Chemical evolution of the Earth

Impact erosion: reconciling chemical differences between Earth and chondrites

It was previously shown that our planet has remarkable isotopic similarities with enstatite chondrites (EC) for the elements whose isotopes do not fractionate during core segregation (O, Ca, N, Mo, Ru, Os, Cr, Ni, Ti and Sr). This makes them prime candidates to represent the primordial building blocks as they are likely formed in the same region of the solar nebula. However, one significant difficulty with a model of Earth accreted with EC is that they present important chemical differences: EC are particularly enriched in silica and alkali elements (such as Li, Na, K and Rb) and depleted in Mg and refractory lithophile elements (Ca, Al, REE) compared with Earth. During my graduate work, I proposed a new integrative model based on experimental data and mass balance calculations that resolves this paradox. I showed that collisional stripping of the Earth’s proto-crust can explain chemical differences between Earth and EC. By conducting melting experiments of enstatite chondrites, I showed that early crusts would be highly enriched in Si and depleted in Mg. Energetic impacts eroding this type of crusts would lead to an increase of the final planetary Mg/Si (Boujibar et al. 2015, Nature Communications).

Chemical differentiation of the Earth’s mantle during the crystallization of the magma ocean

Once the Earth reached ~30% of its present mass, the base of the magma ocean was dominated by equilibria between bridgmanite and both silicate and metal liquids. Accurately determining these chemical equilibria is therefore a key parameter in understanding possible subsequent chemical heterogeneities preserved in the lowermost mantle. Using multi anvil apparatus, I replicated these conditions of simultaneous magma ocean crystallization and core-mantle differentiation and determined how major elements partition between bridgmanite and silicate melt. While partition coefficient for Si stays constant for any degree of melting, it decreases for Fe, Mg and Al as the melt fraction increases. One striking finding is that Fe is compatible at very low degree of melting (F) and becomes more and more incompatible as F increases. As Fe controls the density contrast between liquid and solid in the lower mantle, this has implications for the dynamics of the mantle. It suggests that in regions where the early mantle was highly molten, bridgmanite crystals were likely floating. Conversely, the dominantly solid mantle likely contained buoyant pockets of silicate melt. This lead to important chemical heterogeneities that were progressively attenuated with mantle stirring (Boujibar et al. 2016, American Mineralogist).