A&A manuscript no. (will be inserted by hand later) ASTRONOMY AND Your thesaurus codes are: ASTROPHYSICS 08.05.1; 08.06.2; 08.18.1; 10.15.2 Sco OB2 20.3.2018 High S/N Echelle Spectroscopy in Young Stellar Groups ⋆ II. Rotational Velocities of Early-Type Stars in Sco OB2 A.G.A. Brown@1, W. Verschueren@2,1 @1 Sterrewacht Leiden, P.O. Box 9513, 2300 RA, Leiden, The Netherlands @2 University of Antwerp (RUCA), Astrophysics Research Group, Groenenborgerlaan 171, 2020 Antwerpen, Belgium Received... , accepted... Abstract. We investigate the rotational velocities of 1. Introduction early-type stars in the Sco OB2 association. We measure vsini for 156 established and probable members of the An outstanding problem in theories of star formation association. The measurements are performed with three is the so-called angular momentum problem. Molecular different techniques, which are in increasing order of ex- clouds and cloud cores, even if they spin about their axis pected vsini: 1) converting the widths of spectral lines di- only once per Galactic rotation, contain much more an- rectly to vsini, 2) comparing artificially broadened spectra gular momentum than the stars that form from them (see of low vsini stars to the target spectrum, 3) comparing the e.g., Spitzer 1968). Observations suggest that an inter- i He λ4026 line profile to theoretical models. The sample is stellar parent cloud must lose at least two to four or- extended with literature data for 47 established members ders of magnitude of angular momentum for a relatively of Sco OB2. Analysis of the vsini distributions shows that wide binary system and a single star, respectively, to be- there are no significant differences between the subgroups come dynamically possible (see Mouschovias 1991). How of Sco OB2. We find that members of the binary popu- does angular momentum get redistributed when stars form lation of Sco OB2 on the whole rotate more slowly than and what is the resulting distribution of rotational veloc- the single stars. In addition, we find that the B7–B9 single ities? These questions can best be addressed by studying star members rotate significantly faster than their B0–B6 young stellar groups in which the stars are unevolved and counterparts. We test various hypotheses for the distribu- physically related. It is in these groups that the observed tion of vsini in the association. The results show that we distribution of projected rotational velocities (vsini) may cannot clearly exclude any form of random distribution still reflect the distribution of rotational velocities at the of the direction and/or magnitude of the intrinsic rota- time of star formation. Observations of the present vsini tional velocity vector. We also investigate the effects of distribution provide constraints on the angular momen- rotation on colours in the Walraven photometric system. tum history of forming stars, and contain information on arXiv:astro-ph/9608089v1 14 Aug 1996 We show that positions of B7–B9 single dwarfs above the mechanisms that lead to angular momentum redistribu- main sequence are a consequence of rotation. This estab- tion during star formation. lishes the influence of rotation on the Walraven colours, due primarily to surface gravity effects. Suggested mechanisms for the redistribution of angu- lar momentum, include magnetic braking and the forma- tion of disks. Magnetic braking is important during the Key words: stars: early-type; formation; rotation – open early, diffuse stages of star formation (Mouschovias 1991). clusters and associations: individual: Sco OB2 Stellar disks are important during the proto-stellar and pre main-sequence phases (see e.g., Bodenheimer et al. 1993). Another way of redistributing angular momentum during star formation is through the formation of binary systems. This mechanism can be studied by relating the Send offprint requests to: characteristics of the binary population in stellar groups to A.G.A. Brown, [email protected] the vsini distribution, which may also provide information ⋆ Based on observations obtained at the European about tidal interactions in close binaries. Ultimately, ob- Southern Observatory (ESO), La Silla, Chile in the servations of the vsini distribution in young stellar groups framework of Key Programme 5-005-45K may provide information on the star formation process it- 2 A.G.A. Brown & W. Verschueren: Rotational Velocities of Early-Type Stars in Sco OB2 self. For a recent review on the role of rotation in star jected rotational velocities. Various hypotheses about the formation, see Bodenheimer et al.(1993). true distribution of rotational velocities are tested for the Another important aspect of the study of rotational ve- known members of Sco OB2. We also discuss the interpre- locities of stars is the effect of rotation on observed stellar tation of the data in the context of these hypotheses. In parameters. It is well known that photometry and spectral Sect. 5 we combine photometric data with our rotational classification of stars are affected by rotation, due to sur- velocities and confirm that rotation affects the colours of face gravity effects. The resulting misinterpretation of the the stars in the Walraven photometric system. We discuss observed stellar parameters will bias age determinations the effects of the colour changes on age determinations for stellar groups and the derived mass distributions (e.g., and derivations of mass distributions. Finally, in Sect. 6 Maeder 1971). we summarize our conclusions and suggest future work. OB associations are young and mostly unobscured sites of recent star formation and as such are well suited for 2. Observations studies of the properties of young groups of stars (for re- views see Blaauw 1964, 1991). In this paper we focus on The spectra from which the projected rotational veloc- the Sco OB2 association. Because of its proximity (the dis- ities are derived were all obtained with the ECHELle tance to Sco OB2 is 145 pc, see de Geus et al. 1989) it is + Electronographic Camera (ECHELEC) spectrograph, ∼ easily accessible to proper motion studies and as a result mounted at the Coud´efocus of the ESO 1.52m telescope extensive membership determinations have been carried at La Silla. It consists of a 31.6 grooves/mm echelle grating out for this association. Sco OB2 consists of three well- and a 632 grooves/mm grism as a cross-disperser. The de- known subgroups; Upper Scorpius (US), Upper Centau- tector was a CCD (thinned, back-illuminated, RCA) with rus Lupus (UCL) and Lower Centaurus Crux (LCC). The 640 1024 pixels of 15 15 µm2, and a read-out noise of 65 × × Upper Scorpius subgroup is located near the remnants of e−. In order to increase the signal to noise on read-out the its parental molecular cloud, and some of its stars are still CCD pixels were binned 2 2 (henceforth any reference to partly obscured by gas and dust. We also study a subgroup CCD-pixels is to the binned× pixels). The linear dispersion located southeast of US, although the physical reality of of ECHELEC is 3.1 A/mm˚ at 4000 A˚ as measured for our 1 this subgroup has never been established. data. This implies a velocity scale of 7 km s− per binned The ages of the three main subgroups were deter- pixel and a resolving power of 21500. The spectra cover mined most recently from Walraven photometry by de the wavelength region 3800–4070 A˚ which was recorded in Geus et al.(1989). They found ages of 4–5 Myr, 14–15 Myr 11–12 spectral orders. The formal (i.e., taking only pho- and 11–12 Myr for US, UCL and LCC, respectively. For ton and read-out noise into account) signal to noise ratio, US the kinematic age (i.e., the age determined by trac- measured at the top of the blaze profile in each order, ing the proper motions of the members back in time) was typically varies from 70 to 300 between the blue and determined by Blaauw (1978, 1991) to be 5 Myr. the red end of the spectrum.∼ Further∼ details on the obser- Studies of rotational velocities in Sco OB2 have in the vations are given in Verschueren et al.(1996), which also past concentrated mainly on US. Slettebak (1968) deter- describes the data reduction procedure in detail. mined vsini for 82 stars in US and UCL and found that The data were collected during 1991–1993 in the on average the stars in US rotate faster than those in framework of an ESO Key Programme aimed at the de- 1 1 UCL (174 kms− vs. 119 kms− , but the rms spreads termination of high-precision radial velocities of early-type 1 on these numbers are 100 kms− ) and that the mean stars in young stellar groups (Hensberge et al.1990). The ∼ vsini for B7–A0 stars is larger than that of correspond- stars that were observed are established or probable mem- 1 1 ing field stars (212 km s− vs. 138 kms− ). The B0–B6 bers of Sco OB2. They form a subset of a larger sample stars were found to rotate somewhat more slowly than of stars that was submitted for observation by the HIP- field stars. Rajamohan (1976) derived rotational veloci- PARCOS satellite. The details of the selection of the HIP- ties for 112 members of Sco OB2 and found that for stars PARCOS sample can be found in de Zeeuw, Brown & m with MV < 0. 0 the distributions of rotational velocities Verschueren (1994). For the present study a total of 156 are similar for US and UCL and resemble those of field stars was observed, of which 136 already have a measured m stars. Stars with MV > 0. 0, all of which are located in vsini.