Resonance and conjugation—II Factors determining bond lengths and heats of formation

Abstract

Listen

Translate article

Two previous papers presented interim reports of an investigation into the importance of resonance in conjugated systems. Further work has confirmed our earlier conclusions that resonance effects are unimportant in classical molecules, the bond lengths in them being determined by the state of hybridization of carbon; that the observed stabilization of such molecules is due to changes in bond energy with hybridization rather than to resonance; and that resonance stabilization plays a surprisingly small role even in benzene. A calculation is described which, starting from very limited experimental data, provides values for several CC bond lengths (ethylene, benzene, graphite) which agree excellently with experiment, a value for the sp 2- sp 2 CC bond length which agrees well with the value found in butadiene, diphenyl, etc., and a set of bond energies for CC and CH bonds of various types. Heats of formation for a number of hydrocarbons are calculated from these bond energies and compared with experiment. Agreement is excellent for paraffins, olefines, and acetylenes, confirming the lack of hyperconjugative stabilization in the latter. Deviations in cycloparaffins are ascribed to conformational effects (CHCH dipole repulsions). Butadiene shows an apparent resonance energy of 2 kcal/mole, less than a quarter of the stabilization energy; reasons are given for believing even this value to be much too large. The lack of resonance in butadiene is confirmed by a MO calculation of its effect on the bond lengths. When allowance is made for the effect of σ-bond compression, the resonance energy of benzene is found to be only about 10 kcal/mole, less than a quarter of the stabilization energy. Previous ideas concerning the importance of electron correlation in conjugated systems are clarified: it is shown that the neglect of such correlation invalidates existing methods of calculation, particularly in the case of classical molecules, and a method is suggested whereby such correlation effects might be taken into account. The significance of resonance theory is discussed. Various criticisms made by Mulliken in a recent paper are shown to be unjustified.