The chemistry, mineralogy and temperature of Earth’s interior are the most fundamental properties of our planet. It is these properties combined that control Earth’s evolution, not only of its interior but also of its atmosphere, climate and surface environment over geological time. However, studying the deep Earth is not a straightforward task as direct sampling is limited to ~ 10km depth. As such, we have to rely on petrology and geophysics – particularly seismology – to understand anything except the near surface.

Fortunately, seismic observations are abundant, with lateral and vertical heterogeneities normally assumed to reflect compositional or thermal anomalies. However, seismic observations will only provide useful insight if they can be reliably interpreted in terms of composition, mineralogy temperature and pressure. This is a major sticking point in studies of the deep Earth, because experimental thermoelasticity data at mantle conditions are extremely sparse. In lieu of data, it has become common to use software predictions to interpret mantle seismology. Whilst this approach is effective in theory, there are several limitations including that model predictions do not provide realistic estimates of uncertainty and that many predictions do not match available experimental data. Experiments in the large volume press with/without synchrotron radiation are one approach that can be used to accurately determine the crystallography and thermoelasticity of upper mantle minerals. Here examples of these data from recent and ongoing studies will be presented.