Spinel elasticity and seismic structure of the transition zone of the mantle

THERE is continuing debate about whether the upper mantle is chemically stratified, and whether the seismic discontinuity at 400 km depth represents a chemical boundary, a phase change of the (Mg, Fe) 2 SiO 4 component from the β phase (olivine) to the p phase, or a combination of both. Recent developments in high-pressure synthesis 1,2 and ultrasonic interferometry 3,4 have made possible measurements of elastic-wave velocities in small, poly-crystalline samples of high-pressure phases. By combining our new acoustic measurements on the spinel (γ) phase of Mg 2 SiO 4 with existing data for the α and β phases, we present here velocity profiles for the M2SiO4 component of the mantle (where M represents Mg or Fe) to depths of about 600 km. We find it to be unlikely that any seismologically observable velocity discontinuity at about 520 km depth can be attributed to this component, although the contrast in impedance (the product of density and velocity) might be sufficient for the β→γtransformation to be observed in long-period seismic reflection studies at near-normal incidence 5 . The velocity gradients in the transition zone, particularly for shear waves, are steeper than would be expected for simple adiabatic compression of likely mantle compositions, suggesting that alternative explanations including chemical heterogeneity and anelastic relaxation need to be explored.