Sound velocity measurements of γ-(Mg0.91Fe0.09)2SiO4 show that the ringwoodite to bridgmanite and ferropericlase phase transformation does not produce the seismically observed 660 km discontinuity

The canonical model of Earth's interior directly links the global 410 and 660 km seismic discontinuities to olivine's high pressure phase transitions in a pyrolite mineral assemblage. However, previous studies observe that the expected sound velocities of pyrolite are too low to match 1-dimensional seismic models in the lower mantle transition zone (MTZ). In this study, we report measurements of the elastic properties of (Mg0.9Fe0.1)2SiO4 ringwoodite (Rw90), the dominant component of pyrolite between 520 km and 660 km depths, using pulse-echo ultrasonic interferometry combined with synchrotron X-radiation in the multi-anvil press up to 21 GPa and 1650 K. Our results show that VP and VS of anhydrous Rw90 (0.003-0.035 wt.% H2O, 5.6 +/- 1.2% Fe3+/Fetot) are both higher than predictions at MTZ conditions from previous studies, with a smaller increase for VS. Simultaneous fitting of PVT-VP-VS data yields global fit equation of state (EoS) parameters of V0= 39.69(2) cm3/mol, K0 = 183 (4) GPa, K0 ' = 5.5(3), G0 = 125(2) GPa, G0 ' = 1.3(1), q0 = 0.3(3), gamma 0 = 1.27(4), theta 0 = 1100(100) K and eta S0 = 3.5 (2). Combining results with literature data, our predicted VS for pyrolite produces a sharp discontinuity consistent with seismic models, although VS is slightly lower than observed in the lower MTZ. In contrast, we find that pyrolite would not produce a large and sharp VP discontinuity at 660 km as the jump in VP caused by the decomposition of Rw90 is too small. A homogenous assemblage of pyrolite cannot, therefore, currently explain the seismic features in the lower MTZ. We also find that neither an accumulation of harzburgite nor basaltic crust immediately above or below the 660 km discontinuity explains the observations. Instead, either a heterogeneous mixture of some other chemically distinctive components is required or more prosaically, the elastic properties of all the other mantle phases need reevaluation.