Quantitative spin-trapping studies of weakly illuminated titanium dioxide sols. Implications for the mechanism of photocatalysts

Initial quantum yields phi for the formation of DMPO-OH adducts in clear, aerated, unbuffered TiO2 sols irradiated at 295 nm were determined by kinetic electron paramagnetic resonance spectrometry as function of the 5,5-dimethyl-1-pyrroline N-oxide (DMPO) spin trap concentration, photon irradiance I-o, and added dichlorobenzene DCB, at 298 K. phi steadily increases with [DMPO] in the range 0.3 mM to 0.3 M, extrapolating to phi([DMPO]) (-infinity) = 0.54 +/- 0.10, in contrast with the constant yields obtained for the homogeneous generation of OH radicals in the photodissociation of H2O2 at 254 nm under similar conditions, In TiO2 sols at [DMPO] = 0.6 mM, phi (approximate to 0.002) remains constant over a 200-fold variation of I-o, but decreases upon addition of comparable DCB concentrations. These observations prove that (1) DMPO reacts with photogenerated holes and/or OH radicals on the surface of TiO2 particles rather than in the solution bulk, (2) it is possible to approach quantitative carrier trapping, (3) DCB is as reactive as DMPO toward some of the photogenerated carriers, and (4) the recombination of a single carrier pair-the usual event in small particles al moderate photon fluxes-is a pseudo-first-order process. Photocatalysis in semiconductor sols is a stochastic process not amenable to conventional kinetic analysis based on rate expressions involving continuous concentration variables. Quantum yields are largely controlled by the anodic (h + D --> D+), rather than by the slower (e + O-2 --> O-2(.-)) cathodic charge transfer reaction.