In recent years it has become apparent that the novel properties of nanomaterials may predispose them to a hitherto unknown potential for toxicity. A number of recent toxicological studies of nanomaterials exist, but these appear to be fragmented and often contradictory. Such discrepancies may be, at least in part, due to poor description of the nanomaterial or incomplete characterization, including failure to recognise impurities, surface modifications or other important physicochemical aspects of the nanomaterial. Here we make a case for the importance of good quality, well-characterized nanomaterials for future toxicological studies, combined with reliable synthesis protocols, and we present our efforts to generate such materials. The model system for which we present results is TiO2 nanoparticles, currently used in a variety of commercial products.