Shock-wave Data for Several Minerals and their Implication to Mineral Phases in the Lower Mantle

Abstract

FROM studies of the density and elastic properties of the mantle, Birch¹ showed that the upper mantle is characterized by a zero-pressure density of about 3.3 g cm⁻³ and (K/ρ)_{p = 0} about 35 km² s⁻², where K is the bulk modulus. The corresponding values for the lower mantle, however, are about 4 g cm⁻³ and 55 km² s⁻². Comparing these values with laboratory measurements for a number of materials showed that, while a number of common rock-forming minerals (olivine, pyroxene, garnet, jadeite) have the qualifications of the upper mantle material, none of the familiar silicates have similar properties to the lower mantle. The observation that a few oxides (periclase, corundum, rutile) have densities close to 4 g cm⁻³ and (K/ρ)_p=0 close to 50 km² s⁻² inspired Birch¹ to speculate that one of the possible phase changes in the transition zone is for silicates to break down into closely packed oxides. The popularity, and hence the importance, of this speculation is clearly indicated by the work of Al'tshuler et al.², Ahrens and Syono³ and Anderson⁴. Al'tshuler et al. attempted to construct the equations of state of the high-pressure phases of forsterite and enstatite from the Hugoniot data for quartz and periclase; the constructed curves were compared with the P-ρ curve for the lower mantle. Anderson, on the other hand, calculated on thermodynamic grounds the temperature and pressure range for the breakdown of spinel polymorph of forsterite to form a mixture of closely packed oxides. The result was correlated with the density increase corresponding to one of the steep rises in the seismic velocity profile in layer c. In the light of shock-wave data, the nature of the phase change is examined here.