Comparative assessment of DFT performances in Ru- and Rh-promoted σ-bond activations

In this work, the performances of 19 density functional theory (DFT) methods are calibrated comparatively on Ru- and Rh-promoted σ-bond (C-H, O-H, and H-H) activations. DFT calibration reference is generated from explicitly correlated coupled cluster CCSD(T)-F12 calculations, and the 4s4p core-valence correlation effect of the two 4d platinum group transition metals is also included. Generally, the errors of DFT methods for calculating energetics of Ru-/Rh-mediated reactions appear to correlate more with the magnitude of energetics itself than other factors such as metal identity. For activation energy calculations, the best performing functionals for both Ru and Rh systems are MN12SX < CAM-B3LYP < M06-L < MN12L < M06 < ωB97X < B3LYP < LC-ωPBE (in the order of increasing mean unsigned deviations, MUDs, of less than 2 kcal/mol). For reaction energy calculations, best functionals with MUDs less than 2 kcal/mol are PBE0 < CAM-B3LYP ≈ N12SX. The effect of the DFT empirical dispersion correction on the performance of the DFT methods is beneficial for most density functionals tested in this work, reducing their MUDs to different extents. After including empirical dispersion correction, ωB97XD, B3LYP-D3, and CAM-B3LYP-D3 (PBE0-D3, B3LYP-D3, and ωB97XD) are the three best performing DFs for activation energy (reaction energy) calculations, from which B3LYP-D3 and ωB97XD can notably be recommended uniformly for both the reaction energy and reaction barrier calculations. The good performance of B3LYP-D3 in quantitative description of the energetic trends further adds value to B3LYP-D3 and singles this functional out as a reasonable choice in the Ru/Rh-promoted σ-bond activation processes.