Oxidative Stretching of Metal−Metal Bonds to Their Limits

Oxidation of quadruply bonded Cr2(dpa)4, Mo2 (dpa)4, MoW(dpa)4, and W2(dpa)4 (dpa = 2,2′-dipyridylamido) with 2 equiv of silver(I) triflate or ferrocenium triflate results in the formation of the twoelectron-oxidized products [Cr2(dpa)4]2+ (1), [Mo2(dpa)4]2+ (2), [MoW- (dpa)4]2+ (3), and [W2(dpa)4]2+ (4). Additional two-electron oxidation and oxygen atom transfer by m-chloroperoxybenzoic acid results in the formation of the corresponding metal−oxo compounds [Mo2O(dpa)4]2+(5), [WMoO(dpa)4]2+ (6), and [W2O(dpa)4]2+ (7), which feature an unusual linear M***MO structure. Crystallographic studies of the twoelectron-oxidized products 2, 3, and 4, which have the appropriate number of orbitals and electrons to form metal−metal triple bonds, show bond distances much longer (by >0.5 Å) than those in established triply bonded compounds, but these compounds are nonetheless diamagnetic. In contrast, the Cr−Cr bond is completely severed in 1, and the resulting two isolated Cr3+ magnetic centers couple antiferromagnetically with J/kB= −108(3) K [−75(2) cm−1], as determined by modeling of the temperature dependence of the magnetic susceptibility. Density functional theory (DFT) and multiconfigurational methods (CASSCF/CASPT2) provide support for "stretched" and weak metal−metal triple bonds in 2, 3, and 4. The metal−metal distances in the metal−oxo compounds 5, 6, and 7 are elongated beyond the single-bond covalent radii of the metal atoms. DFT and CASSCF/CASPT2 calculations suggest that the metal atoms have minimal interaction; the electronic structure of these complexes is used to rationalize their multielectron redox reactivity