Deep Water: Hydrous Silicate Melts and the Transition Zone Water Filter

Earth is a wet planet, and its habitability is intrinsically tied to water at the surface. Water also plays a key role inside the Earth because it has the affect of drastically lowering the melting point of mantle rocks, and indeed, the water at the surface ultimately comes from the interior through magmatism. Water is returned to the interior at subduction zones, and over geological time the surface and mantle water inventories are regulated by a deep water cycle. Nearly all of the volcanism we see at the Earth's surface is caused by melting in the shallow mantle, especially above subduction zones, and this constitutes an important part of the deep water cycle. However, water is also transported deeper down into the mantle, and what happens to it, and the controls it has on interior processes, is a mystery.

Deep mantle rocks are generally too cool to melt, but there are regions in the mantle, most notably at about 400 and 700 km, where seismic signals are interpreted to represent partial melting. Water reduces the solidus of the mantle by hundreds of degrees in the upper and lower mantle, but much less so in the mantle transition zone (410-660 km), and this is because water is soluble in the main minerals that make up this region, wadsleyite and ringwoodite, but it is not soluble in the minerals that constitute the regions above and below the transition zone. This means that if there is water in the transition zone, and it potentially can store a couple of oceans worth, when mantle is transported from the transition zone into either the upper or lower mantle it is expected to melt when water is released. This concept was originally applied to the region above the transition zone by Bercovici and Karato in a landmark paper in 2003, and was called the 'transition zone water filter' because of the important predicted affects on mantle geochemistry.

Interestingly, since then seismic evidence has been mounting indicating melted regions above and below the transition zone. If these signals do indeed represent the presence of hydrous melt in these regions, then the transition zone may act as a double-sided mid-mantle water filter, and the melting that occurs at its boundaries could have modulated the chemistry and geodynamics of the mantle throughout its history. Currently we cannot adequately test this model or understand its implications because we do not know accurately the composition of hydrous silicate melts of the mantle at these depths, nor do we know their physical properties, such as the density and viscosity. Because of this, we are currently unable to model accurately the seismic response expected for hydrous silicate melting at these depths, and we cannot model the dynamic behavior of such melts should they exist in these locations. Here we propose to collect this data.

We propose to make high P-T experimental measurements to determine the compositions of hydrous silicate melts in the mantle at depths corresponding to the deep upper mantle, transition zone and upper part of the lower mantle. We will also use novel experiments where we combine diamond anvil cell techniques with synchrotron X-ray scattering methods to determine melt densities. Simultaneously we will use first principles molecular dynamics methods to calculate the physical and seismic properties of these melts, supplemented with experiments to measure their wetting properties. With these data, we will be able for the first time to develop dynamic and seismic models to explicitly test the transition zone water filter model, and make predictions about its chemical and dynamical affects on the mantle, and on the deep water cycle, throughout geological time.

Outputs from this research will be disseminated to academic beneficiaries in the normal way, through publication of results in high-calibre international journals. We will disseminate results to the public via current successful University and Departmental schemes, existing contacts with the local and national media and our web presence. As part of this proposal we will target public outreach to School-aged children by linking in with the Geobus that is part of the Deep Volatile consortium being run by Bristol, UCL and Oxford. We will provide materials for the Geobus in the form models for the Earth's interior that are accessible to School aged children, emphasizing the links between the deep water cycle and the long-term habitability of our planet.