To understand the phase transition from perovskite to postperovskite MgSiO3, total energy is partitioned into atomic energy densities of constituent elements in the oxide, using the energy density analysis. The atomization energies, δEMg for Mg atom, δESi for Si atom, and δEO for O atom, are then evaluated by subtracting the atomic energy density from the energy of the isolated neutral atom,Mg, Si, and O, respectively. It is found that δESi and δEMg are much larger than δEO in the perovskite phase, but δEO is much larger than δE Si and δEMg in the postperovskite phase. This means that most of the energies partitioned into Mg and Si atoms in the perovskite phase transfer to the O atoms in the postperovskite phase during the transition. Such an extremely stable O-atom state is formed by the introduction of edge-shared SiO6 octahedra into the postperovskite structure. This is because the edge-sharing makes the Si-O interatomic distances longer even in very high pressure conditions. The unusual energy balance between atoms is also seen in the other postperovskite, MgGeO3 and NaMgF3.
ASJC Scopus subject areas