Noninnocence of the ligand glyoxal-bis(2-mercaptoanil). The electronic structures of [Fe(gma)]2, [Fe(gma)(py)] x py, [Fe(gma)(CN)]1-/0, [Fe(gma)I], and [Fe(gma)(PR3)(n)] (n = 1, 2). Experimental and theoretical evidence for "excited state" coordination.

The electronic structure of the known iron complexes Fe(gma) (S(t) = 0) (1)(6) and [Fe(gma)(py)].py (S(t) = 1) (2)(7) where H(2)(gma) represents glyoxal-bis(2-mercaptoanil) has been shown by X-ray crystallography, Mössbauer spectroscopy, and density functional theory calculations to be best described as ferric (S(Fe) = 3/2) complexes containing a coordinated open-shell pi radical trianion (gma())(3)(-) and not as previously reported(6,7) as ferrous species with a coordinated closed-shell dianion (gma)(2)(-). Compound 1 (or 2) can be oxidized by I(2) yielding [Fe(III)(gma)I] (S(t) = 1/2) (3). With cyanide anions, complex 1 forms the adduct [(n-Bu)(4)N][Fe(III)(gma())(CN)] (S(t) = 1) (4), which can be one-electron oxidized with iodine yielding the neutral species [Fe(III)(gma)(CN)] (S(t) = 1/2) (5). With phosphines complex 1 also forms adducts(7) of which [Fe(III)(gma())(P(n-propyl)(3))] (S(t) = 1) (6) has been isolated and characterized by X-ray crystallography. [Fe(II)(gma)(P(n-propyl)(3))(2)] (S(t) = 0) (7) represents the only genuine ferrous species of the series. Density functional theory (DFT) calculations at the BP86 and B3LYP levels were applied to calculate the structural as well as the EPR and Mössbauer spectroscopic parameters of the title compounds as well as of the known complexes Zn(gma)(-) and Ni(gma)(-). Overall, the calculations give excellent agreement with the available spectroscopic information, thus lending support to the following electronic structure descriptions: The gma ligand features an unusually low lying LUMO, which readily accepts an electron to give (gma())(3)(-). The one-electron reduction of [Zn(gma)] and [Ni(gma)] is strictly ligand centered and differences in the physical properties of Zn(gma(*)) and Ni(gma(*)) are readily accounted for in terms of a model that features enhanced back-bonding from the metal to the gma LUMO in the case of Ni(gma(*)). In the case of [Fe(gma)(PH(3))], [Fe(gma)(py)], and Fe(gma)(CN) an electron transfer from the iron to the gma LUMO takes place to give strong antiferromagnetic coupling between an intermediate spin Fe(III) (S(Fe) = 3/2) and (gma(*))(3)(-) (S(gma) = 1/2), yielding a total spin S(t) = 1. Broken symmetry DFT calculations take properly account of this experimentally calibrated electronic structure description. By contrast, the complexes [Fe(gma)(PH(3))(2)] and [Fe(PhBMA)] feature closed-shell ligands with a low-spin Fe(II) (S(Fe) = S(t) = 0) and an intermediate spin central Fe(II) (S(Fe) = S(t) = 1), respectively. The most interesting case is provided by the one-electron oxidized species Fe(gma)(py), [Fe(gma)I], and [Fe(gma)(CN)]. Here the combination of theory and experiment suggests the coupling of an intermediate spin Fe(III) (S(Fe) = 3/2) to the dianionic ligand (gma)(2)(-) formally in its first excited triplet state (S(gma) = 1) to give a resulting S(t) = 1/2. All physical properties are in accord with this interpretation. It is suggested that this unique "excited state" coordination is energetically driven by the strong antiferromagnetic exchange interaction between the metal and the ligand, which cannot occur for the closed-shell form of the ligand.