Journal of Chemical Information and Modeling, Год журнала: 2024, Номер unknown

Опубликована: Ноя. 28, 2024

Methods that accelerate the evaluation of molecular properties are essential for chemical discovery. While some degree ligand additivity has been established transition metal complexes, it is underutilized in asymmetric such as square pyramidal coordination geometries highly relevant to catalysis. To develop predictive methods beyond simple additivity, we apply a many-body expansion octahedral and complexes introduce correction based on adjacent ligands (i.e., cis interaction model). We first test model adiabatic spin-splitting energies Fe(II) predicting DFT-calculated values unseen binary within an average error 1.4 kcal/mol. Uncertainty analysis reveals optimal basis, comprising homoleptic mer symmetric complexes. next show solved basis) infers both DFT- CCSD(T)-calculated catalytic reaction 1 kcal/mol average. The predicts low-symmetry with outside range complex energies. observe trans interactions unnecessary most monodentate systems but can be important combinations ligands, containing mixture bidentate ligands. Finally, demonstrate may combined Δ-learning predict CCSD(T) from exhaustively calculated DFT same fraction needed model, achieving around 30% using alone.

Язык: Английский