Metalloenzymes can catalyze a range of oxidation reactions utilizing an ultimate "green" oxidant, dioxygen. The reactions are typically initiated by O2 binding at a reduced, electron-rich metal center, such as Cu(I) or Fe(II). The resulting dioxygen adducts often contain a coordinated superoxide.1 Subsequent reduction of O2 takes place along a well-defined reaction pathway and leads to formation of metal-peroxo and -oxo intermediates. More importantly, nature has developed various enzymes to employ all reduced forms of O2 to functionalize substrates.2
The Fe(II)- and α-ketoglutarate (αKG)-dependent dioxygenases serves as a representative example.3 The proposed mechanism (Scheme 1) involves (1) addition of dioxygen to the square pyramidal quaternary enzyme-Fe(II)-alpha-KG-substrate complex to yield an Fe(III)-superoxo intermediate (I), (2) attack of the uncoordinated O-atom in the O2-moiety on C2 of alpha-KG to form a bicyclic alkylperoxo species (II), (3) cleavage of the O-O bond and decarboxylation resulting in an Fe(IV)-oxo species (III),4 (4) abstraction of an H-atom from the substrate to yield an Fe(III)-hydroxide complex and a substrate radical, (5) hydroxylation via OH rebound, and (6) dissociation of the product. We have employed high-level electronic-structure theory and spectroscopic tools to understand the reactivity of high-valent iron species.5
Synthetic oxo- and nitrido-iron complexes feature various coordination geometries and distinct electronic structures, and therefore exhibit diverse reactivity (Figure 1).6 Synergy from both experimental and theoretical findings can help us to elucidate their bonding and delineate their mechanistic features toward a range of chemical processes, like hydrogen-atom abstraction (HAT),7 oxygen atom transfer (OAT) and electron transfer (ET). In collaboration with Eckhard Bill, we employ various spectroscopic techniques to understand their bonding and electronic structure. Specifically, electron paramagnetic resonance (EPR), Mössbauer and magnetic circular dichroism (MCD) are in the front line of our current research. The MCD spectroscopy, in particular, is an outstanding technique because it is able to provide information about the geometric and electronic properties of the ground state such as oxidation and spin states, spin Hamiltonian parameters and coordination geometry as well as those of excited states. Thus, MCD serves as an invaluable link between ground state spectroscopy, EPR and excited state spectroscopy, electronic absorption (ABS) and resonance Raman spectroscopy.
A wealth of mononuclear oxo-iron(IV) model complexes have been prepared and characterized in the past decades, for which complex 1 serves as a representative example (Scheme 2). However, synthetic analogs of high-valent diiron species are still quite rare. Recent experiments report that a complex with a [FeIV2(μ-O)2] diamond core structure could be generated from open-core diiron(IV) complex 2 (Scheme 2) upon treatment with one equivalent of a strong acid. We performed a detailed MCD study on complexes 1 and 2 in combination with multi-configurational CASSCF/NEVPT2 (complete active space self-consistent field/N-electron valence perturbation theory) calculations to correlate their electronic structure and reactivity.8
The MCD spectra were directly calculated using CASSCF/NEVPT2 method and independent determination of the MCD signs were made based on the associated electron donating orbital (EDO) and electron accepting orbital (EAO) for a given transition. In comparison with experiment, this approach allowed us to make unambiguous assignments of the important transitions of complex 1 and gain more insight into the MCD signs and the temperature-dependent intensity variations (Figure 2a and 2b). Based on MCD/ABS intensity ratios, calculated excitation energies, polarizations, and MCD signs, the key transitions of complex 2 are assigned as ligand-field- or oxo- or hydroxo-to-metal charge transfer transitions. The correlation of the electronic structures of complexes 1 and 2 with their reactivity toward C–H bond oxidation and O-atom transfer reveals that, despite a difference in nuclearity, the two ferryl sites actually have very similar electronic structures that led to similar reactivity (Figure 2c).
We extended our understanding of electronic structure of oxo-iron(IV) complexes to a recently synthesized tetracarbene oxo-iron(IV) species (3) (Figure 3) using a combined experimental and theoretical approach.9 We were able to unambiguously assign the important ligand field transitions through direct computations of MCD spectra with CASSCF/NEVPT2 based methods and independent determination of the MCD C-term signs. In contrast to the majority of triplet ferryl complexes supported by polydentate N-donor ligands (complex 1), complex 3 has been proven to feature a distinct electron configuration in which the dx2-y2 orbital lies higher in energy than dz2. Our detailed electronic-structure analysis by using MCD and infrared photo dissociation (IRPD) spectroscopy clearly show that the tetracarbene ligand does not considerably affect the bonding in the (FeO)2+ core, but strongly destabilizes the dx2-y2 orbital and lifts it above the dz2 orbital in energy. As a result, the HAT reaction with complex 3 is likely to exclusively take place on the triplet surface due to the large quintet-triplet energy gap.
We have been performing theoretical investigation on the C–H activation reactivity exhibited by complexes 1, 2 and 3. The subtle difference in the electronic structures between complexes 1 and 3 results in a decisive consequence to their reactivity. Classical ferryl model compounds, such as complex 1, typically follow a mechanistic scenario of two-state reactivity for C-H bond activation. They usually have a low-lying quintet state (ΔG <3 kcal/mol,) with an electron configuration of (dxy)1(π*-dxz/yz)2(σ*-dx2-y2)1(σ*-dz2)0, and the quintet σ-pathway typically involves a much lower barrier than that for the triplet π-channel, the system hence first undergoes a spin-crossover to the quintet state and the subsequent C-H bond cleavage process takes place predominantly on the quintet surface. Due to swapping of the d-orbital energies in the eg set, the lowest-energy quintet state of complex 3 features an electron configuration of (dxy)1(π*-dxz/yz)2(σ*-dz2)1(σ*-dx2-y2)0. As a consequence, a significantly greater quintet-triplet energy separation (16.6 kcal/mol) is found for complex 3. Thus, the entire reaction is likely to exclusively occur on the triplet surface, viz. single-state reactivity (Figure 4). Our calculated barrier for the rate-determining step of H-atom transfer (18.8 kcal/mol) matches the experimental kinetic data (ΔG‡ (20° C) = 15.2 kcal/mol) very well, providing further credence to our proposed electronic structure.