Edwards et al. in their communication in essence convey the impression that the photoreduction of dinitrogen on heterogeneous photocatalysts is thermodynamically close to impossible and that the ammonia yields observed by previous workers''. 31 may have been due to erroneous measurements of background contamination.
The authors furthermore claim to have conducted numerous experiments with promising photocatalysts under conditions identical to those used by the earlier investigators. This is actually not quite true. For example, Edwards et al. performed their experiments in quartz, rather than Pyrex vessels and employed, a medium-pressure Hg lamp; accordingly, the relative contribution of short UV light was higher than in the experiments conducted by us and most other workers in the field. Furthermore, and more seriously, the majority of the experiments seem to have been done with solids that would be expected to have little or no photocatalytic activity. Thus, Edwards et al. performed most of their experiments with "binary oxide xerogels" composed of TiO,, Fe,O,, NiO, A1,0,, and S O , , which we would judge to have little chance of being photoactive in the first place. Only very few experiments were actually performed with iron-doped TiO, under conditions where nitrogen photoreduction could have taken place. In these experiments, the Fe-TiO, photocatalysts must first be heat-treated at defined temperatures and for known periods of time and subsequently stored in humid air. From the information given, it is our impression that the authors did not spend enough time and effort to optimize the photocatalyst pretreatment conditions. It should be recalled that only homogeneous samples having Fe3+ ions in the lattice of TiO, are photoactive. Heterogeneous samples, that is, those in which the surface is covered by iron oxides, or those with mixed iron-titanium oxides such as Fe,TiO, on the surface, are photochemically completely inactive. The heat treatment temperature must be chosen so that the anatase-rutile conversion does not occur too rapidly. Pro- longed heating at temperatures that are too high produces large crystals of low or zero photocatalytic activity. The optimal heating temperatures and times must be determined for each batch, as both vary from sample to sample, and