It is well known that water (as a source of hydrogen) affects the physical and chemical properties of minerals--for example, plastic deformation and melting temperature--and accordingly plays an important role in the dynamics and geochemical evolution of the Earth. Estimating the water content of the Earth's mantle by direct sampling provides only a limited data set from shallow regions (<200 km depth). Geophysical observations such as electrical conductivity are considered to be sensitive to water content, but there has been no experimental study to determine the effect of water on the electrical conductivity of olivine, the most abundant mineral in the Earth's mantle. Here we report a laboratory study of the dependence of the electrical conductivity of olivine aggregates on water content at high temperature and pressure. The electrical conductivity of synthetic polycrystalline olivine was determined from a.c. impedance measurements at a pressure of 4 GPa for a temperature range of 873-1,273 K for water contents of 0.01-0.08 wt%. The results show that the electrical conductivity is strongly dependent on water content but depends only modestly on temperature. The water content dependence of conductivity is best explained by a model in which electrical conduction is due to the motion of free protons. A comparison of the laboratory data with geophysical observations suggests that the typical oceanic asthenosphere contains approximately 10(-2) wt% water, whereas the water content in the continental upper mantle is less than approximately 10(-3) wt%.

Electrical conductivity of most minerals is sensitive to hydrogen (water) content, temperature, major element chemistry and oxygen fugacity. The influence of these parameters on electrical conductivity of major minerals has been characterized for most of the lower crust, upper mantle and transition zone minerals. When the results of properly executed experimental studies are selected, the main features of electrical conductivity in minerals can be interpreted by the physical models of impurity-assisted conduction involving ferric iron and hydrogen-related defects. Systematic trends in hydrogen-related conductivity are found among different types of hydrogen-bearing minerals that are likely caused by the difference in the mobility of hydrogen. A comparison of experimental results with geophysically inferred conductivity shows: (1) Electrical conductivity of the continental lower crust can be explained by a combination of high temperature, high (ferric) iron content presumably associated with dehydration.(2) Electrical conductivity of the asthenosphere can be explained by a modest amount of water (~10 -2 wt% in most regions, less than 10 -3 wt% in the central/western Pacific). ( 3) Electrical conductivity of the transition zone requires a higher water content (~10 -1 wt% in most regions, ~10 -3 wt% in the southern European transition zone, ~1 wt% in the East Asian transition zone). The majority of observations including those on the lower crust and the asthenosphere can be interpreted without partial melting or any fluids. However, experimental studies on electrical conductivity of lower mantle minerals are incomplete and it is not known if hydrogen enhances the conductivity of lower mantle minerals or not. Some discussions are also presented on the electrical conductivity in other planetary bodies including the Moon and Mars.