Modeling viscosity of (Mg,Fe)O at lowermost mantle conditions

Abstract

The viscosity of the lower mantle results from the rheological behavior of its two main constituent minerals, aluminous (Mg,Fe)SiO_3 bridgmanite and (Mg,Fe)O ferropericlase. Understanding the transport properties of lower mantle aggregates is of primary importance in geophysics and it is a challenging task, due to the extreme time-varying conditions to which such aggregates are subjected. In particular, viscosity is a crucial transport property that can vary over several orders of magnitude. It thus has a first-order control on the structure and dynamics of the mantle. Here we focus on the creep behavior of (Mg,Fe)O at the bottom of the lower mantle, where the presence of thermo-chemical anomalies such as ultralow-velocity zones (ULVZ) may significantly alter the viscosity contrast characterizing this region. Two different iron concentrations of (Mg_(1–x)Fe_x)O are considered: one mirroring the average composition of ferropericlase throughout most of the lower mantle (x = 0.20) and another representing a candidate magnesiowüstite component of ULVZs near the base of the mantle (x = 0.84). The investigated pressure-temperature conditions span from 120 GPa and 2800 K, corresponding to the average geotherm at this depth, to core-mantle boundary conditions of 135 GPa and 3800 K. In this study, dislocation creep of (Mg,Fe)O is investigated by dislocation dynamics (DD) simulations, a modeling tool which considers the collective motion and interactions of dislocations. To model their behavior, a 2.5 dimensional dislocation dynamics approach is employed. Within this method, both glide and climb mechanisms can be taken into account, and the interplay of these features results in a steady-state condition. This allows the retrieval of the creep strain rates at different temperatures, pressures, applied stresses and iron concentrations across the (Mg,Fe)O solid solution, providing information on the viscosity for these materials. A particularly low viscosity is obtained for magnesiowüstite with respect to ferropericlase, the difference being around 10 orders of magnitude. Thus, the final section of this work is devoted to the assessment of the dynamic implications of such a weak phase within ULVZs, in terms of the viscosity contrast with respect to the surrounding lowermost mantle.