High-throughput Computational Workflow for Ligand Discovery in Catalysis with the CSD

A novel semi-automated, high throughput computational workflow for ligand/catalyst discovery based on the Cambridge Structural Database is reported. The transition states of the rate-determining step of the Ullmann-Goldberg reaction were identified and used as a template for a ligand search within the CSD, leading to >33000 potential ligands. The ∆G‡ for catalysts using these ligands were calculated using GFN2-xTB//B97-3c with high success rates and good correlation to benchmarking values from DLPNO-CCSD(T)/def2-TZVPP calculations. Furthermore, machine learning models were developed based on the generated data, leading to accurate predictions of ∆G‡, 70.6-81.5% of prediction falling within ± 4 kcal.mol−1 of the calculated ∆G‡, without the need for the costly calculation the transition state. This accuracy of machine learning models was improved to 75.4-87.8% using descriptors derived from GFN2-xTB//TPSS/def2-TZVP calculations with minimal increase in computational time. This new workflow offers significant advantages over currently used methods due to its faster speed and lower computational cost, coupled with excellent accuracy compared to higher-level methods.