Thermal Conductivity of Lower Mantle Minerals and Heat Flux Across the Core-Mantle Boundary

Abstract:

The thermal conductivity properties of the minerals comprising the Earth’s lowermost mantle control the core-mantle boundary heat flux, and are therefore critical properties for determining the thermal state and evolution of the Earth’s interior. Here we present measurements of the thermal conductivity of lower mantle oxides and silicates as a function of pressure, temperature, and iron content determined in the laser-heated diamond anvil cell using a combination of measurements and 3-D modeling. Our models and measurements demonstrate that the measured steady-state temperature and its increase with increasing laser power depend on the sample thermal conductivity as well as the experimental geometry, enabling measurements of the pressure- and temperature- dependence of lattice thermal conductivity in the laser-heated diamond anvil cell. We applied this technique to iron-bearing silicate perovskites and MgO at lower mantle pressure and temperature conditions. For MgO, we determine the increase in thermal conductivity k with density ρ to be ∂lnk/∂lnρ=4.7±0.6, which is in agreement with results obtained using other experimental and computational techniques. For (Mg_0.8,Fe_0.2)SiO₃ perovskite, we find ∂lnk/∂lnρ=2.9±0.6. We use these values in combination with independent computational and experimental results to determine thermal conductivity of lower mantle minerals up to core-mantle boundary conditions. We combine the mineralogical thermal conductivity estimates in a composite model and include an estimate for the radiative contribution to thermal conductivity. Our new value of the thermal conductivity of the lowermost mantle is ~5-6 W/m/K and is sensitive to the details of the lower mantle assemblage, but is relatively insensitive to pressure and temperature. We combine our mantle thermal conductivity with models for the lower mantle boundary layer to generate a series of two-dimensional maps of core-mantle boundary heat flux, which emphasize the importance of lateral variations in phase and boundary layer thickness. Our values imply a total core-mantle boundary heat flow of 6-8 TW, which is sufficient to drive plumes and convection, is consistent with current geochemical estimates for mantle heat content, and permits a slow growth rate for the inner core.