The experimentally motivated question of the acetylene bromoboration mechanism was addressed in order to suggest possible radical isomerization pathways for the syn-adduct. Addition–elimination mechanisms starting with a bromine radical attack at the “bromine end” or the “boron end” of the C=C bond were considered. Dispersion-corrected DFT and MP2 methods with the SMD solvation model were employed using three all-electron bases as well as the ECP28MWB ansatz. The rate-determining, elimination step had a higher activation energy (12 kcal mol−1) in case of the “bromine end” attack due to intermediate stabilization at both the MP2 and DFT levels. In case of the “boron end” attack, two modes of C–C bond rotation were followed and striking differences in MP2 vs. DFT potential energy surfaces were observed. Employing MP2, addition was followed by either a 180° rotation through an eclipsed conformation of vicinal bromine atoms or by an opposite rotation avoiding that conformation, with 5 kcal mol−1 of elimination activation energy. Within B3LYP, the addition and rotation proceeded simultaneously, with a 9 (7) kcal mol−1 barrier for rotation involving (avoiding) eclipsed conformation of vicinal bromines. For weakly bound complexes, ZPE corrections with MP2 revealed significant artifacts when diffuse bases were included, which must be considered in the Gibbs free energy profile interpretation.
Active learning of potential-energy surfaces of weakly bound complexes with regression-tree ensembles
