The method of Pariser and Parr has been applied to furan, thiophene, and 1,4-dithiadiene with a minimum of additional approximations and assumptions. It is possible to reproduce the spectra, dipole moments, resonance energies and reactivities of these molecules with a small number of adjusted parameters, all of which have reasonable values. The 3d orbitals of sulfur need not be explicitly considered but they may alter the valence state ionization potential of sulfur. A good deal of the previously predicted resonance energy of 1,4-dithiadiene disappears in the more sophisticated calculation. The very strong preference of furan for ~-substitution is predicted and originates primarily in the inductive effect of the oxygen. The lesser specificity of thiophene is also predicted, in this case arising from the form of the wave functions.
A NUMBER of difficulties have impeded the development of an entirely satisfactory theory of conjugated heterocyclic molecules such as furan, thiophene, and 1,4-dithiadiene. These difficulties include uncertainty about appropriate overlap and resonance integrals, effective electronegativities, and the participation in bonding of orbitals which are not occupied in the separated atoms. 2
If one uses the Hiickel approximations 3 or something similar, it develops that there are more disposable parameters than useful data because, for molecules of this type, the Htickel approximations are not expected to reproduce either spectrographic properties or dipole moments reliably. The remaining data are not large in number; they are usually ratios of rate or equilibrium constants, or thermodynamic properties of reaction or activation. These typically leave a good deal of room for maneuver in their interpretation. Moderately good dipole moments can often be obtained by making the field self-consistent in some manner or other, 4 but in order to correlate electronic spectra at all well some configuration interaction as well as specific consideration of electronnuclear attraction and electron-electron repulsion appears to be necessary. 4c
Electronic spectra would seem to be the most abundant and reliable source of data on which to base molecular wave functions. Pariser and Parr 5 have advanced a method 1