Challenges in using ammonia decomposition for hydrogen production include increasing catalyst activity and developing cheaper catalysts. While Ru is the best single metal catalyst, multi-site catalysts, including bimetallics, offer possible improvements. We aim to develop a unified modeling framework for understanding and designing multi-site catalysts to meet these challenges.

Microkinetic models typically assume a homogeneous catalyst surface. A catalyst consists of various crystallographic planes, steps and defects, and multiple chemical components (different metal islands), which form multiple, interacting sites. The kinetics on different sites and their interactions must be captured in predictive models for these spatially distributed situations. These interactions ‘couple' microkinetic models; we show that coupling is subject to thermodynamic constraints, and we develop appropriate multi-site models to compute reaction rates. The applicability of such mean-field models is assessed via kinetic Monte Carlo simulations.

We show an example for ammonia decomposition on a bimetallic catalyst containing Pt and Ru, which have the same rate-determining step and fall on the same side of the “volcano curve”. Fig.1 shows ammonia conversion as a function of Ru fraction. Both non-interacting and weakly interacting (adsorbed species can diffuse to and react on the other catalyst surface) cases are shown. Pure Ru is the best catalyst, as expected for this situation. This approach is being extended to a library of catalysts and best bimetallic catalysts will be discussed. The effect of steps will also be illustrated to underscore their relative importance.