First-Principles Research On N/F-Doped And N-F-Codoped TiO₂(101) Surfaces

In order to enhance the visible light (VIS) sensitivity of TiO₂ photocatalyst materials,people have made many efforts including introducing anionic species for doping. Recently,the research on nitrogen(N)/fluorin(F)-doped and N-F-codoped TiO₂ has made effectiveprogress. Some groups confirmed that N/F-doping and N-F-codoping can enhance thephotocatalysis of TiO₂ under VIS radiation experimentally. However the mechanism of theVIS sensitivity for N/F-doping and N-F-codoping is still a controversial question fortheoretic researchers. One of the reasons may be the differences between the models usedin simulation calculations with the material synthesized in lab. For example, most ofsimulation calculations were performed on TiO₂ bulk structure, whereas the surfaceproperties play an important role on photocatalytic reactions. Moreover, many theoreticworks considered substitutional doping only, neglecting the situation of interstitial dopingas well as adsorption.In our present work, the electronic structures of N/F-doped and N-F-codoped TiO₂anatase (101) surfaces have been investigated by density functional theory (DFT)plane-wave pseudopotential method in order to give a further insight on the mechanism ofthe VIS sensitivity for N/F-doped and N-F-codoped TiO₂. Because the general gradientapproximation (GGA) which was used to describe the exchange-correlation effects alwaysleads to a severe underestimation of the band gap for the case of transition metal oxide,GGA + U (Hubbard coefficient) method was also adopted to calculate the electronicstructures. Our work includes the following parts:First, the model of TiO₂ anatase (101) surface used in simulation calculations has beenbuilt by analyzing surface energy and oxygen vacancy formation energy. Both GGA andGGA+U methods were performed to calculate the electronic structure of pure anatase (101)surface as well as the defective surface (the surface with oxygen vacancies). GGAcalculations demonstrated that the introducing of oxygen vacancy takes little effect onband gap narrowing, whereas GGA +U calculations confirmed that oxygen vacancyreduces the band gap effectively. Second, N/F-doped and N-F-codoped TiO₂ anatase (101) surfaces have beeninvestigated by both GGA and GGA+U methods. Besides the situation of substitutionaldoping, N/F interstitial doping in the surfaces as well as N/F adsorption on the surfaceshas also been taken into account in order to find how different doping styles make effecton the electronic structure of TiO₂. Because it has been confirmed that N/F-doping islikely introducing oxygen vacancies in TiO₂, the N/F-doped TiO₂ anatase (101) defectivesurface has also been analyzed as well as the N-F-codope, d defective surface.GGA calculations demonstrated that F 2p states take no effect on band gap narrowingand N induced states have a positive effect on band gap narrowing by leading anexpansion of VB. Whereas GGA+U calculations gave different results. There is noobvious expansion of VB neither band gap narrowing observed by N-doping besides someisolated N 2p states lying in the gap. And F-doping was found playing an important roleon band gap narrowing. From the comparison of GGA and GGA+U calculations, wefound GGA+U calculations can give a better explanation for the reported experimentalobservations. Additionally, both GGA and GGA+U calculations for interstitial N-dopingand surface N/F adsorption on TiO₂ showed that N dopant/adsorbate introduces severalnew states between VB and CB of TiO₂, and F adsorbate takes no effect on electronicstructure of TiO₂.Lastly, we have investigated several possible adsorption configurations for an isolatedO₂ molecule at different surface oxygen vacancy sites on TiO₂ anatase (101) surface. Theelectronic properties of the most energetic adsorption configuration have been discussedand it was found that the appearance of oxygen adsorbate induced states near the edge ofthe VB may attribute to the improvement of visible light sensitivity of TiO₂ surface.