Astrometric identification of nearby binary stars - II. Astrometric binaries in the Gaia Catalogue of Nearby Stars

Abstract

We examine the capacity to identify binary systems from astrometric deviations alone. We apply our analysis to the Gaia eDR3 and DR2 data, specifically the Gaia Catalogue of Nearby Stars. We show we must renormalize (R)UWE over the local volume to avoid biasing local observations, giving a local unit weight error (LUWE). We use the simple criterion of LUWE>2, along with a handful of quality cuts to remove likely contaminants, to identify unresolved binary candidates. We identify 22 699 binary candidates within 100 pc of the Sun (just under 10 per cent of sources in this volume). We find an astrometric binary candidate fraction of around 20 per cent for giant stars, 10 per cent on the main sequence and lower than 1 per cent for white dwarfs. We also look for Variability Induced Movers, by computing the correlation between photometric variability and astrometric noise - and show that VIMs may dominate the binary population of sub-Solar mass MS stars. We discuss the possibility and limitations of identifying non-luminous massive companions from astrometry alone, but find that our method is insensitive to these. Finally, we compare the astrometric deviations of MS binaries to the simulated sample from paper I, which show excellent agreement, and compute the astrometric candidate binary fraction as a function of absolute magnitude.