Identifying and locating the geochemical and geophysical heterogeneities in the Earth’s interior is one of the most important and challenging tasks for the deep Earth scientists. Subducted oceanic crust metamorphizes into the dense eclogite in the upper mantle and is considered as a major cause of geochemical and geophysical heterogeneities in the deep Earth. In order to detect eclogitic materials inside the Earth, precise measurements of the high pressure-temperature single-crystal elasticity of major minerals in eclogite are thus exceedingly important. Omphacite, a Na,Al-bearing clinopyroxene, constitutes up to 75 vol% of eclogite. In the present study, we performed the first high pressure-temperature single-crystal elasticity measurements of omphacite using Brillouin spectroscopy. Utilizing the finite-strain approach, we obtained the following thermoelastic parameters for omphacite: KS0’ = 4.5(1), G0’ = 1.53(5), ∂KS0/∂T = −0.029(5) GPa/K, ∂G0/∂T = −0.013(5) GPa/K, with KS0 = 123(3) GPa, G0 = 74(2) GPa, and ρ0 = 3.34(1) g/cm3. We found that the seismic velocities of undeformed eclogite are similar to pyrolite at the depths of 200–300 and 410–500 km, thus eclogite is seismically invisible at these depths. Combined with the lattice-preferred orientations of the omphacite in naturally deformed eclogites, we also modeled seismic anisotropy of eclogite at various pressure-temperature conditions. A 10 km thick subducted eclogitic crust can result in ∼0.2 s shear wave splitting in the Earth’s upper mantle.

Hao, M.Zhang, J. S.Zhou, W.-Y., & Wang, Q. (2021). Seismic visibility of eclogite in the Earth’s upper mantle—Implications from high pressure-temperature single-crystal elastic properties of omphaciteJournal of Geophysical Research: Solid Earth126, e2021JB021683. abstract

Experimentally determined velocities of the omphacite crystals with different orientations at 18.4 GPa and 700 K. The dotted lines are calculated from the final Cij model, and the circles represent the measured velocities. The phonon directions of the Vp and Vs measurements were calculated by matching the in-plane measured velocities with velocity dispersion curves predicted by the best-fit Cij model. The uncertainties of the individual velocity measurements are smaller than the symbols. The root-mean-square (RMS) error between the experimental data and the model is 35 m/s.