This article gives an overview of the literature on clean technology diffusion, followed by suggestions for future analysis. Findings from diffusion analysis are presented in the form of 10 stylised facts, helping the reader to see the forest for the trees. The overall conclusion is that the diffusion of clean technology (same as the diffusion of normal innovations) is governed by endogenous mechanism (epidemic learning and learning economies) and by exogenous mechanisms. Policy is important for clean technology diffusion but other factors are important too: the characteristics of the clean technology, absorptive capacities of potential adopters and the age structure of capital. It is often overlooked that companies have a choice: they can choose between an end-of-pipe solution, a process change (adaptation) and a change of process (substitution). This means that the diffusion and evolution of one clean technology will be at the expense of the diffusion of another clean technology, something overlooked in studies on clean technology diffusion. Further research is needed on the influence of public policy on clean technology choice, expectations (about learning economies and prices), adjustment costs, network externalities and complementary innovations on clean technology adoption choices. (c) 2007 Published by Elsevier Ltd.