Mapping the d−d Excited-State Manifolds of Transition Metal β-Diiminato−Imido Complexes. Comparison of Density Functional Theory and CASPT2 Energetics

Abstract

Trigonal-planar, middle transition metal diiminato-imido complexes do not exhibit high-spin states, as might be naively expected on the basis of their low coordination numbers. Instead, the known Fe(III), Co(III), and Ni(III) complexes exhibit S = 3/2, S = 0, and S = 1/2 ground states, respectively. Kohn-Sham DFT calculations have provided a basic molecular orbital picture of these compounds as well as a qualitative rationale for the observed spin states. Reported herein are ab initio multiconfiguration second-order perturbation theory (CASPT2) calculations, which provide a relatively detailed picture of the d-d excited-state manifolds of these complexes. Thus, for a C(2v) Fe(III)(diiminato)(NPh) model complex, two near-degenerate states ((4)B(2) and (4)B(1)) compete as contenders for the ground state. Moreover, the high-spin sextet, two additional quartets and even a low-spin doublet all occur at <0.5 eV, relative to the ground state. For the Co(III) system, although CASPT2 reproduces an S = 0 ground state, as observed experimentally for a related complex, the calculations also predict two exceedingly low-energy triplet states; there are, however, no other particularly low-energy d-d excited states. In contrast to the Fe(III) and Co(III) cases, the Ni(III) complex has a clearly nondegenerate (2)B(2) ground state. The CASPT2 energetics provide benchmarks against which we can evaluate the performance of several common DFT methods. Although none of the functionals examined perform entirely satisfactorily, the B3LYP hybrid functional provides the best overall spin-state energetics.