Hypercoordinate Bonding to First-row Atoms

Examples of 1,3-dipoles include diazoalkanes, nitrones, carbonyl ylides and fulminic acid. Organic chemists typically describe 1,3-dipolar cycloaddition reactions [15] in terms of four out-of-plane 71 electrons from the dipole and two from the dipolarophile. Consequently, most of the interest in the electronic structure of 1,3-dipoles has been concentrated on the distribution of the four Jt electrons over the three heavy atom centres. Of course, a characteristic feature of this class of molecules is that it presents awkward problems for classical valence theories a conventional fashion of representing such systems invokes resonance between a number of zwitterionic and diradical structures [16-19]. Much has been written on the amount of diradical character, with widely differing estimates of the relative weights of the different bonding schemes. 

The mode of bonding which emerges from spin-coupled descriptions on various of 1,3-dipoles and related molecules [22-24] turns out to be closely analogous to the one that we have since established for second-row atoms. Calculations have 

Spin-coupled descriptions analogous to that for CH2N2 have been found to apply to all of the 1,3-dipoles that we have studied, as well as to related systems such as NNO. It is indeed tempting to represent the latter with the structural formula N=N=0, but it must be clearly understood that the singly occupied orbitals involved in a given bond are not strictly localised on individual atoms. 

We initially believed [23] that a similar mode of description carries over to 03, but we now realise that this system is somewhat more complicated than we had first supposed. Calculations for the four out-of-plane n electron, whether using spin-coupled theory or at the CASSCF level, reveal that there exist two solutions that are remarkably close in energy. The one that lies lowest, provided we optimize properly the description of the inactive electrons, corresponds to a singlet diradical whereas the other corresponds to a hypercoordinate central atom [24]. It is clear that neither description carries much conviction on its own, and that we must consider expanding the active space. 

---