Formation of the Earth's Core

Abstract

The formation of the Earth's core is intimately linked to the manner in which the planet accreted. Small bodies (i.e., planetesimals), if they accreted sufficiently early, may have melted and differentiated due to the decay of short-lived radioisotopes such as 26Al. For Earth-sized bodies, the final stages of accretion involved impacts between comparably sized objects, many of which were probably already differentiated. These impacts were sufficiently energetic to melt large portions of the planet, although the lateral extent of the melting and the lifetime of the resulting magma oceans are both currently poorly understood. Metal–silicate segregation, as required for core formation, can occur by several mechanisms depending on the degree of melting. The percolation of molten iron through solid silicates, which depends on the wetting properties of the melt (dihedral angle), melt fraction and shear stress, is unlikely to be an efficient mechanism in an Earth-sized planet. In a magma ocean, on the other hand, the separation of molten iron from molten silicates is very rapid and efficient. Core formation itself releases gravitational energy and thus generates further heating. Thus, once the Earth grew sufficiently large, melting and core formation were inevitable. Because of the stochastic nature of late-stage impacts, large bodies of iron (the cores of planetesimals) were delivered to the Earth at discrete intervals. The extent to which these iron bodies emulsified during impacts is currently uncertain but has important implications for the chemical evolution of the core and mantle. The hafnium–tungsten (Hf–W) chronometer has been used to constrain the core formation time of the Earth to 30–50 Myr after solar system formation. This timescale, though it cannot fully capture the growth of the core by multiple, discrete impacts, is in good agreement with numerical accretion models. The Hf–W results show that the cores of impactors at least partly reequilibrated with the silicate mantle, suggesting that emulsification during impacts is a relatively efficient process. Mantle siderophile element concentrations are higher than would be expected based on low-pressure partition coefficients. Metal–silicate partition coefficients determined at high P–T conditions (20–40 GPa and 2500–4000 K) can explain the abundances of moderately siderophile elements. These partition coefficients are again consistent with metal–silicate equilibration occurring near the base of a magma ocean, though the true picture is likely to involve multiple magma oceans with different (and evolving) conditions of pressure, temperature, and oxygen fugacity. The highly siderophile element (HSE) concentrations require the addition of chondritic material to the mantle (the ‘late veneer’) after core formation ended. The identity of the light element(s) in the core is currently uncertain, due mainly to the paucity of experiments at the required P–T conditions. As for the siderophile elements, the light element content was likely set during core formation through metal–silicate partitioning. Based on cosmochemical arguments and a range of recent studies, the major light element in the core is most likely silicon, with a lesser amount of O and ~ 2 wt%S.