COMPARISON OF EXPERIMENTAL AND CALCULATED GEOMETRIES OF SESQUIBICYCLIC HYDRAZINES

X-ray crystal structures are reported for bis(N,N'-bicyclic) hydrazines 2-4 and 7-9. In contrast to 1-8, which are cis fused at the nitrogens, 9 is trans fused. Molecular mechanics (MM2), semiempirical MO (AM1), and ab initio (6-31G*, on 1-3,5, and 6, and MP2/6-3 1G* on 6) calculations are compared with the experimental geometries. MM2 calculations seriously underestimate the NN bond length but do an excellent job of predicting pyramidality at nitrogen and correctly get the trans fused geometry for 9 as the more stable one. AM1 calculations also underestimate the NN bond length and do not estimate enough torsion at CNNC ring bonds, causing cis fused 9 to be calculated as more stable than the transform. The 6-31G* calculations make about half as large an error in the NN bond lengths as do MM2 and AM1 but are significantly poorer at obtaining the proper pyramidality at nitrogen. The MP2/6-31G* calculation on 6 predicts the X-ray NN bond length and nitrogen pyramidality a(av) to within experimental error. Errors in calculated structures and relative energies of conformations for less constrained hydrazines are discussed.