An empirical calibration to estimate cool dwarf fundamental parameters from H-band spectra

TL;DR

This study develops empirical calibrations using H-band spectra to estimate fundamental parameters of mid K to mid M dwarfs, validated with interferometric data, and applies these to characterize exoplanet host stars.

Contribution

The paper introduces new empirical relationships based on H-band spectral features to determine M dwarf stellar parameters, improving radius and temperature estimates over previous models.

Findings

Calibrations achieve 73K temperature accuracy.

Radius estimates have 0.027 Rsun residuals.

Inferred planet radii are 15% larger than previous estimates.

Abstract

Interferometric radius measurements provide a direct probe of the fundamental parameters of M dwarfs, but is within reach for only a limited sample of nearby, bright stars. We use interferometrically-measured radii, bolometric luminosities, and effective temperatures to develop new empirical calibrations based on low-resolution, near-infrared spectra. We use H-band Mg and Al features to derive calibrations for effective temperature, radius and log luminosity; the standard deviations in the residuals of our best fits are, respectively, 73K, 0.027Rsun, and 0.049 dex (11% error on luminosity). These relationships are valid for mid K to mid M dwarf stars, roughly corresponding to temperatures between 3100 and 4800K. We apply our calibrations to M dwarfs targeted by the MEarth transiting planet survey and to the cool Kepler Objects of Interest (KOIs). We independently validate our calibrations by demonstrating a clear relationship between our inferred parameters and the absolute K magnitudes of the MEarth stars, and we identify objects with magnitudes too bright for their estimated luminosities as candidate multiple systems. We also use our inferred luminosities to address the applicability of near-infrared metallicity calibrations to mid and late M dwarfs. The temperatures we infer for the KOIs agree remarkably well with those from the literature; however, our stellar radii are systematically larger than those presented in previous works that derive radii from model isochrones. This results in a mean planet radius that is 15% larger than one would infer using the stellar properties from recent catalogs. Our results confirm those of previous in-depth studies of Kepler-42, Kepler-45, and Kepler-186.