Our view of the interior of the Earth relies on modeling seismic velocities using physical properties of the constituent minerals. One can resolve 100 Kelvin temperature gradient or a few percent chemical component (such as Al) variation in lateral regions at 600 Km depth by combining modern high resolution seismic tomography and recent mineral data measured at mantle conditions. Elastic properties, which define how fast seismic wave travels, and phase equilibrium, which defines the stable minerals at a given depth, are the key ingredients to simulate seismic velocities. However, the interaction of the seismic wave with phase transitions has been ignored for over four decades. The exact mineralogy is often assumed to be unchanging in the velocity models. If the period of the seismic P wave is comparable to the phase transition rate, and the P wave, as a compressional force wave, drives a small amount of minerals though phase transitions, P wave velocities will be reduced based on behavior of solids. Our pilot experiments suggest this process is important for the Earth. Indeed incompatibilities between seismic models and mineral models persist, particularly in the transition zone. Most regions between 200 and 1000 km depth contain significant amounts of coexisting high- and low-pressure phases. Furthermore, the effective bulk modulus of thermodynamically equilibrated materials undergoing a volume reducing phase transformation is significantly lower than that of the individual phases. If the stress of the P wave itself induces phase transitions, then the P velocity will be reduced in these regions as the P waves sample a relaxed and lower modulus. A comparison between the amount of time required by phase transitions to reach equilibrium and the sampling period thus becomes crucial in order to define the amount of softening or attenuation of P waves within a two-phase zone. This proposal is to conduct an experimental research program aimed at defining the effect of phase transformations on the expected P wave velocity in the depth range of 200 ? 1000 km in the Earth. Using synchrotron and a multi-anvil device, we have developed the capability of measuring stress-strain-time relations at mantle P-T and seismic frequencies. Our pilot experiments which include the kinetics, attenuation, and dispersion during the olivine-spinel phase transition imply that phase transitions will significantly reduce P velocities measured seismically. We will focus on volume changing phase transformations as relaxation processes. Mg-Fe exchange controlled olivine-wadsleyite-ringwoodite-perovskite transition and Al/Si diffusion controlled pyroxene-garnet-perovskite transition are to be evaluated within the context of this model in order to define the effects on seismic velocities and attenuation. The goals of this proposal include (1) Establish a working model that is supported by elastic and anelastic properties of minerals during first-order phase transitions at the time scale of seismic frequencies; and that can be extrapolated to the stress amplitudes of a seismic wave. (2) Measure elastic and anelastic properties of minerals during phase transitions at mantle P-T and seismic frequencies.
