Updated Compositional Models of the TRAPPIST-1 Planets

Abstract: After publication of our initial mass-radius-composition models for the TRAPPIST-1 system in Unterborn et al. (2018), the planet masses were updated in Grimm et al. (2018). We had originally adopted the data set of Wang et al., 2017 who reported different densities than the updated values. The differences in observed density change the inferred volatile content of the planets. Grimm et al. (2018) report TRAPPIST-1 b, d, f, g, and h as being consistent with <5 wt% water and TRAPPIST-1 c and e has having largely rocky interiors. Here, we present updated results recalculating water fractions and potential alternative compositions using the Grimm et al., 2018 masses. Overall, we can only reproduce the results of Grimm et al., 2018 of planets b, d and g having small water contents if the cores of these planets are small (<23 wt%). We show that, if the cores for these planets are roughly Earth-sized (33 wt%), significant water fractions up to 40 wt% are possible. We show planets c, e, f, and h can have volatile envelopes between 0-35 wt% that are also consistent with being totally oxidized and lacking an Fe-core entirely. We note here that a pure MgSiO$_3$ planet (Fe/Mg = 0) is not the true lowest density end-member mass-radius curve for determining the probability of a planet containing volatiles. All planets that are rocky likely contain some Fe, either within the core or oxidized in the mantle. We argue the true low density end-member for oxidizing systems is instead a planet with the lowest reasonable Fe/Mg and completely core-less. Using this logic, we assert that planets b, d and g likely must have significant volatile layers because the end-member planet models produce masses too high even when uncertainties in both mass and radius are taken into account.