Disintegrating Exoplanets: Creating Size Constraints by Statistically Peering Through the Debris

We study two intriguing disintegrating exoplanets, Kepler-1520b and K2-22b, and attempt to constrain the size of the underlying objects. These two planets are being disintegrated by their host stars, spewing dust and debris pulled from their surface into tails that trail and precede the exoplanet in its orbit, making it difficult to discern the true nature of the object. We attempted to peer through the dust cloud to put a constraint on the maximum radii of these exoplanets. While previous studies have done this in the past by selecting shallow transit events, we attempt a new statistical approach to model the intrinsic astrophysical and photon noise distributions simultaneously. We assume that the lightcurve flux distribution is distributed as a convolution of a Gaussian photon noise component and a Raleigh astrophysical component. The Raleigh curve has a finite flux maximum, which we fit with a Hamiltonian Markov Chain. With these methods, a more accurate flux maximum may be estimated, producing a more accurate and better final value for the size of these exoplanets. To determine statistical significance, we used the python package PyMC3 to find the posterior distribution for our data with Gaussian, Rayleigh, and joint function curves and plotting it against our collected flux.

After completing this analysis, we were able to place a constraint on the maximum radius of K2-22b at approximately 12,000km. However, we were unable to constrain the radius of Kepler-1520b due to either transits obscured by forward scattering of dust or a grazing transit.