Assessment of theoretical procedures for hydrogen-atom abstraction by chlorine, and related reactions

We have examined a number of hydrogen-abstraction reactions and assessed various theoretical procedures with regard to their performance for geometry optimization and for calculating barriers and reaction energies. We find that the BH&H-LYP and M05-2X procedures with the 6-31+G(d,p) basis set provide reasonable predictions for the geometries of the transition structures and also yield reasonable imaginary frequencies when compared with our benchmark QCISD/6-31+G(d,p) and CCSD(T)/6-311+G(3df,2p) values. For the calculation of barriers and reaction energies, M05-2X appears to be the most accurate of the hybrid functionals. The double-hybrid functionals, B2K-PLYP, UB2-PLYP-09, ROB2-PLYP, and DSD-B-LYP-D3, when used in combination with an augmented triple-zeta basis set, give very good agreement with the benchmark URCCSD(T)/aug-cc-pVQZ energies. We find that for wavefunction procedures, use of CCSD(T) in combination with an augmented triple-zeta quality basis set is required for the accurate prediction of barriers and reaction energies for these reactions.