Characterising & classifying the local population of ultracool dwarfs with Gaia DR2 and EDR3

Abstract

Ultracool dwarfs (UCDs) are the lowest mass products of star formation and span the end of the stellar main sequence from very-low mass, hydrogen-burning M stars to the coolest brown dwarfs. In this thesis we characterise and classify the ultracool dwarf population in the solar neighbourhood using the accuracy and precision of data from the Gaia space observatory. Combining astrometric (in particular parallax) and photometric data from Gaia DR2 and EDR3 with photometry from UKIDSS, SDSS and 2MASS, we prepare some of the largest and most accurate, near-100% complete volume-limited populations of nearby, field late-M, L and T dwarfs. From these samples we derive key population characteristics such as colour-absolute magnitude relationships, the stellar luminosity function, the binary fraction and the binary mass ratio. Our statistical-based approach differs from much of the UCD literature to date which seeks to prepare meta-catalogues from disparate surveys and individual spectroscopic observations with distance determined by indirect methods. Our approach offers improvements in scale, completeness, and distance accuracy. In particular we use Gaia to update the colour-magnitude relations and derive the stellar luminosity functions in MJ and MG of the UCDs. We calculate the binary fraction of the late-M and early-L dwarfs as a function of spectral type by carefully modelling the over-luminous unresolved binary population and show that late-M dwarf binaries reside almost exclusively in equal-mass pairs or twins. Given the complex spectral features of UCDs, consistent and accurate classification is challenging. We investigate the current traditional methods of classification and evaluate a range of alternative techniques including supervised and unsupervised machine learning. In a separate study we use Gaia data to prepare a large, cylindrical sample of FGK main sequence dwarf stars to calculate the structure of the vertical density distribution close to the galactic plane, in fine detail, as a function of colour. Using our derived colour-dependent thin disk scale height we directly determine the star formation history of the solar neighbourhood by modelling the evolution of stellar populations using state-of-the-art PARSEC isochrones.