Abstract
Many suggestions have been made to improve the original model of exciton decay which leads to the displacement of halogen ions in the alkali lattice. Several of the models involve thermally-activated movement of the crowdion defect (the H centre). In this paper the models and experiments are reviewed. Many of the problems are resolved by considering the possibility of exciton diffusion before the decay of the excitons by luminescence or a non-radiative process involving a replacement collision sequence. In this model, the collision sequence need only require a few replacements to achieve defect stability. The model also covers the production of HA centres in impure alkali halides and the temperature dependence of the process.