Molecular geometry as a source of chemical information for π-electron compounds

Abstract

After a short introduction presenting a general view on the title problem and after information about more general approaches such as e.g. the Bürgi and Dunitz Principle of structural correlation a series of empirical models applying the molecular geometry are shown as methods allowing us to read the molecular geometry in the language of chemistry. First, a model enabling estimation of the heat of formation from CC bond lengths for hydrocarbons is presented. The ring energy content calculated by this model shows that it may vary considerably depending on different reasons: the topological embedding in benzenoid hydrocarbons, topology of substitution and the nature of substituents in polysubstituted benzene derivatives and intermolecular interactions in the crystalline state. Secondly, the model enabling estimation of the canonical structure weights is presented (HOSE model). Based on this model it is shown how geometry of the benzene ring may reflect the substituent effect in terms of varying weights of canonical structure. The non-mesomeric interaction of the nitro group with the ring in nitrobenzene is shown and carefully documented. A new substituent effect due to angular groups (AGIBA-effect) is best illustrated by use of the canonical structure weights calculated from molecular geometry. Next we present the application of the Bent-Walsh rule which shows how intramolecular and intermolecular interactions associated with the charge transfer cause deviations from this rule. Finally, application of the molecular geometry to determine indices of aromaticity is presented. The most recent and interesting result in this field is that the index HOMA based solely on the experimental (or computed) bond lengths may be divided into two independent terms, of which one accounts for the energetic and the other for the geometric contributions to aromaticity. The first one is related to resonance energy of the ring in question, the second one-to the bond length alternation. A few illustrative applications are presented.