Skip to main content
European Commission logo print header

Deep Earth Chemistry of the Core

Final Report Summary - DECORE (Deep Earth Chemistry of the Core)

Executive Summary: Earth’s core formed deep in the early molten Earth (magma ocean) at large depths ( > 1000 km) but no larger that 1700 km. The magma ocean was rather oxidised, maybe twice more oxidized that the present mantle. As the core formed, it sucked the oxygen out of the magma ocean to make it its most abundant element after iron. It left behind a magma ocean (that subsequently froze) that has the present-day redox and siderophile element concentrations of the mantle. It also leaves behind a core that has the right amounts of oxygen and silicon to be consistent with seismic wave speed models. The Earth would typically have accreted from objects with the oxidation state of asteroids like 4-Vesta.

The aim of the project is to (1) build an experimental facility based on high-pressure high-temperature apparatuses (large-volume presses and laser-heated diamond anvil cells) that reproduce the conditions of pressure and temperature that exist in the deep Earth (2) in order to study experimentally the processes that occurred in the first 100 million years after the formation of the Solar System, namely the accretion of the Earth and its primordial differentiation and formation of its core. In collaboration with John Brodholt at University College London, I have pursued a theoretical investigation that complements and supplements the experiments. The aim is to use those to gain more insight in the physics and chemistry of core formation.

The essence of this project is to see how the Earth-forming accretionary material (made of rocks and iron-rich metal) melted and segregated (that is separated the metallic part from the rocky part) to form the core and the molten silicate mantle (magma ocean). Specifically, we want to follow the evolution of siderophile trace elements and light atomic elements incorporation into the core, and how they are linked. The reason behind this is simple: siderophile trace-element concentration in the mantle and light-element concentration in the core are geochemical and geophysical observables and measurables, and provide strong constraints on the process of core formation.

Indeed, seismology shows that the core is lighter than pure iron, so we need to some up with a process of core formation that allows light elements to enter the core, in amounts that are consistent with seismology. Geochemistry shows a pattern of siderophile trace-element depletion from the mantle, and we need to come out with a core formation model that leaves behind this chemical imprint on the present-day mantle. Above all, we need to solve both these constraints simultaneously, as we need to explain all the observations both geophysical and geochemical. We are therefore trying to find an experimental path in a multivariate solution space that reproduces all the observables. The idea is conceptually simple, but from the technical point, this requires state-of-the-art experiments and modelling.

Through the use of new state-of-the-art experiments in the laser-heated diamond anvil cell, and using the focussed ion beam microscope, the scanning electron microscopes, and the electron microprobe, we have shown that the core can be formed in a deep magma ocean no deeper than 70 GPa (1650 km depth). That core formation scenario is consistent with the present day concentration in the mantle of Ni, Co, Cr, V, Mn and Nb. At the same time, large amount of O and Si are incorporated in the metal by virtue of thermodynamic equilibrium with the silicate, which is needed by the seismological models.
The core has been suggested (based on data extrapolation to high P and T) to form in a very reduced proto-Earth, through the accretion of a varying composition of the meteoritic material between the beginning and the end of accretion. We have shown for the first time that this is not a requirement, and even further, we propose that the opposite scenario based on core formation in a relatively oxidized proto-Earth is just as viable for geochemistry, and more suitable for geophysics. In the first case silicon is the major light element entering the core, and in the second it is mostly oxygen with small mounts of silicon.
The scenario of core formation in a oxidised magma ocean that gets reduced over time has the double advantage of providing a viable mechanism for the reduction of the magma ocean through oxygen incorporation in the core, and of providing a better fit to the partitioning of Cr and V. It needs to be noted however that the temperature of the magma ocean plays an important role. In cooler magma ocean, the reduced model works better whereas in hotter magma oceans, the oxidized model takes the advantage.
From the geophysical standpoint, our first principles calculations show that oxygen is the only element that is always required (alone, or along with variable amounts Si, S, or C) in the outer core to match its seismological velocity and density profiles. This geophysical requirement of an oxygen-rich outer core once again justifies the hypothesis of accretion and core formation in a rather oxidised environment. We also note that this oxidized environment is that of the differentiated planetesimals from which the Earth is believed to have accreted, such as asteroid 4-Vesta.