1: \documentclass[12pt,preprint]{aastex}
2: \usepackage{epsfig}
3:
4: \shortauthors{Cui \& Konopelko}
5: \shorttitle{X-ray Counterpart of HESS~J1804-216}
6:
7: \def\be{\begin{equation}}
8: \def\ee{\end{equation}}
9: \def\mdot{\dot{m}}
10: \def\ergs{{\rm\,erg\,s^{-1}}}
11: \def\msun{M_{\odot}}
12: \def\ergscc{\rm \ \ erg \ cm^{-3} \ s^{-1}}
13: \def\gs{\rm \,g\,s^{-1}}
14: \def\ergscc{\rm \,erg\,cm^{-3}\,s^{-1}}
15: \def\ergs{\rm \,erg\,s^{-1}}
16: \def\be{\begin{equation}}
17: \def\ee{\end{equation}}
18: \catcode`\@=11 % This allows us to modify PLAIN macros.
19: \def\@versim#1#2{\vcenter{\offinterlineskip
20: \ialign{$\m@th#1\hfil##\hfil$\crcr#2\crcr\sim\crcr } }}
21: \def\lsim{\mathrel{\mathpalette\@versim<}}
22: \def\gsim{\mathrel{\mathpalette\@versim>}}
23: \def\mpy{M_\odot \ {\rm yr^{-1}}}
24:
25: \begin{document}
26:
27: \title{Chandra View of the Unidentified TeV Gamma-ray Source HESS~J1804-216}
28:
29: \author{Wei Cui and Alexander Konopelko}
30: \affil{Department of Physics, Purdue University, West Lafayette, IN 47907}
31:
32: \begin{abstract}
33:
34: We present high-resolution X-ray images taken with the {\em Chandra X-ray
35: Observatory} of the field that contains the unidentified TeV gamma-ray source
36: HESS~J1804-216. A total of eleven discrete sources were detected with {\it
37: a posteriori} significance of $> 5\sigma$ over the entire field of view.
38: Among them, only one, designated as CXOU~J180351.4-213707, is significantly
39: extended. The source is about 40\arcsec\ away from the radio pulsar
40: PSR~J1803-2137, which was the target of the {\em Chandra} observation
41: but was not detected in X-rays. A natural question is whether the two
42: sources are physically related. While it is conceivable that
43: CXOU~J180351.4-213707 could be associated with a previously unknown supernova
44: remnant (SNR), in which the pulsar was born, it seems equally plausible that
45: it might be a pulsar wind nebula (PWN) that is powered by a different
46: pulsar whose emission is beamed away from us. In either case, we argue that
47: CXOU~J180351.4-213707 is likely the X-ray counterpart of HESS~J1804-216,
48: based on the fact that the Galactic TeV gamma-ray sources are predominantly
49: SNRs or PWNe. The X-ray spectrum of the source can be fitted well with a
50: power law, although the model is not well constrained due to large statistical
51: uncertainties. The spectrum seems to be very hard, with the best-fit photon
52: index $\sim 1.2$. Under the assumption that CXOU~J180351.4-213707 is
53: the X-ray counterpart of HESS~J1804-216, we attempted to model the X-ray and
54: TeV emission as synchrotron and inverse Compton scattered radiation from
55: relativistic electrons. We briefly discuss the results.
56:
57: \end{abstract}
58:
59: \keywords{acceleration of particles --- gamma rays: theory --- pulsars:
60: individual (PSR J1803-2137) --- radiation mechanisms: non-thermal
61: --- supernova remnants --- X-rays: general}
62:
63: \section{Introduction}
64:
65: One of the most exciting recent advances in high energy astrophysics is the
66: detection of various classes of sources at TeV energies with ground-based
67: gamma ray facilities (for recent reviews see Weekes 2006 and Cui 2006). The
68: established TeV gamma-ray emitters now include blazars, radio galaxies,
69: shell-type supernova remnants (SNRs), pulsar wind nebulae (PWNe),
70: microquasars,
71: and Be binaries, all of which have traditionally been standard targets for
72: X-ray/soft gamma-ray observations. Arguably the most significant discovery
73: is, however, the presence of a population of unidentified TeV gamma-ray
74: sources. These sources are probably Galactic in origin, given their
75: concentration around the Galactic plane (though one must take into account
76: a strong observational bias towards the Galactic plane). Some of them have
77: plausible counterparts at longer wavelengths (Aharonian et al. 2005a, 2006a),
78: based mostly on positional coincidence, but others have none at all
79: (Mukherjee et al. 2003; Butt et al. 2003; Aharonian et al. 2005a, 2006a).
80:
81: A number of proposals have been made on the nature of the unidentified TeV
82: gamma-ray sources. The sources might be associated with old shell-type SNRs
83: (Yamazaki et al. 2006). Such systems could have a very high ratio of TeV to
84: X-ray (or radio) fluxes, which makes them difficult to detect at low energies.
85: For instance, HESS J1813-178 was initially ``dark'' (Aharonian et al. 2005a)
86: but was subsequently identified with a shell-type SNR (G12.8-0.0; Brogan et
87: al. 2005). On the other hand, HESS J1825-137 is now positively associated
88: with a PWN (G18.0-0.7; Aharonian et al. 2006b). Therefore, both SNRs and PWNe
89: are viable candidates. It is worth noting that the plausible counterparts of
90: unidentified TeV gamma-ray sources are mostly shell-type SNRs or PWNe
91: (Aharonian et al. 2005a, 2006a). Other possibilities also exist. For instance,
92: HESS J1303-631 is postulated as the remnant of a gamma-ray burst that occurred
93: some tens of thousands of years ago in our galaxy (Atoyan et al. 2006).
94: TeV~J2032+4130 might be associated with shocks produced by colliding winds of
95: massive stars in the dense Cyg OB2 association (Butt et al. 2003).
96:
97: In this Letter, we report the detection of extended X-ray emission near the
98: radio pulsar PSR~J1803-2137. The pulsar has been suggested as a possible
99: counterpart of HESS~J1804-216, one of the brightest unidentified TeV gamma-ray
100: sources (Aharonian et al. 2005a,2006a). While our results cannot definitively
101: establish a physical connection between the extended emission and the pulsar,
102: they have provided evidence to suggest that the former may be the X-ray
103: counterpart of HESS~J1804-216.
104:
105: \section{Data Analysis and Results}
106:
107: The data for this work were derived from an archival {\it Chandra} observation
108: of PSR J1803-2137 (ObsID \#5590). The data were taken with the ACIS detector,
109: with CCDs I2--3 and S0--3 being read out. The aim point is on the S3 chip
110: (with the default Y offset $\Delta Y$=$-20$\arcsec). The total on-source time
111: is about 30 ks. The data were reduced with the standard {\it CIAO} analysis
112: package (version 3.3), along with CALDB 3.2.0. We followed the {\it CIAO}
113: Science Threads\footnote{see http://asc.harvard.edu/ciao/threads/index.html}
114: in preparing the data (including bad pixel removal and data filtering) and
115: constructing images and spectra. Little data were filtered out in this case,
116: so the effective exposure time is also about 30 ks. We found that the Level 2
117: event file (with a processing version of 7.6.7) from the archive had already
118: incorporated the most updated calibrations, so we used it as a starting point
119: for subsequent imaging and spectral analysis.
120:
121: \subsection{Imaging Analysis}
122:
123: We made an X-ray image of the field in the 0.5--10 keV band and carried out a
124: search for discrete sources. The {\it CIAO} tool {\it celldetect} was used.
125: It is based on the sliding-cell algorithm, but the detection cell is allowed
126: to vary in size to match the local point spread function (PSF). The key
127: parameters are the signal-to-noise (S/N) threshold and the size of the
128: detection cell,
129: which are, by default, set at 3 and 80\% of the encircled energy of the PSF,
130: respectively. We also left all other parameters at the default values. The
131: default settings have been shown to be effective against spurious
132: detections.\footnote{see http://asc.harvard.edu/ciao/download/doc/detect\_manual/cell\_false.html.} We assessed the statistical significance of each
133: detection with Eq. (17) in Li \& Ma (1983), based on the output of
134: {\it celldetect} (namely, the sizes of the source and background regions
135: and the total counts in the two regions). Table~1 shows all detections with
136: significance $\gtrsim 5\sigma$. Seven of the sources are on the S3 chip,
137: two on the S2 chip, two on the S1 chip, and one on the I3 chip.
138:
139: A few remarks on the results are necessary. First of all, only statistical
140: uncertainties are shown for the best-fit positions of the sources. It is
141: known that the systematic uncertainty on the absolute position of a source
142: can be much larger, about 0.6\arcsec\ in radius of the 90\% error circle and
143: 0.8\arcsec\ of the 99\% error circle on average (but larger for sources more
144: than 3\arcmin\ from the aim point)\footnote{See http://asc.harvard.edu/cal/ASPECT/celmon/}. Secondly, for extended sources, {\it celldetect} only determines
145: the coordinates of the centroid. Thirdly, the count rates shown are also from
146: {\it celldetect}, except for CXOU J180351.4-213707, which {\it celldetect}
147: failed to adequately characterize, due to its extended nature. To accurately
148: extract source counts from CXOU J180351.4-213707, we used a circular source
149: region that centered on it with a radius of 10 pixels (or about 5\arcsec)
150: and a concentric background annulus with an inner radius of 10 pixels and
151: an outer radius of 25 pixels (chosen to avoid a possible source that is
152: below our detection threshold; see Fig. 1). The true count rate of the
153: source is shown in Table~1.
154:
155: Among the sources that we detected, CXOU J180351.4-213707 is the only one
156: that is significantly extended. Figure~1 shows an expanded view of the source,
157: along with a broad view of PSR~J1803-2137 and its surroundings. The image has
158: been smoothed (in {\it ds9} ) with a Gaussian kernel that is 3 pixels in
159: radius. CXOU J180351.4-213707 is seen to extend roughly 7\arcsec\ along both
160: right accession and declination. It is about 40\arcsec\ away from
161: PSR~J1803-2137. No X-ray
162: emission from the latter is detected. To be more quantitative, we extracted
163: counts from a circular source region that is 5 pixels in radius and centered
164: at the position of the pulsar and from a concentric background annulus with
165: an inner radius of 10 pixels and outer radius of 40 pixels. The resulting
166: net count rate is $(-5\pm 9)\times 10^{-5}$ $cts$ $s^{-1}$.
167:
168: We searched the {\it SIMBAD} and {\it NED} databases for plausible
169: counterparts of the detected sources. Within a 30\arcsec\ search radius,
170: we found only one candidate, 1WGA J1803.6-2140, which is about 12\arcsec\
171: away from CXOU J180341.5-214034. In addition, CXOU J180432.4-214009 appears
172: to be the same source as Suzaku J1804-2140 (Bamba et al. 2006), which was
173: suggested as a plausible X-ray counterpart of HESS J1804-216, although the
174: source is not obviously extended in our case. We should note that this
175: source falls on the I3 chip, which is quite far from the aim point. No
176: X-ray emission was detected at the position of Suzaku J1804-2142 (Bamba et
177: al. 2006), implying that it is either a transient or highly variable
178: source.
179:
180: \subsection{Spectral Analysis}
181:
182: We used the {\it CIAO} script {\it specextract} to construct the X-ray
183: spectrum of CXOU J180351.4-213707. Here, we adopted the same source
184: region as before but used a much larger background annulus, the outer
185: radius of which is 170 pixels. We excluded a circular region (of 9-pixel
186: radius) that is centered on the possible source to the northwest of
187: CXOU J180351.4-213707, just to be on the cautious side. The script
188: produced both the overall and background spectra, as well as the
189: corresponding {\it rmf} and {\it arf} files that are needed for subsequent
190: spectral modeling.
191:
192: For spectral analysis, we excluded data points below 0.3 keV and above 10
193: keV and then rebinned the raw spectrum so that there are at least 15 counts
194: in each bin. We carried out spectral modeling in {\it XSPEC 11.3.2} (Arnaud
195: 1996). The background-subtracted spectrum is shown in Figure~2. It can be
196: fitted well with a simple absorbed power-law model ({\it wabs*powerlaw} in
197: {\it XSPEC}). The best-fit model and residuals are also shown in Fig.~2. The
198: reduced $\chi^2$ of the fit is 0.57 for 8 degrees of freedom. The derived
199: parameters are: hydrogen column density $N_H=8^{+6}_{-3}\times 10^{21}$
200: $cm^{-2}$, photon index $\Gamma=1.2^{+0.5}_{-0.4}$, and normalization
201: $K=1.1^{+1.2}_{-0.5}\times 10^{-5}$ $ph$ $cm^{-2}$ $s^{-1}$ $keV^{-1}$
202: at 1 keV. The errors shown represent 90\% confidence limits. Although the
203: model is not well constrained, due to large statistical uncertainties,
204: CXOU J180351.4-213707 seems to be a very hard X-ray source. The
205: spatially-averaged
206: (absorbed) flux of the source is $1.0^{+29.3}_{-0.9}\times 10^{-13}$ $erg$
207: $cm^{-2}$ $s^{-1}$ (in the 0.3--10 keV band).
208:
209: \section{Discussion}
210:
211: The {\it Chandra} observation has revealed the presence of eleven discrete
212: X-ray sources in the general vicinity of HESS J1804-216. Figure~3 shows
213: the positions of these sources in galactic coordinates, as overlaid over
214: the TeV gamma-ray image of the field (Aharonian et al. 2006a). While it
215: is difficult to be certain as to which is the counterpart of HESS~J1804-216,
216: we argue that CXOU~J180351.4-213707 is most probable. It is the only
217: significantly extended source detected. It is conceivable that the source
218: might be a previously unknown SNR that is associated with PSR~J1803-2137. If
219: so, the spatial offset between the two could be attributed to the proper
220: motion of the pulsar. At the distance of the pulsar ($\sim 4$ kpc; Taylor
221: \& Cordes 1993), the separation between the two is only about 0.78 pc.
222: Assuming the pulsar is born at the ``center'' of CXOU J180351.4-213707, it
223: would only require a transverse speed of $\lesssim 50$ km/s for it to reach
224: the current position in $\sim$16,000 yrs, which is its characteristic
225: spin-down age ($\equiv \frac{P}{2\dot{P}}$; Clifton \& Lyne 1986). This
226: would be easily achievable.
227: However, the measured size of CXOU~J180351.4-213707 is only about 0.14 pc
228: (for an angular size of $\sim$7\arcsec), which seems to be much too small
229: for an SNR that is over $10^4$ years old.
230:
231: On the other hand, {\em Chandra} might have revealed only the brightest part
232: of the hypothesized SNR and its true size might be much larger. In fact,
233: CXOU~J180351.4-213707 might only represent a bright spot of X-ray emission
234: of such an SNR. This scenario can be tested with a much deeper exposure of
235: the field with {\em Chandra} or {\em XMM-Newton} in the future. It should be
236: noted that PSR J1803-2137 was initially thought to be associated with the
237: SNR G8.7-0.1 (Kassim \& Weiler 1990; see also Fig.~3), based on similar
238: estimated distances and ages
239: between the two. It was, however, subsequently realized that an unusually
240: large transverse velocity ($\sim 1700$ $km$ $s^{-1}$) would be required for
241: the pulsar to reach the current position if it was born at the center of
242: G8.7-0.1. This, along with a smaller revised distance to the pulsar and the
243: lack of evidence for such a large transverse velocity, led to a
244: dis-association between the two systems (Frail et al. 1994; see, however,
245: Finley \& \"{O}gelman 1994). There would be no similar issues with the
246: scenario that we are postulating here, because the pulsar might be even
247: closer to the dynamical center of the hypothesized SNR than to
248: CXOU~J180351.4-213707 (about 0.78 pc, at the distance of PSR~J1803-2137).
249:
250: Alternatively, CXOU~J180351.4-213707 might have no physical connection with
251: PSR~J1803-2137. Instead, it could be a PWN that is powered by a different
252: pulsar whose emission is beamed away from us. Again, the true size of such
253: a PWN could be much larger than what has been measured. For instance,
254: CXOU~J180351.4-213707 might be the bright compact core of the PWN. Such a
255: core is known to exist in several PWNe. As an example, a torus-like compact
256: X-ray emitting region was seen around the Vela pulsar with {\it Chandra}
257: (Helfand et al. 2001) inside a much larger nebula (Markwardt \& \"{O}gelman
258: 1995; Aharonian et al. 2006c). Also, a small extended X-ray feature was
259: detected around PSR~J1826-1334 with {\it XMM-Newton}, as part of a more
260: extended nebula (Gaensler et al. 2003), whose asymmetric profile is the key
261: to establishing a physical connection between this PWN and HESS~J1825-137
262: (Aharonian et al. 2005b). In this case, the X-ray emission is much less
263: extended than the gamma-ray emission, which might be attributable to the
264: difference in the cooling times of the X-ray and gamma-ray emitting
265: electrons (Aharonian et al. 2005b). The same might also be true for
266: CXOU~J180351.4-213707, if it is the X-ray counterpart of HESS~J1804-216.
267: Moreover, PSR~J1826-1334 is offset from the gamma-ray peak of HESS~J1825-137,
268: as would be the case for CXOU~J180351.4-213707/HESS~J1804-216 if the PWN
269: scenario holds.
270:
271: Assuming that CXOU~J180351.4-213707 is the X-ray counterpart of
272: HESS~J1804-216, we proceeded to assemble the spectral energy distribution
273: (SED) of the system. Figure~4 shows the results. Since
274: the X-ray emitting region could be much larger, as already discussed,
275: the measured X-ray fluxes should only be taken as lower limits.
276: We attempted to model the SED in a leptonic scenario, in which the X-ray
277: emission is assumed to originate from the synchrotron radiation from
278: relativistic electrons in the region and TeV gamma-ray emission from
279: inverse Compton (IC) scattering of ambient photons by the same electrons.
280: The spectral distribution of the electrons is assumed to be of the form:
281: $Q_e \propto \gamma^{-s} e^{- \gamma / \gamma_{max}}$. We found a fairly
282: good fit to the X-ray SED with $s=0.4$ and $\gamma_{max}\simeq 10^8$,
283: for a spherical emitting region of radius $R=10^{17}$ $cm$ and magnetic
284: field $B=3\mu G$, using the methodology described by Mastichiadis (1996).
285: The synchrotron self-Compton emission from such an electron spectrum peaks
286: at frequencies well above $10^{28}~\rm Hz$. The position of the IC peak
287: remains essentially unchanged with the addition of the cosmic microwave
288: background (CMB) photons ($U_{CMB} = 0.33~\rm eV~cm^{-3}$), as shown
289: in Fig.~4, in conflict with the measured gamma-ray spectrum. Varying model
290: parameters does not fundamentally improve the situation.
291:
292: On the other hand, we could find a reasonable fit to the TeV gamma-ray
293: spectrum (e.g., with $s \simeq 1.5$, $\gamma_{max} \simeq 10^7$, and
294: $B \simeq 3 \mu G$), as also shown in Fig.~4. However, the required
295: electron spectrum deviates significantly from that needed to explain the
296: X-ray emission. This perhaps argues for a multi-zone scenario.
297: Intriguingly, in this case, an extrapolation of the IC spectrum comes
298: very close to the X-ray measurements (see Fig.~4). Of course, one should
299: always keep in mind the possibility that, e.g., CXOU~J180351.4-213707
300: might be the compact core of a much extended PWN and the true X-ray flux of
301: the PWN might thus be much higher. More sophisticated modeling is beyond
302: the scope of this work.
303:
304: We conclude by noting the lack of X-ray emission from PSR~J1803-2137. It is
305: a bit surprising that this Vela-like pulsar could not be detected in a 30 ks
306: ACIS/{\em Chandra} observation. Given the measured $P$ ($=133$ ms) and
307: $\dot{P}$ ($=1.34 \times 10^{-13}$ $s$ $s^{-1}$), the spin-down power of the
308: pulsar ($\equiv -4\pi^2 I \frac{\dot{P}}{P^3}$, where $I$ is the moment of
309: inertia of the neutron star) is
310: about $2.25\times 10^{36}$ $erg$ $s^{-1}$. Becker \& Tr\"umper (1997) showed
311: that rotation-powered pulsars typically radiate away 0.1\% of the spin-down
312: power in the {\em ROSAT} band (0.1--2.4 keV). So, we would expect a
313: 0.1--2.4 keV luminosity of $2.25\times 10^{33}$ $erg$ $s^{-1}$ for
314: PSR~J1803-2137, or a
315: flux of $1.2\times 10^{-12}$ $erg$ $cm^{-2}$ $s^{-1}$, which is several orders
316: of magnitude higher than the {\em Chandra} ($1\sigma$) upper limit of
317: $1.7\times 10^{-15}$ $erg$ $cm^{-2}$ $s^{-1}$ for the total emission.
318:
319: \begin{acknowledgements}
320: We wish to thank the anonymous referee for his/her very useful comments.
321: This research has made use of the NASA/IPAC Extragalactic Database (NED)
322: which is operated by the Jet Propulsion Laboratory, California Institute
323: of Technology, under contract with the National Aeronautics and Space
324: Administration, and of the Simbad Database.
325: We gratefully acknowledge financial support from the Department of Energy.
326:
327: \end{acknowledgements}
328:
329: \begin{table}
330: \caption{Detected X-ray Sources$^\dag$}
331: \begin{tabular}{lllc}\hline\hline
332: Name & Right Ascension & Declination & Count Rate \\
333: & J2000 & J2000 & 10$^{-3}$ $cts$ $s^{-1}$ \\ \hline
334: %CXOU J180338.1-213656 & 18:03:38.11(1) & -21:36:56.0(2) & 0.65 $\pm$ 0.21 \\
335: CXOU J180341.5-214034 & 18:03:41.54(2) & -21:40:34.8(2) & 1.27 $\pm$ 0.22 \\
336: CXOU J180345.3-213038 & 18 03 45.34(1) & -21 30 38.7(2) & 2.42 $\pm$ 0.37 \\
337: %CXOU J180345.0-214050 & 18:03:45.06(20 & -21:40:50.4(3) & 0.82 $\pm$ 0.26 \\
338: %CXOU J180348.7-213034 & 18:03:48.70(2) & -21:30:34.9(3) & 0.76 $\pm$ 0.24 \\
339: CXOU J180349.1-212317 & 18:03:49.19(3) & -21:23:17.2(4) & 4.34 $\pm$ 0.79 \\
340: CXOU J180349.3-214135 & 18:03:49.35(2) & -21:41:35.8(2) & 1.54 $\pm$ 0.26 \\
341: CXOU J180350.9-213837 & 18 03 50.94(1) & -21 38 37.8(1) & 1.48 $\pm$ 0.25 \\
342: CXOU J180351.4-213707 & 18 03 51.411(4) & -21 37 07.37(5) & 4.25 $\pm$ 0.32 \\
343: CXOU J180355.0-213937 & 18 03 55.00(1) & -21 39 37.4(1) & 1.12 $\pm$ 0.17 \\
344: CXOU J180400.7-214251 & 18 04 00.76(2) & -21 42 51.5(2) & 2.41 $\pm$ 0.32 \\
345: CXOU J180401.2-213153 & 18 04 01.217(5) & -21 31 53.48(7) & 6.23 $\pm$ 0.53 \\
346: CXOU J180404.2-213709 & 18 04 04.27(1) & -21 37 09.5(1) & 0.97 $\pm$ 0.19 \\
347: %CXOU J180405.7-213213 & 18 04 05.79(2) & -21 32 13.6(2) & 0.90 $\pm$ 0.28 \\
348: CXOU J180432.4-214009 & 18 04 32.47(2) & -21 40 09.8(3) & 4.58 $\pm$ 0.45 \\ \hline
349: \end{tabular} \\
350: $^\dag$The coordinates and count rates shown are based on the output of
351: {\it celldetect}, except for CXOU J180351.4-213707 (see text). The numbers
352: in parentheses indicate uncertainty in the last digit. Note that only
353: statistical uncertainties are shown. \\
354: \end{table}
355:
356: \clearpage
357:
358: \begin{references}
359: \reference{}Aharonian, F., et al. 2005a, Science, 307, 1938
360: \reference{}Aharonian, F., et al. 2005b, A\&A, 442, L25
361: \reference{}Aharonian, F., et al. 2006a, ApJ, 636, 777
362: \reference{}Aharonian, F., et al. 2006b, A\&A, in press
363: \reference{}Aharonian, F., et al. 2006c, A\&A, 448, L43
364: \reference{}Atoyan, A., Buckley, J., \& Krawczynski, H. 2006, ApJ, 642, L153
365: \reference{}Arnaud, K.A. 1996, Astronomical Data Analysis Software and
366: Systems V, eds. Jacoby G. and Barnes J., ASP Conf. Series, vol. 101, p.17
367: \reference{}Bamba, A., et al. 2006, PASJ, in press (astro-ph/0608310)
368: \reference{}Becker,~W., \& Tr\"{u}mper,~J. 1997, A\&A, 326, 682
369: \reference{}Brogan, C.L., et al. 2005, ApJ, 629, L105
370: \reference{}Butt, Y. M., et al. 2003, ApJ, 597, 494
371: \reference{}Clifton, T. R., \& Lyne, A. G. 1986, Nature, 320, 43
372: \reference{}Cui,~W. 2006, in Frontier Objects in Astrophysics and Particle
373: Physics, eds. F. Giovannelli \& G. Mannocchi, Italian Physical Society,
374: Editrice Compositori, Bologna, Italy, in press (astro-ph/0608042)
375: \reference{}Finley, J. P., \& \"{O}gelman, H. 1994, ApJ, 434, L25
376: \reference{}Frail, D. A., Kassim, N. E., \& Weiler, K. W. 1994, AJ, 107, 1119
377: \reference{}Gaensler, B. M., \& Schulz, N. S., Kaspi, V. M., Pivovaroff, M. J.,
378: \& Becker, W. E. 2003, ApJ, 588, 441
379: \reference{}Helfand, D. J., Gotthelf, E. V., \& Halpern, J. P. 2001, ApJ, 556,
380: 380
381: \reference{}Kassim, N. E., \& Weiler, K. W. 1990, Nature, 343, 146
382: \reference{}Li, T. P., \& Ma, Y. Q. 1983, ApJ, 272, 317
383: \reference{}Markwardt, C. B., \& \"{O}gelman, H. 1995, Nature, 375, 40
384: \reference{}Mastichiadis, A. 1996, A\&A,305, L53
385: \reference{}Mukherjee, R., Halpern, J. P., Gotthelf, E. V., Eracleous, M.,
386: \& Mirabal, N. 2003, ApJ, 589, 487
387: \reference{}Taylor, J. H., \& Cordes, J. M. 1993, AJ, 411, 674
388: \reference{}Weekes, T.C. 2006, in Energy Budget in the High Energy Universe,
389: Kashiwa, Japan (astro-ph/0606130)
390: \reference{}Yamazaki, R., et al. 2006, MNRAS, in press (astro-ph/0601704)
391:
392: \end{references}
393:
394:
395: %fig.1
396: \begin{figure}
397: \epsscale{1.}
398: \plotone{f1.eps}
399: \caption{X-ray View of PSR J1803-2137 and its surroundings. The image was made
400: in the 0.5--10 keV band. It has been smoothed and is shown on a logarithmic
401: scale. The pulsar is not detected; its radio position is indicated by an
402: open circle. On the other hand, the presence of CXOU J180351.4-213707 is
403: apparent. The inset shows an expanded view of the source, with contours (at
404: the levels of 0.50, 0.87, 1.52,and 2.64 counts) overlaid to show its
405: extension. }
406: \end{figure}
407:
408: %fig.2
409: \begin{figure}
410: \includegraphics[width=4.5in,angle=-90]{f2.eps}
411: \caption{X-ray spectrum of CXOU J180351.4-213707. The solid histogram in the
412: top panel shows the best fit to the data with an absorbed power-law model.
413: The residuals are shown in the bottom panel. }
414: \end{figure}
415:
416: %fig.3
417: \begin{figure}
418: \epsscale{1.}
419: \includegraphics[width=4.5in]{f3.eps}
420: \caption{X-ray sources in the vicinity of HESS J1804-216. The TeV gamma-ray
421: image was adapted from Aharonian et al. 2006a. The dashed circle is meant to
422: show the
423: average extent of the gamma-ray emission region. The positions of the
424: detected X-ray sources are indicated by filled circles. Highlighted are
425: CXOU J180351.4-213707 (in red), which might be the X-ray counterpart of
426: HESS J1804-216, and CXOU J180432.4-214009 (in black), which is likely the
427: same source as Suzaku J1804-2140. The {\it ROSAT} contours of G8.7-0.1
428: (Finley \& \"{O}gelman 1994) are also overlaid. The white circle shows the
429: extent of the 20 cm emission as measured with the {\it VLA}. }
430:
431: \end{figure}
432:
433: %fig.4
434: \begin{figure}
435: \includegraphics[scale=0.75]{f4.eps}
436: \caption{The combined spectral energy distribution of CXOU~J180351.4-213707
437: (in triangles) and HESS~J1804-216 (in open circles). The solid curve shows
438: representative results from leptonic calculations that aim at fitting
439: the X-ray data, while the dashed curve shows those that fit the gamma-ray
440: data (see text).
441: }
442: \end{figure}
443:
444:
445: \end{document}
446: