The photooxidative degradation of phenol in aqueous TiO2 dispersions has been revisited to determine the dependencies of the rate on the concentration of phenol and on the photon flow (rho) of the actinic light at 365 nm. The principal objective was to assess the factors that influence the efficiency of the photocatalytic process, the rate of which is described by the function dC/dt(rho, C) = (const) C(n)rho(m) where n and m are orders of the reaction with respect to concentration and photon flow (light intensity), respectively. The reaction order m varies with reagent concentration C, whereas the order n depends on photon flow rho. The description indicates that m -- 1 if n --> 0, whereas n --> 1 if m --> 0. Therefore, the reaction orders m and n of phenol photodegradation are interdependent. A detailed kinetic description of the process is given based on two well-known mechanistic/kinetic models, namely (i) the Langmuir-Hinshelwood (ZH) model whereby the organic reagent is pre-adsorbed on the photocatalyst surface prior to UV illumination, and (ii) the Eley-Rideal (ER) model for which the organic reagent diffuses from the solution bulk onto the photocatalyst surface to interact with the activated state of the photocatalyst. The kinetic treatment infers that it is possible (under certain conditions) to delineate between the LH and ER mechanistic models on the basis of the magnitude of the Langmuir constant K-L at very high photon how i.e. when rho-->infinity (for the LH pathway, K-L-->K; for the ER model K-L-->0), and on the dependence of k(obs), of the process on rho. For the ER model, kobs scales linearly with p at high photon flow, whereas for the LH pathway k(obs), displays a sub-linear dependence on rho and tends toward saturation at high photon flow. (C) 2000 EIsevier Science S.A. All rights reserved.