Metal‐silicate thermochemistry at high temperature: Magma oceans and the “excess siderophile element” problem of the Earth's upper mantle

Abstract

Recent theoretical considerations indicate that Earth should have been at least partly molten at the end of accretion and perhaps sufficiently to produce a magma ocean. Metal‐silicate partition coefficients applicable to a magma ocean (3000 K‐4000 K and up to 130 GPa) would be valuable to test such models against well‐known mantle siderophile element abundances, but they have not been measured. However, an extrapolation of existing low temperature data to these extreme conditions has recently been attempted [Murthy, 1991]. Murthy's results seem to account for many apparent excesses of siderophile elements in the mantle. We examined his extrapolation method and found it to be inconsistent with published temperature dependencies for metal‐silicate partition coefficients. We also attempted an extrapolation to magma ocean temperatures, based on known chemical behavior for several elements, and using published metal‐silicate partition coefficients. If the chemistry quantified by the low‐temperature data is applicable at high temperature, an important assumption, then our results indicate that high temperature alone will not help ameliorate the excess siderophile element problem of the upper mantle, in contrast to the conclusions of Murthy [1991]. For most elements, a modest increase in siderophile behavior is predicted with rising temperature if an iron‐wüstite redox buffer is paralleled. But long‐range extrapolation of experimental data containing even modest experimental errors is hazardous. Extrapolated high‐temperature partition coefficients can differ by orders of magnitude for a given element, even though the input from independent studies is consistent within quoted errors. Direct experimental measurements for at least some of the siderophile elements will be necessary to accurately assess siderophile element behavior in a magma ocean. The excess siderophile element problem of the Earth's upper mantle remains unsolved.