Bayesian optimization- and high-throughput-guided synthesis of monofluoroalkenes via defluorinative C-C bond formation.

Monofluoroalkenes are privileged fluorinated motifs in medicinal chemistry, but their selective synthesis through C-F bond activation remains challenging because of the high strength of C-F bonds and the condition-sensitive nature of defluorinative transformations. To address this challenge, this work establishes a Pd-catalyzed defluorinative C-C bond-forming strategy that couples trifluoroethyl ketones, esters, and amides with diverse organoboron reagents to deliver structurally diverse monofluoroalkenes. The integration of Bayesian optimization with high-throughput experimentation (HTE) enabled rapid condition optimization, identifying productive reaction conditions for the model substrate pair from a 376-condition table with a 53-condition Bayesian-optimization selection budget. To efficiently explore substrate generality under limited experimental budgets, the organoboron reagent space was further stratified on the basis of literature precedent, commercial availability, and structural diversity. Under the optimized condition, 970 curated substrate-reagent reaction entries were evaluated across 11 HTE plates, enabling rapid assessment of substrate compatibility and identification of productive monofluoroalkene-forming reactions. Leveraging the resulting single-condition HTE dataset, a two-stage decision-support model was further developed for reactivity classification and yield prediction, with the XGBoost model achieving an F1-score of 0.932 for classification and an R 2 of 0.852 with a mean absolute error of 4.03% for yield regression on the held-out test set. Overall, this study establishes an experimentally grounded workflow that combines Bayesian optimization, HTE, substrate-space stratification, and data-assisted modeling to accelerate reaction discovery in a condition-sensitive C-F bond activation system.