A Spectroscopic Analysis of the California-Kepler Survey Sample. II. Correlations of Stellar Metallicities with Planetary Architectures

We present independent and self-consistent metallicities for a sample of 807 planet-hosting stars from the California-Kepler Survey from an LTE spectroscopic analysis using a selected sample of Fe I and Fe II lines. Correlations between host-star metallicities, planet radii, and planetary architecture (orbital periods-warm or hot-and multiplicity-single or multiple) were investigated using nonparametric statistical tests. In addition to confirming previous results from the literature, e.g., that overall host-star metallicity distributions differ between hot and warm planetary systems of all types, we report on a new finding: when comparing the median metallicities of hot versus warm systems, the difference for multiple super-Earths is considerably larger when compared to that difference in single super-Earths. The metallicity cumulative distribution functions of hot single super-Earths versus warm single super-Earths indicate different parent stellar populations, while for sub-Neptunes this is not the case. The transition radius between sub-Neptunes and sub-Saturns was examined by comparing the APOGEE metallicity distribution for the Milky Way thin disk in the solar neighborhood with metallicity distributions of host stars segregated based upon the largest known planet in their system. These comparisons reveal increasingly different metallicity distributions as the radius of the largest planet in the systems increases, with the parent stellar metallicities becoming significantly different for R_p > 2.7 R_⊕. The behavior of the p-values as a function of planet radius undergoes a large slope change at R_p = 4.4 ± 0.5 R_⊕, indicating the radius boundary between small and large planets.