1: \documentclass[12pt,preprint]{aastex}
2: \begin{document}
3: \newcommand{\msun}{\mbox{M$_{\odot}$}}
4: \newcommand{\rsun}{\mbox{R$_{\odot}$}}
5: \newcommand{\zsun}{\mbox{Z$_{\odot}$}}
6: \newcommand{\lsun}{\mbox{L$_{\odot}$}}
7:
8: \title{Towards an Accurate Determination of Parameters for Very Massive
9: Stars: the Eclipsing Binary LMC-SC1-105\altaffilmark{1}}
10:
11: \author{Alceste Z. Bonanos\altaffilmark{2}}
12:
13: \affil{Carnegie Institution of Washington, Department
14: of Terrestrial Magnetism, \\ 5241 Broad Branch Rd. NW, Washington D.C.,
15: 20015, USA \\ \tt e-mail: bonanos@dtm.ciw.edu}
16:
17: \altaffiltext{1}{Based on observations obtained with the 2.5 meter
18: DuPont and 6.5 meter Magellan Clay Telescope located at Las Campanas
19: Observatory, Chile.}
20:
21: \altaffiltext{2}{Vera Rubin Fellow}
22:
23: \begin{abstract}
24:
25: This paper presents a photometric and spectroscopic study of the bright
26: blue eclipsing binary LMC-SC1-105, selected from the OGLE catalog as a
27: candidate host of very massive stars ($\geq 30\,\msun$). The system is
28: found to be a double-lined spectroscopic binary, which indeed contains
29: massive stars. The masses and radii of the components are $\rm
30: M_{1}=30.9\pm1.0\;\msun$, $\rm M_{2}=13.0\pm0.7\;\msun$, and $\rm
31: R_{1}=15.1\pm0.2\;\rsun$, $\rm R_{2}=11.9\pm0.2\;\rsun$, respectively.
32: The less massive star is found to be filling its Roche lobe, indicating
33: the system has undergone mass-transfer. The spectra of LMC-SC1-105
34: display the Struve-Sahade effect, with the \ion{He}{1} lines of the
35: secondary appearing stronger when it is receding and causing the
36: spectral types to change with phase (O8$+$O8 to O7$+$O8.5). This effect
37: could be related to the mass-transfer in this system. To date, accurate
38: ($\leq 10\%$) fundamental parameters have only been measured for 15
39: stars with masses greater than 30~\msun, with the reported measurements
40: contributing valuable data on the fundamental parameters of very massive
41: stars at low metallicity. The results of this work demonstrate that the
42: strategy of targeting the brightest blue stars in eclipsing binaries is
43: an effective way of studying very massive stars.
44:
45: \end{abstract}
46:
47: \keywords{binaries: eclipsing -- binaries: spectroscopic -- stars:
48: fundamental parameters, individual (OGLE053448.26-694236.4) -- galaxies:
49: individual (LMC)}
50:
51: \section{Introduction}
52: \label{section:intro}
53: \vspace{-0.4cm}
54: The fundamental parameters of very massive stars ($\geq 30\,\msun$)
55: remain uncertain, despite the large impact massive stars have in
56: astrophysics, both individually and collectively \citep[see review
57: by][]{Massey03}. The equations of stellar structure allow for stars with
58: arbitrarily large masses, however the mechanisms to form massive stars
59: \citep[accretion and mergers; e.g.][]{Bally05} and the associated
60: instabilities \citep[see][and references therein]{Elmegreen00,
61: Zinnecker07} are not well understood, hindering theoretical predictions
62: on the existence of an upper limit on the stellar mass. The ``mass
63: discrepancy'' problem, i.e. the disagreement between masses derived from
64: parameters determined by fitting stellar atmosphere models to spectra
65: and from evolutionary tracks \citep[see e.g.][for a
66: comparison]{Repolust04, Massey05}, still affects studies of single
67: massive stars, even though significant progress has been made in both
68: stellar atmosphere \citep[see review by][]{Herrero07} and stellar
69: evolution models \citep[e.g.][]{Meynet03}. The parameters of single
70: stars also suffer from suspected multiplicity, which in many cases
71: cannot be determined. Pismis 24-1 demonstrates this problem: its
72: inferred evolutionary mass $>200\,\msun$ \citep{Walborn02} contradicted
73: the upper stellar mass limit of $\sim 150\,\msun$ suggested by
74: statistical arguments based on observations
75: \citep{Figer05,Oey05}. \citet{Maiz-Apellaniz07} resolved it into a
76: visual binary with the {\it Hubble Space Telescope}, thereby removing
77: the discrepancy. One of its components is also a spectroscopic binary,
78: illustrating the systematic effects often accompanying ``single'' stars.
79:
80: \vspace{-0.2cm}
81: The only model-independent way to obtain accurate fundamental parameters
82: of distant massive stars and to resolve the ``mass discrepancy'' problem
83: is to use eclipsing binaries \citep[see review by][]{Andersen91}. In
84: particular, double-lined spectroscopic binary systems exhibiting
85: eclipses in their light curves are extremely powerful tools for
86: measuring masses and radii of stars. Specifically, the light curve
87: provides the orbital period, inclination, eccentricity, the fractional
88: radii and flux ratio of the two stars. The radial velocity
89: semi-amplitudes determine the mass ratio; the individual masses can be
90: solved for by using Kepler's third law. Furthermore, by fitting
91: synthetic spectra to the observed ones, one can infer the effective
92: temperatures of the stars, solve for their luminosities and derive the
93: distance \citep[e.g.\@][]{Bonanos06}. The most massive stars measured in
94: eclipsing binaries are galactic Wolf-Rayet stars of WN6ha spectral type:
95: NGC3603-A1 \citep[M$_1=116 \pm 31\; \msun$, M$_2=89 \pm 16\;
96: \msun$;][]{Schnurr08}, and WR~20a \citep[M$_1=83.0 \pm 5.0\; \msun$ and
97: M$_2=82.0 \pm 5.0\; \msun$][]{Rauw04, Bonanos04} in Westerlund~2,
98: presenting a challenge for both stellar evolution and massive star
99: formation models \citep{Yungelson08, Zinnecker07} and raising the issue
100: of the frequency and origin of ``binary twins'' \citep{Pinsonneault06,
101: Lucy06, Krumholz07}. Such systems are of particular interest, since
102: massive binaries might be progenitors of gamma-ray bursts
103: \citep[e.g.\@][]{Fryer07}, especially in the case of Population III,
104: metal-free stars \citep[see][]{Bromm06}.
105:
106: Analogs of these heavyweight champions, if not more massive binaries,
107: are bound to exist in the young massive clusters at the center of the
108: Galaxy (Center, Arches, Quintuplet), in nearby super star clusters
109: (e.g.\@ Westerlund\,1, R136), in Local Group galaxies (e.g.\@ LMC, SMC,
110: M31, M33) and beyond (e.g.\@ M81, M83, NGC 2403). A systematic
111: wide-ranging survey of these clusters and galaxies is currently
112: underway. The goal is to provide data with which to test star formation
113: theories, stellar atmosphere and stellar evolution models for both
114: single and binary stars as a function of metallicity, and the
115: theoretical predictions on the upper limit of the stellar mass. The
116: adopted strategy involves two steps: a variability survey to discover
117: eclipsing binaries in these massive clusters and nearby galaxies, which
118: is followed by spectroscopy to derive parameters of the brightest --
119: thus most luminous and massive -- blue systems. However, characterizing
120: massive stars requires the availability of 8-m class telescopes and high
121: resolution near-infrared spectrographs (since massive stars in the
122: Galaxy are extincted and extragalactic ones are faint) and has only
123: become feasible in the past few years. \citet{Bonanos07} demonstrated
124: that this method efficiently finds massive candidates, by performing the
125: first variability survey of the Westerlund~1 super star cluster and
126: discovering 4 massive eclipsing binary systems.
127:
128: Figure~\ref{massradius} illustrates the extent of our knowledge of
129: precise fundamental parameters of massive stars. It presents published
130: mass-radius measurements from eclipsing binaries, accurate to better
131: than $10\%$ for the more massive component. The zero-age main sequences
132: (ZAMS) at both Z=0.02 \citep{Schaller92} and Z=0.008 \citep{Schaerer93}
133: are overplotted as a reference. The Galactic data are mainly taken from
134: the compilations of \citet{Andersen91} and \citet{Gies03} with additions
135: and updates from \citet{Vitrichenko07}, but also \citet{Gonzalez05} for
136: the Large Magellanic Cloud (LMC), \citet{Harries03} and
137: \citet{Hilditch05} for the SMC, and \citet{Ribas05} for M31. A
138: literature search was done to include all accurate measurements of stars
139: in eclipsing binaries with masses $\geq 30\,\msun$, which are presented
140: in Table~\ref{massradiusdata}. This Table is, to my knowledge, complete
141: at present and consists of only 14 very massive stars with better than
142: $10\%$ mass-radius measurements, located in 3 galaxies. Of these, WR~20a
143: and M33 X-7 \citep{Orosz07} are the most massive and noteworthy. M33 X-7
144: contains a very massive $70.0 \pm 6.9\; \msun$ O-type giant and a
145: record-breaking $15.65\;\msun$ black hole, challenging current
146: evolutionary models, which fail to explain such a large black hole
147: mass. Without accurate measurements for a large sample of massive stars,
148: theoretical models will remain unconstrained.
149:
150: A survey to determine accurate parameters for several massive eclipsing
151: binaries in the low metallicity (Z$=0.008$) LMC was undertaken, with the
152: purpose of increasing the sample and improving our understanding of
153: these rare systems. Several candidates were selected from the OGLE-II
154: catalog of eclipsing binaries in the LMC \citep{Wyrzykowski03} as the
155: brightest systems with $B-V<0$. LMC-SC1-105, or OGLE053448.26-694236.4,
156: is one of the brightest eclipsing binaries with $I_{max}=13.04$~mag,
157: $V_{max}=12.97$~mag, $B_{max}=12.81$~mag and a preliminary semi-detached
158: classification. This work presents the analysis of the follow-up
159: observations obtained for LMC-SC1-105. The paper is organized as
160: follows: \S2 describes the spectroscopy and data reduction, \S3 the
161: spectral classification, \S4 the radial velocity curve, \S5 the light
162: curve analysis, \S6 the evolutionary status, and \S7 the conclusion.
163:
164: \section{Spectroscopy}
165: \vspace{-0.2cm}
166:
167: A total of 9 spectra of LMC-SC1-105 near quadrature phases were acquired
168: over 4 runs on 2 telescopes at Las Campanas Observatory, Chile. In
169: December 2005, January and February 2006, spectra were obtained with the
170: Echelle spectrograph on the 2.5-m DuPont telescope. The $1\arcsec \times
171: 4\arcsec$ slit resulted in a spectral resolution of 10 km~s$^{-1}$
172: ($R=30000$) or 0.15\AA\, at 4500\AA, as measured from the full width
173: half maximum of the comparison lamp lines. In January 2006 and October
174: 2007, additional spectra were obtained with the blue and red sides of
175: the MIKE spectrograph \citep{Bernstein03} on the 6.5-m Magellan Clay
176: telescope. The $1.0\arcsec \times 5.0\arcsec$ slit used in 2006 resulted
177: in a spectral resolution of 9~km~s$^{-1}$ ($R=32000$) or 0.14\AA\, at
178: 4500\AA, as measured from the full width half maximum of the comparison
179: lamp lines. The $0.7\arcsec \times 5.0\arcsec$ slit used in 2007
180: resulted in a spectral resolution of 7~km~s$^{-1}$ ($R=41000$) or
181: 0.11\AA\, at 4500\AA. Table~\ref{speclog} summarizes the log of the
182: observations, specifying the telescope and instrument used, the exposure
183: times and final signal to noise (S/N) ratio per pixel measured on the
184: merged spectra.
185:
186: The Magellan spectra were extracted using the MIKE reduction pipeline
187: written by D. Kelson \citep{Kelson00,Kelson03}. The extracted orders for
188: each star were averaged, normalized and merged. The wavelength coverage
189: of the final merged spectra is $3900-5050$\AA\, in the blue and
190: $5040-7150$\AA\, in the red. The DuPont spectra were reduced, extracted
191: and wavelength calibrated using the $noao.imred.echelle$ package in
192: IRAF\footnote{IRAF is distributed by the NOAO, which are operated by the
193: Association of Universities for Research in Astronomy, Inc., under
194: cooperative agreement with the NSF.}. Cosmic rays were removed from the
195: two dimensional images with the algorithm of \citet{Pych04}. The orders
196: were finally averaged, normalized and merged, yielding a wavelength
197: range $3700-9000$\AA. The heliocentric radial velocity corrections for
198: each star were computed with the IRAF $rvsao.bcvcorr$ routine and taken
199: into account in the subsequent radial velocity determination. Narrow
200: nebular emission, likely originating from a nearby HII region, is
201: present in the spectra and was removed. Finally, the \ion{Ca}{2} H and K
202: and \ion{Na}{1} D lines exhibit multiple absorption features,
203: corresponding to Galactic and LMC interstellar material with radial
204: velocities between 0 and 315 km~s$^{-1}$.
205:
206: \section{Struve-Sahade Effect and Spectral Classification}
207: \label{section:sse}
208:
209: Inspection of the quadrature spectra reveals that LMC-SC1-105 exhibits
210: the {\it ``Struve-Sahade effect''} \citep{Stickland97}. This term
211: describes the variable strength of the spectral lines of the secondary
212: star (or primary star in some cases) in a double-lined spectroscopic
213: binary \citep[see][and references
214: therein]{Howarth97}. Figures~\ref{lmc105ss}, \ref{lmc105ss2}, and
215: \ref{lmc105ss3} present quadrature spectra of LMC-SC1-105 for the most
216: prominent \ion{He}{1} and \ion{He}{2} lines. At phase 0.27, the
217: \ion{He}{1} lines of the secondary are significantly stronger than at
218: phase 0.75, while, of the \ion{He}{2} lines this is the case only for
219: \ion{He}{2} $\lambda4200$. Several mechanisms have been proposed to
220: cause this effect \citep[see][and references therein]{Bagnuolo99,
221: Linder07}, which could, in principle, affect the luminosity ratio and
222: masses derived. In the case of LMC-SC1-105, both lines are clearly
223: resolved and therefore the position of the line centers, and
224: consequently the radial velocities can be unambiguously measured.
225: Studying the Struve-Sahade effect in eclipsing binaries could be
226: valuable for understanding its origin, because the inclination and sizes
227: of stars are additionally known. A thorough investigation of the effect,
228: as undertaken by \citet{Linder07}, is beyond the scope of this paper.
229:
230: Following the criteria of \citet{Walborn90}, the spectral types of the
231: primary and secondary are O8V and O8III-V at phase 0.75,
232: respectively. The luminosity class of the secondary cannot be
233: unambiguously determined, because of the strong emission in the
234: \ion{He}{2} $\lambda4686$ line, superposed on the absorption line of the
235: secondary. The Struve-Sahade effect further causes the spectral types of
236: both stars to change. At phase 0.25 the stars appear to have types O7V
237: and O8.5III-V. \citet{Massey00} have assigned a O8.5III spectral type
238: to LMC-SC1-105 (or W28-22, LH 81-72) from their lower resolution
239: spectra, likely having observed it near or during primary
240: eclipse. According to the calibration of effective temperature (T$_{\rm
241: eff}$) with spectral type by \citet{Mokiem07} for the LMC, the primary
242: should have T$_{\rm eff}=35-40$kK and the secondary T$_{\rm
243: eff}=30-35$kK. A (sub)giant classification would make it $\sim1-2$kK
244: cooler than a dwarf with the same spectral class \citep[see
245: e.g.][]{Martins05}. Figures~\ref{fitmultioct23} and~\ref{fitmultioct25}
246: present the hydrogen and helium lines at each quadrature and {\sc
247: TLUSTY} model atmospheres (see \S\ref{section:rv}) for a range of
248: T$_{\rm eff}$. The combination that best fits the data has T$_{\rm
249: eff1}=35\pm2.5$kK and T$_{\rm eff2}=32.5\pm2.5$kK, with $\log(g)$ fixed
250: to 3.50 and the projected rotational velocities $vsini$ to the
251: synchronous values of 180 km~s$^{-1}$ and 140 km~s$^{-1}$ (as determined
252: in \S\ref{section:lc}). These T$_{\rm eff}$ values are consistent with
253: the spectral type calibration of \citet{Mokiem07}. The T$_{\rm eff}$
254: grid step was adopted as a conservative error. Note that no combination
255: of models can fit all the helium lines satisfactorily, in particular the
256: singlet lines \ion{He}{1} $\lambda4387$ and $\lambda4922$.
257: \citet{Puls05} point out that CMFGEN \citep{Hillier98}, and therefore
258: {\sc TLUSTY} \citep[since both models are consistent, see][]{Bouret03},
259: predicts much weaker \ion{He}{1} singlet lines than FASTWIND
260: \citep{Santolaya97, Puls05}. \citet{Bouret03} also state that ``a
261: simultaneous fit to all \ion{He}{1} and \ion{He}{2} lines is almost
262: never achieved''. Computing a finer grid of {\sc TLUSTY} models to
263: better constrain the T$_{\rm eff}$ of the stars or using a unified model
264: atmosphere, such as FASTWIND, was not pursued, because the determination
265: of masses and radii is independent of the T$_{\rm eff}$.
266:
267: \section{Radial Velocity Curve}
268: \label{section:rv}
269:
270: Two methods were used to measure the orbital parameters of LMC-SC1-105:
271: two dimensional cross correlation (or TODCOR) and spectral
272: disentangling. TODCOR was developed by \citet{Zucker94} and can
273: distinguish small velocity separations even more accurately than one
274: dimensional cross correlation \citep{Tonry79}. Synthetic spectra from
275: the OSTAR2002 {\sc TLUSTY} non-LTE grid \citep{Lanz03} at half-solar
276: metallicity were used for the cross correlation. The microturbulent
277: velocity was fixed to 10 km~s$^{-1}$ and the helium abundance to the
278: solar value, $He/H= 0.1$ by number. The T$_{\rm eff}$ of the models in
279: the grid range from 27500--55000 K in steps of 2500 K. The surface
280: gravity $\log(g)$, depending on the exact T$_{\rm eff}$ value, ranges
281: from 3.00--4.75 dex (cgs) in steps of 0.25 dex and the microturbulent
282: velocity was fixed at 10 km~s$^{-1}$. The models were rotationally
283: broadened (20-400 km~s$^{-1}$ in steps of 20 km~s$^{-1}$) and the
284: instrumental broadening was applied with the $rotin3$ program
285: distributed with the {\sc TLUSTY} grid.
286:
287: Initially, best fit models were computed by minimizing the residuals of
288: the sum of 2 models shifted appropriately to the highest S/N quadrature
289: spectrum from UT 2007 October 24. This procedure yielded the following
290: best fit models (T$_{\rm eff}$, $\log(g)$, $vsini$): (35000, 3.50, 160)
291: for the primary and (30000, 3.00, 140) for the secondary. These models
292: were used as templates in TODCOR to derive initial values for the radial
293: velocities, orbital elements and stellar parameters. Subsequently,
294: $vsini$ and $\log(g)$ were fixed for the estimation of T$_{\rm
295: eff}$. The models used for the final analysis are shown in
296: Figures~\ref{fitmultioct23} and~\ref{fitmultioct25}. The regions around
297: the Balmer lines and $\lambda\lambda4620-4700$ were excluded from the
298: TODCOR analysis, as the former are broad and the latter region contains
299: \ion{He}{2} $\lambda$4686 emission (formed in the wind or interaction
300: region), which the hydrostatic equilibrium {\sc TLUSTY} models cannot
301: reproduce. The resulting TODCOR velocities are given in
302: Table~\ref{rv}.
303:
304: An orbital fit for the systemic $\gamma$ velocity, the semi-major axis
305: and the mass ratio $q$ was performed with PHOEBE \citep[version
306: 0.31a,][]{Prsa05}, which builds on and enhances the capabilities of the
307: Wilson-Devinney program \citep{Wilson71, Wilson79, Wilson90}. The two
308: highest S/N spectra were assigned a weight of 2, while the lowest S/N
309: spectrum was assigned a weight of 0.5; the rest were assigned weights of
310: 1. The values and their formal uncertainties, found by fixing the
311: ephemeris from the OGLE catalog, are: $q=0.42\pm0.02$, $a \sin
312: i=38.9\pm0.5\,\rsun$, $\gamma=284\pm3\, \rm km\; s^{-1}$. These imply
313: semi-amplitudes of $K_{1}=137\pm4$ km~s$^{-1}$ and $K_{2}=326\pm3$
314: km~s$^{-1}$, and minimum masses of $M_{1} \sin^3 (i)=30.9\pm1.0\,\msun$
315: and $M_{2} \sin^3 (i)=13.0\pm0.7\,\msun$. The rms of the fit is 12~$\rm
316: km\; s^{-1}$ for the primary and 8~$\rm km\; s^{-1}$ for the secondary,
317: which are adopted as representative values for the error in each radial
318: velocity measurement. Separate $\gamma$ velocities for each star were
319: also fit for, but yielded values consistent within errors with the value
320: above, therefore were not considered further.
321:
322: Another accurate method for deriving radial velocities of spectroscopic
323: binary stars is spectral disentangling \citep{Simon94}. The program
324: KOREL \citep{Hadrava95} implements the method using Fourier
325: transforms. In order to explore the $\chi^2$ parameter space, KOREL was
326: run for a range of primary semi-amplitudes ($100-160$ km~s$^{-1}$) and
327: mass ratios ($q=M_{2}/M_{1}: 0.2-0.7$), using the range
328: $\lambda\lambda3990-4965$, a 4 km~s$^{-1}$ velocity step, excluding the
329: $\lambda\lambda4620-4700$ region. The resulting values were found to be
330: $K_1=140^{+15}_{-20}$ km~s$^{-1}$, $q=0.45^{+0.06}_{-0.08}$, in
331: agreement with the TODCOR results. The latter have more realistic errors
332: and were adopted in the subsequent analysis. \citet{Southworth07}
333: present the first detailed comparison of the methods developed to derive
334: radial velocities and find that disentangling is the most accurate,
335: however they applied TODCOR on a single line.
336:
337: \section{Light Curve Analysis}
338: \label{section:lc}
339:
340: Besides the OGLE $I-$band light curve, blue and red filter light curves
341: for LMC-SC1-105, roughly corresponding to $V$ and $R$, are also
342: available from the MACHO database (ID 81.8881.21). The MACHO light
343: curves have 331 measurements in the red and 167 in the blue and span 7.2
344: years, but are noisier than the OGLE-II and OGLE-III light curve
345: \citep[kindly made available by I. Soszynski; see][]{Soszynski08}, which
346: has 750 points (after removing outliers) that span more than 11
347: years. The instrumental OGLE-III light curve was offset to match the
348: OGLE-II light curve in the analysis below.
349:
350: Detailed simultaneous modeling of the 3 light curves was performed with
351: PHOEBE. Both the detached mode and the semi-detached mode with the
352: secondary filling its Roche lobe were considered. Note, the primary star
353: (star 1) in both binaries analyzed herein is defined photometrically, as
354: the hotter star producing the deeper eclipse at phase zero. The
355: parameters that were allowed to vary are: the inclination $i$, T$_{\rm
356: eff2}$, the period $P$, and the surface potentials $\Omega_1$ and
357: $\Omega_2$. The time of primary eclipse T$_0$ was fixed to the value
358: determined by \citet{Wyrzykowski03}, while the value of the mass ratio
359: and semi-major axis were fixed to the values determined from the orbital
360: fit. The light ratio was computed rather than fit for, following
361: \citet{Prsa05}. Synchronous rotation, a circular orbit and no third
362: light were assumed, as there was no evidence to the contrary. T$_{\rm
363: eff1}$ was fixed to the value determined from the spectra in
364: \S\ref{section:sse}, while the gravity darkening exponents were set to 1
365: for stars with radiative envelopes (i.e. $T_{\rm eff} \propto g^{0.25}$)
366: and the albedo values to 0.5, following \citet{Hilditch01}. Limb
367: darkening coefficients for the square root law were fixed from
368: \citet{Claret00} and the approximate reflection model was used.
369:
370: The values and errors of the parameters and the stability of the
371: solution were explored using PHOEBE's scripter. The Wilson-Devinney
372: differential corrections minimization was run 1000 times, each time
373: updating the input parameters to the values determined in the previous
374: iteration. The final values for the parameters were determined by
375: calculating the mean and standard deviation of the resulting values from
376: each iteration; the resolution step was adopted as the error for the
377: period. The errors are similar to the conventional errors from
378: Wilson-Devinney. A semi-detached configuration with the secondary
379: filling its Roche lobe yielded a consistent
380: model. Figures~\ref{lmc105i}, \ref{phoebemachorLC}
381: and~\ref{phoebemachobLC} present the phased OGLE and MACHO light curves
382: and best fit model from PHOEBE. The more accurate and better sampled
383: OGLE light curve exhibits a depression before the primary eclipse, which
384: is often seen in Algol-type binaries \citep[e.g., see][]{Hilditch05} and
385: is attributed to a mass-transfer stream. The \citet{O'Connell51} effect
386: explains the residuals near phase 0.25, since the first quadrature is
387: brighter than the second by 0.01 mag. The earlier spectral type derived
388: at this phase for the primary correlates with this extra flux. All bands
389: display deeper eclipses than predicted by the model. This could be due
390: to an inaccurate conversion of the differential flux light curves to
391: magnitudes caused by blending \citep[see][]{Zebrun01}, however it is
392: also seen in the MACHO light curves, obtained with profile fitting
393: photometry. The only way to obtain a model with flat residuals (without
394: changing the mass ratio) would be to additionally model cool spots. This
395: was not pursued, as adding free parameters would improve the fit, but
396: not significantly change the masses and radii.
397:
398: Figure~\ref{lmc105rv} shows the radial velocity curve and best fit model
399: from PHOEBE, while Figure~\ref{lmc105phoebe} presents histograms of the
400: PHOEBE scan results and the final parameters, which are also listed in
401: Table~\ref{wd105}. The ephemeris is:
402:
403: \begin{equation}
404: \rm T_0(HJD)=2450451.90113 + 4.250806(1) \times E,
405: \end{equation}
406:
407: \noindent refining the recent period determinations of \citet{Derekas07}
408: and \citet{Faccioli07} that are based on the MACHO data. The period was
409: also calculated using the analysis of variance algorithm of
410: \citet{Schwarzenberg89} on the OGLE light curve and was found to be
411: 4.25083 days. The uncertainties for the parameters are similar to the
412: formal errors from PHOEBE. T$_{\rm eff2}$ was left as an adjustable
413: parameter in PHOEBE as a proxy for the flux ratio, however, its value
414: and uncertainty are quoted from the spectral type calibration described
415: in \S\ref{section:sse}. The light (or luminosity) ratio is $L_{2}/L_{1}=
416: 0.46\pm0.02$ in $I$, $0.45\pm0.04$ in $R$ and $0.45\pm0.05$ in $V$ and
417: was fixed to 0.45 in the TODCOR analysis. These formal errors were
418: computed by iterating PHOEBE one more time with the primary light levels
419: as additional adjustable parameters, as described by \citet{Prsa05}. The
420: flux ratio is $F_{2}/F_{1}=0.74\pm0.03$, in agreement with the value
421: ($0.739 \pm 0.004$) found for LMC-SC1-105 by \citet{Mazeh06} in their
422: automatic analysis of OGLE eclipsing binaries.
423:
424: The physical parameters for the system are presented in
425: Table~\ref{physpar105}, with the final values for the masses and radii
426: of $\rm M_{1}=30.9\pm1.0\;\msun$ and $\rm R_{1}=15.1\pm0.2\;\rsun$ for
427: the primary, and $\rm M_{2}=13.0\pm0.7\;\msun$ and $\rm
428: R_{2}=11.9\pm0.2\;\rsun$ for the secondary. The $\log(g)$ value for the
429: secondary is lower than that for the primary, in support of the star
430: being evolved with a luminosity class between III-V. Given the physical
431: parameters, the synchronous rotational velocities of the stars are found
432: to be 180 km~s$^{-1}$ and 143 km~s$^{-1}$. Note, the spectra were
433: rotationally broadened to essentially identical values in
434: \S\ref{section:sse}. The measured masses and radii of the components
435: have been determined to $\sim5\%$ and $\sim2\%$ accuracy,
436: respectively. As discussed in \S1, only 14 accurate measurements
437: currently exist for very massive stars, with this work contributing
438: towards increasing this sample to 15 (see Table~\ref{massradiusdata}).
439:
440: \section{Evolutionary Status}
441:
442: The semi-detached configuration of LMC-SC1-105 with the less massive
443: star filling its Roche lobe, along with the main sequence classification
444: of the primary and possible (sub)giant classification of the secondary,
445: point to the system being in a slow-mass-transfer stage of case A binary
446: evolution. The depression in the light curve before primary eclipse
447: (described in \S\ref{section:lc}) also indicates the presence of a
448: mass-transfer stream. The emission in the \ion{He}{2} $\lambda$4686 line
449: and some of the Balmer lines (in particular H$\alpha$) originates near
450: the secondary and can be explained by gas being transfered onto the
451: primary. The velocities of these emission lines coincide with the
452: secondary star but are shifted, indicating an origin near the Lagrangian
453: L1 point (see Figures~\ref{fitmultioct23} and~\ref{fitmultioct25}). The
454: evolutionary state of the stars can also be inferred by comparison with
455: theoretical stellar evolution models. Figures~\ref{lumteff105} and
456: \ref{radmass105} compare the parameters of the binary with the widely
457: used theoretical stellar evolutionary models by \citet{Schaerer93} for
458: single stars and the newer models by \citet{Claret06}, respectively. The
459: plots illustrate that these post-mass transfer stars are oversized and
460: overluminous for their masses. According to the isochrones of
461: \citet{Claret06} in a mass-radius diagram for single stars, the primary
462: of LMC-SC1-105 has an age $\sim5$ Myr, while the age of the secondary is
463: $>10$ Myr (see Figure~\ref{radmass105}), which is not in agreement with
464: them being coeval. This apparent age discrepancy can be explained by
465: mass-transfer. The ages of the stars according to the evolutionary
466: models of \citet{Schaerer93} in a luminosity-temperature diagram are
467: $\sim4$ Myr for the primary and $\sim5$ Myr for the secondary.
468:
469: \citet{Massey00} have estimated the age and mass of LMC-SC1-105 (or
470: LH~81-72, W28-22), which is located in the LH~81 association. They find
471: a range of log(age[yr]) between 5.55-6.77 for the highest mass unevolved
472: stars and a turnoff mass around $70\,\msun$. Figure~\ref{radmass105}
473: indicates an age of $\sim5$ Myr or log(age[yr])$\sim6.7$ for the
474: primary, in agreement with the value of 6.61 derived by \citet{Massey00}
475: for LMC-SC1-105. They also find a range of reddenings $E(B-V)=0.13-0.23$
476: mag for LH~81 and a mean value of 0.15 mag. Late O-stars have
477: $B-V\sim-0.3$, therefore the measured color of LMC-SC1-105
478: ($B-V=-0.16$~mag) implies a reddening value $\sim$0.14
479: mag. Interestingly, LH~81 contains several evolved stars: the WC4 star
480: BR~50, the WN4+OB star BR~53 and the B0I+WN star Sk -$69^\circ$194
481: \citep{Massey00}. LMC-SC1-105 is found by these authors to have a mass
482: of $39\,\msun$ assuming it is single. This value is 26\% greater than
483: the dynamical mass of the primary or 12\% smaller than the total mass of
484: its components, illustrating the systematic error encountered in the
485: study of massive stars, which could also affect estimations of the
486: masses of their parent clusters.
487:
488: Given that LMC-SC1-105 has exchanged mass, a comparison with models of
489: close binary evolution is needed to infer the initial masses of the
490: components, however models at the metallicity of the LMC do not
491: currently exist. According to the solar metallicity evolutionary models
492: of \citet{Nelson01}, LMC-SC1-105 could have had the following initial
493: masses, listed as pairs of primary and secondary masses: (25.1, 22.4),
494: (39.8, 35.5), (50.0, 44.6) \msun, implying an initial mass ratio of
495: 0.90. Lower masses for the initial primary do not result in the measured
496: mass ratio. These models assume conservative mass-transfer, however the
497: sum of predicted inital masses ($>$47\,\msun) is greater than the sum of
498: the final, measured masses (44.1\,\msun). Furthermore, \cite{deMink07}
499: have shown that conservative evolution is not a valid assumption, by
500: comparing 50 SMC eclipsing binaries \citep[from][]{Harries03,
501: Hilditch05} with their grid of models for a range of mass-transfer
502: efficiencies. A similar grid at the metallicity of the LMC is therefore
503: necessary to estimate the initial parameters of the binary.
504:
505: What will the end product of the evolution of LMC-SC1-105 be? The
506: primary has a large fill-out ratio of $F=0.89$, as defined by
507: \citet{Mochnacki72}, implying that it will quickly fill its Roche lobe
508: when it leaves the main sequence and the hydrogen shell burning phase
509: begins. Further modeling of LMC-SC1-105, following \citet{deMink07} or
510: \citet{Petrovic05} for the mass-transfer efficiency, could indicate if
511: the secondary will evolve into a Wolf-Rayet star and subsequently
512: explode as a core-collapse supernova or a common envelope phase will
513: cause the stars to merge.
514:
515: \section{Conclusions}
516:
517: This paper presents accurate (to better than $5\%$) fundamental
518: parameters of LMC-SC1-105, one of the brightest blue eclipsing binary
519: stars in the LMC found by the OGLE survey. The aim of this work is
520: twofold: 1) to demonstrate that targeted surveys of the brightest blue
521: eclipsing binaries in nearby galaxies do indeed select very massive
522: stars and 2) to measure accurate parameters for one of these rare
523: systems. The parameters of LMC-SC1-105 were determined from the light
524: curves available from the OGLE and MACHO surveys and newly acquired high
525: resolution spectroscopy that targeted quadrature phases, in part
526: applying the strategy proposed by \citet{Gonzalez05} to constrain the
527: radial velocity curve with a small number of spectra. The system was
528: found to contain a very massive main sequence primary
529: ($30.9\pm1.0\,\msun$) and a possibly evolved Roche lobe-filling
530: secondary. The spectra display the Struve-Sahade effect, which is
531: present in all the \ion{He}{1} lines, causing the spectral
532: classification to change with phase, and could be related to the mass
533: transfer occurring in the system. LMC-SC1-105 could further be used as a
534: distance indicator to the LMC. However, in addition to accurate radii,
535: accurate flux (i.e. effective temperatures) and extinction estimates are
536: necessary for accurate distances. Eclipsing binaries have been used to
537: derive accurate and independent distances to the LMC
538: \citep[e.g.][]{Guinan98, Fitzpatrick03}, the Small Magellanic Cloud
539: \citep{Harries03,Hilditch05}, M31 \citep{Ribas05} and most recently to
540: M33 \citep{Bonanos06}.
541:
542: The accurate parameters determined herein for LMC-SC1-105 contribute
543: valuable data on very massive stars, increasing the current sample of 14
544: very massive stars with accurate parameters to 15, which despite their
545: importance remain poorly studied. Such data serve as an external check
546: to resolve the ``mass discrepancy'' problem, as \citet{Burkholder97}
547: have shown, and to constrain stellar atmosphere, evolution and formation
548: models. Further systematic studies of massive binaries in nearby
549: galaxies are needed to extend the sample of 50 SMC eclipsing binaries
550: \citep{Harries03, Hilditch05} to higher masses and metallicities and
551: populate the sparsely sampled parameter space (mass, metallicity,
552: evolutionary state) with accurate measurements of their masses and
553: radii. The method of targeting very massive stars in bright blue
554: eclipsing binaries can therefore be employed towards this goal.
555:
556: \acknowledgments{I am very grateful to Kris Stanek for motivating me to
557: undertake this project. I thank the referee, Ian Howarth, for a careful
558: reading of the manuscript, John Debes for obtaining the Magellan spectra
559: in October 2007, Igor Soszynski for making available the OGLE-III light
560: curve prior to publication, Andrej Pr{\v s}a and Phil Massey for
561: valuable comments that improved the manuscript, Guillermo Torres for his
562: program implementing TODCOR, Nolan Walborn for advice on the spectral
563: classification and Ignasi Ribas for his script to merge echelle
564: orders. This paper utilizes public domain data obtained by the MACHO
565: Project, jointly funded by the US Department of Energy through the
566: University of California, Lawrence Livermore National Laboratory under
567: contract No. W-7405-Eng-48, by the National Science Foundation through
568: the Center for Particle Astrophysics of the University of California
569: under cooperative agreement AST-8809616, and by the Mount Stromlo and
570: Siding Spring Observatory, part of the Australian National
571: University. AZB acknowledges research and travel support from the
572: Carnegie Institution of Washington through a Vera Rubin Fellowship.}
573:
574: {\it Facilities:} \facility{Magellan:Clay (MIKE)}, \facility{Du Pont (Echelle)}
575:
576: \begin{figure}[ht]
577: \plotone{fig1.ps}
578: \caption{Mass and radius determinations of stars in eclipsing binaries,
579: accurate to $\leq 10\%$ and complete $\geq 30\,\msun$ from the
580: literature (see \S\ref{section:intro} and Table~\ref{massradiusdata} for
581: references). The solid line is the Z=0.02 ZAMS from \citet{Schaller92};
582: the dashed line is the Z=0.008 ZAMS from \citet{Schaerer93}. Note the
583: small number of measurements for stars with masses greater than
584: $30\,\msun$, all published since 2001.}
585: \label{massradius}
586: \end{figure}
587:
588: \begin{figure}[ht]
589: \plotone{fig2.ps}
590: \caption{Quadrature spectra of LMC-SC1-105, obtained with MIKE/Magellan
591: at phases $\phi=0.75$ and 0.27, displaying the Struve-Sahade effect. The
592: primary (P) and secondary (S) stars are labelled in the first panel. The
593: panels display the \ion{He}{1} $\lambda4009$ and \ion{He}{2}
594: $\lambda4026$, \ion{He}{1} $\lambda4144$, \ion{He}{2} $\lambda4200$,
595: \ion{He}{1} $\lambda4387$, \ion{He}{1} $\lambda4471$ lines,
596: respectively. Note that the lines of the secondary are all stronger at
597: $\phi=0.27$; the systemic velocity is 284 km~s$^{-1}$; tickmarks
598: correspond to 5\,\AA\, in the last panel.}
599: \label{lmc105ss}
600: \end{figure}
601:
602: \begin{figure}[ht]
603: \plotone{fig3.ps}
604: \caption{The Struve-Sahade effect in LMC-SC1-105 quadrature spectra, as
605: in Figure~\ref{lmc105ss}, but for the following lines: \ion{He}{2}
606: $\lambda4541$, \ion{He}{2} $\lambda4686$, \ion{He}{1} $\lambda4713$,
607: \ion{He}{1} $\lambda4922$, \ion{He}{1} $\lambda5016$, respectively. The
608: \ion{He}{2} $\lambda4686$ emission is likely due to gas being transfered
609: from the secondary onto the primary; tickmarks correspond to 5\,\AA\, in
610: the last panel.}
611: \label{lmc105ss2}
612: \end{figure}
613:
614: \begin{figure}[ht]
615: \plotone{fig4.ps}
616: \caption{The Struve-Sahade effect in LMC-SC1-105 quadrature spectra, as
617: in Figure~\ref{lmc105ss}, but for the following lines: \ion{He}{2}
618: $\lambda5411$, \ion{He}{1} $\lambda5876$, H$\alpha$ $\lambda6563$,
619: \ion{He}{1} $\lambda6678$, \ion{He}{1} $\lambda7065$, respectively. Note
620: the narrow Galactic and LMC interstellar \ion{Na}{1} D lines at
621: $\lambda\lambda5890-95$. The S/N of the spectra decreases at redder
622: wavelengths, due to the lower sensitivity of MIKE and the color of the
623: system. Tickmarks correspond to 5\,\AA\, in the second and fourth
624: panels.}
625: \label{lmc105ss3}
626: \end{figure}
627:
628: \begin{figure}[ht]
629: \includegraphics[angle=180,width=6.5in]{fig5.ps}
630: \caption{{\sc TLUSTY} model compared with the MIKE spectrum of
631: LMC-SC1-105 at phase 0.27. The colored lines represent the sum of two
632: models shifted at the appropriate velocity, with the following (T$_{\rm
633: eff1}$,T$_{\rm eff2}$) pairs (in kK): (37.5, 35) in red (dotted line),
634: (35, 32.5) in green (solid line), and (32.5, 30) in blue (dashed
635: line). The green model fits most lines best.}
636: \label{fitmultioct23}
637: \end{figure}
638:
639: \begin{figure}[ht]
640: \includegraphics[angle=180,width=6.5in]{fig6.ps}
641: \caption{Same as Figure~\ref{fitmultioct23}, but for phase 0.75.}
642: \label{fitmultioct25}
643: \end{figure}
644:
645: \begin{figure}[ht]
646: \plotone{fig7.ps}
647: \caption{Phased OGLE $I-$band light curve of LMC-SC1-105. The best fit
648: model from PHOEBE (solid curve) assumes a semi-detached configuration
649: with the secondary filling its Roche lobe. The residuals suggest the
650: presence of an accretion stream and hot spots (not modeled), arising
651: from mass-transfer onto the primary.}
652: \label{lmc105i}
653: \end{figure}
654:
655: \begin{figure}[ht]
656: \plotone{fig8.ps}
657: \caption{Phased MACHO $r$ light curve of LMC-SC1-105 offset to match
658: OGLE $I-$band photometry; the best fit model from PHOEBE (solid curve)
659: is overplotted.}
660: \label{phoebemachorLC}
661: \end{figure}
662:
663: \begin{figure}[ht]
664: \plotone{fig9.ps}
665: \caption{Phased MACHO $b$ light curve of LMC-SC1-105 offset to match
666: OGLE $I-$band photometry; the best fit model from PHOEBE (solid curve)
667: is overplotted.}
668: \label{phoebemachobLC}
669: \end{figure}
670:
671: \begin{figure}[ht]
672: \plotone{fig10.ps}
673: \caption{Radial velocity curve for LMC-SC1-105. The TODCOR measurements
674: are shown as filled circles for the primary and open circles for the
675: secondary; overplotted is the best fit model from PHOEBE, denoted by a
676: solid line for the primary and a dashed line for the secondary. Error
677: bars correspond to the rms of the orbital fit: 11.8 $\rm km\; s^{-1}$
678: for the primary and 8.4 $\rm km\; s^{-1}$ for the secondary.}
679: \label{lmc105rv}
680: \end{figure}
681:
682: \begin{figure}[ht]
683: \plotone{fig11.ps}
684: \caption{Histogram of the results of the heuristic scan for the 4
685: parameters fit for in the semi-detached mode with PHOEBE. The average
686: and standard deviation of each parameter is labeled. The axis for $P$
687: has been scaled for display purposes. The final value for T$_{\rm eff2}$
688: was adopted from the spectral type calibration.}
689: \label{lmc105phoebe}
690: \end{figure}
691:
692: \begin{figure}[ht]
693: \plotone{fig12.ps}
694: \caption{Comparison of the parameters of LMC-SC1-105 with evolutionary
695: tracks (solid lines) and isochrones for single stars at Z=0.008
696: \citep{Schaerer93}. The dotted lines, from left to right, correspond to
697: 1, 3, 5, 7 and 10 Myr isochrones. Both components are overluminous for
698: their masses. The cooler, lower mass secondary appears older than the
699: primary, indicating that the system has undergone mass-transfer.}
700: \label{lumteff105}
701: \end{figure}
702:
703: \begin{figure}[ht]
704: \plotone{fig13.ps}
705: \caption{Comparison of the parameters of LMC-SC1-105 with evolutionary
706: tracks and isochrones for single stars at Z=0.007 \citep{Claret06}. The
707: dotted lines, from the bottom up, correspond to 1, 3, 5, 7 and 10 Myr
708: isochrones. Single star isochrones are not compatible with the measured
709: parameters for the system, which has undergone mass-transfer.}
710: \label{radmass105}
711: \end{figure}
712:
713:
714:
715: %%%%%%%%%%%%%%%% BIBLIOGRAPHY %%%%%%%%%%%%%%%%%%%%%%%%
716: \clearpage
717: \begin{thebibliography}{85}
718: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
719:
720: \bibitem[{{Andersen}(1991)}]{Andersen91}
721: {Andersen}, J. 1991, \aapr, 3, 91
722:
723: \bibitem[{{Bagnuolo} {et~al.}(1999){Bagnuolo}, {Gies}, {Riddle}, \&
724: {Penny}}]{Bagnuolo99}
725: {Bagnuolo}, Jr., W.~G., {Gies}, D.~R., {Riddle}, R., \& {Penny}, L.~R. 1999,
726: \apj, 527, 353
727:
728: \bibitem[{{Bally} \& {Zinnecker}(2005)}]{Bally05}
729: {Bally}, J. \& {Zinnecker}, H. 2005, \aj, 129, 2281
730:
731: \bibitem[{{Bernstein} {et~al.}(2003)}]{Bernstein03}
732: {Bernstein}, R. {et~al.} 2003, in SPIE, Vol. 4841, pp. 1694-1704, ed. M.~{Iye}
733: \& A.~F.~M. {Moorwood}
734:
735: \bibitem[{{Bonanos}(2007)}]{Bonanos07}
736: {Bonanos}, A.~Z. 2007, \aj, 133, 2696
737:
738: \bibitem[{{Bonanos} {et~al.}(2006){Bonanos}, {Stanek}, {Kudritzki},
739: {et~al.}}]{Bonanos06}
740: {Bonanos}, A.~Z., {Stanek}, K.~Z., {Kudritzki}, R.~P., {et~al.} 2006, \apj,
741: 652, 313
742:
743: \bibitem[{{Bonanos} {et~al.}(2004){Bonanos}, {Stanek}, {Udalski},
744: {et~al.}}]{Bonanos04}
745: {Bonanos}, A.~Z., {Stanek}, K.~Z., {Udalski}, A., {et~al.} 2004, \apjl, 611,
746: L33
747:
748: \bibitem[{{Bouret} {et~al.}(2003){Bouret}, {Lanz}, {Hillier},
749: {et~al.}}]{Bouret03}
750: {Bouret}, J.-C., {Lanz}, T., {Hillier}, D.~J., {et~al.} 2003, \apj, 595, 1182
751:
752: \bibitem[{{Bromm} \& {Loeb}(2006)}]{Bromm06}
753: {Bromm}, V. \& {Loeb}, A. 2006, \apj, 642, 382
754:
755: \bibitem[{{Burkholder} {et~al.}(1997){Burkholder}, {Massey}, \&
756: {Morrell}}]{Burkholder97}
757: {Burkholder}, V., {Massey}, P., \& {Morrell}, N. 1997, \apj, 490, 328
758:
759: \bibitem[{{Claret}(2000)}]{Claret00}
760: {Claret}, A. 2000, \aap, 363, 1081
761:
762: \bibitem[{{Claret}(2006)}]{Claret06}
763: ---. 2006, \aap, 453, 769
764:
765: \bibitem[{{de Mink} {et~al.}(2007){de Mink}, {Pols}, \& {Hilditch}}]{deMink07}
766: {de Mink}, S.~E., {Pols}, O.~R., \& {Hilditch}, R.~W. 2007, \aap, 467, 1181
767:
768: \bibitem[{{Derekas} {et~al.}(2007){Derekas}, {Kiss}, \& {Bedding}}]{Derekas07}
769: {Derekas}, A., {Kiss}, L.~L., \& {Bedding}, T.~R. 2007, \apj, 663, 249
770:
771: \bibitem[{{Elmegreen}(2000)}]{Elmegreen00}
772: {Elmegreen}, B.~G. 2000, \apj, 539, 342
773:
774: \bibitem[{{Faccioli} {et~al.}(2007){Faccioli}, {Alcock}, {Cook},
775: {et~al.}}]{Faccioli07}
776: {Faccioli}, L., {Alcock}, C., {Cook}, K., {et~al.} 2007, \aj, 134, 1963
777:
778: \bibitem[{{Fern{\'a}ndez Laj{\'u}s} \& {Niemela}(2006)}]{Fernandez06}
779: {Fern{\'a}ndez Laj{\'u}s}, E. \& {Niemela}, V.~S. 2006, \mnras, 367, 1709
780:
781: \bibitem[{{Figer}(2005)}]{Figer05}
782: {Figer}, D.~F. 2005, \nat, 434, 192
783:
784: \bibitem[{{Fitzpatrick} {et~al.}(2003){Fitzpatrick}, {Ribas}, {Guinan},
785: {et~al.}}]{Fitzpatrick03}
786: {Fitzpatrick}, E.~L., {Ribas}, I., {Guinan}, E.~F., {et~al.} 2003, \apj, 587,
787: 685
788:
789: \bibitem[{{Fryer} {et~al.}(2007){Fryer}, {Mazzali}, {Prochaska},
790: {et~al.}}]{Fryer07}
791: {Fryer}, C.~L., {Mazzali}, P.~A., {Prochaska}, J., {et~al.} 2007, \pasp, 119,
792: 1211
793:
794: \bibitem[{{Gies}(2003)}]{Gies03}
795: {Gies}, D.~R. 2003, in IAU Symposium, Vol. 212, A Massive Star Odyssey: From
796: Main Sequence to Supernova, ed. K.~{van der Hucht}, A.~{Herrero}, \&
797: C.~{Esteban}, 91--+
798:
799: \bibitem[{{Gonz{\'a}lez} {et~al.}(2005){Gonz{\'a}lez}, {Ostrov}, {Morrell}, \&
800: {Minniti}}]{Gonzalez05}
801: {Gonz{\'a}lez}, J.~F., {Ostrov}, P., {Morrell}, N., \& {Minniti}, D. 2005,
802: \apj, 624, 946
803:
804: \bibitem[{{Guinan} {et~al.}(1998){Guinan}, {Fitzpatrick}, {Dewarf},
805: {et~al.}}]{Guinan98}
806: {Guinan}, E.~F., {Fitzpatrick}, E.~L., {Dewarf}, L.~E., {et~al.} 1998, \apjl,
807: 509, L21
808:
809: \bibitem[{{Hadrava}(1995)}]{Hadrava95}
810: {Hadrava}, P. 1995, \aaps, 114, 393
811:
812: \bibitem[{{Harries} {et~al.}(2003){Harries}, {Hilditch}, \&
813: {Howarth}}]{Harries03}
814: {Harries}, T.~J., {Hilditch}, R.~W., \& {Howarth}, I.~D. 2003, \mnras, 339, 157
815:
816: \bibitem[{{Herrero}(2007)}]{Herrero07}
817: {Herrero}, A. 2007, RevMexAA in press (arxiv:0704.3528)
818:
819: \bibitem[{{Hilditch}(2001)}]{Hilditch01}
820: {Hilditch}, R.~W. 2001, {An Introduction to Close Binary Stars} (An
821: Introduction to Close Binary Stars, by R.W.~Hilditch.~ Cambridge University
822: Press, 2001, 392 pp.)
823:
824: \bibitem[{{Hilditch} {et~al.}(2005){Hilditch}, {Howarth}, \&
825: {Harries}}]{Hilditch05}
826: {Hilditch}, R.~W., {Howarth}, I.~D., \& {Harries}, T.~J. 2005, \mnras, 357, 304
827:
828: \bibitem[{{Hillier} \& {Miller}(1998)}]{Hillier98}
829: {Hillier}, D.~J. \& {Miller}, D.~L. 1998, \apj, 496, 407
830:
831: \bibitem[{{Howarth} {et~al.}(1997){Howarth}, {Siebert}, {Hussain},
832: {et~al.}}]{Howarth97}
833: {Howarth}, I.~D., {Siebert}, K.~W., {Hussain}, G.~A.~J., {et~al.} 1997, \mnras,
834: 284, 265
835:
836: \bibitem[{{Kelson}(2003)}]{Kelson03}
837: {Kelson}, D.~D. 2003, \pasp, 115, 688
838:
839: \bibitem[{{Kelson} {et~al.}(2000)}]{Kelson00}
840: {Kelson}, D.~D. {et~al.} 2000, \apj, 531, 159
841:
842: \bibitem[{{Krumholz} \& {Thompson}(2007)}]{Krumholz07}
843: {Krumholz}, M.~R. \& {Thompson}, T.~A. 2007, \apj, 661, 1034
844:
845: \bibitem[{{Lanz} \& {Hubeny}(2003)}]{Lanz03}
846: {Lanz}, T. \& {Hubeny}, I. 2003, \apjs, 146, 417
847:
848: \bibitem[{{Linder} {et~al.}(2007){Linder}, {Rauw}, {Sana}, {et~al.}}]{Linder07}
849: {Linder}, N., {Rauw}, G., {Sana}, H., {et~al.} 2007, \aap, 474, 193
850:
851: \bibitem[{{Lucy}(2006)}]{Lucy06}
852: {Lucy}, L.~B. 2006, \aap, 457, 629
853:
854: \bibitem[{{Ma{\'{\i}}z Apell{\'a}niz} {et~al.}(2007){Ma{\'{\i}}z
855: Apell{\'a}niz}, {Walborn}, {Morrell}, {et~al.}}]{Maiz-Apellaniz07}
856: {Ma{\'{\i}}z Apell{\'a}niz}, J., {Walborn}, N.~R., {Morrell}, N.~I., {et~al.}
857: 2007, \apj, 660, 1480
858:
859: \bibitem[{{Martins} {et~al.}(2005){Martins}, {Schaerer}, \&
860: {Hillier}}]{Martins05}
861: {Martins}, F., {Schaerer}, D., \& {Hillier}, D.~J. 2005, \aap, 436, 1049
862:
863: \bibitem[{{Massey}(2003)}]{Massey03}
864: {Massey}, P. 2003, \araa, 41, 15
865:
866: \bibitem[{{Massey} {et~al.}(2002){Massey}, {Penny}, \& {Vukovich}}]{Massey02}
867: {Massey}, P., {Penny}, L.~R., \& {Vukovich}, J. 2002, \apj, 565, 982
868:
869: \bibitem[{{Massey} {et~al.}(2005){Massey}, {Puls}, {Pauldrach},
870: {et~al.}}]{Massey05}
871: {Massey}, P., {Puls}, J., {Pauldrach}, A.~W.~A., {et~al.} 2005, \apj, 627, 477
872:
873: \bibitem[{{Massey} {et~al.}(2000){Massey}, {Waterhouse}, \&
874: {DeGioia-Eastwood}}]{Massey00}
875: {Massey}, P., {Waterhouse}, E., \& {DeGioia-Eastwood}, K. 2000, \aj, 119, 2214
876:
877: \bibitem[{{Mayer} {et~al.}(2005){Mayer}, {Drechsel}, \& {Lorenz}}]{Mayer05}
878: {Mayer}, P., {Drechsel}, H., \& {Lorenz}, R. 2005, \apjs, 161, 171
879:
880: \bibitem[{{Mayer} {et~al.}(2008){Mayer}, {Harmanec}, {Nesslinger},
881: {et~al.}}]{Mayer08}
882: {Mayer}, P., {Harmanec}, P., {Nesslinger}, S., {et~al.} 2008, \aap, 481, 183
883:
884: \bibitem[{{Mazeh} {et~al.}(2006){Mazeh}, {Tamuz}, \& {North}}]{Mazeh06}
885: {Mazeh}, T., {Tamuz}, O., \& {North}, P. 2006, \mnras, 367, 1531
886:
887: \bibitem[{{Meynet} \& {Maeder}(2003)}]{Meynet03}
888: {Meynet}, G. \& {Maeder}, A. 2003, \aap, 404, 975
889:
890: \bibitem[{{Mochnacki} \& {Doughty}(1972)}]{Mochnacki72}
891: {Mochnacki}, S.~W. \& {Doughty}, N.~A. 1972, \mnras, 156, 51
892:
893: \bibitem[{{Mokiem} {et~al.}(2007){Mokiem}, {de Koter}, {Evans},
894: {et~al.}}]{Mokiem07}
895: {Mokiem}, M.~R., {de Koter}, A., {Evans}, C.~J., {et~al.} 2007, \aap, 465, 1003
896:
897: \bibitem[{{Nelson} \& {Eggleton}(2001)}]{Nelson01}
898: {Nelson}, C.~A. \& {Eggleton}, P.~P. 2001, \apj, 552, 664
899:
900: \bibitem[{{Niemela} {et~al.}(2006){Niemela}, {Morrell}, {Fern{\'a}ndez
901: Laj{\'u}s}, {et~al.}}]{Niemela06}
902: {Niemela}, V.~S., {Morrell}, N.~I., {Fern{\'a}ndez Laj{\'u}s}, E., {et~al.}
903: 2006, \mnras, 367, 1450
904:
905: \bibitem[{{O'Connell}(1951)}]{O'Connell51}
906: {O'Connell}, D.~J.~K. 1951, Publications of the Riverview College Observatory,
907: 2, 85
908:
909: \bibitem[{{Oey} \& {Clarke}(2005)}]{Oey05}
910: {Oey}, M.~S. \& {Clarke}, C.~J. 2005, \apjl, 620, L43
911:
912: \bibitem[{{Orosz} {et~al.}(2007){Orosz}, {McClintock}, {Narayan},
913: {et~al.}}]{Orosz07}
914: {Orosz}, J.~A., {McClintock}, J.~E., {Narayan}, R., {et~al.} 2007, \nat, 449,
915: 872
916:
917: \bibitem[{{Ostrov}(2001)}]{Ostrov01c}
918: {Ostrov}, P.~G. 2001, \mnras, 321, L25
919:
920: \bibitem[{{Ostrov} \& {Lapasset}(2003)}]{Ostrov03}
921: {Ostrov}, P.~G. \& {Lapasset}, E. 2003, \mnras, 338, 141
922:
923: \bibitem[{{Ostrov} {et~al.}(2001){Ostrov}, {Morrell}, \&
924: {Lapasset}}]{Ostrov01b}
925: {Ostrov}, P.~G., {Morrell}, N.~I., \& {Lapasset}, E. 2001, \aap, 377, 972
926:
927: \bibitem[{{Petrovic} {et~al.}(2005){Petrovic}, {Langer}, \& {van der
928: Hucht}}]{Petrovic05}
929: {Petrovic}, J., {Langer}, N., \& {van der Hucht}, K.~A. 2005, \aap, 435, 1013
930:
931: \bibitem[{{Pinsonneault} \& {Stanek}(2006)}]{Pinsonneault06}
932: {Pinsonneault}, M.~H. \& {Stanek}, K.~Z. 2006, \apjl, 639, L67
933:
934: \bibitem[{{Pr{\v s}a} \& {Zwitter}(2005)}]{Prsa05}
935: {Pr{\v s}a}, A. \& {Zwitter}, T. 2005, \apj, 628, 426
936:
937: \bibitem[{{Puls} {et~al.}(2005){Puls}, {Urbaneja}, {Venero}, {et~al.}}]{Puls05}
938: {Puls}, J., {Urbaneja}, M.~A., {Venero}, R., {et~al.} 2005, \aap, 435, 669
939:
940: \bibitem[{{Pych}(2004)}]{Pych04}
941: {Pych}, W. 2004, \pasp, 116, 148
942:
943: \bibitem[{{Rauw} {et~al.}(2004){Rauw}, {De Becker}, {Naze}, {et~al.}}]{Rauw04}
944: {Rauw}, G., {De Becker}, M., {Naze}, Y., {et~al.} 2004, \aap, 420, L9
945:
946: \bibitem[{{Repolust} {et~al.}(2004){Repolust}, {Puls}, \&
947: {Herrero}}]{Repolust04}
948: {Repolust}, T., {Puls}, J., \& {Herrero}, A. 2004, \aap, 415, 349
949:
950: \bibitem[{{Ribas} {et~al.}(2005){Ribas}, {Jordi}, {Vilardell},
951: {et~al.}}]{Ribas05}
952: {Ribas}, I., {Jordi}, C., {Vilardell}, F., {et~al.} 2005, \apjl, 635, L37
953:
954: \bibitem[{{Santolaya-Rey} {et~al.}(1997){Santolaya-Rey}, {Puls}, \&
955: {Herrero}}]{Santolaya97}
956: {Santolaya-Rey}, A.~E., {Puls}, J., \& {Herrero}, A. 1997, \aap, 323, 488
957:
958: \bibitem[{{Schaerer} {et~al.}(1993){Schaerer}, {Charbonnel}, {Meynet},
959: {et~al.}}]{Schaerer93}
960: {Schaerer}, D., {Charbonnel}, C., {Meynet}, G., {et~al.} 1993, \aaps, 102, 339
961:
962: \bibitem[{{Schaller} {et~al.}(1992){Schaller}, {Schaerer}, {Meynet}, \&
963: {Maeder}}]{Schaller92}
964: {Schaller}, G., {Schaerer}, D., {Meynet}, G., \& {Maeder}, A. 1992, \aaps, 96,
965: 269
966:
967: \bibitem[{{Schnurr} {et~al.}(2008){Schnurr}, {Casoli}, {Chen{\'e}},
968: {et~al.}}]{Schnurr08}
969: {Schnurr}, O., {Casoli}, J., {Chen{\'e}}, A.-N., {et~al.} 2008, \mnras, 389,
970: L38
971:
972: \bibitem[{{Schwarzenberg-Czerny}(1989)}]{Schwarzenberg89}
973: {Schwarzenberg-Czerny}, A. 1989, \mnras, 241, 153
974:
975: \bibitem[{{Simon} \& {Sturm}(1994)}]{Simon94}
976: {Simon}, K.~P. \& {Sturm}, E. 1994, \aap, 281, 286
977:
978: \bibitem[{{Soszynski} {et~al.}(2008){Soszynski}, {Poleski}, {Udalski},
979: {et~al.}}]{Soszynski08}
980: {Soszynski}, I., {Poleski}, R., {Udalski}, A., {et~al.} 2008, AcA, submitted
981: (astro-ph/0808.2210)
982:
983: \bibitem[{{Southworth} \& {Clausen}(2007)}]{Southworth07}
984: {Southworth}, J. \& {Clausen}, J.~V. 2007, \aap, 461, 1077
985:
986: \bibitem[{{Stickland}(1997)}]{Stickland97}
987: {Stickland}, D.~J. 1997, The Observatory, 117, 37
988:
989: \bibitem[{{Tonry} \& {Davis}(1979)}]{Tonry79}
990: {Tonry}, J. \& {Davis}, M. 1979, \aj, 84, 1511
991:
992: \bibitem[{{Vitrichenko} {et~al.}(2007){Vitrichenko}, {Nadyozhin}, \&
993: {Razinkova}}]{Vitrichenko07}
994: {Vitrichenko}, E.~A., {Nadyozhin}, D.~K., \& {Razinkova}, T.~L. 2007, Astronomy
995: Letters, 33, 251
996:
997: \bibitem[{{Walborn} \& {Fitzpatrick}(1990)}]{Walborn90}
998: {Walborn}, N.~R. \& {Fitzpatrick}, E.~L. 1990, \pasp, 102, 379
999:
1000: \bibitem[{{Walborn} {et~al.}(2002){Walborn}, {Howarth}, {Lennon},
1001: {et~al.}}]{Walborn02}
1002: {Walborn}, N.~R., {Howarth}, I.~D., {Lennon}, D.~J., {et~al.} 2002, \aj, 123,
1003: 2754
1004:
1005: \bibitem[{{Wilson}(1979)}]{Wilson79}
1006: {Wilson}, R.~E. 1979, \apj, 234, 1054
1007:
1008: \bibitem[{{Wilson}(1990)}]{Wilson90}
1009: ---. 1990, \apj, 356, 613
1010:
1011: \bibitem[{{Wilson} \& {Devinney}(1971)}]{Wilson71}
1012: {Wilson}, R.~E. \& {Devinney}, E.~J. 1971, \apj, 166, 605
1013:
1014: \bibitem[{{Wyrzykowski} {et~al.}(2003){Wyrzykowski}, {Udalski}, {Kubiak},
1015: {et~al.}}]{Wyrzykowski03}
1016: {Wyrzykowski}, L., {Udalski}, A., {Kubiak}, M., {et~al.} 2003, Acta
1017: Astronomica, 53, 1
1018:
1019: \bibitem[{{Yungelson} {et~al.}(2008){Yungelson}, {van den Heuvel}, {Vink},
1020: {et~al.}}]{Yungelson08}
1021: {Yungelson}, L.~R., {van den Heuvel}, E.~P.~J., {Vink}, J.~S., {et~al.} 2008,
1022: \aap, 477, 223
1023:
1024: \bibitem[{{Zebrun} {et~al.}(2001){Zebrun}, {Soszynski}, {Wozniak},
1025: {et~al.}}]{Zebrun01}
1026: {Zebrun}, K., {Soszynski}, I., {Wozniak}, P.~R., {et~al.} 2001, Acta
1027: Astronomica, 51, 317
1028:
1029: \bibitem[{{Zinnecker} \& {Yorke}(2007)}]{Zinnecker07}
1030: {Zinnecker}, H. \& {Yorke}, H.~W. 2007, \araa, 45, 481
1031:
1032: \bibitem[{{Zucker} \& {Mazeh}(1994)}]{Zucker94}
1033: {Zucker}, S. \& {Mazeh}, T. 1994, \apj, 420, 806
1034:
1035: \end{thebibliography}
1036:
1037:
1038: \clearpage
1039: \input{tab1.tex}
1040: \clearpage
1041: \input{tab2.tex}
1042: \input{tab3.tex}
1043: \input{tab4.tex}
1044: \input{tab5.tex}
1045: \end{document}
1046:
1047: