The mantle convection causes various geophysical and geological phenomena such as plate motion, earthquakes and volcanism. Hence it is essential to quantitatively understand dynamics flows in the mantle. For this purpose, knowledge of rheological properties of the mantle minerals is indispensable. However, because deformation experiment is the most difficult kinds of laboratory studies, the results of high P-T deformation experiments have large uncertainty. In addition, strain rates in laboratory deformation experiments are at least 5 orders of magnitude larger than those expected in the mantle. Such huge difference could cause significant of misleading for our understanding about the dynamics of the mantle.

For these reasons, we adopt rather indirect methods to estimate rheological properties under high P-T conditions. One is measurement of self-diffusion coefficients. Diffusion of the slowest species limits creep of minerals. Since the rate-limiting species in silicate minerals is usually Si, we are measuring Si self-diffusion coefficients of the important mantle minerals. In this method, we make an isotope-enriched thin film on a polished surface of a target mineral. The sample is annealed at high pressure and high temperature, to make the isotope diffuse into the substrate. The diffusion profile will be determined by SIMS. We obtain diffusion coefficient from the diffusion profiles with the annealing times.

Another approach is dislocation recovery experiment, which allows measuring dislocation mobility. In this experiment, single crystals are deformed to create dislocation, and the deformed crystals are annealed at high temperature conditions for certain duration. During annealing, the dislocation density should decrease by coalescence of dislocations. The dislocation mobility is obtained by annealing duration and dislocation densities measured before and after the annealing.

At present the most important problem is the effect of water on mantle rheology. If incorporation of water should significantly soften minerals, patterns of mantle convection will be very complex. A number of deformation experiments actually reported significant decrease in creep strength of rocks by adding water. However, this phenomenon might be caused by lubrication of free water on grain boundaries not by water incorporation into mineral structures, because such experiments were conducted under water-oversaturated conditions. Because the mantle should be mostly water-undersaturated, such lubrication will not occur in the majority of the mantle. We have already obtained experimental evidence showing small effects of water on mineral rheology based on the Si self-diffusion study. We are also confirming these results by means of dislocation recovery experiments.

Self-diffusion experiments

• Si and O self-diffusion in hydrous forsterite and iron-bearing olivine from the perspective of defect chemistry [Fei and Katsura, 2016]
• New constraints on upper mantle creep mechanism inferred from silicon grain-boundary diffusion rates [Fei et al., 2016]
• Water has no significant effect on oxygen self-diffusion rate in forsterite [Fei et al., 2015]
• Small effect of water on upper-mantle rheology based on silicon self-diffusion coefficients [Fei et al., 2013]
• High silicon self-diffusion coefficient in dry forsterite [Fei et al., 2012]
• Silicon and magnesium diffusion in a single crystal of MgSiO₃ perovskite [Xu et al., 2012]
• Si self-diffusion coefficient of stishovite [Shatskiy et al., 2010]

Dislocation recovery and other dislocation-related studies

• Water activation of the E-type slip system [Wang et al., 2019]
• Pressure dependence of [100](010) and [001](010) dislocation mobility in natural olivine [Wang et al., 2017]
• Temperature dependence of [100](010) and [001](010) dislocation mobility in natural olivine [Wang et al., 2016]