Constraints on core formation from molybdenum solubility in silicate melts at high pressure

We report results from 43 new molybdenum solubility experiments performed in order to test molybdenum’s compatibility with the magma ocean hypothesis for core formation. A Walker-type multi-anvil press was used for all experiments and we investigated the pressure range of 2.5–12 GPa and temperature range of 1585–2200 °C. Eight different silicate compositions were also employed. Our data show that increasing temperature causes solubility to increase, whereas pressure has a negligible effect over the range investigated here. In general, increasing silicate melt polymerization causes solubility to decrease; however, the effect of silicate composition is best addressed by looking at the effects of individual oxides versus a universal melt parameter such as NBO/T (ratio of non-bridging oxygens to tetrahedrally coordinated oxygens). From our solubility data, we calculated metal–silicate partition coefficients at infinite iron dilution. Parameterization of our data plus data from the literature shows that there is no discrepancy between partition coefficients determined directly from experiments and those calculated from solubility data, so long as all variables are taken into account, i.e. changes in metal phase composition. Additionally, most of the experiments in the literature were conducted at pressures below 2 GPa, therefore the addition of our high pressure data set makes extrapolations to deep magma ocean conditions more accurate. We determined that the observed mantle abundance of Mo can be explained by both single-stage and multi-stage magma ocean models. Previous siderophile element studies have suggested a wide range of possible single-stage core formation conditions, from 10 to 60 GPa along the peridotite liquidus. Our results narrowed this range by constraining the P–T conditions to 40–54 GPa and 3050–3400 K. Our results also further constrained the multi-stage core formation models by limiting the depth of metal–silicate equilibration during the final impacts of accretion to 31–42% of the core–mantle boundary depth.

► Mo solubility in silicate melts is strongly dependent on silicate composition.
► Increasing temperature causes Mo solubility to increase.
► D values from solubility experiments are directly comparable to partitioning data.
► The Mo mantle depletion is consistent with single and multi-stage core formation.