An Exomoon Survey of 70 Cool Giants and the New Candidate Kepler-1708 b-i

Long-period, giant planets are well-suited for forming and retaining exomoons, yet have received little attention previously. Geometric bias greatly attenuates the fraction of such worlds that transit their star, and of those many undergo just 1 or 2 transits even within Kepler’s 4+ year baseline. We curate a sample of 70 such candidate planets from the literature, all found using Kepler data, and subject them to a systematic search for exomoons using photodynamics and Bayesian model comparison. One object is found to pass all of detection tests, as well as a battery of subsequent vetting efforts, for which the host planet is dubbed Kepler-1708b after its successful validation. With a ~2 year period around a 15.7th magnitude Solar-type star, the approximately Jupiter-sized planet undergoes just two transits and has a nearly-circular orbit at 1.6 AU. Light curve fits recover a 4.8-sigma moon-like transit signal intermixed with that of the planet, with the moon-model favoured at a Bayes factor of 11.9. Injection-recovery tests into the target’s light curve reveal the false-positive probability of the signal to be 1%. This 2.6 Earth-radius exomoon candidate appears to be on an approximately coplanar orbit, 12 planetary radii from its host, and provides a challenge to traditional moon formation models. Here, we will also present new calculations predicting the signal’s detectability with HST/JWST, by propagating the posteriors into the future and performing multiple injection-recoveries. We conclude by discussing the believability of this signal and the steps necessary to confirm or refute it’s existence.