Disentangling the influence of observational constraints on Mercury's interior evolution

The forthcoming orbital phase of the ESA/JAXA mission BepiColombo is expected to provide new and updated constraints on Mercury’s interior structure and evolution, highlighting the need to reassess and strengthen existing thermal evolution models. Significant uncertainties remain regarding the thermal, transport and elastic properties of Mercury’s mantle and core. Notably, the mantle composition is poorly known. It could be composed of variable proportions of forsterite and enstatite [1], which exhibit markedly different thermal conductivities and rheological behavior [2, 3]. The core can be approximated as an Fe-Si alloy, although it likely also contains other light elements such as S and C. Yet, the core Si fraction is also uncertain and substantially impacts key properties such as density, liquidus, entropy, and thermal conductivity [4]. To restrict the range of uncertainties regarding these parameters, we conducted an extensive set of Monte-Carlo simulations of the coupled core-mantle thermal evolution of Mercury, filtering the results according to various combinations of observational constraints whose choice can significantly influence the number of successful models and their typical evolution.

We employed a 1D parameterized mantle convection model (TEMPURA [5]) in which we incorporated the influence of pressure- and temperature-dependent thermal conductivity, heat capacity, thermal expansivity, and density for forsterite, enstatite, and diopside, which we considered in variable proportions. We coupled this mantle model with an evolution model of the core simulating inner core crystallization and the development of a thermally stratified layer at the top of the liquid core [6] using  self-consistent Fe-Si thermodynamic properties.

We conducted Monte-Carlo simulations of Mercury’s thermal evolution, varying the following mantle parameters: initial temperature profile, forsterite over enstatite ratio, reference mantle viscosity, and crustal enrichment factor in heat-producing elements. Additionally, we varied some core parameters: starting core-mantle boundary (CMB) temperature, core radius, thermal conductivity, and melting entropy.

We filtered the models based on the following observational constraints: (1) final crust thickness between 15 and 60 km; (2) onset of global contraction within the first 500 Myr; (3) dynamo generation between 500 and 1000 Myr and at the present day; (4) bulk of the volcanic crust produced within the first   Gyr, and (5) contraction accumulating at a decreasing rate over the evolution.

Models that satisfy constraints (1), (2), and (3) exhibit a high reference viscosity (10²² Pa s), low initial CMB temperature (< 2000 K), a  core radius between 2000 and 2010 km, and require a high core melting entropy, and a low core thermal conductivity. The initial mantle temperature, mantle enstatite fraction, and crustal enrichment factor are less decisive. Upon including  the constraint (4) and (5), the results shift to lower reference viscosities (10²¹ Pa s) and large crustal enrichment factors (>7).

These discrepancies can be understood in terms of the competing thermal effects imposed by the different constraints. High reference viscosities inhibit efficient planetary cooling, which facilitates the maintenance of a convective layer in the core for 4.5 Gyr, a prerequisite for the current dynamo (3). However, higher reference viscosities (10²³ Pa s) also suppress melt production, leading to crustal thicknesses that are too small to satisfy the constraint (1), and may even produce early-stage expansion, inconsistent with constraint (2). Low initial CMB temperatures are essential for nucleating the inner core before 1 Gyr, which is an important driver for the ancient dynamo (3). Imposing the additional requirement that the bulk of the crust and contraction forms within the first Gyr (4-5) shifts the preferred solutions to  lower reference viscosities and higher enrichment factors, both of  which promote efficient mantle cooling.

In summary, the different observational constraints favor markedly different mantle and core properties, highlighting the challenge of  reproducing Mercury’s volcanic, tectonic, and magnetic evolution within a single thermal history model. 

Therefore, accurately characterizing Mercury’s interior evolution requires a careful assessment of which observations are most robust and how they should be incorporated into thermal evolution models.

[1] Saracino et al. (2025). Chem. Geol., 683, 122777, doi: 10.1016/j.chemgeo.2025.122777. [2] Zhang et al. (2019). Earth Planet. Sci. Lett., 519, 109-119, doi:10.1016/j.epsl.2019.04.048. [3] Guo et al. (2024). Geochem. Geophys. Geosyst., 25(6), e2023GC011419, doi:10.1029/2023GC011419. [4] Edmund et al. (2022). Nature Comm., 13, 387, doi:10.1038/s41467-022-27991-9. [5] Baumeister et al. (2023). A\&A, 675, A122, doi:10.1051/0004-6361/20224579. [6] Davies et al. (2024). Earth Planet. Sci. Lett., 641, 118812, doi:10.1016/j.epsl.2024.118812.