Elemental Abundances of Kepler Objects of Interest in APOGEE. I. Two Distinct Orbital Period Regimes Inferred from Host Star Iron Abundances

TL;DR

This study uses APOGEE spectra to analyze the chemical abundances of Kepler exoplanet host stars, revealing a correlation between host star metallicity and the orbital period of small planets, with a critical period around 8 days.

Contribution

It confirms the reliability of APOGEE abundance measurements for Kepler host stars and identifies a new correlation between stellar [Fe/H] and planetary orbital periods.

Findings

Small planets with periods less than 8 days orbit more metal-rich stars.

The critical period for this correlation is approximately 8.3 days.

No significant correlation found between host star temperature and orbital period.

Abstract

The Apache Point Observatory Galactic Evolution Experiment (APOGEE) has observed $\sim$600 transiting exoplanets and exoplanet candidates from \textit{Kepler} (Kepler Objects of Interest, KOIs), most with $\geq$18 epochs. The combined multi-epoch spectra are of high signal-to-noise (typically $\geq$100) and yield precise stellar parameters and chemical abundances. We first confirm the ability of the APOGEE abundance pipeline, ASPCAP, to derive reliable [Fe/H] and effective temperatures for FGK dwarf stars -- the primary \textit{Kepler} host stellar type -- by comparing the ASPCAP-derived stellar parameters to those from independent high-resolution spectroscopic characterizations for 221 dwarf stars in the literature. With a sample of 282 close-in ($P<100$ days) KOIs observed in the APOGEE KOI goal program, we find a correlation between orbital period and host star [Fe/H] characterized by a critical period, $P_\mathrm{crit}$= $8.3^{+0.1}_{-4.1}$ days, below which small exoplanets orbit statistically more metal-enriched host stars. This effect may trace a metallicity dependence of the protoplanetary disk inner-radius at the time of planet formation or may be a result of rocky planet ingestion driven by inward planetary migration. We also consider that this may trace a metallicity dependence of the dust sublimation radius, but find no statistically significant correlation with host $T_\mathrm{eff}$ and orbital period to support such a claim.