Automated Exploration of Heterogeneous Catalysis with a Gas–Solid Nanoreactor

We present an automated method, gas–solid nanoreactor molecular dynamics (GS-NMD), designed to explore reaction space and construct reaction networks for complex gas–solid heterogeneous catalysis systems by integrating multiple acceleration techniques. Periodic pulses were used to drive gas-phase molecules toward the catalyst surface, accelerating adsorption and Eley–Rideal reactions. Adsorbed species were then subjected to metadynamics to overcome reaction barriers associated with migration, Langmuir–Hinshelwood-type reactions, and desorption, using the root-mean-square deviations in Cartesian space as collective variables. We demonstrate the efficiency of GS-NMD with the case of N2 dissociation on Fe surfaces, showing its ability to effectively screen for low-barrier reactions within a vast reaction space and distinct catalysts of different performances. Additionally, we illustrate the method's utility in constructing effective reaction networks for heterogeneous catalysis, exemplified by ammonia synthesis, which comprises only low-barrier elementary steps. These results suggest that GS-NMD is a promising and efficient tool for the automated exploration of heterogeneous catalysis, enabling the identification of the most favorable mechanisms and active sites for gas–solid reactions.