Probability of Physical Association of 104 Blended Companions to \textit{Kepler} Objects of Interest Using Visible and Near-Infrared Adaptive Optics Photometry

TL;DR

This study assesses the likelihood that blended companions to Kepler Objects of Interest are physically associated, using adaptive optics photometry, to improve false positive estimates and inform planet formation models.

Contribution

It introduces two methods for calculating photometric parallax to determine physical association of companions, providing new insights into false positives in Kepler data.

Findings

14.5% of companions are likely unassociated with primaries

No increased association likelihood for companions within 1 arcsecond

2.6% of transits suggest large planet radii over 15 Earth radii

Abstract

We determine probabilities of physical association for stars in blended Kepler Objects of Interest, and find that $14.5\%^{+3.8\%}_{-3.4\%}$ of companions within $\sim4\arcsec$ are consistent with being physically unassociated with their primary. This produces a better understanding of potential false positives in the Kepler catalog and will guide models of planet formation in binary systems. Physical association is determined through two methods of calculating multi-band photometric parallax using visible and near-infrared adaptive optics observation of 84 KOI systems with 104 contaminating companions within $\sim4\arcsec$. We find no evidence that KOI companions with separation of less than $1\arcsec$ are more likely to be physically associated than KOI companions generally. We also reinterpret transit depths for 94 planet candidates, and calculate that $2.6\% \pm 0.4\%$ of transits have $R > 15R_{\earth}$, which is consistent with prior modeling work.