High-resolution TEM of grain boundaries in silicon

Abstract

In contrast to grain boundaries in metals, little is known about the structure and the properties of grain boundaries in covalent crystals, especially in semiconductors. The reason for this lack of knowledge is that grain-boundary-free crystals of, e.g., silicon, are easy to grow; commercially used silicon crystals are free of dislocations and thus nearly perfect. It was not until after the use of polycrystalline silicon had been proposed for the large scale production of cheap solar cells that grain-boundary properties in silicon gained considerable interest. In particular their electronic properties and their influence on device performance is important in this case. Moreover, “grain boundary devices”, i.e., devices with a grain boundary as the active element and with rather interesting properties, are conceivable - provided the relationship between the grain boundary structure and their electronic behaviour can be understood (cf. /l/). In addition, the study of grain boundaries in silicon, with an electronic structure and a binding configuration very different from metals, may provide a valuable tool to test competing grain boundary models (see, e.g., /2,3/) and may lead to a deeper insight into the crystal parameters governing the grain boundary properties.