Planet Mercury is an outlier among the terrestrial planets in our solar system. It has the largest core-mass fraction (CMF = ~70%), suggesting a unique formation history. One hypothesis for Mercury's formation is the removal of mantle silicates from a bigger proto-Mercury due to a giant impact. This event would have left behind proto-Mercury's core and deep mantle rocks, which correspond with the core and bulk silicate shell of present-day Mercury. The thermodynamic models commonly used to simulate magma ocean solidification on terrestrial planets are unable to predict phase equilibria at high pressures (~20 GPa) and highly reduced conditions, the likely conditions at which proto-Mercury's mantle and core formed. Further, only very few experiments, none of which with volatile elements, are available to calibrate the existing thermodynamic models at these high pressures under highly reducing conditions. 

The main objectives of this research are to reconstruct the crystallisation products of proto-Mercury's magma ocean. This is needed to be able to estimate the bulk composition of present-day Mercury's silicate shell. 

Multi-anvil and piston cylinder experiments, with graphite capsules and starting compositions based on enstatite chondrite compositions, are performed at pressure and temperature conditions relevant to proto-Mercury's magma ocean (1 – 20 GPa, < 2200°C). Our experimental protocol allows us to simulate progressive fractional crystallisation. The phase equilibria of this first experiment are used to calculate the starting composition of the next experiment. To simulate the evolving sulfur content in the residual melt, the bulk S content is increased in each new experiment until a sulfide phase becomes present in the experimental products. To buffer the oxygen fugacity, each starting composition is prepared with the same Si/SiO2 ratio.