Ce^IV- and HClO₄-Promoted Assembly of an Fe₂^IV(μ-O)₂ Diamond Core from its Monomeric Fe^IV=O Precursor at Room Temperature

Abstract

Diiron(IV)-oxo species are proposed to effect the cleavage of strong C–H bonds by nonheme diiron enzymes such as soluble methane monooxygenase (sMMO) and fatty acid desaturases. However, synthetic mimics of such diiron(IV) oxidants are rare. Herein we report the reaction of (TPA*)Fe^II (1) (TPA* = tris(3,5-dimethyl-4-methoxypyridyl-2-methyl)amine) in CH₃CN with 4 equiv. CAN and 200 equiv. HClO₄ at 20 °C to form a complex with an [Fe^IV₂(μ-O)₂]⁴⁺ core. CAN and HClO₄ play essential roles in this unprecedented transformation, in which the comproportionation of Fe^III–O–Ce^IV and Fe^IV=O/Ce⁴⁺ species is proposed to be involved in the assembly of the [Fe^IV₂(μ-O)₂]⁴⁺ core.

Spontaneous assembly!

The spontaneous assembly of monomeric oxoiron(IV) units into an Fe^IV₂(μ-O)₂ diamond core complex is observed in the presence of CAN and excess HClO₄ at room temperature, an unexpected and unprecedented avenue for accessing high valent diamond core complexes

Graphical Abstract

Soluble methane monooxygenase (sMMO) and fatty acid desaturases are nonheme diiron enzymes that activate O₂ to respectively hydroxylate methane to methanol and convert saturated fatty acids into monounsaturated derivatives.¹ For sMMO, a diferryl intermediate Q has been shown to be responsible for cleaving the substrate C–H bond, generating a substrate radical that then undergoes oxygen rebound to form methanol.² A related but yet unobserved oxidant is proposed for the desaturases that extracts two adjacent H-atoms from the fatty acid substrate to form the unsaturated product.^1d The nature of the highly reactive Q has intrigued many scientists since its discovery as a key intermediate in the catalytic cycle of sMMOH. Mossbauer spectroscopy establishes that it has an antiferromagnetically coupled diiron(IV) core.² Resonance Raman (rR) studies reveal an O-isotope-sensitive vibration at 690 cm⁻¹ that is characteristic of an Fe₂(μ-O)₂ core.³ This notion was supported by early X-ray absorption spectroscopy (XAS) results that found a short Fe•••Fe distance of 2.46 Å consistent with a diamond core,⁴ but more recent efforts of DeBeer and co-workers⁵ using Fe Kα high-energy-resolution fluorescence-detected (HERFD) X-ray absorption spectroscopy (XAS) have cast doubt on this short Fe•••Fe distance and instead support an open Fe^IV–O–Fe^IV=O core structure with a much longer Fe•••Fe distance of 3.4 Å.⁶ These apparently conflicting conclusions from the resonance Raman and HERFD XAS experiments have to date not been resolved.

Currently, there are only a few synthetic diiron(IV) complexes that may serve as synthetic models for Q.⁷ The first complex with a nonheme Fe^IV₂(μ-O)₂ core was reported in 2007.^7a [(TPA*)₂Fe₂^IV(μ-O)₂]⁴⁺ (4) is supported by the tripodal TPA* (tris(3,5-dimethyl-4-methoxy-2-pyridylmethyl)-amine) and generated by the one-electron oxidation of its mixed-valent congener [(TPA*)Fe^III(μ-O)₂Fe^IV(TPA*)]³⁺. A year later, it was found that 4 could be obtained by reaction of H₂O₂ with [(TPA*)(H₂O)Fe^III(μ-O)Fe^III(OH)(TPA*)]³⁺ followed by the addition of 1 equiv. H⁺.^7b The second example of a complex with an Fe^IV₂(μ-O)₂ core has just been reported and is supported by a different tripodal ligand Me₃NTB (tris((1-methyl-1H-benzo[d]imidazol-2-yl)methyl)-amine).⁸ This new diiron(IV) complex is generated in two steps: initial exposure of the bis(μ-hydroxo)diiron(II) precursor to O₂ to form an adduct that is best described as a (μ-oxo)(μ−1,2-peroxo)diiron(III) intermediate, followed by treatment of the latter with Sc(OTf)₃ to cleave the O–O bond and afford the target diiron(IV) complex. In this paper, we report yet another strategy for accessing this elusive species, namely its unprecedented assembly starting from a mononuclear oxoiron(IV) complex with the help of Ce^IV and HClO₄ (Figure 1). In contrast to the earlier studies done at −40 °C,⁷ 4 is in fact stable enough to be generated at 20 °C.

Figure 1.

(A) Formation of 4 (red) from the reaction of 2 (black) by the addition of 4 equiv. CAN in 20 μL D₂O to 1 mM 1 in CH₃CN with 200 equiv. HClO₄ at 20 °C. Treatment of 2 with HClO₄ immediately generates 3 (blue), which then converts to 4 in 50% yield over the course of 100 s. Inset: Zoom in to show isosbestic points at 701 and 763 nm in the conversion of 3 to 4. (B) Absorption changes at 738 nm for 3 and 875 nm for 4 vs time. Note the different y-axis scales for 3 (right) and 4 (left).

In previous work, Nam and Fukuzumi have shown that mononuclear oxoiron(IV) complexes can be generated by the reaction of cerium(IV) ammonium nitrate (CAN) and iron(II) complexes, with the oxo moiety deriving from water.⁹ Following this precedent, we have demonstrated the formation of inner-sphere Fe^III–O–Ce^IV adducts from the reaction with Ce(III) with oxoiron(IV) complexes of the pentadentate ligands N4Py (N,N-bis(2-pyridylmethyl)-N-bis(2-pyridyl)methylamine) and BnTPEN (N-benzyl-N,N′,N′-tris(2-pyridyl-methyl)ethylenediamine) in CH₃CN, which are in water-dependent equilibria with corresponding Fe^IV=O species.¹⁰ Water favors Fe^IV=O formation by displacing nitrate as a ligand to the Ce^IV center, which raises the Ce^IV/Ce^III potential and promotes inner-sphere electron transfer from Fe^III to Ce^IV. In this report, we have extended this chemistry to [(TPA*)Fe^II(CH₃CN)₂](OTf)₂ (1, TPA* = tris-(3,5-dimethyl-4-methoxy-2-pyridylmethyl)amine). Addition of 4 equiv. CAN in 20 μL D₂O to a 1-mM solution of 1 in CH₃CN at 20 °C generates the Fe^III–O–Ce^IV adduct 2 (Figure 1A, black line), reminiscent of [(L)Fe^III–O–Ce^IV(NO₃)₄(H₂O)]⁺ complexes 2(N4Py) for L = N4Py and 2(BnTPEN) for L = BnTPEN) reported earlier.¹⁰ Subsequent treatment of 2 with 200 equiv. HClO₄ in CH₃CN immediately generates [Fe^IV(O)(TPA*)]²⁺ (3) with its characteristic band at 738 nm in close to quantitative yield (ε = 300 M^−1·cm⁻¹).^7a In the course of 100 s, 3 spontaneously converts into 4 in ~50% yield,⁷ as demonstrated by the appearance of intense absorption bands with maxima at 485 and 875 nm (ε = 2200 M^−1·cm⁻¹)^7a characteristic of 4 (Figure 1A). The 875-nm peak of 4 grows concomitantly with the decay of the 738-nm band of 3 at a rate of 0.03 s⁻¹ with isosbestic points at 701 and 763 nm (Figure 1B). To the best of our knowledge, this is the first example for the formation of an Fe^IV₂(μ-O)₂ diamond core from a monomeric oxoiron(IV) precursor at room temperature.

Scheme 1 shows the various species observed in this chemistry, as corroborated by their unique spectroscopic properties. Upon addition of 1 equiv. CAN, the ferrous precursor 1 is fully oxidized to the ferric state (Figure S1) and proceeds to generate 2 with additional CAN (see Figure 1 for a spectrum of 2). Just like the related [(L)Fe^III(O)Ce^IV(NO₃)₄]⁺ complexes (L = N4Py or BnTPEN) we previously reported,¹⁰ 2 exhibits an EPR spectrum at 4 K with signals at g = 8.8 and 4.2, characteristic of a rhombic S = 5/2 Fe^III center (Figure S2). Its resonance Raman (rR) spectrum (λ_exc 515 nm) shows a band at 695 cm⁻¹ (Figure 2A) that is assigned to the Fe–O stretch of the Fe–O–Ce unit, with a frequency flanked by the Fe^III–O stretches of corresponding 2(N4Py) (707 cm⁻¹) and 2(BnTPEN) (685 cm⁻¹) complexes.¹⁰ The 695 cm⁻¹ band of 2 downshifts to 660 cm⁻¹ upon labeling of the oxo bridge with H₂¹⁸O, as predicted by Hooke’s law.

Scheme 1.

Structure of 1 (left panel) and stepwise transformations from 1 to 4 (right panel).

Figure 2.

(A) Resonance Raman spectra (λ_exc 515 nm) of 2 from the reaction of 5 mM 1 in CH₃CN with 4 equiv. CAN in 20 μL H₂O (black) and the corresponding ¹⁸O-labeled complex generated from 4 equiv. CAN in 20 μL H₂¹⁸O (red). (B) Unfiltered Fe K-edge EXAFS spectrum (dotted black line) and best fit (solid red line) of 2 at 10 K. Inset: corresponding unfiltered k-space data (dotted black line) and the best fit (solid red line).

Further insights into the nature of 2 have been obtained in frozen solution at 10 K by X-ray absorption spectroscopy. Complex 2 has an Fe K-edge energy at 7125.1 eV (Figure S3), which is ca. 2 eV higher than that reported for 2(N4Py) (7122.7 eV) species.¹⁰ This 2.3-eV upshift is not unusual for an Fe(TPA*) complex, as 4 and its diiron(IV) precursor have K-edge energies in the 7129–7130-eV range.^{7, 8} Similarly, the pre-edge peak (assigned to a 1s → 3d transition) for 2 is centered at 7115 eV, upshifted by 2 eV from that of 2(N4Py). It has an area of 8.3 units, consistent with values found for other Fe^III(μ-O)Ce^IV complexes.¹⁰ This low pre-edge area is also inconsistent with an Fe^IV oxidation state assignment for 2, as pre-edge areas of Fe^IV=O species typically range from 20 to 45 units;¹¹ for example, the reported pre-edge area for [(TPA)Fe^IV(O)(solv)]²⁺ is 25.4(3) units).¹²

The EXAFS spectrum of 2 shows three sets of features at R + Δ ~ 1.7, 2.5, and 3.3 Å (Figure 2B, Tables S1–2). The peak at shortest distance corresponds to the first shell of scatterers, namely 1 N/O at 1.86 Å (the oxo bridge) and 5 N/O scatterers at 2.11 Å (4 N-atoms from TPA* and 1 from CH₃CN), which are within the range of first-shell distances found for 2(N4Py) and 2(BnTPEN).¹⁰ The one at an intermediate distance is assigned to a carbon shell at 3.01 Å, while the feature at the longest distance arises from strong scattering by the heavy Ce scatterer at 3.66 Å. This scatterer is observed at a significantly shorter distance than the Fe•••Ce distances found in 2(N4Py) (3.825 Å) and 2(BnTPEN) (3.78 Å), indicating a smaller Fe^III–O–Ce^IV angle in the TPA* complex than in the corresponding N4Py and BnTPEN complexes.¹⁰ Assuming the Ce^IV–(μ-O) distance in 2 to be comparable to that found in 2(N4Py) (2.013 Å),¹⁰ 2 would have an estimated Fe^III–O–Ce^IV angle of about 141°. This smaller angle in 2 relative to the 170.3° value found crystallographically for 2(N4Py)¹⁰ is likely due to the tetradentate nature of TPA*, which leaves an open site at the iron center of 2 for a sterically much less demanding CH₃CN ligand.

By analogy to the previously reported effect of adding water that shifts the equilibrium between an Fe^III–O–Ce^IV adduct and the corresponding Fe^IV=O species to favor the latter,¹⁰ adding excess HClO₄ converts 2 immediately to 3 in close to quantitative yield, as evidenced by the characteristic absorption band of 3 at 738 nm (ε = 300 M⁻¹ cm⁻¹)^7a (Figure 1). Addition of high concentrations of HClO₄ raises the Ce^IV/Ce^III potential of CAN as well,¹³ so HClO₄ plays a role similar to that of water in this conversion to increase the Ce^IV/Ce^III potential upon its addition to promote inner-sphere electron transfer from Ce^IV to Fe^III centers, resulting in the facile conversion of 2 to 3.

Complex 3 then converts into 4 over a 100-s time frame at 20 °C, as indicated by its intense charge transfer bands at 485 nm (ε = 9800 M⁻¹ cm⁻¹) and 875 nm (ε = 2200 M⁻¹ cm⁻¹) (Figure 1). In corroboration, the resonance Raman spectrum of 4 exhibits a vibration at 674 cm⁻¹ that downshifts by 28 cm⁻¹ upon ¹⁸O-labeling and shows maximum resonance enhancement with 515-nm excitation (Figures S5 and S6). These results concur with the original report on 4 establishing the presence of an [Fe^IV₂(μ-O)₂]⁴⁺ core.^{7, 8} Its excitation profile (Figure S5) reveals maximum resonance enhancement with 515-nm excitation, consistent with the assignment of the intense band at 485 nm as a charge transfer transition of the Fe^IV₂(μ-O)₂ core. The facile assembly of the diamond-core containing 4 from monoiron precursors reported herein is an intriguing result that would seem initially to defy chemical logic in the combination of two [Fe^IV=O]²⁺ units to form an [Fe^IV₂(μ-O)₂]⁴⁺ product with double the overall charge. Nevertheless, we observe that it does happen in as high as 50% yield (Figure 1) with CAN and HClO₄ playing crucial roles in facilitating the conversion of 3 to 4 (Scheme 1). Unlike our earlier work on 4, which was carried out at −40 °C,⁷ the current studies are done at 20 °C, a temperature at which 4 can persist for 30 min.

The formation of 4 has been investigated as a function of [CAN] while keeping the added HClO₄ fixed at 200 equiv. As shown in Figure 3A, 4 forms rapidly and then decays. With 2 equiv. CAN added, the yield of 4 is only about 25% but doubles to 50% with 4 equiv. CAN added; however no further increase is observed beyond 4 equiv. CAN (Figure S7). It is found that k_formation and k_decay values for 4 respectively remain constant at 0.03 s⁻¹ and 0.002 s⁻¹ (corresponding to a t_1/2 ~ 6 min) (Figure S8). Also observed is a lag phase at the onset of decay for the experiments with 6 and 8 equiv., extending the overall lifetime of 4 to 1 hour with 8 equiv. CAN. This suggests that excess CAN regenerates 4 and must be consumed before the final exponential decay phase can be observed. We surmise that CAN likely plays an additional role besides simply oxidizing Fe(II) to Fe(IV).

Figure 3.

[CAN] dependence on the formation of 4 at RT. Conditions: (A) Addition of 2–8 equiv. CAN (in 20 μL D₂O) to a 1-mM solution of 1 in CH₃CN to generate 2 to which 200 equiv. HClO₄ was added. (B) Addition of 0.5–10 equiv. CAN in 20 μL CH₃CN to a 1-mM solution of 3 (generated by the addition of 1.2 equiv. ArIO in TFE to 1 in CH₃CN).

Figure 4.

(A) [HClO₄] dependence of the conversion of 2 to 4 at 20 °C upon mixing 1 mM 1 in CH₃CN and 4 equiv. CAN (in 20 μL D₂O). (B) Plot of the rates of 4 formation and decay at 20 °C in units of s⁻¹ as a function of [HClO₄].

To probe whether CAN alone is sufficient to convert 3 to 4, we have avoided adding HClO₄ by preparing 3 through the reaction of 1 with the oxo-transfer agent 2-^tBuSO₂-C₆H₄IO (ArIO) in CH₃CN, and then treating it with 0.5–10 equiv. CAN in CH₃CN. In this reaction, 3 does afford 4 in the absence of HClO₄, suggesting that CAN alone enables the conversion of 3 to 4 (Figures 3B, S9, S10). However, under these conditions, a maximum yield of only 10% is obtained at 5 equiv. CAN or higher, a factor of 5 lower than that found for reactions with CAN/HClO₄. This result suggests that, although not essential, HClO₄ plays an important role in facilitating the conversion from 3 to 4. Indeed adding 200 equiv. HClO₄ to this reaction raises the yield of 4 to 50% (Figure S11).

Figure 4 shows how [HClO₄] affects the formation of 4. In Figure 4A, the yield of 4 is only about 13% with the addition of 50 equiv. HClO₄. This is not surprising as insufficient acid has been added at this point to raise the Ce^IV/Ce^III potential sufficiently to allow for quantitative conversion of 2 to 3, let alone 3 to 4 (Figure S12). However, doubling the number of acid equivalents to 100 more than triples the yield of 4 to ~45%, and a maximum yield of ~50% is obtained with more added acid. Figure 4B monitors the formation of 4 at 875 nm as a function of [HClO₄] and shows that the rate of formation increases approximately as a linear function of [HClO₄], with some plateauing found at higher [HClO₄] values. In contrast, the decay rate for 4 is essentially independent of [HClO₄] above 50 equiv. HClO₄, suggesting that HClO₄ likely does not affect 4 decay. While the requirement for a high concentration of HClO₄ to achieve the maximum yield of 4 can at first glance appear puzzling, work reported over 75 years ago¹³ has shown that HClO₄ raises the Ce^IV/Ce^III potential of CAN. In the current study, just slightly over 100 equiv. HClO₄ is the amount necessary to maximize the formation of 4. At the present time, it is not clear to us why the yield of 4 maximizes at 50% yield under our conditions, and further work is required to clarify this question. Nevertheless, the fact that a 50% yield of 4 can be achieved by this unique pathway for assembling an [Fe^IV₂(μ-O)₂]⁴⁺ core at room temperature should be a compelling enough rationale for further investigation.

The accumulated results we have presented above lead to the key question of how to assemble two units of 3 to form 4. To shed light on this puzzle, we have investigated the formation of 4 as a function of [1] and find a linear dependence of the k_obs for the formation of 4 from 3 (Figures 5, S13, Table S3). As this reaction is first order with respect to [Fe], the rate determining step cannot be a simple dimerization of two dicationic Fe^IV=O complexes, but likely requires a step in which 3 becomes activated. One attractive possibility is the formation of an adduct of 3 with CAN (3+Ce). Such an adduct may acquire partial Fe^V=O character to allow it to react with residual 2 in a comproportionation step to form 4 (Scheme 2). A precedent for such an Fe^IV=O/Ce^IV adduct can be found in the reaction of α-[(mcp)Fe^IV(O)]²⁺ (mcp = (N,N’-dimethyl-N,N’-bis(2-pyridylmethyl)-cis-1,2-diaminocyclohexane) with excess CAN by Costas et al. and is proposed to be the species responsible for the catalytic water oxidation observed for this reaction mixture.¹⁴ This intermediate has been characterized by cryospray ionization mass spectrometry and by UV-vis absorption and resonance Raman spectroscopy that lead to its formulation as the {[(mcp)Fe^IV(O)(μ–O)Ce^IV(NO₃)₃]}⁺ ion.¹⁵ Similarly, Karmalkar et al. have more recently demonstrated the formation of a (DPAQ)Mn^IV–O–Ce^IV adduct (DPAQ = 2-[bis(pyridine-2-ylmethyl)]-amino-N-quinolin-8-yl-acetamidate anion), which is supported by the observation of a ν(Mn–O) of 675 cm⁻¹ by resonance Raman spectroscopy as well as an Mn–O distance of 1.69 Å and an Mn•••Ce distance of 3.67 Å by EXAFS analysis.¹⁶ Thus a similarly formulated 3+Ce adduct in this work may represent the key species that reacts with residual 2 to generate 4 (Scheme 2). The high HClO₄ concentrations required also promote 4 formation by raising the Ce^IV/Ce^III potential¹³ to make the more potent 3+Ce oxidant.

Figure 5.

Linear dependence on [1] of the formation rate of 4 at 20 °C. Conditions: 0.5–2.5 mM 1 in CH₃CN + 4 equiv. CAN in 20 μL D₂O + 200 equiv. HClO₄ at 20 °C. Note that [CAN] and [HClO₄] are increased proportionately with [1].

Scheme 2.

Proposed mechanism for the formation of 4.

In summary, we have found reaction conditions that lead to the formation of an Fe^IV₂(μ-O)₂ diamond core from a monoiron precursor with the help of CAN and HClO₄. The notion that two dicationic monoiron species could be assembled into a tetracationic diiron product is counter-intuitive and quite an unexpected outcome. This transformation would not have been discerned, were it not for the intense near IR chromophore associated with 4. Indeed, these results demonstrate how much more there is to learn in the exploration of high-valent iron chemistry.

Experimental Section

Experimental Details. See the Supporting Information for experimental details of synthesis and physical methods.

Table 1.

Properties of complexes relevant to this work

	2a	2(N4Py)10	3a	47
λmax (nm)	500 (sh)	500 (sh)	738	485, 875
v(Fe–O) (cm−1) [Δ18O]	695 [−35]	707 [−31]	830b	677 [−30]
Pre-edge area	8.3	13.7		19
Fe–O (Å)	1.86	1.825	1.65c	1.77
Fe•••M (Å) M =	3.66 (Ce)	3.852 (Ce)	—	2.73 (Fe)
Fe–O–M (o)	141	170.3	—	101

^aThis work.

^bValue for [Fe^IV(O)(TPA)(NCMe)]²⁺, see Figure S4.

^cValue for [Fe^IV(O)(TPA)(NCMe)]²⁺ from reference 12.