8 Helium abundance versus rotational velocity

Blaauw (1993) pointed out that most of the reliable runaways known at the time have high helium abundances ( ) and large rotational velocities (  km s^-1), whereas most of the non-runaways have normal helium abundances and rotational velocities. Only one classical runaway star, AE Aur, does not follow this trend, and this was already suspected to be the result of the DES rather than the BSS. Blaauw suggested that the $(\epsilon,v_{\rm rot} \sin i)$-diagram of the small sample of known runaway stars supported the conjecture that most runaways are produced by the BSS, since these characteristics are natural consequences of close binary evolution (Sect. 1; van den Heuvel 1985).

As shown in Table 3, only five of the 23 runaways listed there have a measurement of $\epsilon$, and six do not have a measured rotational velocity. In order to pursue Blaauw's suggestion, we therefore constructed a sample of O stars with known rotational velocities (Penny 1996) and helium abundances (Kudritzki & Hummer 1990; Herrero et al. 1992). We also determined, based on Hipparcos astrometry and Hipparcos Input Catalogue radial velocities, the space velocities of these stars with respect to their local standard of rest. The $(\epsilon,v_{\rm rot} \sin i)$ diagram in the left panel of Fig. 16 shows that these O stars can roughly be divided into three groups: (i) those with small rotational velocities,  km s^-1, and normal helium abundances, $\epsilon \sim 0.09$, (ii) those with moderate rotational velocities,  km s^-1, and normal to high helium abundances, , and (iii) those with large rotational velocities,  km s^-1, and high helium abundances, . The symbols in the left panel of Fig. 16 are chosen according to the magnitude of the space velocity. The stars represented by filled circles have space velocities $v_{\rm space} \ge 30$ km s^-1, and the open circles have $v_{\rm space} < 30$ km s^-1; the other symbols indicate stars for which no Hipparcos data are available (asterisks) or for stars with insignificant Hipparcos data (starred).

Figure 17: Colour vs. absolute magnitude diagrams of the runaways (stars) and their parent association/cluster (small dots). The association/cluster members have been de-reddened using the Q-method. The colours and absolute magnitudes of the runaways have been determined using their spectral types (Table 3; Schmidt-Kaler 1983). The isochrones are from Schaller et al. (1992) for Solar metallicity and standard mass loss. The ages of the associations are indicated in the top right of each panel (US: Upper Scorpius, UCL: Upper Centaurus Lupus, LCC: Lower Centaurus Crux)

The left panel of Fig. 16 does not show an obvious separation between runaways (filled circles) and non-runaways (open circles). However, based on the available data for each star (position, radial velocity, distance modulus, cluster membership) it is possible to decide whether the star is a runaway. The result is shown in the right panel of Fig. 16; it turns out that many of the stars with insignificant Hipparcos data are runaways (including some classical runaways), and that some of the stars indicated as high-velocity stars in the left panel are members of a cluster which has a peculiar motion. The resulting $(\epsilon,v_{\rm rot} \sin i)$ diagram now shows a clear separation between the runaway stars and the normal O stars. Except for AE Aur (Sect. 4) all runaways have high helium abundances, , and large rotational velocities,  km s^-1. Only one star in this area of the $(\epsilon,v_{\rm rot} \sin i)$ diagram is not indicated as a runaway: the double star HIP 113306 (OV7n). This star is a member of Cep OB3 ( $\epsilon = 0.17$, $v_{\rm rot} \sin i \sim 359$ km s^-1).

We thus confirm Blaauw's conclusion that massive runaways predominantly have high helium abundances and large rotational velocities, suggesting that they are formed mainly by the binary-supernova scenario. However, this conclusion is based on a limited sample which is by no means statistically complete. A systematic survey of the radial velocities, rotational velocities and chemical abundances of the early-type stars in the Solar neighbourhood is highly desirable.