Copolymers of carbon dioxide and nitrogen: an AM1 and ab initio computational study

Extending work by various groups on possible dimers, trimers, etc. of dinitrogen and of carbon dioxide, the authors have studied analogous copolymers of N₂ and CO₂ computationally. Twelve cyclic structures were examined with the AM1, HF/3-21G, HF/6-31G* and MP2(FC)/6-31G* methods, and the acyclic “monomer” to “tetramer” HO(C(O)O–N= N–)_nH, n=1–4, were studied at the AM1 and HF/3-21G levels; the cyclic species included 2-oxa-3,4-diazacyclobut-3-ene-1-one, 2-oxa-3,4,5,6-tetraazacyclohexa-3,5-diene-1-one, and various aza/oxa bicyclo2.2.0] and bicyclo2.2.2] systems. For the cyclic species, it was concluded that only the MP2(FC)/6-31G* results, which differ considerably from those at the other three levels, are likely to be reliable. These MP2 calculations indicate that only seven of the 12 cyclic structures studied are stationary points (one is a transition structure), and none of them is kinetically stable at room temperature. Although some have high energy densities (ca. 7–10 kJ g^?1), their expected low kinetic stabilities seems to make this of little practical value. The acyclic “copolymers” were all relative minima at the AM1 and HF/3-21G levels; unlike the cyclic species, their kinetic stabilities were not investigated directly by comparing the energies of reactants and decomposition transition states. The energy density of the infinite acyclic polymer was found by extrapolation to be 5.1 (AM1) or 5.6 (3-21G) kJ g^?1. The calculated vibrational spectra of the MP2 stationary points and of the acyclic molecules gave some indication of instability by the presence of low-frequency modes leading in the limit to decomposition.