Performance of the semiempirical AM1, PM3, MNDO, and tight-binding methods in comparison with DFT method for the large fullerenes C_(116)–C_(120)

摘要

Are the semiempirical methods still reliable for large fullerenes? To answer this question, in this paper, the semiempirical AM1, PM3, MNDO, and tight-binding calculations were performed on the candidate isomers selected from the complete sets of isolated-pentagon-rule isomers of C_(116)–C_(120) based on the second generation reactive empirical bond order potential. To assess the reliability of these semiempirical methods and predict the ground-state structures more accurately, the density functional theory calculations at the B3LYP/6-31G(d)//B3LYP/3-21G level of theory were also carried out on the top 20 low-energy candidates from each semiempirical method. It was found that the tight-binding potential gives a good agreement for the prediction of the relative energies when compared with the B3LYP/6-31G* results, while AM1, PM3, and MNDO methods behave badly. Furthermore, we also investigated the structures of the B3LYP/6-31G* low-energy isomers for C_(116)–C_(120) to understand the relationship between the stability and some structural factors.