Ab initio nonadiabatic molecular dynamics of wet-electrons on the TiO(2) surface

The electron transfer (ET) dynamics of wet-electrons on a TiO(2) surface is investigated using state-of-the-art ab initio nonadiabatic (NA) molecular dynamics (MD). The simulations directly mimic the time-resolved experiments [Science 2005, 308, 1154] and reveal the nature of ET in the wet-electron system. Focusing on the partially hydroxylated TiO(2) surface with 1-monolayer water coverage, and including electronic evolution, phonon motions, and electron-phonon coupling, the simulations indicate that the ET is sub-10 fs, in agreement with the experiment. Despite the large role played by low frequency vibrational modes, the ET is fast due to the strong coupling between the TiO(2) surface and water. The average ET for the system has equal contributions from the adiabatic and NA mechanisms, even though a very broad range of individual ET events is seen in the simulated ensemble. Thermal phonon motions induce a large fluctuation of the wet-electron state energy, generate frequent crossings of the donor and acceptor states, and drive the adiabatic mechanism. The rapid phonon-assisted NA tunneling from the wet-electron state to the TiO(2) surface is facilitated by the strong water-TiO(2) electronic interaction. The motions of molecular water have a greater effect on the ET dynamics than the hydroxyl vibrations. The former contribute to both the wet-electron state energy and the water-TiO(2) electronic coupling, while the latter changes only the energy and not the coupling. Delocalized over both water and TiO(2), wet-electrons are supported by a new type of state that is created at the interface due to the strong water-TiO(2) interaction and that cannot exist separately in either material. Similar states are present in a number of other systems with strong interfacial coupling, including certain dye-sensitized semiconductors and metal-liquid interfaces. The ET dynamics involving such interfacial states share many universal features, such as an ultrashort time scale and weak-dependence on temperature, surface defects, and other system details.