Growth of epitaxial silicon layers by vacuum evaporation

Abstract

Epitaxial Si layers have been grown on (111) Si substrates by the vacuum evaporation method. A preliminary assessment of these layers was recently reported (Unvala and Booker 1964, Paper I). The present paper describes a transmission electron microscope investigation of the initial nucleation and growth process. The layers examined were grown using an initial pressure of 3 × 10⁻⁶ mm Hg, a substrate temperature of 1200°c and deposition rates in the range 0·1 to 2·5 micron/min.

Three different types of growth were distinguished. At slow rates (<0·5 micron/min), separate growth centres initially formed and gradually joined up to give a worm-like structure and subsequently a continuous layer. The separate centres and all areas of the subsequently formed layer contained stacking faults. At medium rates (0·5 to 1·0 micron/min), separate growth centres initially formed as before. However, at an early stage during the growth, the centres joined up almost instantaneously without any further growth centres being observed. The initial centres contained stacking faults, but the areas in between were completely free from stacking faults. At fast rates (> 1·0 micron/min), no growth centres were observed from the earliest stages, and no stacking faults occurred in any areas of the layer.

A qualitative explanation is proposed to account for these observations. In particular, it is considered that (a) the separate growth centres form, and the stacking faults occur, because of contamination present on the substrate surface, and (b) the growth rate dependence arises because the incident Si atoms are able to act as a getter and remove the contamination, and the effectiveness of the gettering action increases as the deposition rate increases.