Computational study of B–B single-bond formation between a cage-localised carboranyl radical and DMAc-activated bis(catecholato)diboron: a concerted but asynchronous B–B bond-forming step with barrierless DMAc release

The boron-centred carboranyl radical generated in situ through electron donor–acceptor (EDA) photochemistry adds to the DMAc-activated bis(catecholato)diboron adduct B2cat2·DMAc to form an unsymmetrical boryl–carborane B–B single bond, a key step in a recently reported scalable route to boron-rich constructs for boron neutron capture therapy. Using UM062X/def2-SVP calculations with the SMD solvation model for N, N-dimethylacetamide, we show that this step proceeds over an ultra-flat potential-energy ridge on which direct synchronous-transit saddle-point searches (QST2/QST3) fail to converge. A constrained scan along the forming B–B coordinate brackets the barrier with an upper bound of + 37.4 kcal mol⁻¹ at R(B50–B21) ≈ 2.20 Å, whereas relaxation along the DMAc dissociation coordinate is monotonically downhill, indicating a barrierless release of the Lewis base. The reaction is concerted but asynchronous: B–B bond formation advances well ahead of DMAc dissociation, and the product, in which the cage radical has coupled to the coordinated boron, is stabilised by 13.4 kcal mol⁻¹ relative to the separated radical/adduct limit. Bonding analysis with QTAIM and Wiberg indices confirms a genuine covalent B–B bond (ρ = 0.117 au, ∇²ρ = −0.196 au, H(r) = − 0.063 au; Wiberg index 0.67) between the apical cage boron B50 and the adduct boron, while the DMAc oxygen retains only residual interaction with the product (Wiberg index below 0.05). The results rationalise the experimentally observed reactivity and establish the cage apical boron as the operative radical site.