The spin rates of O stars in WR + O binaries. I. Motivation, methodology and first results from SALT
Michael M. Shara, Steven M. Crawford, Dany Vanbeveren, Anthony F.J. Moffat, David Zurek, Lisa Crause
aa r X i v : . [ a s t r o - ph . S R ] A ug Mon. Not. R. Astron. Soc. , 000–000 (0000) Printed 11 October 2018 (MN L A TEX style file v2.2)
The spin rates of O stars in WR + O binaries. I.Motivation, methodology and first results from SALT
Michael M. Shara ⋆ , Steven M. Crawford , Dany Vanbeveren ,Anthony F. J. Moffat , David Zurek and Lisa Crause Department of Astrophysics, American Museum of Natural History, Central Park West at 79th Street, New York, NY 10024, USA South African Astronomical Observatory, P.O. Box 9, Observatory 7935, Cape Town, South Africa Astrophysical Institute, Vrije Universiteit Brussel, Pleinlaan 2, 1050, Brussels, Belgium D´epartement de Physique, Universit´e de Montr´eal, CP 6128 Succ. C-V, Montr´eal, QC H3C 3J7, Canada
Accepted Received
ABSTRACT
The black holes (BH) in merging BH-BH binaries are likely progeny of binaryO stars. Their properties, including their spins, will be strongly influenced by theevolution of their progenitor O stars. The remarkable observation that many singleO stars spin very rapidly can be explained if they accreted angular momentum froma mass-transferring, O-type or Wolf-Rayet companion before that star blew up asa supernova. To test this prediction, we have measured the spin rates of eight Ostars in Wolf-Rayet (WR) + O binaries, increasing the total sample size of such Ostars’ measured spins from two to ten. Polarimetric and other determinations of thesesystems’ sin i allow us to determine an average equatorial rotation velocity from HeI(HeII) lines of v e = 348 (173) km/s for these O stars, with individual star’s v e fromHeI (HeII) lines ranging from 482 (237) to 290 (91) km/s. We argue that the ∼ v e ∼
530 km/s, the observed average v e ofthe O-type stars in our sample is 65% that speed. This demonstrates that, even overthe relatively short WR-phase timescale, tidal and/or other effects causing rotationalspin-down must be efficient. A challenge to tidal synchronization theory is that the twolongest-period binaries in our sample (with periods of 29.7 and 78.5 days) unexpectedlydisplay super-synchronous rotation. Key words: surveys – binaries: massive – stars: Wolf-Rayet – stars:black holes
When massive stars rotate sufficiently rapidly (i.e. withequatorial velocities > ⋆ E-mail: [email protected]
Wolf-Rayet (WR) stars, are found in binary systems. Merg-ing binary black holes (Abbott et al. 2016) that weregenerated in WR + O binaries may have spins deter-mined by the evolution of the binary components. Thosespins will be measurable with Advanced LIGO and Virgo(P¨urrer, Hannam, & Ohme 2016).The theoretically-predicted importance of rotation hasmotivated several observational studies to measure the dis-tribution of rotational velocities of massive stars. Galac-tic massive stars’ rotation is studied in Penny (1996);Howarth et al. (1997); Vanbeveren et al. (1998a), while theVLT Tarantula survey examined similar stars in the LMC(Dufton et al. 2013; Ram´ırez-Agudelo et al. 2013). All of c (cid:13) M. Shara et al. these studies reached a similar conclusion: the distributionfor both the early B-type stars and O-type stars is bi-modal. The majority of these stars are relatively slow ro-tators with an average equatorial velocity v e ∼
100 km/s,while a smaller, but significant fraction of them are rapidrotators with v e >
200 km/s, reaching up to 500-600 km/s.The early Be-type stars obviously belong to the early B-typehigh velocity tail.The Geneva team (Ekstr¨om et al. (2012) and referencestherein) attempted to determine the average equatorial ve-locity < v e > of massive single stars at birth under the as-sumption that the overall bimodal distribution, noted above,is the norm for massive single stars in general. They con-cluded that < v e > ∼
300 km/s. This has been the value usedin most of the massive single star evolutionary calculationspublished by the Geneva team in the last 15 years.
However, it was argued decades ago that a significant frac-tion of all massive stars are close binary components (forextended reviews see van den Heuvel (1993); Vanbeveren(1998b); Vanbeveren et al. (1998a)), and that a significantfraction of the rapid rotators may have a binary origin. Since1998 several extensive observational campaigns have con-firmed that most (and perhaps all) massive stars are born inclose binaries (Mason et al. 1998, 2009; Sana & Evans 2011;Sana et al. 2013). The suggestion that “many rapid rota-tors have a binary origin” has become more and more plau-sible. A theoretical study by de Mink et al. (2013), whichused overall population synthesis tools and included bina-ries, successfully reproduces the observed distribution of ro-tation rates of massive stars.Vanbeveren, De Loore, & Van Rensbergen (1998c)posed the questions of whether the WR components inWR+OB binaries formed by stellar wind mass loss, or bybinary mass loss processes (Roche Lobe Overflow -RLOF- or Common Envelope -CE), and if RLOF happened inWR+OB binary progenitors, was it accompanied by masstransfer and mass accretion? When primaries exceed ∼ M ⊙ they may lose a significant fraction of theirinitial masses via stellar winds, greatly increasing thebinary period. The existence of WR + O binaries withperiods of days can only come about via RLOF or CEphases. The progenitor of a WR+OB binary, where theOB star has luminosity class V, or where the O componentis an early O-type star, most likely underwent RLOFand mass transfer, causing the rejuvenation of the gainer(Vanbeveren, De Loore, & Van Rensbergen 1998c).Rapid rotation in close binaries can arise during Rochelobe overflow (RLOF) when mass lost by the mass donoris accreted by the mass gainer (Vanbeveren 1998b). Massaccretion is accompanied by angular momentum accretion,hence, in this scenario, the mass gainer spins up and becomesa rapid rotator (Packet 1981). When the mass donor ends itslife with an asymmetric supernova explosion, the binary maybe disrupted. This binary supernova scenario successfullypredicts that the escaped component (the mass gainer) willbe a rapidly rotating single runaway star (defined as a starwith peculiar space velocities >
30 km/s (Blaauw 1961)).Most of the known massive runaway stars are rapid rotators,with ζ Pup being the prototype (Vanbeveren 2012). ζ Pup displays spectral class O4 I(n)fp (Sota et al. 2014). Whileuncertainties in its distance (Ma´ız Apell´aniz, Alfaro, & Sota2008; Schilbach & R¨oser 2008) translate directly into uncer-tainties in its radius (14 - 26 R ⊙ ) and mass (22 -56 M ⊙ ),and hence its critical rotation speed (550 -640 km/s) , its ob-served v e sin(i) of 220 km/s (Vanbeveren 2012) is ∼ − The motivation of this and subsequent papers is to test theprediction of RLOF-driven spin-up in massive, Wolf-Rayet(WR) + O binary stars (Petrovic, Langer, & van der Hucht2005). Most classical Wolf-Rayet (WR) stars are core heliumburning objects that have lost their hydrogen-rich layers.This has occurred via stellar wind mass loss, if the WR staris a single star or in a wide, non-interacting binary, and/orby RLOF when the WR star is a close-binary component.O-type companions to WR stars that are spinning super-synchronously (with v e sin i typically > M ⊙ O V star and a 13.4 M ⊙ WR star in a9.555 day orbit that is inclined at 55.3 degrees to the line-of-sight (de La Chevroti`ere, Moffat, & Chen´e 2011). TheMartins, Schaerer, & Hillier (2005) calibration for O-starradii (their table 1) yields 8.11 R ⊙ as the radius of the Ostar in WR 127. If the O star were rotating synchronouslyit would display a v e sin i of 36 km/s. The observedvalue is ∼
250 km/s (Massey 1981), which is highly super-synchronous.A comprehensive study of the rotational velocities ofthe O-type components of WR+O binaries would be avaluable test of the idea that many rapid rotators havea binary origin, and have undergone RLOF mass trans-fer. A literature search reveals that the v e sin i value ofthe O-type components of WR + O binaries have beenmeasured for only two systems, and estimated for onemore. In one carefully measured case (V444 Cyg = WR139(Marchenko, Moffat, & Koenigsberger 1994)), the O-typecomponent is observed to be a rapid rotator, with v e sini = 215 ±
13 km/s. In the second measured case (thebrightest and nearest WR+O binary in the sky = WR11 (Baade, Schmutz, & van Kerkwijk 1990)), the O7.5 giantdisplays v e sin i = 220 ±
20 km/s. For HD 186943 = WR127(Massey 1981), the widths of the O stars’ absorption lineswere estimated to correspond to v e sin i ∼ − v e sin i = 200 - 250 km/s is significantly faster than the v e corresponding to binary synchronicity in all three systems,and suggestive, though not conclusive evidence that RLOFhas been important in the systems’ evolution. Despite theseencouraging early works, a systematic, quantitative study ofO-star spins in WR + O binaries, both from a theoreticaland an observational point of view, has not yet been carriedout. The test of whether O companions in WR+O binariesare spun up by binary accretion can best be accomplishedby measuring the observational line width parameter v. sini of the O component from line broadening of helium lines(Ram´ırez-Agudelo et al. 2013, 2015). Once this is availablethen one wants, ideally, to extract v e , the equatorial rota- c (cid:13) , 000–000 star spin rates tion speed of the star. A binary’s sin i may be obtained frompolarimetry, photospheric or atmospheric eclipses, colliding-wind analysis, by assuming a mass for the O star from itsspectral type if not otherwise known, or from visual binaryorbits (Moffat 2008). If one assumes alignment of the spinaxes of the binary components with the binary axis, thensin i is also known for the O-star axis. However, this maynot always be a good assumption (e.g. Villar-Sbaffi et al.(2005, 2006) for CQ Cep and CX Cep, two very-short pe-riod Galactic WR+O binaries, where the orbital and spinaxes are misaligned). Therefore, one needs a large numberof systems from which one can extract statistical informa-tion from v. sin i to deduce < v e > . This is precisely themethodology adopted in studies of rotation speeds of singlefield O stars (Ram´ırez-Agudelo et al. 2013), and in studies ofO + O star binaries (Ram´ırez-Agudelo et al. 2015). In thisstudy we have succeeded in measuring the v. sin i of eightO stars in WR + O binaries, thereby raising the sample sizewith well-measured spin rates from two to ten.In Section 2 we describe the data and their reductions.The high resolution spectra of the helium lines of the O starswe study are presented in Section 3, as well as these stars’derived rotation rates. In section 4 we discuss the implica-tions of our results for the overall evolution of rotationalvelocities in massive binaries, and we briefly summarize ourresults in Section 5. Observations of the target stars were obtained with theHigh Resolution Spectrograph (Crause et al. 2014) (HRS) ofthe Southern African Large Telescope (SALT). The HRS is adual beam, fiber-fed echelle spectrograph. It was used in lowresolution mode with a 2.23” arcsec diameter fiber to providea spectrum in the blue arm over the spectral range of 3700-5500 ˚A. All of the spectra have a resolving power R ∼ > × τ Scowere obtained on 7 August 2015 to provide a measurementof a source with a known, low value for v e sin i. Basic CCDreductions including bias subtractions, gain corrections, andflat fielding were handled by the ccdproc package (Crawford2015).Spectroscopic reductions of the data were carried outusing the pyhrs package (Crawford 2015). The software cre-ated an order frame from a flat field image that assigned eachpixel in the image to a specific order. Prior to extracting,each order was corrected for spatial and spectral curvature.A second order polynomial was fit to the overall shape of the Observations were taken under SALT Proposal Code: 2015-1-SCI-064. order, which was then removed. In addition, each row of theorder was corrected for a small offset in the vertical directionbetween each of the rows of the order. This offset was wellbelow the size of the resolution element for the low resolu-tion mode. To calculate the wavelength solution, a spectrumwas extracted by summing the rows in the order. The resul-tant spectrum was then passed to the specidentify taskin the PySALT reduction package (Crawford et al 2010) forwavelength identification and calculation of the wavelengthsolution. Next, the target spectrum was extracted from ourobservations. The extraction of the order was performed asdescribed with the final spectrum being a result of a flux-weighted co-addition of all of the illuminated rows in the or-der and the wavelength derived from the solution calculatedfor that order. As a minimum of two exposures were takenfor each observation, the final step combined the extractedspectra from these two exposures to produce the final spec-trum for each star. HRS does have a sky and target fiber,but as the target fiber only was used in our analysis, all stepswere only performed on the target fiber.
The expected FWHM for an unresolved line at λ = 4750 ˚ A is ∼ .
37 ˚ A , corresponding to a resolving power R ∼ λ . ± . A , better than the predictedvalue.As a further test of our methodology and the resolvingpower of HRS, we observed the bright B0.2V star τ Sco =HD 149438. The most recently measured v e sin i of τ Sco =4 km/s (Nieva & Przybilla 2012). This extremely low valueof v e sin i demonstrates that the effects of thermal Dopplerbroadening, Stark effect, and micro and macro-turbulenceare very small in a star of spectral type nearly identicalto that of our coolest observed O star. Of course, this isno guarantee that the O stars’ He lines in the O + WRbinaries that we study below do not undergo line broad-ening from one or more of the above mechanisms. Indeed,Sundqvist et al. (2013) describe two magnetic O stars withrotation periods longer than one year, and hence with v e sin i v e sin i ∼ v e sin i up to ∼ λ τ Sco (see Figure 2). The average v. sin i from the three lines, using the same methodologydescribed below for our program stars, was 24 ± τ Sco c (cid:13) , 000–000 M. Shara et al. to its resolution limit. All of our program stars, discussedbelow, display FWHM of He absorption lines that are 10 to20 times larger than the resolution limit of SALT’s HRS lowresolution mode, corresponding to v. sin i (for HeI lines) thatalways exceed 200 km/s. Simulations show that the blendingof the satellite lines seen in Figure 2 with the 4922 line dueto spin speeds in excess of 150 km/s and micro- or macro-turbulence of 50 km/s cause us to overestimate v. sin i byless than 15% in all cases. To determine the value for v. sin i for the stars in oursample, we followed the process as outlined in detail byRam´ırez-Agudelo et al. (2013) and Ram´ırez-Agudelo et al.(2015). This involved measuring the broadening of thesestars’ HeI and HeII lines via the full width at half max-imum (FWHM). To measure the FWHM of the linesin our stars, we used the astropy modeling software(Astropy Collaboration et al. 2013) to fit Gaussian curvesto the HeI λ λ λ λ The fit to each of the continuum-divided absorption linesof HeI and HeII in each of our eight WR + O binaries ispresented in Figures 3 through 10.
Figure 1.
A SALT High Resolution Spectrograph arc line closeto the HeII 4541 line. The line’s FWHM is 0 . ± . A . The datapoints are in light gray while the model fit is the solid black curvein this and all following figures. Wavelength (Å) F r a c t i o n o f C o n i n t uu m Figure 2.
The HeI 4922 absorption line of the slowly rotatingstar τ Sco on 07 Aug 2015.
In Table 1, we report our measured HeI and HeIIFWHM and velocities for each of eight O stars in WR+ O binaries, for τ Sco, and for WR 127 from Figure 8of de La Chevroti`ere, Moffat, & Chen´e (2011). HeI absorp-tion lines were detected in five O stars by SALT (WR 21,WR 31, WR 42, WR 97 and WR 113), and in WR 127 byde La Chevroti`ere, Moffat, & Chen´e (2011), while the HeII λ . ± . A (for the HeII λ v. sin i of 89 ±
11 km/s for WR 21. Thelargest FWHM measured is 7 . ± . A (for the HeI λ v. sin i of 315 ±
21 km/s forWR 42. For each star with more than one measurementof v. sin i, (because of multiple epochs and/or more thanone line measured), we adopt an average value from all itsmeasurements. The average v. sin i for our sample of eightO stars with measured HeII lines is 137 km/s. The average v. sin i for our sample of six O stars with measured HeI linesis 268 km/s, which is nearly twice as large as the HeII lines’ v. sin i. We return to this large difference below.These data can now be combined with orbital inclina-tions gleaned from the literature. In table 2 we present thesedata, which yields (by far) the largest ensemble of rotational c (cid:13)000
21 km/s forWR 42. For each star with more than one measurementof v. sin i, (because of multiple epochs and/or more thanone line measured), we adopt an average value from all itsmeasurements. The average v. sin i for our sample of eightO stars with measured HeII lines is 137 km/s. The average v. sin i for our sample of six O stars with measured HeI linesis 268 km/s, which is nearly twice as large as the HeII lines’ v. sin i. We return to this large difference below.These data can now be combined with orbital inclina-tions gleaned from the literature. In table 2 we present thesedata, which yields (by far) the largest ensemble of rotational c (cid:13)000 , 000–000 star spin rates Table 1.
Measured FWHM and v. sini of τ Sco and O stars in O+WR star binariesObject Observed HeI λ , v e sin i (km/s) He II λ λ v e sin i (km/s)WR21 2015-05-08 5 . ± .
23 246 ±
20 3 . ± .
04 89 ± . ± .
17 209 ±
16 4 . ± .
03 120 ± . ± .
23 178 ± . ± .
05 216 ±
15 5 . ± .
05 176 ± . ± .
05 315 ±
21 4 . ± .
04 109 ± ... ... . ± .
05 122 ± . ± .
13 90 ± . ± .
06 86 ± . ± .
03 121 ± . ± .
04 121 ± . ± .
06 127 ± . ± .
12 230 ±
17 4 . ± .
03 136 ± . ± .
14 254 ±
18 4 . ± .
05 152 ± . ± .
13 310 ±
22 5 . ± .
07 168 ± . ± .
14 309 ±
22 5 . ± .
06 172 ± . ± . ±
30 ... ... τ Sco 2015-05-08 0 . ± .
06 24 ± velocities of O stars in WR + O binaries yet assembled.The errors in derived v e are dominated by the uncertain-ties in orbital inclinations. In the final column of Table 2 wealso list the synchronous rotation speeds and critical rota-tion speeds of the O-stars in each binary, using the already-measured masses and the spectral type-radius calibrationof Martins, Schaerer, & Hillier (2005). A key result of thispaper is that the average equatorial rotational velocity ofsix O stars, measured from HeI lines, is a highly super-synchronous 348 km/s. The average equatorial rotationalvelocity of eight O stars, measured from HeII lines, is 173km/s, which is still significantly super-synchronous. Ram´ırez-Agudelo et al. (2013) have noted that values of v. sin i for single O stars measured from HeII absorption linesare systematically lower than those measured from HeI linesby 25% (see their Figure 10). They suggested that grav-ity darkening was responsible for this systematic difference.Rapid rotation produces equatorial centrifugal support ofstars, hence oblateness and lower surface gravity at theequator, which is thus cooler and darker. Photospheric re-gions closer to the poles (equators) should then contributemore to the formation of He II (He I) lines. Projected ro-tational velocities derived from He II lines should thus belower than those from HeI lines, as observed in Figure 10 ofRam´ırez-Agudelo et al. (2013).von Zeipel’s theorem (von Zeipel 1924) states that theradiative flux in a uniformly rotating star is proportionalto the local effective gravity. Spectrographic observations,constrained by interferometric measurements of rapidly ro-tating stars show convincingly that latitudinal large tem-perature differences exist in the rapidly rotating stars Al-pha Leo, Alpha Aquila and Achernar (McAlister et al.2005; Monnier et al. 2007; Domiciano de Souza et al. 2014). In particular, Alpha Leo displays equatorial (polar) tem-peratures of 10,314 K (15,400 K) (McAlister et al. 2005).Achernar similarly displays equatorial (polar) temperaturesof 12,673 K (17,124 K) (Domiciano de Souza et al. 2014). Arecent review of the observations of these and other rapidlyspinning stars is given by Claret (2016). These observa-tions demonstrate conclusively that the polar temperaturesof rapidly rotating stars can be 50% hotter than their equa-torial temperatures.Modern theoretical studies of gravity darkening(Espinosa Lara & Rieutord 2011; Claret 2016) concludethat von Zeipel’s theorem is only applicable to slowly ro-tating stars. Espinosa Lara & Rieutord (2011) demonstratethat latitudinal variations in the effective temperature of arapidly rotating star depend on the ratio of the equatorialvelocity to the Keplerian velocity. Their model demonstratesgood agreement with the above-noted observations of AlphaLeo and Alpha Aquila. Claret (2016) obtains good agree-ment with observed gravity darkening indices for six rapidlyrotating stars by considering optical depth effects in thesestars’ atmospheres.While our primary goal in undertaking this obser-vational study was to determine whether RLOF-induced,super-synchronous rotation is seen in the O stars in WR+ O binaries, an unexpected bonus has emerged.
Our ob-servations indicate that not only are the O stars rotatingsuper-synchronously, they also display large variations of ef-fective temperature with latitude, suggesting that they maybe oblate . Since we are interested in the speeds of rotationat the equators of our O stars, it is clear that the velocitieswe derive from the HeI lines, and not those of the HeII linesare the speeds to use.
Mass transfer during RLOF, wherein some of the mass lostby a donor star is accreted by its companion, is accompa-nied by angular momentum transfer, which forces the massgainer to spin-up. It was demonstrated by Packet (1981) c (cid:13) , 000–000 M. Shara et al.
Table 2.
WR + O star binaries’ propertiesSystem Spectral Types Period(d) M(WR+O) sin i ( M ⊙ ) i(deg) HeI v e (km/s) HeII v e (km/s) O-star v(sync)/v(crit)(km/s)WR 21 WN5o + O4-6 8.3 8.4 + 16.3 48 −
62 278 −
331 115 −
138 70/528WR 30 WC6 + O6-8 18.8 15.4 + 31.9 78 - 90 ... 178 −
182 25/818WR 31 WN4o + O8 V 4.8 2.7 + 6.3 40 - 62 244 −
336 199 −
274 89/382WR 42 WC7 + O7 V 7.9 3.7 + 6.2 38 - 44 453 −
511 157 −
177 66/361WR 47 WN6o + O5 V 6.2 40 + 47 67 - 90 ... 88 −
96 93/897WR 79 WC7 + O5-8 8.9 1.8 + 4.9 34 - 45 ... 174 −
220 53/321WR 97 WN5b + O7 12.6 2.3 + 4.1 31 - 85 243 −
470 146 −
279 39/293WR 113 WC8d + O8-9 IV 29.7 10.6 + 22.3 70 - 90 310 −
330 170 −
181 19/623WR 127 WN5o + O8.5 V 9.6 13.4 + 23.9 55 - 90 300 −
366 ... 42/697WR 11 WC8 + O7.5 III 78.5 6.8 + 21.6 63 220 ... 9/545WR 139 WN5o + O6 III-V 4.2 8.8 + 26.3 78 . v e sin i references: This paper, except for WR 127 (de La Chevroti`ere, Moffat, & Chen´e 2011), WR 11(Baade, Schmutz, & van Kerkwijk 1990), and WR 139 (Marchenko, Moffat, & Koenigsberger 1994) that when the RLOF-process in a case B binary is quasi-conservative, then soon after the onset of this process themass gainer is spun up to its critical Keplerian speed. Sincethe observed rotational velocity of the O-type companions inall 10 of the WR binaries in which it has now been measuredis super-synchronous, it is tempting to conclude that masstransfer and spin-up have played important roles during theprogenitor evolution.To further explore this scenario we assume that, verysoon after the RLOF process began in the binaries we nowobserve, the O-type mass gainer was rotating critically. Thiscorresponds to Keplerian rotation speeds of ∼
530 km/s. Theaverage rotation speed of six O-type companions with HeIlines in our observed binary sample is 348 km/s, close to65% of the Keplerian value. The evolutionary timescale of aWR star is typically of the order of a few hundred thousandyears. Thus the 348 km/s observed rotation speed meansthat, even over this short timescale, tidal interactions mustbe efficient, able to cause significant spin-down of the O-typecompanions.We conclude by noting that the two longest periodWR binaries in our sample (WR 113 at P= 29.7 d andWR 11 with P = 78.5 d) each have an O-type companionthat is rotating super-synchronously, but with speeds thatare significantly below critical. We are unaware of a tidal-synchronization theory (e.g., Hut (1981); Tassoul & Tassoul(1996); Zahn (2013)) that is capable of explaining a possiblespin-down from critical rotation to the presently observedsuper-synchronous values in binaries with such large peri-ods. A detailed set of massive star binary evolution mod-els, including RLOF angular momentum transfer and tidalbraking by oblate stellar models, against which to compareour results, is lacking in the literature. Detailed models andpopulation synthesis simulations confronting the observa- A case B binary has an initial orbital period such that RLOFstarts while the mass loser is hydrogen-shell burning, i.e. prior tocore helium burning. The periods of such binaries range betweenabout 10 days and 1000 days. tions reported here will be published in a separate paper(Vanbeveren et al. 2016).
WR + O binaries may be progenitors of ultra-luminoussupernovae, long-duration gamma-ray bursts and BH-BHmergers. Theory predicts that the O stars in WR + O bi-naries must have accreted significant amounts of angularmomentum during RLOF from their companions. Only twoO stars in WR+O binaries have previously published, mea-sured values of v. sin i, and one other has a rough estimate.We report new v. sin i measurements for 8 O stars in WR+ O binaries. Using the literature values of i we then findthe average equatorial rotational value of 348 km/s for 6 Ostars, ranging from 237 to 482 km/s, from HeI lines. Thesevalues are highly super-synchronous. The observed super-synchronicity is in qualitative agreement with the predic-tions of short period, massive binary evolution models whichinclude angular momentum transfer during RLOF and tidalbraking afterwards. However, the super-synchronous natureof the two longest period binaries is a challenge to currenttidal braking theory. We also find that the rotation ratesof O stars in WR + O binaries, measured from HeI lines,are on average twice as large as those measured from HeIIlines. We conclude that we are observing gravity darkeningin these O stars, and predict that they are oblate. ACKNOWLEDGMENTS
All of the new observations reported in this paper were ob-tained with the High Resolution Spectrograph (HRS) ofthe Southern African Large Telescope (SALT). We grate-fully acknowledge the fine support of the astronomers andoperators at the SALT Observatory. The generosity of thelate Paul Newman and the Newman Foundation has madeAMNH’s participation in SALT possible; MMS gratefullyacknowledges that support. S.M.C. acknowledges the SouthAfrican Astronomical Observatory and the National Re-search Foundation of South Africa for support during this c (cid:13) , 000–000 star spin rates Wavelength (Å) F r a c t i o n o f c o n t i nuu m WR212015-05-084905 4910 4915 4920 4925 4930 4935
Wavelength (Å) F r a c t i o n o f c o n t i nuu m WR212015-05-244530 4535 4540 4545 4550
Wavelength (Å) F r a c t i o n o f c o n t i nuu m WR212015-05-084530 4535 4540 4545 4550
Wavelength (Å) F r a c t i o n o f c o n t i nuu m WR212015-05-24
Figure 3. (Top) The HeI 4922 absorption line of WR 21 on 08May 2015. (Second from Top) Same as above but on 24 May 2015.(Third from Top) The HeII 4541 absorption line of WR 21 on 08May 2015. (Bottom) The HeII 4541 absorption line of WR 21 on24 May 2015
Wavelength (Å) F r a c t i o n o f c o n t i nuu m WR302015-05-08
Figure 4.
The HeII 4541 absorption line of WR 30 on 08 May2015.
Wavelength (Å) F r a c t i o n o f c o n t i nuu m WR312015-05-084530 4535 4540 4545 4550
Wavelength (Å) F r a c t i o n o f c o n t i nuu m WR312015-05-08
Figure 5. (Top) The HeI 4922 absorption line of WR 31 on 08May 2015. (Bottom) The HeII 4541 absorption line of WR 31 on08 May 2015. project. We thank Ray Sharples and the Durham Universityteam for construction and delivery of an excellent High Res-olution Spectrograph. This research made use of Astropy, acommunity-developed core Python package for Astronomy(Astropy Collaboration, 2013). We thank Nobert Langer andYong Shao for helpful suggestions. We also thank ProfessorIan Howarth for two careful and thoughtful referee’s reports,which resulted in significant improvements to the paper. c (cid:13) , 000–000 M. Shara et al.
Wavelength (Å) F r a c t i o n o f c o n t i nuu m WR422015-05-084530 4535 4540 4545 4550
Wavelength (Å) F r a c t i o n o f c o n t i nuu m WR422015-05-08
Figure 6. (Top) The HeI 4922 absorption line of WR 42 on 08May 2015. (Bottom) The HeII 4541 absorption line of WR 42 on08 May 2015.
REFERENCES
Abbott B. P., et al., 2016, PhRvL, 116, 241102Astropy Collaboration, et al., 2013, A&A, 558, A33Baade D., Schmutz W., van Kerkwijk M., 1990, A&A, 240,105Blaauw A., 1961, BAN, 15, 265Brott I., et al., 2011, A&A, 530, A116Claret A., 2016, A&A, 588, A15Craig M. W., et al., 2015, ascl.soft, ascl:1510.007Crause L. A., et al., 2014, SPIE, 9147, 91476TCrawford, S. M., Still, M., Schellart, P., et al., 2010, SPIE,7737, 773725Crawford, S. M., and Craig, M. 2015, Astrophysics SourceCode Library, submitted.Crawford, S. M., 2015, Astrophysics Source Code Library,submitted.Crowther P. A., 2015, arXiv, arXiv:1509.00495de La Chevroti`ere A., Moffat A. F. J., Chen´e A.-N., 2011,MNRAS, 411, 635de Mink S. E., Langer N., Izzard R. G., Sana H., de KoterA., 2013, ApJ, 764, 166Domiciano de Souza A., et al., 2014, A&A, 569, A10Dufton P. L., et al., 2013, A&A, 550, A109Espinosa Lara F., Rieutord M., 2011, A&A, 533, A43Ekstr¨om S., et al., 2012, A&A, 537, A146Georgy C., Meynet G., Walder R., Folini D., Maeder A.,2009, A&A, 502, 611Heger A., Langer N., 2000, ApJ, 544, 1016
Wavelength (Å) F r a c t i o n o f c o n t i nuu m WR472015-05-224905 4910 4915 4920 4925 4930 4935
Wavelength (Å) F r a c t i o n o f c o n t i nuu m WR472015-05-21
Figure 7. (Top) The HeII 4541 absorption line of WR 47 on 22May 2015. (Bottom) The HeI 4922 absorption line of WR 47 on21 May 2015
Hirschi R., Meynet G., Maeder A., 2004, A&A, 425, 649Howarth I. D., Siebert K. W., Hussain G. A. J., PrinjaR. K., 1997, MNRAS, 284, 265Hut P., 1981, A&A, 99, 126Kushnir D., Zaldarriaga M., Kollmeier J. A., Waldman R.,2016, MNRAS, 462, 844Lamontagne R., Moffat A. F. J., Drissen L., Robert C.,Matthews J. M., 1996, AJ, 112, 2227Maeder A., Meynet G., 2000, A&A, 361, 159Ma´ız Apell´aniz J., Alfaro E. J., Sota A., 2008, arXiv,arXiv:0804.2553Marchenko S. V., Moffat A. F. J., Koenigsberger G., 1994,ApJ, 422, 810Martins F., Schaerer D., Hillier D. J., 2005, A&A, 436, 1049Mason B. D., Gies D. R., Hartkopf W. I., Bagnuolo W. G.,Jr., ten Brummelaar T., McAlister H. A., 1998, AJ, 115,821Mason B. D., Hartkopf W. I., Gies D. R., Henry T. J.,Helsel J. W., 2009, AJ, 137, 3358Massey P., 1981, ApJ, 244, 157McAlister H. A., et al., 2005, ApJ, 628, 43Moffat A. F. J., 2008, IAUS, 250, 119Monnier J. D., et al., 2007, Sci, 317, 342Nieva M.-F., Przybilla N., 2012, A&A, 539, A143Packet W., 1981, A&A, 102, 17Penny L. R., 1996, ApJ, 463, 737Petrovic J., Langer N., van der Hucht K. A., 2005, A&A,435, 1013P¨urrer M., Hannam M., Ohme F., 2016, PhRvD, 93, 084042Ram´ırez-Agudelo O. H., et al., 2013, A&A, 560, A29 c (cid:13)000
Hirschi R., Meynet G., Maeder A., 2004, A&A, 425, 649Howarth I. D., Siebert K. W., Hussain G. A. J., PrinjaR. K., 1997, MNRAS, 284, 265Hut P., 1981, A&A, 99, 126Kushnir D., Zaldarriaga M., Kollmeier J. A., Waldman R.,2016, MNRAS, 462, 844Lamontagne R., Moffat A. F. J., Drissen L., Robert C.,Matthews J. M., 1996, AJ, 112, 2227Maeder A., Meynet G., 2000, A&A, 361, 159Ma´ız Apell´aniz J., Alfaro E. J., Sota A., 2008, arXiv,arXiv:0804.2553Marchenko S. V., Moffat A. F. J., Koenigsberger G., 1994,ApJ, 422, 810Martins F., Schaerer D., Hillier D. J., 2005, A&A, 436, 1049Mason B. D., Gies D. R., Hartkopf W. I., Bagnuolo W. G.,Jr., ten Brummelaar T., McAlister H. A., 1998, AJ, 115,821Mason B. D., Hartkopf W. I., Gies D. R., Henry T. J.,Helsel J. W., 2009, AJ, 137, 3358Massey P., 1981, ApJ, 244, 157McAlister H. A., et al., 2005, ApJ, 628, 43Moffat A. F. J., 2008, IAUS, 250, 119Monnier J. D., et al., 2007, Sci, 317, 342Nieva M.-F., Przybilla N., 2012, A&A, 539, A143Packet W., 1981, A&A, 102, 17Penny L. R., 1996, ApJ, 463, 737Petrovic J., Langer N., van der Hucht K. A., 2005, A&A,435, 1013P¨urrer M., Hannam M., Ohme F., 2016, PhRvD, 93, 084042Ram´ırez-Agudelo O. H., et al., 2013, A&A, 560, A29 c (cid:13)000 , 000–000 star spin rates Wavelength (Å) F r a c t i o n o f c o n t i nuu m WR792015-05-084530 4535 4540 4545 4550
Wavelength (Å) F r a c t i o n o f c o n t i nuu m WR792015-05-214530 4535 4540 4545 4550
Wavelength (Å) F r a c t i o n o f c o n t i nuu m WR792015-05-22
Figure 8. (Top) The HeII 4541 absorption line of WR 79 on 08May 2015. (Middle) Same as above but on 21 May 2015. (Bottom)Same as above but on 22 May 2015.
Ram´ırez-Agudelo O. H., et al., 2015, A&A, 580, A92Sana H., Evans C. J., 2011, IAUS, 272, 474Sana H., et al., 2013, A&A, 550, A107Schilbach E., R¨oser S., 2008, A&A, 489, 105Shara M. M., Faherty J. K., Zurek D., Moffat A. F. J.,Gerke J., Doyon R., Artigau E., Drissen L., 2012, AJ,143, 149Sota A., Ma´ız Apell´aniz J., Morrell N. I., Barb´a R. H.,Walborn N. R., Gamen R. C., Arias J. I., Alfaro E. J.,2014, ApJS, 211, 10Sundqvist J. O., Sim´on-D´ıaz S., Puls J., Markova N., 2013,A&A, 559, L10Tassoul J.-L., Tassoul M., 1996, FCPh, 16, 337van den Heuvel E. P. J., 1993, SSRv, 66, 309Vanbeveren D., De Loore C., Van Rensbergen W., 1998,A&ARv, 9, 63
Wavelength (Å) F r a c t i o n o f c o n t i nuu m WR972015-05-124905 4910 4915 4920 4925 4930 4935
Wavelength (Å) F r a c t i o n o f c o n t i nuu m WR972015-06-134530 4535 4540 4545 4550
Wavelength (Å) F r a c t i o n o f c o n t i nuu m WR972015-05-124530 4535 4540 4545 4550
Wavelength (Å) F r a c t i o n o f c o n t i nuu m WR972015-06-13
Figure 9. (Top) The HeI 4922 absorption line of WR 97 on 12May 2015. (Second from Top) Same as above but on 13 June2015. (Third from Top) The HeII 4541 absorption line of WR 97on 12 May 2015. (Bottom) The HeII 4541 absorption line of WR97 on 13 June 2015.c (cid:13) , 000–000 M. Shara et al.
Wavelength (Å) F r a c t i o n o f c o n t i nuu m WR1132015-05-244905 4910 4915 4920 4925 4930 4935
Wavelength (Å) F r a c t i o n o f c o n t i nuu m WR1132015-06-074530 4535 4540 4545 4550
Wavelength (Å) C o un t s WR1132015-05-244530 4535 4540 4545 4550
Wavelength (Å) F r a c t i o n o f c o n t i nuu m WR1132015-06-07
Figure 10. (Top) The HeI 4922 absorption line of WR 113 on24 May 2015. (Second from Top) Same as above but on 07 June2015. (Third from Top) The HeII 4541 absorption line of WR 113on 24 May 2015. (Bottom) The HeII 4541 absorption line of WR113 on 07 June 2015.
Vanbeveren D., De Donder E., Van Bever J., Van Rensber-gen W., De Loore C., 1998, NewA, 3, 443Vanbeveren D., Van Rensbergen, W., De Loore, C., 1998,The Brightest Binaries. Boston: KluwerVanbeveren D., 2009, NewAR, 53, 27Vanbeveren D., 2010, ASPC, 435, 115Vanbeveren D., 2010, AIPC, 1314, 361Vanbeveren D., 2010, IAUS, 266, 293Vanbeveren D., 2011, BSRSL, 80, 530Vanbeveren D., 2012, ASPC, 465, 342Vanbeveren D., Mennekens, N., Shara, M., Crawford, S.,Moffat, A.F.J., Zurek, D., Crause, L., 2016, in preparationVillar-Sbaffi A., St-Louis N., Moffat A. F. J., Piirola V.,2005, ApJ, 623, 1092Villar-Sbaffi A., St-Louis N., Moffat A. F. J., Piirola V.,2006, ApJ, 640, 995von Zeipel H., 1924, MNRAS, 84, 665Woosley S. E., Heger A., 2006, ApJ, 637, 914Yoon S.-C., Langer N., 2005, A&A, 443, 643Zahn J.-P., 2013, LNP, 861, 301 c (cid:13)000