Mass-dependent Dynamics of Terrestrial Exoplanets Using ab initio Mineral Properties

We present new modelling results for the internal structure and convective dynamics of large terrestrial (rocky) exoplanets. For upto 20 Earth masses (Me) our results show pressure and temperature (P,T) of several Terapascal (TPa) and 10000 Kelvin in the silicate mantle.
Recent mineral physics predictions show stepwise dissociation into SiO_2 and MgO of the main magnesium-silicate mineral of Earth’s mantle under these P,T conditions (Umemoto et al., 2017).
Using material properties from these models we have modelled the internal structure of planets in the range 1-20 Me with an Earth-like core mass fraction of 0.3. We found that full dissociation into oxides occurs in planets with M > 13 Me, where at the core mantle boundary P > 2.4 TPa.
Our Rayleigh-Benard mantle convection results for planet mass 1-20 Me show strong differences in the internal structure and the convection dynamics between different cases.
First, due to increasing pressure the number of phase transitions increases fom zero in the smallest case to four, for cases > 13 Me. Furthermore, we observe three regimes of convective dynamics, with: 1) smaller planets (< 4 Me), showing vigorous convection, 2) intermediate cases (< 12 Me), with sluggish penetrative convection, concentred in a single shallow zone of higher flow velocity, and 3) large planets, (> 12 Me), with vigorous convection in top and bottom zones,
separated by a high viscosity mid-mantle with sluggish convection.
These regimes are directly related to the pressure dependence of mantle viscosity, first increasing then decreasing due to pressure weakening. Here the planet mass is the control variable because it sets the mantle pressure range.
For the larger planet cases with a bottom layer of oxides, and reduced viscosity, we observe vigorous convection and small scale structure in the deepest part of the mantle that interacts with the dissociation phase boundary. This impacts the heat-flux from the core and the viability of core dynamo processes.