Enabling Automation\nof de Novo Catalyst Design: An\nExperimentally Validated, Multifactor Design Metric for Olefin Metathesis

Automated methods for molecular design navigate chemical\nspace\nby ranking candidate compounds against predefined, numerical design\nmetrics. To date, metrics for homogeneous catalysts focus on catalyst\nactivity as the sole criterion, neglecting performance-critical factors\nsuch as stability and degradation. Here we introduce a general, multifactor\ndesign metric for molecular catalysts, and highlight the opportunities\ncreated by mechanistically-based de novo design by implementing this\nmetric within the showcase application of olefin metathesis. Ruthenium-catalyzed\nolefin metathesis offers a prominent context within which the central\nimportance of catalyst degradation is now widely acknowledged, and\nmechanistic understanding of the decomposition pathways has reached\nan advanced stage. A numerical figure of merit (or “fitness\nscore”) for these catalysts is generated by combining functions\nbased on DFT-calculated relative energies, which describe (i) catalyst\ninitiation, (ii) catalyst activity in the metathesis of terminal olefins,\n(iii) catalyst stability–specifically, resistance to decomposition\nvia β-hydride elimination, (iv) the synthetic accessibility\nof the precatalyst, and (v) its thermodynamic stability in the <i>trans</i>-anionic geometry essential for high activity. By comparing\ncalculated fitness scores with catalytic turnovers measured in benchmark\nolefin metathesis reactions, we demonstrate that this multifactor\nfitness function reproduces the experimental ranking and productivity\ntrend for catalysts known to exhibit profoundly different susceptibilities\nto decomposition. The trend cannot be reproduced by considering in\nisolation any of the individual factors, including catalytic activity\nor resistance to β-hydride elimination. The fitness formulation\npresented here establishes a foundation for automated screening and\ndesign of improved catalysts for olefin metathesis. More broadly,\nit establishes a general strategy for development of multifactor design\nmetrics for molecular catalysts that incorporate mechanistic understanding\nof catalyst activity, stability, and synthetic accessibility.