Solvent Effects in Electron Spin Resonance Spectra

Abstract

It has been shown that a simple model can account for the available data on the solvent dependence of the hyperfine splittings in the ESR spectra of organic free radicals. We have assumed that the changes in splittings arise entirely from a redistribution of the pi-electron spin density, and that the spin density is affected only by localized complexes between the solvent and polar substituents or heteroatoms in the radical. This model predicts that the magnitudes of the changes in proton splittings should often be small, although large fractional changes at positions of small spin density can sometimes occur. The large variations found for the nuclei of many electron atoms are shown to arise because their splittings are very critical functions of the spin density. The effect of the solvent on the proton hyperfine splittings in the semiquinones has been treated by assuming that the solvent alters the electronegativity of the oxygen atoms and by performing molecular-orbital calculations to estimate spin density changes. These calculations reproduce both the directions and magnitudes of the changes in proton splittings with solvent and are in agreement with the very large fractional changes observed at some positions with small splittings. An analysis of the effects on the ESR spectra of the exchange reactions between the different solvent complexes has shown that the spectra normally observed result from systems undergoing rapid exchange. A simple model for a radical with only one functional group which can interact with the solvent (e.g., the nitrobenzene anion), is shown to account very well for the observed variations in hyperfine splittings as a function of the composition of a binary solvent mixture. The treatment of the exchange reactions by use of the modified Bloch equations is compared with the spectral density method. The contribution to the linewidth from solvent-induced fluctuations in the spin-density distribution is calculated for a simple two-jump case, and other factors affecting the linewidths of radicals subjected to random solvent interactions are discussed.