Pre-Spectroscopic False Positive Elimination of Kepler Planet Candidates

TL;DR

This paper details a data validation process for Kepler planet candidates, using multiple metrics and graphics to distinguish true planets from false positives, thereby improving candidate vetting accuracy.

Contribution

It introduces specific data validation metrics and graphical analyses to effectively identify genuine exoplanets among Kepler's transit signals.

Findings

Effective identification of false positives using centroid analysis.

Detection of secondary eclipses and transit depth differences.

Enhanced vetting process for Kepler planet candidates.

Abstract

Ten days of commissioning data (Quarter 0) and thirty-three days of science data (Quarter 1) yield instrumental flux timeseries of ~150,000 stars that were combed for transit events, termed Threshold Crossing Events (TCE), each having a total detection statistic above 7.1-sigma. TCE light curves are modeled as star+planet systems. Those returning a companion radius smaller than 2R_J are assigned a KOI (Kepler Object of Interest) number. The raw flux, pixel flux, and flux-weighted centroids of every KOI are scrutinized to assess the likelihood of being an astrophysical false-positive versus the likelihood of a being a planetary companion. This vetting using Kepler data is referred to as data validation. Herein, we describe the data validation metrics and graphics used to identify viable planet candidates amongst the KOIs. Light curve modeling tests for a) the difference in depth of the odd- versus even-numbered transits, b) evidence of ellipsoidal variations, and c) evidence of a secondary eclipse event at phase=0.5. Flux-weighted centroids are used to test for signals correlated with transit events with a magnitude and direction indicative of a background eclipsing binary. Centroid timeseries are complimented by analysis of images taken in-transit versus out-of-transit, the difference often revealing the pixel contributing the most to the flux change during transit. Examples are shown to illustrate each test. Candidates passing data validation are submitted to ground-based observers for further false-positive elimination or confirmation/characterization.