Hybrid density functional
theory (B3LYP) and density matrix renormalization
group (DMRG) theory have been used to quantitatively compare the degree
of ligand noninnocence (corrole radical character) in seven archetypal
metallocorroles. The seven complexes, in decreasing order of corrole
noninnocent character, are Mn[Cor]Cl > Fe[Cor]Cl > Fe[Cor](NO)
> Mo[Cor]Cl
2
> Ru[Cor](NO) ≈ Mn[Cor]Ph ≈
Fe[Cor]Ph ≈
0, where [Cor] refers to the unsubstituted corrolato ligand. DMRG-based
second-order perturbation theory calculations have also yielded detailed
excited-state energetics data on the compounds, shedding light on
periodic trends involving middle transition elements. Thus, whereas
the ground state of Fe[Cor](NO) (
S
= 0) is best described
as a locally
S
= 1/2 {FeNO}
7
unit antiferromagnetically
coupled to a corrole A′ radical, the calculations confirm that
Ru[Cor](NO) may be described as simply {RuNO}
6
–Cor
3–
, that is, having an innocent corrole macrocycle.
Furthermore, whereas the ferromagnetically coupled
S
= 1{FeNO}
7
–Cor
•2–
state
of Fe[Cor](NO) is only ∼17.5 kcal/mol higher than the
S
= 0 ground state, the analogous triplet state of Ru[Cor](NO)
is higher by a far larger margin (37.4 kcal/mol) relative to the ground
state. In the same vein, Mo[Cor]Cl
2
exhibits an adiabatic
doublet-quartet gap of 36.1 kcal/mol. The large energy gaps associated
with metal–ligand spin coupling in Ru[Cor](NO) and Mo[Cor]Cl
2
reflect the much greater covalent character of 4d−π
interactions relative to analogous interactions involving 3d orbitals.
As far as excited-state energetics is concerned, DMRG-CASPT2 calculations
provide moderate validation for hybrid density functional theory (B3LYP)
for qualitative purposes, but underscore the possibility of large
errors (>10 kcal/mol) in interstate energy differences.