Constraining the style of magma-ocean crystallisation by present-day Earth structure: a coupled thermodynamic-geodynamic approach

Earth began life as a globally molten ball, a "magma ocean", following the vast energy release of its accretion. While evidence for such a magma ocean is preserved on the Moon's surface, there is no direct geochemical or geophysical evidence preserved on the Earth's surface.

As the magma ocean cooled and froze, crystals settled to form the first solid mantle, and the shrinking magma ocean is thought to have become progressively enriched in iron and e.g. radioactive trace elements. Accordingly, the last droplets of the magma ocean, which ultimately stabilized the first crust on Earth, are thought to be extremely enriched in these elements. The differentiation of the rocky Earth in the magma-ocean stage has far-reaching implications for the long-term evolution of the Earth interior, the dynamics of which sustain life-friendly conditions on its surface.

In this project, we propose to address an important, but previously neglected process during magma-ocean crystallization. Recent work has established that the crystal package already undergoes solid-state churning, or convection, while the magma ocean is still in the process of progressive freezing. This convection leads to upwellings of hot material, and the associated pressure decrease will bring about partial melting. However, the consequences of this melting for magma-ocean compositional evolution have not yet been explored. Using newly coupled thermodynamic-geodynamic models, our goal is to quantify these consequences of partial melting of the crystal package, and of related material exchange with the magma ocean.

We hypothesize that this exchange completely changes the compositional structure of the first solid mantle and the chemistry of the primary crust. For example, it may reconcile the rather moderate compositional stratification of the present-day Earth mantle, which is not addressed by previous models of magma-ocean crystallization at all. The predictions of our new models will be tested, for example by interrogating the seismic structure of the deep mantle, which may host the remnants of the primary crust.