The free energies of reaction calculated in Gaussian (1 atm standard state concentration) were converted to 1 M standard state values in order to calculate the equilibrium data in Figure 3. The concentration at 1 atm, [1 atm], at each temperature is given by PV/RT, from which the entropy correction, S corr , is given by -Rln(1/[1 atm]). Applying this correction to each of the reactants leads to the corrected ∆G react values given in Table S1.
Supporting InformationRadkiewicz-Poutsma and co-workers 4 , raised the concern that B3LYP geometry optimization, as implemented in the CBS-QB3 scheme, may lead to a poor geometry for the borane-olefin π-complex, and as such the energetics from CBS-QB3 may be incorrect for the weak R 2 BH-olefin complexes studied herein. This could also have an influence on the free energy barrier for olefin dissociation in these cases. In addressing this, we first discuss the findings of Gilbert and Radkiewicz-Poutsma, and then evaluate the effect of different methods on our results.Gilbert 3 evaluated various DFT methods as well as MP2 for predicting both the geometry and dissociation energy of amineborane dative bonds. It was found that B3LYP predicts B-N bond dissociation energies (BDEs) that are lower than found experimentally, and also predicts B-N bond distances that are too long. Radkiewicz-Poutsma and co-workers 4 extended the study of amine-borane dative bonds. They found that the poor performance of B3LYP in calculating the B-N BDE is due to how this method calculates the energy of the dative bond, rather than being due to the geometry resulting from optimization with the B3LYP method. Bond energies resulting from high level single point calculations on B3LYP geometries were found to be reliable, and CCSD(T) calculations were relatively insensitive to whether an MP2 or B3LYP geometry was used, despite MP2 leading to a tighter B-N bond. This is consistent with the PES corresponding to the dative bond B-N distance being very flat. Both authors found that the MP2 method leads to more reliable B-N bond distances, and MP2 BDEs were closer to experimental or CCSD(T) values. We note that our work has not employed B3LYP energies (apart from thermal corrections to Gibbs free energy); in all cases higher level composite energies on B3LYP geometries have been obtained. We also note that while validating theoretical studies of propene hydroboration (more directly analogous to our work), Singleton 1 found B3LYP to be reliable for energy calculations.We have investigated the effect of MP2 geometry optimization by studying elimination and re-addition of propene from B i Pr 3 , as shown in Figure 5(a) of the main article. CBS-QB3 optimisation in this case led to the loosest post-elimination borane-olefin complex encountered in this work, presumably for steric reasons. The B-C α distance in the i Pr 2 BH••••propene complex is 4.5 Å in this case. The transition structures (elimination and re-addition), π-complex and separated products (propene and i Pr 2 BH) have been re-optimized at the MP2/6-31+G(...