Bridgmanite, the dominant mineral in the Earth’s lower mantle, crystallizes in the perovskite structure and transforms into post-perovskite at conditions relevant for the D layer. This transformation affects the dynamics of the Earth’s lowermost mantle and can explain a range of seismic observations. The thickness over which the two phases coexist, however, can extend over 100 km, casting doubt on the assignment of the observed seismic boundaries. Here, experiments show that the bridgmanite to post-perovskite transition in (Mg,Fe)SiO is fast on geological timescales. The transformation kinetics, however, affects reflection coefficients of and waves by more than one order of magnitude. Thick layers of coexisting bridgmanite and post-perovskite can hence be detected using seismic reflections. Morever, the detection and wave period dependence of D reflections can be used to constrain significant features of the Earth’s lowermost mantle, such as the thickness of the coexistence layer, and obtain information on temperature and grain sizes.

This paper describes a methodology for characterizing the orientation and position of grains of an orthorhombic polycrystalline material at high pressure in a diamond anvil cell. The applicability and resolution of the method are validated by simulations and tested on an experimental data set collected on MgSiO 3 post-perovskite at 135 GPa. In the simulations, $95% of the grains can be indexed successfully with $80% of the peaks assigned. The best theoretical average resolutions in grain orientation and position are 0.02 and 1.4 mm, respectively. The indexing of experimental data leads to 159 grains of postperovskite with 30% of the diffraction peaks assigned with a 0.2-0.4 resolution in grain orientation. The resolution in grain location is not sufficient for in situ analysis of spatial relationships at high pressure. The grain orientations are well resolved and sufficient for following processes such as plastic deformation or phase transformation. The paper also explores the effect of the indexing parameters and of experimental constraints such as rotation range and step on the validity of the results, setting a basis for optimized experiments. research papers J. Appl. Cryst. (2017). 50, 120-130 Christopher Langrand et al. Reliability of multigrain indexing 121