A comparison of high-level theoretical methods to predict the heats of formation of azo compounds

Abstract

The heats of formation (HOFs) of 57 azo compounds were calculated using six high-level theoretical models including variants of Gaussian- n (G n: G2, G2(MP2), and G3) and complete basis set (CBS- n: CBS-4M, CBS-Q, and CBS-QB3). For comparison, Hartree–Fock (HF) and density function theory (DFT) methods were also employed. The results show that the HF and DFT methods are not able to accurately predict the HOFs for the azo compounds. The G2 model presents the most accurate HOFs among the six high-level models, followed by the G2(MP2) and CBS-QB3. The CBS-Q model consistently gives more accurate HOFs than the G2 or G2(MP2) for the compounds with –NO 2. However, the case is quite the contrary for the compounds with –NH 2, –NH– or –CN. The CBS-Q or G2(MP2) model is a compromise between computation efficiency and accuracy for predicting the HOFs of the tile compounds.