DI31B-2599
On the cooling of a deep terrestrial magma ocean: Experimental perspectives

Wednesday, 16 December 2015
Poster Hall (Moscone South)
Giacomo Pesce1, Denis Andrault2, Geeth M Manthilake1, Julien Monteux2, Nathalie Bolfan-Casanova1, Davide Novella1, Nicolas Guignot3 and Julien Chantel4, (1)Laboratoire Magmas et Volcans, Clermont-Ferrand Cedex, France, (2)Laboratoire Magmas et Volcans, Clermont-Ferrand, France, (3)Organization Not Listed, Washington, DC, United States, (4)Laboratoire Magmas et Volcans, Clermont Ferrand, France
Abstract:
Our knowledge about the melting properties of the shallow Earth's mantle is based on few high-pressure experiments performed using the cook-and-look technique. It provided major informations, including reliable solid-liquid partition coefficients for a number of elements. This technique, however, has major drawbacks. One of them is a poor sensibility to low degrees of partial melting, because small pockets of melt can be extremely difficult to evidence, if not impossible, due to melt recrystallization in quenched samples. For this reason, we studied the melting properties of a chondritic mantle in situ in the multi-anvil press, by (i) X-ray diffraction and (ii) electrical conductivity, at pressures up to 30 GPa, which corresponds to ~800 km depth. Both techniques allow detection of small degrees of partial melting and the monitoring of the melting behaviour during the experiments. Both types of measurement suggest that previous studies overestimated by ~300 K the solidus temperature.

There are two major implications for such a low solidus temperature at upper mantle P-T conditions. The first concerns the crystallisation processes in the early magma ocean. Upon decrease with time of the surface potential temperature, due to magma-ocean cooling, partial melting has remained a dominant process in the upper mantle until a thick solid-crust could dump the temperature profile significantly below mantle adiabat at depth as deep as a couple hundred kilometres below the Earth's surface. A second implication concerns the current mantle state. Seismic observations highlighted the presence of low velocity zones atop the 410 km discontinuity. These regions show significant reduction in shear wave velocity, compatible with the presence of partial melting. We combined our new solidus curve with the expected melting temperature depletion due to water, in order to discuss the presence and role of water in the asthenosphere. Our wet melting curves happen to be well compatible with the depth of the seismic anomalies, for a water content of 400-800 ppm in the upper mantle