(672b) Pushing Boundaries in Microkinetics through Automated Mechanism Generation for Fischer-Tropsch Synthesis

The Co-catalyzed Fischer-Tropsch synthesis (FTS) is a crucial process for a sustainable aviation industry, but it is one of the most challenging reactions to model on a first-principles basis. Challenges arise due to the complex active sites on Co, which transform under reaction conditions. Yet, the construction of reaction mechanisms that account for all possible products is far more challenging. The broad product distribution causes an explosion of possible pathways, which is impossible to explore manually. Consequently, microkinetics are truncated at the C₃ chemistry, while products of interest are in the fuel range. Even constructing these “small” mechanisms requires chemical intuition to pre-select pathways from the vast reaction space, which introduces bias. To tackle the complexity of the FTS, we propose an automated generation of microkinetics using the open-source Reaction Mechanism Generator (RMG). RMG explores the entire reaction space for gas-phase or catalyzed reactions but retains only the kinetically relevant pathways using a rate-based algorithm.

We extended RMG with functionalities to identify FTS pathways on a Co(0001) facet. New reaction families, thermo-kinetic databases, and estimation routines are provided using an accelerated workflow with toolkits from the Open Catalyst Project. Crucial new features were implemented that enable the handling of large molecules, including multidentate functionalities and resonance structure generation.

RMG generates microkinetics for the FTS on Co(0001), which contain 12000 reactions, while more than 150000 possible pathways were explored (see Fig. 1). This mechanism includes pathways from the CO-insertion and carbide mechanism. Additionally, the discovered mechanisms consist mostly of abstraction reactions, which are currently not considered in the literature. Reactor simulations with Cantera highlight the importance of coverage effects and multiple active site motifs to accurately predict experiments. The improvements made to RMG in this study provide a universally applicable tool for other challenging catalytic reactions like catalytic plastic waste upcycling.