Mixed‐oxide and perovskite‐structure model mantles from 700–1200 km

Summary. The mantle between 700 and 1200 km is modelled using mixed oxides (MgO + FeO + SiO 2 (stishovite)) having olivine, pyroxene, and peridotite (67 per cent ol‐33 per cent px) stoichiometries and third‐order finite strain theory. It is possible to satisfy the densities, compressional and shear velocities, and bulk and shear moduli of seismic models B1 and PEM and to obtain adiabatic temperature profiles consistent with literature profiles with olivine and peridotite stoichiometries, but not with pyroxenes. Estimates of K and μ for stishovite satisfying Bl and PEM profiles indicate that the stishovite data of Mizutani, Hamano & Akimoto and Liebermann, Ringwood & Major are inconsistent with the assumed mixed‐oxide assemblages, even with a large range of pressure and temperature derivatives for K and μ. Assuming perovskite structure materials below 700 km shows that pyroxenes transformed to the perovskite structure are too dense by 1–2.5 per cent. Taking a wide range of P and T derivatives for the K and μ of MgSiO 3 (perovskite) indicates that the seismic models can be satisfied with olivines and peridotites transformed to perovskites and oxides. The MgSiO 3 (perovskite) K and μ estimates bracket the systematics estimates of Liebermann, Jones & Ring‐wood. We find only marginal evidence for relaxation of the shear modulus. The mean atomic weight calculated for the mixed‐oxide models agrees with the results of Watt, Shankland & Mao for the same region and with previous estimates for the upper mantle. Thus, an increase in iron content below 700 km appears unlikely. This result and the corresponding temperatures (lower than the Clark & Ringwood conduction geotherm) are consistent with convection below 700 km.