Probing the sub-arcsecond regime for visually unresolved systems using Gaia astrometry

Multiplicity is a common phenomenon in the stellar population, and is nigh ubiquitous amongst the most massive stars. The high rate of detecting multiple star systems indicates that this phenomenon is intrinsically linked to the process of star formation, and thus must be investigated in order to obtain a comprehensive understanding of stellar formation and evolution. In the age of high-precision astrometric surveys such as Gaia, large-scale studies into binarity using changes in astrometry over an extended period can be performed. This thesis thus presents three studies into binarity, linked by a common metric - the Proper Motion Anomaly (PMa) - which utilises changes between two different proper motion vectors to probe the sub-arcsecond regime for visually unresolved companions. The first of these studies investigates the multiplicity of B and Be-type stars, utilising two different PMa values and the Gaia-provided RUWE. Therein, a dearth of astrometric binaries at the closest separations probed by these techniques is found, consistent with expectations of post-mass transfer systems. The second provides a study into the binarity of Herbig stars - with 16±2% of the sources within the sample used being detected as binary via the Gaia DR2 - DR3 PMa. Trends in binarity as a function of a variety of properties were also investigated, with a decrease in binarity as these sources age being noted, alongside an increase in the rate of binarity as the mass increases. Furthermore, this study reports 37 systems newly detected as binary. The final study then proceeds to detail the process of generating two large catalogues of PMa values; one based upon a combination of Tycho-2 and Gaia DR3 data, containing 2,372,187 sources, 1.78% of which are detected as binary, and based upon a combination of Gaia DR2 and DR3 data, containing 1,094,962,906 sources, 0.51% of which are detected as binary. These catalogues provide an up to three orders of magnitude expansion over previous, similar studies.