Computer simulation of defects and reactions at oxide surfaces

Abstract Oxide surfaces are technologically important in a range of applications, including catalysis, corrosion, gas sensors, ceramics and high-temperature superconductivity. In recent years, computer simulation methodologies and algorithms have developed to the stage where simulation can now play a complementary role with experiment, aiding in interpretation of experimental data and predicting surface behaviour. Several different approaches to the study of oxide surfaces have been developed and their use demonstrated. Static lattice simulations allow surfaces as well as point and extended defects in highly ionic materials to be investigated using an essentially classical model with simple interionic potentials. Comparison with experimental surface segregation studies shows the great predictive power of these calculations, and indeed much expensive experimental work could now be obviated. These classical models have also proved valuable in understanding the topology of non-planar surfaces. However, when an understanding of the electronic behaviour of the system is required, for example in studying reactivity, we turn instead to quantum mechanical methods. In recent years both cluster approaches (with a suitable cluster termination to represent the bulk lattice) and periodic treatments have been used with success. Particularly in the case of periodic treatments, it is possible to assess the predicted degree of ionicity and thus to confirm or disprove the validity of the static lattice approach for various different systems. With these developments, simulation is poised to make a valuable contribution as full partner to experiment in areas such as ceramics processing and catalyst design.