Leveraging the Power of a Planet Population: Mass-Radius Relation, Host Star Multiplicity, and Composition Distribution of Kepler's Sub-Neptunes

With the advent of large, dedicated planet hunting surveys, the search for extrasolar planets has evolved into an effort to understand the properties and formation of a planet population whose characteristics continue to surprise the provincial perspective we've derived from our own Solar System. The Kepler Mission in particular has enabled a large number of these studies, as it was designed to stare simultaneously at thousands of stars for several years and its automated transit search pipeline enables fairly uniform detection criteria and characterizable completeness and false positive rates. With the detection of nearly 5000 planet candidates, 80% of which are smaller than 4 R_Earth, Kepler has especially illuminated the unexpectedly vast sub-Neptune population. Such a rich dataset provides an unprecedented opportunity for rigorous statistical study of the physics of these planets that have no analogs in our Solar System.Contributing to this endeavor, I present the statistical characterization of several aspects of this population, including the comparison between Kepler's planet candidates and low-mass occurrence rates inferred from radial velocity detections, the relationship between a sub-Neptune's mass and its radius, the frequency of Kepler planet candidate host stars which have nearby visual companions as revealed by follow-up high resolution imaging, and the distribution of gaseous mass fractions that these sub-Neptunes could possess given a rock-plus-hydrogen composition. To do so, I have used sophisticated statistical analyses such as Monte Carlo simulations and hierarchical Bayesian modeling to tie theory more closely to observations and have acquired near infrared laser guide star adaptive optics imaging of 196 Kepler Objects of Interest. I find that even within this sub-Neptune population these planets are very diverse in nature: there is intrinsic scatter in masses at a given radius, the planet host stars have visual companions at a wide range of separations, and the composition distribution spans two orders of magnitude, with a peak at 1% hydrogen and helium by mass. There is much work to be done to explain this diversity quantitatively, and especially to tie these results to various planet formation scenarios; I have no doubt that many more surprises await us