Abstract from Materials Technology'98, University of Waterloo, 13 June 1998
Serge Grabtchak1), Michael Cocivera1), Dolf Landheer2) , 1)Guelph-Waterloo Centre for Graduate Work in Chemistry, Guelph, Ontario N1G 2W1, 2)Institute for Microstructural Sciences, NRC, Ottawa, Ontario K1A 0R6
Thin polycrystalline CdSe films have been studied by the Advanced Method of Transient Microwave Photoconductivity (AMTMP). The essential feature of the method is in obtaining two transient decays at once, the photoconductivity decay, delta sigma(t) and the decay due to changes in the real part of the dielectric constant, delta epsilon'(t) which can be related to trapped electron densities. The analysis of free and trapped electron decays rather than only the photoconductivity decay provides less ambiguous conclusions about the nature of the process observed.
The kinetics were measured in 123-358 K temperature range and in 10^-8-10^-1 s time scale. At all temperatures and intensities delta sigma (t) can be described as t^(a-1) decay truncated by an exponential decay while delta epsilon'(t) decayed more slowly but revealed the similar exponential decay at a tail region. The temperature dependence of a showed an initial decrease with increasing temperature (123-200 K) with a plateau region (210 -270 K) followed by a subsequent increase (280 - 358 K). The forms of both kinetics were simulated using multiple trapping rate equations with a monomolecular recombination term. The forms of experimental decays of delta sigma(t) and delta epsilon'(t) allowed us to rule out exponential, rectangular and linear distributions of localized states as possible distributions in the sample. A peaked Gaussian-like distribution gave very close fit to experimental data transients. Because the parameter is related to a width of distribution as a=T/T0 where T0 is a characteristic temperature it is difficult to explain the decrease of with temperature. To explain these phenomena some authors used a concept of electronic doping which implies a transformation of traps to recombination centers with different capture cross-sections for electrons and holes as temperature increased. This explanation is based on a temperature movement of demarcation levels. We did not observed any significant changes in recombination times during our experiments which would be consistent with electronic doping model. We suggested that at any specific temperature and within a particular time scale the only small part of the real distribution is probed. The Fermi level position would distort the higher energy side of the distribution from the true Gaussian one and give rise to narrowed effective distribution. Therefore, the temperature dependence of will be determined by T/ T0 giving eventually an increase when the temperature is high enough to probe almost the entire distribution. The position of the demarcation level will truncate the continuous distribution at every T giving the last available trap level in the distribution. The temperature dependence of the trap depth extracted from the slow exponential decay follows expected kT dependence supporting the use of demarcation energy concept.