Fig. 3. A database of 25 116 molecular transition-metal catalyst candidates. Each complex consists of one out of six transition metals and a combination of two out of 91 ligands (left) (see the ESI† for details). Each ligand was written as the SMILES notation and all possible L1–M–L2 combinations were constructed (top right-hand corner). SMILES strings were then converted into Cartesian coordinates through the 3D structure generator of the OpenBabel software (bottom right-hand corner). DFT reference results for training and testing of ML models were obtained for a sub-set of 7054 candidates. Those structures were exploited for computing binding energies and for training the ML models.
