Transient Absorption Studies and Numerical Modeling of Iodine Photoreduction by Nanocrystalline TiO2 Films

We have used transient absorption spectroscopy to study the reaction between photogenerated electrons in a dye-free nanocrystalline titanium dioxide film and an iodine/iodide redox couple. Recombination kinetics was measured by recording the transient optical signal following band gap excitation by a UV laser pulse. In the presence of a methanol hole scavenger in the electrolyte, a long-lived (0.1−1 s) red/infrared absorbance is observed and assigned to photogenerated electrons forming Ti3+ species. In the presence of iodine and excess iodide in the electrolyte, the signal decays on a millisecond−microsecond time scale, assigned to reduction of the redox couple by photogenerated electrons in the TiO2. The electron lifetime decreases inversely with increasing iodine concentration, indicating that the back reaction is first order in [I2]. No evidence for I2 - is observed, indicating that the reaction mechanism does not involve the formation of I2 - as an intermediate. The shape of the kinetics evolves from monoexponential at low [I2] to stretched-exponential as [I2] increases. A Monte Carlo continuous-time random walk model is implemented to simulate the kinetics and its [I2] dependence and used to address the order of the recombination reaction with respect to electron density, n. The model incorporates the diffusion of oxidized species from the electrolyte toward the TiO2 surface as well as electron trapping and transport in the TiO2. In the limit of low [I2], the monoexponential kinetics is explained by the recombination reaction being rate limited by the diffusion of the oxidized species in the electrolyte. The stretched-exponential behavior at high [I2] can be explained by the reaction being rate limited by the transport of electrons through a distribution of trap states toward reactive sites at the TiO2−electrolyte interface, similar to the mechanism proposed previously for the kinetics of electron−dye cation recombination. Such trap-limited recombination can also explain the superlinear dependence of electron recombination rate on electron density, which has been reported elsewhere, without the need for a reaction mechanism that is second order in n. In contrast, a second-order reaction mechanism in a trap-free medium cannot explain the observed kinetics, although a second-order mechanism incorporating electron trapping cannot be conclusively ruled out by the data. We propose that the most likely reaction scheme, that is first order in both [I2] and n