Enabling Automation of de Novo Catalyst Design: An Experimentally Validated, Multifactor Design Metric for Olefin Metathesis

Abstract

Automated methods for molecular design navigate chemical space by ranking candidate compounds against predefined, numerical design metrics. To date, metrics for homogeneous catalysts focus on catalyst activity as the sole criterion, neglecting performance-critical factors such as stability and degradation. Here we introduce a general, multifactor design metric for molecular catalysts, and highlight the opportunities created by mechanistically-based de novo design by implementing this metric within the showcase application of olefin metathesis. Ruthenium-catalyzed olefin metathesis offers a prominent context within which the central importance of catalyst degradation is now widely acknowledged, and mechanistic understanding of the decomposition pathways has reached an advanced stage. A numerical figure of merit (or “fitness score”) for these catalysts is generated by combining functions based on DFT-calculated relative energies, which describe (i) catalyst initiation, (ii) catalyst activity in the metathesis of terminal olefins, (iii) catalyst stability–specifically, resistance to decomposition via β-hydride elimination, (iv) the synthetic accessibility of the precatalyst, and (v) its thermodynamic stability in the trans-anionic geometry essential for high activity. By comparing calculated fitness scores with catalytic turnovers measured in benchmark olefin metathesis reactions, we demonstrate that this multifactor fitness function reproduces the experimental ranking and productivity trend for catalysts known to exhibit profoundly different susceptibilities to decomposition. The trend cannot be reproduced by considering in isolation any of the individual factors, including catalytic activity or resistance to β-hydride elimination. The fitness formulation presented here establishes a foundation for automated screening and design of improved catalysts for olefin metathesis. More broadly, it establishes a general strategy for development of multifactor design metrics for molecular catalysts that incorporate mechanistic understanding of catalyst activity, stability, and synthetic accessibility.