A VLT/FLAMES survey for massive binaries in Westerlund 1

Abstract

Context. X-ray emission from massive stars was first reported four decades ago, but the precise physics governing its formation as a function of stellar properties and binarity remains not fully understood. With the recent suggestion that such objects may be important sites of cosmic ray production, a better understanding of their high-energy properties is particularly timely. Aims. The young massive cluster Westerlund 1 provides an ideal testbed for understanding this emission, with over 50 cluster members detected in historical X-ray observations. In the decade since these data were obtained, significant new multi-epoch observations of the cluster have been made, allowing a fundamental reappraisal of the nature of both X-ray bright and dark stars. Methods. Optical spectroscopy permits accurate classification of cluster members, while multi-epoch observations of a sub-set allow identification and characterisation of the binary population. Results. A total of 45 X-ray sources within Wd1 now have precise spectral classifications. Of these, 16 have been identified as candidate or confirmed massive binaries. X-ray emission is confined to O9-B0.5 supergiants, Wolf-Rayets and a small group of highly luminous interacting/post-interaction OB+OB binaries. Despite their presence in large numbers, no emission is seen from earlier, less evolved O stars or later, cooler B super-/hypergiants. A total of 22 stars have X-ray properties that are suggestive of a contribution from emission originating in a wind collision zone. Conclusions. We suppose that the lack of X-ray emission from O giants is due to their comparatively low intrinsic bolometric luminosity if, as expected, they follow the canonical LX/Lbol relation for hot stars. The transition away from X-ray emission for OB supergiants occurs at the location of the bistability jump; we speculate that below this limit, stellar wind velocities are insufficient for internal, X-ray emitting shocks to form. Our results are consistent with recent findings that massive binaries are not uniformly brighter than single stars of comparable luminosity or spectral type, although it is noteworthy that the brightest and hardest stellar X-ray sources within Wd1 are all either confirmed or candidate massive, interacting/post-interaction binaries.