The vast majority of rocks that constitute the Earth are composed of multiple mineral phases and complicated deformation conditions are found everywhere, yet little is known about plastic deformation of polyphase polycrystalline rocks, especially with low symmetry phases and highly contrasting mechanical properties. In particular, plastic deformation of mantle rocks is of interest for its connection to geophysical processes, such as mantle convection, slab subduction and upwelling plumes. Seismic observations show regions in the mantle where seismic waves encounter large scale anisotropy of the rocks they propagate in. Although there are many possible reasons for this large scale anisotropic signature of these particular regions, in the upper mantle it has been connected to the development of crystallographic preferred orientation, hereafter called texture, of the mineral olivine due to plastic deformation produced by specific geodynamical processes. Studies have also connected measured seismic anisotropy to texture in the minerals that compose the core mantle boundary region, in the mantle transition zone and in the core of the Earth. Deformation of lower mantle minerals is still poorly understood. Although large convective cells, upwhelling plumes and subducted slabs are expected, the bulk of the lower mantle remains seismically isotropic. The lower mantle is mainly composed of the mineral bridgmanite and ferropericlase, with a small volume fraction of Ca-Si-perovskite. Understanding plastic deformation of the bridgmanite and ferropericlase mineral aggregate can provide valuable information to better understand lower mantle dynamics, and to explain seismic observations. The objective of this dissertation is to explore plastic deformation of bridgmanite and periclase polycrystalline aggregates from both modeling and experimental approaches. In particular to understand the influence of microstructure in texture development and the role of deformation heterogeneities at the local grain scale.

A finite element framework is used in Chapter 2 to explore plastic deformation by dislocation glide of a virtual two phase polycrystal with a random microstructure. Aggregates are deformed up to 20% strain under uniaxial compression and two different simulations where performed using the FEpX code, one where the yield strength of both phases is the same, and the second where the yield strength of the bridgmanite phase is 8 times higher than the periclase. This approach enabled the investigation of the effects of yield strength contrast in the mechanical response of the phases under uniaxial compression. Trends in texture development, plastic deformation rates and intragranular misorientation distributions are analyzed. Single phase bridgmanite and single phase periclase simulations were also performed for comparison with their two phase counterparts. It is found that both the bridgmanite and periclase phase develop weak texture in both the single and two phase simulations. Distributions of the plastic deformation rate show that as the yield strength contrast is increased in the two phase simulations, heterogeneities in the distribution of the plastic deformation rate across the bulk of the sample increase drastically. However, a wider distribution of the plastic deformation rate is observed in the single phase bridgmanite simulation, than in the two phase simulation where both phases have the same yield strength. This indicates that deformation heterogeneities in bridgmanite are mainly due to the high anisotropy of its single phase mechanical properties. Misorientation of each element in a grain with respect to the grain average orientation is calculated and statistical trends of the grains of the bridgmanite phase and the periclase phase are analyzed separately and compared. While misorientation distributions for the bridgmanite phase remain very similar across the different simulations, misorientation distributions of the periclase phase become much wider and with a larger mean misorientation as the yield strength contrast is increased. In addition, the distribution of misorientations in the periclase phase when deformed on its own is the narrowest of them all, indicating that deformation heterogeneities in the periclase phase are introduced by deforming in the two phase scenario with a harder phase. Slip system activity calculations show most of the deformation by dislocation slip is carried by the periclase phase in the two phase simulation with large yield strength contrast.

A fast Fourier transform formulation implemented in the VPFFT code is used in Chapter 3 to study the effects of microstructure in texture development of a two phase polycrystalline aggregate composed of 75% bridgmanite and 25% periclase, and also single phase simulations for comparison. Using this approach, it is also found that bridgmanite and periclase develop weak texture, even when deformed as single phase aggregates. Furthermore, it is observed that there is little microstructure dependence of the texture and strain rate distributions for a microstructure where periclase is at the cores of bridgmanite grains, one where grains of both phases are randomly distributed across the aggregate and one where periclase grains are at triple junctions of the bridgmanite grains. A microstructure where periclase grains is percolating around bridgmanite grains is found to be the outlier of this study, it develops very strong texture in the bridgmanite phase and presents sharper distributions of the strain rate in this phase than the other tested microstructures, indicating a microstructure dependance of the deformation of bridgmanite in this case. Since grains are discretized into smaller elements, statistical trends in the distribution of strain rates of regions at grain boundaries and regions in interior of grains can be compared. It is found that in two phase simulations, regions at grains boundaries in grains of the periclase phase develop the widest distribution of strain rate values, suggesting that deformation heterogeneities are mainly concentrated in these regions.

High pressure diamond anvil cell experiments were performed on the bridgmanite + periclase two phase system to explore texture development at high pressure due to dislocation slip. Diamond anvil cell experiments were performed in synchrotron x-ray sources, where 2D diffraction patterns are collected and samples are laser heated to induce phase transformations at high pressures. The sample grains sizes that were previously possible to analyze using the traditional Rietveld technique were limited to grain sizes that produced smooth powder diffraction patters in order to analyze intensity variations along Debye rings and determine the texture developed in the sample. Recently the high pressure diamond anvil cell community has adopted a more novel multigrain data collection and analysis technique, where diffraction images with sharp diffraction spots originating from multiple grains can be indexed and clusters in orientation space are searched to correlate multiple diffraction spots with the particular grain they originate from. Applying this data collection technique presents challenges in the diamond anvil cell, in particular when non hydrostatic conditions are desired for texture development studies. The small rotation angle accessible to the diamond anvil cell, peak overlap due to plastic deformation in the grains and low intensity of diffraction spots due to weakly diffracting materials are some of the challenges encountered in these experiments. An overview of the implementation of this data collection and analysis technique to the diamond anvil cell for texture studies, together with some preliminary results, is presented in Chapter 4.