1:
2: \documentclass{emulateapj}
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{\nh}{n_{\rm H}}
14: \newcommand{\edot}{\dot{E}}
15: \def\farcs{\hbox{$.\!\!^{\prime\prime}$}}
16: \def\farcm{\hbox{$.\!\!^{\prime}$}}
17:
18:
19:
20: \def\chan{{\sl Chandra}}
21: \def\xmm{{\sl XMM-Newton}}
22: \def\tpc{T_{\rm pc}}
23: \def\rpc{R_{\rm pc}}
24: \def\tef{T_{\rm eff}}
25: \def\lbol{L_{\rm bol}}
26: \def\lbolpc{L_{\rm bol}^{\rm pc}}
27: \def\epc{\eta^{\rm pc}}
28: \def\enon{\eta^{\rm nonth}}
29: \newcommand{\gapr}{\raisebox{-.6ex}{\mbox{
30: $\stackrel{>}{\mbox{\scriptsize$\sim$}}\:$}}}
31: \newcommand{\lapr}{\raisebox{-.6ex}{\mbox{
32: $\stackrel{<}{\mbox{\scriptsize$\sim$}}\:$}}}
33:
34:
35: \begin{document}
36:
37:
38: \submitted{\today}
39:
40: \title{
41: X-ray emission from the double neutron star binary B1534+12: Powered
42: by the pulsar wind?}
43:
44: \author{
45: O.\ Kargaltsev, G.\ G.\ Pavlov, and G.\ P.\ Garmire}
46: \affil{The Pennsylvania State University, 525 Davey Lab, University
47: Park, PA 16802, USA} \email{pavlov@astro.psu.edu}
48:
49:
50:
51: \begin{abstract}
52: We report the detection of the double neutron star binary (DNSB)
53: B1534+12 (= J1537+1155) with the \chan\ X-ray Observatory. This DNSB
54: ($P_{\rm orb} = 10.1$ hr) consists of the millisecond (recycled)
55: pulsar J1537+1155A ($P_A=37.9$ ms) and a neutron star not detected
56: in the radio. After the remarkable double pulsar binary
57: J0737$-$3039, it is the only other DNSB detected in X-rays. We
58: measured the flux of $(2.2\pm 0.6)\times10^{-15}$ ergs s$^{-1}$
59: cm$^{-2}$ in the 0.3--6 keV band. The small number of collected
60: counts allows only crude estimates of spectral parameters. The
61: power-law fit yields the photon index $\Gamma = 3.2\pm 0.5$ and the
62: unabsorbed 0.2--10 keV luminosity $L_{\rm X}\approx6\times10^{29}\,
63: {\rm ergs}\, {\rm s}^{-1} \approx 3\times 10^{-4} \dot{E}_A$, where
64: $\dot{E}_A$ is the spin-down power of J1537+1155A. Alternatively,
65: the spectrum can be fitted by a blackbody model with $T\approx 2.2$
66: MK and the projected emitting area of $\sim 5\times 10^3$ m$^2$. The
67: distribution of photon arrival times over binary orbital phase shows
68: a deficit of X-ray emission around apastron, which suggests that the
69: emission is caused by interaction of the relativistic wind from
70: J1537+1155A with its neutron star companion. We also reanalyzed the
71: {\sl Chandra} and {\sl XMM-Newton} observations of J0737$-$3039 and
72: found that its X-ray spectrum is similar to the spectrum of
73: B1534+12, and its X-ray luminosity is about the same fraction of
74: $\dot{E}_A$, which suggests similar X-ray emission mechanisms.
75: However, the X-ray emission from J0737$-$3039 does not show orbital
76: phase dependence. This difference can be explained by the smaller
77: eccentricity of J0737$-$3039 or a smaller misalignment between the
78: equatorial plane of the millisecond pulsar and the orbital plane of
79: the binary.
80: \end{abstract}
81: \keywords{pulsars: individual (PSR B1534+12 = PSR J1537+1155, PSR J0737$-$3039A,
82: PSR J0737$-$3039B)
83: --- stars: neutron --- X-rays: stars}
84: \section{Introduction}
85:
86: The sample of known double neutron star binaries (DNSBs) has nearly doubled over the
87: last six years. Currently, eight such binaries are known
88: (they are shown in the $P$-$\dot{P}$ diagram
89: in Fig.\ 1). All of them are found
90: via radio timing of
91: pulsars which provides
92: binary orbital
93: parameters, including the companion masses. DNSBs are thought to be
94: formed in consequent supernova explosions in the course of evolution
95: of massive binary
96: systems.
97: The details of binary evolution depend on
98: initial star masses and binary orbital parameters (see Stairs 2004
99: for a review).
100: It is generally believed
101: that once the first neutron star (NS) is formed in the binary,
102: accretion of matter from the evolved second star
103: spins up the NS so that it becomes a
104: millisecond (recycled) pulsar (MSP).
105: Eventually, the second
106: star explodes as a supernova and, if the explosion does not disrupt
107: the binary or disintegrate the second star,
108: a DNSB (or a NS-black hole binary) is
109: formed.
110:
111: Radio observations of DNSBs allow one to measure the masses of both
112: NSs and test the predictions of General Relativity (GR). The recent
113: discovery of the tight DNSB J0737$-$3039 ($P_{\rm orb}=2.45$ hours;
114: see Burgay et al.\ 2003; Lyne et al.\ 2004; and Tables 1 and 2 for
115: other details), where both NSs are radio pulsars orbiting in the
116: plane seen nearly edge-on (the inclination angle $i$ differs from
117: $90^\circ$ by $0\fdg 29 \pm 0\fdg 14$; Coles et al.\ 2005),
118: dramatically boosted interest in DNSBs. Apart from measuring the GR
119: parameters, the radio observations have shown variations of the
120: pulsed flux with orbital phase, for both pulsars.
121: The radio pulse of the MSP J0737$-$3039A
122: (hereafter J0737A) disappears around its superior
123: conjunction\footnote{
124: At the superior (inferior) conjunction of J0737A, it is located
125: behind (in front of) J0737B along the observer's
126: line-of-sight.},
127: which has been interpreted as an eclipse of J0737A ( Lyne et al.\
128: 2004), caused by synchrotron absorption in the magnetosphere of
129: J0737B (Rafikov \& Goldreich 2005; Lyutikov \& Thompson 2005).
130: Even more interesting is
131: the finding that the generally faint pulsed radio emission of J0737B
132: is strongly enhanced during two short orbital phase intervals of
133: $\sim 0.1$ of the period around the phase of inferior conjunction of
134: J0737B. These changes suggest that the magnetosphere of J0737B
135: responds to the wind or radiation of the more energetic J0737A
136: (Lyne et
137: al.\ 2004;
138: Jenet \& Ransom 2004;
139: Demorest at
140: al.\ 2004;
141: Zhang \& Loeb 2004).
142:
143:
144: The J0737 was
145: the first and, until this work, the only DNSB detected in
146: X-rays.
147: It has been observed with both \chan\ (McLaughlin et al.\ 2004) and
148: \xmm\
149: (Campana, Possenti \& Burgay 2004; Pellizzoni et al.\ 2004).
150: Its luminosity
151: is $L_{\rm X}\sim (2$--$3) \times 10^{30}$ ergs s$^{-1}$ in the 0.2--10 keV
152: band,
153: assuming a 500 pc distance. The X-ray spectrum is rather soft;
154: it can be fitted
155: by either
156: a power-law (PL) model with a photon index
157: $\Gamma \sim 3$--4
158: or a blackbody (BB)
159: model with
160: $kT\approx 0.2$ keV and projected emitting area
161: $A\sim 3\times 10^4$ m$^2$.
162: No X-ray variability has been reported for J0737 so far.
163: A number of different
164: interpretations of the X-ray emission
165: have been suggested. These include emission from
166: the surface and/or magnetosphere of J0737A (similar to
167: solitary
168: MSPs),
169: emission from the J0737B's magnetosphere (surface) energized (heated)
170: by the J0737A's wind (Zhang \& Loeb 2004),
171: and emission by particles accelerated in
172: a bow shock
173: produced by the J0737A's wind in the interstellar medium (ISM)
174: or in the magnetosphere of J0737B (Granot \& M\'esz\'aros 2004).
175: However, because of
176: the small number of X-ray counts collected, it is difficult to
177: discriminate between these models from observations of just one
178: DNSB.
179:
180: In this paper we report the detection of X-rays from
181: the second DNSB, B1534+12 (=J1537+1155; hereafter
182: J1537), discovered by Wolszczan (1991).
183: One of the two NSs in this DNSB is a millisecond pulsar
184: (J1537A hereafter; period $P_{A}=37.9$ ms,
185: characteristic age $\tau_{A}\equiv P_{A}/2\dot{P}_{A}=248$
186: Myr, spin-down power
187: $\dot{E}_{A}=1.8\times10^{33}$ ergs s$^{-1}$) on a 10.1 hour
188: eccentric ($e=0.274$) orbit.
189: As one can see from Figure 1 and Table 2,
190: the J1537A's parameters are generally close to those of
191: J0737A,
192: although J1537A is a factor of 3 less energetic.
193: The dispersion measure of J1537A suggests a distance of about 0.7 kpc.
194: The comparison of the observed orbital period decay
195: with that predicted by GR
196: allowed Stairs et al.\ (2002)
197: to evaluate the distance more accurately: $D=1.02\pm0.05$ kpc.
198: Multi-year timing observations in the radio (Stairs et al.\ 2004)
199: suggest that we are looking at J1537 close to
200: its orbital plane ($i\simeq 77\fdg 2$), and J1537A is
201: an almost orthogonal rotator (i.e., its magnetic axis and
202: the line of sight are nearly orthogonal to
203: the spin axis: $\alpha\approx103^{\circ}$,
204: $\zeta\approx 97^\circ$ at the epoch of our observation\footnote{The angle
205: $\zeta$ is changing with time, $d\zeta/dt \simeq -0.21^\circ /{\rm yr}$
206: as a result of geodetic precession.}) whose spin axis is
207: misaligned with the orbital angular momentum by
208: $\delta = 25^{\circ}\pm 4^\circ$
209: (Stairs et al.\ 2004; Thorsett et al.\ 2005).
210: The second NS (hereafter J1537B)
211: has not been detected
212: as a radio pulsar,
213: but its mass, $M_{B}=(1.345\pm0.001)M_{\odot}$, has been
214: measured (Stairs et al.\ 2002). The detection of X-rays from J0737
215: prompted us to observe J1537, the second nearest DNSB.
216: We describe the {\sl Chandra} observation of J1537 in \S2 and
217: the observational results in \S3, which also includes our reanalysis
218: of the X-ray observations of J0737 and a comparison of the X-ray
219: properties of these DNSBs. We discuss plausible X-ray emission
220: models and compare their predictions with the observed X-ray
221: properties in \S4 and summarize our results in \S5.
222:
223:
224: \section{Observation}
225:
226: We observed J1537 with the Advanced CCD Imaging Spectrometer (ACIS)
227: aboard {\sl Chandra} on 2005 April 10 (start time
228: 53,470.352222 MJD).
229: The useful scientific exposure time was 36,080 s
230: (first 935 s of the 37,015 s total exposure
231: are excluded from the Good Time Intervals [GTIs]). The
232: observation was carried out in Very Faint mode, and the pulsar was
233: imaged on S3 chip of the ACIS-S array. The detector was operated in
234: Full Frame mode which provides time resolution of 3.2 seconds. The
235: data were reduced using the Chandra Interactive Analysis of
236: Observations (CIAO) software (ver.\ 3.2.1; CALDB ver.\ 3.0.3).
237:
238:
239: \begin{figure}
240: % \center
241: \vspace{-1.0cm}
242: \includegraphics[width=2.8in,angle=90]{f1.eps}
243: \caption{ $P$-$\dot{P}$ diagram for $\approx1400$ radio pulsars
244: (small dots) from the ATNF catalog (Manchester et al.\ 2005). Lines
245: of constant pulsar age, magnetic field, and $\dot{E}$ are shown.
246: Eight DNSBs are marked by different symbols. J1537A is marked by the
247: filled star. The J0737 components are marked by open stars connected
248: with the straight dashed line. The rest of the DNSBs are marked by
249: filled circles. {\em See the electronic edition of the Journal for
250: the color version of this figure.} }
251: \end{figure}
252:
253: \section{X-ray image and spectrum}
254:
255: Figure 2 shows the ACIS-S3 image of the J1537's field. The X-ray
256: source is clearly seen
257: at ${\rm R.A.}=15^{\rm h}37^{\rm
258: m}09\fs986$, ${\rm Decl.}=+11^{\circ}55' 55\farcs39$
259: (the combined aspect determination and centroiding uncertainty
260: is 0\farcs7 at a 90\% confidence level).
261: Since this position is within $0\farcs5$ of the J1537A's radio
262: position
263: (Stairs et al.\ 2002),
264: we conclude with confidence that we detected the X-ray
265: emission from J1537. The distribution of source counts in the
266: ACIS image is consistent with that of a point source.
267:
268:
269: \begin{figure}[t]
270: \centering
271: \includegraphics[width=2.6in,angle=0]{f2.eps}
272: \caption{ ACIS-S3 image of the J1537 field. The circle of 1\farcs23
273: radius shows the radio position of PSR J1537+1155 from Stairs et
274: al.\ (2002). }
275: \end{figure}
276:
277: We extracted the pulsar spectrum from a circular aperture of
278: 2.5 ACIS pixels radius ($=1\farcs23$; $\approx90$\% encircled energy radius)
279: using the CIAO {\tt psextract} task. The background
280: was extracted from a $5''<r<22''$ annulus centered on the source.
281: The total number of counts within the source aperture is 16, of
282: which 98.4\% are expected to come from the source. The observed
283: source flux is $(2.2\pm0.6)\times10^{-15}$ erg s$^{-1}$ cm$^{-2}$ in
284: the 0.3--6 keV range.
285: We group the counts
286: into 4 spectral bins, with each bin having 3--5 counts (see Fig.\
287: 3), and fit the spectrum with absorbed BB and PL models using the
288: C-statistic (Cash 1979) implemented in XSPEC, ver.\ 11.3.0 (the
289: fitting parameters are given in Table 3). To obtain constrained
290: fits, we had to freeze the hydrogen column density, $N_{\rm H}$. The
291: pulsar's dispersion measure, DM$=11.6$ cm$^{-3}$ pc, corresponds to
292: $N_{\rm H} =3.6\times 10^{20}$ cm$^{-2}$ (assuming a 10\% ISM
293: ionization). This value is close to the total Galactic absorption in
294: neutral hydrogen, $3.6\times 10^{20}$ cm$^{-2}$ (Dickey \& Lockman
295: 1990) in the direction of J1537 ($l=19\fdg85$, $b=48\fdg34$), which
296: is located well above the Galactic plane ($z=0.76$ kpc at the
297: distance of 1 kpc). Therefore, we adopt $N_{\rm H}=3.6\times
298: 10^{20}$ cm$^{-2}$ in the X-ray fits.
299:
300:
301: \begin{figure}
302: \centering
303: \vbox{
304: \includegraphics[width=2.1in,angle=-90]{f3a.eps}
305: \vspace{0.3cm}
306: \includegraphics[width=2.3in,angle=-90]{f3b.eps}}
307: \caption{ X-ray spectrum of J1537 fitted with PL ({\em top}) and BB
308: ({\em bottom}) models. The contributions of the energy bins into the
309: best-fit C-statistic are shown in lower panels, multiplied by $-1$
310: when the number of data counts is smaller than the number of model
311: counts. }
312: \end{figure}
313:
314: The spectrum fits
315: a soft PL with a photon index
316: $\Gamma=
317: 2.7$--3.7 and an isotropic unabsorbed 0.2--10 keV
318: luminosity $L_{\rm X} \equiv 4\pi D^2 F_{\rm X}^{\rm unabs}
319: \simeq(6.1^{+3.0}_{-2.1})
320: \times 10^{29}$ ergs s$^{-1}$
321: at
322: $D =1$ kpc.
323: The temperature and the projected area of the emitting region
324: obtained from the BB fit are strongly correlated (see Fig.\ 4),
325: which results in large uncertainty of these parameters. The BB model
326: gives the
327: best-fit temperature $T_{\rm BB} = 2.2$ MK and a projected emitting area
328: $A= 5\times 10^{3}$ m$^2$. The projected area is much smaller than
329: that of a NS, $\sim 3 \times10^{8}$ m$^{2}$, suggesting that the
330: thermal radiation might be emitted from small, hot polar caps with an
331: effective radius $R \sim 39 \langle \cos\varsigma\rangle^{-1/2}$ m and a
332: bolometric luminosity $L_{\rm bol}= \sigma T^4 A \langle
333: \cos\varsigma\rangle^{-1}=(6.8\pm2.1)\times10^{28}\langle
334: \cos\varsigma\rangle^{-1}$ ergs s$^{-1}$, where $\langle
335: \cos\varsigma\rangle$ is the time-averaged cosine of the angle between
336: the line of sight and the
337: normal to the visible polar cap ($\langle\cos\varsigma\rangle\approx 0.62$
338: for the axis orientations suggested by Stairs et
339: al.\ 2004, assuming a centered dipole with two equivalent polar caps
340: and neglecting the GR bending of photon trajectories). The X-ray
341: luminosity of J1537 corresponds to $3.4\times10^{-4}\dot{E}_A$ and
342: $3.8\times10^{-5}\langle \cos\varsigma\rangle^{-1}\dot{E}_A =
343: 6.1\times 10^{-5}\dot{E}$ for the PL (0.2--10 keV band) and BB fits,
344: respectively. The scarce statistics does not allow us to
345: discriminate between the BB and PL models.
346:
347:
348: To compare the spectra and light curves of J1537 and J0737,
349: we reanalyzed the {\sl
350: Chandra} ACIS and {\sl XMM-Newton} EPIC observations of J0737, using
351: the same spectral models and the same approach to the data analysis
352: as for J1537. The observations have been described by McLaughlin et
353: al.\ (2004), Campana et al.\ (2004), and Pellizzoni et al.\ (2004).
354: The results of our fits are presented in Table 3 and Figure 5. For
355: the PL fit of the ACIS spectrum, we obtained about the same
356: parameters as McLaughlin et al.\ (these authors did not fit the BB
357: model). As for the EPIC data, we found, in accordance with Campana
358: et al., that the EPIC-pn data are too noisy for a useful analysis because of
359: the very high background in the timing mode. Therefore, we only used
360: the MOS1 and MOS2 data.
361: For each of
362: the MOS detectors, we created a light curve with binning time of 10 s
363: and constructed GTIs excluding the time intervals during which the
364: total chip count rate exceeded 3 counts s$^{-1}$ (the remaining
365: useful time is 24.62 and 24.60 ks for MOS1 and MOS2, respectively).
366: The spectra we extracted from circular apertures
367: of $46''$ radius ($\simeq 90\%$ encircled energy). Comparing the properties
368: of the X-ray radiation from J0737 and J1537, we see that the
369: spectral parameters are very similar: $\Gamma \approx 3$ for the PL fit,
370: and $kT\approx 0.2$ keV for the BB fit.
371:
372:
373: \begin{figure}[]
374: \centering
375: \vbox{
376: \includegraphics[width=2.5in,angle=90]{f4a.eps}
377: \vspace{-0.0cm}
378: \includegraphics[width=2.5in,angle=90]{f4b.eps}}
379: \caption{ Confidence contours (68\% and 90\%) for the PL ({\em top})
380: and BB ({\em bottom}) fits to the spectrum of J1537. The PL
381: normalization is in units of $10^{-7}$ photons cm$^{-2}$ s$^{-1}$
382: keV$^{-1}$ at 1 keV. The BB normalization (vertical axis) is the
383: projected emitting area in units of m$^2$, for $D = 1$ kpc. The
384: lines of constant unabsorbed flux (top panel; in units of 10$^{-15}$
385: ergs cm$^{-2}$ s$^{-1}$, in 0.2--10 keV band) and constant $L_{\rm
386: bol}\langle\cos\varsigma\rangle$ (bottom panel; in units of
387: 10$^{28}$ ergs s$^{-1}$) are plotted as dashed lines, for fixed
388: $N_{\rm H} =3.6\times 10^{20}$ cm$^{-2}$. }
389: \end{figure}
390:
391: \begin{figure}[]
392: \centering
393: \vbox{
394: \includegraphics[width=2.5in,angle=90]{f5a.eps}
395: \vspace{-0.0cm}
396: \includegraphics[width=2.5in,angle=90]{f5b.eps}}
397: \caption{ Same as in Fig.\ 4 but for J0737. Solid and dash-dotted
398: contours are from the {\sl XMM-Newton} EPIC MOS and {\sl Chandra}
399: ACIS observations, respectively. {\em See the electronic edition of
400: the Journal for the color version of this figure.}}
401: \end{figure}
402:
403:
404: In the upper panel of Figure 6 we show the distribution of arrival
405: times for the 16 detected photons
406: over binary orbital phases of J1537
407: ($\phi_{\rm orb} =0$ corresponds to
408: periastron passage)\footnote{To calculate the event orbital phases,
409: we used the binary ephemeris for the epoch of our observation
410: kindly provided by Ingrid Stairs:
411: $P_{\rm orb} = 0.420737299122$ days, epoch of periastron
412: 53470.309046790 MJD, longitude of periastron 290\fdg 004585816.}.
413: Notice that no events arrived during the 12,795 s interval between
414: MJD 53470.46046 and 53470.60855, which corresponds to the phase
415: interval [0.365; 0.717] ($\Delta\phi_{\rm orb} = 0.352$). This
416: interval includes the phase of apastron, $\phi=0.5$, and the phase
417: of superior conjunction of J0537A, $\phi = 0.408$ at the epoch of
418: our {\sl Chandra} observation. The probability of getting zero
419: counts by chance in the 12,795 s interval is 0.0034 (assuming
420: Poisson statistics), i.e.\ the observed ``gap'' has a
421: $\approx3\,\sigma$ significance. For comparison, we plotted similar
422: distributions for the \chan\ and {\sl XMM-Newton} observations of
423: J0737 (middle and bottom panels of Fig.\ 6), which do not show a
424: statistically significant dependence on orbital phase (Fig.\ 7).
425:
426:
427:
428:
429:
430: \begin{figure}
431: \vbox{
432: \hspace{-1.0cm}
433: \center
434: \includegraphics[width=2.5in,angle=90]{f6a.eps}
435: \vspace{-0.3cm}
436: %\hspace{-1.0cm}
437: \includegraphics[width=2.5in,angle=90]{f6b.eps}
438: \vspace{-0.3cm}
439: %\hspace{-1.0cm}
440: \includegraphics[width=2.5in,angle=90]{f6c.eps}}
441: \caption{ Distributions of photon arrival times over binary phase
442: for the {\sl Chandra} ACIS observations of J1537 ({\em top}) and
443: J0737 ({\em middle}), and for the {\sl XMM-Newton} MOS1 and MOS2
444: observation of J0737 ({\em bottom}). Zero phase corresponds to
445: periastron passage. Only first 14.05 ks of the {\sl XMM-Newton}
446: exposure, least affected by the flare background, are used here. The
447: hatched area in the top panel shows the time interval excluded from
448: the GTIs. The dashed vertical lines show the phases of superior (S)
449: and inferior (I) conjunctions for J1537A and J0737A at the epochs of
450: the {\sl Chandra} and {\sl XMM-Newton} observations.
451: }
452: \end{figure}
453:
454:
455: \begin{figure}
456: \center
457: \includegraphics[width=2.5in,angle=90]{f7.eps}
458: \caption{ X-ray light curve of J0737 folded with the binary period.
459: We used both {\sl Chandra} ACIS and {\sl XMM-Newton} MOS data taken
460: over the length of a single binary revolution in each case (most of
461: the remaining 51-ks-long {\sl XMM-Newton} exposure suffered from
462: strong background flares.). The horizontal dotted line shows the
463: background level.
464: }
465: \end{figure}
466:
467: \begin{figure}
468: \centering
469: \includegraphics[width=3.4in,angle=0]{f8.eps}
470: \caption{Configuration of the J1537 binary at the epoch of the {\sl
471: Chandra} observation (53,470 MJD). View from above the orbital plane
472: with the Earth to the right is shown. The ellipses show the orbits
473: of J1537A and J1537B around the center of mass of the system. The
474: filled and open circles show the positions of the components at
475: superior and inferior conjunction of J1537A. The vertical solid line
476: is the line of nodes, and the dashed straight line is the apsidal
477: line. The ticks on the J1537B's orbit show the distribution of the
478: 16 detected photons over the orbit, while the shaded part of the
479: orbit (around the line of nodes) corresponds to the excluded time
480: interval at the beginning of the observation. (Notice that the
481: distribution of photons over the orbit looks different from that
482: over phases in Fig.\ 6 because of the substantial eccentricity; in
483: particular, the gap near the periastron is much larger because of
484: the increased orbital velocity.) The shaded sectors show the ranges
485: of true anomalies $\theta$ (counted clockwise from periastron) in
486: which the lack of detected photons could be explained by a
487: misalignment of the A's spin axis and the binary angular momentum,
488: assuming an equatorial outflow of the A's wind (see \S4.1.1). }
489: \end{figure}
490:
491:
492: \section{Discussion.}
493:
494: The small number of detected counts does not allow one to determine
495: the nature of the X-ray emission from J1537 unequivocally.
496: Among viable models are intrinsic emission
497: from the surface or magnetosphere of the NS(s) (just as in
498: solitary pulsars), the synchrotron emission resulting from the
499: interaction between the pulsar winds or between the wind of J1537A
500: and the magnetosphere of J1537B, and
501: emission from the
502: inner magnetosphere or surface of J1537B
503: triggered by the relativistic particles from the J1537A's wind.
504: Below we briefly discuss these interpretations,
505: with emphasis on
506: different dependences of the X-ray flux on binary phase.
507:
508:
509: \subsection{
510: Phase-dependent X-ray emission powered by the wind
511: from pulsar A}
512:
513: The
514: apparent lack of J1537's X-ray emission near the apastron suggests that a
515: phase-dependent interaction mechanism is responsible for the X-ray
516: emission in close DNSBs.
517: The interaction
518: can proceed via several different
519: mechanisms, and the {\em pulsar winds} are expected to
520: play an important role in each of them. Below we consider
521: plausible
522: scenarios.
523:
524:
525: \subsubsection{
526: Interaction of pulsar A's wind with the outer magnetosphere of pulsar B}
527:
528: As discussed in many papers on pulsars in binaries
529: (e.g., Arons \& Tavani 1993),
530: phase-dependent
531: synchrotron emission
532: can be
533: produced
534: when the
535: pulsar's wind collides with the wind or magnetosphere of
536: the secondary companion.
537: For a DNSB,
538: the position of the interaction site
539: (e.g., of the bow-shock head) on the line connecting
540: the two pulsars
541: can be estimated by balancing the
542: pressure of the pulsar A's wind with the pressure of pulsar B's
543: wind
544: or magnetosphere:
545: \begin{eqnarray}
546: \frac{\dot{E}_{A} f_A}{4\pi c (d_{AB}-r_B)^2 c}
547: \approx\frac{\dot{E}_B}{4\pi c r_B^2} \left\{ \begin{array}{ll}
548: f_B & \mbox{if $r_B \gg R_{lc,B}$}\\
549: g_B \left(\frac{R_{lc,B}}{r_B}\right)^{4} & \mbox{if $r_B \ll R_{lc,B}$}
550: \end{array} ,
551: \right.
552: \end{eqnarray}
553: where $d_{AB}$ is the distance between the binary
554: companions, $r_B$ is the distance between the interaction site and
555: pulsar B, and $R_{lc,B}=4.8\times 10^9 P_B$ cm is the light
556: cylinder radius for pulsar B (approximately equal to
557: the size of undisturbed magnetosphere).
558: The factors $f_A$
559: and $f_B$
560: take into account possible
561: anisotropy of the pulsar winds
562: ($f=1$ for an isotropic wind). They depend on the
563: colatitudes,
564: $\vartheta_A$ and $\vartheta_B$,
565: of the wind interaction site with respect to the A's and B's
566: spin axes because the pulsar wind should be
567: axially symmetric;
568: for instance,
569: $f = (3/2) \sin^2\vartheta$ for an angular distribution
570: similar to that of magnetic dipole radiation
571: (cf.\ Arons et al.\ 2004), or
572: \begin{eqnarray}
573: f =\left\{ \begin{array}{ll}
574: (\sin \Delta)^{-1} & \quad \mbox{if $|\vartheta -\pi/2| < \Delta$}\\
575: 0 & \quad \mbox{otherwise}
576: \end{array}
577: \right.
578: \end{eqnarray}
579: for an equatorial outflow with a uniform energy flux
580: within sharp boundaries.
581: The factor $g_B=(3/4)(1+3\cos^2\Theta_B)$
582: takes into account the
583: dependence of
584: the magnetic pressure on the polar angle $\Theta_B$ with respect
585: to the B's magnetic axis.
586:
587: Since B is expected to be
588: a slowly rotating, old (perhaps even dead) pulsar,
589: we can assume
590: $\dot{E}_{B}\ll \dot{E}_{A}$
591: (e.g., $\dot{E}_B/\dot{E}_A = 2.8\times
592: 10^{-4}$ for J0737),
593: which gives
594: \begin{eqnarray}
595: r_B \sim \left\{ \begin{array}{ll}
596: d_{AB}(\dot{E}_{B}f_B/\dot{E}_{A}f_A)^{1/2} & \quad \mbox{for $r_B>R_{lc,B}$}\\
597: d_{AB}^{1/3}R_{lc,B}^{2/3}(\dot{E}_{B}g_B /\dot{E}_{A}f_A)^{1/6} & \quad \mbox{for $r_B<R_{lc,B}$}
598: \end{array} .
599: \right.
600: \end{eqnarray}
601: The interaction may occur outside or
602: inside the B's light cylinder, depending on the separation
603: and the unknown properties of
604: pulsar B. In particular, $r_B < R_{lc,B}$
605: (i.e., the B's magnetosphere is compressed by the A's wind
606: on the ``dayside'')
607: if
608: \be
609: P_B > 0.66\, d_{AB,11} (\dot{E}_{B,30} g_B/\dot{E}_{A,33} f_A)^{1/2}\, {\rm s},
610: \ee
611: which corresponds to
612: \begin{eqnarray}
613: P_B > \left\{ \begin{array}{ll}
614: 0.3\, (g_B/f_A)^{1/2}\,{\rm s} & \quad \mbox{for J0737}\\
615: 1.1\,(d_{AB}/a) (\dot{E}_{B,30} g_B/f_A)^{1/2}\,{\rm s} & \quad \mbox{for J1537}
616: \end{array}\, ,
617: \right.
618: \end{eqnarray}
619: where
620: $a$ is the semimajor axis of the binary relative orbit
621: ($a=2.28\times 10^{11}$ cm and $8.78\times 10^{10}$ cm for J1537
622: and J0737, respectively).
623:
624: The shocked A's wind can generate
625: synchrotron X-ray emission if the wind particles
626: are sufficiently energetic and the magnetic field
627: $B$ is strong
628: enough in the emission region:
629: \be
630: \gamma^2 B \gtrsim
631: 10^{12} (E/5\,{\rm keV})\,,
632: \ee
633: where $\gamma$
634: is the Lorentz factor,
635: and $E$ is the photon energy.
636: For the plausible case $r_B < R_{lc,B}$, the magnetic field at
637: the interaction site can be estimated from the pressure balance
638: between the A's wind and the B's magnetic pressure:
639: \begin{eqnarray}
640: B =
641: \left(\frac{2\dot{E}_Af_A}{c d_{AB}^2}\right)^{1/2}
642: =\frac{a}{d_{AB}} f_A^{1/2} \left\{ \begin{array}{ll}
643: 7.1\, {\rm G} & \quad \mbox{for J0737}\\
644: 1.5\, {\rm G} & \quad \mbox{for J1537}
645: \end{array} \, .
646: \right.
647: \end{eqnarray}
648: At such fields, we need electrons with Lorentz
649: factors $\gtrsim 10^5$--$10^6$ to produce synchrotron X-rays.
650: The recent pulsar wind models (e.g., Kirk \& Skj{\ae}raasen 2003 and
651: references therein)
652: suggest that the initially
653: Poynting-flux-dominated wind of pulsar A
654: converts most of its electromagnetic energy into particle
655: kinetic energy at a distance
656: much greater than the binary separation, which implies that
657: the bulk Lorentz factor upstream of the interaction site
658: remains the same as acquired in the pulsar A's magnetosphere.
659: Although the predictions of the current pulsar models
660: are rather uncertain in this regard (in particular,
661: the Lorentz factor depends on poorly known ``multiplicity''
662: of the pair cascades; e.g., Hibschman \& Arons 2001),
663: the required values of $\gamma$ are likely too large
664: to be produced in the A's magnetosphere, which means
665: that additional acceleration must occur at the interaction site
666: (e.g., caused by reconnection of the magnetic fields of the
667: A's wind and B's magnetosphere, similar to the magnetic reconnection
668: at the interaction of the solar wind with the Earth's magnetosphere).
669:
670: If the shocked/accelerated A's wind is the sole source of the X-ray emission,
671: the X-ray luminosity should be
672: a fraction of the A's wind power:
673: \be
674: L_{\rm X} = \epsilon_\Omega \epsilon_e \epsilon_{\rm rad} \epsilon_{\rm X} \dot{E}_A\,
675: \ee
676: where $\epsilon_\Omega\approx f_A (r_\perp/2d_{AB})^2$ is the
677: fraction of A's power intercepted by the B's magnetosphere
678: with a transverse size $r_\perp$ that depends on geometry
679: of the interaction region (crudely, $r_\perp \sim R_{lc,B}$),
680: $\epsilon_e$ is the fraction of energy of the shocked/accelerated
681: wind carried by particles (electrons and/or positrons),
682: $\epsilon_{\rm rad} \approx {\rm min}(1;t_{\rm flow}/t_{\rm rad})$
683: is the radiative efficiency, i.e.
684: the fraction of intercepted energy emitted as (synchrotron)
685: radiation ($t_{\rm rad}$ and $t_{\rm flow}$ are the time of
686: synchrotron losses and the flow time in the emission region),
687: and $\epsilon_{\rm X}$ is the fraction of synchrotron power
688: emitted in the X-ray range. The intercepted fraction can be
689: estimated as $\epsilon_\Omega \sim 6\times 10^{-3} f_A
690: (r_\perp/R_{lc,B})^2 $ and $\sim 5\times 10^{-4} f_A
691: (r_\perp/10^{10}\,{\rm cm})^2 (a/d_{AB})^2$ for J0737 and J1537,
692: respectively. The fraction $\epsilon_e$ depends on acceleration
693: mechanism (e.g., magnetic reconnection); the latter is currently
694: unknown, but the assumption that $\epsilon_e$ is not much lower than
695: unity seems plausible. The fraction $\epsilon_{\rm rad}$ depends on
696: the particle energy spectrum, geometry of emitting region, and flow
697: velocity. For a particle of given energy, the time of synchrotron
698: losses is $\tau_{\rm rad} = 510\, \gamma_6^{-1} B^{-2}\, {\rm s}$,
699: while the flow time is $t_{\rm flow} \sim l/v_{\rm flow} \sim 0.3\,
700: l_{10} \beta^{-1}\,{\rm s}$, where $l = 10^{10} l_{10}\,{\rm cm}$ is
701: the travel length, and $\beta = v_{\rm flow}/c$. This gives
702: $\epsilon_{\rm rad} \sim {\rm min}(1; 7\times 10^{-4} l_{10}\gamma_6
703: B^2 \beta^{-1})$. We see that electrons emit a very small fraction
704: of their energy unless the travel length substantially exceeds a
705: typical size of the B's magnetosphere. Finally, the fraction
706: $\epsilon_{\rm X}$ cannot be reliably estimated without knowing the
707: particle energy distribution. However, even for $\epsilon_{\rm X}$
708: (and $\epsilon_e$) close to unity, the above estimates of
709: $\epsilon_\Omega$ and $\epsilon_{\rm rad}$ suggest the A's wind
710: should be substantially anisotropic ($f_A\gg 1$) and/or the emitting
711: particles should spend a long time ($\gtrsim 10^2\,{\rm s}$) in the
712: emitting region to explain the ratio $L_{\rm X}/\dot{E}_A \sim
713: 3\times 10^{-4}$ observed\footnote{We should note that the $L_{\rm
714: X}$ estimates in \S3 are based on the assumption of isotropic X-ray
715: emission, which is not necessarily true; the actual $L_{\rm X}$ can
716: be lower or higher than estimated.}
717: in J0737 and J1537.
718:
719: The interaction mechanism and the properties of the A's wind can be
720: inferred from the analysis of the dependence of the X-ray emission
721: on orbital phase. The collision of the A's wind with the B's
722: magnetosphere is expected to form a bow-shock, convex toward pulsar
723: A, with the shocked A's wind flowing in a magnetosheath engulfing
724: the B's magnetosphere (Arons et al.\ 2004; Lyutikov 2004).
725: It is tempting to explain the
726: X-rays observed from J0737 and J1537 as synchrotron emission from
727: the shocked A's wind in the magnetosheath (as discussed by Granot \&
728: M\'esz\'aros 2004). Such an explanation implies an orbital
729: dependence of the X-ray flux caused by the relativistic beaming
730: (Doppler boost), showing a minimum flux at the phase of inferior
731: conjunction of pulsar A, when the wind in the sheath is flowing away
732: from the observer, and a maximum flux around the superior
733: conjunction. For a reasonable flow speed, $v_{\rm flow} = \beta c$,
734: the Doppler boost modulation can be quite large, up to $F_{\rm
735: max}/F_{\rm min} = [(1+\beta)/(1-\beta)]^{\Gamma +2}$, especially at
736: the observed large photon index, $\Gamma \sim 3$. However, the
737: observed distributions of photons over orbital phase do not show a
738: statistically significant flux deficit at the phase of A's inferior
739: conjunction, in neither J0737 nor J1537 (see Figs.\ 6 and 7).
740: Therefore, such an explanation can be ruled out. It seems surprising
741: that no orbital modulation caused by the anisotropy of the flow in
742: the putative magnetosheath is seen (unless, of course, our
743: assumption, that the X-rays are produced by interaction of the A's
744: wind with the B's magnetosphere, is wrong). We can only speculate
745: that either the magnetosheath does not form because the A's wind is
746: strongly magnetized or the X-ray emitting particles accelerated by
747: magnetic reconnection do not flow into the magnetosheath.
748:
749: Having ruled out the magnetosheath as the source of the observed
750: X-rays, we can, nevertheless, suggest two
751: explanations for the observed orbital dependence of X-ray emission
752: in J1537, such that they do not contradict to the lack of such
753: dependence in J0737. First, this difference may be connected with
754: the fact that J1537 has a substantial eccentricity while J0737 is on
755: an almost circular orbit. At apastron, the separation $d_{AB}$ is
756: larger than at periastron by a factor $(1+e)/(1-e)$, which is 1.755
757: for J1537, vs.\ 1.192 for J0737. The decrease of the magnetic field
758: and particle number density, $n_e$, at the interaction site with
759: increasing separation should lead to a lower X-ray luminosity at
760: apastron in eccentric DNSBs. Crudely, assuming $\epsilon_{\rm X}$
761: does not depend on $d_{AB}$ substantially,
762: the luminosity behaves as
763: $L_{\rm X} \sim \epsilon_{\rm X} n_e V {\cal P}_{\rm syn} \propto
764: V d_{AB}^{-4}$,
765: where ${\cal P}_{\rm syn} \propto B^2 \propto
766: d_{AB}^{-2}$ is the total synchrotron power per electron , and
767: $V$
768: is the emission volume. The dependence of the latter on $d_{AB}$ is
769: rather uncertain, but it seems reasonable to assume
770: $V\propto r_B^3$. Under this assumption,
771: we obtain $V\propto d_{AB}^3$ or $V\propto d_{AB}$ for
772: $r_B> R_{lc,B}$ or $r_B< R_{lc,B}$, respectively, so the increase of $V$
773: with $d_{AB}$
774: does not compensate for the decrease caused by decreasing $B$ and $n_e$.
775: In particular, $L_{\rm X} \propto d_{AB}^{-3}$
776: for $r_B< R_{lc,B}$, which means that we can expect
777: the luminosity at apastron to be a factor of 5.4 lower than
778: at periastron, for J1537.
779: Although the effect of varying binary separation can explain
780: the observed deficit around apastron in J1537,
781: we do not see an increased count rate at periastron,
782: predicted by this model;
783: however, we cannot rule it out because of the scarce statistics
784: of our data.
785:
786: In addition to the varying binary separation,
787: the phase dependence of the X-ray flux in J1537 can be caused by
788: a misalignment between the binary orbital plane and the equatorial
789: plane of pulsar A, where most of the A's wind is expected to be
790: confined. In this case, one should expect brighter X-ray emission
791: at orbital phases when pulsar B plunges in the dense wind flowing in the
792: equatorial plane of pulsar A.
793: For instance, if the equatorial outflow is confined
794: between two conical surfaces, within an angle
795: $\pm\Delta$ from the equatorial plane, and the A's spin axis
796: is inclined at an angle $\delta$ to the direction of the orbital
797: angular momentum,
798: then, at $\delta > \Delta$,
799: the wind ``misses'' pulsar B in two segments of the orbit:
800: $|\psi| < \eta$ and $|\pi -\psi| < \eta$, where $\psi$ is
801: the azimuthal angle
802: counted from the
803: projection of the A's spin
804: onto the orbital plane, and $\eta$ is given by the equation
805: \be
806: \cot\eta = \tan\Delta\,
807: (\tan^2\delta -\tan^2\Delta)^{-1/2} (\cos\delta)^{-1}\,.
808: \ee
809: In other words, one should expect a photon deficit in two
810: phase intervals corresponding to two sectors symmetric with
811: respect to the A's spin projection onto the orbital plane.
812: If such a phase dependence is observed, then $\eta$
813: can be measured, and $\Delta$ can be found from
814: the equation
815: \be
816: \tan^2\Delta = \cot^2\eta\,\sin^2\delta\,(1+\cot^2\eta\,\cos^2\delta)^{-1}.
817: \ee
818: For J1537, we determined the true anomaly $\theta$ for each of the 16 events
819: and marked the events on the J1537B orbit in Figure 8. Because of the
820: higher (lower) NS velocities near
821: the periastron (apastron), a relatively
822: small phase segment
823: around $\phi =0$
824: with no photons detected, $-0.114 < \phi < 0.043$,
825: translates into a sector of true anomalies,
826: $-68^\circ < \theta < 28^\circ$,
827: whose width, $\Delta\theta_{\rm per} = 95^\circ$,
828: is even larger than
829: that of the photon-free sector at apastron, $\Delta\theta_{\rm ap}=80^\circ$
830: ($150^\circ < \theta < 231^\circ$).
831: If we interpret the photon deficit
832: in these sectors as
833: caused by the anisotropy of the wind
834: outflow, then, from the symmetry requirement, the sectors in which
835: the J1537A's wind misses J1537B
836: (shaded in Fig.\ 8) are
837: $-30^\circ < \theta < 28^\circ$
838: and
839: $150^\circ < \theta < 208^\circ$, i.e.,
840: the projection of the A's spin onto the orbital plane
841: almost coincides with the apsidal line,
842: and $\eta \leq 29^\circ$.
843: Since $\delta = 25\degr\pm4\degr$ for J1537 (Stairs et al.\ 2002),
844: equation (10) gives a lower limit on the half-opening angle
845: of the equatorial outflow, $\Delta \geq 18^\circ$--$25^\circ$.
846: Being consistent with the observed phase dependence,
847: such an explanation
848: is also attractive because it
849: implies an increased particle and energy supply in
850: the two segments of orbit where the J1537A's wind encounters J1537B
851: ($f_A$ can be as large as 2.4--3.2, according to eq.\ [2]), a higher
852: magnetic field at the interaction region (by a factor of
853: $f_A^{1/2}$), and, correspondingly, a smaller ratio $t_{\rm
854: flow}/t_{\rm rad}$ (by a factor of $f_A$). These consequences of the
855: wind anisotropy alleviate the energetic and geometric restrictions
856: inherent in the interpretation of the X-ray emission as caused by
857: the interaction of the A's wind with the B's magnetosphere. We
858: should emphasize, however, that the X-ray flux deficit around
859: periastron is not statistically significant, and this hypothesis can
860: be confirmed only by a deeper observation.
861: For J0737,
862: the angle $\delta$ is rather uncertain ($0 < \delta < 60\degr$;
863: Manchester et al.\ 2005), so one can assume that the lack of a clear
864: orbital phase dependence means that $\delta < \Delta$, i.e., the
865: J0737A's wind interacts with J0737B throughout the whole orbit.
866:
867: In addition to the energetics and geometry (luminosity and its phase
868: dependence), some information on the interaction mechanism can be
869: obtained from the X-ray spectra. In this regard, we should note that
870: the spectra observed in J1537 and J0737
871: are surprisingly soft
872: ($\Gamma \sim 3$) in comparison with, e.g., the spectra
873: observed in pulsar wind nebulae ($\Gamma \sim 1$--2).
874: The models of particle acceleration by ultrarelativistic MHD shocks
875: also suggest much flatter spectra,
876: $\Gamma\simeq1.6$, corresponding to the slope $p\simeq2.2$
877: of the particle energy spectrum
878: (Achterberg et al.\ 2001). We can speculate, however, that
879: acceleration via relativistic magnetic reconnection might produce
880: softer spectra. Also, we do not exclude the possibility that
881: the observed spectra consist of two components: a hard PL component
882: from the shocked wind and a soft thermal component from the polar caps
883: of pulsar A (see \S4.2).
884:
885:
886: \subsubsection{
887: Emission from B's inner magnetosphere and/or surface
888: induced by captured A's wind}
889: There is a possibility that some fraction of the A's
890: wind
891: is captured by the B's magnetosphere
892: (e.g., similar to capturing solar wind in the Earth magnetosphere).
893: The captured wind can increase the magnetospheric pair density
894: or even precipitate onto the NS surface
895: (such a possibility has been discussed by Zhang \& Loeb 2004
896: to explain
897: the flaring radio emission of J0737B).
898: The captured relativistic particles lose a substantial fraction of their
899: energy to synchrotron and inverse Compton radiation, thereby increasing
900: the radiative efficiency, $\epsilon_{\rm rad}$,
901: of the intercepted wind. The particles
902: traveling along the open field lines may reach the NS polar caps and
903: heat them to X-ray temperatures, up to a few million kelvins,
904: so that we could see not only nonthermal but also thermal X-ray emission,
905: which might explain the softness of the X-ray spectra observed in
906: J0737 and J1537. The total luminosity from these emission mechanisms
907: is, of course, a small fraction of $\dot{E}_A$; it can be described
908: by an equation similar to equation (8) but with an additional factor
909: in its right-hand side, $\epsilon_{\rm cap}$, which is the fraction of
910: intercepted wind particles captured by the B's magnetosphere.
911: Similar to the emission produced by the wind interaction with the
912: outer magnetosphere (\S4.1.1), the X-ray luminosity should decrease
913: with increasing $d_{AB}$ (i.e., it should be lower near apastron)
914: and vanish during the phase intervals when pulsar B is far from the
915: equatorial plane of pulsar A\footnote{We should mention that such
916: orbital phase dependencies are expected not only for the nonthermal
917: but also for the thermal component because the
918: polar cap cooling time scales
919: are very short (about a few microseconds for $B\sim 10^{12}$ G;
920: e.g., Gil et al.\ 2003).}. In addition, the fraction $\epsilon_{\rm
921: cap}$ should depend on the angle between the B's magnetic axis and
922: the line connecting the two pulsars. This may lead to an additional
923: modulation of $L_{\rm X}$ with the orbital period and the period of
924: pulsar B
925: unless the B's magnetic
926: axis is perpendicular to the orbital plane, as expected if the main
927: braking mechanism for pulsar B is its interaction with the A's
928: wind (Arons et al.\ 2004; Lyutikov 2004).
929: In addition, the X-ray emission caused by the captured wind
930: should show pulsations with the B's spin period.
931:
932: There is also an interesting possibility that the wind's particles
933: that reach the B's surface and heat the polar caps can facilitate
934: the extraction of electrons or ions from the polar cap surface by
935: the pulsar's electric field, triggering the pair cascade in pulsar
936: B. In this case, the A's wind works as a catalyst of B's X-ray (and
937: possibly radio) emission, providing an additional luminosity
938: associated with the loss of the pulsar B's spin energy\footnote{If
939: such a mechanism works, we should expect that $\dot{E}_B$ (and
940: $\dot{P}_B$) are increased during the ``activation phases''.}. The
941: orbital phase dependence of this luminosity should be similar to
942: that of the above-discussed captured wind. In particular, even if B
943: is a dead pulsar when no wind blows on it (e.g., when it is far from
944: the A's equatorial plane), it may become ``resurrected'' on the
945: orbit segments where it captures the wind.
946:
947: Quantitative predictions of the properties of the X-ray emission
948: induced by the captured wind are highly uncertain, depending on
949: the unknown wind properties,
950: structure of the magnetic field near the interaction
951: region, etc. We can crudely estimate the number of A's wind particles
952: captured by B per unit time:
953: $\delta\dot{N}_{A}=\kappa_A\dot{N}_{{\rm GJ},A}
954: \epsilon_\Omega \epsilon_{\rm cap}$, where
955: $\dot{N}_{{\rm GJ},A} \sim
956: (3\dot{E}_Ac/2e^2)^{1/2}$
957: is
958: the Goldreich-Julian rate of particle ejection from the A's magnetosphere,
959: $\kappa_A$ the pair production multiplicity, and
960: $\epsilon_\Omega \sim f_A (r_B/2d_{AB})^2$.
961: Assuming $r_B < R_{lc,B}$, we obtain
962: $\delta \dot{N}_A \sim 2.2\times 10^{27} \kappa_A \epsilon_{\rm cap}
963: (a/d_{AB})^{4/3} P_B^{4/3} (f_A^2g_B \dot{E}_{B,30})^{1/3}\,{\rm s}^{-1}$
964: for J1537, and $\delta \dot{N}_A \sim 4.4\times 10^{28} \kappa_A
965: \epsilon_{\rm cap} (f_A^2 g_B)^{1/3}\,{\rm s}^{-1}$ for J0737.
966: If the captured wind is the only source of X-rays, the product
967: $\kappa_A \epsilon_{\rm cap}$ should be large enough to provide the
968: observed X-ray luminosity: e.g., $\kappa_A \epsilon_{\rm cap} >
969: 3\times 10^3 (f_A^2 g_B)^{-1/3} P_{B}^{-4/3} \dot{E}_{B,30}^{-1/3}
970: (a/d_{AB})^{-4/3} (10^5 m_ec^2/\bar{\cal E})$ in J1537
971: and $\kappa_A \epsilon_{\rm cap} >
972: 0.8\times 10^3 (f_A^2 g_B)^{-1/3} (10^5 m_ec^2/\bar{\cal E})$ in J0737,
973: where $\bar{\cal E}$ is an average energy per captured particle
974: reradiated in X-rays (Zhang \& Loeb 2004 assume $\kappa_A \sim
975: 10^6$, $\epsilon_{\rm cap} \sim 0.1$ in their estimates for J0737;
976: as we have mentioned above, at such large $\kappa_A$ no X-rays can be
977: expected from the wind that was not captured). For $\kappa_A \sim
978: 10^2$ obtained in standard cascade theory, we have to invoke the
979: above-mentioned activation of pulsar B by the captured wind to
980: explain the observed energetics.
981:
982: Note that a large value of $\kappa_A\epsilon_{\rm cap}$
983: is also required for $\delta \dot{N}_{A}$ to exceed
984: the intrinsic B's pair production rate,
985: $\dot{N}_{B}=\kappa_B\dot{N}_{{\rm GJ},B}=
986: 4.4\times 10^{29}\dot{E}_{B,30}^{1/2}\kappa_{B}\,{\rm s}^{-1}$.
987: It occurs
988: when
989: $\kappa_A\epsilon_{\rm cap}/\kappa_B >
990: 200\, (d_{AB}/a)^{4/3} P_B^{-4/3} (f_A^2g_B)^{-1/3} \dot{E}_{B,30}^{1/6}\, {\rm s}^{-1}$
991: and
992: $13\, (f_A^2 g_B)^{-1/3}$ for J1537 and J0737, respectively.
993:
994: Thus, we can
995: conclude that energizing the B's magnetosphere or heating
996: its surface by the particles from the A's wind
997: can explain the observed dependence of the X-ray
998: flux on binary phase. However, it would require
999: a rather high pair production multiplicity
1000: in pulsar A and considerable anisotropy of the A's wind.
1001:
1002: \subsection{
1003: X-ray emission from individual pulsars}
1004:
1005:
1006: Since solitary pulsars are known to emit X-rays, at least some
1007: fraction of the detected X-rays could come from one or both of the
1008: DNSB components, even if the X-ray emission caused by the
1009: interaction of the A's wind with the B's magnetosphere were
1010: negligible. In particular, MSPs can exhibit both nonthermal
1011: (magnetospheric) and thermal (polar cap) components in their X-ray
1012: spectra, the latter dominating in MSPs with $\dot{E} \lesssim
1013: 10^{34}$ ergs s$^{-1}$. Typical thermal conversion efficiencies,
1014: $\eta_{\rm th}\equiv L_{\rm bol}/\dot{E}$, are
1015: (0.3--$3)\times10^{-4}$ (Zavlin 2006, and references therein). The
1016: efficiencies inferred from the BB fits of the J1537 and J0737
1017: spectra, $6.1\times 10^{-5}$ and $4.6\times
1018: 10^{-5}\langle\cos\varsigma\rangle^{-1} D_{500}^2$, respectively,
1019: are within that range\footnote{Since the orientations of the axes is
1020: poorly known for J0737A (Demorest et al.\ 2004; Manchester et al.\
1021: 2005), the value of $\langle\cos\varsigma\rangle$ is very
1022: uncertain.}. The polar cap radii derived from the BB fits, are
1023: smaller than the conventional PC radii [$r_{\rm pc} \equiv (2\pi
1024: R_{NS}^{3}/cP)^{1/2} = 740$ m and 960 m, respectively), similar to
1025: other pulsars (e.g., Zavin \& Pavlov 2004; Zavlin 2006). Thus, the
1026: luminosities and the soft spectra of J1537 and J0737 are consistent
1027: with being emitted from polar caps of J1537A and J0737A, but of
1028: course such emission would not show any dependence on binary phase.
1029:
1030: The X-ray emission in J1537 might also originate
1031: from J1537B if this NS is not a ``dead pulsar'', i.e.\ it is
1032: still capable of producing $e^{+}e^{-}$ pairs
1033: emitting
1034: curvature and synchrotron radiation in the magnetosphere and heating
1035: its polar caps. Indeed, the X-ray spectrum of J1537 resembles the
1036: spectra of
1037: old
1038: ($\tau \sim 1$--10 Myr) solitary radio pulsars (e.g., PSR B0950+08;
1039: Zavlin and Pavlov 2004). However, the J1537's luminosity (e.g.,
1040: $L_X= 8.5\times 10^{28}$ ergs s$^{-1}$ in the 1--10 keV band, for
1041: the PL fit) is rather high for such an old pulsar\footnote{Based on
1042: the J1537's proper motion and distance from the Galactic plane,
1043: Thorsett et al.\ (2005) argue that the SN explosion that produced
1044: J1537B occurred $\gtrsim 10$ Myrs ago.}; it would either imply a
1045: high spin-down luminosity, $\dot{E}_{\rm J1537B}\gtrsim 10^{32}$
1046: ergs s$^{-1}$,
1047: or suggest that the X-ray efficiencies of old
1048: pulsars strongly grow with decreasing $\dot{E}$,
1049: which is hardly supported
1050: by observations
1051: (cf.\ Fig.\ 5 in Kargaltsev
1052: et al.\ 2006).
1053: Moreover, such an interpretation is firmly excluded for J0737,
1054: whose X-ray luminosity exceeds $\dot{E}_{\rm J0737B} = 1.6\times 10^{30}$
1055: ergs s$^{-1}$.
1056:
1057: Thus, if the observed J1537's emission originates from the
1058: surface or magnetosphere of one of the binary companions, J1537A
1059: is a much more viable candidate than J1537B.
1060: The assumption that most of the observed X-ray emission is generated
1061: at the polar caps (or magnetosphere) of J1537A would be strongly
1062: supported if X-ray pulsations at the J1537A period are found in future
1063: observations.
1064: An indirect support for this interpretation would also be provided by
1065: detection of X-ray pulsations from J0737 at the
1066: J0737A's period.
1067: On the other hand, confirming the reduced count rate near
1068: apastron would virtually rule out this interpretation
1069: and put an interesting upper limit on the MSP luminosity.
1070:
1071:
1072: \subsection{Emission from the A's wind interacting with the interstellar
1073: medium}
1074:
1075: The velocities of J1537 and J0737 in the plane of the sky,
1076: $v_{\perp}
1077: \simeq 120$ and $70$
1078: km s$^{-1}$, respectively,
1079: are large enough to assume that these DNSBs
1080: are moving through the ISM
1081: with supersonic speeds, i.e., $v >
1082: c_s = 15\, (\mu/0.6)^{-1/2} T_4^{1/2}$ km s$^{-1}$,
1083: where $\mu$ and $T=10^4 T_4$ K are the molecular weight and
1084: temperature. At a supersonic speed, the pulsar's wind forms a
1085: termination shock (TS) in the ISM. For an isotropic wind,
1086: the TS acquires
1087: a bullet-like shape (Bucciantini, Amato, \& Del Zanna 2005),
1088: with a distance $R_{\rm h} \simeq (\dot{E}_A/4\pi c \rho v^2)^{1/2}
1089: =4.0\times 10^{15} \dot{E}_{A,33}^{1/2} n^{-1/2} v_7^{-1}$ cm
1090: between the pulsar and the bullet head
1091: [here $n=\rho/(1.66\times 10^{-24}\,{\rm g})$ is
1092: the ISM number density in atomic mass units per cm$^3$,
1093: and $v_7$ is the pulsar speed in units of 100 km/s].
1094: For large enough $\gamma^2 B$ in the shocked plasma
1095: (see equation [6]),
1096: it can generate X-rays,
1097: as discussed by Granot \& M\'esz\'aros (2004) for J0737.
1098:
1099: The interpretation of the observed X-rays as generated by the
1100: A's wind interaction with the ISM rather than with the B's magnetosphere
1101: (\S4.1) has two apparent advantages for explaining the energetics
1102: (but not the putative dependence on binary phase in J1537):
1103: a much larger fraction of the A's wind is intercepted
1104: (e.g., $\epsilon_\Omega \sim 0.5$),
1105: and, possibly, a larger fraction of the A's wind energy is
1106: carried by particles ($\epsilon_e$ is closer to unity)
1107: because a larger fraction of the wind's magnetic energy
1108: can be converted into particle energy at such large distances,
1109: $R_{\rm h}\gg d_{AB}$.
1110: On the other hand, this interpretation may have its own
1111: problems, even with the X-ray energetics.
1112: First of all,
1113: since the wind's magnetic field is much lower
1114: at the TS than at the site of interaction
1115: with the B's magnetosphere,
1116: the synchrotron
1117: power and radiative efficiency, $\epsilon_{\rm rad}$, are much lower,
1118: which requires much higher particle energies for the radiation
1119: to be emitted in the X-ray range.
1120: Assuming that the magnetization parameter (i.e., the ratio of the Poynting
1121: flux to the kinetic energy flux) is small, $\sigma \ll 1$, we can,
1122: following Kennel \& Coroniti (1986), estimate
1123: the post-shock magnetic field at the shock head:
1124: $B \sim 3\sigma^{1/2} (4\pi\rho v^2)^{1/2} =
1125: 14\, \sigma_{-2}^{1/2} n^{1/2}
1126: v_7\,\, \mu{\rm G}$, where $\sigma_{-2} =\sigma/10^{-2}$.
1127: For the synchrotron radiation in such a field to be emitted in the
1128: X-ray range, we should have enough electrons in the energy range
1129: $\gamma \sim (1$--$3)\times 10^8 \sigma_{-2}^{-1/4} n^{-1/4} v_7^{-1/2}$.
1130: Such energies are too high to be acquired
1131: in the pulsar
1132: magnetosphere.
1133: The wind electrons can be further accelerated
1134: in the pre-shock wind
1135: by the conversion of the electromagnetic
1136: wind's energy into kinetic energy. However,
1137: irrespective of the (poorly understood) conversion mechanism, the maximum energy
1138: is limited by the value
1139: $\gamma_{\rm max} = (e^2/2 m_ec^3) \dot{N}_{{\rm GJ},A} \kappa_A^{-1}$
1140: (see Lyubarsky \& Kirk 2001),
1141: which gives $\gamma_{\rm max} =0.9\times 10^8 \kappa_A^{-1}$ and
1142: $1.6\times 10^8 \kappa_A^{-1}$ for J1537A and J0737A, respectively.
1143: These limiting energies are sufficiently large only at
1144: very small multiplicities ($\kappa \lesssim 1 $ for reasonable
1145: $\sigma$, $n$, and $v$),
1146: much lower that $\kappa \sim 100$ expected from the pulsar models.
1147: Too low electron energies and magnetic
1148: fields may be responsible for the fact that no TSs
1149: around supersonically moving MSPs have been
1150: convincingly detected in X-rays
1151: (e.g., Zavlin et al.\ 2002), in contrast to
1152: forward bow shocks seen in H$_\alpha$.
1153:
1154: It is worth mentioning another problem with the wind-ISM interaction
1155: as the source of X-ray emission from
1156: close binary pulsars,
1157: such as J1537 and J0737. The conventional model for the shock
1158: (and the above consideration) assume that the pulsar's velocity
1159: relative to ISM is constant. However, a binary pulsar rotates
1160: around the center of mass, and the velocity of
1161: this motion may even exceed the binary systemic velocity.
1162: For instance, the A's orbital velocity varies between 150 and 263 km
1163: s$^{-1}$ in J1537 and between 275 and 328 km s$^{-1}$ for J0737.
1164: This means that both the absolute magnitude and the direction of the
1165: A's velocity relative to ISM vary with the orbital period. Moreover,
1166: using the pulsar's velocity relative to ISM in the above
1167: consideration would result in TS changing its position and
1168: orientation in the ISM (e.g., the TS head could appear {\it behind}
1169: the moving binary during some fraction of the binary period). On the
1170: other hand, the characteristic time of propagation of perturbations
1171: in the ultrarelativistic wind,
1172: $\gtrsim R_{\rm h}/c \sim 30$--100 hours, exceeds the
1173: binary period, so that we do not expect the wind-ISM interaction
1174: to create a rotating (or wobbling) TS ``bullet''. Nevertheless,
1175: even if this interaction results in a TS whose global properties,
1176: averaged over the orbital period, resemble those of a solitary
1177: pulsar moving with a constant speed $v$, this structure should respond to
1178: the perturbations caused by the fast binary motion, and these perturbations
1179: should manifest themselves in periodic variations in the radiation
1180: from the shocked plasma.
1181: However, no such variations have been seen in the X-ray radiation from J0737.
1182: If future X-ray observations of J0737 put more stringent limit on the orbital
1183: variation,
1184: it will be an additional argument against the interpretation
1185: in terms of the wind-ISM interaction.
1186:
1187:
1188:
1189:
1190:
1191:
1192:
1193: \section{Conclusions}
1194: We have detected the DNSB J1537 in X-rays, analyzed its X-ray spectrum
1195: and light curve,
1196: and compared the properties of its X-ray emission
1197: with those
1198: of the double pulsar J0737. Although the X-ray spectra and
1199: luminosities of these systems are similar, J1537 shows a gap in its
1200: light curve during an interval of 0.35 of orbital period around
1201: apastron while J0737 does not show show a significant orbital phase
1202: dependence. There are two viable interpretations of the X-ray
1203: emission: it can be caused by interaction of the pulsar A's wind
1204: with pulsar B (\S4.1), including the capture of the wind particles
1205: (\S4.1.2), or it can be emission from the magnetosphere or, more
1206: likely, hot polar caps of pulsar A ( similar to solitary MSPs with
1207: $\dot{E}< 10^{34}$ ergs s$^{-1}$; \S4.2). The former interpretation
1208: can explain the putative orbital phase dependence in J1537 if
1209: the A's wind is mostly concentrated in the equatorial plane
1210: around the A's spin axis
1211: inclined to the orbital plane. The phase dependence can also be
1212: associated with the large eccentricity of J1537 (much larger
1213: than in J0737) because the X-ray luminosity depends on binary
1214: separation. On the contrary,
1215: no orbital phase dependence is expected if the X-rays are produced
1216: by pulsar A.
1217:
1218: Unfortunately, the scarce count statistics
1219: does not allow us to discriminate between the different
1220: interpretations. Further observations of both J1537 and J0737, which
1221: have different orbital parameters and orientations with respect to
1222: the observer, should provide a clue to the nature of the X-ray
1223: radiation from DNSBs and the properties of pulsars and their winds.
1224:
1225: If future observations of J1537
1226: confirm the deficit of X-ray flux during a
1227: substantial part of
1228: orbit,
1229: this would prove that the X-rays are produced via interaction between
1230: the binary companions. If a similar deficit is observed at the
1231: opposite part of the orbit,
1232: it would strongly
1233: support the hypothesis that the A's wind is concentrated
1234: towards the equatorial plane, and it would allow
1235: one to measure the opening angle of the equatorial outflow.
1236: If, on the contrary, the X-ray flux
1237: shows a maximum at periastron, it would mean that the binary separation
1238: is the main cause of the variable X-ray flux.
1239: For J0737, the dependence of the
1240: X-ray flux on orbital phase should be much weaker due to
1241: a smaller angle between the orbital angular momentum
1242: and the A's spin or/and the much
1243: smaller eccentricity.
1244: Confirming the interaction with the B's outer magnetosphere as
1245: the source of X-rays would provide strong evidence that
1246: the A's wind is accelerated
1247: up to energies $\gtrsim 10^6 m_ec^2$ at the interaction site, likely
1248: by magnetic reconnection.
1249: In such a scenario, one can expect
1250: X-ray pulsations with B's period,
1251: especially if the wind is captured by pulsar B. If the emission spectrum
1252: is found to be softer at the pulsations maxima,
1253: we would conclude that a fraction
1254: of the captured A's wind is
1255: channeled onto B's polar
1256: caps, heating them to X-ray temperatures.
1257: Detection of radiation caused by the capture of the A's wind
1258: would put a stringent lower limit on pair multiplicities in
1259: pulsar A.
1260:
1261:
1262:
1263:
1264: If
1265: future observations rule out
1266: any orbital phase dependence in the DNSBs
1267: (including the flux deficit near apastron in J1537),
1268: but
1269: pulsations at
1270: the A's period are detected,
1271: then we will have to conclude that the X-ray emission is
1272: produced by pulsar A itself, while the interaction of the A's wind
1273: with pulsar B is insignificant.
1274: This would
1275: mean
1276: that conversion of the
1277: intercepted wind's power
1278: into X-ray luminosity is inefficient, possibly because
1279: the particles are not accelerated to high enough energies
1280: (which results in small values of $\epsilon_{\rm X}$ and $\epsilon_{\rm rad}$
1281: in eq.\ [8]).
1282: In this case, studying the X-ray spectrum and pulsations would
1283: better characterize pulsar A, allowing one to compare its
1284: properties with those of ``ordinary'' MSPs, which have shorter
1285: periods and larger characteristic ages.
1286:
1287: Finally, it is quite plausible that the observed emission
1288: consists of two components: a soft thermal component from pulsar A
1289: and a hard nonthermal component from the A's wind interaction with
1290: pulsar B, as suggested by the X-ray spectra of J0737 and J1537.
1291: If this is the case, we expect to see periodic variations
1292: with both the orbital and A's periods. Also, the relative contributions
1293: of these two components are expected to vary with orbital phase.
1294: The changes can be seen through phase-resolved spectroscopy
1295: because the wind emission is expected to be substantially harder than
1296: the MSP emission.
1297:
1298: To conclude, further deep
1299: observations of J1537 and J0737, and perhaps other DNSBs,
1300: would be of great interest
1301: as they will constrain both the properties of
1302: pulsar
1303: winds, including their interaction
1304: with the
1305: neutron star companions, and the properties of pulsars themselves.
1306:
1307: \acknowledgements
1308: We thank Ingrid Stairs for providing the updated ephemeris for
1309: J1537 and Firoza Sutaria for the help with the analysis of the
1310: {\sl XMM-Newton} data.
1311: This research was supported by NASA grants NAG5-10865 and
1312: NAS8-01128 and {\sl Chandra} award SV4-74018.
1313:
1314: \begin{thebibliography}
1315:
1316: \bibitem[]{1261}
1317: Achtenberg, A., Gallant, Y.\ A., Kirk, J.\ G., \& Guthmann, A.\ W.
1318: 2001, MNRAS, 328, 393
1319:
1320: \bibitem[]{1265}
1321: Arons, J., Backer, D.\ C., Spitkovsky, A., \& Kaspi, V.\ M. 2004,
1322: in Binary Radio Pulsars, eds.\ F.\ A.\ Rasio \& I.\ H. Stairs,
1323: ASP Conf.\ Ser., in press (astro-ph/0404159)
1324:
1325: \bibitem[]{1270}
1326: Arons, J., \& Tavani, M. 1993, ApJ, 403, 249
1327:
1328: \bibitem[]{1273}
1329: Bucciantini, N., Amato, L. \& del Zanna, L. 2005, A\&A, 423, 253
1330:
1331: \bibitem[]{1276}
1332: Burgay, M., et al.\ 2003, Nature, 426, 531
1333:
1334: \bibitem[]{1279}
1335: Campana, S., Possenti, A., \& Burgay, M. 2004, ApJ, 613, L53
1336:
1337: \bibitem[]{1280}
1338: Cash, W. 1979, ApJ , 228, 939
1339:
1340: \bibitem[]{1282}
1341: Coles, W.\ A., McLaughlin, M.\ A., Rickett, B.\ J., Lyne, A.\ G.,
1342: \& Bhat, N.\ D.\ R. 2005, ApJ, 623, 392
1343:
1344: \bibitem[]{1286}
1345: Demorest, P., Ramachandran, R., Backer, D.\ C., Ransom, S.\ M.,
1346: Kaspi, V., Arons, J., \& Spitkovsky, A. 2004, ApJ, 615, L137
1347:
1348: \bibitem[]{1290}
1349: Dickey, J.\ M., \& Lockman, F.\ J. 1990, ARA\&A, 28, 215
1350:
1351: \bibitem[]{1293}
1352: Gil, J., Melikidze, G.\ I., \& Geppert, U. 2003, A\&A, 407, 315
1353:
1354: \bibitem[]{1296}
1355: Granot, J., \& M\'esz\'aros, P., 2004, ApJ, 609, L17
1356:
1357: \bibitem[]{1299}
1358: Hibschman, J.\ A., \& Arons, J. 2001, ApJ, 560, 871
1359:
1360: \bibitem[]{1302}
1361: Jenet, F.\ A., \& Ransom, S.\ M. 2004, Nature, 428, 919
1362:
1363: \bibitem[]{1305}
1364: Kargaltsev, O.\ Y., Pavlov, G.\ G., \& Garmire, G.\ P. 2006, ApJ,
1365: 636, 406
1366:
1367: \bibitem[]{1309}
1368: Kennel, C.\ F., \& Coroniti, F.\ V. 1984, ApJ, 283, 694
1369:
1370: \bibitem[]{1312}
1371: Kirk, J., \& Skj{\ae}raasen, O. 2003, ApJ, 591, 366
1372:
1373: \bibitem[]{1315}
1374: Konacki, M., Wolszczan, A., \& Stairs, I.\ H. 2003, ApJ, 589, 495
1375:
1376: \bibitem[]{1318}
1377: Lyne, A.\ G., et al. 2004, Science, 303, 1153
1378:
1379: \bibitem[]{1321}
1380: Lyutikov, M. 2004, MNRAS, 353, 1095
1381:
1382: \bibitem[]{1324}
1383: Lyutikov, M., \& Thompson, C. 2005, ApJ, 634, 1223
1384:
1385: \bibitem[]{1327}
1386: Manchester, R.\ N., et al. 2005, ApJ, 621, L49
1387:
1388: \bibitem[]{1330}
1389: McLaughlin, M.\ A., et al. 2004, ApJ, 605, L41
1390:
1391: \bibitem[]{1333}
1392: Lyubarsky, Y., \& Kirk, J.\ G. 2001, ApJ, 547, 437
1393:
1394: \bibitem[]{1336}
1395: Pellizzoni, A., De Luca, A., Mereghetti, S., Tiengo, A., Mattana, F.,
1396: Caraveo, P., Tavani, M., \& Bignami, G.\ F. 2004, ApJ, 612, L49
1397:
1398: \bibitem[]{1340}
1399: Rafikov, R.\ R., \& Goldreich, P. 2005, ApJ, 631, 488
1400:
1401: \bibitem[]{1343}
1402: Stairs, I.\ H. 2004, Science, 304, 547
1403:
1404: \bibitem[]{1346}
1405: Stairs, I.\ H., Thorsett, S.\ E., Taylor, J.\ H., \& Wolszczan, A. 2002,
1406: ApJ, 581, 501
1407:
1408: \bibitem[]{1350}
1409: Stairs, I.\ H., Thorsett, S.\ E., \& Arzoumanian, Z. 2004,
1410: Phys.\ Rev.\ Lett., 93, 141101
1411:
1412: \bibitem[]{1354}
1413: Stappers, B.\ W., Gaensler, B.\ M., Kaspi, V.\ M., van der Klis, M.,
1414: \& Lewin, W.\ H.\ G. 2003, Science, 299, 1372
1415:
1416: \bibitem[]{1358}
1417: Thorsett, S.\ E., Dewey, R.\ J., \& Stairs, I.\ H. 2005,
1418: ApJ, 619, 1036
1419:
1420: \bibitem[]{1362}
1421: Wolszczan, A. 1991, Nature, 350, 688
1422:
1423: \bibitem[]{1365}
1424: Zavlin, V.\ E. 2006, ApJ, 638, 951
1425:
1426: \bibitem[]{1368}
1427: Zavlin, V.\ E., \& Pavlov, G.\ G. 2004, ApJ, 616, 452
1428:
1429: \bibitem[]{1371}
1430: Zavlin, V.\ E., Pavlov, G.\ G., Sanwal, D., Manchester, R.\ N.,
1431: Tr\"umper, J., Halpern, J.\ P., \& Becker, W. 2002, ApJ, 569, 894
1432:
1433: \bibitem[]{1375}
1434: Zhang, B., \& Loeb, A. 2004, ApJ, 614, L53
1435:
1436: \end{thebibliography}
1437:
1438: \clearpage
1439:
1440:
1441: \begin{deluxetable}{lll}
1442: \tabletypesize{\scriptsize} \tablewidth{0pt}
1443: \setlength{\tabcolsep}{0.005in} \tablecaption{ Observed and derived
1444: parameters for J1537+1155 and J0737-3039 DNSBs}
1445:
1446: \tablehead{ \colhead{Parameter} & \colhead{J1537+1155} & \colhead{J0737-3039} } % \\
1447: \startdata
1448: Binary period, $P_{b}$ (days)......................... & 0.421 & 0.102 \\
1449: Relative semimajor axis, $a$ ($10^{11}$ cm) .... & 2.28 & 0.878 \\
1450: Eccentricity, $e$ ........................................ & 0.274 & 0.0877 \\
1451: Distance, $d$ (kpc).................................... & 1.0 & $\approx 0.5$ \\
1452: Orbital inclination, $\sin i$ ......................... & 0.975 & 1.000 \\
1453: Transverse velocity, $v_\perp$ (km/s)................ &120 &70\\
1454: Mass, $M$ ($M_\odot$)...................................... & 2.68 & 2.59 \\
1455:
1456: \enddata
1457: \tablecomments{Based on the data from Stairs et al.\ (2002, 2004),
1458: Konacki et al.\ (2003), Lyne et al.\ (2004), Manchester
1459: et al.\ (2005), and Cole et al.\ (2005).
1460: The transverse velocoty in the plane of the sky,
1461: $v_{\perp}$, is uncorrected for the solar system motion. }
1462: \end{deluxetable}
1463:
1464:
1465:
1466: \begin{deluxetable}{llll}
1467: \tabletypesize{\scriptsize} \tablewidth{0pt}
1468: \setlength{\tabcolsep}{0.005in} \tablecaption{ Observed and derived
1469: parameters for J1537+1155A, J0737-3039A and J0737-3039B pulsars.}
1470: \tablehead{\colhead{Parameter} & \colhead{J1537+1155A} & \colhead{J0737-3039A} & \colhead{J0737-3039B} } % \\
1471: \startdata
1472: Spin period, $P$~(ms)........................... &37.90 & 22.70 &2773.46 \\
1473: Period derivative, $\dot{P}$ ($10^{-17}$).............. &
1474: 0.242 & 0.174 & 88 \\
1475: Dispersion measure, DM (cm$^{-3}$~pc)\,...
1476: & 11.6 & 48.9 & 48.7\\
1477: Angle $\delta$ (deg) ..................................... & $25\pm4$ (or $155\pm4$) &$0-60$ (or $120-180$) & ... \\
1478: Angle $\alpha$ (deg) .................................... &
1479: 103 & $0-60$ & ... \\
1480: Mass ($M_\odot$) ....................................... & 1.33 & 1.34 & 1.25 \\
1481: Semimajor axis, $a$ ($10^{11}$ cm) .............. & 1.12 & 0.424 & 0.454 \\
1482: Surface magnetic field, $B_s$~($10^{10}$ G)....
1483: & 0.97 & 0.64 & 158\\
1484: Spin-down power, $\dot{E}$~($10^{33}$ erg s$^{-1}$)....
1485: & 1.8 & 5.8 & 0.0016 \\
1486: Age, $\tau=P/(2\dot{P})$, (Myr).....................
1487: & 248 & 210 & 50 \\
1488: Light cylinder radius, $R_{lc}$ ($10^{8}$ cm) ... & 1.8& 1.1 & 130 \\
1489: Magnetic field at $R_{lc}$, $B_{lc}$ (G) ......... & 1660 & 5050 & 0.69 \\
1490: Polar cap radius, $R_{\rm pc}$ (m) ..................& 740 &960 &87 \\
1491: Primary pair flux, $\dot{N}_{\rm GJ}$ ($10^{30}$ s$^{-1}$)...... & 19 & 34 & 0.6\\
1492:
1493: \enddata
1494: \tablecomments{Based on the data from Stairs et al.\ (2002, 2004), Lyne et al.\ (2004),
1495: and Manchester et al.\ (2005).
1496: Angle $\delta$ is between the pulsar's spin and orbital
1497: rotation axis. Angle $\alpha$ is between the pulsar's spin and
1498: magnetic axis. The surface magnetic field, $B_s$, is assumed to be
1499: that of a dipole and given at the equator. }
1500: \end{deluxetable}
1501:
1502:
1503: \begin{table}[]
1504: \caption[]{Parameters of PL and BB models fitted to the {\sl
1505: Chandra} and {XMM-Newton} spectra of J1537 and J0737}
1506: \vspace{-0.5cm}
1507: \begin{center}
1508: \begin{tabular}{cccccccc}
1509: \tableline\tableline Instrument & Model &
1510: $N_{H,20}$\tablenotemark{a}&Norm.\tablenotemark{b} &
1511: $\Gamma$ or $kT$\tablenotemark{c} & C/dof & Q\tablenotemark{d} & $L_{\rm X}$ or $L_{\rm bol}\langle\cos\varsigma\rangle$\tablenotemark{e} \\
1512: \tableline
1513: & ~~~~ & ~~~ & ~~~~~~~~~~~{\bf J1537+1155} & & \\
1514: ACIS & PL & $3.6$&$5.7^{+1.6}_{-1.5}\times 10^{-7}$ & $3.17^{+0.52}_{-0.52}$ & 1.1/2 & 12\% & $6.1^{+3.0}_{-2.0}\times 10^{29}$ \\
1515: ACIS & BB & $3.6$&$4.8^{+5.5}_{-3.0}\times 10^{7}$ & $193_{-30}^{+37}$ & $5.3/2$ & 76\% & $6.8^{+2.1}_{-2.1} \times 10^{28}$ \\
1516: \tableline
1517: & ~~~~ & ~~~ & ~~~~~~~~~~~~{\bf J0737--3039}
1518: & & \\
1519: ACIS & PL & $5.0$&$1.01^{+0.14}_{-0.16}\times10^{-5}$ & $2.93^{+0.29}_{-0.31}$ & 1.7/4 & 5\% & $2.27_{-0.39}^{+0.42}\times 10^{30}$ \\
1520: ACIS & BB & $5.0$&$1.91_{-0.71}^{+0.99}\times 10^8$ & $192^{+17}_{-15}$ & 5.7/4 & 52\% & $2.67^{+0.33}_{-0.37}\times 10^{29} $ \\
1521: MOS1+2 & PL & $5.0$&$0.98^{+0.08}_{-0.10}\times10^{-5}$ & $3.35^{+0.19}_{-0.19}$ & 10.0/14 & 12\% & $3.03_{-0.49}^{+0.50}\times 10^{30}$ \\
1522: MOS1+2 & BB & $5.0$&$3.95_{-1.0}^{+1.5}\times 10^8$ & $169^{+11}_{-13}$ & 16.9/14 & 62\% & $3.28^{+0.40}_{-0.38}\times 10^{29} $ \\
1523: \tableline
1524: \end{tabular}
1525: \end{center}
1526: \tablecomments{ The errors are at the 68\% confidence level for one
1527: interesting parameter. } \tablenotetext{a}{Fixed values of hydrogen
1528: column density, in units of $10^{20}$ cm$^{-2}$.}
1529: \tablenotetext{b}{Normalization: Spectral flux in photons cm$^{-2}$
1530: s$^{-1}$ keV$^{-1}$ at 1 keV or projected area of the emitting
1531: region in cm$^2$, for the PL and BB models, respectively.}
1532: \tablenotetext{c}{Photon index or BB temperature in eV.}
1533: \tablenotetext{d}{Percentage of 10,000 Monte Carlo simulations,
1534: drawn from the best-fit model, which give a C-statistic value lower
1535: than the best-fit value. Although the BB fits are formally worse
1536: than the PL fits, they are statistically acceptable. }
1537: \tablenotetext{e}{ $L_{\rm X}$ is the unabsorbed PL luminosity in
1538: the 0.2--10 keV band, $L_{\rm bol}$ the bolometric luminosity for
1539: the BB fits. }
1540: \end{table}
1541:
1542:
1543:
1544:
1545:
1546:
1547:
1548:
1549: \end{document}
1550: