Phase change memory materials: Why are alloys of Ge, Sb, and Te the materials of choice?

Rewritable optical storage is dominated by alloys of a small number of elements, particularly Ge, Sb, and Te, and Ge/Sb/Te alloys in the composition range (GeTe) 1− x (Sb 2 Te 3) x (0 ≤ x ≤ 1) have been the materials of choice: all have metastable rock salt structures that change little over decades at archival temperatures, and all contain vacancies (cavities). The special status arises from the close similarity of the valence orbitals of Ge, Sb, and Te, which arises from the irregular changes in atomic orbitals and properties as the atomic number increases ("secondary periodicity"). The instability of cubic (metallic) Bi to a (semimetallic) rhombohedral structure (H. Jones, 1934) can be adapted to Ge/Sb/Te alloys and explains the crucial metastable (rock salt) structures of these compounds. Vacancies almost always occur next to Te atoms, which form one sublattice of the rock salt structure. The crystallization of amorphous bits in phase change materials (e.g. DWDRW, BD-RE) is astonishingly rapid (of the order of nanoseconds) and is a rare case where parameter-free density functional simulations can follow the phase change on the physical time scale. Alloys of Ge, Sb, and Te are the favourites: the atoms have the same size and have average valence 5, and the crucial formation of a metastable rock salt can be understood in terms of a straightforward, 90-year-old theory (H. Jones). [Display omitted] • Alloys of Ge, Sb, Te containing vacancies are optimal phase change memory materials. • These alloys (including vacancies) have an average valence-electron occupancy of 5 s,p-electrons per site. • Simple nearly-free electron explanation of structure in terms of Jones zones. • Existence of accessible metastable rock-salt structure crucial. [ABSTRACT FROM AUTHOR]