Beyond gas dwarfs and water worlds: hybrid envelopes among the M-dwarf sub-Neptunes characterised with JWST

arXiv:2609.34635 · astro-ph.EP · Submitted 2026-09-28 · Read on arXiv

astro-ph.EP

Submitted: 2026-09-28

Updated: 2026-10-05

Comments: Submitted to Icarus. 4 figures

License: http://creativecommons.org/licenses/by/4.0/

The gist: Sub-Neptunes orbiting M dwarfs are commonly interpreted either as rocky cores beneath primordial H/He envelopes (gas dwarfs) or as volatile-rich bodies formed beyond the water-ice line (water worlds).

Terminology

Abstract

Sub-Neptunes orbiting M dwarfs are commonly interpreted either as rocky cores beneath primordial H/He envelopes (gas dwarfs) or as volatile-rich bodies formed beyond the water-ice line (water worlds). Because the two are degenerate in mass and radius once envelope evolution is treated self-consistently, the atmospheric mean molecular weight μ is the more direct discriminant. We compile the eight M-dwarf sub-Neptunes with published JWST transmission spectra and a constraint on μ, grade each constraint by the quality of the underlying evidence, and combine it with the restricted Jeans parameter and with energy-limited escape integrated over the system lifetime, using XUV histories from M-dwarf activity-age relations. Planets are classified by what their present composition implies about their origin. At the μ= 3.8 amu threshold adopted in the gas-dwarf literature, four planets fit neither end-member: their envelopes (3.8 < μ 18 amu) are too heavy for primordial H/He, yet neither their composition nor their escape history requires a volatile-rich origin. Two of these assignments rest on molecular detections (LP 791-18 c and TOI-270 d) and two on bounded or conditional metallicities. One planet, LHS 1140 b, requires a volatile-rich origin; one, TOI-776 c, is tentatively a gas dwarf; two remain ambiguous. We propose the term hybrid worlds for the intermediate class, whose envelopes can be produced by magma-ocean equilibration, supercritical miscibility, preferential hydrogen loss or accretion interior to the ice line. The count depends on the adopted threshold (three planets at 5 amu, one at 10 amu), but no threshold leaves both end-member classes physically consistent. The sample is set by JWST target selection, so these fractions are not occurrence rates; a homogeneous survey testing whether μ is bimodal would test the proposed class directly.

Sources

Related papers