Bandgap narrowing of TiO$_{2}$ via codoping for enhanced photocatalytic reactions*

We investigated the growth of Cr$-$N codoped single-crystal anatase TiO$_{2}$(001) (A-TiO$_{2})$ thin films using pulsed laser deposition with a target of Cr$_{2}$O$_{3}$ and TiN mixture. N concentrations were finely tuned under different growth temperatures and oxygen pressures, and high quality films with atomically flat terraces were obtained. UV$-$Vis absorption measurements showed that the band-gap of the codoped A-TiO$_{2}$ film is significantly narrowed in comparison with the undoped and monoelement doped films. We further systematically investigated the structures and the activity of the oxidized and reduced (1 $\times$ 4) reconstructed surfaces of A-TiO$_{2}$ epitaxially grown on SrTiO$_{3}$ using scanning tunneling microscopy/spectroscopy, X-ray/ultraviolet photoemission spectroscopy and first-principles calculations. Quite unexpectedly, it is found that the perfect (1 $\times$ 4) surface of A-TiO$_{2}$ is not even active for H$_{2}$O and O$_{2}$ adsorption at room temperature. Two types of intrinsic point defects are identified, among which only the Ti$^{3+}$ defect site on the reduced surface demonstrates considerable activity for H$_{2}$O and O$_{2}$ adsorption. The perfect surface itself should be fully oxidized, but shows no obvious activity. We thus propose an oxidized ridge model for the reconstructed (1 $\times$ 4) surface, where the Ti atoms at the normal ridge sites are sixfold coordinated. The Ti-rich point defects on reduced surface are fourfold-coordinated. This model provides consistent explanations for our experimental observations We have compared the results with those from rutile TiO$_{2}$(001)-(1 $\times$ 1) surface in our investigations. Our findings suggest that the activity of the A-TiO$_{2}$ surface should depend on its reduction status, similar to that of rutile TiO$_{2}$ surfaces.