Platinum partitioning between metal and silicate melts: Core formation, late veneer and the nanonuggets issue

International audienceHigh-pressure, high-temperature experiments have been performed at ∼1.2 GPa and 1360–2100 °C to investigate the partitioning of Pt between a silicate melt and a metallic melt. Our experiments indicate that nanonuggets encountered in previous experiments are experimental artifacts, formed at high temperature by oversaturation caused by high oxygen fugacity during the initial stages of an experiment. Experiments at high-acceleration using a centrifuging piston-cylinder show that nanonuggets can be removed by gravity during the experiment. Formation of nanonuggets can also be avoided by using initially reduced starting materials. The presence of iron is also a key element in reducing the formation of nanonuggets. Our nanonugget-free data are broadly consistent with previous nanonuggets-filtered data, and suggest that Pt partitioning becomes independent of oxygen fugacity below an oxygen fugacity of at least IW+2. Pt is thus possibly dissolved as a neutral species (or even an anionic species) at low fO2, instead of the more common Pt2+ species present at higher fO2. Due to low concentration, the nature of this species cannot be determined, but atomic Pt or Pt− are possible options. Under core-formation conditions, Pt partitioning between metal and silicate is mostly independent of oxygen fugacity, silicate melt composition, and pressure. Partition coefficient during core formation can be expressed by the following equation: logDPt^M(metal/silicate) = 1.0348+14698/T (in weight units). Calculations indicate that the Pt content (and by extension the Highly Siderophile Elements content) of the Earth’s mantle cannot be explained by equilibrium partitioning during core formation, requiring further addition of HSE to the mantle. The mass of this late veneer is approximately 0.4% of the total mass of the Earth (or 0.6% of the mass of the mantle)