The positional probability and true host star identification of TESS exoplanet candidates

Abstract

ABSTRACT We present a method for deriving a probabilistic estimate of the true source of a detected TESS transiting event. Our method relies on comparing the observed photometric centroid offset for the target star with models of the offset that would occur if the event was either on the target or any of the Gaia identified nearby sources. The comparison is done probabilistically, allowing us to incorporate the uncertainties of the observed and modelled offsets in our result. The method was developed for TESS Full Frame Image light curves produced from the TESS Science Processing Operations Center (SPOC) pipeline, but could be easily adapted to light curves from other sources. We applied the method on 3226 TESS Objects of Interest (TOIs), with a released light curve from SPOC. The method correctly identified 96.5 per cent of 655 known exoplanet hosts as the most likely source of the eclipse. For 142 confirmed Nearby Eclipsing Binaries (NEBs) and Nearby Planet Candidates (NPCs), a nearby source was found to be the most likely in 96.5 per cent of the cases. For 40 NEBs and NPCs where the true source is known, it was correctly designated as the most likely in 38 of those. Finally, for 2365 active planet candidates, the method suggests that 2072 are most likely on-target and 293 on a nearby source. The method forms a part of an in-development vetting and validation pipeline, called RAVEN, and is released as a standalone tool.