2025ApJ...993..235R

Exoplanet mass and radius inferences fundamentally rely on host star mass and radius inferences. Despite the importance of host star mass, radius, and elemental abundance inferences for the derivation of exoplanet internal structure constraints, published constraints have often been based on inferences that are not self-consistent. For 24 dwarf stars hosting terrestrial exoplanets, we use astrometric and photometric data plus high-resolution spectroscopy to infer accurate, precise, homogeneous, and physically self-consistent photospheric and fundamental stellar parameters, as well as elemental abundances. We infer updated planetary masses and radii using these data plus Doppler and transit observables and then use the complete data set to derive constraints on the CMFs of these terrestrial exoplanets. We find that the population of resonant or likely formerly resonant terrestrial exoplanets represented by Kepler-36 b and Kepler-105 c has a significantly lower mean CMF than the rest of the terrestrial exoplanets in our sample. Their resonant configurations suggest that they migrated inward from more distant formation locations, and we attribute their low densities to the incorporation and retention of significant amounts of water during their formation. We confirm that the ultra-short-period exoplanets 55 Cnc e and WASP-47 e have densities inconsistent with pure-rock compositions. We propose that they are both the stripped cores of mini-Neptunes, and we associate their low densities with the presence of significant amounts of hydrogen, helium, water, and/or other volatiles in their interiors. We verify that our results are independent of stellar parameter and elemental abundance inference approach and therefore robust.