Presentation #105.01 in the session Binary Stars.
The most ubiquitous oddity of stars is that they are not a single object, but a binary (or higher multiple) — indistinguishable in many ways but with a whole panoply of new dynamical behaviours and paths for evolution. Precise astrometry can resolve the slight excess motion caused by a companion, and the thousand-fold increase in precision afforded by the Gaia survey can be expected to yield a similar thousand-fold increase in detectable astrometric binaries. Fitting a single-star model is easy, so much so that we do it to every source by default. An unresolved binary companion will perturb this fit, biasing the inferred position, motion and parallax, and introducing extra error irreconcilable with the single-star model. Thus we can detect, constrain and start to classify a huge number of astrometric binary systems, with periods from ~a month to a decade, from the already available Gaia data, its error, and the small shifts between data releases. In this talk I’ll introduce in more detail the character of the astrometric contribution of binary systems, and thus how the observability depends on the properties of the system. I will also introduce simple methods for selecting relatively reliably binary candidates and present the properties of such nearby candidates in Gaia eDR3.