Quantitative Modeling of Electron Dynamics and the Effect of Diffusion in Photosensitized Semiconductor...

Photosensitized semiconducting nanomaterials have received considerable attention because of their applications in photocatalytic and photoelectronic devices. In such systems, photoexcited electrons with sufficiently high energies can be injected into the conduction band (CB) of an adjacent semiconductor. These excited electrons are subjected to various physical processes that can lead to their annihilation before exercising their catalytic/electric functions, and the efficiency of the photosensitized functions depends on the quantity of CB electrons produced and how long they remain near the surface region of the semiconductor. The rise and decay of photoexcited electrons in the semiconductor CB can be probed with transient IR absorption (TA), which was first demonstrated by Lian and co-workers. Results from various laboratories have since revealed that electrons appear in the CB following the excitation of the photosensitizer in tens to hundreds of femtoseconds and that the decay of the CB electrons typically exhibits multiple exponentials on varying ultrafast time scales. The size of the semiconductor nanoparticle appears to influence the diffusion of the CB electrons and thus their lifetimes. In all studies reported, the observed multiexponential decays have been analyzed and interpreted using purely phenomenological models, in which the individual decays were intuitively assigned to one specific relaxation or loss process. In reality, however, each exponential decay can be a convolution of multiple physical processes. In this Account, we report a universally applicable physical model, constructed by including all known electron dynamic processes, to quantitatively account for the multiexponential decays. We characterize the model as universal, as it can be used to analyze our own TA measurements, as well as data acquired in other laboratories. In our study of TiO₂ nanorods photosensitized by Ag platelets, we demonstrate that each of the observed triple-exponential decays corresponds to a convolution of several physical decay processes occurring on similar time scales. The rate of each of the processes can be deconvoluted and determined to construct a complete, physically based model to assess the most important question: How many CB electrons are near the semiconductor surface region and what is their lifetime?The size of the semiconductor is an important consideration. Intuitively, as the semiconductor volume increases, there is more room for CB electrons to diffuse around, which increases their lifetime as annihilation occurs primarily at the surface. Indeed, Tachiya and co-workers previously reported that this lifetime increases with particle size. Nevertheless, while CB electrons live longer in the bulk of the particle, they are only useful when they are at the surface. Overall, what really matters is the CB electrons near the surface region, where the photosensitized functions actually occur. In applying our model to analyze the previously reported size-dependent Au/TiO₂ results, we successfully reproduced the observation that larger semiconductor nanoparticles lengthen the lifetime of CB electrons because of diffusion into the bulk. More importantly, however, our model reveals that the size of the semiconductor has almost no influence on the retention of CB electrons near the semiconductor surface. This information is only revealed when all physical processes are quantitatively taken into account for the observed electron dynamics, which is not feasible with a phenomenological approach.