Second Generation Visible-Light-Active Photocatalysts: Preparation, Optical Properties, and Consequences of Dopants on the Band Gap Energy of TiO₂

Abstract

First generation metal-oxide photocatalysts based mostly on nominally pure, pristine titanium dioxide have been the object of great debate in the past 30 years with regard (i) to the nature of the oxidative agent (^•OH radicals vs. holes h⁺); (ii) to the site at which the reaction takes place (surface vs. bulk solution); (iii) to whether TiO₂ is indeed a photocatalyst since turnover numbers are difficult to determine owing to the nature of the particle surface; and (iv) to how the process efficiency can be ascertained, among many other issues yet to be resolved satisfactorily. One issue that has taken some time to be resolved is the notion of how we can make better use of sunlight’s visible radiation seeing that the absorption edge of TiO₂ is at 387 nm (ca. 3.2 eV – the band gap energy) for the anatase polymorph. A successful strategy that is gaining some momentum is to dope this metal oxide with suitable dopants (e.g., metal ions and/or non-metals) to shift the absorption edge to longer wavelengths. Doping has been achieved using various physical and chemical strategies, which have led to materials whose absorption edges have been red-shifted to wavelengths ~550 nm (and beyond in some cases). The debate that now occupies discussions of doped-TiO₂ materials regards the causes for this red shift. Several reports, based on density functional theory (DFT), have asserted that the band gap of doped-TiO₂ is narrowed because of interactions between the dopant states and the O 2p states of the valence band, thereby pushing the valence band edge upward. Others have proposed isolated dopant states located within the band gap to explain the red shift of the absorption edges of doped-TiO₂ systems through excitation of the electrons in these states to the conduction band of TiO₂. Absorption spectra, calculated from several diffuse reflectance spectra (DRS) reported in the literature for both metal ion-doped TiO₂s and systems doped with non-metals (e.g., carbon, sulfur, nitrogen, and fluorine), are remarkably similar if not identical in the visible spectral region. The broad spectral envelope observed at wavelengths greater than 400 nm can be deconvoluted into 2–3 single bands, which indicate different species give rise to these bands. This chapter is therefore concerned, albeit in a very restrictive way, with the various strategies used to dope TiO₂, with their modeling by DFT methods, and finally with their optical properties with which we shall argue that the absorption edge red-shift originates from a singular source involving mostly the formation of (additional to existing) oxygen vacancies in the metal-oxide lattice (both surface and bulk) that can act as electron traps to yield F-type color centers and/or Ti³⁺ color centers.