Thermal and chemical evolution of Ganymede's primitive core

Ganymede is the largest icy moon in the solar system. The Galileo spacecraft discovered the presence of a magnetic field that is generated in an iron-rich core (Kivelson et al., 1996). Gravity data from the Galileo mission suggest a low value of the reduced Moment of Inertia (MoI) of 0.3115 (Anderson et al., 1996), which indicates a high degree of differentiation. Schubert et al. (1996) proposed a three-layer structure: an iron-rich liquid core, coherent with magnetic data, a silicate mantle, and a hydrosphere. The present work investigates the thermal evolution of Ganymede’s rocky core to assess the conditions under which the rocky interior can differentiate, leading to the formation of an iron core. The first step of the study addresses the possible end-members for the rocky composition. Two chondritic compositions have been proposed for Ganymede’s rocky interior: LL chondrite (Kronrod, Kuskov, 2006) that can explain the low Fe/Si ratio inferred from the density derived from mass and moment of inertia (MoI) or a carbonaceous chondrite (CI type) which is more probable during the accretion beyond the snow line (Néri et al., 2020). Using the bulk elementary composition and the abundances of the different phases, the elementary composition of each phase is calculated. The silicate fractions have a very similar Mg# and the main difference is the much larger fraction of iron-rich phases in the CI chondrite. Perple_X,a thermodynamic calculation tool that determines the phases from elementary composition,was used to study the mineralogical evolution of the silicate phase in the (P,T) domain relevant to Ganymede. In the case of the LL composition, we consider the hydration of the rock during the accretion and differ- entiation between the rocky core and the hydrosphere. The second step simulates the thermal evolution of the rocky core starting after the differentiation of the hydrosphere (primitive core). The decay of long-lived radioactive elements (K, U, Th) provides thermal energy that heats up the interior. Thermal energy can be transferred by either conduction or convection. The onset of convection in a solid material depends on its viscosity (ratio of stress to strain rate) that depends on parameters such as pressure, temperature, grain size, and stress. Hydrated silicates have a viscosity that strongly depends on stress whereas dry silicates have a viscosity that depends mainly on temperature. More than 30 numerical simulations were performed to investigate the effect of parameters including the time of formation of the primitive core (or time of differentiation of the hydrosphere), the type of chondrite origin (LL chondrites have a larger amount of radioactive elements), the initial temperature profile, and the sensitivity to the rheology parameters. The effect of numerical parameters such as initial perturbations in temperature and strain rate and maximum viscosity contrasts were also investigated.Simulations performed in 3D spherical geometry show that, in the domain of investigated parameters, convection does not happen in the hydrated silicates before dehydration. Dehydration occurs in the center leading to a structure in two layers: an upper layer, about 150 km thick, of hydrated silicates sitting on top of a dry silicates core. As temperature increases in the core, the eutectic temperature of the Fe-FeS sys- tem (Buono, Walker, 2011) is reached before convection in the dry silicates starts. Such an event happens between 1.5 and 2.5 Gyr, leading to a potential formation of the iron-rich liquid core. The percolation of the iron-rich liquid phase would have significant effect on the core dynamics as the deep silicates would become less dense than the upper layer of hydrated silicates. Even without this effect, we observe an onset of convection when the temperature in the core reaches 1600 K when dry silicates have a viscosity low enough for convection to start. We note that convection does not start in all models: when the differentiation of the hydrosphere is late the rocky interior remains in a con- ductive state until present, the rocky interior remains in a conductive state until present. The differentiation of the iron core and its effect on the interior dynamics is not yet implemented in the model. It would lead to an earlier onset of the convection. The convection step is short on geological time steps (a few 10s to 100s of Myr). It has two consequences. First, it dehydrates the upper layer and eventually allows for silicate melt- ing. Second, it cools down the interior very efficiently, reducing the temperature and stopping the convection process. By bringing the newly dehydrated iron-rich silicates in the center, it may lead to a second step in the formation of the iron core. It also creates a pulse in heat flux that may have a major consequence on Ganymede’s global interior dynamics. The upcoming ESA mission JUICE that will orbit Ganymede will provide additional information on the interior structure of Ganymede that will help understand the evolution of its core. Such models can be extended to the evolution of the silicate cores of other icy moons such as Titan and Europa that will be visited by the Dragonfly mission and the Europa Clipper mission, respectively.  Anderson JD, Lau EL, Sjogren WL, Schubert G, Moore WB., 1996. Nature. 384, 6609. 541– 543.Buono AS, Walker D, 2011.Geochimica et Cosmochimica Acta.  75, 8. 2072–2087.Kivelson MG, Khurana KK, Russell CT, Walker RJ, Warnecke J, Coroniti FV, Polanskey C, Southwood DJ, Schubert G. 1996. Nature. 384, 6609. 537–541.Kronrod VA, Kuskov OL., 2006. Geochemistry International. 44, 6. 529–546.Néri A, Guyot F, Reynard B, Sotin C, 2020. Earth and Planetary Science Letters. 530. 115920.Schubert G, Zhang K, Kivelson MG, Anderson JD, 1996. Nature. 384, 6609. 544–545.