Skip to main content

Magmas at Depth: an Experimental Study at Extreme Conditions

Final Report Summary - MAD-ESEC (Magmas at Depth: an Experimental Study at Extreme Conditions.)

The project aimed at measuring the physical and chemical properties of magmas at extreme pressure conditions in order to answer the fundamental scientific questions related to the presence of magmas at depth in planetary interiors. Those questions range from the formation of planetary reservoirs (core, mantle, crust and atmosphere) that occurred while planets were mostly or totally molten, to current-day magmatic processes. Magmas are indeed essentially high pressure objects, being formed at depth. However, the existing database on the properties of magmas at high pressures was extremely limited at the start of this project, due to the challenging nature of the required experiments, with high temperatures involved and liquids being poor scatterers of light (from IR to X-rays wavelengths) due to their disordered nature. The development of methodologies and the opening of new synchrotron beamlines prior to and during the project made it feasible.
The nature of the continental lithosphere-asthenosphere boundary has long been elusive as the presence of melts could explain its geophysical properties but melts were thought to be gravitationally unstable and rise through the lithosphere. Our density measurements on alkali basalts showed that in fact, those are neutrally stable at the P-T conditions of the LAB, providing an answer to the longevity of cratons. Another density trap was found for lunar high Ti basalts at the bottom of the lunar mantle. The viscosity of magmas has been measured for the most depolymerized melts (i.e. mantle compositions), showing a strong effect of pressure even for these compositions. The main implication is that such melts will be most mobile at depth were they are produced, providing an explanation for their fast ascent. In the early Earth, our density data up to lower mantle P-T conditions support the hypothesis of a mantle consisting of two magma oceans separated by a crystalline layer, thus explaining the longevity of the magma ocean era. The compression mechanisms underpinning the densification of melts have been unveiled by structural studies: collapse of the intermediate range order (<5 GPa), followed by an increase in coordination number of the major cations. After completion of the 4Si to 6Si transition (>35 GPa), melt densification only results from bond compaction, therefore closely following the equation of state of crystalline silicates.
The structural environment of key tracers of planetary differentiation has been probed, revealing: 1) the formation of Xe-O bonds in haplogranitic melts, pointing to the lower crust as a reservoir for the missing Xe in the atmospheres of the Earth and Mars; 2) the bonding environment of Lu-O was found to change abruptly at 5 GPa in basalt analogues, with a coordination number increasing from 6 to 8-fold. This occurs over where a change in the P-dependence in the rare Earth element mineral/melt partition coefficients is observed; 3) the change of oxidation state of W in magmas circa 2-3 GPa, reflected in increased crystal/melt partitioning coefficient. These results show that standard models for predicting trace elements partitioning at depth must incorporate the effect of P-induced structural transformations in the melt in order to realistically model planetary differentiation. From the combination of the density and structural data, we propose a comprehensive model of planetary differentiation that satisfies all available isotopic data that previously required hidden reservoirs and/or ad-hoc meteoritic impacts to be explained.
Although the vast majority of magmas are molten silicates, carbonate melts are also produced at depth and have large economical impacts, being enriched in elements of mining interest. Density and structural data were collected on natural carbonate melts and simple end-members, from a few GPa to over a megabar. A transition from isolated carbonate molecules to polymeric state is found, accompanied by a large viscosity increase.