1: %documentclass[12pt,preprint]{aastex}
2: \documentclass{article}
3: \usepackage{emulateapj}
4: \usepackage{float,epsfig,psfig}
5:
6: \newcommand{\LA}{\mbox{\raisebox{-0.6ex}{$\stackrel{\textstyle<}{\sim}$}}}
7: \newcommand{\GA}{\mbox{\raisebox{-0.6ex}{$\stackrel{\textstyle>}{\sim}$}}}%%%%%%
8: \newcommand{\cxo}{{\sl Chandra}}
9: \newcommand{\xmm}{{\sl XMM-Newton}}
10: \newcommand{\ros}{{\sl ROSAT}}
11: \newcommand{\ngc}{{NGC~4214}}
12: \newcommand{\msun}{M$_{\odot}$}
13: \newcommand{\ergl}{ergs~s$^{-1}$}
14: \newcommand{\ctss}{cts~s$^{-1}$}
15: \newcommand{\cxou}{CXOU~J121538.2+361921}\newcommand{\cxous}{CXOU~J121538}
16: \newcommand{\cX}{X-1}
17: \newcommand{\ergcms}{ergs~cm$^{-2}$~s$^{-1}$}
18: \newcommand{\ha}{H$\alpha$}
19: \newcommand{\hii}{H{\sc ii}}
20: \newcommand{\mdot}{$\dot{M}$}
21: \newcommand{\hst}{{\sl Hubble}}
22: \newcommand{\etal}{et al.}
23:
24: \slugcomment{Submitted to Astrophysical Journal}
25:
26: \begin{document}
27:
28: \title{Discovery of a 3.6-hr Eclipsing Luminous X-Ray Binary in the
29: Galaxy NGC 4214}
30:
31: \author{
32: Kajal K. Ghosh\altaffilmark{1},
33: Saul Rappaport\altaffilmark{2},
34: Allyn F. Tennant\altaffilmark{3},
35: Douglas A. Swartz\altaffilmark{1},
36: David Pooley\altaffilmark{4,5}, and
37: N. Madhusudhan\altaffilmark{2}
38: }
39: \altaffiltext{1}{Universities Space Research Association,
40: NASA Marshall Space Flight Center, VP62, Huntsville, AL, 35805}
41: \altaffiltext{2}{Dept. of Physics and Kavli
42: Institute for Astrophysics and Space Research,
43: Massachusetts Institute of Technology, 77
44: Massachusetts Ave., Cambridge, MA, 02139}
45: \altaffiltext{3}{Space Science Department,
46: NASA Marshall Space Flight Center, VP62, Huntsville, AL, 35805}
47: \altaffiltext{4}{Astronomy Department, University of California Berkeley,
48: 601 Campbell Hall, Berkeley, CA, 94720}
49: \altaffiltext{4}{\cxo\ Fellow}
50:
51: \begin{abstract}
52: We report the discovery of an eclipsing X-ray binary with a 3.62-hr
53: period within 24\arcsec\ of the center of the dwarf starburst galaxy
54: NGC 4214. The orbital period places interesting constraints on the
55: nature of the binary, and allows for a few very different interpretations.
56: The most likely possibility is that the source lies within NGC 4214
57: and has an X-ray luminosity, $L_x$, of up to $7 \times 10^{38}$~\ergl.
58: In this case the binary may well be comprised of a
59: naked He-burning donor star with a neutron-star accretor,
60: though a stellar-mass black-hole accretor cannot be completely excluded.
61: There is no obvious evidence for a strong stellar wind in the X-ray orbital
62: light curve that would be expected from a massive He star; thus, the
63: mass of the He star should be $\lesssim 3-4$~\msun.
64: If correct, this would represent a new class of very luminous
65: X-ray binary~--~perhaps related to Cyg X-3.
66: Other less likely possibilities include a conventional low-mass X-ray
67: binary that somehow manages to produce such a high X-ray luminosity
68: and is apparently persistent over an interval of years; or a
69: foreground AM Her binary of much lower luminosity that fortuitously
70: lies in the direction of NGC 4214.
71: Any model for this system must accommodate the lack of an
72: optical counterpart down to a limiting magnitude of 22.6 in the visible.
73:
74: \end{abstract}
75:
76: \keywords{galaxies: individual (NGC 4214) ---
77: galaxies: starburst ---
78: X-rays: binaries --- binaries: general ---
79: binaries: eclipsing --- stars: Wolf-Rayet}
80:
81:
82: \section{Introduction}
83:
84: With the advent of the sub-arcsec X-ray imaging capability of the \cxo\
85: X-ray Observatory (Weisskopf \etal\ 2003), observations of galaxies
86: out to the Virgo Cluster routinely detect tens
87: to hundreds of X-ray sources above
88: detection limits of $\sim 10^{37\pm1}$~\ergl\
89: (Fabbiano \& White 2006).
90: In analogy with our own Milky Way, the majority of these bright sources are
91: likely X-ray binaries.
92: The precise nature of many of these objects is
93: usually difficult to quantify because
94: of a lack of high-quality X-ray spectra and light curves.
95:
96: In a few instances, periodic dips are apparent in
97: the observed X-ray light curves
98: of individual sources.
99: By interpreting these dips as eclipses by a companion star, the light curves
100: can be used to constrain the system orbital parameters and even the mass of
101: the compact object if suitable additional information is available such as
102: an estimate of the companion star mass via
103: spectral and luminosity-class typing (see, e.g.,
104: Weisskopf et al. 2004; Pietsch \etal\ 2004; Pooley \& Rappaport 2005;
105: Fabbiano et al. 2006).
106:
107: \ngc\ is a dwarf starburst galaxy, morphological type IAB(s)m, located 3.5~Mpc
108: distant (1\arcsec$=$17~pc). The brightest X-ray source in \ngc, \cxou\
109: (hereafter, \cxous), shows distinct
110: behavior indicative of an eclipse at a period of 3.62~hrs. This periodicity is
111: visible in each of 5 X-ray observations taken over a 10-yr timespan though no
112: more than two cycles are evident in any single observation. The X-ray light
113: curves and spectra of the source are presented in \S~2.
114: Upper limits from a search for an optical counterpart are given in \S~3.
115: An analysis of the X-ray source and its companion based on the
116: observational evidence is derived in \S~4. In
117: \S~5 we discuss the interpretation of the
118: observations and the implications for the nature
119: of the binary.
120:
121: \section{X-ray Timing and Spectra}
122:
123: Table~1 lists the X-ray observations of \ngc\
124: that include the variable source \cxous.
125: Previous analysis of the initial \cxo\ and the
126: \xmm\ observations are reported in
127: Hartwell \etal\ (2004).
128: They list source \cxous\ as source~11 at
129: $\alpha=12^h15^m38.^s25$,
130: $\delta=36^{\circ}19\arcmin21.\arcsec4$
131: (J2000 coordinates).
132: It is by far the brightest point source among the
133: 20 discrete sources they detect,
134: with a 0.3--8.0~keV flux of 3.0$\times$10$^{-13}$~\ergcms.
135:
136: \begin{center}
137: %\small{
138: \begin{tabular}{llcc}
139: \multicolumn{4}{c}{{\sc Table 1}} \\
140: \multicolumn{4}{c}{X-ray observations of \cxou} \\
141: %
142: \hline \hline
143: %
144: \multicolumn{1}{c}{Date of} & Instrument & Exposure & Counts$^a$ \\
145: \multicolumn{1}{c}{observation}& & (ks) & ($\pm$~error) \\
146: \hline
147: 1994-12-10 &ROSAT/HRI & 42.6 & 91$\pm$12 \\ %& RH600741N00
148: 2001-10-16 &Chandra/ACIS-S & 26.4 & 943$\pm$31 \\ %& 2030
149: 2001-11-22 &XMM-Newton/PN & 16.7 & 453$\pm$36 \\ %& 0035940201
150: 2004-04-03 &Chandra/ACIS-S & 27.2 & 201$\pm$16 \\ %& 4743
151: 2004-07-30 &Chandra/ACIS-S & 28.6 & 843$\pm$30 \\ %& 5197
152: \hline
153: \multicolumn{4}{l}{$^a$background-subtracted in 0.5--8.0 keV energy band}
154: \end{tabular}
155: %} %end \small
156: \end{center}
157:
158:
159: \begin{figure*}[t]
160: \begin{center}
161: \includegraphics[angle=-90,width=0.97\columnwidth]{f1a.eps} %0.47
162: \includegraphics[angle=-90,width=0.97\columnwidth]{f1b.eps}
163: \includegraphics[angle=-90,width=0.97\columnwidth]{f1c.eps}
164: \includegraphics[angle=-90,width=0.97\columnwidth]{f1d.eps}
165: \figcaption{
166: Background-subtracted \cxo\ and \xmm\ X-ray light curves of \cxous.
167: The solid curve traces the best-fit 7-knot spline fit to the folded light
168: curve (see text).
169: \label{fig:alllc}}
170: \end{center}
171: \end{figure*}
172:
173: \begin{center}
174: \includegraphics[angle=-90,width=0.47\textwidth]{f2.eps} %0.97
175: \figcaption{Folded light curve of the X-ray dataset made by combining the
176: two brightest \cxo\ observations obtained on 2001-10-16 and 2004-7-30.
177: The heavy curve denotes the best-fitting 7-knot spline model to this
178: light curve.
179: \label{fig:foldedlc}}
180: \end{center}
181:
182: Our locally-developed analysis software suite,
183: {\tt LExtrct} (Tennant 2006), was
184: used to extract and analyze
185: the source and background light curves and
186: spectra from the \ros\ and \cxo\ data.
187: The \xmm\ Science Analysis System (version is
188: 6.5.0) was used for the \xmm\ data.
189: The nearest discrete source is 23.\arcsec 7 to the north and is only 10\% of
190: the flux of \cxous.
191: Thus, source \cxous\ is easily isolated from other
192: discrete sources even in the \xmm\ and
193: \ros\ data.
194: For the \cxo\ data, a circular extraction region of radius 5\arcsec\ was used
195: for the source and a nearby region of area 170 square arcsecs was chosen
196: for the background.
197: For the \xmm\ data, a circular source region of radius 15\arcsec\ and a
198: background of 1385 square arcsecs were used for analysis.
199:
200: The 0.5--8.0~keV \cxo\ and \xmm\ light curves of \cxous\ are shown in
201: Figure~\ref{fig:alllc} (the \ros\ light curve is
202: sparsely sampled because of the
203: 90~min orbit of the satellite, and is not shown here). Prominent dips
204: are seen in all four light curves.
205: Using the longest exposure data, the \cxo\ observation on 2004-7-30,
206: the light curve was folded
207: on trial frequencies ranging from
208: 6.4$\times$10$^{-5}$ to 1.2$\times$10$^{-4}$~Hz
209: (some noise is apparent at lower frequencies).
210: The resulting $\chi^2$ statistic {\sl vs.} trial
211: frequency was fit to a Gaussian
212: distribution. The peak of the Gaussian is at 7.682(60)$\times$10$^{-5}$~Hz
213: corresponding to a period of 13020(100)~s.
214: A repeat of this procedure for the first \cxo\ dataset, from
215: 2001-10-16, resulted in a
216: best-fit period of 12900(170)~s.
217:
218: Since both datasets have the same period, within the uncertainties,
219: we combined them to increase the number of counts per phase bin.
220: The minima of the two datasets were aligned via a $\chi^2$ fitting
221: procedure before they were added together.
222: The folded light curve made from this combined dataset (in 64 phase
223: bins, Fig.~\ref{fig:foldedlc}) was fit using a 7-knot spline with
224: periodic boundary conditions to create a model light curve.
225: This model was then applied to all the light
226: curves as shown in Fig.~\ref{fig:alllc}.
227: For these individual exposures, only the amplitude, phase, and DC level
228: of the spline model were
229: allowed to vary in the model fitting.
230: A period of 13020(100)~s is consistent with all the observations.
231: However, since the uncertainty in the period is of order 1\%, phase
232: coherence is totally lost in 100 cycles or
233: approximately two weeks; much shorter
234: than the time between observations.
235: For this analysis, no barycenter corrections were applied because
236: they should be at most
237: $\sim$30 km~s$^{-1} \times 10^4$ s or about 1 lt-sec,
238: which is well within our measurement
239: errors.
240:
241: %\pagebreak
242: \begin{center}
243: \includegraphics[angle=0,width=0.47\textwidth]{f3.eps} %0.97
244: \figcaption{Three-color \hst\ image of an 8\arcsec $\times$8\arcsec\
245: region around \cxous\ in \ngc.
246: Data are from 1997 (MacKenty \etal\ 2000) and include
247: F656N (red), F555W (green), and F336W (blue) filters.
248: The error region for \cxous, with a 1/3$''$ radius,
249: is shown as a circle near the image center. North is indicated.
250: \label{fig:hst}}
251: \end{center}
252:
253: Several models provide acceptable fits to the spectra of \cxous\ extracted from
254: the individual \cxo\ and \xmm\ observations.
255: The model fit parameters are consistent with the values obtained by
256: Hartwell \etal\ (2004); namely, for an absorbed power law model,
257: $N_{\rm H} = (1.8^{+0.6}_{-0.7})\times 10^{21}$~cm$^{-2}$ and
258: $\Gamma = 1.8^{+0.3}_{-0.2}$.
259: The average flux in the 0.5--8.0~keV band ranges from
260: (2.0$\pm$0.1)$\times$10$^{-14}$~\ergcms\ (on 2004-04-03) to
261: (3.41$\pm$0.34)$\times$10$^{-13}$~\ergcms\ (on 2001-10-16).
262: The flux during the bright phases on 2001-10-16 is
263: (4.77$\pm$0.48)$\times$10$^{-13}$~\ergcms.
264: Fewer than 1000 source counts were accumulated during
265: any individual X-ray observation
266: of \cxous\ (Table~1).
267: Therefore, no other important constraints on the nature of the source can
268: be deduced from these spectra.
269:
270: If the dips in the light curve are due to an eclipse, then the spectral shape
271: during time intervals near the minimum in the light curve may differ from
272: the shape at other times due to changes in
273: absorption or scattering of X-rays by the
274: atmosphere or wind of the companion star.
275: To test for this possibility, we compared spectra extracted from high-
276: ($>$0.035~c~s$^{-1}$) and low-count-rate ($<$0.025~c~s$^{-1}$)
277: phases of the 2001-10-16 \cxo\ observation.
278: A Kolmogorov-Smirnoff test showed these spectra are consistent with being
279: drawn from the same parent distribution at the 40\% level.
280: Examination of the
281: soft (0.5--2.0) and hard (2.0--8.0~keV) light curves also show
282: no evidence for enhanced absorption near or away from eclipse.
283:
284: In summary,
285: periodicity is apparent in all the observations of \cxous. We surmise the
286: periodicity is due to an eclipse of the X-ray source by a companion
287: star and that the orbital period of the system is
288: $P_{\rm orb} = 13020(100)~{\rm s} = 3.62(3)~{\rm h}$.
289:
290: \section{Optical Counterpart Search} \label{sec:optical}
291:
292: \hst\ observations of \ngc\ were carried out in 1997 in several filters
293: as reported by MacKenty \etal\ (2000).
294: The dynamical center of \ngc\ (\cxo\ source~10 in Hartwell \etal\ 2004)
295: and \cxous\ are both located on the WPC2 camera~3 in these
296: observations making the registration between the optical and X-ray images
297: accurate.
298: We estimate the radius of the X-ray error circle to be $\sim$0.\arcsec 3.
299: There are several optical sources within this circle in the \hst\ images
300: (Figure~\ref{fig:hst}).
301: Taking the total light within the error circle as a conservative upper limit
302: to any potential counterpart, we deduce the following observed magnitudes:
303: $m_{\rm F336W}=23.6$~mag, $m_{\rm F555W}=23.7$~mag,
304: $m_{\rm F702W}=22.4$~mag,
305: and $m_{\rm F814W}=22.0$~mag.
306: Since the different sources within the error circle contribute different
307: amounts in the different bandpasses, optical colors
308: based on the total light are not meaningful.
309: An estimated reddening correction can be made from the fitted X-ray absorption
310: column, which averages to $\sim$2$\times$10$^{21}$~cm$^{-2}$,
311: corresponding to
312: A(V)$\sim$1.1~mag.
313: The resulting upper limit to the optical
314: counterpart to \cxous\ is $m_{\rm F555W}=22.6$~mag.
315: By way of comparison, we note that this is about equivalent to
316: a late OV~star at the distance of \ngc.
317:
318: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
319: \section{Orbital Constraints}
320:
321: \subsection{Orbital Period}
322:
323: In order to understand what this eclipse discovery implies, we start
324: by examining the constraints on the donor star set by the measured
325: orbital period, $P_{\rm orb} =3.62$ hr. First we take Kepler's 3rd
326: law:
327: \begin{equation}
328: \frac{GM_T}{a^3} = \left(\frac{2\pi}{P_{\rm orb}}\right)^2 ~~,
329: \end{equation}%
330: where $M_T$ and $a$ are the total binary mass and orbital separation,
331: respectively, and utilize the relation: $R_L = r_L a = R_{\rm don}/f$,
332: where $R_L$ is the Roche-lobe radius of the donor
333: and $f$ is the fraction of the Roche lobe that is filled by the donor
334: star of radius $R_{\rm don}$. Combining these, we find:
335: %
336: \begin{equation}
337: \frac{{R}^{3/2}_{\rm
338: don}}{f^{3/2}{M}_T^{1/2}r_L^{3/2}} \simeq 0.36
339: \left(\frac{P_{\rm orb}}{1\,{\rm hr}}\right)~~,
340: \end{equation}
341: %
342: where ${M}_T$ and ${R}_{\rm don}$ are in solar units. Finally, we make
343: use of Eggelton's (1983) expression for $r_L$ to find:
344: %
345: \begin{equation}
346: \frac{{R}^{3/2}_{\rm don}}{{M}_{\rm don}^{1/2}} \simeq 0.12
347: \xi(q)f^{3/2}\left(\frac{P_{\rm orb}}{1\,{\rm hr}}\right)~~,
348: \end{equation}
349: %
350: where
351: %
352: \begin{equation}
353: \xi(q) \equiv
354: \frac{q^{1/2}\sqrt{1+q}}{\left[0.6q^{2/3}+\ln(1+q^{1/3})\right]^{3/2}}~~,
355: \end{equation}
356: %
357: and $q$ is the mass ratio $M_{\rm don}/M_{\rm acc}$. Here, $M_{\rm
358: acc}$ is the mass of the accreting star. Equation (3)
359: is analogous to the classical radius-mass relation (e.g.,
360: eq.\,[3] of Pooley \& Rappaport 2005; see also
361: the discussion in Weisskopf et al.\,2004) for
362: Roche-lobe filling donor
363: stars, but is generalized via Eggleton's (1983) expression for $r_L$
364: for virtually any mass ratio. The function $\xi(q)$ differs by less
365: than 10\% from unity for $0.01 \lesssim q \lesssim 2$, and increases
366: to a value of only 1.35 as $q \rightarrow 10$. An additional feature
367: of eq.\,(3) is that the factor $f$ allows for donor stars that
368: underfill their Roche lobes.
369:
370: \begin{center}
371: \includegraphics[angle=0,width=0.47\textwidth]{f4.eps} %0.97
372: \figcaption{Constraints on the eclipsing source \cxous\ in \ngc.
373: Red curves: radius-mass
374: constraints set by the value of $P_{\rm orb}$; solid, dashed and dot-dashed
375: curves correspond to the mass ratio,
376: $q \equiv M_{\rm don}/M_{\rm acc}$,
377: and Roche-lobe filling parameter, $f$, in the following pairs
378: (1,1), (10,1) and (1,3/4), respectively. Green
379: curves: H-burning stars; solid (dashed) curve represents the zero
380: (terminal) age main sequence. Blue curves: He-burning stars;
381: solid (dashed) curve represents the zero (approximate terminal)
382: age main sequence. Orange curve: degenerate (He) dwarfs.
383: \label{fig:RofM}}
384: \end{center}
385: %\end{figure}
386:
387: A plot of eq.\,(3), with $P_{\rm orb} = 3.62$ hr, is shown in
388: Figure~\ref{fig:RofM} in the radius--mass plane of the donor star.
389: The solid (dashed) red curve is for all $q \lesssim 2$ ($q = 10$);
390: both curves are for $f=1$ (i.e., a Roche-lobe filling donor). The lower
391: red dot-dashed curve represents the case where the donor star fills
392: only 3/4 of its Roche lobe, and $q=1$. The solid (dashed) green curve
393: is a simple $R(M)$ relation for zero age (terminal age) H-burning
394: main-sequence stars, while the solid (dashed) blue curve is the $R(M)$
395: relation for zero age (terminal age) He-burning main-sequence stars
396: (see, e.g., Paczy\'nski 1971; Kato \& Iben 1992; Justham \& Podsiadlowski
397: 2006, private communication). For completeness, we show, as an orange
398: curve, the $R(M)$ relation for degenerate He stars.
399:
400: From an inspection of Figure~\ref{fig:RofM},
401: and the intersection of the red curves
402: with the green and blue curves, we can draw some basic conclusions
403: about the nature of the donor star.
404:
405: \vspace{0.3cm}\centerline{{\em
406: Hydrogen-burning companion
407: star}}\vspace{0.3cm}
408:
409: One obvious interpretation
410: of the nature of the eclipsing binary is a normal
411: main-sequence donor star with a mass of $\sim$0.4~\msun\
412: (the intersection of the green and red curves of Figure~\ref{fig:RofM}).
413: This,
414: coupled with an orbital period of 3.62 hr, evokes either a
415: conventional low-mass X-ray binary (LMXB) in NGC 4214, or a
416: cataclysmic variable binary (CV) with a white dwarf accretor in our
417: own Galaxy. In the former case, an important question would be how
418: an unevolved, low-mass donor star could drive a sufficiently high
419: rate of mass transfer to account for $L_x \simeq 7 \times 10^{38}$~\ergl.
420: Mass transfer from an unevolved 0.4~\msun\ star is induced by
421: orbital decay through gravitational
422: radiation and magnetic braking
423: (see, e.g., Rappaport, Verbunt, \& Joss 1983).
424: The mass transfer rate would then be low
425: and we estimate the corresponding X-ray luminosity would range from only
426: $\sim 2 \times 10^{36}$~\ergl\ (for gravitational radiation and a
427: neutron star accretor) to
428: $\sim 6 \times 10^{37}$~\ergl\ (for magnetic braking and a
429: stellar-mass black hole accretor).
430:
431: In the case that \cxous\ is a CV in our own Galaxy,
432: the shape of the X-ray eclipse (see, Figs.~\ref{fig:alllc}
433: and~\ref{fig:foldedlc}) is reminiscent of that of AM Her systems
434: (Heise \etal\ 1985).
435: If CVs are confined to the Galactic disk with a scale height $H\sim300$~pc,
436: then \cxous, in this interpretation, would be within
437: $\sim$1~kpc, corresponding to an X-ray luminosity of
438: $\sim$6$\times$10$^{31}$~\ergl, or about 1/3 of the average luminosity
439: for AM~Her systems (see, e.g., Ramsay \& Cropper 2003).
440:
441: There are no physical solutions for extremely~--~or even
442: moderately~--~evolved normal main-sequence
443: stars which would more
444: naturally drive the high rate of mass transfer
445: implied for a system
446: at the distance of \ngc.
447:
448: \vspace{0.3cm}\centerline{{\em Helium-burning
449: companion star}}\vspace{0.3cm}
450:
451: Another obvious possibility for the nature of the
452: binary is a He-burning star in a binary with either a neutron-star or
453: stellar-mass black-hole accretor (see also Weisskopf et al. 2004).
454: The intersection of the various orbital constraint (red) curves with
455: the He-star (blue) curves in Fig.\,4 yield a range of possible
456: combinations of He-star mass and accretor mass.
457: For example, a $\sim$12~\msun\ He-ZAMS star could fill its Roche with an
458: accretor of comparable mass (i.e., $q \sim 1$), e.g., a
459: $\sim$12~\msun\ black hole. Alternatively, a more massive
460: He-ZAMS donor of $\sim$23~\msun\ would fill its Roche lobe with
461: a much less massive accretor, e.g., a neutron star. Or a $\sim$7~\msun\
462: He-ZAMS donor could underfill its Roche lobe (by a factor of, e.g., 3/4) and
463: still achieve substantial mass transfer via a stellar wind. However, for
464: reasons discussed below, perhaps the most promising interpretation is
465: an intermediate-mass He star (e.g., $\sim$2--3~\msun) that is somewhat evolved, with approximately twice its main-sequence radius, or near
466: the He terminal-age main sequence (TAMS).
467:
468: \subsection{Eclipse Duration}
469:
470: For the case where the X-ray eclipse is due to
471: the compact accretor going behind the companion
472: star, we can also use the eclipse duration to
473: learn more about the constituent masses of the
474: binary. Note that these constraints would
475: therefore pertain to the case of the He-star
476: companion, but probably not to an AM Her-type eclipse.
477: We modify the expression for the eclipse half
478: angle, $\theta_{\rm ecl}$, from eq.\,(4) of
479: Pooley \& Rappaport (2005; see also Weisskopf et al. 2004)
480: to allow for arbitrary mass ratios, $q$,:
481: %
482: \begin{equation}
483: \theta_{\rm ecl} =
484: %\sin^{-1}\left[\frac{0.49q^{2/3}}{0.6q^{2/3}+\ln(1+q^{1/3})}\right]
485: \cos^{-1}\left[\frac{1}{\sin
486: i} \sqrt{1-r_L^2}\right]
487: ~~,
488: \end{equation}
489: %
490: where $i$ is the orbital inclination angle, and
491: %
492: \begin{equation}
493: r_L = \frac{0.49q^{2/3}}{\left[0.6q^{2/3}+\ln(1+q^{1/3})\right]}~~,
494: \end{equation}
495: %
496: (Eggleton 1983). A plot of $\theta_{\rm ecl}(q)$ is shown in
497: Fig.~\ref{fig:eclipse} for inclination angles of $90^\circ$,
498: $80^\circ$, and $70^\circ$. The horizontal
499: lines denote our estimates for the maximum and minimum observed
500: eclipse half angles. We have not performed any detailed
501: eclipse calculations for the case where a stellar wind accounts for
502: an appreciable portion of the X-ray modulation. Given the limited
503: statistical precision associated with the X-ray
504: light curves, we can say only that the drop in
505: intensity more closely resembles a classical X-ray eclipse, i.e.,
506: as in a high-mass X-ray binary, than a wind-modulated light curve,
507: e.g., X1908+075 (Levine et al. 2004).
508:
509: %\begin{figure}[t]
510: \begin{center}
511: \includegraphics[angle=0,width=0.47\textwidth]{f5.eps} %0.97
512: \figcaption{Geometric eclipse half-angle (in degrees) as a function of
513: the mass ratio, for several values of the inclination angle, $i$ (eq.\,[5]).
514: The horizontal lines are estimates for the observed upper and lower limits on
515: $\theta_{\rm ecl}$.
516: \label{fig:eclipse}}
517: \end{center}
518: %\end{figure}
519:
520: From the intersection of the two horizontal curves with the
521: $\theta_{\rm ecl}(q)$ curves in Fig.~\ref{fig:eclipse}, we conclude
522: that the mass ratio in this system is
523: %
524: \begin{equation}
525: 1 \lesssim q %= \frac{M_{\rm don}}{M_{\rm acc}}
526: \lesssim 10 ~~.
527: \end{equation}
528: %
529: This range of $q$ values is consistent with the illustrative system
530: components suggested above. In particular, unevolved He stars of mass
531: $\sim7-18$~\msun\ could have a black-hole accretor of
532: $\sim7-12$~\msun\ (intersection of solid blue with solid and
533: dot-dashed red curves in Fig.\,4). Higher mass He stars (i.e.,
534: $\gtrsim$22~\msun) would require much less massive accretors (e.g.,
535: neutron stars; intersection of solid blue and dashed red curves).
536: Finally, somewhat evolved He stars of $\sim2-3$~\msun\ could have
537: neutron-star or very low-mass black hole accretors.
538:
539: \section{Discussion}
540:
541: Our observations of a 3.62-hr eclipse in the source
542: \cxous\ have yielded three distinct possibilities
543: for the nature of the binary. We discuss the
544: pros and cons of each in turn.
545:
546: \subsection{Conventional LMXB}
547:
548: As discussed above, in the context of the orbital constraints, if the
549: donor is a H-rich, low-mass star it cannot be very evolved.
550: Thus, it is unclear how an unevolved, low-mass donor star could drive a
551: sufficiently high rate of mass transfer to account for values of
552: $L_x$ that approach $7 \times 10^{38}$ ergs s$^{-1}$ during several different observations spanning nearly 3 years~--~especially if
553: the accretor were a neutron star.
554: Conventional black hole LMXBs,
555: on the other hand, can have high luminosities but
556: only during short transient outbursts.
557: Systems with short orbital periods like \cxous\
558: should either not be transients or, if they
559: are, should not have such prolonged intervals of high $\dot M$ (see, e.g.,
560: King \etal\ 1996; Kalogera \etal\ 2004; Fabbiano \etal\ 2006).
561: Moreover,
562: if the accretor were a stellar-mass black hole then $q \ll 1$ and
563: the eclipse duration would be much shorter than observed
564: (Figure~\ref{fig:eclipse}).
565:
566: In discussing a possible short orbital period ULX, Liu \etal\ (2002)
567: suggested beaming of the X-radiation to reduce the actual value of
568: $L_x$. Given that \cxous\ is an eclipsing system this possibility seems
569: remote, though not impossible if the accretor were a rapidly spinning black
570: hole, and a large kick during its formation oriented the orbital
571: plane perpendicular to the spin axis.
572:
573: \subsection{CV/AM Her system}
574:
575: For the case of an AM Her system,
576: the eclipse would be caused by periodic occultations of
577: the accreting magnetic polar cap as it rotates
578: around the spin axis of the white dwarf. In this
579: case, we cannot infer anything further about the
580: mass ratio of the binary constituents, as in
581: Fig.~\ref{fig:eclipse}. Thus, we focus here on the constraints
582: set by the HST image (Fig.~\ref{fig:hst}), and an evaluation
583: of the probability of finding a high Galactic
584: latitude ($b \simeq 78^\circ$) AM Her binary
585: within the $D_{25}$ area of \ngc. As discussed in
586: \S\ref{sec:optical}, the limiting visual magnitude
587: for any counterpart to \cxous\ is 22.6. However,
588: at $P_{\rm orb} \simeq 3.62$ hr, we expect the
589: donor star in an AM Her system to be roughly of
590: intrinsic spectral type $M0~V-M3~V$ with visual
591: magnitude $8.9 \lesssim M_{\rm V} \lesssim 10.5$.
592: From the optical limit, this requires the source to be
593: at least 2.6 to 5.5~kpc away.
594: Thus, unless the hypothesized AM Her system is in
595: the Galactic halo, the observed constraint on its
596: magnitude seems difficult to reconcile with its
597: expected brightness.
598:
599: It also seems
600: somewhat unusual that an AM Her system would be
601: found by chance so close ($\sim 25''$) to a
602: particular galaxy that is being observed with
603: \cxo\ for the purpose of detecting
604: luminous X-ray sources. The $D_{25}$ ellipse region
605: associated with \ngc~has semi-minor and
606: semi-major axes of $197''$ and $255''$,
607: respectively. This corresponds to a solid angle
608: of $\sim$44 square arcmin, or $0.012$ square
609: degrees. If there are $n_{\rm CV}$ CVs pc$^{-3}$ with
610: $P_{\rm orb} \gtrsim 3$ hr, then the surface
611: density, $S$, toward the NGP is $2 n_{\rm CV} H^3$~sr$^{-1}$, for
612: an assumed exponential distribution of CVs from
613: the Galactic plane with scale height $H$. For
614: illustrative parameters of $n_{\rm CV} \sim 10^{-6}$
615: pc$^{-3}$ (see, e.g., from observations, Patterson 1984;
616: Howell \etal\ 2002; Schmidt et al. 2005; Gansicke et al. 2005;
617: Szkody et al. 2006; and from population models,
618: de Kool 1992; Kolb 1993; Han et al. 1995; and Politano,
619: Howell, \& Rappaport 1998) and $H \sim 300$ pc,
620: we have $S \sim 55$ per steradian, or
621: $\sim$0.016 per square degree. Thus, the a
622: priori probability of finding a CV with $P_{\rm
623: orb} \gtrsim 3$ hr within the 0.012 square degree
624: area of the \ngc\ $D_{25}$ ellipse seems quite small, i.e.,
625: $\sim$2$\times$10$^{-4}$.
626:
627: Finally, in regard to finding a CV close to a
628: galaxy of particular interest, we note that a
629: similar circumstance has arisen in the case of CG
630: X-1 in the field of the Circinus galaxy
631: (Weisskopf et al.\,2004). This particular
632: eclipsing source had an orbital period of 7.5 hr,
633: which already is quite long for an AM Her system.
634: The Circinus galaxy is at $b \simeq -3.8^\circ$
635: (compared to $78^\circ$ for \cxous), and
636: therefore the probability of finding a
637: chance foreground CV is higher than near the NGP.
638: Nonetheless, these two eclipsing sources, when
639: taken together, reduce the likelihood that both
640: are foreground CVs.
641:
642: Overall, we find the hypothesis that \cxous\ is a
643: foreground AM Her system to be implausible.
644:
645: \subsection{Naked He-burning donor star}
646:
647: If we interpret the \cxous\ system as
648: having a naked He-burning donor star of
649: mass somewhere in the range of $\sim 2-20$~\msun,
650: the requisite mass-transfer
651: rate, $\dot M$, could be achievable via either
652: direct Roche-lobe overflow or via Bondi-Hoyle
653: capture (see, e.g., Bondi \& Hoyle 1944) of a
654: modest fraction of the stellar wind from the He
655: star (Dewi et al.\,2002; Justham et al.\,2006). For the
656: lower-mass He stars within this range, direct Roche-lobe
657: overflow would probably be required to supply
658: the mass transfer rate. The nuclear lifetimes of He stars
659: in this mass range
660: are quite short ($\sim$3$\times$10$^5$ to 3$\times$10$^6$ yr;
661: Paczy\'nski 1971; Habets 1986; Langer 1989; Kato \& Iben
662: 1992; Dewi et al.\,2002; Justham \& Podsiadlowski 2006,
663: private communication) so that $\dot M$ values of
664: $\gtrsim 10^{-7}$~\msun\ yr$^{-1}$ are easily attained.
665:
666: If the donor star is indeed a He-burning star, there is an additional
667: constraint that the X-ray light curve neither appears
668: to be dominated by modulation in a dense stellar
669: wind, nor suffers from any obvious photoelectric
670: (i.e., strongly energy-dependent) absorption.
671: For a simple $1/r^2$ stellar wind profile, the
672: column density as a function of orbital phase (at
673: $i = 90^\circ$) has a simple analytic form:
674: %
675: \begin{equation}N = n_0 a \phi/\sin \phi ~~~,
676: \end{equation}
677: %
678: except for values of $\phi$ corresponding to direct geometric
679: eclipse (e.g., Levine et al. 2004), where $n_0$ is the wind density
680: at the orbit of the compact accretor, $a$ is the orbital radius, and $\phi$
681: is the phase angle (with $\phi = 180^\circ$ being
682: defined as superior conjunction). For some
683: illustrative parameters, we find:
684: %
685: \begin{equation}N \simeq 10^{23} \dot M_{-6}
686: \left(\frac{a}{3\,R_\odot}\right)^{-1}
687: v_{1000}^{-1} ~ {\rm cm}^{-2}~~~,
688: \end{equation}
689: %
690: where the units for $\dot M$ and $v$ are
691: $10^{-6}$~\msun\ yr$^{-1}$ and 1000 km
692: s$^{-1}$, respectively. Thus, unless the stellar
693: wind is essentially completely ionized, the
694: optical depth to soft X-rays would be enormous.
695: For stellar-wind density profiles that start with
696: zero velocity and infinite density at the stellar
697: surface (e.g., Lucy \& Solomon 1970; Castor,
698: Abbott, \& Klein 1975), the column densities
699: would be larger, and the discrepancy with
700: the observed light curve would be even greater.
701: Therefore, if the donor star in \cxous\ is
702: ultimately identified with a He star with $M \gtrsim 3-4$~\msun,
703: this apparent lack of a dense stellar wind will have to be
704: reconciled with the binary stellar model.
705:
706: To quantify the constraint on the He-donor star set by the
707: apparent lack of a strong stellar wind, we utilize the following
708: expression for stellar-wind rates as a function of luminosity by
709: Hamann, Koesterke, \& Wessolowski (1995):
710: %
711: \begin{equation}
712: \log \dot M (M_\odot\, {\rm yr}^{-1}) = -11.95 + 1.5 \log (L/L_\odot)
713: ~~,
714: \end{equation}
715: %
716: where $L$ is the bolometric luminosity of the He star. More
717: recently, Hamann \& Koesterke (1998) and Petrovic, Langer \& van der
718: Hucht (2005) have suggested that the wind loss rates are lower than
719: given by eq.\,(10) by factors of a few. Only for luminosities
720: $\lesssim 10^{4}\,L_\odot$ does the donor star avoid producing a
721: highly attenuating stellar wind (e.g., with
722: $\dot M \lesssim 10^{-6}$~\msun\ yr$^{-1}$). This luminosity,
723: in turn corresponds to He star masses $\lesssim 3.5$~\msun\
724: (Kato \& Iben 1992; Justham \& Podsiadlowski 2006, private communication).
725: Our limit on the optical
726: counterpart of 22.6 magnitudes, is also much more consistent with
727: lower mass He stars (i.e., $2-3$~\msun).
728:
729: \section{Summary and Conclusions}
730:
731: The brightest X-ray source in the field of the
732: galaxy \ngc, \cxou, shows distinct
733: eclipse-like intensity dips at a period of 3.62~hr.
734: These eclipses are present in each of five X-ray
735: observations spanning 1994 to 2004.
736: Assuming that these eclipses are indicative of the
737: orbital period, we can rather tightly constrain the
738: nature of the binary system.
739:
740: We have considered the possibility that \cxous\ is a
741: conventional LMXB that somehow produces a very high luminosity for a
742: sustained interval of time. The mass transfer rate that can be
743: driven by a low-mass, unevolved donor star is too low to supply the
744: requisite mass transfer rate, and it is unlikely that this source is,
745: or should be, a transient.
746:
747: We have also evaluated quantitatively
748: the possibility that the \cxous\
749: system is a foreground CV (of the AM Her
750: subtype). Major difficulties with this scenario
751: are (i) the improbability of finding an AM Her system
752: aligned so closely, by chance, with a galaxy of
753: interest (even after taking into account the fact
754: that {\em Chandra} has observed a substantial
755: number of such galaxies with equal exposure), and
756: (ii) the lack of an optical counterpart brighter than
757: 22.6~mag, requiring the CV be a halo object.
758: We consider this
759: possibility unlikely.
760:
761: The alternative is that the donor is a He-burning star.
762: He-donor star masses $M\gtrsim 4$~\msun\ are difficult
763: to reconcile with the apparent lack of a dense stellar wind.
764: We believe that the most likely interpretation of this system is a
765: $\sim2-3$~\msun\ slightly evolved (e.g., TAMS) naked He star
766: transferring matter (stably) through the inner Lagrange point to a
767: neutron star. This type of binary has been extensively modeled by
768: Dewi et al.\,(2002). If confirmed, this would open
769: exciting possibilities for both stellar evolution
770: studies (see, e.g., Dewi et al.\,2002; Justham et al. 2006) and the
771: interpretation of ultraluminous
772: X-ray sources at the lower end of their luminosity
773: function (see, e.g., Colbert \& Miller 2004;
774: Fabbiano \& White 2006).
775: In the case of \cxous, the peak observed luminosity is only
776: about twice the Eddington limit for accretion of He onto a neutron star.
777: If the accretor in such a system, i.e., with a He donor star, were a
778: stellar-mass black hole, then the Eddington limit
779: would be $\sim 4 \times 10^{39}$~\ergl; well within
780: the range of ultraluminous X-ray source luminosities.
781:
782: Finally, we note that \cxous\ may be the first known
783: {\sl immediate} progenitor of a compact double neutron
784: star binary; i.e., beyond the second
785: common envelope phase and prior to collapse
786: of the He star core.
787: A perhaps unverifiable prediction is that
788: within some $10^5-10^6$ years the evolved core of the He
789: star will collapse to produce a type Ib supernova explosion
790: (see, e.g., Podsiadlowski, Joss, \& Hsu 1992; Woosley, Langer,
791: \& Weaver 1995) and leave a binary radio pulsar
792: (if the natal kick given to the pulsar is not so large as to unbind
793: the system; see, e.g., Pfahl et al. 2002 and references therein;
794: Podsiadlowski et al. 2004). The putative neutron star that is currently in
795: the system is in the process of being spun up via accretion
796: torques and will be the rapidly rotating pulsar member of the
797: binary. It is also interesting to note that this predicted
798: event has actually occurred already (given the $\sim$10
799: Myr light travel time for the information to reach us).
800:
801: \acknowledgements
802: We thank Stephen
803: Justham, Philipp Podsiadlowski, Eric Pfahl, and
804: Steve Howell for very helpful discussions. KKG is
805: supported in part by NASA under grant NNG04GC86G issued
806: through the Office of Space Science.
807: DP gratefully acknowledges
808: support provided by NASA through Chandra Postdoctoral
809: Fellowship
810: grant number PF4-50035 awarded by the Chandra X-ray
811: Center, which
812: is operated by the Smithsonian Astrophysical
813: Observatory for NASA
814: under contract NAS8-03060. SR received some
815: support from Chandra
816: Grant TM5-6003X.
817:
818:
819: %==============================================================
820: %references
821: \begin{thebibliography}{}
822: \bibitem[Bondi \& Hoyle(1944)]{BH}
823: Bondi, H., \& Hoyle, F. 1944, MNRAS, 104, 273
824: \bibitem[]{1}
825: Castor, J. I., Abbott, D. C., \& Klein, R. I. 1975, ApJ, 195, 157
826: \bibitem[]{1}
827: Colbert, E. J. M., \& Miller, M. C. 2004, in
828: ``Tenth Marcel Grossmann Meeting on General Relativity'',
829: eds. M. Novello, S. Perez-Bergliaffa and R. Ruffini,
830: (World Scientific: Singapore) [astro-ph/0402677]
831: \bibitem[]{1}
832: de Kool, M. 1992, A\&A, 261, 188
833: \bibitem[]{1}
834: Dewi, J. D. M., Pols, O. R., Savonije, G. J., \& van~den~Heuvel, E. P. J.
835: 2002, MNRAS, 331, 1027
836: \bibitem[]{1} Eggleton, P. P. 1983, ApJ, 268, 368
837: \bibitem[]{1} Fabbiano, G., \& White, N. E. 2006, in ``Compact Stellar X-Ray
838: Sources'', eds. W. H. G. Lewin and M. van der Klis,
839: (Cambridge:Cambridge Univ. Press), 475 [astro-ph/0307077]
840: \bibitem[]{1} Fabbiano, G., et al. 2006, submitted to ApJ
841: \bibitem[]{1} Gansicke, B. T., et al. 2005, MNRAS, 361, 141
842: \bibitem[]{1} Habets, G. M. H. J. 1986, 167, 61
843: \bibitem[]{1} Hamann, W. -R., Koesterke, L., \& Wessolowski, U.
844: 1995, A\&A, 299, 151
845: \bibitem[]{1} Hamann, W. -R., \& Koesterke, L. 1998, A\&A, 335, 1003
846: \bibitem[]{1} Han, Z., Podsiadlowski, Ph., \&
847: Eggleton, P. 1995, MNRAS, 272, 800
848: \bibitem[]{1} Hartwell, J. M., Stevens, I. R., Strickland, D. K.,
849: Heckman, T. M., \& Summers, L. K. 2004, MNRAS, 348, 406
850: \bibitem[]{1} Heise, J., Brinkman, A. C., Gronenschild, E.,
851: Watson, M., King, A. R., Stella, L., \& Kieboom, K. 1985, A\&A, 148, L14
852: \bibitem[]{1} Howell, S.B., Mason, E., Huber, M., \& Clowes, R.
853: 2002, A\&A, 395, L47
854: \bibitem[]{1} Justham, S., Madhusudhan, N., Podsiadlowski, Ph., \& Rappaport,
855: S. 2006, in preparation for ApJ
856: \bibitem[]{1} Kalogera, V.,
857: Henninger, M., Ivanova, N., \& King, A. R. 2004, ApJ, 603, L41
858: \bibitem[]{1} Kato, M., \& Iben, I., Jr. 1992, ApJ, 394, 305
859: \bibitem[]{1} King, A. R., Kolb, U., \& Burderi, L. 1996, MNRAS, 464, L127
860: \bibitem[]{1} Kolb, U. 1993, A\&A, 271, 149
861: \bibitem[]{1} Langer, N. 1989, A\&A, 210, 93
862: \bibitem[]{1} Levine, A. M., Rappaport, S., Remillard, R., \& Savcheva, A.
863: 2004, ApJ, 617, 1284
864: \bibitem[]{1} Lucy, L. B., \& Solomon, P. M. 1970, ApJ, 159, 879
865: \bibitem[]{1} MacKenty, J. W., Maz-Apell\'{a}niz, J., Pickens, C. E.,
866: Norman, C. A., Walborn, N. R. 2000, AJ, 120, 3007
867: \bibitem[]{1} Nugis, T., \& Lamers, H. J. G. L. M. 2000, A\&A, 360, 227
868: \bibitem[]{1} Paczy\'nski, B. 1971, Acta Astr, 21, 1
869: \bibitem[]{1} Patterson, J. 1994, ApJ Suppl., 54, 443
870: \bibitem[]{1} Petrovic, J., Langer, N., \& van der Hucht, K.A.
871: 2005, A\&A, 435, 1013
872: \bibitem[]{1} Pietsch, W., Mochejska, B. J., Misanovic, Z., Haberl, F., Ehle, M., \& Trinchieri, G. 2004, A\&A, 413, 879
873: \bibitem[]{1} Pfahl, E., Rappaport, S.,
874: Podsiadlowski, Ph., \& Spruit, H. 2002, ApJ, 574, 364
875: \bibitem[]{1} Podsiadlowski, Ph., Joss, P. C., \& Hsu, J. J. L. 1992,
876: ApJ, 391, 246
877: \bibitem[]{1}
878: Podsiadlowski, Ph., Langer, N., Poelarends, A. J. T., Rappaport, S.,
879: Heger, A., \& Pfahl, E. 2004, ApJ, 612, 1044
880: \bibitem[]{1}
881: Politano, M., Howell, S. B., \& Rappaport, S. 1998, ASP Conference
882: Series, 137, 207
883: \bibitem[]{1} Pooley, D., \& Rappaport, S. 2005, ApJ Lett., 634, 85
884: \bibitem[]{1} Ramsay, G., \& Cropper, M. 2003, MNRAS, 338, 219
885: \bibitem[]{1} Rappaport, S., Verbunt, F., \& Joss, P.C. 1983, ApJ, 275, 713
886: \bibitem[]{1} Schmidt, G.D., et al. 2005, ApJ, 630, 1037
887: \bibitem[]{1} Szkody, P., et al. 2006, AJ, 131, 973
888: \bibitem[]{1} Tennant, A. F. 2006, AJ in press
889: \bibitem[]{1} van Kerkwijk, M. H., et al. 1992, Nature, 355, 703
890: \bibitem[]{1} Weisskopf, M. C., \etal\ 2003, ExA, 16, 1
891: \bibitem[]{1} Weisskopf, M. C., Wu, K.,
892: Tennant, A. F., Swartz, D. A., \& Ghosh, K. K. 2004, ApJ, 605, 360
893: \bibitem[]{1} Woosley, S. E., Langer, N., \& Weaver, T. A. 1995, 448, 315
894:
895: \end{thebibliography}
896:
897: \end{document}
898: