1: % SAMPLE1.TEX -- AASTeX sample paper with minimal markup.
2: %
3: % V0.1 - SM 30/11/2006
4: % V0.2 - SM 3/12
5: % V0.22 - AT + SM 4/12
6: % V0.4 - SM+AT - submitted 6/12 ~23 UT
7: % V1.0 - edited ms.tex downloaded from ApJ site 23/12
8: % V1.1 - SM 28/12 resubmitted
9: % V2.0 - edited ms.tex downloaded from ApJ site 16/1/2007
10: % V3.0 - edited ms.tex downloaded 25/1 from ApJ site, changed and resubmitted - 25/1
11: %\documentstyle[12pt,aasms4]{article}
12:
13: \documentclass[12pt,preprint]{aastex}
14:
15: \begin{document}
16:
17: %%%%%%%%%%%%%% MY DEFINITIONS
18: \def\pdot {\dot P}
19: \def\Omdot {\dot \Omega}
20: \def\ltsima{$\; \buildrel < \over \sim \;$}
21: \def\lsim{\lower.5ex\hbox{\ltsima}}
22: \def\gtsima{$\; \buildrel > \over \sim \;$}
23: \def\gsim{\lower.5ex\hbox{\gtsima}}
24: \def\msole{~M_{\odot}}
25: \def\mdot {\dot M}
26: \def\rxj {RX~J1856.5--3754}
27: \def\cha {\textit{Chandra~}}
28: \def\xmm {\textit{XMM-Newton~}}
29: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
30: \slugcomment{Accepted for publication on The Astroph. Journal
31: Letters}
32:
33: \title{XMM-Newton discovery of 7 s pulsations in the isolated neutron star \rxj\ }
34: \author{Andrea Tiengo \& Sandro Mereghetti}
35: \affil{INAF - Istituto di Astrofisica Spaziale e Fisica Cosmica - Milano, \\
36: v.Bassini 15, I-20133 Milano, Italy \\
37: tiengo@iasf-milano.inaf.it, sandro@iasf-milano.inaf.it}
38: %\authoremail{sandro@ifctr.mi.cnr.it}
39:
40: \begin{abstract}
41:
42: Thanks to the high counting statistics provided by a recent
43: XMM-Newton observation of \rxj\, we have discovered that this
44: isolated neutron star pulsates at a period of 7.055 s. This
45: confirms that \rxj\ is similar in nature to the other six
46: thermally emitting, nearby neutron stars discovered in soft
47: X--rays with ROSAT. The pulsations are detected at consistent
48: periods in several XMM-Newton observations spanning from April
49: 2002 to October 2006, yielding an upper limit of
50: $\pdot<1.9\times10^{-12}$ s s$^{-1}$ (90\% c.l.) on the period
51: derivative. This implies a surface magnetic field smaller than
52: 1.2$\times10^{14}$ G, under the usual assumption of vacuum dipole
53: magnetic braking. The pulse profile is nearly sinusoidal with a
54: pulsed fraction in the 0.15-1.2 keV range of only
55: $\sim$1.2\%, the smallest ever seen in an isolated X-ray pulsar.
56:
57: \end{abstract}
58:
59: \keywords{stars: individual (RX J1856.5-3754) -- stars: neutron --
60: X--rays: pulsars}
61:
62: \section{Introduction}
63:
64:
65: The soft X--ray source \rxj\ belongs to a small group of seven
66: nearby, isolated neutron stars discovered with the ROSAT
67: satellite and often dubbed X--ray Dim Isolated Neutron Stars
68: (XDINS, see \citet{haberl2006} for a recent review). These
69: sources are characterized by thermal spectra with blackbody
70: temperature in the range 40--110 eV, X--ray luminosity of
71: 10$^{30}$--10$^{32}$ erg s$^{-1}$, faint optical counterparts
72: (V$>$25), and absence of radio emission. The initial suggestion
73: that XDINS could be powered by accretion from the interstellar
74: medium \citep{walter1996} has been ruled out by the measurement,
75: in three of them, of large proper motions that imply an accretion
76: rate far too small to power the observed luminosity (see, e.g.,
77: \citet{motch2005} and references therein). Pulsations have been
78: found in five (or possibly six) XDINS, with periods in the 3--12
79: s range, pulsed fractions between 4\% and 18\%, and, in
80: two cases, period derivatives of the order of 10$^{-13}$ s
81: s$^{-1}$ \citep{kaplan2005a,kaplan2005b}. These spin-down values
82: yield characteristic ages of $\sim$1--2 Myrs, spin-down
83: luminosities of 4--5$\times$10$^{30}$ erg s$^{-1}$, and, with the
84: canonical dipole radiation braking assumptions, magnetic fields of
85: a few 10$^{13}$ G. It is thus believed that the X--ray emission in
86: XDINS is powered by the neutron star cooling.
87:
88: \rxj\ received particular interest since it is the brightest XDIN,
89: its X--ray flux does not show any apparent contamination from non
90: thermal magnetospheric emission, and a parallactic distance
91: (117$\pm$12 pc) has been reported \citep{walter2002}. These
92: properties make \rxj\ an ideal target to derive information on the
93: neutron star radius and thus constrain the equation of state of
94: matter at super nuclear density through a detailed modelling of
95: its surface thermal emission. Quite surprisingly, even the X--ray
96: spectra with the highest resolution and the best statistics,
97: obtained in long Chandra and XMM-Newton observations, do not show
98: any line feature and can be fit by a simple Planckian with
99: temperature kT=63 eV \citep{burwitz2003}.
100:
101: Despite extensive searches, no pulsations have ever been detected
102: in \rxj\ \citep{pons2002,ransom2002,drake2002,burwitz2003}. The
103: tight upper limits on the pulsed fraction ($<$1.3\% at 2$\sigma$)
104: obtained for periods between 0.02 and 1000 s \citep{burwitz2003},
105: led to the speculation that \rxj\ could be a millisecond pulsar,
106: not yet seen to pulsate owing to the reduced sensitivity for short
107: periods of the observations carried out up to now.
108: Recently, an upper limit on the pulsed fraction of 2.1\%
109: (at 1$\sigma$) was reported for periods in the range 1-20 ms
110: \citep{zavlin2006}. Here we report the discovery of pulsations at
111: 7 seconds, obtained thanks to the very high counting statistics
112: provided by the most recent XMM-Newton observation of \rxj .
113:
114:
115: \section{Timing analysis and results}
116:
117: A 70 ks long observation of \rxj\ was carried out with the
118: XMM-Newton satellite starting on 2006 October 24 at 00:30 UT. We
119: used the data of the EPIC instrument (0.1-12 keV), consisting of
120: two MOS and one pn cameras \citep{turner01,struder01}. All the
121: cameras mounted the thin filter and were operated in Small Window
122: mode, yielding time resolutions of 6 ms and 0.3 s for the pn and
123: MOS, respectively. We processed the data using version 7.0 of the
124: \emph{XMM-Newton Science Analysis System}; the event files were
125: created using the tasks \emph{epproc} and \emph{emproc} with
126: default options and filtered to exclude the time intervals of high
127: particle background, resulting in net (dead-time corrected)
128: exposure times of 47 ks and 68 ks for the pn and MOS,
129: respectively.
130:
131: We started the timing analysis using the data of the pn camera and
132: selecting the energy range 0.15-1.2 keV. Due to the very soft
133: spectrum of \rxj , practically no source photons are detected at
134: higher energies. We used a circular extraction region centered at
135: the source position and with radius 40$''$. This resulted in
136: $\sim$356,600 counts, of which $\sim$2,700 can be ascribed to the
137: background. The arrival times were corrected to the Solar System
138: barycenter. In order to search for periods of a few seconds, as
139: seen in the other sources of this class, we rebinned the counts at
140: $\sim$0.278 s and computed the Fourier power spectrum. A
141: significant peak in the power spectrum was found at the frequency
142: of 0.14174 Hz. Taking into account the number of searched periods
143: (131072), the peak value of 38.46 corresponds to a chance
144: probability of 6$\times10^{-4}$. To better estimate the period, we
145: used the epoch folding technique and fitted the peak in the
146: $\chi^2$ versus trial period distribution as described in
147: \citet{leahy87}, obtaining P=7.05514$\pm$0.00007 s. The
148: corresponding folded light curve is shown in the top panel of
149: Fig.~1. No significant differences in the pulse profiles were
150: found by dividing the counts in soft and hard energy intervals.
151: The phase averaged spectrum is well fit with absorbed blackbody
152: parameters fully consistent with those of the previous XMM-Newton
153: observations of \rxj\ \citep{haberl2006}. The pulsed fraction,
154: derived by fitting a sinusoid to the background subtracted pn
155: light curve, is (1.6$\pm$0.2)\%. This small pulsed fraction
156: explains why the periodicity was not discovered in previous
157: observations and is not significantly detected in the MOS data.
158: The two MOS together collected only $\sim$137,900 counts from a
159: 40$''$ radius extraction region, less than half of the pn ones. It
160: is thus not surprising that the MOS 0.15-1.2 keV folded light
161: curve (bottom panel of Fig.~1) is statistically consistent with a
162: constant flux.
163:
164:
165:
166: In order to confirm the periodicity seen in the pn data of
167: October 2006, we analyzed the other XMM-Newton observations listed
168: in Table 1. After reducing the data as described above, we
169: searched for the 7 s pulsations with the Z$^2$ test
170: \citep{buccheri83}, which has a good sensitivity for sinusoidal
171: signals. For each observation, we restricted our search to the
172: range of periods obtained by a backward extrapolation from the
173: current value, assuming a conservative period derivative
174: $|\pdot|<5\times10^{-10}$ s s$^{-1}$. During the April 2004
175: observation the pn was in Timing mode and the two MOS were in Full
176: Frame mode with a time resolution of only 2.6 s, that caused
177: significant photon pile-up in the MOS data. Therefore we used only
178: the pn data, selecting the energy range 0.28-1.2 keV in order to
179: avoid the soft flares produced by high-energy particles in the pn
180: Timing mode \citep{burwitz2004}. In all the other observations we
181: used the sum of the 0.15-1.2 keV counts from the pn and the two
182: MOS\footnote{For the 2002 observation the counts below 0.18 keV in
183: the MOS1 were excluded in order to eliminate the electronic noise
184: present in the MOS timing mode}.
185:
186: The distributions that we obtained for the Z$^2$ statistics
187: are shown in Fig.~2, where it can be seen that a maximum
188: of Z$^2$ is always present at the expected pulse period. To
189: evaluate the statistical significance of these detections, we must
190: take into account the number of periods searched in each
191: observation. The most significant detection is that of April 2002,
192: where, considering the 335 independent periods in the considered
193: range (6.98--7.13 s), the peak Z$^2$=32.22 corresponds to a chance
194: probability of 335$\times e^{-32.22/2}$= 3$\times10^{-5}$. In a
195: similar way we derived the following chance probabilities for the
196: other observations: 7$\times10^{-5}$ (April 2004),
197: 5$\times10^{-4}$ (September 2004), 2$\times10^{-3}$ (September
198: 2005), and 2$\times10^{-2}$ (March 2006).
199:
200: These values indicate that the periodicity is detected in
201: all the observations, although sometimes with small Z$^2$ values,
202: as expected for the case of a small pulsed fraction. The
203: relatively low significance of the March 2006 detection, despite a
204: number of counts and duration similar to the October 2006
205: observation, is not particularly surprising. In fact, with about
206: half a million counts, a $\sim$1\% modulation has a probability of
207: ~10\% of giving a Z$^2$ as small as the observed value of 15.5
208: \citep{brazier1994}.
209:
210:
211: The best periods and pulsed fractions, determined in each
212: observation as described above for the October 2006 data, are
213: given in Table~1, while the folded light curves are presented in
214: Fig.~3. Although this figure seems to suggest some variations in
215: the pulse profiles, we verified by means of a Kolmogorov-Smirnov
216: test that, taking into account the unknown relative phase
217: alignment, all the light curves are fully compatible with the same
218: profile. We determined the relative phase shifts of the
219: six observations by fitting sinusoidal functions to the folded
220: light curves. After applying these shifts, we summed all the pn
221: and MOS data taken in Small Window mode to obtain the averaged
222: light curves shown in Fig.~4 for the total (0.15-1.2 keV), soft
223: (0.15-0.26 keV) and hard (0.26-1.2 keV) energy ranges. The
224: corresponding pulsed fractions are (1.17$\pm$0.08)\%,
225: (0.88$\pm$0.11)\%, and (1.5$\pm$0.11)\%.
226:
227:
228: All the period measurements, spanning more than four years, are
229: consistent with an average value of P=7.05515$\pm$0.00004 s. With
230: a linear fit to the period values we can limit the period
231: derivative to $-1.2\times10^{-12} < \pdot <1.9\times10^{-12}$ s
232: s$^{-1}$ (90\% c.l.).
233:
234: We searched for the \rxj\ pulsations also in the long
235: Chandra observation performed with the LETG/HRC-S instrument in
236: October 2001. Our search, based on about 91,600 counts extracted
237: from the zero-order image, gave a negative result. This is not
238: surprising, since for this number of counts and a $\sim$1.2\%
239: pulsed fraction, the expectation value for the Z$^2$ statistics is
240: only 6.6. Furthermore, the long time span of the observation
241: yields an intrinsic frequency resolution a factor ~10 better than
242: that of the best XMM-Newton observation. This means a tenfold
243: increase in the number of statistically independent periods that
244: have to be examined and thus a higher threshold for a significant
245: detection.
246:
247:
248: \section{Conclusions}
249:
250: Our discovery of the long sought periodicity in \rxj\ and the
251: upper limit derived for $\pdot$ can be used to infer a
252: characteristic age $\tau>6\times10^4$ yr, a magnetic field
253: B$<1.2\times10^{14}$ G, and a spin-down luminosity
254: $<2\times10^{32}$ erg s$^{-1}$. These values are fully consistent
255: with those of the other six XDINS.
256:
257: The pulsed fraction of \rxj\ is significantly smaller than that of
258: the other members of the group, which have pulsed fractions in the
259: range from 4\% to 18\%. The XMM-Newton data give also
260: evidence for a pulsed fraction increasing as a function of the
261: energy. The small pulsed fraction in \rxj\ might simply result
262: from geometrical orientation effects or by a particularly small
263: gradient in the surface temperature distribution. Detailed
264: phase-resolved spectral modelling, that might discriminate between
265: the different possibilities, will be reported elsewhere. Here we
266: simply note that the October 2006 pn spectra for the maximum and
267: minimum of the folded light curve are both consistent with the
268: best-fit absorbed blackbody model of the phase-averaged spectrum,
269: simply re-scaled in normalization. Letting all the fit parameters
270: free to vary, the maximum allowed difference in the temperatures
271: of the two spectra is of kT$_{\rm BB}$=1.2 eV (at the 90\%
272: c.l.). Repeating the same exercise for 5 phase bins, we find again
273: no significant spectral variations, with all the parameters within
274: the following ranges: N$_{\rm H}=$(4--8)$\times$10$^{19}$
275: cm$^{-2}$ and kT$_{\rm BB}$=61.4--63.2 eV.
276:
277: Our results indicate that, except for the striking absence of
278: deviations from a pure blackbody in its X--ray spectrum, \rxj\
279: shares most of the properties of the other members of the XDINS
280: group. More importantly, the discovery of pulsations opens the
281: way for the determination of the timing-dependent parameters that
282: are crucial for a more detailed modelling of the star's surface
283: emission properties.
284:
285:
286:
287:
288:
289: \acknowledgments
290:
291:
292: Based on observations obtained with XMM-Newton, an ESA science
293: mission with instruments and contributions directly funded by ESA
294: Member States and NASA. This work has been supported by the
295: Italian Space Agency under contract ASI/INAF/I/023/05/0.
296:
297:
298:
299: \acknowledgments
300:
301: \begin{thebibliography}{16}
302: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
303:
304: \bibitem[{{Buccheri} {et~al.}(1983){Buccheri}, {Bennett}, {Bignami}, {Bloemen},
305: {Boriakoff}, {Caraveo}, {Hermsen}, {Kanbach}, {Manchester}, {Masnou},
306: {Mayer-Hasselwander}, {Ozel}, {Paul}, {Sacco}, {Scarsi}, \&
307: {Strong}}]{buccheri83}
308: {Buccheri}, R., {Bennett}, K., {Bignami}, G.~F., {et~al.} 1983, \aap, 128, 245
309:
310:
311: \bibitem[{{Brazier} (1994) }]{brazier1994}
312: {Brazier}, K.T.S. 1994, \mnras, 268, 709
313:
314: \bibitem[{{Burwitz} {et~al.}(2004){Burwitz}, {Haberl}, {Freyberg}, {Dennerl},
315: {Kendziorra}, \& {Kirsch}}]{burwitz2004}
316: {Burwitz}, V., {Haberl}, F., {Freyberg}, M.~J., {et~al.} 2004, in X-Ray and
317: Gamma-Ray Instrumentation for Astronomy XIII. Edited by Flanagan, Kathryn A.;
318: Siegmund, Oswald H. W. Proceedings of the SPIE, Volume 5165, pp. 123-130
319: (2004)., ed. K.~A. {Flanagan} \& O.~H.~W. {Siegmund}, 123--130
320:
321: \bibitem[{{Burwitz} {et~al.}(2003){Burwitz}, {Haberl}, {Neuh{\"a}user},
322: {Predehl}, {Tr{\"u}mper}, \& {Zavlin}}]{burwitz2003}
323: {Burwitz}, V., {Haberl}, F., {Neuh{\"a}user}, R., {et~al.} 2003, \aap, 399,
324: 1109
325:
326: \bibitem[{{Drake} {et~al.}(2002){Drake}, {Marshall}, {Dreizler}, {Freeman},
327: {Fruscione}, {Juda}, {Kashyap}, {Nicastro}, {Pease}, {Wargelin}, \&
328: {Werner}}]{drake2002}
329: {Drake}, J.~J., {Marshall}, H.~L., {Dreizler}, S., {et~al.} 2002, \apj, 572,
330: 996
331:
332: \bibitem[{{Haberl}(2006)}]{haberl2006}
333: {Haberl}, F. 2006, ArXiv Astrophysics astro-ph/0609066
334:
335: \bibitem[{{Kaplan} \& {van Kerkwijk}(2005{\natexlab{a}})}]{kaplan2005a}
336: {Kaplan}, D.~L. \& {van Kerkwijk}, M.~H. 2005{\natexlab{a}}, \apjl, 628, L45
337:
338: \bibitem[{{Kaplan} \& {van Kerkwijk}(2005{\natexlab{b}})}]{kaplan2005b}
339: {Kaplan}, D.~L. \& {van Kerkwijk}, M.~H. 2005{\natexlab{b}}, \apjl, 635, L65
340:
341: \bibitem[{{Leahy}(1987)}]{leahy87}
342: {Leahy}, D.~A. 1987, \aap, 180, 275
343:
344: \bibitem[{{Motch} {et~al.}(2005){Motch}, {Sekiguchi}, {Haberl}, {Zavlin},
345: {Schwope}, \& {Pakull}}]{motch2005}
346: {Motch}, C., {Sekiguchi}, K., {Haberl}, F., {et~al.} 2005, \aap, 429, 257
347:
348: \bibitem[{{Pons} {et~al.}(2002){Pons}, {Walter}, {Lattimer}, {Prakash},
349: {Neuh{\"a}user}, \& {An}}]{pons2002}
350: {Pons}, J.~A., {Walter}, F.~M., {Lattimer}, J.~M., {et~al.} 2002, \apj, 564,
351: 981
352:
353: \bibitem[{{Ransom} {et~al.}(2002){Ransom}, {Gaensler}, \& {Slane}}]{ransom2002}
354: {Ransom}, S.~M., {Gaensler}, B.~M., \& {Slane}, P.~O. 2002, \apjl, 570, L75
355:
356: \bibitem[{{Str{\"u}der} {et~al.}(2001){Str{\"u}der}, {Briel}, {Dennerl},
357: {Hartmann}, {Kendziorra}, {Meidinger}, {Pfeffermann}, {Reppin}, {Aschenbach},
358: {Bornemann}, {Br{\"a}uninger}, {Burkert}, {Elender}, {Freyberg}, {Haberl},
359: {Hartner}, {Heuschmann}, {Hippmann}, {Kastelic}, {Kemmer}, {Kettenring},
360: {Kink}, {Krause}, {M{\"u}ller}, {Oppitz}, {Pietsch}, {Popp}, {Predehl},
361: {Read}, {Stephan}, {St{\"o}tter}, {Tr{\"u}mper}, {Holl}, {Kemmer}, {Soltau},
362: {St{\"o}tter}, {Weber}, {Weichert}, {von Zanthier}, {Carathanassis}, {Lutz},
363: {Richter}, {Solc}, {B{\"o}ttcher}, {Kuster}, {Staubert}, {Abbey}, {Holland},
364: {Turner}, {Balasini}, {Bignami}, {La Palombara}, {Villa}, {Buttler},
365: {Gianini}, {Lain{\'e}}, {Lumb}, \& {Dhez}}]{struder01}
366: {Str{\"u}der}, L., {Briel}, U., {Dennerl}, K., {et~al.} 2001, \aap, 365, L18
367:
368: \bibitem[{{Turner} {et~al.}(2001){Turner}, {Abbey}, {Arnaud}, {Balasini},
369: {Barbera}, {Belsole}, {Bennie}, {Bernard}, {Bignami}, {Boer}, {Briel},
370: {Butler}, {Cara}, {Chabaud}, {Cole}, {Collura}, {Conte}, {Cros}, {Denby},
371: {Dhez}, {Di Coco}, {Dowson}, {Ferrando}, {Ghizzardi}, {Gianotti}, {Goodall},
372: {Gretton}, {Griffiths}, {Hainaut}, {Hochedez}, {Holland}, {Jourdain},
373: {Kendziorra}, {Lagostina}, {Laine}, {La Palombara}, {Lortholary}, {Lumb},
374: {Marty}, {Molendi}, {Pigot}, {Poindron}, {Pounds}, {Reeves}, {Reppin},
375: {Rothenflug}, {Salvetat}, {Sauvageot}, {Schmitt}, {Sembay}, {Short},
376: {Spragg}, {Stephen}, {Str{\"u}der}, {Tiengo}, {Trifoglio}, {Tr{\"u}mper},
377: {Vercellone}, {Vigroux}, {Villa}, {Ward}, {Whitehead}, \& {Zonca}}]{turner01}
378: {Turner}, M.~J.~L., {Abbey}, A., {Arnaud}, M., {et~al.} 2001, \aap, 365, L27
379:
380: \bibitem[{{Walter} \& {Lattimer}(2002)}]{walter2002}
381: {Walter}, F.~M. \& {Lattimer}, J.~M. 2002, \apjl, 576, L145
382:
383: \bibitem[{{Walter} {et~al.}(1996){Walter}, {Wolk}, \& {Neuhauser}}]{walter1996}
384: {Walter}, F.~M., {Wolk}, S.~J., \& {Neuhauser}, R. 1996, \nat, 379, 233
385:
386:
387: \bibitem[{{Zavlin}(2006)}]{zavlin2006}
388: {Zavlin}, V.~E. 2006, ArXiv Astrophysics,
389: astro-ph/0608210
390:
391:
392: \end{thebibliography}
393:
394: \clearpage
395:
396: \begin{table}
397: \caption{Log of \xmm\ observations of \rxj. The exposure
398: times take into account the filtering for high background time
399: intervals and the instrumental dead-time. The time resolution of
400: the small window (SW) mode is 6 ms for the pn and 0.3 s for the
401: MOS, while for the Timing (TI) mode it is 0.03 ms for pn and 1.5
402: ms for MOS. The period values are obtained with the sum of pn and
403: MOS counts. The pulsed fractions are computed from sinusoidal fits
404: of the background subtracted pulse profiles (PF=A/C, where the
405: count rate is given by C+Asin($\phi-\phi_0$)). Errors are at
406: 1$\sigma$.} \smallskip
407: \begin{tabular}[c]{ccccrcc}
408: \hline \hline
409: Date & Instr. & Mode & Exp. & Counts & Period & PF \\
410: & & & (ks) & & (s) & (\%)\\
411: \hline \hline
412: 2002-04-08&pn&SW&40&304,052 &7.05510$\pm$0.00013&1.4$\pm$0.2\\
413: &MOS1&TI&57&53,529&&1.1$\pm$0.5\\
414: &MOS2&SW&56&82,837& &0.8$\pm$0.4\\
415: \hline
416: 2004-04-17&pn&TI&31&109,232& 7.05502$\pm$0.00012&2.4$\pm$0.3\\
417: \hline
418: 2004-09-24&pn&SW&23&178,718&7.05524$\pm$0.00010&1.1$\pm$0.2\\
419: &MOS1&SW&68&46,284&&2.0$\pm$0.5\\
420: &MOS2&SW&68&52,865& &2.0$\pm$0.5\\
421: \hline
422: 2005-09-24&pn&SW&23&169,803&7.05518$\pm$0.00027&1.6$\pm$0.2\\
423: &MOS1&SW&27&26,936&&0.8$\pm$0.6\\
424: &MOS2&SW&27&29,706& &1.5$\pm$0.6\\
425: \hline
426: 2006-03-26&pn&SW&49&365,050&7.05526$\pm$0.00016&0.8$\pm$0.2\\
427: &MOS1&SW&67&64,895&&1.4$\pm$0.4\\
428: &MOS2&SW&67&74,821& &0.7$\pm$0.4\\
429: \hline
430: 2006-10-24&pn&SW&47&356,605&7.05514$\pm$0.00007&1.6$\pm$0.2\\
431: &MOS1&SW&68&65,941&&1.1$\pm$0.4\\
432: &MOS2&SW&68&71,973& &1.1$\pm$0.4\\
433: \hline
434:
435: \end{tabular}
436: \label{tab:axp_obs}
437: \end{table}
438:
439: \clearpage
440:
441: \begin{figure}
442: \epsscale{.7}
443: \plotone{f1.ps}
444: \caption{Folded light curves of
445: \rxj\ in the 0.15-1.2 keV energy
446: range obtained in the October 2006
447: XMM-Newton observation. Top: EPIC pn, Bottom: EPIC MOS1+MOS2.
448: %The Z$^2$ statistics for
449: %the sum of the two MOS is 8.1, corresponding to a chance probability of
450: %1.7\%.
451: Fitting the light curves with a constant value gives
452: $\chi^2$ values of 56.8 (pn) and 8.6 (MOS1+MOS2) for 9 d.o.f.,
453: corresponding to chance probabilities of 6$\times10^{-9}$ and
454: 0.47, respectively.
455: \label{f1}}
456: \end{figure}
457:
458: \clearpage
459:
460: \begin{figure}
461: \epsscale{1.0}
462: \plotone{f2.ps}
463: \caption{Distributions of the
464: Z$^2$ statistics for the XMM-Newton observations
465: of \rxj . The period ranges shown in the first five panels correspond to those used in
466: the period search and derived from a conservative $\pdot$ assumption (see text for details).
467: \label{f2}}
468: \end{figure}
469:
470: \clearpage
471:
472: \begin{figure}
473: \plotone{f3.ps}
474: \caption{Background subtracted light curves of
475: \rxj\ folded at the best-fit periods (see Table~1). The light
476: curves are normalized to the average count rate. The phase
477: alignment between the different panels is arbitrary. All the light
478: curves are from the sum of pn and MOS counts in the 0.15-1.2 keV
479: range, except for the 2004 April observation where only the
480: 0.28-1.2 keV pn counts have been used. \label{f3}}
481: \end{figure}
482:
483: \clearpage
484:
485: \begin{figure}
486: \plotone{f4.ps}
487: \caption{Light curves of \rxj\ in the total
488: (0.15-1.2 keV), soft (0.15-0.26 keV), and hard (0.26-1.2 keV)
489: energy ranges obtained by summing all the pn and MOS data in small
490: window mode. \label{f4}}
491: \end{figure}
492:
493: \end{document}
494: