1: %\documentclass[12pt,preprint]{aastex}
2: %\documentclass[manuscript]{aastex}
3: %\documentclass[preprint2]{aastex}
4: %\documentclass[preprint2,longabstract]{aastex}
5: \documentstyle[emulateapj,psfig]{article}
6: %\def\astroph{1}
7:
8: %% Use these for ApJL submission
9: %\ifnum\astroph=0
10:
11: %\documentclass[12pt,preprint]{aastex} % Single column, single spaced
12: %\documentclass[manuscript]{aastex} % Single column, double spaced
13: %\documentclass[preprint2]{aastex} % Double column, single spaced
14:
15: %% Use these for ApJL emulation, preparing for astro-ph
16: %\else
17: %\documentclass{emulateapj}
18: %\fi
19: %\usepackage{epsfig}
20:
21:
22: \newcommand{\vdag}{(v)^\dagger}
23: \newcommand{\be}{\begin{equation}}
24: \newcommand{\ee}{\end{equation}}
25: \newcommand{\ba}{\begin{eqnarray}}
26: \newcommand{\ea}{\end{eqnarray}}
27: \newcommand{\siml}{\lower4pt \hbox{$\buildrel < \over \sim$}}
28: \newcommand{\simg}{\lower4pt \hbox{$\buildrel > \over \sim$}}
29: \def\paren#1{\left( #1 \right)}
30: \def\brace#1{\left\{ #1 \right\}}
31: \def\bra#1{\left[ #1 \right]}
32: \def\angle#1{\left\langle #1 \right\rangle}
33:
34: \def\xmm{{\it XMM}-Newton}
35:
36: \slugcomment{Submitted to ApJ Letters}
37:
38: %\shorttitle{XMM-Newton observation of PSR~B0943+16}
39: %\shortauthors{Zhang, Sanwal \& Pavlov}
40:
41: \begin{document}
42:
43: \title{XMM-Newton observation of
44: the drifting pulsar B0943+10}
45:
46: \author{Bing Zhang$^1$, Divas Sanwal$^2$ \& George G. Pavlov$^2$}
47: \affil{$^1$ Department of Physics, University of Nevada, Las Vegas,
48: Las Vegas, NV 89154; bzhang@physics.unlv.edu \\
49: $^2$ Department of Astronomy \& Astrophysics, Pennsylvania State
50: University, 525 Davey Lab., University Park, PA 16802;
51: divas@astro.psu.edu, pavlov@astro.psu.edu}
52:
53: \begin{abstract}
54: Radio pulsar subpulse drifting has been interpreted as rotation of
55: sub-beams (sparks) of pair plasma produced by intermittent breakdowns
56: of an inner vacuum gap above the pulsar polar cap. This model also
57: predicts strong thermal X-ray emission from the polar cap caused by
58: inflowing particles created in spark discharges. We have observed the
59: best-studied drifting pulsar B0943+10 with {\sl XMM-Newton} and
60: detected a point source coincident with the radio pulsar position.
61: Its spectrum could be fitted with a thermal blackbody model, although
62: a power-law model is also acceptable. The thermal fit gives a
63: bolometric luminosity $L_{\rm bol}\approx 5 \times 10^{28}\,{\rm
64: erg\,s^{-1}}$ and a surface area $A \approx 10^3 (T/3\,{\rm
65: MK})^{-4}\,{\rm m}^2$, much smaller than the conventional polar cap
66: area, $6\times 10^4\,{\rm m}^2$. Such thermal
67: radiation can be interpreted as emitted from footprints of sparks
68: drifting in an inner gap of a height $h \sim0.1$--$0.2\, r_{\rm pc}$, where
69: $r_{\rm pc}$ is the polar cap radius. However, the original vacuum
70: gap model by Ruderman and Sutherland requires some modification
71: to reconcile the X-ray and radio data.
72: \end{abstract}
73:
74: \keywords{pulsars: individual (PSR B0943+10) - stars: neutron -
75: X-rays: stars}
76:
77: \section{Introduction}
78:
79: The mechanism of radio pulsar emission remains a mystery after decades
80: of study (Melrose 2004). It is generally agreed that there exists an
81: inner magnetospheric charge acceleration region (or inner gap) near
82: the pulsar polar cap. Ruderman \& Sutherland (1975, hereafter RS75)
83: proposed that the gap is nearly vacuum, based on the assumption that
84: ions cannot be stripped off the surface by the electric field
85: component parallel to the magnetic field lines. Such an assumption
86: was disputed by later workers (e.g., Jones 1986; K\"ossl et al.\
87: 1988), who argued that the surface charges can be freely supplied into
88: the magnetosphere.
89: The inner accelerator flow is therefore expected to
90: be space charge limited (Arons \& Scharlemann 1979; Harding \&
91: Muslimov 1998,2001,2002; Harding et al.\ 2002). Nevertheless, the
92: vacuum gap model is widely used to interpret radio data. Suggestions
93: to amend the binding energy problem have been made (e.g., Usov \&
94: Melrose 1996; Xu et al.\ 1999; Gil \& Melikidze 2002), and some radio
95: emission models heavily rely upon the existence of such an inner
96: vacuum gap (e.g., Qiao \& Lin 1998; Melikidze et al.\ 2000). To
97: resolve this controversy, one should look for independent
98: observational signatures of the vacuum gaps.
99:
100: A distinguishing property of the vacuum gap model is its high polar
101: cap heating rate. In this model, the vacuum gap breaks down
102: intermittently due to pair production discharges (RS75), and the
103: number of the outflowing particles is comparable to that of the
104: inflowing particles, which results in a high luminosity of polar caps
105: heated up to X-ray temperatures (e.g., Zhang et al.\ 2000). On the
106: other hand, the space-charge-limited flow (SCLF) model predicts a much
107: lower polar cap heating rate because only a small fraction of the
108: positrons (for an electron accelerator) turn around and bombard the
109: surface (Arons \& Scharlemann 1979; Zhang \& Harding 2000; Harding \&
110: Muslimov 2001,2002). Therefore, measuring the thermal X-ray luminosity
111: from heated polar caps can provide a clue to verify the existence of
112: the vacuum gaps.
113:
114: The strongest support for the vacuum gap model comes from its ability
115: to explain the regular sub-pulse drifting observed in some
116: long-period, old pulsars with the so-called ``conal'' emission beam
117: (Rankin 1986, and references therein). The phenomenon is naturally
118: interpreted as the ${\bf E \times B}$ drift of polar cap ``sparks''
119: circulating around the magnetic pole (RS75). One of the best-studied
120: drifting pulsars is PSR B0943+10 ($P=1.10$ s, $\dot{E}=1.0\times
121: 10^{32}$ erg s$^{-1}$, $\tau = 5.0$ Myr, $B_p=4.0\times 10^{12}$ G).
122: Deshpande \& Rankin (1999, 2001) monitored the drifting pattern and
123: revealed the ``polar cap map'' of this pulsar. They identified 20
124: sparks rotating counterclockwise with a period of 37 rotation periods
125: (i.e., $\hat P_3 \simeq 37 P$, where $P$ is the spin period, and $\hat
126: P_3$ is the minimum period for the drifting pattern to repeat itself),
127: a conclusion confirmed by an analysis of low-frequency radio data
128: (Asgekar \& Deshpande 2001). The result was claimed to be consistent
129: with the RS75 model (e.g., Deshpande \& Rankin 1999, 2001; Gil \&
130: Sendyk 2003), although some modifications to the model may be needed
131: (Gil et al.\ 2002; Asseo \& Khechinashvili 2002; Gil et al.\ 2003).
132:
133: According to the Galactic electron density model (Cordes \& Lazio
134: 2002), the distance to PSR~B0943+10 determined from its dispersion
135: measure (DM$ =15.35 ~{\rm pc\,cm^{-3}}$) is $0.63\pm 0.10$ kpc. The
136: expected high X-ray luminosity and the small distance make it the best
137: target to test the predictions of the vacuum gap polar cap heating
138: model. Here we report on observations of PSR B0943+10 with {\sl
139: XMM}-Newton, aimed at detecting the thermal X-ray emission from the
140: polar cap region.
141:
142: \begin{figure*}[ht]
143: \hskip 0.3cm
144: %\psfig{figure=bbcont2.ps,height=6.9cm}
145: \psfig{figure=bbcont.ps,height=6.9cm,angle=-90}
146: %\psfig{figure=plcont2.ps,height=6.9cm}
147: \psfig{figure=pl_new2.ps,height=6.9cm}
148: \caption{\footnotesize{
149: Confidence contours (68\%, 90\%, and 99\%) for the blackbody (left)
150: and power-law (right) model fits to the EPIC-pn spectrum of PSR
151: B0943+10. The BB normalization (vertical axis) is the projected
152: emitting area in units of m$^2$, for $d=630$ pc. The PL normalization
153: is in units of $10^{-6}$ photons cm$^{-2}$ s$^{-1}$ keV$^{-1}$ at 1
154: keV. The lines of constant bolometric luminosity (left panel; for a polar
155: cap close to the center of stellar disk, in units of $10^{28}$ erg
156: s$^{-1}$) and constant flux (right panel; in units of $10^{-15}$ erg
157: cm$^{-2}$ s$^{-1}$) are plotted as dashed lines, for fixed $n_{\rm H}
158: =4.3\times 10^{20}$ cm$^{-2}$.
159: }}
160: \label{fig1}
161: \vskip -0.2cm
162: \end{figure*}
163:
164: \section{Observations and Data Analysis }
165:
166: PSR B0943+10 was observed twice with the European Photon Imaging
167: Camera (EPIC) aboard \xmm. The EPIC-pn (hereafter PN) and two
168: EPIC-MOS (hereafter MOS) detectors were operated in extended full
169: frame and full frame modes, respectively. Thin filters were used for
170: all the three EPIC detectors. The initial observation of 2003 May 7
171: was interrupted by a strong solar flare. The exposure time for this
172: observation, compromised by a high background, was 15 ks. Second
173: observation was taken on 2003 December 2--3, for 20 ks. We processed
174: the data using \xmm\ Science Analysis Software (SAS), v.\ 6.0. We
175: removed the intervals of high background and retained only patterns
176: 0--4 and 0--12 for the PN and MOS detectors, respectively. We also
177: applied energy filters to keep the events only in the energy range
178: 0.5--8.0 keV. The resulting good exposure times are 24.3 ks for PN,
179: 32.8 ks for MOS1, and 32.9 ks for MOS2.
180:
181: We created a composite image using all the data (PN+MOS) in the 0.5--8
182: keV band. Using a modified version of the CIAO {\em wavdetect}
183: tool\footnote{{\tt http://cxc.harvard.edu/ciao/threads/wavdetect}}, we
184: found 22 point sources in the $20'\times 20'$ field. We searched for
185: counterparts of the field sources in the 2MASS and USNO-B1.0 catalogs
186: and found three matches in 2MASS and one match in USNO-B1.0, within
187: $2''$ of the X-ray source positions. (There are no other matches
188: closer than $6''$.) Assuming these are the counterparts of the X-ray
189: sources, we found average offsets of the X-ray with respect to
190: optical/NIR positions of $+0\farcs1$ in RA\ and $-0\farcs9$ in Dec,
191: with an rms error of about $0\farcs5$ in each coordinate (which does
192: not include the uncertainty of the X-ray source position
193: determination). Visual inspection of the image shows a faint source
194: at the (boresight-corrected) position $\alpha = 09^{\rm h}46^{\rm
195: m}07\fs 4$, $\delta = 09^\circ 51'54''$ (J2000.0), with an uncertainty
196: of about $5''$ in each coordinate. The main source of the uncertainty
197: is a large error in centroiding the very faint X-ray source. Since
198: the separation of 2\farcs7 between the X-ray source position and the
199: radio pulsar position ($\alpha = 09^{\rm h}46^{\rm m}07\fs 652$,
200: $\delta = 09^\circ 51' 55\farcs52$) is smaller than the source
201: position uncertainty, we conclude that the observed source is most
202: likely the X-ray counterpart of the radio pulsar.
203:
204: We measured background-subtracted source count rates of $2.52\pm0.48$,
205: $0.95\pm0.19$, and $1.03\pm0.21$ counts ks$^{-1}$ for the PN, MOS1,
206: and MOS2 detectors, respectively. These count rates were measured in
207: the 0.5--8 keV band within a $16''$ radius aperture, which contains
208: $\approx 67\%$ and 71\% of total point source counts for the PN and
209: MOS detectors, respectively.
210:
211: We were unable to find a satisfactory way to combine the PN and MOS
212: data for spectral analysis of this faint source as the responses of
213: these detectors are very different. Since the number of source counts
214: in the MOS data is very small, we used only the data from the most
215: sensitive PN detector. We extracted the source+background spectrum
216: (102 counts) for the combined PN data from a $16''$ radius circle and
217: the background from a nearby $32''$ radius source-free region. We
218: binned the spectrum to have a minimum of 10 counts per energy bin.
219: Spectral fitting was done using the XSPEC package (v.\ 11.3.0).
220:
221: Because of the scarce count statistics, we fixed the absorbing column
222: to the value corresponding to the pulsar's DM (assuming 10\% ionized
223: ISM): $n_{\rm H} =4.3\times10^{20}$ cm$^{-2}$. The spectrum fits
224: equally well with an absorbed blackbody (BB) or an absorbed power-law
225: (PL) model ($\chi_\nu^2 = 0.9$ and 0.7, respectively, for 8 dof). The
226: BB model gives a best-fit temperature $T = 3.1$ MK and a projected
227: emitting area of $A\sim 10^3 d_{630}^2$ m$^2$, corresponding to an
228: effective radius of about $18 d_{630}$ m of an equivalent emitting
229: disk, where $d_{630} = d/(630\,{\rm pc})$.
230: This thermal radiation could be interpreted as
231: emitted from a hot polar cap; in this case the bolometric luminosity
232: is $L_{\rm bol} = A
233: % \pi r_{\rm pc}^2
234: \sigma T^4 = 4.9^{+0.6}_{-1.6}\times
235: 10^{28} d_{630}^2 \langle \cos \theta\rangle^{-1}$ erg s$^{-1}$, where
236: $\langle \cos \theta\rangle$ is a time-averaged cosine of the angle
237: between the magnetic axis and the line of sight ($\langle \cos
238: \theta\rangle = 0.97$ for the axis orientations inferred by Deshpande
239: \& Rankin 2001). The errors here and below are determined from the
240: 68\% confidence contours for joint variation of two fitting
241: parameters. Although the strongly correlated temperature and radius
242: vary in broad ranges, the luminosity is much more restricted, as
243: demonstrated in the left panel of Figure~1.
244:
245: The PL model gives a photon index
246: $\Gamma = 2.6_{-0.5}^{+0.7}$ and
247: a flux $F_{0.5-8\,{\rm keV}} =
248: 4.4^{+1.8}_{-1.5}\times 10^{-15}$ erg cm$^{-2}$
249: s$^{-1}$.
250: This flux corresponds to
251: an isotropic luminosity $L_{0.5-8\,{\rm keV}}=
252: 4\pi d^2 F_{0.5-8\,{\rm keV}}^{\rm unabs}=
253: 2.4^{+0.8}_{-0.7}\times 10^{29} d_{630}^2$
254: erg s$^{-1}$.
255:
256: The inferred parameters of the X-ray source do not contradict to the
257: hypothesis that it is the counterpart of the pulsar (see \S3). The
258: most direct way to prove the identification would be detection of
259: X-ray pulsations. Our timing analysis of the PN data (time resolution
260: 0.2 s) did not reveal any significant signal at the pulsar's period of
261: 1.1 s. This result, however, is not restrictive: the radiation should
262: be about 100\% pulsed to find pulsations with only 102 counts detected
263: in two observations separated by about 7 month (of which only about
264: 61 counts are expected to come from the source). Thus,
265: given the uncertainty of the spectral results and a lack of period
266: detection, we cannot completely rule out the possibility that we
267: detected a field X-ray source very close to the pulsar position. In
268: this case, the above flux/luminosity estimates should be considered as
269: upper limits.
270:
271: \section{Physical implications}
272:
273: As described in \S2, the quality of the data is too poor to
274: discriminate between the thermal (BB from a polar cap) and nonthermal
275: (PL from the magnetosphere) spectral fits of the putative X-ray
276: counterpart of PSR B0943+10. The photon index for the PL fit is
277: somewhat larger than, but not inconsistent with, those observed in
278: other pulsars ($\Gamma=1$--2), given its large uncertainty (in
279: particular, $\Gamma$ decreases if a smaller $n_{\rm H}$ value is
280: assumed). The luminosity found from the PL fit allows one to estimate
281: the ``X-ray efficiency'' of the magnetospheric emission, $\eta \equiv
282: L_X/\dot{E}$. It follows from the above results that $\eta =2.4
283: (-1.2,+1.8)\times 10^{-3}$ and $0.8 (-0.6,+1.5)\times 10^{-3}$ in the
284: 0.5--8 and 2--10 keV bands, respectively, where the distance
285: uncertainty is included (the latter band is often used for
286: characterization of the nonthermal pulsar radiation; e.g., Possenti et
287: al.\ 2002). Although the inferred efficiency
288: is rather high (close to the upper limit suggested by Possenti et al.\
289: 2002), it is within the range of efficiencies found recently for other
290: old pulsars (Zavlin \& Pavlov 2004).
291:
292: Within the thermal interpretation of the observed radiation,
293: particularly interesting implications follow from the comparison with
294: the predictions of polar cap heating models, especially the vacuum gap
295: model (RS75). In this model, a vacuum gap of height $h$ breaks down
296: due to run-away electron-positron pair production, forming ``sparks''
297: along the magnetic field lines.
298: A half of the relativistic particles produced in the spark discharge
299: stream back toward the neutron star and heat the ``spark footprints''
300: on the polar cap surface. Since the strong magnetic field suppresses
301: the heat conduction across the field lines, and the cooling timescales
302: are very short (about a few $\mu$s for $B=$ a few $\times
303: 10^{12}$\,G; Gil et al. 2003), only the spark footprints emit thermal
304: X-rays while the rest of the polar cap remains cold. The BB fit of
305: the observed spectrum allows one to estimate the bolometric
306: luminosity, $L_{\rm bol} \approx 5\times 10^{28}$ ergs s$^{-1}$, and
307: the total X-ray emitting area that is strongly correlated with the
308: temperature, $A \approx 10^3 (T/3\,{\rm MK})^{-4}$ m$^2$ ($A \approx
309: 300$--5000 m$^2$, $T\approx 2.0$--4.2 MK, at 68\% confidence). Even
310: the largest area allowed by the fit is much smaller than the
311: conventional polar cap area for a dipole magnetic field, $A_{\rm
312: pc}^{\rm dip} =\pi (r_{\rm pc}^{\rm dip})^2 \simeq 6\times 10^4$
313: m$^2$, where $r_{\rm pc}^{\rm dip} = (2\pi R^3/c P)^{1/2} \simeq 138$
314: m, for a neutron star radius $R=10$ km. The spark footprints fill a
315: fraction of the polar cap, with a filling factor $f=A/A_{\rm pc} \sim
316: 0.017 b A_3 $, where $A_3 = A/(10^3\,{\rm m}^2$), and
317: the factor $b = B_s / B_p \sim
318: A_{\rm pc}^{\rm dip}/A_{\rm pc} >1$
319: takes into account that the magnetic field is non-dipolar
320: (i.e., the field in the gap, $B_s$, may exceed the dipolar value $B_p$,
321: and the polar cap area may be smaller
322: than $A_{\rm pc}^{\rm dip}$; Gil et al.\ 2002, 2003).
323: The radio sub-pulse drifting was interpreted as caused by 20
324: sparks drifting along the circular outer boundary of the polar cap
325: (Deshpande \& Rankin 1999,2001). Since some sparks may not be seen in
326: the radio (but should be seen in more isotropic thermal X-rays), the
327: total number of sparks $N$ can exceed 20 (e.g., Gil \& Sendyk 2000
328: suggest $N\simeq 42$).
329: A characteristic spark radius, $r_{\rm sp}$, can be estimated as
330: (assuming all sparks have the same size)
331: \be
332: r_{\rm sp} \sim (A/\pi N)^{1/2} = 4\, A_3^{1/2} (N/20)^{-1/2} {\rm m}.
333: \label{rsp}
334: \ee
335:
336: Another constraint on the gap and spark parameters
337: comes from the comparison of the bolometric
338: and spin-down luminosities. The bolometric
339: polar cap luminosity can be
340: estimated as $L_{\rm bol} \sim
341: (1/2) \Phi c
342: \rho_{_{\rm GJ}} A$,
343: where $\Phi \sim 2\pi B_s h^2/cP$ is the
344: potential drop across the gap (along the magnetic field lines), and
345: $\rho_{_{\rm GJ}}
346: %\propto B_s$
347: \sim B_s/cP$
348: is the Goldreich-Julian charge density.
349: On the other hand, the total spin-down luminosity
350: can be written as $\dot E \sim 2 \Phi_{\rm max} c \rho_{_{\rm
351: GJ}}^{\rm dip} A_{\rm pc}^{\rm dip}= 2 \Phi_{\rm max} c \rho_{_{\rm
352: GJ}} A_{\rm pc}$,
353: % \pi
354: %r_p^2$,
355: where
356: $\Phi_{\rm max} = B_p A_{\rm pc}^{\rm dip}/cP = B_s A_{\rm pc}/cP$
357: is the maximum possible
358: potential drop along a magnetic field line (RS75). This gives
359: $L_{\rm bol}/\dot{E} \sim (f/4) (\Phi/\Phi_{\rm max}) =
360: (1/2) N (r_{\rm sp}/r_{\rm pc})^2 (h/r_{\rm pc})^2$.
361: Using the observational value $L_{\rm bol}/\dot{E}\approx 5\times 10^{-4}$,
362: we obtain
363: \be
364: (h/r_{\rm pc}) (r_{\rm sp}/r_{\rm pc}) \sim 7\times 10^{-3}
365: (N/20)^{-1/2}.
366: \label{hrsp/rpc2}
367: \ee
368: Adopting the assumption that the size of the spark is
369: comparable to the gap height
370: (i.e., $r_{\rm sp} \sim h/2$; Gil \& Sendyk 2000),
371: we obtain from equation (2)
372: \be
373: h/r_{\rm pc} \sim 0.12\, (N/20)^{-1/4}\,.
374: \label{h/rpc}
375: \ee
376: This gives $h \sim 16\, b^{-1/2} (N/20)^{-1/4}$ m,
377: and $r_{\rm sp} \sim 8\, b^{-1/2} (N/20)^{-1/4}$ m.
378: Using this estimate for $r_{\rm sp}$ and equation (1), we obtain
379: $A \sim 4000 b^{-1} (N/20)^{1/2}$ m$^2$, which is consistent with the
380: X-ray data at reasonable values of $b$ and $N$ (see Fig.\ 1).
381:
382: Further constraints can be obtained from the drifting period.
383: According to RS75, for sparks at the radial distance $r_\perp$
384: from the pole,
385: it can be estimated as $\hat{P}_3 \simeq 2\pi r_\perp/v
386: \simeq 2\pi r_\perp B_s/E_\perp c$, where
387: $v=|{\bf E \times B_s}| c/B_s^2$ is the drift velocity,
388: and $E_\perp = |dV(r_{\perp})/dr_\perp|$ is the electric field component
389: perpendicular to the magnetic field lines at the top of the gap.
390: In the case $h\ll r_{\rm pc}$, $V(r_\perp)$ is approximately uniform
391: throughout the polar cap except for near the cap edge, $r_{\rm
392: pc}-r_\perp \lesssim h$ (Appendix II of RS75). This means that the
393: ``inner sparks'' drift very slowly,
394: while $E_\perp \sim \Phi/h$, $v\sim 2\pi h/P$, and
395: $\hat{P}_3/P \sim r_{\rm pc}/h$ for the sparks near the cap edge.
396: The measured $\hat{P}_3/P \simeq 37$ thus gives $h/r_{\rm pc} \sim
397: 0.027$, or $h\sim 4\, b^{-1/2}$ m
398: and $r_{\rm sp} = 36\, b^{-1/2} (N/20)^{-1/2}$ m (from eq.\ [2]).
399: Such a spark radius substantially exceeds the gap height, and it
400: corresponds to the emitting area, $A \sim 8\times 10^4 b^{-1}$ m$^2$,
401: well above the range allowed by the X-ray fit.
402: Therefore, the original vacuum gap model of RS75 can hardly be
403: reconciled
404: with the X-ray data. However, we cannot rule out that a modification of the
405: RS75 model could explain both the X-ray and radio data in a consistent way.
406: For instance, Gil et al.\ (2003) suggested a modification assuming
407: thermal emission of charged particles from the polar cap surface
408: which can partially screen the gap potential: $\Phi \to \zeta\Phi$, where
409: $\zeta < 1$ is the screening parameter. In such a modified RS75 model,
410: $\hat{P}_3/P \sim r_{\rm pc}/\zeta h$.
411: For $\zeta = 0.17$, inferred by Gil et al.\
412: for PSR B0943+10, we obtain $h/r_{\rm pc} \sim 0.16$, very
413: close the value inferred from the X-ray data (eq.\ [\ref{h/rpc}]).
414: However, a more detailed study of the dynamics of a partially screened
415: unsteady polar gap is desirable for
416: an accurate treatment of the problem.
417: We should also caution that the BB model used for
418: fitting the spectrum gives only crude estimates for the area and
419: temperature in the case of strong magnetic
420: fields (Zavlin \& Pavlov 2004).
421: To conclude, the thermal interpretation of the X-ray data
422: is not inconsistent with the general idea that the X-ray radiation
423: is emitted from spark footprints, but the original vacuum gap model
424: of RS75 needs some modification to explain both the X-ray and radio
425: data.
426:
427: The results of our X-ray observations also constrain other models.
428: Motivated by solving the binding energy problem, Xu et al.\ (1999)
429: suggested that PSR 0943+10 and other drifting pulsars might be strange
430: quark stars covered by a very thin layer of normal matter. In this
431: case, the heat deposited at the polar cap spreads over the entire star
432: surface because of a high thermal conductivity of the surface layer
433: (Xu et al.\ 2001). This results in a nearly uniform, low surface
434: temperature ($\lesssim 0.1$ MK), being inconsistent with the thermal
435: interpretation of the observed X-ray radiation (but of course this
436: hypothesis cannot be excluded if the observed radiation is
437: nonthermal).
438:
439: Polar cap heating in the SCLF model has been studied by Harding \&
440: Muslimov (2001,2002). Assuming a dipole field configuration near the
441: surface, the timing parameters of PSR B0943+10 indicate that the gap
442: is controlled by inverse Compton scattering (ICS) rather than
443: curvature radiation (CR), and that the gap is in the ``saturated''
444: regime. Using eqs.\ (64) and (66) of Harding \& Muslimov (2002), the
445: predicted polar cap luminosity for this pulsar is $8.8(1.7) \times
446: 10^{27}\,{\rm erg\,s^{-1}}$ for a resonant (non-resonant) ICS gap.
447: Since these luminosities are much lower than observed, the SCLF model
448: is inconsistent with the thermal interpretation of the X-ray data,
449: but is consistent with the nonthermal interpretation. If future
450: observations show that the X-ray radiation is, in fact, predominantly
451: nonthermal ($L_{\rm bol}\ll 5\times 10^{28}$ ergs s$^{-1}$), then the
452: regular drifting behavior observed in PSR B0943+10 and other drifting
453: pulsars will have to be interpreted by models other than bunched
454: plasma sparks in the inner gap (e.g., Kazbegi et al.\ 1996; Wright
455: 2003).
456:
457: \section{Summary}
458: To test the existence of an inner vacuum gap in PSR B0943+10, we
459: observed the pulsar with {\sl XMM-Newton} and detected a point source
460: at the radio pulsar position. The best-fit 0.5--8 keV band isotropic
461: luminosity is $\sim 2.4\times 10^{29}~{\rm erg~s^{-1}}$ for the PL
462: fit, while the bolometric BB luminosity is about $5\times 10^{28}$ erg
463: s$^{-1}$. Within the thermal interpretation, the X-ray radiation is
464: emitted from a heated area much smaller than the conventional polar
465: cap area. In the framework of the sparking pulsar model, this implies
466: that only spark footprints are heated up to X-ray temperatures while
467: the rest of the polar cap remains relatively cold. The thermal model,
468: supplemented with the results of observations of subpulse drifting, is
469: consistent with the presence of an inner gap of a height $h \sim
470: 0.1$--$0.2\, r_{\rm pc}$, and $N\geq 20$ rotating sparks whose
471: footprints fill $\sim 0.04$--0.07 of the polar cap area. The thermal
472: model disfavors the strange quark star model of PSR 0943+10, and it is
473: also inconsistent with the SCLF polar cap heating model. Both models
474: are, however, allowed if the spectrum is non-thermal.
475:
476: The nonthermal (magnetospheric) interpretation of the observed radiation
477: cannot be ruled out. In this interpretation, the X-ray efficiency,
478: $\eta\sim 10^{-3}$ in the 2--10 keV band, is comparable to that of
479: other old pulsars. Much deeper observations are required to
480: firmly establish the nature of the X-ray radiation from PSR B0943+10.
481:
482: \acknowledgements
483: This work was supported by NASA grants NAG5-13539 and NAG5-10865.
484:
485: \begin{thebibliography}{}
486: \bibitem[]{} Arons, J. \& Scharlemann, E. T. 1979, ApJ, 231, 845
487: \bibitem[]{} Asgekar, A. \& Deshpande, A. A. 2001, MNRAS, 326, 1249
488: \bibitem[]{} Asseo, E. \& Khechinashvili, D. 2002, MNRAS, 334, 743
489: \bibitem[]{} Cordes, J. M. \& Lazio, T. J. W. 2002, astro-ph/0207156
490: \bibitem[]{} Deshpande, A. A. \& Rankin, J. M. 1999, ApJ, 524, 1008
491: \bibitem[]{} -----. 2001, MNRAS, 322, 438
492: \bibitem[]{} Gil, J. \& Melikidze, G. I. 2002, ApJ, 577, 909
493: \bibitem[]{} Gil, J., Melikidze, G. I. \& Geppert, U. 2003, A\&A, 407,
494: 315
495: \bibitem[]{} Gil, J., Melikidze, G. I. \& Mitra, D. 2002, A\&A, 388,
496: 246
497: \bibitem[]{} Gil, J. \& Sendyk, M. 2000, ApJ, 541, 351
498: \bibitem[]{} -----. 2003, ApJ, 585, 453
499: \bibitem[]{} Harding, A. K. \& Muslimov, A. G. 1998, ApJ, 508, 328
500: \bibitem[]{} -----. 2001, ApJ, 556, 987
501: \bibitem[]{} -----. 2002, ApJ, 568, 862
502: \bibitem[]{} Harding, A. K., Muslimov, A. G. \& Zhang, B. 2002, ApJ,
503: 576, 366
504: \bibitem[]{} Jones, P. B. 1986, MNRAS, 218, 477
505: \bibitem[]{} Kazbegi, A., Machabeli, G., Melikidze, G. \& Shukre, C.
506: 1996, A\&A, 309, 515
507: \bibitem[]{} K\"ossl, D., Wolff, R. G., M\"uller, E. \& Hillebrandt,
508: W. 1988, A\&A, 205, 347
509: \bibitem[]{} Melikidze, G. I., Gil, J. \& Pataraya, A. D. 2000, ApJ,
510: 544, 1081
511: \bibitem[]{} Melrose, D. 2004, in Young Neutron Stars and
512: Their Environments (IAU Symp.\ 218, ASP Conf.\ Proc.),
513: eds.\ F.\ Camilo and B.\ M.\ Gaensler, p.349
514: \bibitem[]{} Possenti, A., Cerutti, R., Colpi, M. \& Mereghetti, S.
515: 2002, A\&A, 387, 993
516: \bibitem[]{} Qiao, G. J. \& Lin, W. P. 1998, A\&A, 333, 172
517: \bibitem[]{} Rankin, J. M. 1986, ApJ, 301, 901
518: \bibitem[]{} Ruderman, M. \& Sutherland, P. G. 1975, ApJ, 196, 51 (RS75)
519: \bibitem[]{} Usov, V. V. \& Melrose, D. B. 1996, ApJ, 464, 306
520: \bibitem[]{} Wright, G. A. E. 2003, MNRAS, 344, 1041
521: \bibitem[]{} Xu, R. X., Qiao, G. J. \& Zhang, B. 1999, ApJ, 522, L109
522: \bibitem[]{} Xu, R. X., Zhang, B. \& Qiao, G. J. 2001, APh, 15, 101
523: \bibitem[]{} Zavlin, V.\ E., \& Pavlov, G.\ G. 2004, ApJ, 616, 452
524: \bibitem[]{} Zhang, B. \& Harding, A. K. 2000, ApJ, 532, 1150
525: \bibitem[]{} Zhang, B., Harding, A. K. \& Muslimov, A. G. 2000, ApJ,
526: 531, L135
527: \end{thebibliography}
528:
529:
530: \end{document}
531:
532: