Abstract:
Here we report the phase boundary of pyrope garnet (Mg3Al2Si3O12) to Al-bearing silicate perovskite plus corundum, with the highest transition pressure determined by in situ measurements in a multi-anvil apparatus at high temperature. The consistency of the pressure scales by different standards of Au, NaCl, Pt, W, and Mo at high temperature was also evaluated by in situ X-ray measurements. Our results, together with recent in situ measurements of the post-spinel transition in Mg2SiO4 [Irifune et al., Science 279 (1998) 1698-1700] and the ilmenite-perovskite transition in MgSiO3 [Ono et al., Geophys. Res. Lett. (2000) submitted], show that pressures determined in conventional quench experiments [Ito and Takahashi, J. Geophys. Res. 94 (1989) 10637-10646] could have been overestimated by more than 2 GPa at pressures corresponding to the bottom of the transition zone. On the basis of the in situ measurements, the post-spinel transition occurs at a depth (~600 km) that is too shallow to match with the 660-km seismic discontinuity in the Earth's mantle. Therefore, an olivine dominant mantle compositional model may be inconsistent with the seismic observations. Alternatively, we propose a pyroxene-garnet dominant transition zone with an appropriate Al2O3 content (ca. 6-8 mol%), in which majorite garnet transforms to perovskite at the depth of the 660-km discontinuity. Any alternative models would have to consider chemical stratification in the mantle.