The Mechanism of Solid Phase Epitaxy

Abstract

The enhancement of the solid phase epitaxial growth (SPEG) rate in Si and Ge by hydrostatic pressure, and the reduction in the rate by uniaxial compression, place severe constraints on the kinds of point defects that can be responsible for thermal SPEG. These measurements are interpreted in terms of an activation strain tensor, the nonhydrostatic analogue of the activation volume, which results from an extension of transition state theory to nonhydrostatic stress states. These results and those of other experiments allow us to rule out all mechanisms in which the rate-limiting step is thermal generation of point defects in the bulk of either phase, and the migration of these defects to the crystal-amorphous interface. All experimental results are semi-quantitatively consistent with the Spaepen-Turnbull interfacial dangling bond mechanism. The structural aspects of the Williams-Elliman interfacial kink site model are shown to be a special case of the dangling bond mechanism. The electronic aspect of the Williams-Elliman model has been generalized to take into account more recent experiments on the doping-dependence of the SPEG rate. It is compared to the fractional ionization model of Walser and Jeon. They both account for the enhancements due to low, but not high, dopant concentrations. The relevance to models for the effects of ion irradiation on SPEG is also discussed.