Photoluminescence decay time studies of type-II GaAs/AlAs quantum well structures

This is a comment on a paper with the same title by Dawson et al. [J.Appl. Phys. 65, 3606 (1989)], which argues that the nonexponential decay of X-point excitons in type-II GaAs/AlAs superlattices previously reported by one of us [E. Finkman, M. D. Sturge, and M. C. Tamargo, Appl. Phys. Lett. 49, 1299 (1986)] is due to spectral diffusion within the inhomogeneous line, and that where diffusion is absent, the decay is exponential. We report time decay measurements as a function of excitation intensity on the samples used by Dawson et al., and on similar ones from a different source. We find that at very low intensities the results agree with those of Finkman et al., while at high intensities they agree with Dawson et al. We argue that spatial fluctuations in the transition matrix element due to irregularities in the interfaces are responsible for the nonexponential behavior, and that the change at high intensity is due to delocalization of excitons and consequent spatial averaging of the transition matrix element.