Our modern information society is characterized by a steadily increasing demand for powerful data storage devices, which calls for the development of innovative storage concepts. Consequently, memories based on phase-change materials have attracted considerable interest in recent years. These materials can reversibly be switched between an amorphous and a crystalline state. Since both phases exhibit significantly different physical properties, in particular reflectivity and conductivity, they enable optical and electrical storage applications. However, the properties of the employed phase-change material limit the performance of such devices. By now, only few suitable materials have been identified by empirical means. Therefore, the present work aims at developing a theoretical understanding of the material physics of phase-change materials. It is divided into four parts. First, the current state of research is reviewed, which motivates the research questions that this work is concerned with. In the second part, resonant bonding in the crystalline state is identified as a generic property of phase-change materials based on experimental results on optical properties and crystal structure. It causes the contrast between the phases that is employed in phase-change applications. The resonance is endangered, however, by Peierls-like atomic distortions that shift atoms out of the symmetry-positions of the crystal. By means of density functional theory-calculations, it is shown that in phase-change materials the resonance character is weakened by these distortions, but prevails. Subsequently, this newly gained understanding is employed to develop a design-scheme for suitable materials in form of a map. The coordinates of a material, which reflect the ionicity and tendency towards hybridization of the bonding, enable the identification of materials that are characterized by resonant bonding. The map successfully locates suitable materials in a confined region. This design-principle, its predictions, but also the limits of its validity are investigated in the third part. Therefore, density functional theory-calculations are performed on a wide range of materials in order to study and to quantify the structure and the bonding. The results support the principal validity of the map and its predictions regarding property trends. Yet, the calculations also reveal some effects that are not incorporated in the simple map-scheme. Among these are variations in the distortion patterns in the non-binary phase-change materials which are traced back to the local structure in the fourth part of this thesis. Moreover, the volume- or pressure-dependence, respectively, of the structure and the bonding is calculated and discussed. Finally, the potential energy surface is investigated to gain insight not only into the static distortions, but also into the lattice-dynamics.
展开▼
