States in the gap and recombination in amorphous semiconductors

Abstract

The paper examines states in the gap in amorphous silicon and chalcogenides and their effect on photoconductivity, luminescence and drift mobility. It is supposed that carriers in an ‘ideal’ glassy semiconductor without defects would move by hopping at the band edge at low temperatures and by excitation to a mobility edge at high temperatures, and that the carriers do not form polarons; the results of Spear and co-workers (e.g. Spear 1974 a) for glow-discharge-deposited silicon and of Nagels, Callearts and Denayer (1974) for quenched As₂Te₃ containing silicon are considered. The effectively zero value of the Hall coefficient in the hopping regime is discussed. States in the gap are supposed to be due to dangling bonds which may form pairs at divacancies; if the concentration is high, these may have a predominating effect on the conductivity and in this case polaron-type hopping could occur, both for chalcogenides and for silicon. For the chalcogenides (in contrast to silicon), it is proposed, adapting a model due to Anderson (1975), that all dangling bonds are positively or negatively charged due to a large distortion energy associated with the former state; the absence of Curie paramagnetism and variable-range hopping is thereby explained; a.c. conductivity is also discussed. The reason why chalcogenides differ in this respect from silicon and germanium lies in the differing natures of the upper parts of the valence bands, which in the former case arise from non-bonding lone-pair states. It is suggested that the same conclusions may be valid for oxide glasses. The recombination mechanisms active in photoconduction and photoluminescence are described; for non-radiative transitions we use a method due to Englman and Jortner (1970). It is emphasized that hydrogen can greatly accelerate multiphonon recombination.