Ab initio thermochemistry and kinetics for carbon-centered radical addition and beta-scission reactions

A quantitative comparison of ab initio calculated rate coefficients using five computational methods and five different approaches of treating hindered internal rotation and tunneling with experimental values of rate coefficients for nine carbon-centered radical additions/beta scissions at 300, 600, and 1000 K is performed. The high-accuracy compound methods, CBS-QB3 and G3B3, and the density functionals, MPW1PW91, BB1K, and BMK, have been evaluated using the following approaches: (i) the harmonic oscillator approximation; (ii) the hindered internal rotor approximation for the internal rotation about the forming/breaking bond in the transition state and product; and the hindered internal rotation approximation combined with (iii) Wigner, (iv) Skodje and Truhlar, and (v) Eckart zero-curvature tunneling corrections. The density functional theory (DFT) based values for beta-scission rate coefficients deviate significantly from the experimental ones at 300 K, and the DFT methods do not accurately predict the equilibrium coefficient. The hindered rotor approximation offers a significant improvement in the agreement with experimental rate coefficients as compared to the harmonic oscillator treatment, especially at higher temperatures. Tunneling correction factors are smaller than 1.40 at 300 K and 1.03 at 1000 K. For both the CBS-QB3 method, including the hindered rotor treatment but excluding tunneling corrections, and the G3B3 method, including hindered rotor and Eckart tunneling corrections, a mean factor of deviation with experimentally observed values of 3 is found.