1: %% Modified 2004 January 9
2: %%
3: %% This is a sample manuscript marked up using the
4: %% AASTeX v5.x LaTeX 2e macros.
5:
6: %% The first piece of markup in an AASTeX v5.x document
7: %% is the \documentclass command. LaTeX will ignore
8: %% any data that comes before this command.
9:
10: \documentclass{emulateapj}
11: %\documentclass[12pt,preprint]{aastex}
12: \usepackage{graphicx,times}
13: %
14: \slugcomment{Received 2007 August 6; accepted 2007 September 23}
15:
16: \shorttitle{Millisecond pulsation in Aql X-1}
17: \shortauthors{P.Casella et al.}
18: \begin{document}
19: %\def\new#1{{\bf #1}}
20: \def\new#1{{#1}}
21: %
22: \title{Discovery of coherent millisecond X-ray pulsations in Aql X-1}
23:
24: \author{P. Casella\altaffilmark{1}, D. Altamirano\altaffilmark{1},
25: A. Patruno\altaffilmark{1}, R. Wijnands\altaffilmark{1}, and M. van der Klis\altaffilmark{1}}
26: \altaffiltext{1}{Astronomical Institute, ``Anton Pannekoek'', University of Amsterdam, Kruislaan 403, 1098 SJ Amsterdam, The Netherlands. E-mail: casella@science.uva.nl}
27: \thispagestyle{empty}
28:
29: \begin{abstract}
30:
31: We report the discovery of an episode of coherent millisecond X-ray
32: pulsation in the neutron star low-mass X-ray binary Aql X-1. The
33: episode lasts for slightly more than 150 seconds, during which the
34: pulse frequency is consistent with being constant. No X-ray burst or
35: other evidence of thermonuclear burning activity is seen in
36: correspondence with the pulsation, which can thus be identified as
37: occurring in the persistent emission. The pulsation frequency is
38: 550.27 Hz, very close (0.5 Hz higher) to the maximum reported
39: frequency from burst oscillations in this source. Hence we identify
40: this frequency with the neutron star spin frequency. The pulsed
41: fraction is strongly energy dependent, ranging from $<$1\% at 3-5 keV
42: to $>$10\% at 16-30 keV. We discuss possible physical interpretations
43: and their consequences for our understanding of the lack of pulsation
44: in most neutron star low-mass X-ray binaries. If interpreted as
45: accretion-powered pulsation, Aql X-1 might play a key role in
46: understanding the differences between pulsating and non-pulsating
47: sources.
48:
49: \end{abstract}
50: \keywords{pulsars: individual
51: (Aql X-1) --- stars: neutron --- X-rays: binaries }
52:
53: % additional keywords: binaries: general ---
54:
55: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
56:
57: \section{Introduction}
58:
59: Accretion-powered millisecond X-ray pulsars (hereinafter AMSPs) had
60: been predicted in the early 1980s as the progenitors of millisecond
61: radio pulsars \citep{backeretal82,alparetal82}. The first
62: observational indication that neutron stars in low-mass X-ray binaries
63: (LMXBs) rotate rapidly came in 1996 with the discovery of slightly
64: drifting in frequency millisecond oscillations during thermonuclear
65: X-ray bursts \citep[for a review see][]{strobild06}. However it was
66: not until 1998 that the first AMSP was discovered
67: \citep{wvdk98}. Since then, a total of eight AMSPs have been found out
68: of the $>150$ LMXBs known up to date \citep{liuetal07}.
69:
70: Since the theoretical prediction of the existence of AMSPs was made,
71: the main issue remained to explain the lack of pulsation in the
72: persistent X-ray emission of the majority of LMXBs. In recent decades
73: many theoretical efforts have been made to explain this, the main
74: question remaining whether the pulsation is hidden from the observer
75: or not produced at all. At present, the scenarios most often
76: considered are: (a) the magnetic field in non-pulsating LMXBs is too
77: weak to allow channeling of the matter onto the magnetic poles;
78: (b) the magnetic field has comparable strength inside most LMXB
79: neutron stars, but in the large majority of them it has been
80: ``buried'' by accretion, resulting in a very low surface magnetic
81: field \citep[e.g.,][]{cummingetal01} which again is too weak to allow
82: channeling of matter. After the accretion stops, the magnetic field
83: eventually emerges from the neutron star surface and assumes its
84: intrinsic value; (c) pulsations {\it are} produced in all LMXBs, but
85: in most of them they are attenuated by a surrounding scattering medium
86: that washes out the coherent beamed pulsation
87: \citep{brainerdlamb87,kylafis87,titarchuketal02}; (d) the pulsations
88: are attenuated by gravitational lensing from the neutron star
89: \citep[e.g.,][]{meszarosetal88}.
90:
91: From an observational point of view, since the discovery of the first
92: AMSP efforts have been focused on finding differences between the
93: sources showing pulsation and those that do not, in order to test
94: different theoretical hypotheses. Possible observed differences so far
95: are the orbital period \citep[which is on average shorter in AMSPs
96: than in other LMXBs, see e.g.,][]{kaaretetal06} and the time-averaged
97: accretion rate \citep[which is considered to be on average smaller in
98: AMSPs than in other LMXBs, see e.g.,][]{galloway06}. However,
99: although it is probable that orbital period and time-averaged
100: accretion rate play an important role in the determination of AMSP
101: properties, the reason for the lack of pulsation in most of LMXBs
102: still has to be found.
103:
104: The properties of the seventh discovered AMSP \citep[HETE
105: J1900.1-2455,][]{kaaretetal06,gallowayetal07} gave new insights to
106: this issue. This source has a much higher inferred time-averaged
107: accretion rate than in the other AMSPs. The pulsation became
108: undetectable after two months from the beginning of the outburst, at
109: strong variance with the other AMSPs in which pulsations were always
110: detectable until the end of the outbursts. This has been interpreted
111: as evidence for burying of the magnetic field by the accreted material
112: during the outburst \citep{gallowayetal07}.
113:
114: Transient highly coherent pulsations were also observed in 4U 1636-53
115: \citep{stromark02} during an $\sim$800 s interval. However, the
116: pulsations were detected at the flux maximum of a superburst, hence
117: they were likely nuclear-powered. \new{This is different compared to
118: the seven AMSP, where pulsations were interpreted in terms of an
119: hotspot resulting from magnetic channeling of matter onto the neutron
120: star surface.}
121:
122: We are searching the full Rossi X-ray Timing Explorer (RXTE) archive
123: data for coherent pulsations \citep[see also][]{altamiranoetal07}.
124: In this letter we report the
125: discovery of an episode of coherent millisecond X-ray pulsarions in
126: the neutron star binary Aql X-1. The LMXB recurrent X-ray transient
127: Aql X-1 is an atoll source \citep{reigetal00} wich shows kHz
128: quasi-periodic oscillations, X-ray bursts and burst oscillations
129: \citep{zhangetal98b}. An orbital period of 18.95 hr was obtained by
130: optical measurments \citep{chevalier91,welshetal00}. The presumed
131: onset of the propeller effect \citep{illarionov75,stellaetal86}
132: allowed \citet{campanaetal98} to estimate a value for the magnetic
133: field of $1-3\times10^8 ({D\over2.5~kpc})$ G, consistent with the
134: magnetic field expected (and later on measured) in AMSPs \cite[see
135: also][]{zhangetal98a}.
136:
137: We analyzed RXTE archive data of Aql X-1 and found that the source
138: showed coherent pulsation at a frequency near the neutron star spin
139: frequency (as inferred from burst oscillations) in its persistent
140: emission for 150 seconds during the peak of its 1998 outburst. The
141: discovery of this pulsation episode may provide new insights on the
142: issue of why some LMXBs pulse and some do not, after many years of
143: debate.
144:
145: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
146:
147: \section{Data Analysis} \label{dataanalysis}
148:
149: % One column figure
150: %-----------------------------------------------------------------------
151: \begin{figure}
152: \plotone{f1.eps}
153: \caption{Power spectrum of the last 128s data interval of
154: observation 30188-03-05-00 (see Fig. \ref{lc}). The power
155: spectrum was obtained from the full bandpass event mode data. The
156: frequency resolution is 1/128 Hz. The pulse is clearly visible at
157: $\sim$550 Hz. In the inset we plot the 2-60 keV light curve
158: folded at the pulsation period. Two cycles are plotted for
159: clarity. The solid line shows the best sinusoidal fit.}
160: \label{pds}
161: \end{figure}
162:
163: %_______________________________________________________________________
164:
165: We analyzed the whole PCA public archive of Aql X-1, searching
166: for coherent pulsation in its persistent emission. A total of 363
167: observations were analysed. In addition to the two Standard Modes
168: always present, the PCA data in different observations were obtained
169: in two different high-time resolution modes (either Event Mode
170: (125$\mu$s) or GoodXenon($\sim1\mu$s)); to all data we applied
171: barycenter correction for Earth and satellite motion. For each
172: observation we computed fast Fourier transforms from 128s data
173: intervals (after filtering for the presence of X-ray bursts) and
174: obtained power spectra with a Nyquist frequency of 4096 Hz and a
175: frequency resolution of 1/128 Hz. We searched all power spectra for
176: significant power at frequencies close to the expected pulse frequency
177: ($\sim$550 Hz). Only one detection was found with a Leahy power of 120
178: \citep{leahyetal83}, corresponding to a single trial chance
179: probability of $9\times10^{-27}$ at a frequency of 550.27 Hz (see
180: Fig. \ref{pds}).
181:
182: The single trial significance of the pulsation is 11 $\sigma$. Once we
183: take into account the number of trials (i.e. the number of frequency
184: bins times the number of 128 s power spectra we searched) the
185: significance becomes 9 $\sigma$ ($3\times10^{-17}$ chance
186: probability). Combined with its frequency being near that of the
187: previously observed burst oscillations, we conclude that the pulsation
188: is real and intrinsic to Aql X-1 and represents the spin of the
189: neutron star in this source. A full description of the total data
190: analysis is beyond the scope of this letter and will be the subject of
191: a forthcoming publication. Here we focus on the detailed analysis
192: performed of the single observation where pulsation was discovered.
193:
194: The pulsation was detected at the end of a 1600-second long
195: observation starting on 1998 March 10 at 22:28 UT (ObsId:
196: 30188-03-05-00). All five proportional counter units were active
197: during the whole observation, with a total 2-60 keV count rate varying
198: between 4500 and 5500 counts/sec. During this observation high-time
199: resolution data were available in Event Mode with 125$\mu$s time
200: resolution and 64 energy channels over the 2-60 keV instrument
201: bandpass. The light curve of the observation is shown in the top panel
202: of Figure \ref{lc}.
203:
204: The high power was detected in the last 128s interval of the
205: observation. By using an epoch folding technique we obtain a period
206: estimate of 1.8172746 ms. The folded profile was well modeled with a
207: sinusoidal component with 1.9$\pm$0.2 \% amplitude (see inset of
208: Fig. \ref{pds}). No second harmonic could be detected, with a 95\%
209: upper limit on the amplitude of 0.8 \%. To study the evolution and the
210: properties of the signal we extracted a light curve with maximum time
211: resolution, divided it into 16-second segments and determined the
212: pulsation amplitude and phase in each segment by an epoch folding
213: technique (see Fig. \ref{lc}). The pulsation is detected for $\sim150$
214: seconds at the end of the observation; its amplitude gradually
215: increases and then decreases again over these 150 seconds. To estimate
216: the accuracy of our period measurement we performed a linear fit to
217: the phases in the interval where pulsations are detected. Over this
218: interval the phase is consistent with being constant. From a 3
219: $\sigma$ upper limit of 10$^{-3}$ s$^{-1}$ on the slope, we derive a
220: best period measurement of 1.817275(3) ms. The short duration of the
221: pulsation did not allow us to obtain any useful upper limit on the
222: frequency drift. From the number of cycles we were able to phase
223: connect, we can estimate the coherence of the pulsation as Q $\ge
224: 8\times10^4$. \new{No correction for the binary orbital motion was
225: possible, given the small expected drift ($<$ 2 mHz in 150 seconds)
226: and the large uncertainity on the orbital phase \citep[$\sim$80\%,
227: applying the orbital solution from][]{welshetal00}}.
228:
229: In order to study the energy dependence of the pulsation we extracted
230: light curves with maximum time resolution in different energy ranges
231: and determined the pulsation amplitude and phase in each of them. In
232: Figure \ref{ampene}, we show the amplitude vs. energy diagram. The
233: pulsations have a strong energy dependence, being not detected at low
234: energies (with an upper limit on its amplitude of $\sim$1\% below 5
235: keV) and strong at high energies.
236:
237: % One column figure
238: %-----------------------------------------------------------------------
239: \begin{figure}
240: \plotone{f2.eps}
241: \caption{{\it Top panel}: 2-60 light curve with 16 seconds bin
242: size. {\it Bottom panels}: Amplitude ({\it middle}) and phase
243: ({\it bottom}) by epoch folding technique every 16 seconds,
244: vs. time. Values of amplitude and phase between 0 and $\sim$1400 s are
245: consistent with those expected from Poissonian noise.}
246: \label{lc}
247: \end{figure}
248:
249: %_______________________________________________________________________
250:
251: % One column figure
252: %-----------------------------------------------------------------------
253: \begin{figure}
254: \plotone{f3.eps}
255: \caption{Amplitude of the pulsation as a function of energy.
256: 1 $\sigma$ errors are shown.}
257: \label{ampene}
258: \end{figure}
259:
260: %_______________________________________________________________________
261:
262:
263: \section{Discussion} \label{discussion}
264:
265: We discovered millisecond X-ray pulsations in the persistent X-ray
266: emission of the LMXB Aql X-1. The pulsation is transient, present for
267: only $\sim$150 seconds and has an average fractional amplitude (over
268: the full 2-60 keV instrumental bandpass) of $\sim$2\%, increasing up
269: to $>$10\% at energies above $\sim$16 keV. It appears and disappears
270: gradually, on a time scale of a few tens of seconds, with a 3
271: $\sigma$ upper limit on its amplitude of 0.7\% during the $\sim$128
272: seconds before its appearance.
273:
274: This pulsating episode appears to be unique in the whole RXTE archive
275: of Aql X-1 observations. In one third of the observations
276: (corresponding to an exposure of 500 ks) the count rate is high enough
277: to allow the detection (3-$\sigma$, single trial) of a similar pulsating
278: episode as the one we observed. Even considering only these data
279: (instead of the total analysed exposure of 1300 ks) we obtain a
280: recurrence rate smaller than $3\times10^{-4}$. Such extreme rarety
281: in itself is extremely informative. Any physical interpretation must
282: not only in fact explain the appearance of the pulsation, but also its
283: extremely low occurrence rate.
284:
285: The frequency of the pulsation is 550.27 Hz, $\sim$0.53 Hz higher than
286: the reported asymptotic frequencies for burst oscillations in this
287: source \new{ \citep{zhangetal98b}}. The light curve shows no evidence of an
288: X-ray burst or other obvious feature within 1500 seconds before the
289: appearance of the pulsation. Together with its high coherence, this
290: leads us to conclude that the observed pulsation arises from a hot
291: spot spinning with the neutron star. This makes Aql X-1 the
292: millisecond X-ray pulsar with the longest orbital period and with the
293: highest time-average accretion rate known at present.
294:
295: The detection of this fast appearing pulsation in a LMXBs leads once
296: again to the same question: is the pulsation in Aql X-1 usually hidden
297: from the observer, by some scattering or screening medium that very
298: rarely disappears (or strongly reduces its otpical thickness) for
299: $\sim$150 seconds? Or is the pulsation usually absent, the episode we
300: discovered being the result of an occasional and rare asymmetry
301: on the neutron star surface?
302:
303: In the following we analyse different possible scenarios:
304:
305: \subsection{Permanent pulsation} \label{discus:permanent}
306:
307: If a hot spot is normally present on the neutron star surface, the
308: question is what made it observable only for $\sim$150 seconds.
309: Different authors have argued against or in favor of the hypothesis
310: that in non-pulsating sources the pulsation is washed out by some
311: optically thick scattering media. The main issue is the value of the
312: optical depth $\tau$ needed to attenuate the beamed
313: oscillations. \citet{psaltischakrabarty99} discuss the apparent
314: similarities between values of $\tau$ in pulsating and non-pulsating
315: sources, and discard this hypothesis. \citet{titarchuketal02} report
316: spectral fits giving other values of $\tau$ (for different
317: sources), and conclude that the data {\it do} support this
318: hypothesis. More recently, \citet{gogusetal07} report new,
319: independent spectral fits and conclude that in non-pulsating LMXBs
320: $\tau$ is not large enough to cause the pulsations to disappear
321: \citep[unless the electron temperature is very low, see
322: also][]{titarchuketal07}.
323:
324: This issue appears to be difficult to resolve, mainly because of a
325: substantial degeneracy between optical depth and electron temperature
326: in the spectral fits. However, the pulsation episode in Aql X-1 now
327: allows us to provide new insights into this issue. Even though, because
328: of their degeneracy, the absolute values of spectral parameters can not
329: be properly constrained, one would expect them to change when the
330: pulsation appears. In particular, the appearance of the pulsation, if
331: caused by a temporary (total or local) decrease of the optical depth
332: of the scattering media, is expected to be accompanied by a decrease
333: of the comptonization parameter, hence by a softening of the energy
334: spectrum. However, our spectral fits show that a two-component model
335: (disk plus boundary layer) is sufficient to fit the data, without the need
336: for any additional comptonized component. Moreover, neither spectral
337: fitting nor analysis of hardness light curves show any evidence of
338: spectral variability associated with the appearance of the
339: pulsation. Furthermore, the similarities between the spectral
340: properties of Aql X-1 and many other LMXBs would lead one to expect
341: similar episodes of pulsation in most of LMXBs. On the other hand,
342: the very hard spectrum of the observed pulsation (see
343: Fig. \ref{ampene}) suggests that the beamed radiation is strongly
344: comptonized \citep[for a discussion on the expected energy dependence
345: in the case of the screening scenario see][]{falanga07}.
346:
347: An alternative could be to assume a non-isotropic screening medium: an
348: occasional and rare hole in the medium or a reflecting cloud could
349: have allowed the radiation from the hot spot to avoid the screening
350: medium, thus becoming visible. Finally, the hot spot itself could have
351: moved becoming temporarily visible, either because of a different
352: absorption through the line of sight or because of a change in the
353: rotational of gravitational lensing geometry. We note that all these
354: hypotheses could in principle be consistent with the observed gradual
355: increase and decrease of the pulsation amplitude, but require very
356: special and fine-tuned geometries.
357:
358: \subsubsection{Estimate of the magnetic field} \label{discus:B}
359:
360: Under the standard dipolar magnetic channeling hypothesis for creating the hot
361: spot, we can give an estimate of the magnetic field. The magnetic field
362: must be strong enough to locally disrupt the disk flow. The pulsation
363: episode happens when the source is in the soft state, close to the
364: peak of the outburst. The 2-10 keV flux is $\sim 8.8\times10^{-9}$ erg
365: s$^{-1}$ cm$^{-2}$ which corresponds, assuming a distance of 5 kpc
366: \citep{jonkernelemans04} and a bolometric correction of 1.4
367: \citep{migliarifender06}, to a bolometric luminosity of $\sim
368: 3.7\times10^{37}$ erg s$^{-1}\sim0.15L_{\rm E}$. Given the relatively
369: high luminosity, by imposing that the radius at which this happens is
370: larger than the neutron star radius we can derive a stringent lower
371: limit on the magnetic dipole moment \citep{psaltischakrabarty99} as $
372: \mu \ge 0.47 \times 10^{26}~\mbox{G~cm}^3\;$ which corresponds to a
373: magnetic field of $\sim 10^8$ G at the pole.
374:
375: This lower limit is consistent with the upper limit of
376: $\sim1.8\times10^9$ G obtained by \citet{disalvoburderi03} from the
377: quiescent luminosity and with the estimate of $2\sim6\times10^8 $ G
378: obtained by \citet[assuming a corrected distance of 5 kpc instead of
379: 2.5 kpc]{campanaetal98} from the reported onset of the propeller
380: effect. Furthermore, our lower limit rules out the hypothesis of
381: residual accretion onto the neutron star to explain the quiescent X-ray
382: luminosity in Aql X-1, for which an upper limit for the magnetic field of
383: $\sim4\times10^7$ G had been derived (hypothesis a1 of
384: \citet{disalvoburderi03}).
385:
386: \subsection{Transient pulsation} \label{discus:transient}
387:
388: The second possibility is that of a temporary asymmetry on the
389: neutron star surface. Let us examine possible causes:
390:
391: \noindent {\it a) Magnetic channeling}
392:
393: Accretion rate and magnetic field are the two key physical quantities
394: playing a role in channeling of matter. A transient episode of
395: channeling of matter could arise due to a temporary change in the
396: accretion rate (although the X-ray flux, which is thought to be a good
397: indicator of $\dot{M}$, does not show any variability correlated with
398: the appearance of the pulsation) or a change in the accretion
399: geometry.
400:
401: An alternative is the occurrence of some variability of the magnetic
402: field. In the ``burying'' scenario, timescales for the magnetic field
403: to emerge are set by the Ohmic diffusion time and are at least of the
404: order of 100-1000 yr \citep{cummingetal01}, obviously far larger than
405: the few tens of seconds observed in Aql X-1. However a local temporary
406: decrease of the Ohmic diffusion timescale could in principle let the
407: magnetic field emerge. Possible causes could be a temporary local
408: disruption of the screening currents, a starquake, or magnetic
409: reconnections. Perhaps, a temporary local magnetic loop might form,
410: somewhat similar to those observed in the solar corona, where hot
411: coronal plasma ``populates'' different field lines. The magnetic field
412: in this loop might be {\it locally} strong enough to disturb the flow
413: of matter onto the neutron star surface. The resulting hot spot (non
414: necessarily on the magnetic pole) creates the observed pulsation.
415: However multipolar magnetic fields in neutron stars have not been
416: studied in detail, and these phenomena have to our knowledge not been
417: theoretically predicted.
418:
419: \noindent {\it b) Nuclear burning}
420:
421: The hot spot might be associated to a confined nuclear burning event.
422: After its onset, nuclear burning is expected to spread over the neutron
423: star surface in a very short timescale. Hence one would not expect the
424: resulting pulsation to last more than a few seconds. However the burst
425: oscillations observed in many sources during the decay of X-ray bursts
426: strongly challenge this simplified picture. Moreover, the discovery of
427: a very coherent pulsation during a superburst in 4U 1636-53
428: \citep{stromark02} showed that long-lived, very coherent pulsations
429: might arise due to nuclear burning. However the lack of any other
430: X-ray signature of nulcear burning (such as an X-ray burst) associated
431: with the pulsation in Aql X-1 strongly argues against this hypothesis.
432:
433: \noindent {\it \new{c) Surface wave}}
434:
435: \new{An asymmetry on the neutron star surface might result from an
436: unstable surface mode, as the ones suggested to explain burst
437: oscillations \citep[see e.g.,][]{pirobildsten04,narayancooper07}. Since
438: the amplitudes of surface modes have been predicted to increase with
439: energy \citep{pirobildsten06}, the hard energy dependence observed in
440: Aql X-1, more similar to that of burst oscillations than that of
441: accretion-powered pulsations, supports this possibility.}
442:
443: \noindent {\it d) Other possibilities}
444:
445: The extreme rarity of this pulsation episode leads us to examine also
446: more exotic physical explanations. Possible origins for a transient
447: hot spot could be a self-luminous magnetic event on the neutron star
448: surface, as magnetic reconnections or annihilations, or a sudden
449: release of potential energy in the deep crust during a starquake.
450:
451: Finally, we mention the possibility that some of the discussed
452: instabilities could arise due to a collision with an external body
453: \citep[see e.g.,][]{colgate81}. The fluence of the observed pulsation
454: is $\sim10^{35}$ergs, which corresponds to the free fall kinetic
455: energy of a mass of $\sim$15 km$^3$ of iron. However, the observed
456: gradual increase in the pulsation amplitude would be difficult to
457: explain within this scenario. The probability to have such an impact
458: is presumably extremely low. On the other hand, the observed event
459: {\it is} very rare.
460:
461: \new{Whatever is the physical origin of the observed pulsations, their
462: discovery implies the possible occurrance of transient pulsations in
463: any of the {\it apparently} non-pulsating LMXBs.}
464:
465: \acknowledgments
466:
467: The authors would like to thank A. G{\"u}rkan, G. Israel, and
468: L. Stella for very useful comments and stimulating discussions. We
469: also thank L. Bildsten and R. Cooper for pointing out the surface mode
470: possibility.
471: %PC acknowledges support from an NWO Post-doctoral
472: %Fellow. This work was partialy supported by an NWO Spinoza grant to
473: %M. van der Klis and by the Netherlands Research School for Astronomy
474: %(NOVA). This research has made use of data obtained through the High
475: %Energy Astrophysics Science Archive Research Center Online Service,
476: %provided by the NASA/Goddard Space Flight Center.
477:
478: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
479:
480: %\newpage
481:
482: \begin{thebibliography}{}
483:
484: \bibitem[Alpar et al.(1982)]{alparetal82}
485: Alpar, M. A., Cheng, A. F., Ruderman, M. A., \& Shaham, J., 1982, Nature, 300, 7128
486:
487: \bibitem[Altamirano et al.(2007)]{altamiranoetal07}
488: Altamirano, D., Casella, P., Patruno, A., van der Klis, M., Wijnands, R., 2007, ApJ, submitted
489:
490: \bibitem[Backer et al.(1982)]{backeretal82}
491: Backer, D. C., Kulkarni, S. R., Heiles, C., Davis, M. M., \& Goss, W. M., 1982, Nature, 300, 615
492:
493: \bibitem[Brainerd \& Lamb(1987)]{brainerdlamb87}
494: Brainerd, J., \& Lamb, F. K., 1987, ApJ, 317, L33
495:
496: \bibitem[Campana et al.(1998)]{campanaetal98}
497: Campana, S., Stella, L., Mereghetti S., Colpi, M., Tavani, M., Ricci, D., Dal Fiume, D., \& Belloni, T., 1998, ApJ, 499, L65
498:
499: \bibitem[Chakrabarty \& Morgan(1998)]{chakmorgan98}
500: Chakrabarty, D., \& Morgan, E. H., 1998, Nature, 394, 346
501:
502: \bibitem[Chakrabarty(2004)]{chakrabarty04}
503: Chakrabarty, D., 2004, ASP Conference, astro-ph/0408004
504:
505: \bibitem[Chevalier \& Ilovaisky(1991)]{chevalier91}
506: Chevalier, C., \& Ilovaisky, S. A., 1991, A\&A, 251, L11
507:
508: \bibitem[Colgate \& Petschek(1981)]{colgate81}
509: Colgate, S. A., \& Petschek, A. G., 1981, ApJ, 1981, 248, 771
510:
511: \bibitem[Cumming et al.(2001)]{cummingetal01}
512: Cumming, A., Zweibel, E., \& Bildsten, L., 2001, ApJ, 557, 958
513:
514: \bibitem[Di Salvo \& Burderi(2003)]{disalvoburderi03}
515: Di Salvo, T., \& Burderi, L., 2003, A\&A, 2003, 397, 723
516:
517: \bibitem[Falanga \& Titarchuk(2007)]{falanga07}
518: Falanga, M., \& Titarchuk, L., 2007, ApJ, 661, 1084
519:
520: \bibitem[Galloway et al.(2002)]{gallowayetal02}
521: Galloway, D. K., Chakrabarty, D., Morgan, E. H., \& Remillard, R. A., 2002, ApJ, 576, L137
522:
523: \bibitem[Galloway et al.(2005)]{gallowayetal05}
524: Galloway, D. K., Markwardt, C. B., Morgan, E. H., Chakrabarty, D., \& Strohmayer, T. E., 2005, ApJ, 622, L45
525:
526: \bibitem[Galloway(2006)]{galloway06}
527: Galloway, D. K., 2006, in AIP Conf. Proc. 840, The Transient Milky Way: A Perspective for MIRAX, ed. J. Braga, F. D'Amico, \& R. E. Rothschild (Melville: AIP), 50
528:
529: \bibitem[Galloway et al.(2007)]{gallowayetal07}
530: Galloway, D. K., Morgan, E. H., Krauss, M. I., Kaaret, P., \& Chakrabarty, D., ApJ, 654, L73
531:
532: \bibitem[Gogus et al.(2007)]{gogusetal07}
533: Gogus, E., Alpar, M. A., \& Gilfanov, M., 2007, ApJ, 659, 580
534:
535: \bibitem[Illarionov \& Sunyaev(1975)]{illarionov75}
536: Illarionov, A. F., \& Sunyaev, R. A., 1975, A\&A, 19, 185
537:
538: \bibitem[Jonker \& Nelemans(2004)]{jonkernelemans04}
539: Jonker, P., Nelemans, G., 2004, MNRAS, 354, 355
540:
541: \bibitem[Kaaret et al.(2006)]{kaaretetal06}
542: Kaaret, P., Morgan, E. H., Vanderspek, R., \& Tomsick, J. A., 2006, ApJ, 638, 963
543:
544: \bibitem[Kylafis \& Klimis(1987)]{kylafis87}
545: Kylafis, N., \& Klimis, G. S., 1987, ApJ, 323, 678
546:
547: \bibitem[Leahy et al.(1983)]{leahyetal83}
548: Leahy, D. A., Darbro, W., Elsner, R. F., Weisskopf, M. C., Kahn, S., Sutherland, P. G., Grindlay, J. E, 1983, ApJ, 266, 160
549:
550: \bibitem[Liu et al.(2007)]{liuetal07}
551: Liu, Q. Z., van Paradijs, J., van den Heuvel, E. P. J., 2007, A\&A, 469, 807
552:
553: \bibitem[Markwardt et al.(2002)]{markwardtetal02}
554: Markwardt, C. B., Swank, J. H., Strohmayer, T. E., in't Zand, J. J. M., \& Marshall, F. E., 2002, ApJ, 575, L21
555:
556: \bibitem[Markwardt et al.(2003)]{markwardtetal03}
557: Markwardt, C. B., Smith, E., \& Swank, J. H., 2003, IAU Circ. 8080
558:
559: \bibitem[Markwardt \& Swank(2003)]{markwardtswank03}
560: Markwardt, C. B., \& Swank, J. H., 2003, IAU Circ. 8144
561:
562: \bibitem[Markwardt et al.(2007)]{markwardtetal07}
563: Markwardt, C. B., Krimm, H. A., \& Swank, J. H., 2007, ATel 118
564:
565: \bibitem[Meszaros et al.(1988)]{meszarosetal88}
566: Meszaros, P., Riffert, H., \& Berthiaume, G., 1988, ApJ, 325, 204
567:
568: \bibitem[Migliari \& Fender(2006)]{migliarifender06}
569: Migliari, S., \& Fender, R., 2006, MNRAS, 366, 79
570:
571: \bibitem[Narayan \& Cooper(2007)]{narayancooper07}
572: Narayan, R., \& Cooper, R. L., 2007, ApJ, 665, 628
573:
574: \bibitem[Piro \& Bildsten(2004)]{pirobildsten04}
575: Piro, A. L., \& Bildsten, L., 2004, ApJ, 603, 252
576:
577: \bibitem[Piro \& Bildsten(2006)]{pirobildsten06}
578: Piro, A. L., \& Bildsten, L., 2006, ApJ, 638, 968
579:
580: \bibitem[Psaltis \& Chakrabarty(1999)]{psaltischakrabarty99}
581: Psaltis, D., \& Chakrabarty, D., 1999, ApJ, 521, 332
582:
583: \bibitem[Reig et al.(2000)]{reigetal00}
584: Reig, P., Mendez, M., van der Klis, M., \& Ford, E. C., 2000, ApJ, 530, 916
585:
586: %\bibitem[Rutledge et al.(1999)]{rutledgeetal99}
587: %Rutledge, R. E., Bildsten, L., Brown, E. F., Pavlov, G. G., \& Zavlin V. E., 1999, ApJ, 577, 346
588:
589: \bibitem[Stella et al.(1986)]{stellaetal86}
590: Stella, L., White, N. E., Rosner, R., 1986, ApJ, 308, 669
591:
592: \bibitem[Strohmayer \& Bildsten(2006)]{strobild06}
593: Strohmayer, T. E., \& Bildsten, L., 2006, in ``Compact stellar X-ray sources'', eds. W.H.G. Lewin and M. van der Klis, Cambridge University Press, p. 113-156
594:
595: \bibitem[Strohmayer \& Markwardt(2002)]{stromark02}
596: Strohmayer, T. E., \& Markwardt, C. B., 2002, ApJ, 577, 337
597:
598: \bibitem[Titarchuk et al.(2003)]{titarchuketal02}
599: Titarchuk, L., Cui, W., \& Wood, K. S., 2002, ApJ, 576, L49
600:
601: \bibitem[Titarchuk et al.(2007)]{titarchuketal07}
602: Titarchuk, L., Kuznetsov, S., \& Shaposhnikov, N., 2007, ApJ, 667, in press
603:
604: \bibitem[Thorstensen et al.(1978)]{thorstensenetal78}
605: Thorstensen, J., Charles, P., \& Bowyer, S., 1978, ApJ, 220, L131
606:
607: \bibitem[Welsh et al.(2000)]{welshetal00}
608: Welsh, W. F., Robinson, E. L., \& Young P., 2000, AJ, 120, 943
609:
610: \bibitem[Wijnands \& van der Klis(1998)]{wvdk98}
611: Wijnands, R., \& van der Klis, M., 1998, Nature, 394, 344
612:
613: \bibitem[Zhang et al.(1998a)]{zhangetal98a}
614: Zhang, S. N., Yu, W., \& Zhang, W., 1998a, ApJ, 494, L71
615:
616: \bibitem[Zhang et al.(1998b)]{zhangetal98b}
617: Zhang, W., Jahoda, K., Kelley, R. L., Strohmayer, T. E., Swank, J. H., \& Zhang, S. N., 1998b, ApJ, 495, L9
618:
619: \end{thebibliography}
620:
621: \newpage
622:
623:
624: \end{document}
625: %
626: %_______________________________________________________________________
627: