A review of the literature indicates that a reasonable estimate of the composition of the earth's core is iron with Ni_0–5, Si_10–25 (wt.%). Thermodynamic calculations and comparison of chondritic with terrestrial abundances indicate that 1 wt.% each of Mn, P, and Cr might also be present. A core of this composition was probably in chemical equilibrium with the mantle at the time of core formation because:

1.

(1) The reactions 2Fe + SiO₂ = 2FeO + Si and Fe₂SiO₄ + 2Ni = Ni₂SiO₄ + 2Fe proceed further to the right at the T and P values prevailing at the core-mantle boundary than at lower temperatures, thus supporting the presence of Si in the core and the relatively high Ni concentration of the mantle;

2.

(2) the Fe3+Fe2+ ratios in mantle materials indicate oxygen fugacity values close to that of the Fe-Fe_1−xO buffer; and

3.

(3) the apparent partitioning of Au and similar elements between the core and the mantle is close to that of pallasites.

The anomalously high abundance of Cu in the upper mantle can be explained by enrichment through partial melting. Volcanic gases are not likely to represent the composition of volatile elements at the core-mantle boundary, and hence cannot be regarded as valid criteria of disequilibrium at the boundary. Available data on reaction kinetics suggest that a disequilibrium state would be unlikely during core formation.