A Terminal Fe^III-Oxo in a Tetranuclear Cluster: Effects of Distal Metal Centers on Structure and Reactivity

Abstract

Tetranuclear Fe clusters have been synthesized bearing a terminal Fe^III-oxo center stabilized by hydrogen bonding interactions from pendant tert-butyl amino pyrazolate ligands. This motif was supported in multiple Fe oxidation states, ranging from [Fe^II₂Fe^III₂] to [Fe^III₄]; two oxidation states were structurally characterized by single crystal X-ray diffraction. The reactivity of the Fe^III-oxo center in proton coupled electron transfer (PCET) with X–H (X = C, O) bonds of various strengths was studied in conjunction with analysis of thermodynamic square schemes of the cluster oxidation states. These results demonstrate the important role adjacent metal centers have on modulating the reactivity of a terminal metal-oxo.

Graphical Abstract

Authors are required to submit a graphic entry for the Table of Contents (TOC) that, in conjunction with the manuscript title, should give the reader a representative idea of one of the following: A key structure, reaction, equation, concept, or theorem, etc., that is discussed in the manuscript. Consult the journal’s Instructions for Authors for TOC graphic specifications.

Terminal metal-oxo moieties are invoked as key intermediates in both natural and synthetic catalysts of mid-first-row transition metal ions (Mn, Fe, and Co).¹ For example in photosynthesis, water is oxidized in photosystem II by a CaMn₄O₅ cluster known as the oxygen evolving complex (OEC);² numerous computational studies of the catalytic mechanism have proposed a high-valent Mn-oxo playing a key role in O–O bond formation.³ Similarly, a number of synthetic water oxidation catalysts employing various multinuclear scaffolds have been reported, where a terminal metal-oxo is implicated as a key intermediate (Figure 1).^{1e–g, 4}

Figure 1.

Multinuclear catalysts with proposed terminal metal-oxo intermediates (top), and structurally characterized terminal Fe^III-oxo complexes (bottom)

Studies of synthetic transition metal-oxo complexes have been integral for understanding these reactive moieties in catalytic systems.^{1a, 5} However, there is a paucity of literature concerning multinuclear complexes bearing well-characterized terminal metal-oxo motifs.⁶ In a rare example where the effects of a neighboring metal oxidation state on a terminal metal-oxo could be interrogated, Que and coworkers reported that the spin state of an Fe^IV-oxo center would change depending on the oxidation state of a neighboring Fe in a μ₂-O bridged bimetallic complex (L’₂OFe₂(OH)(O)^2+/3+).^6c The authors demonstrated that structural and spin-state changes due to reduction of this secondary Fe leads to a thousand-fold activation of the [Fe₂] complex towards C–H oxidation.

To gain further insights into these multimetallic effects, our group has examined well-defined tetranuclear clusters of Fe and Mn, which facilitate intramolecular oxygen atom transfer reactions; however, a terminal metal-oxo intermediate could not be observed.⁷ Inspired by reports of mononuclear terminal metal-oxo motifs stabilized by second coordination sphere hydrogen bonding interactions,⁸ our group has previously used this strategy to access a terminal Mn^III–OH moiety as part of a [Mn₄] cluster.⁹ Herein, we describe the synthesis, structural characterization, and reactivity studies of clusters bearing a terminal Fe^III-oxo motif, stabilized by tert-butyl-amino-pyrazolates, to probe the significance of a multinuclear scaffold on structural and reactivity aspects of a terminal metal-oxo.

Treating the reported LFe₃(OAc)(OTf)₂ cluster (⁻OTf, triflate = trifluoromethane sulfonate)¹⁰ with three equivalents of potassium tert-butyl-amino-pyrazolate (KPzNHtBu) and iodosylbenzene (PhIO), followed by addition of iron (II) triflate bis-acetonitrile (Fe(OTf)₂ • 2 MeCN) and excess potassium hydroxide in tetrahydrofuran produces the neutral [Fe^II₃Fe^III] cluster, 1 (Scheme 1). Single crystal X-ray diffraction (XRD) studies of 1 reveal a structure similar to our previously reported [Mn₄] cluster bearing a terminal hydroxide ligand (Figure 2A);⁹ the apical metal displays a trigonal bipyramidal geometry, with the terminal hydroxide ligand hydrogen bonded to each amino-pyrazolate (N–O distances of 2.826(1), 2.765(1), 2.789(1) Å for 1). The relatively short distance between the apical Fe and the interstitial μ₄-O (Fe4–O1), 1.837(1) Å, is consistent with an Fe^III in the apical position of the cluster, with the remaining Fe centers being Fe^II.^{7b, 11}

Scheme 1.

Synthesis of [Fe₄] clusters. (Inset) 1,3,5-triarylbenzene ligand platform (L^3-) and tert-butyl amino pyrazolate ligand (PzNHtBu⁻).

Figure 2.

Crystal structures of tetranuclear Fe hydroxide cluster, 1 (A), and oxo cluster, 5 (B). Ellipsoids shown at the 50% probability level with solvent molecules, and hydrogen atoms (except for N–H moieties) omitted for clarity. (C) Mӧssbauer spectrum of 2 (black dots) with simulated parameters: (i) δ = 1.12 mm/s, |ΔE_q| = 3.20 mm/s (solid blue), (ii) δ = 1.10 mm/s, |ΔE_q| = 2.76 mm/s (dashed blue), (iii) δ = 0.52 mm/s, |ΔE_q| = 0.81 mm/s (orange), (iv) δ = 0.41 mm/s, |ΔE_q| = 2.17 mm/s (green). (D) Mӧssbauer spectrum of 5 (black dots) with simulated parameters: (i) δ = 1.12 mm/s, |ΔE_q| = 3.14 mm/s (solid blue), (ii) δ = 1.10 mm/s, |ΔE_q| = 2.87 mm/s (dashed blue), (iii) δ = 0.52 mm/s, |ΔE_q| = 1.13 mm/s (orange), (iv) δ = 0.43 mm/s, |ΔE_q| = 3.04 mm/s (green).

The electrochemistry of the [Fe₄] hydroxide clusters in THF features three quasi-reversible events assigned to the [Fe^II₃Fe^III]→[Fe^II₂Fe^III₂] (−1.53 V; all potentials vs. Fc/Fc⁺), [Fe^II₂Fe^III₂]→[Fe^IIFe^III₃] (−0.68 V), and [Fe^IIFe^III₃]→[Fe^III₄] (−0.10 V) redox couples (Figure S36). Each of the corresponding oxidation states of the cluster could be isolated (Scheme 1). Mössbauer spectra of the oxidized clusters 2, 3, and 4 are consistent with oxidations occurring at the Fe^II centers in the tri-iron core and the Fe–OH moiety remaining Fe^III (Figures 2C, S42, S46, and S47).

Access to a terminal Fe^III-oxo moiety was achieved by deprotonation of the [Fe^II₂Fe^III₂] hydroxide cluster, 2, with potassium tert-butoxide (KOtBu; Scheme 1). The resulting compound, 5, was crystallographically characterized (Figure 2B); deprotonation of the hydroxide ligand leads to structural changes to the apical Fe in 5. The Fe4–O2 distance contracts to 1.817(2) Å, compared to the distances in 1 (1.937(1) Å) and the precursor 2 (1.907(3) Å); this bond length matches closely with the structurally characterized Fe^III-oxo complexes reported by Borovik and Fout.^{8e, 8h, 8i} Compound 6, prepared by deprotonating 3, also displays a short Fe4–O2 distance (1.795(8) Å). Furthermore, the apical Fe-μ₄-O distance (Fe4–O1) elongates to 1.965(2) Å in 5 and 2.049(7) Å in 6, from 1.890(3) Å in 2 and 1.948(2) Å in 3, which is consistent with a greater trans influence exerted by the terminal oxo ligand. The Mӧssbauer spectra of 5 and 6 are consistent with the [Fe^III₂Fe^II₂] and [Fe^III₃Fe^II] oxidation state assignments, respectively (Figure 2D and S54). The quadrupole doublet assigned to the apical Fe^III-oxo centers in 5 and 6 have parameters distinct from the other previously reported data for [(H₃beau)Fe(O)]^2-, and most other terminal Fe-oxo complexes (Table 1).^{8e, 12} Further spectroscopic studies of these Fe^III-oxo clusters are underway to understand the source of their atypical Mӧssbauer parameters.

Table 1.

Selected Bond Distances and Angles, Structural Index Parameter, and Mӧssbauer Parameters of Reported Fe^III-Oxo Complexes

	5	6	[(H3beau)Fe(O)]2– 8e	[N(afaCy)3Fe(O)]+ 8h
Fe–O (Å)	1.817(2)	1.795(8)	1.813(3)	1.806(1)
Fe–Nequatorial (Å)	2.104(2), 2.098(2), 2.093(2)	2.100(8), 2.085(9), 2.087(9)	2.030(4), 2.060(4), 2.082(4)	2.049(1), 2.049(1), 2.052(1)
Fe–Ltrans (Å)	1.965(2) (L=O2-)	2.049(7) (L=O2-)	2.271(4) (L=NR3)	2.276(1) (L=NR3)
N–O (H-bond; Å)	2.647, 2.717, 2.685	2.718, 2.790, 2.750	2.732, 2.702, 2.686	2.641, 2.645, 2.673
∠Nequatorial–Fe–O (°)	96.3, 92.8, 92.0	93.6, 97.5, 96.3	103.3, 99.7, 100.8	102.6, 103.1, 103.1
Fe–N|N’|N”equatorial (Å)	0.14	0.22	0.42	0.45
Structural Index Parameter (τ)a	0.9	0.8	0.5	0.4
Mossbauer parameters (mm/s)	δ = 0.43, |ΔEq| = 3.04	δ = 0.47, |ΔEq| = 2.53	δ = 0.30, |ΔEq| = 0.91	-

^aτ = [Σ (∠N_equit.–Fe–N’_equit.) - Σ (∠N_equit.–Fe–O)]/90

Terminal Fe^III-oxo complexes are rare, and typically stabilized through hydrogen bonding interactions.^{8e, 8h, 8i, 13} The structures of 5 and 6 display comparable hydrogen bonding distances to other structurally characterized Fe^III-oxo complexes, [(H₃beau)Fe(O)]^2- and [N(afa^Cy)₃Fe(O)]⁺, along with similar equatorial Fe–N distances (Table 1). However, the μ₄-O distances in 5 (1.965(2) Å) and 6 (2.049(7) Å) are significantly shorter than the Fe–N distances for the amine trans to the oxo in the mononuclear systems (~2.27 Å). This is likely a result of greater ligand flexibility in the mononuclear systems; the geometry of these Fe^III-oxo complexes display greater deviations from ideal trigonal bipyramidal geometry compared to the apical Fe in 5 and 6, based on a structural index parameter (τ; ideal trigonal bipyramidal geometry = 1.0). For the clusters reported here, the rigid geometry of the pyrazolate ligands prevents significant distortion of the apical Fe out of the equatorial plane.

The hydroxide ligand in 2 was determined to be very basic in THF (pK_a = 30.1; Table S1). Analogous equilibrium studies were performed on 3 and, as expected, oxidation of the cluster reduces the basicity of the Fe^III-oxo moiety (pK_a = 23.0 for 3; Table S2). Attempts to deprotonate 4 with various bases, even at low temperatures, only resulted in decomposition, so a pK_a value for this oxidation state was not measured. These data were combined with electrochemical information for clusters 1 (vide supra) and 5 (Figure S38), to produce thermodynamic square schemes according to equation 1 (Figure 3):¹⁴

$${\text{BDE}}_{\text{O}-\text{H}}=23.06E°+1.37\text{p}{K}_a+C$$ (1)

Similar to our previously reported studies on [Fe₃Mn] hydroxide and aquo clusters, the bond dissociation enthalpy of the O–H bond (BDE_O–H) increases upon oxidation of the distal Fe centers, ranging from 72 kcal/mol in 1 to 84 kcal/mol in 3.¹⁵

Figure 3.

Thermodynamic cycles to evaluate the BDE_O–H values of the hydroxide clusters 1 – 3. Reduction potentials (horizontal lines) are references to Fc/Fc⁺. pK_a values (vertical lines) are based on relative pK_a values of cationic acids in THF. Diagonal lines are the BDE_O–H values calculated from these parameters according to the Bordwell equation (eq 1). Approximate values (~) have been extrapolated from the Bordwell equation.

The three distal Fe oxidation states have a dramatic effect on the reactivity of the Fe^III-oxo center through modifying the pK_a and BDE_O–H values. For example, 5 is incapable of performing proton coupled electron transfer (PCET) reactions^16,17 with substituted phenols over a range of phenol BDE_O–H values (79 – 85 kcal/mol); only proton transfer to generate 2 is observed as expected from the combination of low BDE_O–H for 1 and high pK_a of 2 (Figure 3, Table 2 and Figure S13). Oxidation of the remote Fe centers in 6 and 7 enables PCET reactivity with these phenols (Figures S14 and S16), resulting in the formation of 2 and 3, respectively.

Table 2.

Reactivity of the [Fe₄]-Oxo Clusters, 5 – 7.

	BDE (kcal/mol)	Reactivity Observeda
		5 (FeII2FeIII2)	6 (FeIIFeIII3)	7 (FeIII4)
9,10-dihydroanthracene	78	PCET	PCET	PCET
fluorene	82	PCET	PCET	PCET
2,4,6-tBu3-PhOH	82	PT	PCET	PCET
PMe3	-	NR	NR	OAT

^aPT = proton transfer, PCET = proton-coupled electron transfer (based on cluster products), OAT = oxygen atom transfer, NR = no reaction observed.

^bSecond-order rate constant.

³¹P NMR and GC/MS analyses suggest that 7 is capable of transferring an oxygen atom to trimethylphoshine (PMe₃), where the other Fe^III-oxo clusters display no reaction towards the phosphine on similar timescale (see SI). The difference in reactivity is likely due to the low reduction potentials of 5 and 6 precluding efficient oxygen atom transfer reactivity. A more oxidizing cluster, through oxidations of the distal Fe centers, 7 can undergo OAT.

The kinetics of C–H activation by these clusters was investigated. The reaction between 5 and 9,10-dihydroanthracene (DHA; BDE_C–H = 78 kcal/mol)^14c displays an expected first order dependence on substrate concentration, with an overall second order rate constant of 87 M⁻¹ s⁻¹, and a considerable kinetic isotope effect (KIE) of 7 with d₄-DHA. These data are consistent with a rate-limiting C–H bond activation for the PCET process to form 1 and anthracene. The second-order rate constants between 5 and C–H bonds of varying BDE_C–H and pK_a values were measured and display a linear dependence of the PCET reaction rate on the pK_a of the organic substrate (Figure 4), suggesting either a concerted or stepwise pK_a-driven process.¹⁸ Reactions between DHA and 6 or 7 produce the corresponding hydroxide-clusters and anthracene in yields comparable to 5 (Table S3) indicating PCET processes, but complex kinetics precluded the determination of rate constants and further insights into the mechanism of these reactions.

Figure 4.

Plot of log k₂ (normalized to number of reactive C-H bonds) versus reported pK_a values of the organic substrates in DMSO for PCET reactions with 5.

Overall, this report offers a rare systematic study of the effects of neighboring redox active metals on structural and reactivity aspects of a terminal metal-oxo. Because it is part of a cluster, the reactivity of the terminal metal-oxo motif can be tuned without changing the formal redox state of the metal supporting it; however, redox events at distal centers have significant effect on the acidity and BDE of the corresponding O-H bond. Clearly, the cluster as an assembly is essential for reactivity beyond the structural aspects of the isolated metal-oxo motif. Further development of multinuclear model systems is necessary to fully understand the nature and amplitude of these effects.