All oceans on the surface of the Earth are ultimately the product of the degassing of the Earth´s interior. Over geologic time volcanic activity as well as some non magmatic processes transported water and other volatile components to the Earth´s surface. Until quite recently, it was believed that this is essentially a one way process, water that once had reached the hydrosphere would never make it back into the Earth´s mantle. While there definitively is some recycling of crustal material into the mantle occurring in subduction zones, it appeared unlikely that water could be recycled beyond the stability limit of hydrous phases, such as amphibole. Recent experimental evidence, however, suggests that some water can be retained in the subducted slab even beyond the stability limit of hydrous phases. This is due to the small, but non-zero solubility of water in nominally anhydrous minerals, such as garnets and pyroxenes. Even a few hundred ppm of water dissolved in these minerals would lead to a recycling of several percent of the combined mass of all oceans since the Proterozoic eon, assuming present-day subduction rates. Water solubility in pyroxenes can reach thousands of ppm and subduction may have been much faster during early Earth history. Accordingly, it is possible that a large part of the hydrosphere had actually been recycled into the mantle during the evolution of the Earth.
The likely existence of a real global water cycle makes it more important than ever to understand the properties of hydrous fluids at extreme pressures and temperatures. These fluids include aqueous salt solutions as well as hydrous silicate and carbonatite melts. Studying the properties of fluids at high P and T usually requires in-situ observations of phase equilibria and spectroscopic measurements. The technology for making such measurements under the P,T-conditions of the upper mantle has only recently been developed and is largely based on externally-heated diamond anvil cells.