Computer simulation of the Mg 2SiO 4 phases with application to the 410 km seismic discontinuity

Abstract

Molecular dynamics (MD) simulation is used to predict the structure and elasticity of the olivine (α), modified-spinel (β) and spinel (γ) forms of Mg 2SiO 4 at high temperatures and high pressures. The interionic potential is taken to be the sum of pairwise additive Coulomb, van der Waals attraction, and repulsive interactions. In order to take account of non-central forces in crystals, the breathing shell model (BSM) is used for simulation, in which the repulsive radii of O ions are allowed to deform isotropically under the effects of other ions in the crystal. The same potential model is used for the three Mg 2SiO 4 phases. Required energy parameters, including the breathing parameters of O ions, were obtained empirically using the measured structural and elastic properties of the three phases. The MD simulation with BSM is found to be very successful in reproducing accurately the observed values for the three phases, including the structural parameters and individual elastic constants at ambient conditions, the temperature and pressure derivatives of bulk and shear moduli, and the volume thermal expansivity and volume compression over wide temperature and pressure ranges. We further apply MD simulation to predict the density and seismic velocity contrasts between α- and β-Mg 2SiO 4 at high temperature and high pressure conditions corresponding to the 410 km mantle discontinuity, and compare the simulated results with seismologically observed data. The simulated velocity contrasts for both P- and S-waves support a previous estimate of maximum 50% by volume for the olivine content at the 410 km discontinuity. In contrast, the simulated density difference between the two phases, when compared with the seismologically reported density jump (PREM) at the 410 km discontinuity, requires implausibly too high olivine content of about 90% by volume at this discontinuity. This inconsistency between the simulated density and seismic velocity data may indicate that the bulk composition at the 410 km depth mantle is different from that usually considered such as the pyrolite model, or that the seismic density data at the 410 km mantle contains significant overestimation.