ELECTRON-PARAMAGNETIC RESONANCE AND THEORETICAL-STUDY OF DIBENZYLMERCURY AND DIPHENYLMERCURY RADICAL CATIONS

Radical cations of dibenzylmercury and its p-X derivatives (X = Me, MeS, MeO, Cl) and of diphenylmercury and its p-X derivatives (X = Cl, Et, Me, MeO) have been generated by gamma-radiolysis of dilute solutions of the parent compounds in CFCl3 matrices at 77 K. The EPR results for the dibenzylmercury series show all members to be sigma-u-radical cations with singly occupied molecular orbitals approximating to the form (I). The parallel and perpendicular \ + 1/2 > components of the Hg-199 spectrum are almost coincident, giving an intense low-field isotropic line: this shows a substructure typical of an anisotropic doublet coupling, and arises from a single solvent fluorine atom, but is unprecedentedly large for radical-cation superhyperfine couplings in freon matrices. These additional couplings were absent when the experiments were repeated using a CCl4 matrix which confirms the F-19 assignment. The diphenylmercury radical cation is also of the sigma-u-type, along with its dimethyl and diethyl derivatives, although the spin density on mercury is reduced to ca. 50% of that in the dibenzyl series. In contrast, the dimethoxy and dichloro derivatives are formed in pi-states with apparently negligible spin density on the mercury atom. Thus the sigma-u- and pi-levels must be very close in energy, so that an electronic crossover can be induced by appropriate substitution. MNDO calculations were also performed on these radicals. While there is an overall accord with the experimental results, there are important differences: for instance, F-sigma* complexes are not predicted to be stable 'gas-phase' energy minima. However, constraints placed on the geometry such that the R2Hg.+ radical cation must be within ca. 2-4 A of a CFCl3 molecule, as would be the case in the solid matrix, can be modelled, and under these conditions appreciable F-19 couplings are predicted.