacement of $\sim 7' n^{-1/2} d_{3.5}^{-2}
1315: f_\Omega^{1/2} \sin^2 i$ in the plane of the sky and, for $n^{-1/2}
1316: d_{3.5}^{-2} f_\Omega^{1/2} \sin^2 i\sim 0.6$, places the sky
1317: projection of the pulsar's birthplace close to the apparent center of
1318: the radio SNR G11.18+0.11 (see Fig.\ 3{\em b}), suggesting that J1809
1319: and G11.18+0.11 were created by the same supernova explosion.
1320: This conjecture implies that the pulsar has overtaken the
1321: SNR shell and left the high-pressure SNR interiors, which is consistent
1322: with the assumption $p_{\rm amb}\ll p_{\rm ram}$, used to obtain
1323: the pulsar velocity estimate. This could explain why J1809 looks so
1324: different from those Vela-like PWNe that do not show cometary
1325: morphology because they are moving subsonically in high-pressure
1326: SNR interiors (see KPG07a for examples). However, if G11.18+0.11 is
1327: at $d\approx 3.5$ kpc (the presumed distance to J1809), then its size,
1328: $D\approx 8 d_{3.5}$ pc, would be surprisingly small for a 50 kyr old
1329: SNR, unless it is expanding in an unusually dense environment. To get
1330: more consistent SNR age and size, we have to assume a larger
1331: distance and a smaller pulsar's true age, which would require a lower
1332: ambient density around J1809 to match the pulsar's birthplace with
1333: the center of G11.18+0.11, resulting in a higher inferred pulsar
1334: velocity and Mach number. For instance, if we assume $d=5$ kpc (i.e.,
1335: $D\sim 11$ pc) and the true age = 20 kyr, then $n\sim 0.1 f_\Omega
1336: \sin^4i$ cm$^{-3}$ is required, which corresponds to $v\sim 300(\sin
1337: i)^{-1}$ km s$^{-1}$, ${\mathcal M} \sim 25 T_4^{-1/2}\mu^{1/2}(\sin
1338: i)^{-1}$. Thus, given the uncertainty of the pulsar's true age and
1339: distance, we cannot rule out the possibility that G11.18+0.01 is the
1340: host SNR for J1809.
1341:
1342: It is also possible that G11.18+0.01 is just a background SNR
1343: accidentally projected near the pulsar. In this case, the real host SNR
1344: might be associated with the extended emission around J1809 seen in
1345: the {\sl Chandra} and {\sl ASCA} images (Fig.\ 1, {\em bottom right}
1346: and Fig.\ 3{\em c}). The faintness of
1347: the large-scale X-ray emission does not allow one to determine
1348: its origin (thermal or nonthermal) and, therefore,
1349: thermal plasma emission from the SNR interior cannot be
1350: excluded until better quality data are obtained (see \S2.2.1).
1351: However, the size of this putative SNR is too
1352: small, and, more importantly, the observed PWN morphology does not
1353: look consistent with this hypothesis. Therefore, it seems more plausible
1354: that the extended emission is synchrotron radiation related to the PWN
1355: (see \S3.1.2.2) rather than thermal emission from hot gas in SNR interior.
1356: Measuring the X-ray spectrum of the extended emission in a deeper observation
1357: would distinguish between these possibilities.
1358:
1359: One could also speculate that the host SNR is not seen because its size
1360: is larger than the field-of-view of our observation ($\gtrsim 20$ pc at
1361: $d=3.5$ kpc), and the SNR interior has become cold enough
1362: ($T\lesssim 10^5$ K) to provide ${\mathcal M}\gg 1$ and be
1363: undetectable in X-rays. This option remains quite viable if the pulsar's
1364: true age is close to (or exceeds) its spindown age.
1365:
1366: \subsection{
1367: Origin of HESS\,J1809}
1368:
1369: TeV radiation can be produced by the inverse Compton scattering
1370: (ICS) of low-frequency radiation (e.g., the cosmic microwave
1371: background radiation [CMBR]) off relativistic electrons. Alternatively,
1372: it can be generated by the $\pi^{0}\rightarrow\gamma + \gamma$
1373: decay, the $\pi^{0}$ mesons being produced
1374: when relativistic protons
1375: interact with the
1376: ambient matter.
1377: Therefore, to understand the nature of a TeV source, one should
1378: identify the {\em source} of relativistic particles and the
1379: {\em target} with which
1380: these particles interact.
1381:
1382:
1383:
1384:
1385: \subsubsection{Possible targets where the TeV radiation is produced}
1386:
1387: The omnipresent target for relativistic electrons producing
1388: TeV photons by the ICS
1389: is the CMBR, with the energy density $U_{\rm CMBR} = 0.26\,\, {\rm
1390: eV\,\, cm}^{-3}$. To produce TeV photons with energy $E_\gamma$
1391: by upscattering the CMBR photons
1392: with energy $\epsilon \sim 3kT\sim 4\times 10^{-4}$ eV, electrons
1393: with the Lorentz factor $\gamma \sim 5\times 10^7 E_{\rm
1394: TeV}^{1/2}$ [i.e. $E_e \sim 25 E_{\rm TeV}^{1/2}$ TeV] are required,
1395: where $E_{\rm TeV}=E_\gamma/(1\, {\rm TeV})$. As $\gamma
1396: \epsilon \sim 20$--$100\, {\rm keV} \ll m_ec^2$ for $E_\gamma\sim
1397: 1$--20 TeV, the ICS occurs in the Thompson regime. The same
1398: relativistic electrons produce synchrotron photons with energy
1399: %
1400: \be E_{\rm syn} \sim \gamma^2h\nu_{\rm cyc} \sim 5 \gamma_7^2
1401: B_{-5}\,\, {\rm eV} \sim 0.1 \frac{ E_{\rm TeV}
1402: B_{-5}}{\epsilon/(4\times 10^{-4}\, {\rm eV})}\, {\rm keV}\,, \ee
1403: %
1404: where $\gamma_7=\gamma/10^7$ and $B_{-5}=B/(10\,\mu{\rm G})$.
1405:
1406:
1407: Close to the Galactic plane, where HESS\,J1809 is situated, a factor of
1408: a few higher radiation energy density can be provided by IR emission
1409: from interstellar dust and Galactic starlight.
1410: For instance,
1411: the models of interstellar radiation
1412: field by Strong et al.\ (2000) give the energy densities of 0.6 and 2.7
1413: eV cm$^{-3}$ for these two components, respectively, at the
1414: galactocentric distance of $\sim 4$ kpc. Since the ICS proceeds in the
1415: Klein-Nishina (K.-N.) regime for $\epsilon \gtrsim 1 E_{\rm TeV}^{-1}$
1416: eV (i.e., the K.-N. effects become important at $\epsilon \gtrsim {\rm
1417: a\,\, few}\, \times 10^{-2}$ eV for the high-energy end of the TeV
1418: photon spectrum), we should use a more general formula for
1419: estimating the electron Lorentz factor:
1420: %
1421: \be \gamma \sim
1422: 10^6\left[E_{\rm TeV} + \left(E_{\rm TeV}^2 +E_{\rm
1423: TeV}/\epsilon_{\rm eV}\right)^{1/2}\right]\,, \ee
1424: %
1425: where
1426: $\epsilon_{\rm eV}=\epsilon/(1\, {\rm eV})$. The electrons with such
1427: a Lorentz factor generate synchrotron photons with energy $E_{\rm
1428: syn} \sim 0.05 \left[E_{\rm TeV} + \left(E_{\rm TeV}^2 +E_{\rm
1429: TeV}/\epsilon_{\rm eV}\right)^{1/2}\right]^2 B_{-5}$ eV. For instance,
1430: electrons with $\gamma\sim 2\times 10^7$ (i.e. $E_e \sim 40$ TeV)
1431: produce photons with $E_\gamma= 10$ TeV by upscattering
1432: background starlight photons with a typical energy $\epsilon\sim 1$
1433: eV, and the same electrons produce synchrotron radiation in a far-UV
1434: range, $E_{\rm syn}\sim 6$ eV, in an interstellar field of 3 $\mu$G.
1435: Although the radiation energy density of the starlight can substantially
1436: exceed that of the CMBR, the emissivity is reduced by the smaller
1437: Compton cross section in the K.-N. regime and smaller photon
1438: number density.
1439:
1440:
1441: The radiation energy density and photon number density (hence TeV
1442: emissivity) can be enhanced in the vicinity of very bright sources. The
1443: radiation energy density at an angular distance $\theta$ from a
1444: source of radiation can be estimated as $U = F/(c\theta^2)$,
1445: where $F$ is the source energy flux observed at Earth;
1446: it exceed the ambient radiation energy
1447: density, $U_{\rm amb}$, at
1448: %
1449: \be
1450: \theta > (F/c U_{\rm amb})^{1/2}.
1451: \ee
1452:
1453: The multiwavelength data show two very bright radio-IR objects near
1454: the center of HESS\,J1809 -- Source A and Source B (see Fig.\ 3), which
1455: are likely molecular-dust complexes in star-forming regions. An
1456: approximate integration of the spectral flux of Source A gives its total
1457: flux $F_A \sim 10^{-7}$ ergs cm$^{-2}$ s$^{-1}$ (i.e. the luminosity
1458: $L_A \sim 1.5\times 10^{38} d_{3.5}^2$ ergs s$^{-1}$). This means
1459: that the radiation energy density near this source exceeds $U_{\rm
1460: amb}$ at $\theta < 5' (U_{\rm amb}/1\,{\rm eV\, cm}^{-3})^{-1/2}$.
1461: For instance, if the TeV emission is produced by the ICS of IR photons
1462: with energies around $\epsilon \sim 0.02$ eV (which corresponds to
1463: the maximum spectral flux of Source A), and $U_{\rm amb}\sim 0.1$
1464: eV cm$^{-3}$ in this energy range, then $U>U_{\rm amb}$ at $\theta
1465: \lesssim 15'$. This value is close to the observed radius of
1466: HESS\,J1809, which suggests that the sphere of $\sim 10$ pc of the enhanced
1467: radiation field around Source A might
1468: be the target where the TeV emission is produced by the ICS off the
1469: relativistic electrons supplied by a nearby source
1470: [Lorentz factors $\gamma \sim
1471: (0.5$--$6)\times 10^7$ are required to upscatter the IR photons to
1472: $E_\gamma = 0.5$--20 TeV]. Similar estimates hold for Source B,
1473: whose flux is a factor of 2.5 lower, and the size $\theta$ is a factor of
1474: 1.6 smaller, than those for Source A. The hypothesis that the ``photon
1475: sphere'' of Source A or Source B
1476: is the site where the
1477: TeV emission
1478: is produced
1479: via ICS
1480: could be verified by detection of the accompanying synchrotron
1481: radiation. However, most of the energy range of the
1482: corresponding synchrotron emission, $E_{\rm syn} \sim (1$--$200) B_{-5}$ eV,
1483: is subject to the strong interstellar absorption.
1484:
1485: In principle, the TeV emission could be initiated by high-energy
1486: nucleons supplied by some source (e.g., a pulsar), which produce
1487: decaying $\pi^0$ mesons in collisions with nucleons of the ambient
1488: medium (e.g., Horns et al.\ 2006). The target in this case would the
1489: nucleonic component of the circumpulsar medium, and we should
1490: expect enhanced TeV emission in dense, cold clouds, if there are such
1491: clouds close to the source of high-energy nucleons. The lack of an
1492: extended IR or radio counterpart of a size similar to that of
1493: HESS\,J1809
1494: (which would indicate the
1495: presence of a large molecular cloud) suggests that the TeV source is
1496: not associated with a localized target for high-energy nucleons.
1497:
1498:
1499: \subsubsection{Possible sources of relativistic particles}
1500:
1501: In the case of HESS\,J1809, the natural source of relativistic electrons
1502: (and possibly protons) is the J1809 pulsar. In addition to that, there
1503: are the Ch1 and Ch2 X-ray sources and the radio SNR G11.03-0.05, all
1504: projected close the center of HESS\,J809 (see Fig.\ 3). Finally,
1505: relativistic particles might be produced by acceleration mechanisms in
1506: the forward shock of the putative host SNR of the J1809 pulsar.
1507:
1508:
1509: The soft X-ray spectrum of Ch2 and
1510: the positional coincidence with
1511: a field star suggest that
1512: the X-ray emission comes from an active stellar corona.
1513: Therefore, we conclude that Ch2 is
1514: not related to HESS\,J1809.
1515:
1516: If the shell-like radio SNR G11.03--0.05 were a powerful source of
1517: relativistic electrons, its synchrotron radiation would have been seen
1518: in the X-ray range. Since no X-rays at the SNR location are detected in
1519: our {\sl Chandra} observation, this SNR is not a viable candidate for
1520: the source of relativistic particles that powers HESS\,J1809.
1521:
1522: As we have discussed in \S3.1.2 and \S3.3, the faint,
1523: large-scale X-ray emission south of the J1809 pulsar is likely
1524: not the host SNR of this pulsar, but it is rather related to the J1809
1525: PWN. Moreover, TeV $\gamma$-ray emission from an SNR shock is
1526: usually associated with the SNR shell, not the interior.
1527: Current radio, X-ray and $\gamma$-ray data provide no evidence of such
1528: a shell around J1809. Therefore,
1529: even if the large-scale X-ray emission belongs to the host SNR,
1530: HESS\,J1809 is unlikely to be powered by
1531: particles accelerated in the forward shock of this SNR.
1532:
1533:
1534: Among the X-ray sources detected with {\sl Chandra}, Ch1 is the
1535: closest to the projected HESS\,J1809 center (Fig.\ 3). The lack of an
1536: IR-optical counterpart and the large lower limit on the X-ray-to-optical
1537: flux ratio, $F_X/F_{\rm opt}\gtrsim 3$, mean that the X-ray emission
1538: from Ch1 comes not from a usual field star. On the other hand, the
1539: absorbing hydrogen column density, $n_{\rm H,22} = 1.2\pm 0.4$, is
1540: close to that of the J1809 PWN, which suggests that Ch1 is a Galactic
1541: object, not an AGN observed through the Galactic plane. The slope of
1542: the Ch1 spectrum, $\Gamma = 1.4\pm 0.4$, suggests that it may be
1543: either an X-ray binary or a remote PWN unresolved because of the far
1544: off-axis location. With the optical extinction estimated from the
1545: measured $n_{\rm H}$ value (e.g., $A_V\sim 4$--9), the limits on the
1546: optical-NIR magnitudes (see \S2.5) virtually exclude a high mass X-ray
1547: binary. Although a low mass X-ray binary (LMXB) in a quiescent state
1548: could have an X-ray luminosity comparable
1549: to that of Ch1,
1550: $L_{\rm Ch1}\sim 3\times 10^{32}(d/4~{\rm kpc})^{2}$ ergs s$^{-1}$,
1551: LMXBs are not known to be sources of TeV emission.
1552: If Ch1 is a young pulsar with a PWN, then it might
1553: provide relativistic particles needed for generating the TeV emission.
1554: (We should note, however, that the TeV-to-X-ray flux ratio, $\sim
1555: 150$, would be much higher than the values, $\sim 0.01$--3, inferred
1556: for most of the other TeV PWNe, except for the B1800--21, for which
1557: the ratio is about 100; see Table 2 in Kargaltsev et al.\ 2007b,
1558: hereafter KPG07b, and references therein). To assess the
1559: %likelyhood
1560: likelihood
1561: of Ch1 being the source of relativistic particles for HESS\,J1809, a
1562: deeper {\sl Chandra} observation is needed, with Ch1 imaged close to
1563: the telescope's optical axis.
1564:
1565:
1566: Another plausible physical counterpart to HESS\,J1809 is the J1809
1567: pulsar/PWN. To date, young pulsars have been found in the vicinity
1568: of $\sim10$ extended TeV sources (e.g., de Jager 2006; Gallant 2007),
1569: and the likelihood of this happening by chance is very low (see, e.g.,
1570: KPG07b). The ratio of the TeV flux of HESS~J1809 to the X-ray flux the
1571: J1809 PWN (including the outer component of the compact PWN) is
1572: about 100, higher than those observed in most TeV plerions (KPG07b).
1573: However, if one adds in the X-ray flux of the large-scale extended
1574: emission (\S2.2.1), then the ratio decreases down to $\sim20$ (or
1575: even lower if one includes the diffuse emission of a larger extent,
1576: seen by {\sl ASCA}; see Fig.\ 3{\em c}).
1577:
1578:
1579:
1580: The center of
1581: %
1582: the brightest part of HESS~J1809 is offset from
1583: J1809
1584: by $\approx
1585: 8'$, almost in the direction of the presumed pulsar motion. Such an
1586: offset is comparable to the $10'$--$20'$ offsets found in most of TeV
1587: plerions (including the most secure associations
1588: PSR\,B0833--45/HESS\,J0835--455 [Vela] and
1589: PSR\,B1823--13/HESS\,J1825--137). An obvious explanation for such
1590: an offset is that the photons for the ICS are provided by a compact
1591: source (e.g., Source A) whose center lies at some distance from the
1592: pulsar ($\geq 8 d_{3.5}$ pc for Source A). Also, the offset might be
1593: due to the anisotropic supply of relativistic particles through the
1594: putative front jet (see \S3.1.2).
1595:
1596: If the TeV emission is produced by the ICS of an approximately
1597: uniform background radiation, such as the CMBR or Galactic starlight,
1598: then the origin of the offsets and the asymmetries of the extended
1599: X-ray and TeV PWN components could be attributed to the reverse
1600: SNR shock that had propagated through the nonhomogeneous SNR
1601: interior and reached one side of the PWN sooner than the other side,
1602: crushing the PWN and sweeping its contents along (Blondin et al.\
1603: 2001). One can assume that the TeV source HESS\,J1809 is powered
1604: by this relic crushed PWN, whose relativistic electrons have not lost all
1605: their energy to synchrotron radiation because of, e.g., a lower
1606: magnetic field, while the compact PWN in the pulsar vicinity is created
1607: by fresh electrons, recently injected from the pulsar magnetosphere.
1608: As the electrons responsible for the TeV emission have been
1609: accumulating during a substantial fraction of the pulsar's lifetime, and
1610: the pulsar's spindown power was higher in the past, this hypothesis
1611: could explain the uncomfortably large ratio, $\sim 0.01 d_{3.5}^2$, of
1612: the TeV luminosity to the {\em current} spindown power. (A similar
1613: explanation has been suggested by Aharonian et al.\ 2006 for the
1614: PSR\,B1823--13/HESS J1825--137 association.) This hypothesis can be
1615: tested with deep radio observations, which should be able to detect
1616: the synchrotron emission from the relic electrons of the crushed
1617: PWN.
1618:
1619:
1620: In the above picture, the asymmetric extended X-ray emission
1621: south of the J1809 pulsar could also be associated with the crushed
1622: PWN and considered as an X-ray (synchrotron) counterpart of the TeV
1623: (IC) source HESS\,J1809. The small size of the extended X-ray
1624: emission, in comparison with the size of the TeV source, and its offset
1625: from the center of HESS\,J1809 could be attributed to the lower
1626: sensitivity of the ACIS chip S2 (where the central part of HESS\,J1809 is
1627: imaged -- see Fig.\ 3{\em a}) or to a lower magnetic field in the central
1628: part. A deeper X-ray observation could distinguish between these two
1629: possibilities.
1630:
1631:
1632:
1633: The large-scale morphology of the X-ray emission around J1809 is
1634: remarkably similar to that around PSR B1823--13, which powers the
1635: X-ray PWN G18.0--0.7 ($L_{X}\sim 3 \times 10^{33}$ ergs s$^{-1}$,
1636: angular size $\gtrsim5'$; Gaensler et al.\ 2003). In addition to
1637: the extended low-surface-brightness component,
1638: the B1823--13 PWN has a much more compact
1639: ($5''$--$10''$) brighter core,
1640: resolved by {\sl Chandra}
1641: (Teter et al.\ 2004).
1642: The TeV emission from HESS~1825--137 covers an area much larger
1643: than the X-ray PWN area, extending up to $1^{\circ}$ southward from
1644: the pulsar (Aharonian et al.\ 2006). However, just as in the case of
1645: PSR~J1809/HESS~J1809, both the TeV and the low-surface-brightness
1646: X-ray emission have similar shapes, and they are offset in the same
1647: direction with respect to the pulsar position.
1648:
1649:
1650: Finally, we would like to point out that there are two other young
1651: pulsars, J1811--1925 and J1809--1943, at distances of $\approx 24'$
1652: and $\approx 19'$, respectively, from the HESS\,J1809 center. PSR
1653: J1811--1925 can not be related to HESS\,J1809 since it has been
1654: associated with G11.2--0.3 (Kaspi et al.\ 2001), whose diameter,
1655: $\approx 4'$, is
1656: substantially smaller than the offset from HESS~J1809.
1657: PSR J1809--1943 is the radio counterpart of the transient Anomalous
1658: X-ray Pulsar (AXP) XTE J1810--197 (Halpern et al.\ 2005, and
1659: references therein).
1660: Its association with HESS\,J1809
1661: does not look very plausible because of the large spatial offset and
1662: because its properties
1663: are quite different from those of the young pulsars found in the vicinity
1664: of other TeV sources (no TeV emission
1665: from AXPs have been reported so far).
1666:
1667: Thus,
1668: based on the current data, only two objects,
1669: the J1809 pulsar/PWN and, less likely, Ch1,
1670: are plausible candidates for the source of relativistic particles
1671: powering HESS\,J1809. In both cases, the TeV emission is generated
1672: by ICS of either IR photons from Source A (or Source B) or CMBR
1673: photons. To discriminate between different possibilities, the nature of
1674: Ch1 should be established in deep, high-resolution {\sl Chandra}
1675: observations, and the extended X-ray PWN emission should be
1676: studied in a deep {\sl XMM-Newton} exposure. The synchrotron
1677: emission from the alleged relic PWN could be detected in deep radio
1678: observations.
1679:
1680:
1681: \section{Conclusion}
1682:
1683: We have detected the X-ray emission from PSR B1809--19 and its synchrotron
1684: nebula. The X-ray efficiency and spectrum of the PWN are similar to
1685: those of many other compact PWNe, both younger and older than
1686: J1809. The cometary shape of the compact PWN suggests that the
1687: pulsar is moving supersonically, but no extended tail is
1688: detected behind the pulsar, perhaps because the Mach number is
1689: not large enough or the exposure time is too short. Our analysis
1690: of the PWN morphology suggests an anisotropic pulsar wind outflow,
1691: possibly including jets oriented along the direction of pulsar motion.
1692: To establish the nature of the compact PWN unambiguously, proper
1693: motion measurements and a deeper {\sl Chandra} observation are
1694: required.
1695:
1696: The compact PWN is immersed in
1697: an extended emission of lower surface brightness. This extended
1698: emission is offset with respect to the pulsar in the direction opposite
1699: to that of the compact PWN (i.e.\ in the direction of the alleged proper
1700: motion). If the extended emission is indeed powered by the
1701: supersonically moving pulsar, then the radiating particles could be
1702: supplied through the
1703: front jet of the compact PWN or brought by the
1704: northern part of the reverse SNR shock
1705: that
1706: overtook the pulsar moving southward.
1707:
1708:
1709: The spectrum of the
1710: pulsar
1711: can be described by a two-component BB+PL model. For reasonable
1712: $n_{\rm H}$ values, the parameters of this
1713: model
1714: and the corresponding component luminosities resemble those
1715: of Vela-like pulsars.
1716:
1717: The J1809 pulsar and its PWN are located in the vicinity of the
1718: unidentified extended TeV source HESS~J1809--193, which can be
1719: powered by ultrarelativistic electrons accelerated in the J1809 PWN.
1720: In addition to J1809, we found another X-ray source, Ch1,
1721: within the central part of HESS\,J1809, which might also be an
1722: unresolved pulsar/PWN. Although the true nature of Ch1 remains
1723: elusive, we cannot rule out the possibility that Ch1 is an alternative
1724: X-ray counterpart of HESS\,J1809. Whatever of the two objects, J1809
1725: or Ch1, supply the relativistic electrons, the TeV emission is likely
1726: produced by the ICS of either CMBR or IR photons from bright IRAS
1727: sources (Source A or Source B) projected near HESS\,J1809.
1728:
1729:
1730:
1731:
1732: \acknowledgements Our thanks are due to Divas Sanwal, PI of the {\sl
1733: Chandra} observation, who participated in the initial analysis of the
1734: {\sl Chandra} data. We also thank Eric Gotthelf for providing us with
1735: the {\sl ASCA} images with the latest calibration corrections applied.
1736: We are grateful to Kostya Getman and Leisa Townsley for useful
1737: discussions
1738: about the optical and IR properties of starforming regions.
1739: Support for this work was provided by the National Aeronautics and
1740: Space Administration through Chandra Award Number GO3--4075X
1741: issued by the Chandra X-ray Observatory Center,
1742: which is operated by the Smithsonian Astrophysical Observatory for and
1743: on behalf of the National Aeronautics Space Administration under
1744: contract NAS8-03060.
1745: This work was also partially supported by NASA grant NAG5-10865.
1746:
1747: \begin{thebibliography}
1748:
1749:
1750: \bibitem[]{} Aharonian, F., et al. 2006, ApJ, 636, 777
1751:
1752: \bibitem[]{} Aharonian, F., et al. 2007, A\&A, submitted (arXiv:0705.1605v1 [astro-ph])
1753:
1754:
1755: \bibitem[]{} Arons, J. 2004, Adv.\ in Sp.\ Res., 33, 466
1756:
1757:
1758: \bibitem[]{}Bamba, A., Ueno, M., \& Koyama, K., \& Yamauchi, S.
1759: 2003, ApJ, 589, 253
1760:
1761: \bibitem[]{} Blondin. J., Chevalier, R., \& Frierson, D. 2001, ApJ, 563, 806
1762:
1763:
1764: \bibitem[]{} Brogan, C.\ L., Gelfand, J.\ D., Gaensler, B.\ M., Kassim,
1765: N.\ E., \& Lazio, T.\ J. 2004, AJ, 127, 355
1766:
1767: \bibitem[]{} Bucciantini, N., Amato, E., \& Del Zanna, L. 2005, A\&A,
1768: 434, 189 (B05)
1769:
1770:
1771: \bibitem[]{}
1772: Caraveo, P.\ A., Bignami, G.\ F., De Luca, A., Mereghetti, S., Tur, A., \& Becker, W. 2003, Science, 301, 1345
1773:
1774: \bibitem[]{} Case, G.~L., \& Bhattacharya, D. 1998, ApJ, 504, 761
1775:
1776: \bibitem[]{} Cash, W. 1979, ApJ, 228, 939
1777:
1778:
1779: \bibitem[]{} Cordes, J.\ M., \& Lazio, T.\ J. 2002, astro-ph/0207156
1780:
1781: \bibitem[]{} de Jager, O. 2006, 26-th IAU Meeting, On the Present and Future of Pulsar
1782: Astronomy, Joint Discussion 2, 16-17 August,
1783: 2006, Prague, Czech Republic, JD02, \#53
1784:
1785:
1786: \bibitem[]{} Dickey, J.\ M., \& Lockman. F.\ J. 1990, ARA\&A, 28, 215
1787:
1788: \bibitem[]{} Dodson, R., Legge, D., Reynolds, J.\ E., \& McCulloch, P.\
1789: M. 2003a, ApJ,
1790: 596, 1137
1791:
1792: \bibitem[]{} Dodson, R., Lewis, D., McConnell, D., \& Despande, A. 2003b, MNRAS,
1793: 343, 116
1794:
1795:
1796: \bibitem[]{} Gaensler, B..\ M., \& Slane, P.\ O. 2006, ARA\&A, 44, 17
1797:
1798:
1799: \bibitem[]{} Gaensler, B.\ M., van der Swaluw, Camilo, F., Kaspi, V.\
1800: M., Baganoff, F.\ K., Yusef-Zadeh, F., \& Manchester, R.\ N. 2004,
1801: ApJ, 616, 383
1802:
1803: \bibitem[]{}
1804: Gallant, Y., for the HESS collaboration 2007, Ap Space Sci., in press
1805: (arXiv:astro-ph/0611720v1)
1806:
1807:
1808: \bibitem[]{} Gotthelf, E.\ V. 2003, ApJ, 591, 361
1809:
1810: \bibitem[]{}
1811: Halpern, J.\ P., Gotthelf, E.\ V., Becker, R.\ H., Helfand, D.\ J., \& White, R.\ L. 2005, ApJ, 632, L29
1812:
1813: \bibitem[]{}
1814: Helfand, D.\ J., Becker, R.\ H., White, R.\ L., Fallon, A., \& Tuttle, S. 2006, AJ, 131, 2525
1815:
1816:
1817: \bibitem[]{} Hobbs, G., Lorimer, D.\ R., Lyne, A. G., \& Kramer, M.
1818: 2005,
1819: MNRAS, 360, 974
1820:
1821: \bibitem[]{} Horns, D., Aharonian, F., Hoffmann, A.\ I.\ D., \&
1822: Santangelo, A. 2007, Ap\&SS, published online, (astro-ph/0609386)
1823:
1824: \bibitem[]{} Kargaltsev, O., \& Pavlov, G.\ G. 2006, Astrophys.\ Space
1825: Sci, in press (astro-ph/0609656)
1826:
1827: \bibitem[]{} Kargaltsev, O., Pavlov, G.\ G. \& Garmire G.~P. 2007a,
1828: ApJ, 660, 1413 (KPG07a)
1829:
1830: \bibitem[]{} Kargaltsev, O., Pavlov, G.\ G. \& Garmire G.~P. 2007b,
1831: ApJ,
1832: submitted,
1833: (astro-ph/0701069) (KPG07b)
1834:
1835: \bibitem[]{} Kaspi, V.\ M., Gotthelf, E.\ V., Gaensler, B.\ M., \&
1836: Lyutikov, M. 2001a, ApJ, 562, L163
1837:
1838:
1839: \bibitem[]{} Kaspi, V.\ M., Roberts, M.\ S.\ E., \& Harding, A.\ K. 2006,
1840: In Compact Stellar X-ray Sources, ed.\ W.\ H.\ G.\ Lewin \& M.\ van
1841: der Klis (Cambridge: Cambridge Univ.\ Press), 279
1842:
1843:
1844:
1845: \bibitem[]{} Kennel, C.\ F., \& Coroniti, F.\ V. 1984, ApJ, 283, 694
1846:
1847: \bibitem[]{} Komissarov, S.\ S., \& Lyubarsky, Y.\ E. 2004, MNRAS, 349,
1848: 779
1849:
1850:
1851: \bibitem[]{}
1852: Maccacaro, T., Gioia, I.\ M., Wolter, A., Zamorani, G., \& Stocke, J.\ T.
1853: 1988, ApJ, 326, 680
1854:
1855:
1856: \bibitem[]{}
1857: McGowan, K.\ E., Vestrand, W.\ T., Kennea, J.\ A., et al.\ 2006, ApJ,
1858: 647, 1300
1859:
1860: \bibitem[]{} Mori, K., Tsunemi, H., Miyata, E., Baluta, C. J., Burrows, D.
1861: N., Garmire, G. P., \& Chartas, G. 2001, in ASP Conf.\ Ser.\ 251, New
1862: Century of X-Ray Astronomy, eds. H.\ Inoue \& H. Kunieda (San
1863: Francisco: APS), 576
1864:
1865: \bibitem[]{} Ng, C.-Y., \& Romani, R.\ W. 2004, ApJ, 601, 479
1866:
1867:
1868: \bibitem[]{} Pavlov, G.\ G., Shibanov, Yu.\ A., Zavlin, V.\ E., \& Meyer,
1869: R.\ D. 1995, in The Lives of the Neutron Stars, eds.\ M.A.\ Alpar, U.\
1870: Kiziloglu, \& J.\ van Paradijs (Kluwer: Dordrecht), 71
1871:
1872: \bibitem[]{} Pavlov, G.\ G., Zavlin, V.\ E., Sanwal, D., Burwitz, V., \&
1873: Garmire, G.\ P. 2001a, ApJ, 552, L129
1874:
1875: \bibitem[]{} Pavlov, G.\ G., Kargaltsev, O.\ Y., Sanwal, D., \& Garmire,
1876: G.\ P. 2001b, ApJ, 554, L189
1877:
1878: \bibitem[]{} Pavlov, G.\ G., Teter, M.\ A., Kargaltsev, O., \& Sanwal, D.
1879: 2003, ApJ, 591, 1157
1880:
1881: \bibitem[]{}
1882: Pavlov, G.\ G., Sanwal, D., \& Zavlin, V.\ E. 2006, ApJ, 643, 1146
1883:
1884: \bibitem[]{} Pestalozzi, M.\ R., Minier, V., \& Booth, R.\ S.
1885: 2005, A\&A 432, 737
1886:
1887:
1888: \bibitem[]{} Romani, R.\ W., Ng, C.-Y., Dodson, R., \& Brisken, W.
1889: 2005, ApJ, 631, 480
1890:
1891: \bibitem[]{}
1892: Sanwal, D., Pavlov, G.\ G., \& Garmire, G.\ P. 2005, BAAS, 37, 497
1893:
1894: \bibitem[]{}
1895: Skrutskie, M.\ F., et al. 2006, AJ, 131, 1163
1896:
1897: \bibitem[]{}
1898: Strong, A.\ W., Moskalenko, I.\ V., \& Reimer, O. 2000, ApJ, 537, 763
1899:
1900: \bibitem[]{}
1901: Sugizaki, M., Mitsuda, K., Kaneda, H., Matsuzaki, K., Yamauchi, S.,
1902: \& Koyama, K. 2001, ApJS, 134, 77
1903:
1904: \bibitem[]{} Taylor, J. H. and Cordes, J. M. 1993, ApJ, 411, 674
1905:
1906: \bibitem[]{} Teter, M.\ A., Sanwal, D., Pavlov, G.\ G., \& Tsuruta, S.
1907: 2003, BAAS, 35, 706
1908:
1909: \bibitem[]{} Tsunemi, H., Mori, K., Miyata, E., Baluta, C. J., Burrows, D.
1910: N., Garmire, G. P., \& Chartas, G. 2001, ApJ, 554, 496
1911:
1912: \bibitem[]{} Ueno, M., Yamaguchi, H., Koyama, K., Bamba, A., Yamauchi, S.,
1913: \& Ebisawa, K. 2005, in X-ray and Radio Connections,
1914: eds.\ L.\ O.\ Sjouwerman \&
1915: K.\ K.\ Dyer), published electronically by NRAO,
1916: http://www.aoc.nrao.edu/events/xraydio
1917:
1918:
1919: \bibitem[]{} Yakovlev, D.\ G., \& Pethick, C.\ J. 2004, ARA\&A, 42, 169
1920:
1921: \end{thebibliography}
1922:
1923:
1924:
1925:
1926:
1927:
1928:
1929:
1930:
1931:
1932:
1933:
1934:
1935:
1936:
1937:
1938:
1939:
1940:
1941:
1942: \end{document}
1943: