Electron Spin Resonance of Azulene Anion Radicals

Abstract

The electron spin resonance spectra of the anion radicals of azulene, azulene‐1,3‐d₂, 4,6,8‐trimethylazulene, and 4,6,8‐trimethylazulene‐1,3‐d₂ have been examined. These hydrocarbons are nonalternant and have aromatic ring structures with C–C–C bond angles that depart from 120°. The radicals were generated electrolytically in N,N‐dimethylformamide using tetra‐n‐propylammonium perchlorate as supporting electrolyte. Sigma‐pi‐interaction calculations of the parameter Q_CH^H [in the relation a_i^H=Q_CH^Hρ_i^π between the isotropic proton hyperfine splitting (a_i^H) and pi‐electron spin density (ρ_i^π)] have been performed as a function of the bond angles at the carbon atom bonded to the proton. These calculations indicate a strong dependence of this parameter on angle. By using recent x‐ray data on azulene, predictions could be made of the Q's at the different positions in the radicals. In contrast, the proton splitting at a methyl‐group substituent, determined by the parameter QCCH₃^H, is predicted to be independent of the bond angles at the carbon atom of the aromatic system. Calculations of the pi‐electron spin density were made in both the simple Hückel and the McLachlan extended self‐consistent field molecular‐orbital theory approximations. These calculations were used with the experimental splitting constants to make a comparison with the predicted bond‐angle variation of the sigma‐pi interaction parameters (Q's). The McLachlan procedure gives much more satisfactory results than the Hückel calculations, but the agreement with experiment, involving use of the predicted dependence of Q_CH^H on bond angle, is only approximate. The theoretical estimates of the variation of Q_CH^H with bond angle, and of the constancy of QCCH₃^H, appear to be qualitatively correct. Applications of the theory for the change of Q_CH^H with angle may in some instances necessitate taking into account bent bonds.