Sometime early in its history the Earth underwent two gigantic redistributions of its chemistry: a gargantuan collision that formed the Moon; separation of a metal plus sulfide core from a silicate remainder. These ‘set the scene’ for all subsequent geological (and perhaps biological) evolution. The current theory about core formation stems from a marked disparity between Hf-W and U-Pb geochronology of the mantle. The first suggests a metal-secreting event about 30 Ma after formation of the Solar System – tungsten is siderophile and would have become depleted in the mantle following segregation of a metallic core. The second points to lead partitioning into a sulfide mass descent to the core around 20-100 Ma later; assuming that lead is chalcophile. The key to explaining the disparity and validating the dual core formation hypothesis lies in establishing just how chalcophile lead is, relative to other metals that are present in the mantle (Lagos, M. et al. 2008. The Earth’s missing lead may not be in the core. Nature, v. 456, p. 89-92). The German and Russian geochemists set up experiments to determine directly the partition coefficients of lead and the other ‘volatile’ elements cadmium, zinc, selenium and tellurium between metal, sulfide and silicate melts at mantle pressures. They found that Pb and Cd are moderately chalcophile and lithophile, but never siderophile; Zn favours silicate melts, and is exclusively lithophile under mantle conditions; Se and Te are both chalcophile and siderophile, so would enter the core in both molten sulfide and metal.
The measured partition coefficients give a basis for comparing the relative proportions of the volatile elements estimated in the mantle with those predicted by the two-event model of core formation. This elegant approach strongly suggests that sulfide or iron-nickel metal segregation from the mantle to the core can explain neither the mantle abundances of the five ‘volatile’ elements nor the lead-isotope ratios in the mantle. It even questions the existence of terrestrial sulfur in the core. The postulated Moon-forming mega-impact alone could have produced the measured geochemical features of the mantle as a result of vaporisation of ‘volatile’ elements.
Mantle heat transfer by radiation
After some early speculation about efficient heat transfer in the mantle by radiation, it became generally accepted that convection and conduction dominate at depth in the Earth. Yet the Stefan-Boltzmann law has the radiant energy flux of a body increasing proportionally to the fourth power of its absolute temperature. So at deep mantle temperatures of up to 4300 K radiation ought to be significant unless mantle minerals become opaque at high pressures. Mantle mineralogy is dominated by iron-magnesium silicates that adopt the perovskite structure. High-pressure experiments with perovskites reveal surprisingly high transparency to visible and near-infrared radiation (Keppler, H. et al. 2008. Optical absorption and radiative thermal conductivity of silicate perovskite to 125 gigapascals. Science, v. 322, p. 1529-1532). It seems that a higher than expected radiative contribution to heat transfer should stabilise large plume structures in the zone above the core-mantle boundary.