1: %\documentstyle[epsfig,mncite,times]{mn}
2: %\documentstyle[epsfig,mncite]{mn}
3: \documentclass[useAMS,usenatbib]{mn2e}
4: \usepackage{rotating}
5: \usepackage{journals}
6: \usepackage{times}
7:
8: %\renewcommand{\baselinestretch}{2}
9: \newcommand\msun {M$_{\odot}$}
10: \def\approxgt{\ifmmode \rlap{$>$}{}_{{}_{{}_{\textstyle\sim}}} \else%
11: $\rlap{$>$}{}_{{}_{{}_{\textstyle\sim}}}$\fi}
12: \def\approxlt{\ifmmode \rlap{$<$}{}_{{}_{{}_{\textstyle\sim}}} \else%
13: $\rlap{$<$}{}_{{}_{{}_{\textstyle\sim}}}$\fi}
14: \def\farcm{\hbox{$.\mkern-4mu^\prime$}}
15: \def\farcs{\hbox{$.\!\!^{\prime\prime}$}}
16: \def\degr{\hbox{$^\circ$}}
17: \def\arcmin{\hbox{$^\prime$}}
18: \def\arcsec{\hbox{$^{\prime\prime}$}}
19: \def\fdg{\hbox{$.\!\!^\circ$}}
20: \def\src{1A~1246--588}
21: \def\ecs{erg~cm$^{-2}$s$^{-1}$}
22: \def\es{erg s$^{-1}$}
23:
24:
25: \LARGE \normalsize \title[A kilohertz QPO in \src]{Detection of a 1258 Hz high--amplitude kilohertz
26: quasi--periodic oscillation in the ultra-compact X--ray binary \src}
27:
28: \author[Jonker et al.] {P.G.~Jonker$^{1,2,3}$\thanks{email :
29: p.jonker@sron.nl}, J.J.M. in 't Zand$^{1,3}$, M. M\'endez$^{1,3,4}$, M. van der Klis$^{4}$\\
30: $^1$SRON, Netherlands Institute for Space Research, Sorbonnelaan 2, 3584~CA, Utrecht, The Netherlands\\
31: $^2$Harvard--Smithsonian Center for Astrophysics, 60 Garden Street, Cambridge, MA~02138, Massachusetts,
32: U.S.A.\\
33: $^3$Astronomical Institute, Utrecht University, P.O.Box 80000, 3508 TA, Utrecht, The Netherlands\\
34: $^4$Astronomical Institute ``Anton Pannekoek'', University of Amsterdam, Kruislaan 403, 1098 SJ Amsterdam\\
35: }
36:
37: \begin{document}
38:
39: \maketitle
40:
41: \begin{abstract} \noindent We have observed the ultra-compact low--mass X--ray binary (LMXB) \src\ with the {\it
42: Rossi} X--ray Timing Explorer (RXTE). In this manuscript we report the discovery of a kilohertz quasi--periodic
43: oscillation (QPO) in \src. The kilohertz QPO was only detected when the source was in a soft high--flux state
44: reminiscent of the lower banana branch in atoll sources. Only one kilohertz QPO peak is detected at a relatively high
45: frequency of 1258$\pm$2 Hz and at a single trial significance of more than 7~$\sigma$. Kilohertz QPOs with a higher
46: frequency have only been found on two occasions in 4U~0614+09. Furthermore, the frequency is higher than that found for
47: the lower kilohertz QPO in any source, strongly suggesting that the QPO is the upper of the kilohertz QPO pair often
48: found in LMXBs. The full--width at half maximum is 25$\pm$4 Hz, making the coherence the highest found for an upper
49: kilohertz QPO. From a distance estimate of $\approx$6 kpc from a radius expansion burst we derive that \src\ is at a
50: persistent flux of $\approx$0.2--0.3 per cent of the Eddington flux, hence \src\ is one of the weakest LMXBs for which
51: a kilohertz QPO has been detected. The root--mean--square (rms) amplitude in the 5--60 keV band is 27$\pm$3 per cent,
52: this is the highest for any kilohertz QPO source so far, in line with the general anti--correlation between source
53: luminosity and rms amplitude of the kilohertz QPO peak identified before. Using the X--ray spectral information we
54: produce a colour--colour diagram. The source behaviour in this diagram provides further evidence for the atoll nature
55: of the source.
56:
57: \end{abstract}
58:
59: \begin{keywords} stars: individual (\src) ---
60: accretion: accretion discs --- stars: binaries --- stars: neutron
61: --- X-rays: binaries
62: \end{keywords}
63:
64: \section{Introduction}
65:
66: Ultra-compact X--ray binaries (UCXBs) are low--mass X--ray binaries (LMXBs) that have an orbital period shorter
67: than $P_{\rm orb}\approx$1~hr, implying such a small Roche lobe that the donor in an UCXB will have lost (most of)
68: its hydrogen (\citealt{1986ApJ...304..231N}; \citealt{1986AA...155...51S}). For 12 LMXBs $P_{\rm orb}$ has been
69: measured to be in the ultra-compact regime (\citealt{2006astro.ph..5722N}).
70:
71: In the power spectra of $\sim$20 accreting neutron star LMXBs kilohertz quasi--periodic oscillations
72: (QPOs) have been discovered (see \citealt{2006csxs.book...39V} for a review). Kilohertz QPOs are thought to be
73: caused by motion of matter a few kilometres above the surface of accreting neutron stars. Even though there is as
74: yet no agreement about the exact physical mechanism causing the X--ray light curves to be modulated at kilohertz
75: frequencies, most models agree that the frequency of one of the observed kilohertz QPOs reflects the frequency of
76: orbital motion at the inner edge of the accretion disc. Hence, the kilohertz QPOs potentially allow one to detect
77: effects of the strong gravitational fields and to constrain the neutron star mass--radius relation
78: (e.g.~\citealt{milaps1998}).
79:
80: Often two kilohertz QPO peaks separated by $\Delta\nu$=200--360\,Hz are found (again see \citealt{2006csxs.book...39V}
81: for a review). The highest frequencies have been observed in the UCXB 4U~0614+09 ($1329\pm4$~Hz, \citealt{vafova2000}
82: and 1273.6$\pm$9.5\,Hz; \citealt{2002ApJ...568..912V}). At high ($\approxgt$400 Hz) neutron star spin frequencies $\Delta
83: \nu$ has been found to be equal to half the spin frequency (e.g.~in SAX~J1808.4--3658; \citealt{2003Natur.424...44W}),
84: whereas in sources where the neutron star spin is $\approxlt$350 Hz $\Delta \nu$ is found to be close to the spin
85: frequency (see \citealt{2006AdSpR..38.2675V}, \citealt{2006csxs.book...39V} and \citealt{2006csxs.book..113S} for
86: reviews).
87:
88: \src\ was discovered with the Ariel--V observatory in the mid 1970s at a level of roughly
89: $(1-2)\times10^{-10}$~\ecs\ (2--10 keV) by \citet{1976MNRAS.177P..13S}, but it had also been detected by UHURU a few
90: years earlier (\citealt{1978ApJS...38..357F}). The Ariel--V data exhibited a flare with a peak 5 times above the
91: lowest level (\citealt{1977MNRAS.179P..27C}). On Feb. 16, 1985, the first pointed observation of \src\ was taken
92: with EXOSAT. From the HEASARC standard products we found that the light curve is featureless and the 2--10 keV
93: flux is $9.5\times10^{-11}$~\ecs. In 1997, the first X--ray burst was detected with the BeppoSAX Wide Field
94: Cameras (WFCs; \citealt{1997IAUC.6538....2P}) from a position consistent with that of \src\
95: (\citealt{1997IAUC.6546....1B}). The increased localisation accuracy provided by the WFCs enabled an
96: identification with a ROSAT source and thus the position is known with an accuracy of 9\arcsec\ (Boller et al.
97: 1997). This enabled \citet{2006A&A...446L..17B} to identify the optical counterpart and determine an UCXB nature
98: through the ratio of the X--ray to optical flux. In August 2006 a type--I X--ray burst was detected with the
99: Swift Burst Alert Telescope (\citealt{2006GCN..5436....1R}; \citealt{2006ATel..875....1K}) and the satellite
100: autonomously slewed to the target within 193~s. The burst was still detectable with the Swift X--ray telescope
101: and the total duration could be accurate determined to be 28 minutes. Such a duration is similar to the long
102: bursts from other UCXBs (see \citealt{2007astro.ph..1810I}).
103:
104: Here, we report on our analysis of {\it Rossi} X--ray Timing Explorer (RXTE) observations of \src.
105:
106:
107: \section{Observations, analysis and results}
108:
109: We have used proportional counter array (PCA) data from 55 short RXTE observations of \src. All data obtained between MJD
110: 53720--54108 (UTC; Dec 16, 2005--Jan.~15, 2007) amounting to a total of $\approx45$ ksec were taken from program P90042. In
111: the {\it top panel} of Figure~\ref{rxtelc} we have plotted the light curve obtained with the RXTE All Sky Monitor (ASM).
112: Besides ASM count rate we show the source flux in \ecs; for this we used the fact that the \src\ PCA X--ray spectrum in the
113: low--flux state resembles that of the Crab. The source is persistently detected. It remained at a level of roughly 0.37 ASM
114: counts s$^{-1}$ which is 5~mCrab (2--12 keV), except for a long period between Oct.~1998 (MJD~51100) and Oct.~2001
115: (MJD~52200) when it exhibited much more variability and the flux went up to 2.4 ASM counts~s$^{-1}$ (32 mCrab) and
116: similarly after mid 2004. There are flare--like features visible with peak fluxes that are $\approx$3 times higher than
117: normal, last a couple of weeks and have a recurrence time between $\approx$20 and 100 days. In the {\it bottom panel} of
118: Figure~\ref{rxtelc} we show the absorbed 2--10 keV flux as derived from fitting an absorbed power--law plus blackbody model
119: to the X--ray spectra extracted from the RXTE/PCA Standard 2 data (using {\sl ftools 6.0.4} and {\sl xspec 12.2.1};
120: \citealt{ar1996}).
121:
122: \begin{figure}
123: \includegraphics[angle=0,width=8cm,clip]{asm.ps}
124:
125: \caption{{\it Top panel:} The RXTE All Sky Monitor 7--day--average light curve of \src. The ordinate on the
126: left/right hand side gives the source flux in ASM counts per second/2--10 keV in \ecs. The flare episodes where
127: the count rate increases by a factor $\approx$3 can be discerned. {\it Bottom panel:} The observed (absorbed)
128: 2--10 keV flux derived from the PCA monitoring observations of \src\ fitting an absorbed blackbody plus
129: power--law model to the Standard 2 data. We have also indicated during which observations the kilohertz QPO was
130: significantly detected. Note that the abscissa of the top and bottom panel are different and that the size of
131: the data points is larger than the error on the data points in the {\it bottom panel}.}
132:
133: \label{rxtelc}
134: \end{figure}
135:
136: Using 16~s--long segments of high time resolution PCA data (122$\mu$s resolution), we calculated power spectra up to a
137: Nyquist frequency of 4096~Hz in the full energy band of 2--60 keV. We first combined all the power spectra from all data.
138: The high--frequency (256--4096 Hz) part of the average power spectrum was searched for the presence of kilohertz QPOs. We
139: did not detect a kilohertz QPO. However, we next selected the data for which the total PCA count rate was above $\sim$150
140: counts s$^{-1}$. A strong kilohertz QPO was discovered. Subsequent subselections showed that the kilohertz QPO was
141: significantly present only in the flare data before MJD 53800 (see Fig.~\ref{rxtelc}).
142:
143: We fitted the 64--2048 Hz part of the average power spectrum of the three observations indicated in Fig.~1 with "kHz QPO"
144: combined. A fit function consisting of the sum of a constant, to represent the Poisson noise, and a Lorentzian to represent
145: the kilohertz QPO was used. With a reduced $\chi^2=0.86$ for 207 degrees of freedom the fit was good. The kilohertz QPO has
146: a frequency of 1258$\pm$2 Hz and a full--width at half maximum of 25$\pm$4 Hz, see Fig.~\ref{qpo} (here and below we report
147: 1~$\sigma$ single parameter errors determined using $\Delta\chi^2=1.0$). The single trial significance in the full 2--60
148: keV band is 7.5~$\sigma$. Even when taking into account the $\approxlt 1000$ trials we performed, the significance is still
149: 6.8~$\sigma$. The rms amplitude expressed as a fraction of the source count rate in the 2--60 keV band is 17.2$\pm$1.5 per
150: cent. In order to compare the rms amplitude of the QPO with previous work of e.g.~\citet{2001ApJ...553..335J} we also
151: calculated power spectra in the 5--60 keV energy band. The fractional rms amplitude in that band is 27$\pm$3 per cent. The
152: average source count rate in the 2-60 keV band in the three observations containing the kilohertz QPO is 85$\pm$1 counts
153: per second whereas it is 52$\pm$1 counts per second in the 5--60 keV band. As we will argue in the Discussion the detected
154: kilohertz QPO can most likely be associated with the upper of the kilohertz QPO pair found in several LMXBs. In order to
155: put an upper limit on the presence of a lower kilohertz QPO we determine a 95 per cent confidence upper limit on the
156: presence of a kilohertz QPO at a frequency 200--400 Hz less than 1258 Hz for a QPO full--width at half maximum of 50, 100,
157: or 150 Hz. The upper limit on the presence of such a QPO is 20 per cent (2--60 keV). For completeness we did the
158: same for a kilohertz QPO at a frequency 200-400 Hz higher than 1258 Hz. The upper limit on the presence of such a QPO is 13
159: per cent (2--60 keV). We also inspected the low--frequency (0.1--256 Hz) part of the average power spectrum containing the
160: kilohertz QPO. We found a strong (27.4$\pm$1.0 per cent; 2--60 keV) band--limited noise component. We characterised the
161: band--limited noise component by fitting a Lorentzian to the averaged low--frequency power spectrum. Its central frequency
162: was 24$\pm$1 Hz, whereas the full--width at half maximum was 24$\pm$2 Hz.
163:
164:
165: \begin{figure}
166: \includegraphics[angle=0,width=8cm,clip]{1246kilohertzqpo.ps}
167:
168: \caption{ Part of the \citet{ledael1983} normalised average 2--60 keV power spectrum of the three observations indicated
169: in Figure~\ref{rxtelc}. The kilohertz QPO is clearly visible. It has a more than 7~$\sigma$ single trial
170: significance. The solid line represents the best fitting model consisting of a constant plus a Lorentzian.}
171:
172: \label{qpo}
173: \end{figure}
174:
175: Using the 16~s resolution Standard 2 data we further created colour--colour diagrams (CDs; Figure~\ref{cd}). A soft
176: colour is plotted vs.~a hard colour; the soft and hard colour are defined as the ratio between the count rates in the
177: 3.5--6.4 keV and 2.0--3.5 keV and that between the 9.7--16.0 keV and the 6.4--9.7 keV band, respectively. The
178: observed count rates have been corrected for deadtime effects and we have subtracted the background for each
179: proportional counter unit (PCU) separately. We further corrected the colours for small changes in the instrumental
180: response (the gain) and for differences in response between the various PCUs, using Crab observations close in time
181: and assuming the Crab colours to be constant (for a full description of the correction see
182: \citealt{2003ApJ...596.1155V}). The CD of \src\ bears hallmarks of that of the Island and lower Banana branch of an
183: 'atoll' source (\citealt{hava1989}; see Figure~\ref{cd}). Clearly present is the Island State in the top left part of
184: Fig.~3 where the source, as in other atoll sources, spends much of its time since motions in the CD are generally
185: slower here. Due to the low count rate and the fact that often only one or two PCUs were active, we have plotted the
186: average colour of each observation. Given the short duration of each observation (on average less than 1 ksec), the
187: chance that the source moved over a large area in the CD is small. The large dots indicate two of the observations
188: when the kilohertz QPO was present. The elevation above the Earth's limb for the first PCA observation, when the
189: kilohertz QPO was also present, was too low to create a reliable colour--colour point. The timing properties of that
190: observation are not affected by the close presence of the Earth's limb.
191:
192:
193: \begin{figure}
194: \includegraphics[angle=0,width=8cm,clip]{kleurkleurperOBSID.ps}
195:
196: \caption{ Colour--colour diagram of the Standard 2 PCA data. The small dots are the average colour of the
197: source during one PCA observation. The large dots at a soft colour between 1.1--1.2 are the
198: observations when the kilohertz QPO was detected. The soft and hard colour are defined as the ratio between the count
199: rates in the 3.5--6.4 keV and 2.0--3.5 keV and that between the 9.7--16.0 keV and the 6.4--9.7 keV band,
200: respectively.}
201:
202: \label{cd}
203: \end{figure}
204:
205:
206: \section{Discussion}
207:
208: Using RXTE PCA data we have discovered a kilohertz QPO in the X--ray emission of the ultra--compact X--ray
209: binary \src. At a frequency of 1258$\pm$2 Hz the kilohertz QPO frequency is rather high in comparison with those
210: found in other sources (for an overview see \citealt{2006csxs.book...39V}). Furthermore, the fractional rms
211: amplitude of the kilohertz QPO measured in the 5--60 keV band is the highest found in any kilohertz QPO source.
212: From the properties of a photospheric radius expansion burst In 't Zand et al.~(in preparation) derive a
213: distance to the source of $\approx$6 kpc. This implies a persistent source luminosity of $\approx$0.2--0.3 per
214: cent of the Eddington luminosity, making \src\ one of the weakest LMXBs for which a kilohertz QPO has been
215: detected; the other similarly weak source is 2S~0918--549 (\citealt{2001ApJ...553..335J}). The high frequency
216: and high coherence of the kilohertz QPO peak suggests that it can be identified as the upper kilohertz QPO. The
217: lower kilohertz QPO has never been detected at frequencies higher than $\approx$1000 Hz when both kilohertz QPO
218: peaks were detected simultaneously (\citealt{2006csxs.book...39V}) and the coherence of the upper kilohertz QPO
219: is found to increase at frequencies above $\approx$1100 Hz (\citealt{2006MNRAS.371.1925M,2006MNRAS.370.1140B}).
220: The spectral properties suggest that \src\ belongs to the class of 'atoll' sources. The kilohertz QPO was only
221: significantly detected when the source was in the lower banana branch similar to what has been found in other
222: atoll sources (\citealt{mevafo1999}; \citealt{vafova2000}).
223:
224: The high frequency yields the exciting possibility to obtain constraints on the neutron star. For instance, if a
225: second simultaneous kilohertz QPO is discovered, the neutron star spin rate can be constrained. In the 8 sources
226: where both twin kilohertz QPOs and neutron star spins were directly measured, $\Delta\nu\approx\nu_{\rm spin}$ or
227: $\Delta\nu\approx\nu_{\rm spin}/2$ (see review by \citealt{2006csxs.book...39V}). Another constraint involves the
228: neutron star mass and radius. The measured 1258 Hz frequency is high. Only in one object has a higher frequency ever
229: been measured (see Section 1). If one interprets the observed QPO frequency as a Keplerian frequency and requires
230: that the neutron star is smaller than the radius associated with that Keplerian frequency a neutron star
231: mass--dependent limit can be put on the neutron star radius. In addition if one assumes that the innermost stable
232: circular orbit has a radius also smaller than the associated Keplerian radius a limit on the neutron star mass is
233: obtained (\citealt{milaps1998}). Assuming a non--rotating neutron star this yields the following limits on the mass
234: and radius of the neutron star for the 1258 Hz QPO: M${\rm _{NS}}<$1.75 M$_\odot$ and R$<$15.5 km.
235:
236: As mentioned above the fractional rms amplitude of the kilohertz QPO is the highest found to date. Similarly, the
237: fractional rms amplitude of the broad low--frequency noise component is very high at 27 per cent (2--60 keV). This all
238: fits--in with the trend of increasing rms amplitude for a decreasing source luminosity as shown for the upper kilohertz
239: QPO before by \citet{2001ApJ...553..335J}. Using the data from \citet{2001ApJ...553..335J} and the data on \src\ we
240: have fitted the relation between the fractional rms amplitude of the upper kilohertz QPO and source luminosity. A
241: power--law fit gives an index of $-0.35\pm$0.01. From figure 3 in \citet{2006MNRAS.371.1925M} it seems as if the
242: increase in the maximum fractional rms amplitude of a source with decreasing source luminosity levels off at 18--20 per
243: cent (measured in the 2--60 keV band). However, the kilohertz QPO that we found in \src\ does not seem to be compatible
244: with this trend. Even though the measured amplitude in the 2--60 keV band of 17.2$\pm$1.5 per cent is close to this
245: saturation level, this amplitude is measured while the QPO frequency was very high (1258 Hz). In other sources the rms
246: amplitude of the upper kilohertz QPO decreases strongly at upper kilohertz QPO frequencies above 700--800 Hz. Hence,
247: either the relation between kilohertz QPO frequency and fractional rms amplitude is significantly different in \src\
248: from that in other sources or the increase in fractional rms amplitude and source luminosity does not level off at
249: 18--20 per cent but keeps increasing.
250:
251: \section*{Acknowledgments}
252:
253: \noindent PGJ acknowledges support from the Netherlands Organisation for Scientific Research and R.~Cornelisse for
254: comments on an earlier version of the Manuscript.
255:
256:
257:
258:
259: \begin{thebibliography}{27}
260: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
261:
262: \bibitem[{{Arnaud}(1996)}]{ar1996}
263: {Arnaud}, K.~A., 1996, in ASP Conf. Ser. 101: Astronomical Data Analysis
264: Software and Systems V, vol.~5, p.~17
265:
266: \bibitem[{{Barret} et~al.(2006){Barret}, {Olive}, \&
267: {Miller}}]{2006MNRAS.370.1140B}
268: {Barret}, D., {Olive}, J.-F., {Miller}, M.~C., 2006, \mnras, 370, 1140
269:
270: \bibitem[{{Bassa} et~al.(2006){Bassa}, {Jonker}, {in't Zand}, \&
271: {Verbunt}}]{2006A&A...446L..17B}
272: {Bassa}, C.~G., {Jonker}, P.~G., {in't Zand}, J.~J.~M., {Verbunt}, F., 2006,
273: \aap, 446, L17
274:
275: \bibitem[{{Boller} et~al.(1997){Boller}, {Haberl}, {Voges}, {Piro}, \&
276: {Heise}}]{1997IAUC.6546....1B}
277: {Boller}, T., {Haberl}, F., {Voges}, W., {Piro}, L., {Heise}, J., 1997,
278: \iaucirc, 6546, 1
279:
280: \bibitem[{{Carpenter} et~al.(1977){Carpenter}, {Eyles}, {Skinner}, {Wilson}, \&
281: {Willmore}}]{1977MNRAS.179P..27C}
282: {Carpenter}, G.~F., {Eyles}, C.~J., {Skinner}, G.~K., {Wilson}, A.~M.,
283: {Willmore}, A.~P., 1977, \mnras, 179, 27P
284:
285: \bibitem[{{Forman} et~al.(1978){Forman}, {Jones}, {Cominsky}, {Julien},
286: {Murray}, {Peters}, {Tananbaum}, \& {Giacconi}}]{1978ApJS...38..357F}
287: {Forman}, W., {Jones}, C., {Cominsky}, L., {Julien}, P., {Murray}, S.,
288: {Peters}, G., {Tananbaum}, H., {Giacconi}, R., 1978, \apjs, 38, 357
289:
290: \bibitem[{{Hasinger} \& {van der Klis}(1989)}]{hava1989}
291: {Hasinger}, G., {van der Klis}, M., 1989, \aap, 225, 79
292:
293: \bibitem[{{in 't Zand} et~al.(2007){in 't Zand}, {Jonker}, \&
294: {Markwardt}}]{2007astro.ph..1810I}
295: {in 't Zand}, J.~J.~M., {Jonker}, P.~G., {Markwardt}, C.~B., 2007, ArXiv
296: Astrophysics e-prints
297:
298: \bibitem[{{Jonker} et~al.(2001)}]{2001ApJ...553..335J}
299: {Jonker}, P.~G., et~al., 2001, \apj, 553, 335
300:
301: \bibitem[{{Kong}(2006)}]{2006ATel..875....1K}
302: {Kong}, A.~K.~H., 2006, The Astronomer's Telegram, 875, 1
303:
304: \bibitem[{{Leahy} et~al.(1983){Leahy}, {Darbro}, {Elsner}, {Weisskopf}, {Kahn},
305: {Sutherland}, \& {Grindlay}}]{ledael1983}
306: {Leahy}, D.~A., {Darbro}, W., {Elsner}, R.~F., {Weisskopf}, M.~C., {Kahn}, S.,
307: {Sutherland}, P.~G., {Grindlay}, J.~E., 1983, \apj, 266, 160
308:
309: \bibitem[{{M{\'e}ndez}(2006)}]{2006MNRAS.371.1925M}
310: {M{\'e}ndez}, M., 2006, \mnras, 371, 1925
311:
312: \bibitem[{{M{\'e}ndez} et~al.(1999){M{\'e}ndez}, {van der Klis}, {Ford},
313: {Wijnands}, \& {van Paradijs}}]{mevafo1999}
314: {M{\'e}ndez}, M., {van der Klis}, M., {Ford}, E.~C., {Wijnands}, R., {van
315: Paradijs}, J., 1999, \apjl, 511, L49
316:
317: \bibitem[{{Miller} et~al.(1998){Miller}, {Lamb}, \& {Psaltis}}]{milaps1998}
318: {Miller}, M.~C., {Lamb}, F.~K., {Psaltis}, D., 1998, \apj, 508, 791
319:
320: \bibitem[{{Nelemans} \& {Jonker}(2006)}]{2006astro.ph..5722N}
321: {Nelemans}, G., {Jonker}, P.~G., 2006, ArXiv Astrophysics e-prints
322:
323: \bibitem[{{Nelson} et~al.(1986){Nelson}, {Rappaport}, \&
324: {Joss}}]{1986ApJ...304..231N}
325: {Nelson}, L.~A., {Rappaport}, S.~A., {Joss}, P.~C., 1986, \apj, 304, 231
326:
327: \bibitem[{{Piro} et~al.(1997)}]{1997IAUC.6538....2P}
328: {Piro}, L., et~al., 1997, \iaucirc, 6538, 2
329:
330: \bibitem[{{Romano} et~al.(2006)}]{2006GCN..5436....1R}
331: {Romano}, P., et~al., 2006, GRB Coordinates Network, 5436, 1
332:
333: \bibitem[{{Savonije} et~al.(1986){Savonije}, {de Kool}, \& {van den
334: Heuvel}}]{1986AA...155...51S}
335: {Savonije}, G.~J., {de Kool}, M., {van den Heuvel}, E.~P.~J., 1986, \aap, 155,
336: 51
337:
338: \bibitem[{{Seward} et~al.(1976){Seward}, {Page}, {Turner}, \&
339: {Pounds}}]{1976MNRAS.177P..13S}
340: {Seward}, F.~D., {Page}, C.~G., {Turner}, M.~J.~L., {Pounds}, K.~A., 1976,
341: \mnras, 177, 13P
342:
343: \bibitem[{{Strohmayer} \& {Bildsten}(2006)}]{2006csxs.book..113S}
344: {Strohmayer}, T., {Bildsten}, L., 2006, {New views of thermonuclear bursts},
345: Compact stellar X-ray sources, p. 113
346:
347: \bibitem[{{van der Klis}(2006{\natexlab{a}})}]{2006AdSpR..38.2675V}
348: {van der Klis}, M., 2006{\natexlab{a}}, Advances in Space Research, 38, 2675
349:
350: \bibitem[{{van der Klis}(2006{\natexlab{b}})}]{2006csxs.book...39V}
351: {van der Klis}, M., 2006{\natexlab{b}}, {Rapid X-ray Variability}, Compact
352: stellar X-ray sources, p.~39
353:
354: \bibitem[{{van Straaten} et~al.(2000){van Straaten}, {Ford}, {van der Klis},
355: {M{\'e}ndez}, \& {Kaaret}}]{vafova2000}
356: {van Straaten}, S., {Ford}, E.~C., {van der Klis}, M., {M{\'e}ndez}, M.,
357: {Kaaret}, P., 2000, \apj, 540, 1049
358:
359: \bibitem[{{van Straaten} et~al.(2002){van Straaten}, {van der Klis}, {di
360: Salvo}, \& {Belloni}}]{2002ApJ...568..912V}
361: {van Straaten}, S., {van der Klis}, M., {di Salvo}, T., {Belloni}, T., 2002,
362: \apj, 568, 912
363:
364: \bibitem[{{van Straaten} et~al.(2003){van Straaten}, {van der Klis}, \&
365: {M{\'e}ndez}}]{2003ApJ...596.1155V}
366: {van Straaten}, S., {van der Klis}, M., {M{\'e}ndez}, M., 2003, \apj, 596, 1155
367:
368: \bibitem[{{Wijnands} et~al.(2003){Wijnands}, {van der Klis}, {Homan},
369: {Chakrabarty}, {Markwardt}, \& {Morgan}}]{2003Natur.424...44W}
370: {Wijnands}, R., {van der Klis}, M., {Homan}, J., {Chakrabarty}, D.,
371: {Markwardt}, C.~B., {Morgan}, E.~H., 2003, \nat, 424, 44
372:
373: \end{thebibliography}
374:
375: \end{document}
376:
377:
378:
379:
380:
381:
382:
383:
384: