Ligand driven heterolytic O-O bond cleavage in a non-haem phenolato-Fe(III)-OOH complex to yield a formal Fe(V)O intermediate

Abstract

Fe(V)O species can be generated by the heterolytic cleavage of the O-O bond of corresponding Fe(III)-OOH species. In haem complexes the redox non-innocence of the ligand facilitates such heterolytic cleavage, however non-haem iron complexes generally show homolytic cleavage to form an Fe(IV)O species and a hydroxyl radical. The hydroxyl radical formed is undesirable due to its non-selective reactivity. Here we show that the redox non-innocence of a phenolato ligand moiety in the complex [LFe(III)(μ-O)Fe(III)L]²⁺, where L is 2-(((di(pyridin-2-yl)methyl)(pyridin-2-ylmethyl)amino)methyl)phenolate, facilitates heterolytic O-O bond cleavage, similar in manner to that observed with haem Fe(III)-OOH species, to yield a formal Fe(V)O intermediate. Although not observed directly, the intermediacy of an Fe(V)O species is manifested in the immediate appearance of a doubly oxidised bis-phenolato bridged complex observed by time resolved UV/vis absorption and resonance Raman spectroscopy. This complex is formed by C-C coupling at the para position of the phenolato moiety of the ligand. The pathways to form the final complex via various Fe(IV)O and Fe(V)O intermediates are investigated by DFT methods, which indicate that the impact of the phenolato moiety is due to its redox non-innocence primarily. The ability of the phenolato moiety to transfer charge and spin density induces a switch in the mechanism of O-O bond cleavage from homolytic to heterolytic manifested in the radical character at the para-position needed for C-C bond formation and the high oxidation state of the first observed product.

Fig. 1. Ligands for iron complexes discussed in the text.

Fig. 2. The reaction of 1 with acid (HA) with stoichiometric and excess H₂O₂.

Fig. 3. Structures of mononuclear Fe(iii) complexes discussed in the text and redox chemistry of the biphenolato bridged 3.

Fig. 4. UV/vis absorption spectrum of 1 (0.25 mM) in CH₃CN (dark blue) and with stepwise addition of 0.3 equiv. CF₃SO₂H up to 1.5 equiv. (red).

Fig. 5. Thermodynamic driving forces for formation of 2 from 1 and subsequent exchange of the conjugate base ligands or H₂O by H₂O₂ in kJ mol⁻¹ to form 2c from these species.

Fig. 6. UV/vis absorption spectrum of 1 (0.25 mM) (a) with CF₃SO₃H (0.75 equiv., black) and over the first 5 min following addition of H₂O₂ (1 equiv.), and (b) between 8.5 and 75 min after. (c) Initial spectrum (black) before addition of H₂O₂, and spectra at times where a maximum absorbance at 465 nm (red) and 844 nm (blue) is reached. (d) Absorbance at selected wavelengths over time.

Fig. 7. MCR analysis of UV/vis absorption spectra over time following addition of H₂O₂ (1 equiv.) to 1 (0.25 mM) with CF₃SO₃H (0.75 equiv.). MCR components and their contribution over time are shown (a/b) 0 to 400 s and (c/d) 500 to 4200 s.

Fig. 8. Raman spectra at (a) λ_ex 785 nm (black, 165 s, Fig. 9) and (b) λ_ex 473 nm (red, 2380 s, Fig. 10) of 1 with CF₃SO₃H (0.75 equiv.) after addition of H₂O₂ (1 equiv.) in CH₃CN, with spectra obtained upon electrochemical oxidation of poly-1 at (a) λ_ex 785 nm (blue) and (b) λ_ex 488 nm (purple).

Fig. 9. Raman spectrum at λ_exc 785 nm (a) before (black) and 2380 s (red) after addition of H₂O₂ (1 equiv.) to 1 (0.25 mM) with CF₃SO₃H (0.75 equiv.). Spectra are vertically offset and rescaled for clarity (see Fig. 6 for UV/vis absorption spectra). (b) From 0 (black) to 2380 s. Spectra were normalised with respect to the intensity of the Raman band of CH₃CN at 920 cm⁻¹, and baseline corrected.

Fig. 10. Raman spectrum at λ_exc 473 nm over time following addition of H₂O₂ (1 equiv.) to 1 (0.25 mM) with CF₃SO₃H (0.75 equiv.) (see Fig. 6 for UV/vis absorption spectra) (a) from 0 (black) to 570 s and (b) from 570 to 3290 s. Note that although the spectrum is dominated by Raman scattering from 3c (e.g., band at 1620 cm⁻¹), the characteristic band of 3b at 1608 cm⁻¹ is observed in the final spectrum (red). Spectra were normalised with respect to the intensity of the Raman band of CH₃CN at 920 cm⁻¹, and baseline corrected.

Fig. 11. Raman intensity over time following addition of H₂O₂ (1 equiv.) to 1 (0.25 mM) with 0.75 equiv. CF₃SO₃H in CH₃CN at 1620 cm⁻¹ (blue, λ_exc 473 nm) and 1608 cm⁻¹ (red λ_exc 785 nm). Intensities are corrected for inner filter effects.

Fig. 12. The experimentally established species present during the reaction of 1 and CF₃SO₃H with H₂O₂.

Fig. 13. (a) Possible pathways following exchange of H₂O with H₂O₂ and subsequent dissociation of the O–O bond in 2c and 2d. (b) Calculated energies for species formed by homolytic and heterolytic cleavage of the O–O bond of 2c. All energies are relative to the aqua complex (2a).

Fig. 14. (left) Calculated Gibbs free energies (kJ mol⁻¹) for the (a) homolytic, (c) homolytic (non-protonated H₂O₂) and (e) heterolytic O–O bond cleavage pathway and subsequent coupling pathways possible thereafter. All energies are relative to 2a. (right) Proposed pathway for (a) homolytic, (c) homolytic (non-protonated H₂O₂) and (e) heterolytic O–O bond cleavage and the subsequent coupling pathways possible thereafter. Energies indicated in (a), (c) and (d) are in kJ mol⁻¹. The second y-axis indicates energy in kcal mol⁻¹ for reference.

Fig. 15. The Mulliken charge densities summed for fragments (top) and spin densities (bottom, shown in blue transparent, isosurface 0.003) of the complexes generated upon O–O bond cleavage in the hydroperoxido intermediates. The MDC-m spin density on the para phenolate carbon (where C–C bond formation takes place) is indicated in the top right corner of each structure. The structures are shown schematically, in order from left to right: 4b, 4a and 5.