High-pressure polymorphism of the iron oxides

High-pressure shock-wave data for hematite (Fe2O3) and magnetite (Fe3O4) are remarkable for the very large density increases (∼ 10 per cent at 100 GPa) associated with shock-induced transformation from the relatively close-packed corundum and spinel structures. The recently reported data of Jeanloz & Ahrens for wüstite (Fe0.94O) also reveal a major shock-induced phase change which they interpret as B1 ⇌ B2 polymorphism. Reassessment of the wüstite data, however, indicates that the zero-pressure density increase, associated with shock-induced polymorphism, is at least 10–16 per cent and very possibly as great as 18–28 per cent. The density increases for all three iron oxides thus appear to be too large to be explained in terms of geometric rearrangement (e.g. B1 ⇌ B2) of the usual Fe2+, Fe3+ and O2- ions, unless it is assumed that these ions possess substantially smaller effective radii in the high-pressure phases. In this connection, it is proposed that the more covalently bonded nickel arsenide (NiAs) and derivative structures might accommodate very dense iron oxide phases of all stoichiometries at high pressures. Systematics, involving the volume and free energy changes associated with NaCl ⇌ NiAs polymorphism in various transition metal chalcogenides, suggest a B1 ⇌ NiAs transformation pressure of ∼ 31 GPa for FeO with an estimated zero-pressure density increase of ∼ 11 per cent. An alternative, or additional, mechanism by which Fe—O bonds might be shortened significantly involves spin-pairing of 3d electrons. Calculations, based on structural data for Co3+ oxides, provide an estimated spin-pairing pressure of ∼ 50 GPa for FeO with a density increase of 9 per cent at the transition. While the occurrence of substantial ‘mixed-phase’ regimes along the Fe2O3 and FeO Hugoniots seems to preclude an exclusively electronic interpretation of the observed phase changes, it is quite possible that spin-pairing augments the density increases due to NaCl ⇌ NiAs polymorphism. Comparative studies of the high-pressure phase behaviour of MnO, FeO, CoO and NiO should resolve the relative roles of structural and electronic transitions. The high-pressure stability of iron oxides in covalently bonded structures of the NiAs type would enhance their solubility in molten Fe, and thus provide an attractive scenario for the incorporation of oxygen into the Earth's core.