Limits on forming coreless terrestrial worlds in the TRAPPIST-1 system

TL;DR

The paper addresses why TRAPPIST-1 planets exhibit a density deficit and tests whether a coreless interior can explain the data. It develops a thermodynamic framework for oxygen partitioning between metal cores and silicate mantles, yielding a DO(P,T) relationship derived from high-pressure experiments, and couples this to interior structure models to track FeO formation in cores. The main finding is that $D_O$ increases with pressure and temperature, driving core oxidation and precluding coreless interiors up to about $3.5\,M_\oplus$, with TRAPPIST-1 planets expected to host modest core-oxide contents ($\sim 0.01$–$0.05$ wt% under Earth-like bulk abundances). Across the M–R diagram, self-consistent models show coreless interiors are unnecessary unless stellar Fe/Si is significantly sub-solar, a conclusion supported by JWST-era atmospheric constraints that favor thin or absent atmospheres. The work links planetary interior chemistry to formation environments and stellar metallicity, with implications for atmospheric degassing and water inventories in M-dwarf systems.

Abstract

With seven temperate Earth-sized planets revolving around an ultracool red dwarf, the nearby TRAPPIST-1 system offers a unique opportunity to verify models of exoplanet composition, differentiation, and interior structure. In particular, the low bulk densities of the TRAPPIST-1 planets, compared to terrestrial planets in our solar system, require either substantial amount of volatiles to be present or a corefree scenario where the metallic core is fully oxidised. Here, using an updated metal-silicate partitioning model, we show that during core-mantle differentiation oxygen becomes more siderophile (iron-loving) implying larger planet radii. For the seven TRAPPIST-1 planets, however, we find that they are not sufficiently large to oxidise all the iron in the core, if they differentiate from an Earth-like composition. Oxygen partitioning in rocky worlds precludes coreless planets up to about 4 Earth masses. The observed density deficit in the TRAPPIST-1 planets, and more generally in M dwarf systems if confirmed by future observations, may be explained by system-dependent element budgets during planet formation, which are intrinsically linked to their stellar metallicity.

Figures (3)

• Figure 1: Core–mantle partition coefficient of oxygen as a function of pressure and temperature. Left: Experimental data for metal–silicate partitioning of O obtained at high pressures up to 104 GPa. Horizontal dashed line located at $D_O$=0.83 indicates full oxidation of the core (see the main text), and therefore distinguishes 'terrestrial' from 'coreless' planets. It is clear that, within the investigated pressure range, $D_O$ asymptotically approaches the dashed line, but never reaches the core-free regime, albeit higher pressures and temperatures promote the sequestration of O into the metallic core, i.e. larger $D_O$. Right: The same dataset for O partitioning plotted as a function of reciprocal temperature. Dashed lines are fitted results (Eq. \ref{['eq7']}) showing the dependence of O partitioning on temperature at a given pressure or vice versa.
• Figure 2: Equilibrated core oxygen mass fraction as a function of planetary mass. Left: Results within the pressure range constrained by experimental data. Calculated core mass fractions are compared with independent estimates for Earth and Mars, and the mass range of the TRAPPIST-1 planets is also indicated. Right: Extrapolation of the metal–silicate equilibration model to higher masses and pressures. The oxygen mass fraction approaches an upper limit of 0.22, corresponding to complete oxidation of the core as FeO. In reality, this limit would be approached asymptotically rather than a sharp transition shown at 3.5 M$_\oplus$. Light grey areas illustrate the uncertainty on oxygen partitioning (Eq. \ref{['eq7']}).
• Figure 3: Left: Mass-radius diagram compared to the TRAPPIST-1 planets. Right: mass-normalized density diagram. Densities are normalized to the nominal model of a pure Fe core with solar Fe/Si abundances (orange line). The metal–silicate-equilibrated curve where oxygen partitioning is self-consistently modeled for solar Fe/Si (blue solid line) plots between the end-members of a pure Fe core (orange line) and core-free model (pink line). When the stellar Fe/Si estimate from Unterborn2018 is used, the line shifts to lower densities (blue dash-dotted line), reasonably fitting most TRAPPIST planets.