On the Development and Applications of Automated Searches for Eclipsing Binary Stars

TL;DR

This paper presents a large-scale automated pipeline for identifying and analyzing eclipsing binary stars from extensive photometric datasets, leading to new discoveries and detailed characterization of rare low-mass systems.

Contribution

Developed a systematic pipeline for detecting and modeling eclipsing binaries in large datasets, resulting in new candidate discoveries and detailed analysis of low-mass systems.

Findings

Identified over a dozen new low-mass eclipsing binary candidates.

Spectroscopically confirmed masses of five low-mass binaries.

Discovered a long-period binary consistent with magnetic disruption hypothesis.

Abstract

Eclipsing binary star systems provide the most accurate method of measuring both the masses and radii of stars. Moreover, they enable testing tidal synchronization and circularization theories, as well as constraining models of stellar structure and dynamics. With the recent availability of large-scale multi-epoch photometric datasets we were able to study eclipsing binary stars en masse. In this thesis, we analyzed 185,445 light curves from ten TrES fields, and 218,699 light curves from the OGLE II bulge fields. In order to manage such large quantities of data, we developed a pipeline with which we systematically identified eclipsing binaries, solved for their geometric orientations, and then found their components' absolute properties. Following this analysis we assembled catalogs of eclipsing binaries with their models, computed statistical distributions of their properties, and located rare cases for further follow-up, including T-Cyg1-03378, which has unusual eclipse timing variations. Of particular importance are low-mass eclipsing binaries, which are rare yet critical for resolving the ongoing mass-radius discrepancy between theoretical models and observations. To this end, we have discovered over a dozen new low-mass eclipsing binary candidates and spectroscopically confirmed the masses of five of them. One of these confirmed candidates, T-Lyr1-17236, is especially interesting because of its long orbital period. We examined T-Lyr1-17236 in detail and found that it is consistent with the magnetic disruption hypothesis. Both the source code of our pipeline and the complete list of our candidates are freely available.