The direct–indirect kinetic model in photocatalysis: A reanalysis of phenol and formic acid degradation rate dependence on photon flow and concentration in TiO2 aqueous dispersions

Abstract

Literature data concerning TiO2 photocatalytic oxidation of two model organic substrates, phenol and formic acid, are shown to be incompatible with the behavior predicted by the Langmuir–Hinshelwood (L–H) kinetic model. These data are reanalyzed in detail from a kinetic/mechanistic point of view to the light of the direct–indirect (D–I) model (D. Monllor-Satoca, R. Gomez, M. Gonzalez-Hidalgo, P. Salvador, Catal. Today 129 (1–2) (2007) 247), developed as an alternative to the L–H model. Two interfacial charge transfer mechanisms are considered by the (D–I) model: the indirect transfer (IT) mechanism, which is concerned with the adiabatic transfer of holes trapped at TiO2 terminal oxygen ions to dissolved substrate species, and the direct transfer (DT) mechanism dealing with the inelastic transfer of free holes to specifically adsorbed substrate species. While IT is the only mechanism taking place in the absence of specific adsorption, both DT and IT mechanisms actuate simultaneously under specific adsorption, although DT prevails on IT at high enough UV illumination intensity. On the experimental basis that in aqueous solution phenol is not specifically adsorbed on TiO2, it is shown that the experimental photodegradation rate dependence on photon flux (ρ) and phenol concentration (C) is not compatible with a Langmuir type expression, but can be fitted by the photooxidation rate expression (dC/dt)=[(aC)2+2k0aρC]1/2−aC, as predicted by the D–I model for IT, where a is an experimental parameter involving the rate constants for electron–hole recombination and transfer of photogenerated electrons and holes at the semiconductor–electrolyte interface. An apparent dependence of a on ρ and C is observed experimentally. Two possible hypothesis are invoked in order to explain this behaviour:a shift of the TiO2 energy levels, due to an accumullation of electric charge at the semiconductor surface, and/or the partial contribution of inelastic DT of holes to incipiently adsorbed phenol species. Evidence is given that the photooxidation rate of formic acid, which is shown to be specifically adsorbed on TiO2 in the presence of water, depends linearly on ρ, even at high enough photon flux values, as predicted by the D–I model when DT prevails on IT, but in contradiction with the behavior predicted by the L–H model.