Symmetry-breaking phenomena in metalloporphyrin pi-cation radicals

Abstract

Density functional theory (DFT) calculations of the energetics, molecular structures, and spin density profiles of metalloporphyrin pi-cation radicals suggest that the common practice of describing these radicals in terms of a universal A(1u)/A(2u) dichotomy is often not justified, confirming a possibility first foreseen by Prendergast and Spiro (ref 15) over a decade ago on the basis of vibrational spectroscopy and semiempirical calculations. Because of near-degeneracy of the a(1u) and a(2u) HOMOs of many metalloporphyrins, the cation radicals derived from these compounds undergo a pseudo-Jahn-Teller (pJT) distortion and are, therefore, best described as (2)A(u) with reference to the C(4h) point group, rather than as (2)A(1u) (D(4h) or (2)A(2u) (D(4h)). We find that the porphyrin cation radicals undergo a pJT distortion if the energy difference between the (2)A(1u) and (2)A(2u) pi-cation radicals, optimized under D(4h) symmetry constraints, is less than 0.15 eV. According to this criterion, metallo-porphine and metallo-OEP pi-cation radicals should always be pJT-distorted and metallo-meso-tetrahalogenoporphyrin radicals should not. For [Zn(TPP(*))](+), the (2)A(1u)/(2)A(2u) energy difference is almost exactly at the threshold of 0.15 eV, consistent with the experimental observation of both symmetry-broken and undistorted structures for this species. The (2)A(1u)/(2)A(2u) energy difference (when the molecular geometries are optimized under a D(4h) symmetry constraint) also appears to govern whether the real pJT-distorted cation radical is more A(1u)- or A(2u)-like in terms of its spin density profile. Because many metalloporphyrin pi-cation radicals exist as cofacial dimers in the crystalline phase, we examined the symmetries and structures of the model compounds [[Zn(P)](2)](+,2+) by means of DFT geometry optimizations. The results showed that dimerization has relatively little impact on the bond length alternation in the individual rings. A final interesting result, consistent with experiment, is that the bond length alternation in the delocalized mixed-valence dimer [[Zn(P)](2)](+) is about half that found for [[Zn(P)](2)](2+).