Chain-breaking reactions against lipid peroxidation performed by carotenes, including beta-carotene (beta-CAR) and lycopene (LYC), have been studied using density functional theory. We chose linoleic acid (LAH) as the lipid model and examined two mechanisms: hydrogen abstraction and addition. Our computed reaction diagrams reveal that the addition mechanism is able to offer a larger extent of chain-breaking protection than hydrogen abstraction. In the case of hydrogen abstraction, the resulting carotene radical CAR(-H)(*) has a smaller O(2) affinity than the linoleic acid radical (LA(*)). Formation of the addition adduct radical ROO-CAR(*) is energetically favorable, and it has an even smaller tendency to react with O(2) than CAR(-H)(*). Comparatively, ROO-beta-CAR(*) is less likely to react with O(2) than ROO-LYC(*). Both the hydrogen abstraction and addition radicals (CAR(-H)(*) and ROO-CAR(*)) react readily with a second ROO(*) radical via either hydrogen abstraction or addition.

A computational approach has been designed for accurately determining enthalpies of formation (ΔH(f)) for the carotene species. This approach, named correlation corrected atomization (CCAZ), is based on the concept of bond and group additivity, and is applied along with density functional theory (DFT). Corrections to the deficiencies in DFT were divided into 1,2-, 1,3-, and 1,4- atomic interactions, which were determined by comparisons with the G3 data of the training set. When comparing predictions from CCAZ combined with two different DFT methods (B3LYP and MPWB1K), fairly accurate prediction is expected. In contrast, DFT using the atomization and isodesmic schemes resulted in poor predictions of ΔH(f). The equivalent methods, atomic equivalent (AEQ) and group equivalent (GEQ) provide improved predictions; however, the accuracies are lower than that of CCAZ.