In this article, I review four topics of my past researches that are intended to bring new insights about structure and dynamics of the Earth's mantle. Firstly, the phase boundary of dissociation of ringwoodite to bridgmanite plus periclase, so called post spinel transition, in Mg 2 SiO 4 was determined by means of in situ X ray diŠraction. This study showed that the Clapeyron slope of this reaction was -0.4 MPa/K, which is much smaller than previously considered. This shallow slope suggests

more »

... that seismically observed depression of the 600 km discontinuity under subduction zones cannot be explained by a temperature eŠect only, and that the 660 km discontinuity will not be a barrier of mantle convection. Secondly, the thermal expansion coe‹cients of four major mantle minerals were determined by using a new multi anvil press with an oscillation system, which enables us to obtain high quality diŠraction patterns at high temperatures. Together with the pressures of the olivine wadsleyite transition in (Mg,Fe) 2 SiO 4 , an adiabatic geotherm in the whole mantle was estimated. This proˆle shows that adiabatic gradients in the upper and lower mantle, respectively are 0.4 0.6 and 0.3  0.5 K/km. Thirdly, the hopping and proton conductions of olivine, wadsleyite and ringwoodite were measured under conditions of their stabilityˆelds. The magnitudes of hopping and proton conductions, respectively, are larger and smaller than previously considered, which suggests that the mantle transition zone is essentially dry. Fourthly, water content dependence of Si lattice and grain boundary self diŠusion in forsterite was measured to show that their water contents are around 0.3, which is signiˆcantly smaller than those obtained by deformation experiment. In addition, water content exponent of lattice self diŠusion coe‹cient of oxygen is zero within uncertainty. These water content exponents suggest that eŠects of water on mantle rheology could be limited.