Molecular structure and ionization energies of azides: an ab initio study of hydrazoic acid, methyl azide and ethyl azide

Abstract

Ab initio calculations for hydrazoic acid HN 3, methyl azide CN 3 and ethyl azide CH 3CH 2N 3, and for the corresponding singly ionized structures HN + 3, CH 3N + 3 and CH 3CH 2N 3 + are reported. Complete geometry optimizations at several theoretical levels have been carried out. A detailed analysis of the dependence of the results on the basis set level and a consideration of electronic correlation effects is presented. Concerning the structure, dipole moments, ionization energies and conformational equilibrium, very good agreement with experiment is observed. It has been verified that the observed deviation from linearity of the azide chain is enhanced when correlation effects are included in the geometry optimization procedures. This deviation is slightly more important for the ionized structures. First ionization energies based on Koopmans' theorem are in excellent agreement with experimental results. However, Δ E calculations of these energies depend on the consideration of electronic correlation effects. Rotational barriers for methyl azide and ethyl azide are very sensitive to the basis set quality and to the consideration of electron correlation. For methyl azide a very good agreement between theory (2.9 kJ mol −1) and experiment (2.98 kJ mol −1) is observed for calculations at the MP2//6-311G ∗* level. For ethyl azide, our results suggest that a correct theoretical evaluation of the rotational barrier between the anti and gauche conformers cannot be carried out at the Hartree-Fock (HF) level. Thus, the present HF values for this barrier are in the 0.8–1.4 kJ mol −1 range. Introduction of correlation effects at the MP2/6-31G level yields -1.2 kJ mol −1. Further calculations using the previous geometry at the MP4/6-311G ∗* level gives -1.35 kJ mol −1 and comparison with the experimental value of −0.6 kJ mol −1 (calculated for ethyl azide solvated in a nitrogen matrix) suggests that the difference could be related to solvent effects.