Constraints from metal-silicate partitioning on accretion, core formation and volatile addition to the growing earth (Invited)

The currently accepted model of planetary formation is that a large number of moon to Mars-sized planetary embryos with different metal/silicate/volatile ratios grew from the dust and gas surrounding the young sun in ~1 M.yr. As Earth accreted from such embryos through a succession of impacts, culminating in the moon-forming giant impact, the metallic core segregated and there may have been some loss of the most volatile elements. The aim of this presentation is to show how the chemical and isotopic composition of silicate Earth may, in conjunction with high pressure-high temperature experiments, be used to deduce the nature and timing of these different processes. Silicate Earth is, for example, depleted in siderophile elements such as Ni, Co, W and Mo relative to undifferentiated meteorites because these elements were partitioned into the core. We now have a large number of experimental data on partitioning of elements between metal and silicate which enable us to determine how Earth segregated its core. If we begin with the assumptions that core segregation was continuous during accretion and that metal and silicate fully equilibrated, simultaneous consideration of the depletion factors of a large number of elements in silicate Earth lead to the following general conclusions: (1) The average pressure of core segregation on Earth was high >30 GPa , implying depths of >800 km. (2) Earth began as a small, strongly reduced body and became more oxidised as it grew. (3) Si (~5%) and S (~2%) are major components of the “light” element in Earth’s core. If we relax the assumption of full equilibration (ie allow for the addition of metal to the core without reaction with the mantle) then the average pressure of core segregation increases but the other conclusions are largely unaffected. There has been considerable recent discussion of the timing of addition of moderately volatile elements (elements such as Pb, In, Tl which would condense at temperatures of 500-1000K from a solar gas) to the Earth. The pattern of depletion of these elements in the silicate Earth is consistent with their having been added as the core formed, during the main phase of terrestrial accretion rather than as a "late veneer" after core formation ceased. In that case the differences between accretionary timescales given by 182Hf-182W, 235,238U - 207,206Pb and 205Pb-205Tl isotopic systems can be explained either by addition of a small amount of core (~10%) during the moon-forming impact or by a certain amount of metal-silicate disequilibrium during accretion.