An (RO)B3LYP/LANL2DZdp//B3LYP/LANL2DZ model for the prediction of the homolytic bond dissociation enthalpy (BDE) and adiabatic ionisation potential (IP) of phenolic antioxidants containing heavy chalcogens has been developed. The model has been used to probe the relationship between geometry, chalcogen substitution and activity for a series of alpha-tocopherol analogues of varying ring size. From this, a series of design principles for cyclic antioxidants has emerged, embodied by the compound 4-hydroxy-2,2,3,5,6-pentamethylbenzoselenete (4c). This compound is predicted to have a BDE comparable to alpha-tocopherol, and should act in a dual chain-breaking and hydroperoxide-decomposing manner, by analogy with other selenide antioxidants. The stability of chalcogen-substituted benzoxetes was considered, and the as yet unsynthesised benzotelluretes are predicted to be stable. Finally, an attempt was made to determine antioxidant mechanism by considering calculated BDE and IP data together with experimental rate data.