The removal efficiency for 220 micropollutants was studied at the scale of a municipal wastewater treatment plant (WWTP) upgraded with post-ozonation followed by sand filtration. During post-ozonation, compounds with activated aromatic moieties, amino functions, or double bonds such as sulfamethoxazole, diclofenac, or carbamazepine with second-order rate constants for the reaction with ozone >10(4) M-1 s(-1) at pH 7 (fastreacting) were eliminated to concentrations below the detection limit for an ozone dose of 0.47 g O-3 g(-1) dissolved organic carbon (DOC). Compounds more resistant to oxidation by ozone such as atenolol and benzotriazole were increasingly eliminated with increasing ozone doses, resulting in >85% removal for a medium ozone dose (similar to 0.6 g O-3 g(-1) DOC). Only a few micropollutants such as some X-ray contrast media and triazine herbicides with second-order rate constants <10(2) M-1 s(-1) (slowly reacting) persisted to a large extent. With a medium ozone dose, only 11 micropollutants of 55 detected in the secondary effluent were found at > 100 ng L-1. The combination of reaction kinetics and reactor hydraulics, based on laboratory and full-scale date, enabled a quantification of the results by model calculations. This conceptual approach allows a direct upscaling from laboratory- to full-scale systems and can be applied to other similar systems. The carcinogenic by-products N-nitrosodimethylamine (NOMA),(<= 14 ng L-1) and bromate (<10 mu g L-1) were produced during ozonation, however their concentrations were below or in the range of the drinking water standards. Furthermore, it could be demonstrated that biological sand filtration is an efficient additional barrier for the elimination of biodegradable compounds formed during ozonation such as NOMA. The energy requirement for the additional post-ozonation step is about 0.035 kWh m(-3), which corresponds to 12% of a typical medium-sized nutrient removal plant (5 g DOC m(-3)).