1: \documentclass[12pt,preprint]{aastex}
2: \usepackage{graphicx,lscape}
3: \newcommand\ea{\it et al. \rm}
4: \newcommand\exo{{\sl EXOSAT }}
5: \newcommand\ein{{\sl Einstein}}
6: \newcommand\uhu{{\sl Uhuru }}
7: \newcommand\ros{{\sl ROSAT }}
8: \newcommand\cha{{\sl Chandra}}
9: \newcommand\asca{{\sl ASCA}}
10: \newcommand\xmm{XMM-{\sl{Newton}}}
11: \newcommand\ergsec{\hbox{erg s$^{-1}$}}
12: \newcommand\ergcm{\hbox{erg cm$^{-2}$ s$^{-1}$}}
13: \newcommand\kevcs{\hbox{keV cm$^{-2}$ s$^{-1}$}}
14: \newcommand\am{$^{\prime}$ }
15: \newcommand\as{$^{\prime\prime}$}
16: \shorttitle{Extended emission around PSR B0355+54}
17: \shortauthors{Tepedelenlio\v{g}lu \& \"{O}gelman}
18:
19: \begin{document}
20: \title{\bf Discovery of extended emission around the pulsar B0355+54}
21:
22: \author{E. Tepedelenlio\v{g}lu\altaffilmark{1}
23: \email{emre@cow.physics.wisc.edu}
24: \and
25: H. \"{O}gelman\altaffilmark{1,2}}
26: \email{ogelman@cow.physics.wisc.edu}
27: \altaffiltext{1}{Department of Physics, University of
28: Wisconsin-Madison, 1150 University Ave., Madison, WI 53703, USA}
29: \altaffiltext{2}{Faculty of Engineering and Natural Sciences,
30: Sabanc\i University, Orhanl\i Tuzla, \.{I}stanbul 34956, Turkey}
31:
32:
33: \begin{abstract}
34: PSR B0355+54 is one of the handful of pulsars that has been observed
35: with both \cha\ and \xmm. The analysis of the archival data has
36: revealed the pulsar and a $\sim30$\arcsec\ compact nebula
37: surrounding it. \xmm\ also has detected a trail which extends
38: $\sim$6\arcmin\ and similar to the compact nebula is also
39: counter-aligned with the proper motion of the pulsar. The spectrum
40: of both the pulsar and the extended emission are well described by
41: an absorbed power-law model. The measured flux corresponds to an
42: efficiency of converting the spin-down luminosity into X-rays in the
43: 2-10 keV band of $\sim$0.01\% and $\sim$1\% for the pulsar and
44: the extended emission, respectively. From the \xmm\ data we have
45: detected pulsations at the expected radio period. The energetic and
46: the extent of the extended emission can be explained by a Bow-shock
47: formed by the motion of the pulsar through the interstellar medium.
48:
49: \end{abstract}
50:
51: \keywords{pulsars:individual (\objectname{PSR B0355+54}) ---
52: stars:neutron --- X-rays:stars}
53:
54:
55: \section{Introduction}
56:
57: PSR B0355+54 is a middle age ($\tau\sim$0.56 Myr) pulsar with a
58: spin-down energy loss rate of $\dot{E}=4.5\times 10^{34}$ \ergsec,
59: and a relatively close distance of $d=1.04^{+0.21}_{-0.16}$ kpc
60: \citep{chatterjee}. It was observed with \ros\ for a total of 19 ks
61: and was detected with a non-thermal luminosity (2-10 keV) of
62: $1.2\times 10^{31}$ \ergsec\ \citep{slane}. Taking into account the
63: extended emission discussed in this paper, one should consider this
64: as only an upper-limit to the pulsar's X-ray emission, considering
65: \ros\ does not have a sufficient angular resolution to resolve these
66: features.
67:
68: Pulsar wind nebulae (PWNe) are useful probes of pulsars and their
69: surroundings. In the case of older ($\tau>10^{5}$ years) isolated
70: neutron stars such as Geminga the PWN is thought to be the Bow-shock
71: created by the pulsar proper motion through the interstellar medium
72: (ISM) \citep{deluca,caraveoa}. The distance, the energy loss rate,
73: and the proper motion of some of the pulsars is well known through
74: radio observations. The knowledge of such parameters make pulsar
75: Bow-shocks particularly appealing to study. In this letter we report
76: the discovery of an extended emission associated with the pulsar
77: B0355+54, and investigate the possible implications.
78:
79:
80:
81: \section{Observations}\label{observations}
82: \subsection{\cha}\label{chandra}
83:
84: PSR B0355+54 was observed using the Advanced CCD Imaging
85: Spectrometer (ACIS) instrument aboard the \cha\ X-ray Observatory on
86: 2004 July 16. Data were collected in the nominal timing mode, with
87: 3.241 s exposures between CCD readouts, and "FAINT" spectral
88: mode. The standard \cha\ screening criteria produced a total usable
89: exposure time of 67.13 ks. The ACIS image reveals a point source
90: near the pulsar radio position and a relatively faint diffuse
91: emission that extends $\sim30$\arcsec\ towards the south-east side,
92: which we label as the compact nebula (CN). The detected point
93: source is located at $\alpha$=$03^{\mbox{h}}58^{\mbox{m}}53\fs72$,
94: $\delta$= +54\arcdeg13\arcmin13\farcs9 (J2000.0) which is only
95: 0\farcs2 away from the proper motion corrected pulsar radio
96: position. This source is detected with an ACIS-S countrate of
97: 0.0031$\pm$0.0002 cps in the 0.2-10 keV band. The errors quoted in
98: this work are in the 1$\sigma$ confidence range, unless otherwise
99: noted.
100:
101: After removing the contributions from the point sources, we smoothed
102: the image to examine the extended emission. The resulting image is
103: given in Figure~\ref{smoothimage}. As can be seen, there is excess
104: emission towards south-east of the pulsar. We compared the surface
105: brightness of this extended emission to that of the background's and
106: calculated the countrate of the extended emission from an annulus
107: situated at the pulsar position that extends from 3\arcsec to
108: 35\arcsec. The background countrate was measured from an annulus
109: that extends from 37\arcsec\ to 70\arcsec. This annulus was chosen
110: to be co-centric with the extraction region used for the extended
111: emission. The observed extended emission gives an excess of
112: 714$\pm$69 counts over the background. This corresponds to a net
113: surface brightness of 670$\pm$61 counts arcmin$^{-2}$. We also
114: compared the background used for this analysis to a region without
115: any sources taken away from the pulsar, towards the chip
116: boundary. This resulted in no discrepancies between the surface
117: brightness for the two regions.
118:
119: \subsection{\xmm}\label{xmm}
120:
121: PSR B0355+54 was observed with \xmm\ on 2002 February 10 for 29 ks
122: and 28 ks with EPIC-MOS and EPIC-pn instruments, respectively. The
123: EPIC-pn EPIC-MOS1, and EPIC-MOS2 instruments were operated in the
124: small-window, full-window (imaging), and timing modes which provide
125: a temporal resolution of 6 ms, 2.6 s, and 1.5 ms, respectively. The
126: medium filters were used for MOS, the thin filter was used on
127: pn. The EPIC-pn and -MOS1 cameras detect a point a source consistent
128: with the \cha\ and the radio position.
129:
130: The CN detected with \cha\ is also visible in the \xmm\ data (see
131: Figure~\ref{smoothimage}). Although \xmm\ data has resolved the
132: extended emission it is difficult to perform any spectral analysis
133: on the CN due to relatively low spatial resolution. Consequently,
134: this region in EPIC-MOS data (similar for EPIC-pn) contains
135: significant amount of counts from the pulsar. For example, in an
136: annular region that is centered on the source of interest and
137: extends from 10\arcsec-35\arcsec\ contains $\sim$35\% of the pulsar
138: counts.
139:
140: We should also note that in the \xmm\ data only there is extended
141: emission that extends up to $\sim 7$\arcmin\ that is in the opposite
142: direction of the proper motion, we label this feature as the trail
143: (see Figure~\ref{smoothimage}). This structure is aligned with the
144: CN and at an initial glance seems fainter. To quantify this, we
145: extracted counts from an elliptical region that is aligned with the
146: pulsar's proper motion and is 3\farcm1 wide and 5\farcm8 long which
147: was located 1\farcm5 south-east of the pulsar. This region contained
148: 316$\pm$48 background-subtracted source counts corresponding to a
149: surface brightness of 22.7 counts arcmin$^{-2}$. When we analyzed
150: the Chandra data no such emission was detected. This is not
151: surprising since \cha\ is less sensitive with respect to \xmm. For
152: example, the \cha\ observation of Geminga pulsar was unable to
153: confirm the two tails detected with the \xmm\ observatory
154: \citep{deluca,pavlov}.
155:
156: \section{Timing}
157:
158: PSR B0355+54 has a period of 0.156 s. When \cha\ ACIS is operated in
159: imaging mode its temporal resolution is 3.2 seconds which is not
160: sufficient to resolve the periodicity of the pulsar. When operated
161: in the small window mode \xmm\ EPIC-pn camera has a timing
162: resolution of 6 ms. This becomes 1.5 ms for the EPIC-MOS2 which was
163: operated in the timing mode, where imaging is made only in one
164: dimension to obtain a higher temporal resolution.
165:
166: In order to check for any pulsations we first extracted source
167: counts from a circle centered on the pulsar with a radius
168: 20\arcsec. Such a small radius was used to reduce the background due
169: to the unpulsed extended emission surrounding the pulsar. The
170: extracted event times were barycenter corrected. We then calculated
171: the value of the $Z_{1}^{2}$ \citep{buccheri} at the predicted
172: pulsar frequency (see Table~\ref{tab1}). The obtained value of 7.9
173: corresponds to a probability of chance occurrence of 0.019. We also
174: applied the same criteria to the EPIC-MOS2 data. Since the majority
175: of the data in the vicinity of the pulsar is below 0.2 keV we took
176: all photons with energies below 2 keV. The $Z_{1}^{2}$ for this set
177: at the expected frequency was 17.5, which has a probability of
178: 0.00016 of occurring by chance.
179:
180:
181: At the radio period we constructed the pulse profiles for both MOS2 and
182: pn event times (Figure~\ref{pf}). Both profiles show a double peak
183: nature, however the smaller peak is not significant, specifically in
184: the MOS2 data. As was discussed by \citet{beckera} this kind of a
185: pulse profile would not be observed from a neutron star emitting
186: from its hot polar caps. Together with spectral data this is
187: suggestive of magnetosphere being the main source of X-ray emission
188: for PSR B0355+54.
189:
190:
191:
192: \section{Spectral analysis}
193: \subsection{Pulsar}
194: For the \cha\ observation the pulsar's energy spectrum was extracted
195: from a 2\arcsec\ circular region centered on the pulsar
196: position. Whereas the background was extracted from an annular
197: region extending from 37\arcsec\ to 70\arcsec. The reason for
198: selecting the background away from the pulsar is due to the extended
199: emission surrounding the immediate vicinity of the pulsar. The
200: source region contains $\ga$95\% of the source counts.
201:
202: The spectrum extracted from the MOS1 data was done by selecting all
203: events detected in a circle of radius 20\arcsec\ centered on the
204: pulsar position. Using the \xmm\ EPIC-MOS model point spread
205: function (PSF), 75\% of all source counts are within this
206: region. The reason for selecting such a small source region, and in
207: turn fractional encircled energy, was to minimize the contamination
208: due to the events originating from the extended emission. The
209: background spectrum for the MOS1 data was extracted from an annular
210: region centered on the pulsar position which extends from
211: 55\arcsec-200\arcsec. For the EPIC-pn data we also used an
212: extraction radius of 20\arcsec\ centered on the pulsar. This
213: selection radius includes 75\% of the point-source flux. Since the
214: point source was close to the edge of the CCD, we extracted the
215: background spectrum from a source free region around the pulsar from
216: a circular region with a radius of 45\arcsec. The extracted spectra
217: for PSR B0355+54 was binned so as to have 15, 25 and 20 counts per
218: bin, for ACIS, EPIC-pn and EPIC-MOS, respectively.
219:
220: Among the single-component spectral models fitted to \cha\ spectra a
221: power-law model was found to give the best ($\chi^{2}$=7.6 for 9
222: degrees of freedom [dof]) representation of the observed energy
223: spectrum of the pulsar. A single blackbody model for the pulsar gave
224: neither a statistically nor a physically acceptable fit
225: ($\chi^{2}$=21.9 for 9 dof). The parameters for the best fit
226: power-law model are given in Table~\ref{tab2}. The value obtained
227: for the absorbing column is consistent with the value derived
228: through the basic assumption that there is 10 neutral hydrogen atoms
229: for each free electron, giving a value of $N_{\mbox{H}}\sim 2\times
230: 10^{21}$ cm$^{-2}$. The measured X-ray luminosity corresponds to an
231: efficiency of converting the spin-down luminosity into X-rays in the
232: 2-10 keV band of 0.009\%.
233:
234: The EPIC-MOS and -pn spectra were fit simultaneously to get the best
235: fit parameters. As was done for \cha\ spectrum we tried the
236: single-component models, power-law and blackbody. It was found that
237: absorbed power-law model gave the best representation of the data
238: (see Table~\ref{tab2}). Single blackbody model did not give a
239: statistically acceptable ($\chi^{2}$=51.6 for 33 dof) spectral
240: fit. We also fitted an absorbed power-law model all pulsar spectra
241: (ACIS, EPIC-MOS and -pn) simultaneously. This fit yielded a higher
242: photon index but the best values for the parameters are consistent
243: with single \cha\ and simultaneous \xmm\ fits (see
244: Table~\ref{tab2}).
245:
246:
247: \subsection{Extended emission}
248: The spectrum for the CN was selected from an annular region
249: co-centric with the source region that has inner and outer radii
250: 3\arcsec\ and 35\arcsec, respectively. The background for this
251: spectrum was chosen to be the same as the pulsar's spectrum. For the
252: analysis of the trail we extracted the spectrum from the region
253: described in $\S$~\ref{xmm}. The background for this spectrum was
254: extracted from an off-source region south-west of the trail.
255:
256: The extracted spectra for the CN and the trail were binned so as to
257: have 25 and 50 counts per bin, respectively. We limited the
258: generated spectra to the energy interval 0.2-10 keV. The absorbed
259: power-law model gave a statistically acceptable fit with
260: $\chi^{2}$=24.7 for 42 dof for the CN and $\chi^{2}$=15.9 for 22 dof
261: for the trail. Table~\ref{tab2} lists the best fit parameter values
262: for the extended emission. The measured values given in
263: Table~\ref{tab2} for the luminosities of both the CN and the tail
264: account for only 0.74\% of the pulsar spin-down. We also tried a
265: thermal bremsstrahlung model to fit the trail. This model also gave
266: a statistically acceptable fit with a $\chi^{2}$=16.1 for 25
267: dof. The characteristic temperature of the region is
268: $kT=6.5^{+18.3}_{-3.0}$ keV. The absorbing column is consistent with
269: that of the power-law fit, $N_{\mbox{H}}=4.4^{+2.2}_{-1.4}\times
270: 10^{21}$ cm$^{-2}$. The resulting emission measure ($\int
271: n_{e}n_{H}dV$) is $\sim 1.9\times 10^{54}$ cm$^{-3}$, at a distance
272: of 1.04 kpc.
273:
274:
275:
276: \section{Discussion}
277: The \cha\ and \xmm\ observations of PSR B0355+54 have proven the presence of extended emission associated with the pulsar.
278: This emission has two components which are $\sim$30\arcsec\ and
279: $\sim$6\arcmin\ long (see $\S$~\ref{observations}). The recently
280: detected proper motion direction (see Table~\ref{tab1}) being
281: counter aligned with the trail together with the observed spectrum
282: being well fitted with a power-law model suggests that the observed
283: extended emission may be the result of synchrotron radiation from a
284: wind-driven nebula supported by the ram pressure created by the
285: proper motion through the ISM. This hypothesis have been found to be
286: consistent with the size and energetics of the X-ray trails of
287: Geminga \citep{caraveoa} and PSR B1929+10 \citep{beckera}.
288:
289: There are two relevant time scales that need to be considered. The
290: first is the $\tau_{flow}$, the timescale for the passage of the
291: pulsar over the length of its X-ray trail. The other one is
292: $\tau_{syn}$, the synchrotron lifetime of the radiation. This is
293: given by the equation \citep[see e.g.][]{gaensler}
294: $$ \tau_{syn} =39 B_{\mu G}^{-3/2} \left( \frac{h\nu}{\mbox{keV}}
295: \right)^{-1/2} \mbox{kyr},
296: $$ where $B_{\mu G}$ is the magnetic field in the
297: emission region in units of $\mu$G. The first one of these can be
298: estimated through the relation $\tau_{flow}=r_{t}/v_{p}$, where
299: $r_{t}$ is the extent of the trail which is measured to be
300: $\sim6$\arcmin. This at a distance of 1.04 kpc corresponds to a
301: linear distance of 1.8 pc. The proper motion, $v_{p}$, is measured
302: by \citet{chatterjee} and is 61$^{+12}_{-9}$ km s$^{-1}$. This,
303: together with the linear extent of the trail, gives the time it took
304: PSR B0355+54 to travel across the trail as $\sim 34$ kyr. Assuming
305: these two time scales are comparable the estimate for the magnetic
306: field around the emitting region is $B_{\mu G}\sim$1 for an X-ray
307: photon of energy $h\nu=1$ keV. We should note that this
308: calculation ignores the possible inclination of the pulsar's proper
309: motion with respect to the line of sight.
310:
311: Assuming that the ISM in the vicinity of the pulsar carries a
312: magnetic field that does not differ significantly from that of the
313: average of our galaxy which is $\sim 4$ $\mu$G \citep{beck}. This
314: value should be the characteristic magnetic field for the trail
315: where the X-ray emitting particles are no longer in the shocked
316: region. For the CN, where the electrons are in the shocked region,
317: we need to consider the change in density due to the shock and hence
318: the change in the magnetic field. We can use standard hydrodynamical
319: arguments to show that the density in the shock region is given by
320: the relation
321: $\rho_{\mbox{\small{shock}}}=4\rho_{\mbox{\small{ISM}}}$, which
322: would imply a magnetic field of
323: $B_{\mbox{\small{shock}}}=4B_{\mbox{\small{ISM}}}$ or
324: $B_{\mbox{\small{shock}}}=16$ $\mu$G, given the average ISM magnetic
325: field. This value, given the large uncertainties, is in agreement
326: with the value found here.
327:
328: The termination shock radius is given by the balance of the ram
329: pressure between the wind particles and the ISM,
330: i.e. $$\frac{\dot{E}}{4\pi c R_{s}^{2}}=\frac{1}{2}\rho v_{p}^{2},$$
331: where $\rho$ is the density of the ISM and $v_{p}$ is the proper
332: motion velocity. From here one can estimate the radius of the
333: termination shock as $R_{s}\sim3\times
334: 10^{6}\dot{E}_{34}^{1/2}n^{-1/2}v_{p,100}^{-1}$ cm, where $n$ and
335: $v_{p,100}$ are the number density in the ISM in units of cm$^{-3}$
336: and the pulsar space velocity in units of 100 km s$^{-1}$,
337: respectively. Adopting $n=1$ cm$^{-3}$ the shock termination radius
338: becomes 1.1$\times 10^{17}$ cm. This corresponds to an angular size
339: of $\sim$7\arcsec\ at a distance of 1.04 kpc. This value is
340: consistent with the data given the observational constraints.
341:
342: In order to check the energy dependence of the extent of the CN we
343: constructed two background-subtracted linear profiles along the
344: proper motion direction, which contain photons from the intervals
345: 0.2-2 keV and 2-10 keV (Figure~\ref{radial}). For comparison also
346: the point source contribution is plotted. Although the structure of
347: the CN is not identical in both energy ranges the overall extent
348: within the statistical uncertainties are equal. This is consistent
349: with the identification of this region as the termination
350: shock. This region is expected to show a constant X-ray spectrum
351: across its extent. Due to the pressure difference between the
352: regions ahead and behind the pulsar's motion the termination shock
353: is not of uniform radius around the pulsar and is extended along the
354: proper motion direction \citep{bucciantini}. The larger structure
355: seen by \xmm\ (trail) can be thought of as the emission originating
356: from the wind particles that flow around the edge of the shocked
357: region. The magnetic field in this region should be an order of
358: magnitude less than the nebular magnetic field in the shocked
359: region, if equipartition holds. The two components of the extended
360: emission could be explained by the Bow-shock theory.
361:
362:
363: PSR B0355+54 is yet another pulsar with known proper motion and a
364: collimated outflow in the opposite direction. Such flow, also
365: observed in Geminga, could be interpreted as a pulsar jet
366: \citep{pavlov}. This is reminiscent of the argument proposed for the
367: Vela pulsar \citep{markwardt} where the outflow in opposite
368: direction is proposed as the mechanism to boost pulsar velocities in
369: contrast to birth-kicks.
370:
371:
372: \begin{thebibliography}{}
373: \bibitem[Beck \ea(2003)]{beck} Beck, R., Shukurov, A., Sokoloff, D.,
374: \& Wielebinski, R. 2003, \aap, 411, 99
375: \bibitem[Becker \ea(2005a)]{beckera} Becker, W., Kramer, M.,
376: Jessner, A., \ea, 2005, submitted to \apj, astro-ph/0506545
377: \bibitem[Buccheri \ea(1983)]{buccheri} Buccheri, R., Bennett, K.,
378: Bignami, G. F., \ea, 1983, \aap, 128, 245
379: \bibitem[Bucciantini \ea(2005)]{bucciantini} Bucciantini, N., Amato,
380: E., \& Del Zanna, L., 2005, \aap, 434, 189
381: \bibitem[Caraveo \ea(2003)]{caraveoa} Caraveo, P. A., Bignami,
382: G. F., DeLuca, A., \ea, 2003, Sci., 301, 1345
383: \bibitem[Chatterjee \ea(2004)]{chatterjee} Chatterjee, S., Cordes,
384: J.M., Vlemmings, W.H.T., \ea, 2004, \apj, 604, 339
385: \bibitem[De Luca \ea(2005)]{deluca} De Luca, A., Caraveo, P.A.,
386: Mattana, F., \ea, 2005, to be published in \aap, astro-ph/0511185
387: \bibitem[Duncan \ea(1995)]{dunkan}Duncan, A.R., Haynes, R.F.,
388: Stewart, R.T., \ea, 1995, \mnras, 227, 319.
389: \bibitem[Gaensler \ea(2002)]{gaensler} Gaensler, B.M., Arons, J.,
390: Kaspi, V. M., \ea, 2002, \apj, 569, 878
391: \bibitem[Markwardt \& \"{O}gelman(1995)]{markwardt} Markwardt, C.B., \&
392: \"{O}gelman, H., 1995, Nature, 375, 40
393: \bibitem[Pavlov \ea(2005)]{pavlov} Pavlov, G.G., Sanwal, D., \&
394: Zavlin, V.E., 2005, submitted to \apjl, astro-ph/0511364
395: \bibitem[Slane(1994)]{slane} Slane, P., 1994, \apj, 437, 458
396:
397: \end{thebibliography}
398:
399: \clearpage
400:
401: \input{tab1.tex}
402:
403: \clearpage
404:
405: \begin{figure}
406: \includegraphics[width=16cm]{fig1.ps}
407: \caption{Gray-scale plots of the 0.2-10 keV image centered on
408: the pulsar PSR B0355+54 ({\it{left}}) by \xmm\ EPIC-MOS and
409: ({\it{right}}) \cha\ ACIS. The X-ray point source contribution
410: is removed, the images are smoothed with a Gaussian of width
411: 2\arcsec. Also shown are the isophoto contours. On the left
412: panel the box shows the zoomed in region shown on the right
413: panel (\cha\ image). The direction of the pulsar proper motion
414: is shown on the right panel with a black arrow and the pulsar
415: position is indicated by a white cross ($\times$). The length of
416: the proper motion vector is the distance the pulsar would travel
417: in 1000 years.}
418: \label{smoothimage}
419: \end{figure}
420:
421: \clearpage
422:
423: \begin{figure}
424: \includegraphics[width=11cm,angle=-90]{f2.eps}
425: \caption{X-ray pulse profiles of PSR B0355+54, constructed from
426: EPIC-MOS2 {\it{top}} and EPIC-pn {\it{bottom}}. The pulse
427: fractions are given in the figure. Notice the double-peak nature
428: of the pulse profiles. The peaks are $\sim$0.3 of a period
429: apart. Two cycles are shown or clarity.}
430: \label{pf}
431: \end{figure}
432: \clearpage
433:
434: \input{tab2.tex}
435:
436: \clearpage
437:
438: \begin{figure}
439: \includegraphics[width=12cm,angle=-90]{f3.eps}
440: \caption{The background-subtracted linear profile of the CN, along
441: the proper motion direction as seen by \cha\ in the energy range
442: 0.2-2 keV ({\it{filled circles}}) and 2-10 keV
443: ({\it{diamonds}}). The solid line is the contribution from the
444: pulsar B0355+54 in the 0.2-2 keV range.}
445: \label{radial}
446: \end{figure}
447:
448:
449:
450: \end{document}
451: