The properties and dynamic behaviour of the Earth are controlled to a great extent by the physical and chemical properties of the minerals present. These are in turn controlled by the crystal structures and structural state of these constituent minerals. Consequently, the characterisation of crystal structures and the determination of the changes in properties with phase transitions, especially at the high pressures relevant to the Earth's interior are a focal point of many research programmes at the Bayerisches Geoinstitut.
Two general experimental philosophies are pursued. The first is to synthesise phases at high pressures and temperatures relevant to the Earth's interior (or to recover such phases from natural samples that have been subjected to such high pressures and temperatures in meteorites) and then to characterise the crystal chemistry of the recovered samples under ambient conditions. One of the most important aspects to emerge from such studies is the coordination environment of the cations, especially that of silicon, because in the crystal structures of silicates formed at low pressures typical of the Earth's crust the silicon atoms are normally coordinated (i.e. surrounded by) four oxygen atoms that form a tetrahedron. At higher pressures, typical of the transition zone and lower mantle of the Earth, the minerals contain silicon coordinated by six oxygen atoms in the form of an octahedron. This coordination change generally leads to a closer packing of the atoms and thus to a significant increase in density. Just such a density change accompanying the transformation of silicate spinel (4-coordinate Si) to silicate perovskite (6-coordinate Si) is responsible for the seismic discontinuity that marks the boundary between the transition zone and the lower mantle at a depth of around 670 km in the Earth. In order to determine the coordination of silicon, and other cations, in samples a number of techniques have been employed ranging from single-crystal X-ray diffraction studies that determine the complete structure of a crystal to the use of techniques such as ELNES, Mössbauer spectroscopy and NMR that probe the local environment of specific atoms within the structure. During this year such studies, carried out at the Bayerisches Geoinstitut, have led to some exciting discoveries, including the first crystal structure of an oxide phase to contain silicon in five-fold coordination and the first natural occurrence of MgSiO3 with the ilmenite structure.
The second experimental approach is to take synthesised samples, to re-pressurise them to high pressures in a diamond-anvil pressure cell and to determine their crystal structures and properties in-situ at high pressures, usually by single-crystal X-ray diffraction. Such studies provide not only much-needed equation of state data for mantle minerals (i.e. the variation of density with pressure) but also the change in internal structure that accompanies compression. Such changes in structure can be either continuous or discontinuous at a phase transition (=change of crystal structure, and possibly coordination numbers). Although the three studies of this type described here all showed that no such transitions occurred in the materials studied, these "negative" results are still of interest as they indicate the great stability of some structure types (e.g. braunite) at high pressures, and also the subtlety of the factors that control electronic transitions in structures (as illustrated by the example of TiS2). Clearly much more development of experimental probes and many more systematic studies of crystal chemistry of high-pressure phases of relevance both to the Earth's interior and to physics must be carried out before we can develop a full understanding of crystal chemistry at high pressures and temperatures, and cease to be "surprised" by some of the phenomena that we find.