Comparison of phenomenological kinetics and kMC modelling of CO oxidation at RuO2(110)

Neglecting the explicit spatial distribution of the chemicals at the catalyst surface is a frequently employed concept in phenomenological kinetics (PK). We scrutinize this concept by comparing with results from kinetic Monte Carlo (kMC) simulations, which do not need to rely on this mean-field approximation. For the model system CO oxidation at RuO2(110) we compute the surface composition and activity of the catalyst surface in reactive environments ranging from ultra-high vacuum to technologically relevant conditions (pressures of several atmospheres and elevated temperatures). Using the same first-principles rate constants in both approaches, we find the PK modeling to be in serious error, and even failing to identify the correct dominant reaction mechanism. Approaches to improve on this situation, by explicitly including diffusion events and site correlations in the PK modeling, will be discussed