An Fe^IVO complex of a tetradentate tripodal nonheme ligand

Abstract

The reaction of [Fe^II(tris(2-pyridylmethyl)amine, TPA)(NCCH₃)₂]²⁺ with 1 equiv. peracetic acid in CH₃CN at −40°C results in the nearly quantitative formation of a pale green intermediate with λ_max at 724 nm (ɛ ≈ 300 M⁻¹⋅cm⁻¹) formulated as [Fe^IV(O)(TPA)]²⁺ by a combination of spectroscopic techniques. Its electrospray mass spectrum shows a prominent feature at m/z 461, corresponding to the [Fe^IV(O)(TPA)(ClO₄)]⁺ ion. The Mössbauer spectra recorded in zero field reveal a doublet with ΔE_Q = 0.92(2) mm/s and δ = 0.01(2) mm/s; analysis of spectra obtained in strong magnetic fields yields parameters characteristic of S = 1 Fe^IVO complexes. The presence of an Fe^IVO unit is also indicated in its Fe K-edge x-ray absorption spectrum by an intense 1-s → 3-d transition and the requirement for an O/N scatterer at 1.67 Å to fit the extended x-ray absorption fine structure region. The [Fe^IV(O)(TPA)]²⁺ intermediate is stable at −40°C for several days but decays quantitatively on warming to [Fe₂(μ-O)(μ-OAc)(TPA)₂]³⁺. Addition of thioanisole or cyclooctene at −40°C results in the formation of thioanisole oxide (100% yield) or cyclooctene oxide (30% yield), respectively; thus [Fe^IV(O)(TPA)]²⁺ is an effective oxygen-atom transfer agent. It is proposed that the Fe^IVO species derives from O—O bond heterolysis of an unobserved Fe^II(TPA)-acyl peroxide complex. The characterization of [Fe^IV(O)(TPA)]²⁺ as having a reactive terminal Fe^IVO unit in a nonheme ligand environment lends credence to the proposed participation of analogous species in the oxygen activation mechanisms of many mononuclear nonheme iron enzymes.

Mechanisms for the activation of dioxygen at iron centers typically invoke the heme paradigm in which a high-valent iron-oxo intermediate serves as the oxidizing species (1). Such species have been trapped in the catalytic cycles of heme enzymes such as cytochrome P450 (2, 3) and peroxidases (where these species are referred to as compounds I and II) (1). The generation and characterization of corresponding biomimetic oxo-iron(IV) porphyrin complexes have significantly enhanced our understanding of their structural, spectroscopic, and reactivity properties (4, 5). A diiron(IV) intermediate called Q has been identified with Mössbauer spectroscopy in the catalytic cycle of the nonheme diiron enzyme methane monooxygenase (6–8); this intermediate has been proposed from extended x-ray absorption fine structure (EXAFS) evidence to have an Fe^IV₂(μ-O)₂ core structure (9). Iron(IV)-oxo species have also been proposed as the oxidizing species for nonheme monoiron enzymes that require pterin or α-ketoglutarate cofactors (10, 11), but evidence for the analogous high-valent iron-oxo species is at best only indirect.

Synthetic efforts in the past 10 years have demonstrated the accessibility of the iron(IV) oxidation state in a nonheme ligand environment, particularly in coupled iron(III)iron(IV) complexes (12–18); indeed one of these, [Fe₂(μ-O)₂(5-Et₃-tris(2-pyridylmethyl)amine, TPA)₂](ClO₄)₃, has been crystallographically characterized (19). However, only recently has evidence for synthetic mononuclear iron(IV)-oxo species been obtained. Grapperhaus et al. (20) reported the generation of such a species, formulated as [Fe^IV(O)(1-carboxymethyl-1,4,8,11-tetraazacyclotetradecane or cyclam-acetate)]⁺, by the reaction of its iron(II) precursor with O₃ at −80°C, but its instability and low yield precluded detailed characterization beyond its Mössbauer spectrum. More recently, we investigated the reaction of the iron(II) complex of a closely related macrocyclic ligand [Fe^II(1,4,8,11-tetramethyl-1,4,8,11-tetraazacyclotetradecane or tetra(N-methyl)cyclam, TMC)(trifluoromethylsulfonate or triflate anion, OTf)₂] with PhIO or H₂O₂ and generated [Fe^IV(O)(TMC)(NCCH₃)](OTf)₂ in high yield. The remarkable stability of this complex has allowed us to obtain the high-resolution structure of a complex with a terminal Fe^IVO unit (21). In this article, we demonstrate that the Fe^IVO unit can also be supported by the tetradentate tripodal ligand TPA. In this case, [Fe^IV(O)(TPA)]²⁺ is formed nearly quantitatively from the stoichiometric reaction of its iron(II) precursor with a peracid and is a more reactive species. Its greater oxygen-atom transfer capabilities toward substrates like thioanisole and cyclooctene emphasize the important role that ligand structure can play in modulating the reactivity of the Fe^IVO unit.

Materials and Methods

Materials and General Procedures.

All reagents and solvents were purchased from commercial sources and used as received, unless noted otherwise. Solvents were dried according to published procedures and distilled under Ar before use (22). meta-Chloroperbenzoic acid was purified by washing with phosphate buffer (pH 7.5) and subsequent recrystallization of the residue from CH₂Cl₂/Et₂O (22). Labeled water (95% ¹⁸O) was purchased from ICON Isotopes (Summit, NJ). Preparation and handling of air-sensitive materials were carried out under an inert atmosphere by using either standard Schlenk and vacuum line techniques or a glove box. Elemental analysis was performed by Atlantic Microlab (Norcross, GA).

Preparation of Iron(II) Precursors.

[Fe(TPA)(NCCH₃)₂](ClO₄)₂ (1a) was synthesized from stoichiometric amounts of Fe(ClO₄)₂⋅6H₂O, TPA⋅3HClO₄ (23), and NEt₃ following a literature procedure (24). Caution: perchlorate salts of metal complexes with organic ligands are potentially explosive and should be handled with care! [Fe(TPA)(OTf)₂] (1b) was prepared under an inert gas atmosphere by adding a solution of Fe(OTf)₂⋅2CH₃CN (25) (0.436 g, 1 mmol) in 2 ml of tetrahydrofuran to a solution of TPA (0.305 g, 1.05 mmol) in 2 ml of tetrahydrofuran. (The free base ligand TPA was obtained from TPA⋅3HClO₄ and NEt₃, extracted with CH₂Cl₂, and recrystallized from ethyl acetate after work-up.) After stirring for 12 h, the solution was evaporated to obtain a yellow solid. Recrystallization from CH₂Cl₂/Et₂O afforded the product as fine needles (76%). Analysis for [Fe(TPA)(OTf)₂]⋅3H₂O or C₂₀H₂₄F₆FeN₄O₉S₂, calculated (found): C, 34.40 (34.70); H, 3.46 (3.53); F, 16.32 (16.28); N, 8.02 (8.12); S, 9.18 (9.13). Dissolution of [Fe(TPA)(OTf)₂] in CH₃CN formed the low-spin [Fe(TPA)(NCCH₃)₂]²⁺ ion, whose NMR spectrum was indistinguishable from that of 1a. The preparation of [⁵⁷Fe(TPA)(OTf)₂] was carried out analogously by using ⁵⁷Fe(OTf)₂⋅2CH₃CN.

Physical Methods.

UV/visible spectra were recorded on an Hewlett–Packard 8453A diode array spectrometer with samples maintained at low temperature by using a cryostat from Unisoku Scientific Instruments, Osaka. ¹H NMR spectra were recorded on a Varian Inova VXR-300 spectrometer at ambient temperature. Chemical shifts (ppm) were referenced to the residual solvent peaks. Electrospray MS was performed on a Finnigan-MAT (San Jose, CA) LCQ ion trap mass spectrometer with samples of 0.33 mM Fe complex. The spectrum of the thermally labile intermediate was obtained by direct introduction of the solution of the intermediate generated at −40°C into the injector port of the spectrometer. The capillary heater on the instrument was turned off and the flow of the atomizing gas was increased to minimize thermal decomposition of the unstable intermediate.

Mössbauer spectra were recorded with two spectrometers, using Janis Research (Wilmington, MA) SuperVaritemp dewars that allow studies in applied magnetic fields up to 8.0 T in the temperature range from 1.5 to 200 K. Mössbauer spectral simulations were performed by using the wmoss software package (WEB Research, Minneapolis). Isomer shifts are quoted relative to Fe metal at 298 K.

X-ray absorption spectroscopy (XAS) data were collected at beamline X9 of the National Synchrotron Light Source at Brookhaven National Laboratory (Upton, NY). X-ray absorption spectra at the iron K-edge were collected between 6.9 and 8.0 keV, and the monochromator was calibrated by using the edge energy of iron foil at 7,112.0 eV. The data were obtained in fluorescence mode [A_exp (C_f/C₀)] at 13(1) K. Our XAS data analysis protocol has been described (26).

Results and Discussion

The reaction of [Fe^II(TPA)(NCCH₃)₂]²⁺ (1) with 1 equiv. CH₃CO₃H (32 wt %) in CH₃CN at −40°C produces a pale green intermediate 2 with a λ_max at 724 nm (ɛ ≈ 300 M⁻¹⋅cm⁻¹) within 3 min (Fig. 1A). The same intermediate is formed with 1 equiv. m-chloroperbenzoic acid. This species is clearly distinct from the more intense purple (λ_max 538 nm; ɛ ≈ 1,000 M⁻¹⋅cm⁻¹) and blue (λ_max 600 nm; ɛ ≈ 2,000 M⁻¹⋅cm⁻¹) chromophores observed on addition of 10 equiv. H₂O₂ and ^tBuOOH, respectively, to 1 in CH₃CN at −40°C; the latter have been respectively identified as [Fe^III(TPA)OOH]²⁺ (27, 28) and [Fe^III(TPA)OO^tBu]²⁺ (24). Intermediate 2 is stable for several days at −40°C, but it decays on warming to a stable species 3, whose visible spectrum strongly resembles that of [Fe^III₂(μ-O)(μ-OAc)(TPA)₂]³⁺ (Fig. 1B; ref. 29). The absorbance of 3 indicates that the diiron(III) product is formed quantitatively from its iron(II) precursor. Because an isosbestic point is observed in the conversion of 2 to 3, intermediate 2 must represent the major component of the reaction mixture.

Fig 1.

(A) Conversion of 2 mM 1 to 2 in CH₃CN at −40°C by addition of 1 equiv. CH₃CO₃H (32 wt %), as monitored by UV-visible spectroscopy. (B) Subsequent conversion of 2 to 3 at 10°C. For reasons yet undetermined, the addition of 5 μl of H₂O to the 3-ml solution increased the stability of 2 by a factor of 3 without affecting λ_max and ɛ_max; thus, all subsequent samples were prepared in this manner.

Electrospray MS provides important insights into the above reaction. The spectrum of 1a shows a prominent feature at m/z 445, corresponding to the [Fe(TPA)(ClO₄)]⁺ ion (data not shown). Direct introduction of solutions of 2, maintained at −40°C, into the mass spectrometer results in the appearance of a prominent ion at m/z 461, whose mass and isotope distribution pattern correspond to [Fe(O)(TPA)(ClO₄)]⁺ (Fig. 2). Thus, the reaction with peracid has introduced one oxygen atom into the principal ion. This oxygen atom does not exchange with solvent water, as no mass shift is observed when H₂¹⁸O is added before the addition of CH₃CO₃H to 1a or after formation of 2. When 2 is allowed to decay, the m/z 461 feature disappears and is replaced by a major feature at m/z 965 that corresponds to the {[Fe₂(O)(OAc)(TPA)₂](ClO₄)₂}⁺ ion of 3 (data not shown). Analogous features were observed with 1b as starting material. Thus 2 is a transient species whose formulation requires the introduction of an oxygen atom into 1, resulting in either oxygen–atom incorporation into the TPA ligand or the oxidation of the iron(II) center to the Fe^IVO state.

Fig 2.

Electrospray mass spectrum of a solution of 1a in CH₃CN maintained at −40°C 3 min after addition of CH₃CO₃H.

Mössbauer Studies.

Figs. 3A and 4 show Mössbauer spectra of 2, dissolved in CH₃CN, recorded between 4.2 and 100 K in applied fields up to 7.0 T. The zero field spectrum of Fig. 3A consists of a doublet, accounting for ≈80% of the Fe, with quadrupole splitting ΔE_Q = 0.92(2) mm/s and isomer shift δ = 0.01(2) mm/s. Approximately 18% of the absorption arises from a diamagnetic diiron(III) complex with ΔE_Q = 1.48(5) mm/s and δ = 0.45(2) mm/s, associated with the decay product 3 as its spectral features are identical to those observed in Fig. 3B. These are also the parameters observed for independently synthesized [Fe₂O(O₂CR)(TPA)₂]³⁺ complexes (30). For clarity we have removed the contribution of 3 from the spectra of Fig. 4. The remainder of the absorption in Fig. 3A is an unidentified minor component (≈3–5%), discernible at 4.2 K at velocities <−1.5 mm/s.

Fig 3.

The 4.2-K Mössbauer spectra were recorded in zero field. (A) Spectrum of 2 from the reaction of 1b with CH₃CO₃H. The solid line is a spectral simulation using the ΔE_Q and δ values of Table 1. Eighty percent of the Fe in the sample belongs to 2, and the remainder is a diiron(III) species with properties identical to that observed in B. The spectrum in B was observed after 2 was allowed to decay. This species, 3, is diamagnetic according to a spectrum (not shown) recorded in an applied field of 7.0 T; ΔE_Q = 1.48(5) mm/s, δ = 0.45(2) mm/s, η = 0.5 (approximately equivalent sites). (C) Spectrum observed after addition of thioanisole to 2. The major doublet (75%) is a low-spin iron(II) (S = 0) species with ΔE_Q = 0.35 mm/s and δ = 0.44 mm/s. The remainder of the absorption is mainly a diiron(III) species, most likely the same species present in the sample before addition of thioanisole.

Fig 4.

Mössbauer spectra of 2 (same sample as that used for Fig. 3A) recorded in magnetic fields applied parallel to the observed γ-rays at temperatures and fields indicated. Solid lines are spectral simulations based on Eq. 1 using the parameters listed in Table 1.

The Mössbauer spectral features of 3 compare well with those of S = 1 Fe^IVO complexes, and therefore we have fitted the data with the spin Hamiltonian

where all of the symbols have their conventional meaning. The solid lines drawn through the data in Figs. 3A and 4 are theoretical spectra computed with the parameters listed in Table 1. A few comments about the simulations are in order. The parameters D, ΔE_Q, η, A_z, and δ are determined with good precision. However E/D, A_x, and A_y are strongly correlated, which is not surprising because the spectra depend on the magnetic hyperfine field, B_hf(i) = −〈S_i〉A_i/g_nβ_n (i = x, y) in the x-y plane. The observed hyperfine field in the x-y plane is nearly isotropic and, because E/D determines the anisotropy of the expectation values of S, 〈S_x,y〉, one can trade A_x and A_y for E/D. Table 1 lists the parameter set for E/D = 0, but essentially equivalent fits are obtained for any value 0 ≤ E/D ≤ 1/3. For E/D = 1/3 we obtained A_x/g_nβ_n = −35 T and A_y/g_nβ_n = −19 T; the values for A_x and A_y depend nearly linearly on E/D.

Table 1.

Comparison of [Fe^IV(O)L]_n complexes with S = 1 sites

L =	TPA (2)	TMC*	Cyclam-acetate†	BPMCN‡
n	1	1	1(?)	2
λmax, nm	724	820	676	656, 845
ɛmax, M−1⋅cm−1	300	400		4,200, 3,700
XAS pre-edge peak area	30 (4)	30 (4)		10 (2)
r(Fe−O), Å	1.67 (2) (EXAFS)	1.646 (3) (x-ray)		1.79 (2) (EXAFS)
ν(Fe=O), cm−1		834 (IR)		653, 679, 686 (Raman)
Mössbauer parameters
D (cm−1)	28 (2)	28	23	19
E/D§	0§	0	0	0.15
Ax,y,z/gnβn (T)	−23.5,§ −23.5,§ −5	−25, −20, −3	−23, −23, −10	−22, −17.1, −1.1¶
ΔEQ, mm/s	0.92 (2)	1.24	1.37	1.75¶
η	0.9 (3)	0.5	0.8	0.8
δ, mm/s	0.01 (2)	0.17	0.01	0.10

^*From ref. 21.

^† From ref. 20.

^‡ From ref. 17.

^§ A_x, A_y, and E/D are strongly correlated; see text.

^¶ The site symmetries of this complex are low as indicated by the observation that both the A tensor and the EFG tensor are rotated relative to the zero-field splitting tensor; the rotation angles are quoted in the paper by Costas et al. (17).

The Mössbauer spectra of 2 are incompatible with a high-spin (S = 2) Fe^IV assignment. Thus, fitting the data to an S = 2 site yields an unreasonably large D value (D ≈ 25 cm⁻¹) and an A_iso = (A_x + A_y + A_z)/3 that is 2.5 to three times smaller than A_iso values for S = 2 Fe^IV sites. At 100 K the electronic spin relaxes rapidly and the temperature dependence of the magnetic hyperfine field, B_hf, follows the Curie law. The splitting of the low energy feature reveals that B_hf ≈ S(S + 1){g_⊥βB/3 kT}{A_⊥/g_nβ_n} ≈ −1.6 T, yielding A_⊥/g_nβ_n ≈ −7.3 T for S = 2. It is the presence of the larger spin factor S(S + 1) for S = 2 that requires the small A values. Adjusting to S = 1 increases A_⊥ to values commonly observed for S = 1 complexes. B_hf increases at 4.2 K nearly linearly with the applied field, implying a very large D no matter whether S is assumed to be 1 or 2.

Whereas D values of 20–30 cm⁻¹ are quite common for S = 1 Fe^IV sites, available data directly and indirectly indicate much smaller D values for high-spin Fe^IV sites. The only structurally characterized S = 2 compound, the five-coordinate chloroiron(IV) complex of a macrocyclic tetraamidate reported by Collins et al. (31) has D = −2.6 cm⁻¹ and A_iso/g_nβ_n = −12 T. Exchange-coupled S = 1/2 Fe^IIIFe^IV complexes, designed as models for the diiron clusters of methane monooxygenase and ribonucleotide reductase, contain high-spin Fe^IV sites with A_iso/g_nβ_n = −(15.8–18.5) T [obtained from the raw data after correcting for the spin projection factor; A_iso = (−4/3)a_iso, where a_iso is the intrinsic a value of the Fe^IV site] (13, 14, 16, 18). Moreover, the S = 1/2 ground states of the Fe^IIIFe^IV complexes exhibit g values confined to the range between 1.99 and 2.01, suggesting that mixing of excited orbital states into the ground state by spin-orbit coupling is very small. As long as only states derived from ⁵D are considered the zero-field splitting tensor is proportional to g-2 (32), and thus D is expected to be small.

XAS.

Structural insight into 2 is provided by XAS. Complex 2 exhibits an intense 1-s → 3-d preedge transition with an area of 30(4) units (Fig. 5). The value observed here is much higher than is commonly seen for six-coordinate iron complexes (4–10 units; refs. 33–35) but comparable to those observed for [Fe^IV(O)(TMC)(CH₃CN)](OTf)₂ [30(4) units] (Fig. 5) and for high-valent iron-oxo porphyrin complexes (27–38 units) (36, 37). Such a transition becomes more intense as the metal environment deviates from centrosymmetry, consistent with the presence of a terminal oxo ligand that imposes a significant distortion on the iron coordination environment.

Fig 5.

X-ray absorption near-edge features (Fe K-edge, fluorescence excitation) of 2 (solid line), [Fe^IV(O)(TMC)(CH₃CN)](OTf)₂ (dotted line), and [Fe^IV₂(μ-O)₂(BPMCN)₂](OTf)₄ (dashed line).

EXAFS analysis of 2 shows that its first coordination sphere consists of one O/N at 1.67 Å, four N/O at 1.99 Å, and one N/O at 2.20 Å (Fig. 6, Table 2). The 1.67-Å distance, assigned to the oxo ligand, compares well with corresponding distances in synthetic oxo-iron(IV) porphyrins and heme peroxidase compounds I and II determined from EXAFS analysis (36, 38); more importantly, it closely matches the 1.646(3)-Å distance recently found in the crystal structure of [Fe^IV(O)(TMC)(NCCH₃)](OTf)₂ (21). The other scatterers are associated with the TPA nitrogens and the yet unidentified sixth ligand (possibly the acetate from CH₃CO₃H, a triflate counterion, or CH₃CN and designated as Y in Scheme ). For example, the 1.99-Å distance is comparable to the average Fe–N distance of 2.025 Å found for the low-spin iron centers in the crystal structure of [Fe^IIIFe^IV(μ-O)₂(5-Et₃-TPA)₂]³⁺ (19), whereas the 2.20-Å distance may arise from the sixth ligand or the tertiary amine nitrogen of TPA. Thus the XAS analysis supports the minimal formulation of 2 as an S = 1 [Fe^IV(O)(TPA)]²⁺ complex (Scheme ).

Fig 6.

Fourier transform of the Fe K-edge EXAFS data [k³χ(k)] and Fourier-filtered EXAFS spectrum [k³χ′(k), Inset] of 6 mM 2 in frozen CH₃CN solution at T = 13(1) K, obtained by fluorescence detection and prepared from a 2:1 mixture of [Fe(TPA)(OTf)₂] and [⁵⁷Fe(TPA)(OTf)₂]. This same sample was used for the Mössbauer analysis shown in Figs. 3A and 4. Fourier-transform range k = 2–15 Å⁻¹; back-transformation window indicated by dashed vertical lines; experimental data (dotted line) and best fit (solid line). Fitting: one O/N at 1.67 Å (Δσ², 0.0004 Ų), four N/O at 1.99 Å (0.0016), one N/O at 2.20 Å (0.003), six C at 2.89 Å (0.0028).

Table 2.

EXAFS fitting results for 2

Fe—O/N			Fe—N/O			Fe—N/O			Fe—C			GOF
n	r, Å	Δσ2	n	r, Å	Δσ2	n	r, Å	Δσ2	n	r, Å	Δσ2	ɛ2 ×103
			6	2.00	5.9							1.86
1	1.67	−0.3	5	1.99	3.6							1.25
1	1.67	0.6	4	1.99	1.5	1	2.19	1				1.06
1	1.67	0.4	4	1.99	1.6	1	2.20	3	6	2.89	2.8	0.77

Fourier-transformed range k = 2–15 Å⁻¹ (resolution 0.12 Å). r is in units Å, Δσ² in 10⁻³ Ų. Back-transformation range: r′ = 0.60–3.20 Å. The 1.67-Å shell was fit by using a scatterer with oxygen parameters, whereas the 1.99- and 2.20-Å shells were fit by using scatterers with nitrogen parameters. Note, however, that backscatterers differing in Z by 1 unit cannot be distinguished by EXAFS. GOF, goodness of fit.

Scheme 1.

Reactivity Studies.

Intermediate 2 is stable in CH₃CN at −40°C for several days but converts stoichiometrically to [Fe^III₂(μ-O)(μ-OAc)(TPA)₂]³⁺ (3) on warming (Scheme ). It is not clear what becomes oxidized in this transformation. On the other hand, adding a 10-fold excess of thioanisole to a solution of 2 at −40°C results in the immediate bleaching of the pale green chromophore and the formation of a stoichiometric amount of the corresponding sulfoxide product. Mössbauer studies show that 3 is not formed under these conditions; instead a low-spin iron(II) species with δ = 0.44(1) mm/s; ΔE_Q = 0.35(1) mm/s (Fig. 3C) is observed with parameters identical to [Fe^II(TPA)(NCCH₃)₂]²⁺ (1) (data not shown). Also 2 reacts with a 100-fold excess of cyclooctene at −40°C, decaying over a 4-h period and affording the corresponding epoxide in 30% yield. Thus 2 is an effective oxygen-atom transfer agent at −40°C.

When the oxidation of thioanisole by 2 is carried out at −40°C in the presence of H₂¹⁸O, no ¹⁸O is incorporated into the thioanisole oxide product. This result demonstrates that the oxo ligand of 2 does not readily exchange with H₂¹⁸O, in agreement with the mass spectral results described above. Related iron(IV) complexes [Fe^IV(O)(TMC)(NCCH₃)]²⁺ and [Fe^IV₂(μ-O)₂(N,N′-bis(2-pyridylmethyl)-N,N′-dimethyl-trans-1,2-diaminocyclohexane, BPMCN)₂]⁴⁺ are also comparably inert to exchange with solvent water (17, 21), as are heme peroxidase compounds II at alkaline pH (39). In contrast, exchange of solvent water is facile for peroxidase compounds II at pH 7 because of hydrogen bonding to the oxo atom (39) and for [Fe^IV(O)(porphyrin radical)]⁺ complexes via oxo-hydroxo tautomerism (40–43).

Finally, we point out that 2 forms nearly quantitatively from the reaction of 1 with the addition of only 1 equiv. of peracid. This is analogous to the reaction of [Fe^II(meso-tetramesitylporphinate dianion, TMP)] with m-chloroperbenzoic acid to form [Fe^IV(O)(TMP)] (44). Given the observed stoichiometry, the most plausible mechanism for the formation of 2 involves heterolytic cleavage of a bound peracid anion (Scheme ). This reaction course differs from that observed for the reactions of 1 with other peroxides such as H₂O₂ or ROOH, which require at least 1.5 equiv. In these cases, either an Fe^III—OOH or Fe^III—OOR intermediate is observed instead (24, 27, 45).

Comparisons with Other Iron(IV)-Oxo Complexes.

Although oxo-iron(IV) porphyrin complexes have been extensively investigated as models for intermediates in heme enzymes, only very recently have corresponding complexes with nonporphyrin ligands been observed, providing us with an unprecedented opportunity to investigate the electronic properties of the Fe^IVO unit without complications from a supporting porphyrin ligand. Table 1 compares the properties of 2 with those of three other nonheme iron(IV) complexes with S = 1 sites: the crystallographically characterized [Fe^IV(O)(TMC)(NCCH₃)](OTf)₂ (21), the metastable [Fe^IV(O)(cyclam-acetate)]⁺, which can be generated in only ≈25% yield at −80°C (20), and the dinuclear [Fe^IV₂(μ-O)₂(BPMCN)₂](OTf)₄ (17). Both 2 and [Fe^IV(O)(TMC)(NCCH₃)](OTf)₂ (and perhaps the cyclam-acetate complex as well) exhibit near-IR bands of comparable intensity, suggesting that this transition may be characteristic of a mononuclear S = 1 Fe^IVO center. The low intensities (ɛ ≈ 300–400 M⁻¹⋅cm⁻¹) of these bands make an oxo-to-iron(IV) charge transfer assignment unlikely and suggest instead a ligand field transition. The latter assignment would be consistent with our inability thus far to observe a resonance-enhanced Fe^IVO Raman vibration by excitation into this band. In contrast, [Fe^IV₂(μ-O)₂(BPMCN)₂](OTf)₄ has bands in the same spectral region that are 1 order of magnitude more intense, implicating a significant change in the electronic structure of the metal-oxo unit on dimerization of the Fe^IVO unit to form an Fe^IV₂(μ-O)₂ diamond core. Indeed the related [Fe₂(μ-O)₂(5-Et₃-TPA)₂](ClO₄)₃ complex, which has a valence-delocalized Fe^IIIFe^IV(μ-O)₂ diamond core, also exhibits low energy bands of comparable intensity (12), which are proposed to arise from primarily metal-to-metal transitions (A. J. Skulan, M. A. Hanson, and E. I. Solomon, personal communication). In-depth studies of the nonheme iron(IV)-oxo complexes of Table 1 like those carried out on [Fe₂(μ-O)₂(5-Et₃-TPA)₂](ClO₄)₃ will be required to identify the nature of the near-IR transition(s).

The mononuclear oxo-iron(IV) TPA and TMC complexes can also be distinguished from the dinuclear BPMCN complex by their intense x-ray absorption 1-s → 3-d preedge transitions (Fig. 5) and short Fe—O bonds (Table 1), features that stem from the presence of the terminal Fe^IVO bond. Interestingly however, these three Fe^IV complexes, and the cyclam-acetate complex as well, exhibit remarkably similar spin Hamiltonian parameters, obtained from Mössbauer analysis, despite differences in the nature of the supporting ligands and the nuclearity of the complexes. Similar parameters are also found for compounds I and II of heme peroxidases and synthetic S = 1 oxo-iron(IV) porphyrin and porphyrin radical complexes (46). These parameters can be related to the more fundamental crystal field description of the Fe^IV site by Oosterhuis and Lang (47), who have considered the t configuration of Fe^IV in which the one-electron orbitals |xy〉, |xz〉, and |yz〉 are mixed by spin-orbit coupling. In their model the variables are the tetragonal, Δ, and rhombic, V, components of the ligand field and the one-electron spin-orbit coupling constant, ζ ≈ 400 cm⁻¹. The D value of 2, D ≈ 28(2) cm⁻¹, according to figure 3 of Oosterhuis and Lang (47), suggests that Δ/ζ ≈ 4.5. The resulting small tetragonal splitting predicts an orbital contribution to the magnetic hyperfine tensor that would reduce A_x,y to ≈ −16 T for 2, that is, to values substantially below those obtained experimentally, namely A_x,y ≈ −23.5 T. Similar problems arise for other S = 1 Fe^IV complexes with large zero-field splittings such as compound I of chloroperoxidase (D ≈ 37 cm⁻¹, ref. 48) and some oxo-iron(IV) porphyrin complexes (D ≈ 25 cm⁻¹, ref. 46). These observations suggest that the simple theory has to be amended, most likely by inclusion of low-lying S = 2 states (49). [Preliminary density functional theory calculations (A.S., V. Vrajmasu, and E.M., unpublished results) indicate that the first excited state of 2, obtained for a geometry optimized structure, is indeed S = 2, suggesting immediately that the Oosterhuis and Lang model (47) is not applicable. Presently, however, we do not know whether mixing with this state will produce the right magnitude for D. Moreover, other contributions to D, described in a comprehensive analysis by Neese and Solomon (49), have to be explored as well.] Such excited states can be admixed into the ground state by spin-orbit coupling, and this admixture would affect the D value but would not lead to a reduction of A_x and A_y. It has recently been shown that 40% of the zero-field splitting of the S = 2 state of Fe^II rubredoxin is attributable to mixing with excited S = 1 configurations and that such mixing does not affect the A tensor (50).

Lastly, a comparison of reactivity properties of the iron(IV) complexes shows that the supporting ligand does in fact play an important role in modulating the stability of the iron(IV)-oxo unit. At −40°C the TPA complex is stable for several days, whereas the TMC complex persists for at least a month; their estimated half-lives at 10°C are 1 h and 1 d, respectively. The analogous tetramesitylporphin complex [Fe^IV(O)(TMP)], reported to be stable in benzene for a few minutes at room temperature, falls within this stability range (51). In contrast, [Fe^IV₂(μ-O)₂(BPMCN)₂](OTf)₄ can be obtained only at −80°C and decomposes readily on warming (17). The cyclam-acetate complex appears similarly unstable, but its behavior is not well characterized (20). The relative stabilities of the TPA, TMC, and BPMCN complexes are inversely correlated with their oxidative reactivity, that is, the less stable iron(IV) complexes are the stronger oxidants. The most stable of the three complexes, [Fe^IV(O)(TMC)(NCCH₃)](OTf)₂, can transfer its oxygen atom to PPh₃ at −40°C but not to thioanisole (21). In contrast, 2 oxidizes thioanisole readily at −40°C and epoxidizes cyclooctene. Complex 2 is thus even more reactive than its porphyrin (TMP) analogue, which is stable in the presence of olefins <0°C (51). [Fe^IV₂(μ-O)₂(BPMCN)₂](OTf)₄, the least stable of the three, is able to oxidize the C—H bonds of adamantane in nearly quantitative yield on warming to −40°C (17). Although the mechanisms and energetics of these reactions remain to be investigated in detail, this comparison emphasizes a key point, namely that the ligand, and perhaps the core nuclearity, can play a large role in controlling the reactivity of the Fe^IVO moiety. These iron(IV)-oxo complexes can thus serve as a useful starting point for understanding the mono- or di-iron(IV) oxidation chemistry that is proposed to occur in the nonheme active sites of monoiron enzymes such as pterin-dependent phenylalanine hydroxylase and α-ketoglutarate-dependent prolyl hydroxylase and diiron enzymes such as alkane monooxygenases and fatty acid desaturases (10, 11).

Abbreviations

• EXAFS, extended x-ray absorption fine structure

• BPMCN, N,N′-bis(2-pyridylmethyl)-N,N′-dimethyl-trans-1,2-diaminocyclohexane

• cyclam-acetate, 1-carboxymethyl-1,4,8,11-tetraazacyclotetradecane

• OTf, trifluoromethylsulfonate or triflate anion

• TMC, 1,4,8,11-tetramethyl-1,4,8,11-tetraazacyclotetradecane or tetra(N-methyl)cyclam

• TMP, meso-tetramesitylporphinate dianion

• TPA, tris(2-pyridylmethyl)amine

• XAS, x-ray absorption spectroscopy