1: \documentclass[12pt,preprint]{aastex}
2: \begin{document}
3: \newcommand{\msun}{\mbox{M$_{\odot}$}}
4: \newcommand{\rsun}{\mbox{R$_{\odot}$}}
5: \title{WR~20a is an Eclipsing Binary: Accurate Determination of Parameters
6: for an Extremely Massive Wolf-Rayet System\altaffilmark{1}}
7:
8: \author{A. Z. Bonanos, K. Z. Stanek}
9: \affil{Harvard-Smithsonian Center for Astrophysics, 60 Garden St.,
10: Cambridge, MA~02138}
11: \affil{\tt e-mail: abonanos@cfa.harvard.edu, kstanek@cfa.harvard.edu}
12:
13: \author{A.~Udalski, L.~Wyrzykowski\altaffilmark{2}, K.~\.Zebru\'n,
14: M.~Kubiak, M.~K.~Szyma\'nski, O.~Szewczyk,
15: G.~Pietrzy\'nski\altaffilmark{3}, I.~Soszy\'nski}
16:
17: \affil{Warsaw University Observatory, Al. Ujazdowskie 4, 00-478
18: Warszawa, Poland}
19: \affil{\tt e-mail: \{udalski, wyrzykow, zebrun, mk, msz,
20: szewczyk, pietrzyn, soszynsk\}@astrouw.edu.pl}
21:
22:
23: \altaffiltext{1}{Based on observations obtained with the 1.3 m Warsaw
24: telescope at Las Campanas Observatory, which is operated by the
25: Carnegie Institute of Washington.}
26: \altaffiltext{2}{School of Physics and Astronomy and Wise Observatory,
27: Tel-Aviv University, Tel~Aviv~69978, Israel}
28: \altaffiltext{3}{Universidad de Concepci{\'o}n, Departamento de Fisica,
29: Casilla 160-C, Concepci{\'o}n, Chile}
30:
31: \begin{abstract}
32:
33: We present a high-precision $I$-band light curve for the Wolf-Rayet
34: binary WR~20a, obtained as a sub-project of the Optical Gravitational
35: Lensing Experiment. Rauw et al. have recently presented spectroscopy
36: for this system, strongly suggesting extremely large minimum masses of
37: $70.7 \pm 4.0\; \msun$ and $68.8 \pm 3.8\; \msun$ for the component
38: stars of the system, with the exact values depending strongly on the
39: period of the system. We detect deep eclipses of about $0.4\;$mag in
40: the light curve of WR~20a, confirming and refining the suspected
41: period of $P=3.686$ days and deriving an inclination angle of
42: $i=74.5\pm 2.0\deg$. Using these photometric data and the radial
43: velocity data of Rauw et al., we derive the masses for the two
44: components of WR~20a to be $83.0 \pm 5.0\; \msun$ and $82.0 \pm 5.0\;
45: \msun$. Therefore, WR~20a is confirmed to consist of two extremely
46: massive stars and to be the most massive binary known with an accurate
47: mass determination.
48:
49: \end{abstract}
50: \keywords{binaries: eclipsing -- binaries: spectroscopic -- stars:
51: Wolf-Rayet -- stars: fundamental parameters -- stars: individual:
52: WR~20a}
53:
54:
55: \section{Introduction}
56:
57: Measuring accurate masses for the most massive stars in our Galaxy and
58: beyond is important for constraining star formation and stellar
59: evolution theories, which have indirect implications for many objects
60: that are not well understood, such as supernovae, gamma-ray bursts and
61: Population III stars. The most massive candidates in the Milky Way are
62: LBV 1806-20 \citep{Eikenberry04} and the Pistol Star \citep{Figer98},
63: which have inferred masses up to $\sim 200\;$\msun. HDE~269810 in the
64: LMC is another massive candidate $\sim 150\;\rm \msun$
65: \citep{Walborn04}. Finally, $\eta$ Carinae, one of the most studied
66: massive stars has an estimated mass of $\sim 120\;$\msun, yet it
67: remains unknown whether it is a rapid rotator or a close binary system
68: \citep[see review by][]{Davidson97}. However, all these masses are
69: only indirect estimates and thus have large uncertainties associated
70: with them. The only direct way of measuring accurate masses of distant
71: stars is through double lined spectroscopic binary systems with
72: eclipses present in their light curves. The limited number of existing
73: mass measurements for massive stars ($>20\; \msun$) can be explained
74: by their intrinsic rarity and the observationally demanding process of
75: discovering them in eclipsing binary systems.
76:
77: Until recently, the most massive stars ever weighed in binaries were:
78: R136-38 (O3V+O6V) in the LMC with a primary mass of $56.9\pm 0.6
79: \;\msun$ \citep{Massey02}, WR 22 (WN7+abs$\;+\;$O), with a minimum
80: primary mass of $55.3\pm 7.3\; \msun$ \citep{Rauw96,Schweickhardt99}
81: and Plaskett's star with a minimum primary mass of $51\; \msun$
82: \citep{Bagnuolo92}. However, WR~20a, a Wolf-Rayet (WR) binary in the
83: compact cluster Westerlund~2, seems to be the new record
84: holder. \citet{Rauw04} obtained spectroscopy for WR~20a and measured
85: extremely large minimum masses of $70.7 \pm 4.0$ and $68.8 \pm
86: 3.8\;\msun$ for the components. The final masses strongly depend on
87: both the period and inclination of the binary, which can only be
88: measured from the light curve. They derived a period of $3.675\;$days
89: from the absolute values of the radial velocity differences and
90: assumed a circular orbit to derive the masses of the components. In
91: their discussion of WR~20a, \citet{Rauw04} stressed the importance and
92: necessity of photometric monitoring of this binary.
93:
94: \citet{Shara91} were the first to make the discovery that WR~20a was a
95: WR star, by obtaining spectroscopy of the system. \citet{Moffat91}
96: obtained $UBV$ photometry, however, as \citet{Rauw04} suggest, it is
97: likely affected by photometric eclipses. WR~20a was classified as a
98: candidate binary along with 30 other WR stars in the VIIth Catalogue
99: of Galactic Wolf-Rayet Stars \citep{vanderHucht01}, because of its
100: comparatively weak emission lines, which were thought to be diluted by
101: the continuum light of an OB-type companion. \citet{Rauw04}
102: spectroscopically surveyed these candidate WR binaries and WR~20a
103: clearly emerged as an interesting object, which they followed up. Due
104: to its favorable location in the sky, this object was included in the
105: target list covered by the Optical Gravitational Lensing Experiment
106: (OGLE) photometric survey. We have quickly obtained a precise light
107: curve of WR~20a and thus confirm and refine, by measuring the period
108: and inclination, the remarkable masses of the WR~20a components.
109:
110: In this paper, we present photometry for WR~20a, which clearly shows
111: the presence of eclipses. In Section 2 we describe the observations,
112: in Section 3 we present the light curve and the analysis and in
113: Section 4 we discuss the results.
114:
115: \section{OGLE Observations}
116:
117: WR~20a $\rm(=SMSP2=THA 35-II-036)$ is located in the constellation
118: Carina at $\alpha=10$:23:58.0, $\delta= -57$:45:49 (J2000.0).
119: Observations were carried out with the 1.3 m Warsaw telescope at Las
120: Campanas Observatory, Chile, which is operated by the Carnegie
121: Institute of Washington. The telescope is equipped with a mosaic CCD
122: camera with 8192$\times$8192 pixels. WR~20a is very bright, $I\sim
123: 11.0$, so very short exposures of 10 seconds each were used to avoid
124: saturating the CCD. The observations presented here started on May 1,
125: 2004 and consist of 83 measurements obtained on 17 consecutive nights.
126:
127: The absolute calibration of our photometry was obtained on one
128: photometric night with a couple of standards from Landolt (1992). The
129: uncertainty in the zero points of the photometry is about 0.03 mag.
130:
131: The data were reduced using the standard OGLE-III data pipeline, as
132: described by \citet{Udalski03}. The photometry was derived using
133: difference imaging analysis. This method is the current
134: state-of-the-art for photometric accuracy in crowded fields
135: \citep{Alard00}. Our analysis resulted in a set of time-series
136: photometry in the $I$-band, presented in Table~1. The formal error in
137: the photometry is as low as $0.001-0.002\;$mag, but we have adopted a
138: conservative error of $0.01\;$mag for the differential light
139: curve. $V$-band images were also obtained on one night during
140: maximum. We measured the following magnitudes for WR~20a at maximum:
141: $V=13.45 \pm 0.025$, and $I=10.65 \pm 0.05$.
142:
143: \section{Light Curve and Radial Velocity Curve Analysis}
144:
145: The first goal of our observations was to confirm and refine the
146: orbital period of the WR~20a binary. As discussed extensively by
147: \citet{Rauw04}, besides the main period of $P=3.675\;$days, their
148: spectroscopic data also allowed for a much shorter period of
149: $1.293\;$days, which would result in reduced minimum masses of about
150: $25\; \msun$ for each star in the system. A longer period of
151: $4.419\;$days was also possible, resulting in increased minimum masses
152: of above $80 \;\msun$ for each star. Using the analysis of variance
153: technique of \citet{Schwarzenberg89}, the most significant period
154: derived from our data is $P=1.843 \pm 0.005$ days, where the error is
155: a conservative 3$\sigma$ estimate. However, we know that the system
156: has to have two eclipses, which results in a true orbital period twice
157: as long. We find an ephemeris for the primary (deeper) eclipse of
158:
159: \begin{equation}
160: \rm T_{prim}=2453124.569 + 3.686 \times E,
161: \end{equation}
162:
163: \noindent thus confirming and refining the orbital period of WR~20a
164: and the remarkable masses of its components.
165:
166: In addition, we can derive an inclination angle $i$ for the system
167: from our well-sampled light curve. To first order, the value of the
168: inclination angle is given by the depth of the eclipses ($\sim 0.4$
169: mag), and we find that the exact value of $\;i\;$ only weakly depends
170: on the details of the model fit, or even which eclipsing binary model
171: we use to fit. To demonstrate this, we first fit the light curve with
172: a simple model of two spherical stars with limb darkening (J. Devor
173: 2004, private communication), as well as with the more complex EBOP
174: model \citep{Nelson72,Popper81} and finally, with the Wilson-Devinney
175: (WD) code \citep{Wilson71,Wilson79,vanHamme03} for modeling distorted
176: stars. We find that indeed the inclination angle is insensitive to
177: the model, and the best fit value with WD is $i=74.5\pm 2.0\deg$.
178:
179: We ran WD in the overcontact mode (MODE 3), fixing T$_{\rm
180: eff1}=42\,000$K \citep[since WR 20a is intermediate between Mk42 and
181: WR 47c, see][]{Rauw04,Crowther98}, using linear limb darkening
182: coefficients from \citet{vanHamme03}, values for gravity darkening
183: exponents and albedos from theoretical values for radiative envelopes,
184: and a mass ratio determined from the radial velocity curve (see
185: below). We fit for the inclination $i$, T$_{\rm eff2}$, the luminosity
186: of the primary and the surface potential ($\Omega_{1}=\Omega_{2}$) and
187: defined convergence to be reached after five consecutive iterations
188: for which the corrections for all adjusted parameters were smaller
189: than their respective standard (statistical) errors. The results of
190: the fit are given in Table~2. In Figure~\ref{lc} we show the result of
191: the $I-$band light curve model fit for WR~20a. The uneven eclipse
192: depths suggest slightly different effective temperatures and, thus,
193: different spectral types for the components. Using differential
194: photometry we can measure a difference between the two stars as small
195: as 1-2\%, which is not possible with spectral decomposition. However,
196: multi-band photometry is necessary to resolve the effective
197: temperature -- radius degeneracy. Finally, we derive the following
198: best fit fractional radii for both stars: polar radius $\rm
199: r_{pole}=0.34$, radius to the inner Lagrangian point $\rm
200: r_{point}=0.40$, $\rm r_{side}=0.35$, and $\rm r_{back}=0.37$.
201:
202: Armed with the refined period and the exact value of the inclination
203: angle, we re-analyzed the radial velocity data of
204: \citet{Rauw04}. Fixing the eccentricity to $0$, as the light curve
205: confirms, and the period to $3.686$ days, we fit the radial velocity
206: data applying equal weights and rejecting points with radial velocity
207: measurements smaller than $80\;\rm km\;s^{-1}$. These measurements
208: were made by fitting Gaussians to lines that are not resolved and thus
209: include errors much larger than $10-15\;\rm km\; s^{-1}$ and are also
210: subject to non-Keplerian effects. Our fit is shown in Figure~\ref{rv},
211: and the best fit values with their statistical errors are presented in
212: Table~3. We note that due to the refined period, as \citet{Rauw04}
213: caution, their two earliest RV epochs are now ``flipped''. We derive
214: slightly larger velocity semi-amplitudes than \citet{Rauw04}, which in
215: turn yield larger masses. Specifically, the values are $\rm
216: K_{1}=362.2\;\rm km\; s^{-1}$ and $\rm K_{2}=366.4\;\rm km\; s^{-1}$,
217: the mass ratio $\rm q=m_{2}/m_{1}=0.99$, the systemic velocity
218: $\gamma=8\; \rm km\; s^{-1}$, the semi-major axis $\rm a \;sin\, \it
219: i=\rm 53\; \rsun$, and finally the masses $\rm m_{1}\;sin^{3}\, \it
220: i=\rm 74.3\; \msun$ and $\rm m_{2}\;sin^{3}\, \it i= \rm 73.4\;
221: \msun$. Knowing the inclination, we calculate the masses of the stars
222: to be $83.0 \pm 5.0\; \msun$ and $82.0 \pm 5.0\; \msun$. These erros
223: include the error in the minimum mass and the error in the
224: inclination.
225:
226: \begin{figure}[h]
227: \plotone{lc.epsi}
228: \caption{Wilson-Devinney model fit of a near contact binary to the
229: $I$-band light curve of WR~20a. The period is $3.686$ days, the
230: eccentricity is $0$ and the inclination $i=74.5 \deg$.}
231: \label{lc}
232: \end{figure}
233:
234: \begin{figure}[h]
235: \plotone{rv.epsi}
236: \caption{Radial velocity curve fit to measurements of WR~20a made by
237: \citet{Rauw04}. The filled circles correspond to the primary (more
238: massive) star. The new semi-amplitude values are $\rm K_{1}=362\;\rm
239: km\; s^{-1}$ and $\rm K_{2}=366\;\rm km\; s^{-1}$ and the masses are
240: $\rm m_{1}\,sin^{3}\, \it i=\rm 74.3\, \msun$ and $\rm
241: m_{2}\,sin^{3}\, \it i= \rm 73.4\, \msun$, which, combined with
242: $i=74.5\deg$ yield mass values for the components of $83.0\; \msun$
243: and $82.0\; \msun$.}
244: \label{rv}
245: \end{figure}
246:
247: \section{Discussion}
248:
249: As discussed at length by \citet{Rauw04}, the extreme mass values of
250: WR~20a make it a ``cornerstone system for future investigations of
251: massive star evolution''. We confirm and refine these extreme values,
252: by measuring the masses of the components to be $83.0 \pm 5.0\; \msun$
253: and $82.0 \pm 5.0\; \msun$. If the WN6ha spectral type is confirmed,
254: WR~20a will be one of few known WR binaries with a mass ratio near
255: unity. Thus, it becomes very important now to obtain high resolution
256: spectra and multiband photometry of WR 20a, corrected for eclipses.
257:
258: Such future observations will allow an accurate reddening measurement
259: and furthermore, a direct, independent and accurate distance
260: determination to the system. This will resolve the discrepancy in the
261: distance to Westerlund 2, which ranges from $2.5-8$ kpc \citep[see
262: discussion in][]{Churchwell04}. Several multiband measurements of
263: WR~20a already exist. As measured by us (Section 2), WR~20a has
264: $V=13.45$ and $I=10.65$ at maximum. Our $V$ is different from
265: $V=13.58$ measured by \citet{Moffat91}, but their measurement could
266: have been affected by the eclipses. The Two Micron All Sky Survey
267: \citep[{\it 2MASS\rm},][]{Skrutskie97} has measured $J=8.86, H=8.08$
268: and $K=7.59$ for WR~20a. Recent infrared observations of the HII
269: region RCW 49 with the $Spitzer\; Space\; Telescope$ \citep{Whitney04}
270: contain measurements of the binary. \citet{Belloni94} have obtained
271: $ROSAT$ observations of RCW 49, tentatively identifying WR 20a with
272: one of their point sources. However, further observations are needed
273: to determine the extinction to WR~20a.
274:
275: Analogs of WR~20a, possibly with even higher mass components, are
276: bound to exist in nearby galaxies, such as M31 and M33. There, they
277: would be first found by variability searches such as DIRECT
278: \citep{Stanek98,Bonanos03} as bright eclipsing binaries (about 18th
279: magnitude for $\rm M_{V}=-6.5$), which can then be followed by
280: spectroscopy to obtain radial velocities. It is our experience
281: (A.Z. Bonanos et al., in preparation) that accurate radial velocities
282: can be obtained for such systems with the existing 6.5-10 meter
283: telescopes. The advantage of studying eclipsing binaries in these
284: galaxies would be their well known distances and low reddening, as
285: systems with high reddening would be too faint to follow-up.
286:
287: We conclude that WR~20a truly is a ``cornerstone system'' that
288: deserves detailed additional studies. Not only will studies of the
289: most massive binaries constrain models, but also will help understand
290: other problems, such as the nature of mass loss and winds of early
291: type stars, as well as the connection of these stars to supernovae and
292: gamma ray bursts \citep{Stanek03,Maund04}. WR~20a now overtakes the
293: previous record holders WR 22 \citep{Rauw96,Schweickhardt99} and
294: R136-38 \citep{Massey02} in the most massive star competition, by
295: becoming ``the most massive star ever weighed.''
296:
297:
298: \acknowledgments{We thank G. Torres for a careful reading of and
299: comments on the manuscript and D. Sasselov and J. Devor for useful
300: discussions. We also thank the referee, Eric Gosset, for his prompt
301: and careful reading of the manuscript and his useful comments. AZB and
302: KZS were partially supported by HST Grant HST-GO-09810.06-A. Support
303: for Proposal number HST-GO-09810.06-A was provided by NASA through a
304: grant from the Space Telescope Science Institute, which is operated by
305: the Association of Universities for Research in Astronomy,
306: Incorporated, under NASA contract NAS5-26555. Partial support to the
307: OGLE project was provided with the following grants: Polish KBN grant
308: 2P03D02124 to A. Udalski and NSF grant AST-0204908 and NASA grant
309: NAG5-12212 to B.~Paczy{\'n}ski. }
310:
311:
312: %%%%%%%%%%%%%%%% BIBLIOGRAPHY %%%%%%%%%%%%%%%%%%%%%%%%
313: \clearpage
314: \bibliographystyle{apj}
315: \bibliography{ref}
316: \clearpage
317: \input{table1.tex}
318: \input{table2.tex}
319: \input{table3.tex}
320: \end{document}
321:
322: