Is the mantle chemically stratified? Insights from sound velocity modeling and isotope evolution of an early magma ocean

- Fe-Mg partitioning in lower mantle phases is strongly controlled by Al-bridgmanite.
- Minor elements in lower mantle VS calculations affect composition interpretation.
- Nd-Hf isotope data of upper mantle and crust are inconsistent with a layered mantle.
- The lower mantle is pyrolitic.
- The mantle is unlikely to be chemically stratified.

The upper mantle is widely accepted to be pyrolitic, but the bulk composition of the lower mantle remains highly disputed. Recent modeling of the lower mantle shear wave velocity profile has suggested that the lower mantle is enriched in bridgmanite, therefore implying a higher Si/Mg than that of the upper mantle. We improve upon such modeling by taking into account Ca-perovskite and considering the distribution of Fe between bridgmanite and ferropericlase, more appropriate for Al-bearing systems. Using available experimental data, we derive a means to constrain Fe-Mg partitioning for bridgmanite and ferropericlase, constrain suitable values for the lower mantle, and apply these to lower mantle shear wave velocity calculations. Calculations that consider the effects of minor chemical components such as Ca and Al suggest that the lower mantle shear wave velocities can resolve PREM for a pyrolitic composition to within 1%. We also model chemical fractionations of the 147Sm-143Nd and 176Lu-176Hf systems induced by a crystallizing magma ocean that would produce a putative Si-enriched lower mantle. The comparison of the calculated 143Nd/144Nd and 176Hf/177Hf signatures with those of the terrestrial array shows that a Si-enriched lower mantle, if ever formed, no longer exists. Both mineralogical calculations and geochemical magma ocean modeling support the idea that the Earth's lower mantle is likely pyrolitic and that the mantle as a whole need not be chemically stratified.