Work on the design of new TiO2 based photocatalysts is described. The key concept is the formation of composite structures through the modification of anatase and rutile TiO2 with molecular-sized nanoclusters of metal oxides. Density functional theory (DFT) level simulations are compared with experimental work synthesizing and characterizing surface modified TiO2. DFT calculations are used to show that nanoclusters of metal oxides such as TiO2, SnO/SnO2, PbO/PbO2, ZnO and CuO are stable when adsorbed at rutile and anatase surfaces, and can lead to a significant red shift in the absorption edge which will induce visible light absorption; this is the first requirement for a useful photocatalyst. The origin of the red shift and the fate of excited electrons and holes are determined. For p-block metal oxides the oxidation state of Sn and Pb can be used to modify the magnitude of the red shift and its mechanism. Comparisons of recent experimental studies of surface modified TiO2 that validate our DFT simulations are described. These nanocluster-modified TiO2 structures form the basis of a new class of photocatalysts which will be useful in oxidation reactions and with a correct choice of nanocluster modified can be applied to other reactions.