Oxygen vacancies and nonmetal dopant species in anatase TiO2: A lesson learned?

Visible-light sensitization of TiO2 is a holy grail of research on photocatalysis. With the aim of better exploiting solar radiation to activate the photocatalytic process, numerous strategies have been reported to extend TiO2 light-response to the visible region, often with limited success. Among them, N-doping has been one of the most popular1. Although the benefits of N-doping in terms of visible light absorption are often outweighed by a stability loss and faster charge carrier recombination, the debate over these materials has spurred more fundamental questions about the nature of TiO2 defects and the visible light absorption mechanism. These questions prompted us to develop a combined experimental and theoretical approach to overcome the limitations of individual techniques2. We investigated a broad range of N-doped sol-gel TiO2 samples from several N-sources (NH3, urea, TEA) and nominal content. The materials structure and defectivity was studied by combining HR-XRD, EXAFS, EPR, DRS, PL spectroscopy, and DFT calculations of structural and electronic features. EXAFS and DFT results showed huge differences in the local environment of Ti centers as a function of the N-source type, mirroring a different interplay of N species location and O vacancies. Differential DRS spectra obtained before and after light irradiation3, supported by DFT electronic structure calculations and by PL spectra, provided further insight into the diverse structural defects in the three families of N-doped materials. Our results offered an interpretative basis for the different stability in time of samples’ paramagnetic species and for their photocatalytic activity, as determined by UV and visible-light photocatalytic tests for the gas phase degradation of VOCs and of antibiotics in water.