- © The Mineralogical Society Of America
When it was realized that nominally anhydrous minerals could be a major storage site of water in the mantle (Smyth 1987; Bell and Rossman 1992), it also became rather obvious that many mantle samples probably have lost most of their water during ascent (e.g., Ingrin and Skogby 2000). While analyses of natural samples therefore in many circumstances only provide a lower limit of the actual water content in the mantle, measurements of water solubility give an upper limit of the amount of water that might be stored in a mineral. Experimental measurements of water solubility have therefore naturally evolved in the last decades as a major tool for understanding the water storage capacity of the mantle.
Closely related to studies of water solubility are studies of water partitioning among minerals. If the water solubility in two minerals is known under given conditions, the partition coefficient is just the ratio of the water solubilities under the same conditions, provided that the composition of the fluid coexisting with the two minerals is the same. Since the pioneering work of Bai and Kohlstedt (1992) and of Kohlstedt et al. (1996), the water solubility in all major upper mantle minerals has been studied as well as the partitioning of water between the minerals of the lower mantle and the transition zone.
Solubility studies by themselves do not directly give the water content in the mantle. They constrain the storage capacity of the mantle, i.e., the maximum amount of water that may be stored, if the mantle were water-saturated. However, as most of the mantle is very likely water-undersaturated, the actual water content is probably far below the storage capacity.
Estimates of the actual abundance and distribution of water in the mantle may be obtained by combining solubility and partitioning data with direct …