1: %\documentstyle[epsfig,mncite,times]{mn}
2: %\documentstyle[epsfig,mncite]{mn}
3: \documentclass[useAMS,usenatbib]{mn2e}
4: \usepackage{rotating}
5: \usepackage{journals}
6:
7: %\renewcommand{\baselinestretch}{2}
8: \newcommand\msun {M$_{\odot}$}
9: \def\approxgt{\ifmmode \rlap{$>$}{}_{{}_{{}_{\textstyle\sim}}} \else%
10: $\rlap{$>$}{}_{{}_{{}_{\textstyle\sim}}}$\fi}
11: \def\approxlt{\ifmmode \rlap{$<$}{}_{{}_{{}_{\textstyle\sim}}} \else%
12: $\rlap{$<$}{}_{{}_{{}_{\textstyle\sim}}}$\fi}
13: \def\farcm{\hbox{$.\mkern-4mu^\prime$}}
14: \def\farcs{\hbox{$.\!\!^{\prime\prime}$}}
15: \def\degr{\hbox{$^\circ$}}
16: \def\arcmin{\hbox{$^\prime$}}
17: \def\arcsec{\hbox{$^{\prime\prime}$}}
18: \def\fdg{\hbox{$.\!\!^\circ$}}
19:
20:
21: \LARGE \normalsize \title[The companion star of Cir~X--1?]{Detection of the radial velocity
22: curve of the B5--A0 supergiant companion star of Cir~X--1?}
23:
24: \author[Jonker, Nelemans, \& Bassa] {P.G.~Jonker$^{1,2,3}$\thanks{email :
25: p.jonker@sron.nl. Based on observations made with ESO telescopes at the Paranal Observatories under
26: programme ID 274.D--5047(A)}, G. Nelemans$^{4}$, C.G.~Bassa$^3$ \\
27: $^1$SRON, Netherlands Institute for Space Research, Sorbonnelaan 2, 3584~CA, Utrecht, The Netherlands\\
28: $^2$Harvard--Smithsonian Center for Astrophysics, 60 Garden Street, Cambridge, MA~02138, Massachusetts,
29: U.S.A.\\
30: $^3$Astronomical Institute, Utrecht University, P.O.Box 80000, 3508 TA, Utrecht, The Netherlands\\
31: $^4$Department of Astrophysics, IMAPP, Radboud University Nijmegen, Toernooiveld 1, 6525 ED, Nijmegen, The Netherlands\\
32: }
33:
34: \begin{document}
35:
36: \maketitle
37:
38: \begin{abstract} \noindent In this Paper we report on phase resolved $I$--band optical spectroscopic and photometric
39: observations of Cir~X--1 obtained with the Very Large Telescope. The spectra are dominated by Paschen absorption
40: lines at nearly all orbital phases except near phase zero (coinciding with the X--ray dip) when the absorption lines
41: are filled--in by broad Paschen emission lines. The radial velocity curve of the absorption lines corresponds to an
42: eccentric orbit ($e=0.45$) whose period and time of periastron passage are consistent with the period and phase
43: predicted by the most recent X--ray dip ephemeris. We found that the $I$--band magnitude decreases from 17.6 to $\sim
44: 16.8$ near phase 0.9--1.0, this brightening coincides in phase with the X--ray dip. Even though it is likely that the
45: absorption line spectrum is associated with the companion star of Cir~X--1, we cannot exclude the possibility that
46: the spectrum originates in the accretion disc. However, if the spectrum belongs to the companion star, it must be a
47: supergiant of spectral type B5--A0. If we assume that the compact object does not move through the companion star at
48: periastron, the companion star mass is constrained to $\approxlt$10 M$_\odot$ for a 1.4~M$_\odot$ neutron star,
49: whereas the inclination has to be $\approxgt 13.7^\circ$. Alternatively, the measured absorption lines and their
50: radial velocity curve can be associated with the accretion disc surrounding a 1.4~M$_\odot$ neutron star and its
51: motion around the centre of mass. An absorption line spectrum from an accretion disc is typically found when our
52: line--of--sight passes through the accretion disc rim implying a high inclination. In this scenario the companion
53: star mass is found to be $\sim$0.4 M$_\odot$. However, from radio observations it was found that the angle between
54: the line--of--sight and the jet axis is smaller than 5$^\circ$. This would mean that the jet ploughs through the
55: accretion disc in this scenario, making this solution less probable.
56:
57: \end{abstract}
58:
59: \begin{keywords} stars: individual (Cir~X--1) ---
60: accretion: accretion discs --- stars: binaries --- stars: neutron
61: --- X-rays: binaries
62: \end{keywords}
63:
64: \section{Introduction}
65:
66: \noindent X--ray binaries are binary systems harbouring a neutron star or black hole compact object that
67: accretes matter from either a low-- or a high--mass companion star (LMXB and HMXB, respectively). LMXBs
68: are typically old systems whereas the early type companion star of HMXBs precludes systems older than a
69: few times 10$^7$ yr. Recently, it has been realised that many X--ray binaries might have started--off as
70: intermediate mass X--ray binaries (cf.~Cyg~X--2; \citealt{1999A&A...350..928T};
71: \citealt{2002ApJ...565.1107P}). One intriguing system that so far has defied classification is Cir~X--1.
72: It has a 16.6~day orbital period (\citealt{1976ApJ...208L..71K}; see \citealt{2004MNRAS.348..458C} for
73: the latest X--ray ephemeris). Owing to the detection of type~I X--ray bursts, the compact object in
74: Cir~X--1 likely is a neutron star (\citealt{1986MNRAS.219..871T}; \citealt{1986MNRAS.221P..27T}).
75:
76: The X--ray and radio behaviour of the source is complex. The long--term X--ray lightcurve of the source has been
77: discussed in detail by \citet{2003HEAD....7.1723S}. One of the striking features of the X--ray lightcurve is the
78: periodic appearance of dips. In radio, an arc minute scale radio nebula was found (\citealt{1993MNRAS.261..593S}).
79: Furthermore, a relativistic outflow has been detected on scales of arc seconds that is aligned with the arc minute
80: scale jet (\citealt{1998ApJ...506L.121F}; \citealt{2004Natur.427..222F}). The inclination of the jet with respect
81: to the line--of--sight has to be less than 5$^\circ$ (\citealt{2004Natur.427..222F}). The presence of relativistic
82: outflows detectable as synchrotron emission in the radio band argues against a high magnetic field neutron star in
83: Cir~X--1 (cf.~\citealt{2000MNRAS.317....1F}). Furthermore, the detection of type~I X--ray bursts from Cir~X--1
84: suggests that the neutron star has a low magnetic field, since the thermonuclear instability giving rise to type
85: I X--ray bursts is suppressed by dipole magnetic fields $\approxgt 10^{12}$ Gauss (\citealt{1980ApJ...238..287J};
86: \citealt{1995ApJ...438..852B}). In accordance with this, type I X--ray bursts are not found in HMXBs. The low
87: magnetic field of the neutron star in Cir~X--1 suggests that it is old and hence the companion star is not an
88: early type star.
89:
90: So far, clear spectroscopic evidence on the nature of the companion star is lacking. It has been suggested
91: that Cir~X--1 has a supergiant companion (\citealt{1980A&A....87..292M}). However, it can be argued
92: (cf.~\citealt{1999MNRAS.308..415J}) that the source is too faint in the optical bands for the distance of
93: 8--10 kpc that has been derived from the type~I X--ray bursts (\citealt{2004MNRAS.354..355J}). Nevertheless,
94: \citet{2003A&A...400..655C} found emission lines in the near--infrared spectrum that are consistent with a
95: mid--B super giant. However, these authors concluded that these features must have arisen in the accretion
96: disc and/or in outflows from the disc.
97:
98: In this Manuscript we present phase resolved Very Large Telescope (VLT) photometric and spectroscopic observations
99: obtained with the FOcal Reducer/low dispersion Spectrograph 2 (FORS2).
100:
101:
102: \section{Observations, analysis and results}
103:
104: We have obtained VLT/FORS2 spectra of the peculiar X--ray binary Cir~X--1 using the 1028z holographic grism
105: with a slit width of 1\arcsec. The observations were obtained in service mode on 21 different nights in the
106: period ranging March 15--May 15, 2005 (MJD 53446--53507). In order to sample the $\approx$16.6 day long
107: binary orbital period of Cir~X--1 we obtained one spectrum per night with an exposure time of 1730 seconds.
108: Exceptions were May 13 and April 4. On May 13 we obtained 9 spectra with an exposure time of 1675 seconds
109: since that night Cir~X--1 was close to periastron (according to the ephemeris of
110: \citealt{2004MNRAS.348..458C}). On April 4, three spectra were obtained since the seeing was higher than the
111: specified conditions on 2 of the 3 occasions. To minimise the light coming from an unrelated nearby field
112: star (star 2 in \citealt{1992A&A...260L...7M}), observations were obtained under good seeing conditions
113: (seeing between 0\farcs4 and 1\farcs3~as measured from the point spread function
114: full--width--at--half--maximum of the acquisition images). The dispersion was 0.86~\AA~per pixel. With the
115: 1\arcsec~slit width the resolution is about 120 km s$^{-1}$ at 8800~\AA.
116:
117: The spectra have been reduced with \textsc {iraf}\footnote{\textsc {iraf} is distributed by the National
118: Optical Astronomy Observatories}. We used the overscan area of the Charge Coupled Device (CCD) for bias
119: subtraction. The data were flatfield corrected and optimally extracted (\citealt{1986PASP...98..609H}).
120: Wavelength calibration was done using lines from He, Ar \& Ne lamp spectra that were obtained during
121: daytime the day after the observations with the same instrument set--up, as is customary for VLT Service
122: mode observations. The rms scatter of the wavelength calibration was in the range of 0.05--0.08~\AA.
123:
124: The extracted spectra were further reduced and analysed using the software package \textsc {molly}. We
125: corrected the wavelength calibration for potential shifts caused by flexure by cross correlating the spectra
126: over the wavelength ranges 9050--9160~\AA~and 9300--9500~\AA~with the first object spectrum. That part of the
127: spectrum is dominated by features from the night sky which should have the same wavelength in each spectrum.
128: Note that this does not account for uncertainties in the wavelength calibration caused by centroiding or star
129: tracking inaccuracies. We did this for both the spectra of Cir~X--1 and that of star 2 that was also in the
130: slit. Next, the observation times were corrected to the Heliocentric Julian Date time frame (using UTC times)
131: and we normalised and rebinned the spectra to a uniform velocity scale removing the Earth's velocity.
132: Cross--correlation of the spectra of star 2 with the first spectrum of that star over the range
133: 8000--9000~\AA~shows that the remaining rms velocity differences are 2.5 km s$^{-1}$.
134:
135: In Fig.~\ref{fig:sample} we show the normalised spectra as a function of the orbital period. To fold the spectra
136: we used our best--fit epoch of periastron (T, see below) and an orbital period of ${\rm P_{orb}=16.54}$ days as
137: found by extrapolating the X--ray dip ephemeris of \citet{2004MNRAS.348..458C}. We binned the data in 15 phase
138: bins. However, none of the spectra fall in the phase range 0.60--0.85, hence only 11 phase bins are shown. The
139: most striking feature is the large change in the profile of the Paschen lines. At phase zero, large, broad
140: (possibly double peaked) emission lines are present. Superposed is an absorption line spectrum that could
141: originate in the companion star. This absorption line spectrum can be seen clearly at phases 0.2--0.6. Besides the
142: Paschen absorption lines that are often seen in B and A stars, small absorption lines often seen in supergiants
143: can be seen as well, e.g.~near 8777~\AA~(\citealt{1999A&AS..137..521M}; \citealt{2005A&A...434..949C}; compare the
144: Cir~X--1 spectrum with that of the supergiant system XTE~J1739--302 in \citealt{2006ApJ...638..982N}). This
145: feature is most likely due to He~I.
146:
147: We investigated the behaviour of the equivalent width of the luminosity indicator Paschen--12 at 8750~\AA~as
148: a function of the orbital phase. At phase 0.4--0.6 the equivalent width of the absorption line is largest.
149: Assuming that the Paschen lines are formed in the companion star, the contamination of the accretion disc at
150: phase 0.4--0.6 is thus minimal but not necessarily zero. Hence, the measured equivalent width of 2~\AA~at
151: phase 0.4--0.6 is a lower limit. Using \citet{1994PASP..106..382D}, \citet{1995AJ....109.1379A},
152: \citet{1992ApJS...81..865S} and \citet{1999A&AS..137..521M} we find that the spectra at phase 0.4--0.6
153: resemble most that of late B/early A stars. As can be seen in Fig.~\ref{fig:comp}, the lines are very narrow
154: and from this we conclude that the star must be a supergiant. However, we cannot exclude that the absorption
155: line spectrum is from an accretion disc. Nevertheless, in order to test what spectral type would describe the
156: observed spectrum best under the assumption that the absorption lines are stellar, we obtained template
157: stellar spectra from \citet{2003A&A...402..433L}. We optimally subtracted the spectra of 9 stars (see
158: Table~\ref{tab:ref}) with spectral type ranging from B1Ib--F5Ia from the Cir~X--1 $\approx
159: 8250-9300$~\AA~spectrum at phase 0.4--0.6. The optimal subtraction is performed on normalised spectra,
160: minimising the residuals using the following recipe: ${\rm f^{CirX-1}_\lambda=A+bf^{temp}_\lambda}$, where A
161: is the assumed accretion disc contribution and b is the fraction of light from the companion star. From the
162: residuals of the optimal subtraction it was clear that the Ia supergiants matched the narrow absorption lines
163: better than those of the Ib stars. Between the Ia supergiants we found that the spectra of those of spectral
164: type mid--B provided the best fit. The multiplicative factor in the optimal subtraction was consistent with
165: unity. However, systematic uncertainties prevent us from obtaining formal (reduced chi--squared,
166: $\chi_\nu^2$) fits. Since the template star spectra and the Cir~X--1 spectrum are not obtained with the same
167: telescope--instrument combination the instrumental broadening of the line profile is different. Furthermore,
168: the broadening due to the rotational velocity of the stars is different for the different stars.
169:
170: \begin{table}
171: \caption{Template stars from \citet{2003A&A...402..433L}.}
172: \label{tab:ref}
173: \begin{center}
174: Spectral type \& Henry Draper Identification \\
175: \begin{tabular}{cccc}
176: \hline
177: B1Ib & HD~091316 & B2Ia & HD~268623 \\
178: B5Ib & HD~164353 & B3Ia & HD~271163 \\
179: B7Iab& HD~268749 & B9Ia & HD~032034 \\
180: A0Ib & HD~087737 & A3Ia & HD~033579 \\
181: -- & -- & F5Ia & HD~269697 \\
182: \end{tabular}
183: \end{center}
184: \end{table}
185:
186:
187: \begin{figure} \includegraphics[angle=0,width=8cm,clip]{in15phasebinsavern.ps} \caption{VLT/FORS2 spectra of
188: Cir~X--1 ranging from 8300--8900~\AA~for the different orbital phases as indicated on the right hand side in
189: the plot. The Paschen lines are easily discernible. They change from absorption to strong emission lines
190: (with an absorption core) near phase zero. An arbitrary offset has been applied to the spectra for display
191: purposes. The X--es mark regions where cosmic ray hits have been incompletely corrected for. The absorption
192: line spectrum likely reveals the companion star. The line at 8619.5~\AA~is due to a diffuse interstellar
193: absorption band (DIB). The feature near 8650~\AA~is also due to DIB although an He~I line at 8648~\AA~found
194: in B stars may contribute to the broad nature of the line. }
195:
196: \label{fig:sample}
197: \end{figure}
198: \begin{figure} \includegraphics[angle=0,width=8cm,clip]{CirX-1_A0III-HD77350_A0Ia-HD033579_A0Ib-HD87737.ps}
199:
200: \caption{From {\it bottom} to {\it top}: Spectra ranging from 8300--8900~\AA~of Cir~X--1 (at $\sim$phase
201: 0.5; this work), the template star HD~33579 (spectral type A3Ia), HD~87737 (spectral type A0Ib) and
202: HD~77350 (spectral type A0III; the latter three spectra are from \citealt{2003A&A...402..433L}). If the
203: absorption line spectrum is from the companion star of Cir~X--1, the narrow Paschen lines show that the
204: companion star must be a supergiant. The X--es mark regions where cosmic ray hits have been incompletely
205: corrected for. }
206:
207: \label{fig:comp}
208: \end{figure}
209:
210: Next, we investigated the velocity of the absorption line spectrum as a function of the binary orbital
211: phase. Since we did not obtain early type supergiant template star spectra we cross correlate the spectra of
212: Cir~X--1 with that of one of the spectra. In the cross correlation it is vital that the continuum level has
213: been determined accurately. Furthermore, one has to avoid regions of the spectrum affected by skylines. In
214: order to avoid problems with the continuum normalisation and skylines we have cross--correlated the region
215: between 8700--8900~\AA. In Fig.~\ref{fig:radvel} we show the observed radial velocity curve. The solid line
216: in Fig.~\ref{fig:radvel} is the best--fit elliptical orbit (see Table~\ref{tab:radvel} and ~\ref{tab:ell}).
217: However, in the fit shown in Fig.~\ref{fig:radvel} we have increased the errors by 6.5 km s$^{-1}$ in order
218: to obtain a $\chi_\nu^2$ of 1.0. As mentioned above the rms residual velocities measured by cross
219: correlating the spectra of the nearby star with itself are 2.5 km s$^{-1}$. We used this 2.5 km s$^{-1}$ as
220: a measure of the amplitude of systematic effects. This systematic error dominates the statistical error of
221: the cross--correlation of the Cir~X--1 spectra. However, even when we take this uncertainty into account,
222: the formal $\chi^2_\nu\approx 6$ for the 25 degrees of freedom. Possible other systematic effects that are
223: not included are: (i) the effects of differences in X--ray heating on the absorption line spectrum. E.g.~it
224: can be seen comparing the {\it top and bottom left panel} of Fig.~\ref{fig:radvel} that the first three
225: velocity measurements are somewhat below the fit, these observations were performed when the X--ray flux was
226: still relatively high (ii) from Fig.~\ref{fig:sample} it can be seen that near phase zero strong Paschen
227: emission lines are present. These emission lines are redshifted with respect to their rest wavelengths and
228: since the Paschen absorption lines at those orbital phases are red shifted by a different amount, the
229: emission lines fill--in the absorption in an asymmetric way, possibly skewing the cross correlation
230: velocities (iii) slit centroiding and tracking errors can have introduced errors in the wavelength
231: calibration.
232:
233:
234: \begin{table}
235: \caption{Relative radial velocities of the counterpart of Cir~X--1 as displayed in Fig.~\ref{fig:radvel}a. To
236: convert these velocities to absolute velocities add the systemic velocity of $\sim -26\pm 3$ km s$^{-1}$.}
237: \label{tab:radvel}
238: \begin{center}
239: \begin{tabular}{cccc}
240: \hline
241: MJD & Orbital phase & Orbital phase & Radial velocity \\
242: & ($P_{orb}=16.53$ d) & ($P_{orb}=16.68$ d) & (km s$^{-1}$)$^a$ \\
243: 53445.3868 & 0.323 & 0.326 & -8.2$\pm$1.0\\
244: 53447.3749 & 0.443 & 0.446 & -18.1$\pm$1.0\\
245: 53448.2275 & 0.495 & 0.497 & -17.7$\pm$1.1\\
246: 53455.3966 & 0.929 & 0.926 & 11.1$\pm$0.8\\
247: 53456.3680 & 0.988 & 0.985 & 35.3$\pm$1.0\\
248: 53457.4003 & 0.050 & 0.047 & 23.4$\pm$1.0\\
249: 53459.3877 & 0.171 & 0.166 & 0.4$\pm$1.2\\
250: 53461.1724 & 0.279 & 0.273 & -8.9$\pm$1.1\\
251: 53465.3265 & 0.530 & 0.522 & -9.5$\pm$1.3\\
252: 53465.3517 & 0.531 & 0.523 & -6.9$\pm$1.4\\
253: 53465.3763 & 0.533 & 0.525 & -7.3$\pm$1.1\\
254: 53466.1605 & 0.580 & 0.572 & -12.4$\pm$1.1\\
255: 53471.2432 & 0.888 & 0.876 & 22.0$\pm$0.8\\
256: 53472.2192 & 0.947 & 0.935 & 30.4$\pm$0.8\\
257: 53473.2753 & 0.011 & 0.998 & 49.7$\pm$1.1\\
258: 53491.1779 & 0.094 & 0.072 & 27.2$\pm$1.0\\
259: 53492.2715 & 0.160 & 0.137 & 8.1$\pm$1.1\\
260: 53493.2927 & 0.222 & 0.198 & 11.5$\pm$1.6\\
261: 53495.2467 & 0.340 & 0.316 & -3.9$\pm$0.9\\
262: 53496.2661 & 0.402 & 0.377 & -4.3$\pm$1.0\\
263: 53498.3358 & 0.527 & 0.501 & -5.7$\pm$1.2\\
264: 53504.0955 & 0.875 & 0.846 & 1.5$\pm$0.9\\
265: 53504.1153 & 0.877 & 0.847 & 2.4$\pm$0.8\\
266: 53504.1351 & 0.878 & 0.848 & 3.5$\pm$0.9\\
267: 53504.1549 & 0.879 & 0.850 & 6.1$\pm$0.8\\
268: 53504.1748 & 0.880 & 0.851 & 7.1$\pm$0.8\\
269: 53504.1946 & 0.881 & 0.852 & 9.8$\pm$0.9\\
270: 53504.2144 & 0.883 & 0.853 & 14.9$\pm$0.9\\
271: 53504.2343 & 0.884 & 0.854 & 14.7$\pm$1.0\\
272: 53504.2541 & 0.885 & 0.856 & 17.5$\pm$1.0\\
273: 53506.3040 & 0.009 & 0.978 & 38.0$\pm$0.8\\
274: \end{tabular}
275: \end{center}
276: {\footnotesize$^a$ The error has to be increased by 6.5 to obtain a reduced $\chi^2$ of $\approx$1 in
277: the fit (see text).}\\
278: \end{table}
279:
280:
281: Since we cross correlated the spectra with a spectrum of Cir~X--1 itself the systemic velocity that is derived
282: from the elliptical orbit fit (3.7$\pm$1.3 km s$^{-1}$) is not the true systemic velocity. In fact, as one
283: would expect, within 3~$\sigma$ it is consistent with 0. The systemic velocity is $\sim -26\pm 3$ km s$^{-1}$.
284: We derived this by comparing the best--fit central wavelength of the Gaussian absorption lines with rest
285: wavelengths of 8467.25, 8598.4, 8750.5 and 8862.8~\AA. The Gaussians were fitted to the spectrum that was taken
286: as our template spectrum in the cross--correlation (we have excluded the three Paschen lines that are possibly
287: blended with absorption lines of the Ca~II triplet). Hence, in order to convert the relative velocities in
288: Fig.~\ref{fig:radvel} and Table~\ref{tab:radvel} to absolute velocities one has to add the systemic velocity.
289:
290: \begin{figure*} \includegraphics[angle=-90,width=13cm,clip]{orbfit.ps}
291:
292: \caption{{\it Left panels:} (a) The radial velocity curve derived by cross--correlating the Cir~X--1 spectra
293: over the range 8700--8900~\AA~(note that the velocities are relative to that of the template spectrum).
294: Overplotted are the best--fit elliptical orbit treating the orbital period as a free parameter (light grey
295: line) and that using the predicted orbital period from the X--ray dip ephemeris of Clarkson et al.~(2004;
296: dark grey line, P$_{orb}=$16.53 days; Note that our best--fit orbital period is consistent with that derived
297: from the ephemeris of Clarkson et al.~2004). (b) The evolution of the $I$--band magnitude as derived from
298: psf--fitting on the $I$--band acquisition images. We have plotted the relative errors in the magnitudes, in
299: addition there is the $\sim 0.2$ magnitude uncertainty in the zeropoint (i.e.~in the ordinate) (c) The
300: contemporaneous X--ray flux as observed by the All Sky Monitor (ASM) on board the {\it Rossi} X--ray Timing
301: Explorer. The points with horizontal error bars at an ASM count rate of 40 counts s$^{-1}$ indicate the
302: predicted time of the X--ray dip according to the X--ray dip ephemeris of Clarkson et al.~(2004). {\it Right
303: panels:} (d) The radial velocity curve folded on the X--ray dip orbital period of 16.54 days using the T
304: from our best--fit to the radial velocity curve. (e) The phase folded $I$ band magnitude. (f) The phase
305: folded but unbinned ASM lightcurve using data 10 cycles before and after MJD 53475 (dwell data from
306: MJD~53140--53807). The X--ray dip phase and T from the radial velocity curve are consistent at the
307: 2~$\sigma$ level.}
308:
309: \label{fig:radvel}
310: \end{figure*}
311:
312:
313: \begin{table} \caption{The best--fit orbital parameters. We provide both the solution for the fit with the
314: orbital period as a free fit--parameter and that for where it was fixed to the orbital period extrapolated
315: from the ephemeris of \citet{2004MNRAS.348..458C}. }
316:
317: \label{tab:ell}
318: \begin{center}
319: \begin{tabular}{lcc}
320: \hline
321: P (days)& 16.68$\pm$0.15 & 16.53 (fixed)\\
322: K (km s$^{-1}$) & 25$\pm$2& 25$\pm$2\\
323: e & 0.45$\pm$0.07& 0.47$\pm$0.06\\
324: $\omega$ (deg) & 2$\pm$12 & -7$\pm$11\\
325: T (+53,475.0 days; MJD) & -1.7$\pm0.4$& -1.9$\pm0.4$\\
326: a $\sin$ {\it i} (lightseconds) & 16.9$\pm$1.2 & 16.8$\pm$2.0\\
327: f(m) (M$_\odot$) & 0.019$\pm$0.007 & 0.019$\pm$0.007\\
328:
329: \end{tabular}
330: \end{center}
331: %{\footnotesize$^c$ NA = Not applicable.}\\
332: \end{table}
333:
334: We have used the 32 $I$--band images that were obtained to acquire the source to construct a lightcurve.
335: The images were corrected for bias using the values from the overscan regions and flatfielded using sky
336: flats taken within one or two days from the science images. To determine instrumental magnitudes of
337: Cir~X--1 and stars in its relative crowded vicinity, we used the point--spread--function (psf) fitting
338: routines from \textsc {DAOPHOT~II} (\citealt{1987PASP...99..191S}), running inside \textsc {midas}. The
339: instrumental magnitudes were placed in a common photometric system by removing small magnitude differences
340: between the different images, due to e.g.\ differences in the psf and the exposure times, by matching stars
341: between these images and comparing their magnitudes. This common system was then calibrated using
342: observations of standard star fields (PG\,0942$-$029 and L110; using the calibrated magnitudes by
343: \citealt{2000PASP..112..925S}), which were imaged during photometric nights (March 27, April 11 and 12,
344: 2005). As Cir~X--1 is only imaged in the $I$--band, we have not determined colour terms. The uncertainty in
345: the zeropoint is estimated to be about 0.2\,mag. The derived evolution of the $I$--band magnitude as a
346: function of time and orbital phase is shown in the {\it middle panels} of Fig.~\ref{fig:radvel}. It is
347: clear that the $I$--band magnitude is constant except for a brief period that coincides with the X--ray dip
348: and periastron passage.
349:
350: \section{Discussion}
351:
352: We have obtained VLT phase resolved optical $I$--band spectra and images of the X--ray binary Cir~X--1. The
353: observed X--ray flux was low compared to that observed during other earlier spectroscopic observations
354: (e.g.~\citealt{1999MNRAS.308..415J}; \citealt{2001MNRAS.328.1193J}; \citealt{2003A&A...400..655C}). If this lower
355: X--ray flux corresponds to a lower intrinsic X--ray luminosity the optical light from the accretion disc thought to
356: be caused in large by reprocessing of the X--ray luminosity might be lower as well. Hence, the putative companion
357: star might be more readily observable now compared with high X--ray flux episodes. The spectrum we observed varies
358: strongly as a function of orbital phase. Strong emission lines appear near periastron passage, filling--in the
359: absorption line spectrum observable at other orbital phases. Cross--correlation of the absorption features in the
360: 8700--8900~\AA~range gives a radial velocity curve with an orbital eccentricity of $e=0.45$, $a\,\sin\,i=16.9$
361: lightseconds and a mass function of 0.019 M$_\odot$. The question is; is this the radial velocity curve of the
362: companion star, is it associated with the compact object/neutron star, or are the absorption lines caused by a
363: circumbinary disc?
364:
365: The detected absorption line spectrum is consistent with a stellar spectrum if that star is a B5--A0 supergiant. This
366: provides evidence for a supergiant companion star in Cir~X--1. Similarly, \citet{2003A&A...400..655C} found emission
367: lines in the near--infrared spectrum of Cir~X--1 that often occur in mid--B supergiants. Those results are consistent
368: with our findings (but \citealt{2003A&A...400..655C} favoured a scenario where the emission features arise in the
369: accretion disc/flow.) Previously, \citet{1980A&A....87..292M} proposed that the companion star of Cir~X--1 is an early
370: type supergiant. However, the then known counterpart was later resolved into three stars reducing the magnitude
371: associated with the companion star (\citealt{1992A&A...260L...7M}) which led people to discard the supergiant model for
372: Cir~X--1. On the other hand, an interstellar extinction of A$_V=5$ was used. This value was derived on
373: the basis of the lowest N$_H$ measured in the seventies. That N$_H$ was determined from an X--ray spectrum obtained from
374: Aerobee rocket data. No error was given on N$_H$ (\citealt{1971ApJ...169L..23M}). More recent N$_H$ measurements from
375: X--ray spectral fits with more sensitive satellites with a good soft response such as ROSAT, ASCA and {\it Chandra} give
376: an N$_H$ in the range 1.6--2.2$\times10^{22}$ cm$^{-2}$ (\citealt{1995A&A...293..889P}; \citealt{1996MNRAS.283.1071B};
377: \citealt{1999ApJ...511..304S}; \citealt{2002ApJ...572..971S}). However, \citet{2005ApJ...619..503I} again found a
378: significant lower value for N$_H$ modelling {\it Beppo}--SAX data. As we will show below there are three independant
379: measurements that are consistent with the higher N$_H$ for Cir~X--1. First of all, as mentioned by
380: \citet{1996MNRAS.283.1071B}, a significantly lower value for N$_H$ would not be compatible with the strength of the dust
381: scattering halo in Cir~X--1 as found by \citet{1995A&A...293..889P}. Secondly, the distance that would be derived from
382: the low N$_H$ value is $\sim$4 kpc (\citealt{2005ApJ...619..503I}). This distance is too low to explain the type I X--ray
383: bursts, those require a lower limit to the distance of d=7.8--10.5 kpc (\citealt{2004MNRAS.354..355J}). Finally, the
384: equivalent width of the diffuse interstellar band at 8620~\AA~that we find in the optical spectrum is correlated with
385: $E_{(B-V)}$ (\citealt{2000msl..work..179M}). This provides an A$_V=12.2$ for Cir~X--1 (taking R=3.1), consistent with the
386: higher value derived from N$_H$ which using the conversion of N$_H$ to A$_V$ from \citet{1995A&A...293..889P} gives
387: $9<{\rm A}_V<12$. For these reasons and since the modelling of the X--ray spectrum is model dependent as mentioned in
388: \citet{2005ApJ...619..503I} we think that the higher value for N$_H$ and hence A$_V$ is more likely to be right.
389:
390: From the distance to Cir~X--1 from the type I X--ray bursts (d=7.8--10.5 kpc), A$_V\approx 9-12$, and $I$=17.6, one gets an
391: absolute magnitude $-4.9<{\rm M}_V<-2.5$ (taking into account that $V-I$ is nearly 0 for mid--B/early A type stars, and
392: using the relative extinction from \citealt{scfida1998} A$_I=0.6\times$A$_V$). An absolute magnitude of -2.5 would be
393: too low and even -4.9 is on the low side for standard B5--A0 supergiants (\citealt{1974MNRAS.166..203B}). However, considering the preceding binary
394: evolution involving significant mass transfer from the neutron star progenitor to the initially less massive companion
395: star and the current X--ray heating especially at periastron the companion star of Cir~X--1 is not likely to be a
396: standard supergiant.
397:
398:
399: Let us investigate the consequences of the assumption that the absorption line spectrum and the radial velocity curve
400: track the companion star of Cir~X--1. Kepler's third law, the measured a$_{opt}\, \sin$ {\it i} and P$_{orb}$ define
401: a relation between the inclination of the orbit with respect to the line--of--sight and the mass of the companion
402: star (the solid lines in Fig.~\ref{fig:massinc}). From the spectral classification B5--A0I we find $2.1\approxlt
403: \log\,g\, \approxlt 2.4$ (\citealt{2000asqu.book.....C}). This gives
404: $\frac{R_{comp}}{R_\odot}=C\sqrt\frac{M_{comp}}{M_\odot}$ (where $C$ is in the range of 10--15). To avoid the neutron
405: star going through the companion star at periastron, we constrain the radius of the companion star to be smaller than
406: or equal to the periastron distance $a\,(1-e)$, with $a\,\sin\,i={\rm \frac{M_X+M_2}{M_X}} a_{opt}\,\sin\,i$ and
407: $a_{opt}\,\sin\,i =7.28$ R$_\odot$. From Fig.~\ref{fig:massinc} this yields ${\rm M_{comp} \approxlt 18.9\,M_\odot}$,
408: $\approxlt 15\,{\rm M_\odot}$, $\approxlt 11.9\,{\rm M_\odot},$ and $\approxlt 10.4\,{\rm M_\odot}$ for
409: M$_{comp}=10$, 5, 2 and 1.4 M$_\odot$, respectively. The inclination is constrained to be $ 7.1^\circ\approxlt i
410: \approxlt 14.9^\circ$, $8.9^\circ\approxlt i \approxlt 23.5^\circ$, $12.1^\circ\approxlt i \approxlt 50^\circ$ and $i
411: \approxgt 13.7^\circ$ for a 10, 5, 2, and 1.4~M$_\odot$ compact object, respectively. A normal B5--A0 supergiant has
412: a mass of $\sim$10 M$_\odot$ which would hence fit--in with a neutron star compact object of Cir~X--1. On the other
413: hand it is conceivable that if the companion star of Cir~X--1 has indeed spectral type B5--A0I it is no ordinary star
414: due to the preceding evolution (as mentioned above).
415:
416: The observed type~I X--ray bursts (\citealt{1986MNRAS.219..871T}) are evidence of a neutron star nature of the
417: compact object. Furthermore, as mentioned in the Introduction the neutron star has to have a magnetic field
418: $\approxlt 10^{12}$ Gauss. Also the presence of a strong radio jet is not compatible with a high magnetic field
419: neutron star. These findings seem to be at odds with a supergiant (thus young) companion star. We conclude that if
420: the presence of a supergiant companion star in Cir~X--1 is confirmed the neutron star magnetic field either
421: decayed quickly to below $\approxlt 10^{12}$ Gauss or the neutron star was born with such a low field.
422:
423: Assuming that a neutron star kick at birth did not give the binary system a large systemic velocity, one can
424: calculate the radial velocity for Local Standards of Rest along the direction of Cir~X--1 (see for instance
425: \citealt{2001ApJ...555..364B}) to find that the systemic radial velocity of $\sim -26\pm 3$ km s$^{-1}$ at the
426: location of Cir~X-1 gives a distance of either $\sim 1.6$ kpc or $\sim$11.8 kpc. A distance of 11.8 kpc is close to
427: the distance derived from the observed type I X--ray bursts (\citealt{1986MNRAS.221P..27T}; 7.8--10.5 kpc
428: \citealt{2004MNRAS.354..355J}).
429:
430: The optical I--band lightcurve (see Fig.~\ref{fig:radvel}e) shows a clear brightening near periastron. This is similar in shape to
431: the lightcurves published by \citet{1992A&A...260L...7M} and \citet{1994MNRAS.268..742G}, although one has to bear in mind that
432: those lightcurves were obtained when the observed X--ray luminosity of the source was much higher and that the data was phase folded
433: using different ephemerides. The increase in I--band light corresponds to the phase where Paschen emission lines start to become
434: apparent in the spectrum. Hence, it is probable that the enhanced mass transfer rate near periastron is responsible for both
435: effects. For instance, the emission lines could be formed in the accretion stream from the companion star to the compact object and
436: the enhanced I--band emission could be both due to these emission lines as well as due to enhanced continuum emission related to the
437: stream impact site.
438:
439: \begin{figure} \includegraphics[angle=0,width=7cm,clip]{massincl.ps}
440:
441: \caption{The mass of the companion star as a function of the orbital inclination with respect to the
442: line--of--sight. The drawn lines are derived assuming that the compact object is from left to right, a 10, 5, 2,
443: or a 1.4 M$_\odot$ compact object. The dots are solutions if imposing the radius of the companion star to be
444: smaller than or equal to the periastron distance $a\,(1-e)$ (we took the least constraining case of B5I, the
445: radius of a A0I is somewhat larger hence the dots would all move to lower inclinations). For each of the solid
446: line -- dot combination the allowed parameter space is that on the solid line and below the dot. This yields the
447: following constraints on the companion star mass for a 10, 5, 2 and 1.4 M$_\odot$ compact object; ${\rm M_{comp}
448: \approxlt 18.9\,M_\odot}$, $\approxlt 15\,{\rm M_\odot}$, $\approxlt 11.9\,{\rm M_\odot},$ and $\approxlt
449: 10.4\,{\rm M_\odot}$, respectively. The inclination is constrained to be $ 7.1^\circ\approxlt i \approxlt
450: 14.9^\circ$, $8.9^\circ\approxlt i \approxlt 23.5^\circ$, $12.1^\circ\approxlt i \approxlt 50^\circ$ and $i
451: \approxgt 13.7^\circ$ for a 10, 5, 2, and 1.4~M$_\odot$ compact object, respectively. Taking an AOI star instead
452: of a B5I would change the inclination and companion star mass constraint to $ 7.1^\circ\approxlt i \approxlt
453: 10^\circ, 7.3$ M$_\odot$ for the 10 M$_\odot$ compact object. Under the assumption that the measured $a\, \sin\,i$
454: is not $a_{opt}\, \sin\, i$ but $a_{NS}\, \sin\, i$, the constraint on the companion star mass as a function of
455: inclination is given by the dotted line (taking ${\rm M_{NS}=1.4\,M_\odot}$).}
456:
457: \label{fig:massinc}
458: \end{figure}
459:
460: In principle it cannot be ruled out that the absorption line spectrum originates in the accretion disc.
461: E.g.~the $I$--band spectrum of the high inclination (accretion disc corona) source 2S~0921--630 shows Paschen
462: absorption lines at orbital phases near 0.9 that could well be caused by the line of sight passing through the
463: accretion disc rim (\citealt{2005MNRAS.356..621J}). Those Paschen absorption lines are not present in
464: the K1III spectral type that has been derived for the companion star in 2S~0921--630 (although X--ray heating
465: effects may result in different spectral types being observed at different orbital phases, cf.~the observed
466: spectral type in Cyg~X--2 changes from A5--F2 \citealt{1979ApJ...231..539C}). However, in radio observations
467: of Cir~X--1 superluminal motion has been observed which limits the inclination of the jet axis to the line of
468: sight to $i<5^\circ$ (\citealt{2004Natur.427..222F}; the limit $i<5^\circ$ has been derived assuming a
469: distance of 6.5 kpc, if Cir~X--1 is indeed further away as is indicated by the burst properties, then the
470: limit on $i$ is more stringent still). If our line--of--sight also goes through the accretion disc rim (at all
471: orbital phases) it implies that the jet--axis is nearly in the plane of the orbit. Hence, it implies that the
472: jet ploughs through the accretion disc if it originates close to the compact object.
473:
474: Nevertheless, if we assume that the $a\, \sin\,i$ we measured is associated with the binary motion of the
475: accretion disc/neutron star we can derive a limit on the mass of the companion star assuming the neutron
476: star mass is 1.4 M$_\odot$. This limit is a function of the binary inclination and is given by the dotted
477: line in Fig.~\ref{fig:massinc}. From the discussion above it would seem that the companion star mass is
478: $\approx$0.4 M$_\odot$ since the binary inclination must be high in this scenario. A star of such a mass
479: cannot have evolved off the main--sequence in a Hubble time unless the star was more massive initially (at
480: least 0.8 M$_\odot$) and more than 0.4 M$_\odot$ has been transfered already. However, it is unclear whether
481: such an amount of mass can have been transfered while the orbital eccentricity is still 0.45.
482:
483: Finally, in this paragraph we consider other possible formation scenarios for the absorption line spectrum.
484: E.g.~in the peculiar X--ray binary SS~433 evidence for a circumbinary disc has been found
485: (\citealt{2001ApJ...562L..79B}), perhaps the absorption lines are formed in such a disc. However,
486: \citet{1988AJ.....96..242F} have obtained I--band spectra of SS~433 and their spectra are markedly different
487: from those that we observe from Cir~X--1. In SS~433 the Paschen lines are double peaked and in emission
488: whereas we find them to be in absorption except near periastron passage. On the other hand the inclinations
489: at which we observe SS~433 and Cir~X--1 are different. In this respect it is interesting to note that
490: according to \citet{2006astro.ph..7612T} we will be observing the system through the material that is swept
491: up by the approaching radio jet. However, it is difficult to imagine that that material is dense enough and
492: has a velocity low enough to cause the narrow absorption lines. Finally, in all these scenario's the
493: velocity changes of the absorption lines with orbital phase are difficult to explain.
494:
495: A potential way to determine whether the measured radial velocity curve is associated with the companion star is
496: to obtain a high resolution spectrum. Such a spectrum could reveal narrow spectral features and allow for a better
497: spectral classification e.g.~the observed absorption feature near 8685~\AA~consists of a blend of several narrow
498: lines if the spectrum is from the companion star (\citealt{1999A&AS..137..521M}) and if the companion star is an
499: A--star the Ca~II absorption triplet can be separated from the Paschen lines. Furthermore, the rotational
500: broadening of the spectral lines could be measured. Finally, due to precession of the orbit (i.e.~apsidal motion),
501: the anomalous period as measured by the time between periastron passages is different from the period that would
502: be measured by eclipse timing (i.e.~the sidereal period). Hence, if the X--ray dips in Cir~X--1 are due to
503: (grazing) eclipses the sidereal period will over time differ from the anomalous period.
504:
505:
506:
507: \section*{Acknowledgments}
508:
509: \noindent The authors are grateful to the referee, Dr.~D.~Gies, for his comments that helped improve the
510: Paper. We would like to thank the Director of ESO for approving these DDT observations. The use of the
511: spectral analysis software package \textsc {molly} written by prof.~Tom Marsh is acknowledged. PGJ
512: acknowledges support from NASA grants NNG05GN20G and NNG05GN27G. PGJ, GN and CGB acknowledge support from
513: the Netherlands Organisation for Scientific Research.
514:
515:
516: \begin{thebibliography}{46}
517: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
518:
519: \bibitem[{{Allen} \& {Strom}(1995)}]{1995AJ....109.1379A}
520: {Allen}, L.~E., {Strom}, K.~M., 1995, \aj, 109, 1379
521:
522: \bibitem[{{Balona} \& {Crampton}(1974)}]{1974MNRAS.166..203B}
523: {Balona}, L., {Crampton}, D., 1974, \mnras, 166, 203
524:
525: \bibitem[{{Berger} \& {Gies}(2001)}]{2001ApJ...555..364B}
526: {Berger}, D.~H., {Gies}, D.~R., 2001, \apj, 555, 364
527:
528: \bibitem[{{Bildsten}(1995)}]{1995ApJ...438..852B}
529: {Bildsten}, L., 1995, \apj, 438, 852
530:
531: \bibitem[{{Blundell} et~al.(2001){Blundell}, {Mioduszewski}, {Muxlow},
532: {Podsiadlowski}, \& {Rupen}}]{2001ApJ...562L..79B}
533: {Blundell}, K.~M., {Mioduszewski}, A.~J., {Muxlow}, T.~W.~B., {Podsiadlowski},
534: P., {Rupen}, M.~P., 2001, \apjl, 562, L79
535:
536: \bibitem[{{Brandt} et~al.(1996){Brandt}, {Fabian}, {Dotani}, {Nagase}, {Inoue},
537: {Kotani}, \& {Segawa}}]{1996MNRAS.283.1071B}
538: {Brandt}, W.~N., {Fabian}, A.~C., {Dotani}, T., {Nagase}, F., {Inoue}, H.,
539: {Kotani}, T., {Segawa}, Y., 1996, \mnras, 283, 1071
540:
541: \bibitem[{{Clark} et~al.(2003){Clark}, {Charles}, {Clarkson}, \&
542: {Coe}}]{2003A&A...400..655C}
543: {Clark}, J.~S., {Charles}, P.~A., {Clarkson}, W.~I., {Coe}, M.~J., 2003, \aap,
544: 400, 655
545:
546: \bibitem[{{Clark} et~al.(2005){Clark}, {Negueruela}, {Crowther}, \&
547: {Goodwin}}]{2005A&A...434..949C}
548: {Clark}, J.~S., {Negueruela}, I., {Crowther}, P.~A., {Goodwin}, S.~P., 2005,
549: \aap, 434, 949
550:
551: \bibitem[{{Clarkson} et~al.(2004){Clarkson}, {Charles}, \&
552: {Onyett}}]{2004MNRAS.348..458C}
553: {Clarkson}, W.~I., {Charles}, P.~A., {Onyett}, N., 2004, \mnras, 348, 458
554:
555: \bibitem[{{Cowley} et~al.(1979){Cowley}, {Crampton}, \&
556: {Hutchings}}]{1979ApJ...231..539C}
557: {Cowley}, A.~P., {Crampton}, D., {Hutchings}, J.~B., 1979, \apj, 231, 539
558:
559: \bibitem[{{Cox}(2000)}]{2000asqu.book.....C}
560: {Cox}, A.~N., 2000, {Allen's astrophysical quantities}, Allen's astrophysical
561: quantities, 4th ed.~ Publisher: New York: AIP Press; Springer, 2000.~ Editedy
562: by Arthur N.~Cox.~ ISBN: 0387987460
563:
564: \bibitem[{{Danks} \& {Dennefeld}(1994)}]{1994PASP..106..382D}
565: {Danks}, A.~C., {Dennefeld}, M., 1994, \pasp, 106, 382
566:
567: \bibitem[{{Fender} et~al.(1998){Fender}, {Spencer}, {Tzioumis}, {Wu}, {van der
568: Klis}, {van Paradijs}, \& {Johnston}}]{1998ApJ...506L.121F}
569: {Fender}, R., {Spencer}, R., {Tzioumis}, T., {Wu}, K., {van der Klis}, M., {van
570: Paradijs}, J., {Johnston}, H., 1998, \apjl, 506, L121
571:
572: \bibitem[{{Fender} et~al.(2004){Fender}, {Wu}, {Johnston}, {Tzioumis},
573: {Jonker}, {Spencer}, \& {van der Klis}}]{2004Natur.427..222F}
574: {Fender}, R., {Wu}, K., {Johnston}, H., {Tzioumis}, T., {Jonker}, P.,
575: {Spencer}, R., {van der Klis}, M., 2004, \nat, 427, 222
576:
577: \bibitem[{{Fender} \& {Hendry}(2000)}]{2000MNRAS.317....1F}
578: {Fender}, R.~P., {Hendry}, M.~A., 2000, \mnras, 317, 1
579:
580: \bibitem[{{Filippenko} et~al.(1988){Filippenko}, {Romani}, {Sargent}, \&
581: {Blandford}}]{1988AJ.....96..242F}
582: {Filippenko}, A.~V., {Romani}, R.~W., {Sargent}, W.~L.~W., {Blandford}, R.~D.,
583: 1988, \aj, 96, 242
584:
585: \bibitem[{{Glass}(1994)}]{1994MNRAS.268..742G}
586: {Glass}, I.~S., 1994, \mnras, 268, 742
587:
588: \bibitem[{{Horne}(1986)}]{1986PASP...98..609H}
589: {Horne}, K., 1986, \pasp, 98, 609
590:
591: \bibitem[{{Iaria} et~al.(2005){Iaria}, {Span{\`o}}, {Di Salvo}, {Robba},
592: {Burderi}, {Fender}, {van der Klis}, \& {Frontera}}]{2005ApJ...619..503I}
593: {Iaria}, R., {Span{\`o}}, M., {Di Salvo}, T., {Robba}, N.~R., {Burderi}, L.,
594: {Fender}, R., {van der Klis}, M., {Frontera}, F., 2005, \apj, 619, 503
595:
596: \bibitem[{{Johnston} et~al.(1999){Johnston}, {Fender}, \&
597: {Wu}}]{1999MNRAS.308..415J}
598: {Johnston}, H.~M., {Fender}, R., {Wu}, K., 1999, \mnras, 308, 415
599:
600: \bibitem[{{Johnston} et~al.(2001){Johnston}, {Wu}, {Fender}, \&
601: {Cullen}}]{2001MNRAS.328.1193J}
602: {Johnston}, H.~M., {Wu}, K., {Fender}, R., {Cullen}, J.~G., 2001, \mnras, 328,
603: 1193
604:
605: \bibitem[{{Jonker} \& {Nelemans}(2004)}]{2004MNRAS.354..355J}
606: {Jonker}, P.~G., {Nelemans}, G., 2004, \mnras, 354, 355
607:
608: \bibitem[{{Jonker} et~al.(2005){Jonker}, {Steeghs}, {Nelemans}, \& {van der
609: Klis}}]{2005MNRAS.356..621J}
610: {Jonker}, P.~G., {Steeghs}, D., {Nelemans}, G., {van der Klis}, M., 2005,
611: \mnras, 356, 621
612:
613: \bibitem[{{Joss} \& {Li}(1980)}]{1980ApJ...238..287J}
614: {Joss}, P.~C., {Li}, F.~K., 1980, \apj, 238, 287
615:
616: \bibitem[{{Kaluzienski} et~al.(1976){Kaluzienski}, {Holt}, {Boldt}, \&
617: {Serlemitsos}}]{1976ApJ...208L..71K}
618: {Kaluzienski}, L.~J., {Holt}, S.~S., {Boldt}, E.~A., {Serlemitsos}, P.~J.,
619: 1976, \apjl, 208, L71
620:
621: \bibitem[{{Le Borgne} et~al.(2003)}]{2003A&A...402..433L}
622: {Le Borgne}, J.-F., et~al., 2003, \aap, 402, 433
623:
624: \bibitem[{{Margon} et~al.(1971){Margon}, {Lampton}, {Bowyer}, \&
625: {Cruddace}}]{1971ApJ...169L..23M}
626: {Margon}, B., {Lampton}, M., {Bowyer}, S., {Cruddace}, R., 1971, \apjl, 169,
627: L23
628:
629: \bibitem[{{Moneti}(1992)}]{1992A&A...260L...7M}
630: {Moneti}, A., 1992, \aap, 260, L7
631:
632: \bibitem[{{Munari}(2000)}]{2000msl..work..179M}
633: {Munari}, U., 2000, in {Porceddu}, I., {Aiello}, S., eds., Molecules in Space
634: and in the Laboratory, p. 179
635:
636: \bibitem[{{Munari} \& {Tomasella}(1999)}]{1999A&AS..137..521M}
637: {Munari}, U., {Tomasella}, L., 1999, \aaps, 137, 521
638:
639: \bibitem[{{Murdin} et~al.(1980){Murdin}, {Jauncey}, {Lerche}, {Nicolson},
640: {Kaluzienski}, {Holt}, \& {Haynes}}]{1980A&A....87..292M}
641: {Murdin}, P., {Jauncey}, D.~L., {Lerche}, I., {Nicolson}, G.~D., {Kaluzienski},
642: L.~J., {Holt}, S.~S., {Haynes}, R.~F., 1980, \aap, 87, 292
643:
644: \bibitem[{{Negueruela} et~al.(2006){Negueruela}, {Smith}, {Harrison}, \&
645: {Torrej{\'o}n}}]{2006ApJ...638..982N}
646: {Negueruela}, I., {Smith}, D.~M., {Harrison}, T.~E., {Torrej{\'o}n}, J.~M.,
647: 2006, \apj, 638, 982
648:
649: \bibitem[{{Podsiadlowski} et~al.(2002){Podsiadlowski}, {Rappaport}, \&
650: {Pfahl}}]{2002ApJ...565.1107P}
651: {Podsiadlowski}, P., {Rappaport}, S., {Pfahl}, E.~D., 2002, \apj, 565, 1107
652:
653: \bibitem[{{Predehl} \& {Schmitt}(1995)}]{1995A&A...293..889P}
654: {Predehl}, P., {Schmitt}, J.~H.~M.~M., 1995, \aap, 293, 889
655:
656: \bibitem[{{Saz Parkinson} et~al.(2003)}]{2003HEAD....7.1723S}
657: {Saz Parkinson}, P.~M., et~al., 2003, AAS/High Energy Astrophysics Division, 7,
658: astroph0303402
659:
660: \bibitem[{{Schlegel} et~al.(1998){Schlegel}, {Finkbeiner}, \&
661: {Davis}}]{scfida1998}
662: {Schlegel}, D.~J., {Finkbeiner}, D.~P., {Davis}, M., 1998, \apj, 500, 525
663:
664: \bibitem[{{Schulz}(1999)}]{1999ApJ...511..304S}
665: {Schulz}, N.~S., 1999, \apj, 511, 304
666:
667: \bibitem[{{Schulz} \& {Brandt}(2002)}]{2002ApJ...572..971S}
668: {Schulz}, N.~S., {Brandt}, W.~N., 2002, \apj, 572, 971
669:
670: \bibitem[{{Silva} \& {Cornell}(1992)}]{1992ApJS...81..865S}
671: {Silva}, D.~R., {Cornell}, M.~E., 1992, \apjs, 81, 865
672:
673: \bibitem[{{Stetson}(1987)}]{1987PASP...99..191S}
674: {Stetson}, P.~B., 1987, \pasp, 99, 191
675:
676: \bibitem[{{Stetson}(2000)}]{2000PASP..112..925S}
677: {Stetson}, P.~B., 2000, \pasp, 112, 925
678:
679: \bibitem[{{Stewart} et~al.(1993){Stewart}, {Caswell}, {Haynes}, \&
680: {Nelson}}]{1993MNRAS.261..593S}
681: {Stewart}, R.~T., {Caswell}, J.~L., {Haynes}, R.~F., {Nelson}, G.~J., 1993,
682: \mnras, 261, 593
683:
684: \bibitem[{{Tauris} \& {Savonije}(1999)}]{1999A&A...350..928T}
685: {Tauris}, T.~M., {Savonije}, G.~J., 1999, \aap, 350, 928
686:
687: \bibitem[{{Tennant} et~al.(1986{\natexlab{a}}){Tennant}, {Fabian}, \&
688: {Shafer}}]{1986MNRAS.221P..27T}
689: {Tennant}, A.~F., {Fabian}, A.~C., {Shafer}, R.~A., 1986{\natexlab{a}}, \mnras,
690: 221, 27P
691:
692: \bibitem[{{Tennant} et~al.(1986{\natexlab{b}}){Tennant}, {Fabian}, \&
693: {Shafer}}]{1986MNRAS.219..871T}
694: {Tennant}, A.~F., {Fabian}, A.~C., {Shafer}, R.~A., 1986{\natexlab{b}}, \mnras,
695: 219, 871
696:
697: \bibitem[{{Tudose} et~al.(2006){Tudose}, {Fender}, {Kaiser}, {Tzioumis}, {van
698: der Klis}, \& {Spencer}}]{2006astro.ph..7612T}
699: {Tudose}, V., {Fender}, R.~P., {Kaiser}, C.~R., {Tzioumis}, A.~K., {van der
700: Klis}, M., {Spencer}, R., 2006, ArXiv Astrophysics e-prints
701:
702: \end{thebibliography}
703:
704: \end{document}
705:
706:
707:
708:
709:
710:
711:
712:
713: