Isolation and Characterization of a Dihydroxo-Bridged Iron(III,III)(μ-OH)₂ Diamond Core Derived from Dioxygen

Abstract

Dioxygen addition to coordinatively unsaturated [Fe(II)(O^Me2N₄(6-Me-DPEN))](PF₆) (1) is shown to afford a complex containing a dihydroxo-bridged Fe(III)₂(μ-OH)₂ diamond core, [Fe^III(O^Me2N₄(6-Me-DPEN))]₂(μ-OH)₂(PF₆)₂•(CH₃CH₂CN)₂ (2). The diamond core of 2 resembles the oxidized methane monooxygenase (MMOox) resting state, as well as the active site product formed following H-atom abstraction from Tyr-OH by ribonucleotide reductase (RNR). The Fe-OH bond lengths of 2 are comparable with those of the MMOHox suggesting that MMOHox contains a Fe(III)₂(μ-OH)₂ as opposed to Fe(III)₂(μ-OH)(μ-OH₂) diamond core as had been suggested. Isotopic labeling experiments with ¹⁸O₂ and CD₃CN indicate that the oxygen and proton of the μ-OH bridges of 2 are derived from dioxygen and acetonitrile. Deuterium incorporation (from CD₃CN) suggests that an unobserved intermediate capable of abstracting a H-atom from CH₃CN forms en route to 2. Given the high C–H bond dissociation energy (BDE= 97 kcal/mol) of acetonitrile, this indicates that this intermediate is a potent oxidant, possibly a high-valent iron oxo. Consistent with this, iodosylbenzene (PhIO) also reacts with 1 in CD₃CN to afford the deuterated Fe(III)₂(μ-OD)₂ derivative of 2. Intermediates are not spectroscopically observed in either reaction (O₂ and PhIO) even at low-temperatures (−80 °C), indicating that this intermediate has a very short life-time, likely due to its highly reactive nature. Hydroxo-bridged 2 was found to stoichiometrically abstract hydrogen atoms from 9,10-dihydroanthracene (C-H BDE= 76 kcal/mol) at ambient temperatures.

Introduction

The activation of O₂ is involved in many fundamental biochemical processes promoted by iron metalloenzymes. These transformations typically involve reactive iron-superoxo, -peroxo or high-valent oxo intermediates,^1–12 which are, in some cases, capable of activating strong C–H bonds.^1,13–15 For example, cytochrome P450 binds and subsequently activates O₂ en route to an Fe(IV)-OH porphyrin radical cation intermediate (Compound I).^13,15 Non-heme iron enzymes, such as α-ketoglutarate taurine dioxygenase (TauD)^11,16 and halogenase SyrB2,¹⁷ activate O₂ to form high-valent Fe(IV)-oxo intermediates. The reduced binuclear Fe(II)₂ active site of methane monooxygenase (MMO)^1,10,18–21 and ribonucleotide reductase (RNR)^10,22–25 react with O₂ to afford highly reactive oxidized binuclear Fe(IV)₂(μ-O)₂ (A) or M(III)(μ-O)(μ-OH)M'(IV) (M, M'= Mn, Fe; B) diamond cores^26–28, respectively, (Scheme 1) which abstract H-atoms from either CH₄ or TyrOH, respectively, to afford Fe(III)(μ-OH)(μ-O)Fe(IV) (B) or M(III)(μ-OH)₂M'(III) (C) species.^{3,19,21,25,29,30} The highly reactive nature of these enzymatic intermediates has prompted numerous investigations aimed at establishing benchmark structural, spectroscopic, and reactivity properties of these species.^{1,4,10–12,16,18–28,31–36} With RNR, the introduction of an electron along with O₂ affords structure B (Scheme 1), which in Chlamydia tranchomatis has been shown to contain an oxo/hydroxo bridged Fe(III)Mn(IV) dimer based on Mn and Fe K-edge EXAFS and DFT calculations.²⁵ It's likely that the diiron RNRs have an analogous Fe(III)Fe(IV) core B that converts to core C upon reaction with Tyr-OH substrate (eqn (1)). The significantly stronger C-H bond of CH₄ requires the higher valent core A (Scheme 1) of methane monooxygenase (MMO) to abstract a H-atom. It has been debated as to whether the structure of the MMO oxidized resting state (MMOH_ox) contains a bis-hydroxo bridged Fe(III)(μ-OH)₂Fe(III) (C) or a hydroxo/aquo bridged Fe(III)(μ-OH)(μ-OH₂)Fe(III) (D) core.^21,37–39 One would expect the metal ion Lewis acidity of Fe(III) and Fe(IV) to disfavor protonation of a hydroxide especially if it is bridging. Water has been observed to bridge between two less Lewis acidic Fe(II) ions, however.⁴⁰

Scheme 1.

Many of the insights regarding the reactivity structures and properties of these oxidized bimetallic diamond cores have been provided by synthetic analogues.^4,12,31–36 With synthetic complexes, solubility in non-aqueous solvents with low freezing points and small molecule crystallography can help provide mechanistic details, such as the probable identity and spectroscopic properties of reaction intermediates^18,41–45 or the location of and transfer of key protons.⁴⁶ Research efforts in our group have recently focused on the O₂ reactivity of biologically-relevant first-row transition metal complexes.^47,48 Herein we describe the reactivity of a new coordinatively unsaturated Fe(II) complex with O₂ to afford a dihydroxo-bridged binuclear Fe(III) complex. Spectroscopic evidence has been obtained which suggests that an unobserved intermediate formed during this reaction abstracts a hydrogen atom from CH₃CN solvent.

Experimental Section

General Methods

All manipulations were performed using Schlenk line techniques or under a N₂ atmosphere in a glovebox. The highest purity reagents and solvents available were purchased and used without further purification unless otherwise noted. CH₃OH and CH₃CN (CaH₂) were dried and distilled prior to use. Et₂O was purified using solvent purification columns housed in a custom stainless steel cabinet and dispensed by a stainless steel schlenk-line (GlassContour). CD₃CN and CD₂Cl₂ were purchased from Cambridge Isotope Laboratories, Inc. CD₃CN was either dried over CaH₂, vacuum transferred and stored in a dry box, or obtained from sealed ampules (Cambridge Isotopes (99.8%)) containing volumes small enough for a single experiment. All solvents were rigorously degasses prior to use. ¹H NMR spectra were recorded on either a Bruker AV 301 or Bruker AV 300 FT NMR spectrometer at ambient temperature and were referenced to residual deuterated solvent. Infrared spectra were recorded on NaCl salt plates as Nujol mulls with a Perkin Elmer 1720 FT-IR spectrometer. Electrospray-ionization mass spectra were obtained on a Bruker Esquire Liquid Chromatograph-Ion Trap mass spectrometer. Gas chromatography – mass spectrometry data was recorded using an HP 5971A gas chromatograph interfaced with a quadrupole mass spectrometer equipped with a 7673A autosampler. Electronic absorption spectra were recorded on a Varian Cary 50 spectrophotometer equipped with a fiber optic cable connected to a “dip” ATR probe (C-technologies). A custom-built two-neck solution sample holder equipped with a threaded glass connector was sized specifically to fit the “dip” probe. Cyclic voltammograms were recorded in CH₃CN (100 mM (ⁿBu₄N)(PF₆) supporting electrolyte) on a PAR 263A potentiostat with a glassy carbon working electrode, Ag/AgNO₃ reference electrode, and a platinum auxiliary electrode. EPR spectra were recorded on a Varian CW-EPR spectrometer at 15 K equipped with an Oxford helium cryostat. Solution magnetic moments were calculated by the Evans method,⁴⁹ with temperature correction made in the manner described by Van Geet.⁵⁰ Pascal's constants were used to correct for diamagnetic contributions to the experimental magnetic moment. X-ray crystallography data was recorded on a Bruker APEX II single crystal X-ray diffractometer with Mo-radiation. Elemental analyses were performed by Atlantic Microlabs, Norcross, GA. Experiments involving iodosylbenzene were conducted using a 2 mM stock solution in CH₃OH. N,N-bis(6-methyl-2-pyridilmethyl)ethane-1,2-diamine (6-Me-DPEN) was prepared according to literature procedures.⁴⁸

Synthesis of [Fe^II(O^Me2N₄(6-Me-DPEN))](PF₆) (1)

FeCl₂ (0.094 g, 0.74 mmol), N,N-bis(6-methyl-2-pyridilmethyl)ethane-1,2-diamine (0.20 g, 0.74 mmol), 3-hydroxy-3-methyl-2-butanone (0.09 g, 0.86 mmol), NaOCH₃ (0.40 g, 0.74 mmol), and NaPF₆ (0.12 g, 0.74 mmol) were combined in CH₃OH (10 mL) and allowed to stir under an inert atmosphere in a glovebox at room temperature for two days. All volatiles were then removed in vacuo to yield a crude yellow solid, which was recrystallized from CH₃CN/Et₂O at 258 K to yield the title compound as a bright yellow solid in 47 % yield (0.20 g, 0.35 mmol). Electronic absorption spectrum: λ_max (nm) (ε (M⁻¹cm⁻¹)) (CH₃CN): 412 (250). Redox potential (CH₃CN vs. Fc^+/0, 298 K): E_pa = +230 mV, E_pc = +10 mV, ΔE = 220 mV. Magnetic moment (299 K, CD₃OD solution): 4.71 B.M. ESI-MS: Expected m/z for [C₂₁H₂₉N₄OFe]⁺ = 409.3, found m/z = 409.2. Elemental analysis results were obtained with the BPh₄⁻ salt of 1, which was synthesized by the same procedure described above except that NaBPh₄ was used instead of NaPF₆. Elemental analysis for C₄₅H₄₉N₄BOFe; Calculated C, 72.06; H, 7.16; N, 6.86. Found C, 71.11; H, 7.60; N, 6.41.

Synthesis of [Fe^III(O^Me2N₄(6-Me-DPEN))]₂(μ-OH)₂(PF₆)₂•(CH₃CH₂CN)₂ (2)

Complex 1 (0.20 g, 0.36 mmol) was dissolved in CH₃CN (5 mL) under an inert atmosphere in a glovebox at room temperature. The resulting solution was placed into a gas tight syringe, removed from the glovebox, and injected into a custom-made two-neck vial equipped with a septum cap that had previously been purged with argon and contained a stir bar. Dioxygen, from a pressurized tank (Praxair, 99.9999%, <0.5 ppm H₂O) and run through a column of anhydrous CaSO₄ (Drierite), was gently bubbled into the stirring solution for approximately five minutes. The reaction mixture was allowed to continue stirring at room temperature for two hours, during which the solution had turned from yellow to dark orange. All volatiles were then removed in vacuo to yield a dark orange solid, which was recrystallized from CH₃CN/Et₂O (1/6 vol/vol) at 253 K to afford the title compound in 96 % yield (0.19 g, 0.17 mmol). X-ray quality crystals of 2 were grown from a concentrated CH₃CH₂CN solution at −80 °C, as well as from a concentrated CH₃CN solution at room temperature. FT-IR (NaCl plates, Nujol mull, 298 K): ν(O-H) = 3610 cm⁻¹; ν(O-D) = 2730 cm⁻¹. Elemental analysis for C₄₂H₅₈N₈O₄F₁₂P₂Fe₂; Calculated C, 44.50; H, 5.69; N, 10.61. Found C, 45.84; H, 5.33; N, 10.40.

Characterization of 2 in CH₃CN solution

Electronic absorption spectrum: λ_max (nm) (ε (M⁻¹cm⁻¹)): 310 (610), 375 (br, 440) (extinction coefficients calculated per Fe(III)). Redox potential (CH₃CN vs. Fc^+/0, 298 K): E_pa = +235 mV, E_pc = +26 mV. Magnetic moment (300 K, CD₃CN solution): 5.82 B.M/Fe(III). ESI-MS: Expected m/z for [C₂₁H₃₀N₄O₂Fe]⁺ = 426.3, found m/z = 426.2. Expected m/z for [C₂₁H₂₉DN₄O₂Fe]⁺ = 427.3, found m/z = 427.2. Expected m/z for [C₂₁H₃₀N₄O¹⁸OFe]⁺ = 428.3, found m/z = 428.2. EPR spectrum (1:1 CH₃CN/toluene glass, 14 K): g = 1.96, 4.07, 10.2.

Synthesis and Isolation of Deuterium Labeled [Fe^III(O^Me2N₄(6-Me-DPEN))]₂(μ-OD)₂(PF₆)₂•(CH₃CH₂CN)₂ (2) for FT-IR Experiments

Method 1

A solution of 1 (0.25 g, 0.45 mmol) was prepared in dry CD₃CN (30 mL; Cambridge Isotopes (99.8%)) under an inert atmosphere in a glovebox. The resulting solution was placed into a gas tight syringe, removed from the glovebox, and injected into a custom-made two-neck vial equipped with a septum cap that had previously been purged with argon and contained a stir bar. Dioxygen, derived from a pressurized tank (Praxair, 99.9999%, <0.5 ppm H₂O) and run through a column of anhydrous CaSO₄ (Drierite), was gently bubbled into the stirring solution for approximately five minutes. The reaction was allowed to continue to stir at room temperature for 30 minutes, after which all volatiles were removed in vacuo to afford 2 as a dark orange solid. The resulting solid was redissolved in a minimal amount of CH₃CH₂CN and crystallized at −80 °C. Crystalline samples of 2 were harvested from the mixture via filtration through a fine fritted filter, briefly dried in vacuo at room temperature, and subsequently used for FT-IR experiments.

Method 2

A solution of 2 (0.090 g, 0.16 mmol) was prepared in dry CH₃CN (10 mL) under aerobic conditions at room temperature. D₂O (0.1 mL) was added to this solution, and the resulting reaction mixture was allowed to gently stir at room temperature overnight. All volatiles were then removed in vacuo to afford a dark orange solid, which was left under vacuum for 72 hours. The solid was then redissolved in a minimal amount of CH₃CH₂CN, crystallized at −80 °C, harvested via filtration, and subsequently used for FT-IR experiments.

Analysis of Products From Reactions Between Complex 2 and 9,10-dihydroanthracene in CH₃CN

In a typical reaction, 2 was dissolved in CH₃CN (3–4 mL, ~0.5 mM solution of 2) at room temperature under an inert atmosphere in a glovebox. The resulting solution was placed into a gas tight syringe, removed from the glovebox, and injected into a custom-made two-neck vial containing a stir bar and equipped with a septum cap and threaded dip-probe feed-through adaptor that had previously been purged with Ar. 9,10-dihydroanthracene (0.5 equivalents per Fe(III)) was then added to the solution in a similar fashion from a concentrated stock solution that had been prepared in CH₃CN in a glovebox. Each reaction mixture (the reaction was repeated three times and shown to be reproducible in all three cases) was allowed to stir under anaerobic conditions for 10–15 minutes, and then dried in vacuo. The solid products were then washed with Et₂O (3 mL) and the Et₂O soluble components were subsequently analyzed by GC/MS and quantified using standard calibration curves.

X-Ray Crystallography

A yellow crystal of 1 with dimensions 0.05 × 0.04 × 0.4 mm³ was mounted on a glass capillary with oil. Data was collected at −143 °C. The crystal-to-detector distance was set to 30 mm and exposure time was 60 seconds per degree for all sets of exposure. The scan width was 2.0°. Data collection was 99.4% complete to 25.0° in ϑ. A total of 56,116 partial and complete reflections were collected covering the indices, h = −21 to 21, k = −19 to 19, l = −32 to 32. 69,506 reflections were symmetry independent and the R_int = 0.125 indicated that the data was poor. Indexing and unit cell refinement indicated an orthorhombic P lattice with the space group P b c a (No. 61).

A pale red plate of 2 with dimensions 0.20 × 0.10 × 0.10 mm³ was mounted on a glass capillary with oil. Data was collected at −173 °C. The crystal-to-detector distance was set to 40 mm and exposure time was 30 seconds per degree for all sets of exposure. The scan width was 0.5°. Data collection was 99.9% complete to 25.0° in ϑ. A total of 51,770 partial and complete reflections were collected covering the indices, h = −15 to 14, k = −15 to 15, l = −17 to 17. 7,264 reflections were symmetry independent and the R_int = 0.0512 indicated that the data was good. Indexing and unit cell refinement indicated a triclinic P lattice with the space group P -1 (No. 2).

All data was integrated and scaled using SAINT, SADABS within the APEX2 software package by Bruker. Solutions by direct methods (SHELXS, SIR97) produced complete heavy atom phasing models which were consistent with each proposed structure. Scattering factors are from Waasmair and Kirfel.⁵¹ Crystal data for 1 and 2 are provided in Table 1. All non-hydrogen atoms were refined anistropically by full-matrix least-squares methods. Most hydrogen atoms were placed using a riding model, however the hydrogen atoms on the bridging hydroxide ligands in 2 were refined from the difference map. Crystallographic data is summarized in Table 1. Selected distances and angles are summarized in Table 2.

Table 1.

Crystal data for [Fe^II(O^Me2N₄(6-Me-DPEN))](PF₆) (1) and [Fe^III(O^Me2N₄(6-Me-DPEN))]₂(μ-OH)₂(PF₆)₂ •(CH₃CH₂CN)₂ (2).

	1	2
Formula	C21H29F6FeN4OP	C48H70F12Fe2N10O4P2
MW	553.40	1252.78
T, K	130(2)	100(2)
Unit Cella	Orthorhombic	Triclinic
a, Å	15.9090(11)	11.237(2)
b, Å	17.1316(12)	11.403(2)
c, Å	17.6610(13)	13.334(2)
α, deg	90	70.554(9)
β, deg	90	66.281(9)
γ, deg	90	87.391(10)
V, Å3	4813.3(6)	1467.3(5)
Z	8	1
d(calc), g/cm3	1.530	1.418
Space group	P b c a	P −1
Rb	0.0592	0.0366
Rwc	0.1548	0.0796
GOF	0.928	1.007

Table 2.

Selected Bond Distances (Å) and Bond Angles (deg) for [Fe^II(O^Me2N₄(6-Me-DPEN))](PF₆) (1) and [Fe^III(O^Me2N₄(6-Me-DPEN))]₂(μ-OH)₂(PF₆)₂ •(CH₃CH₂CN)₂ (2).

	1	2
Fe(1)-O(1)	1.925(3)	1.8721(13)
Fe(1) -N(1)	2.086(4)	2.1121(15)
Fe(1) -N(2)	2.273(5)	2.2556(15)
Fe(1) -N(3)	2.139(4)	2.2144(16)
Fe(1) -N(4)	2.121(4)	N/A
Fe(1)–O(2)	N/A	1.9715(13)
Fe(1')–O(2)	N/A	2.0037(14)
Fe(1)•••Fe(1')%	N/A	3.172
O(1)- Fe -N(1)	78.97(16)	78.34(6)
O(1)- Fe -N(2)	154.82(15)	148.61(5)
O(1)- Fe -N(3)	117.97(16)	98.17(6)
O(1)- Fe -N(4)	116.24(16)	N/A
N(1)- Fe -N(3)	116.03(17)	107.26(6)
N(1)- Fe -N(4)	122.64(17)	N/A
N(3)- Fe -N(4)	104.41(15)	N/A
O(1)– Fe–O(2)	N/A	106.24(5)
O(1)– Fe –O(2')	N/A	104.88(6)
N(1)– Fe –O(2)	N/A	163.94(6)
Fe(1)–O(2)–Fe(1')	N/A	105.88(6)
O(2)-Fe(1)-O(2#)	N/A	74.12(6)

^%Fe(1) and Fe(1') and O(2) and O(2') are equivalent in 2 and related by a crystallographically imposed inversion center, ie only one half of the dimer is crystallographically independent.

Results and Discussion

Synthesis and Structure of Coordinatively Unsaturated Fe(II) Alkoxide Complex

The alkoxide-ligated complex [Fe^II(O^Me2N₄(6-Me-DPEN))](PF₆) (1) was synthesized via an Fe(II)-templated Schiff-base condensation between 6-Me-DPEN⁴⁸ and 3-hydroxy-3-methyl-2-butanone in CH₃OH under an inert atmosphere at ambient temperature. X-ray quality crystals of 1 were obtained via vapor diffusion of Et₂O into a saturated CH₃CN solution at room temperature. As shown in the ORTEP diagram of Figure 1, the Fe(II) ion of 1 is found in a distorted trigonal bipyramidal coordination environment (τ = 0.61)⁵² comprised of an alkoxide oxygen (O(1)), imine nitrogen (N(1)), tertiary amine (N(2)), and two pyridine nitrogens (N(3) and N(4)) (Figure 1). Coordinatively unsaturated 1 is high-spin (S = 2, μ_eff = 4.71 B.M. in CD₃OD) and exhibits a quasi-reverisble Fe^III/II redox couple in CH₃CN (under nitrogen) with E_1/2 = +120 mV vs. Fc^+/0 (ΔE = 220 mV, Figure S-1). The room temperature electronic absorption spectrum of 1 in CH₃CN contains a single absorption band with λ_max = 412 nm (ε = 250 M⁻¹cm⁻¹, Figure S-2).

Figure 1.

ORTEP diagram (50% probability) of [Fe^II(O^Me2N₄(6-Me-DPEN))]⁺ (1) with hydrogen atoms and counterion omitted for clarity.

Reactivity with Dioxygen to Afford Dihydroxo-Bridged Iron(III,III)(μ-OH)₂ Diamond Core

If excess amounts of dry O₂, derived from a pressurized tank (Praxair, 99.9999%, <0.5 ppm H₂O) and dried via a column of anhydrous CaSO₄ (Drierite), is bubbled into a CH₃CN solution of 1 for ~2–3 minutes at ambient temperature the solution changes from yellow to dark orange, and a broad band at λ_max = 310 nm (Figure S-3) grows in in the electronic absorption spectrum, and the broad band at 412 nm (Figure S-2) disappears. Low temperature reactions were also performed at either −40 °C in CH₃CN or −80 °C in CH₃CH₂CN, however no transient intermediates were observed under these conditions. X-ray quality crystals of the orange product obtained from this reaction, [Fe^III(O^Me2N₄(6-Me-DPEN))]₂(μ-OH)₂(PF₆)₂ •(CH₃CH₂CN)₂ (2), were obtained in high yield (96 %) via vapor diffusion of Et₂O into a concentrated CH₃CH₂CN solution at −80 °C (Figure S-4). As shown in the ORTEP diagram of Figure 2, both Fe(III) ions of 2 are located in a distorted octahedral environment consisting of an alkoxide oxygen (O(1)), two hydroxide bridging ligands (O(2) and O(2')), an imine nitrogen (N(1)), a tertiary amine (N(2)), and a pyridine nitrogen (N(3)) (Figure 2). The location and refinement of a hydrogen atom on each of the bridging oxygen atoms (O(2)) in the X-ray structure, along with the presence of two PF₆⁻ counterions per diiron complex, conclusively identified the bridging ligands of 2 as hydroxides. Hydrogen-bonding interactions ((O(2)-H⋯N(21)) = 2.200 Å; O(2)-H-N(21)= 171.29°) between the bridging hydroxide ligands and co-crystallized CH₃CH₂CN nitrogens (Figure S-4) further verified this assignment.

Figure 2.

ORTEP diagram (50% probability) of the Fe(III)(μ-OH)₂Fe'(III) diamond core and Fe coordination sphere of dicationic{[Fe^III(O^Me2N₄(6-Me-DPEN))]₂(μ-OH) ₂²⁺ (2). The two halves of the molecule are related by a crystallographic inversion center. An ORTEP diagram of the entire dicationic complex is shown in Figure S-4.

The hydroxide ligands in 2 are asymmetrically bridging with Fe(1)-O(2) and Fe(1)-O(2') bond lengths of 1.9715(13) Å and 2.0037(14) Å, respectively (Table 2). These distances are comparable to other structurally characterized Fe(III)₂(μ-OH)₂ and Fe(III)₂ (μ-OH)₂ (μ-O₂CR) complexes (1.88–2.08 Å).^{35,36,53–64} The Fe(1)⋯Fe(1') distance in 2 (3.172 Å) is, on the other hand, longer than the reported Fe(III)₂(μ-OH)₂ core range (3.07–3.12 Å), while the Fe(1)-O(2)-Fe(1') bridging angle 105.88(6)° is at the larger end of the previously reported range (92.3°–105.8°).^53–63 This likely results from steric congestion around each of the two Fe(III) ions, as is further indicated by the dissociation of a pyridine arm away from each metal ion in both halves of the dimer (Figure S-4).

Isotopic Labeling Experiments

Isotopic labeling studies conducted with ¹⁸O₂ confirmed that the two bridging hydroxide ligands in 2 are indeed derived from O₂ (Scheme 2; Figure S-5–Figure S-7). As shown by the Fe-isotope distribution pattern in the ESI mass spectra of Figure 3a, the μ-hydroxo bridge cleaves in the mass spectrometer to afford a monomeric monocationic Fe-hydroxo species with a parent ion m/z peak at 426.2 (Figure S-5), which corresponds to half of the {[Fe^III(O^Me2N₄(6-Me-DPEN))]₂(μ-OH)₂}²⁺ dimer 2, [Fe^III(O^Me2N₄(6-Me-DPEN))(OH)]⁺, and is 17 mass units greater than that (m/z= 409.2) of 1 (Figure S-6). The parent ion peak (at m/z= 426.2; Figure 3a) shifts by two mass units (to m/z= 428.2; Figure 3c) upon the incorporation of ¹⁸O from ¹⁸O-labeled dioxygen (Figure S-7). In order to determine the source of the hydrogen atoms found on the hydroxide bridging ligands in 2, deuterium labeling studies were conducted. If O₂ is added to 1 in deuterated acetonitrile (CD₃CN), then a parent ion peak at m/z= 427.2 (Figure 3b), which is shifted by one mass unit (Figure S-8) relative to that observed at m/z= 426.2 (Figure 3a) in the CH₃CN reaction (Figure S-5), is observed consistent with the incorporation of deuterium. Again, the Fe-isotope distribution pattern indicates that the bis-hydroxo bridged diiron complex cleaves in the mass spectrometer to afford a monomeric [Fe^III(O^Me2N₄(6-Me-DPEN))(OD)]⁺. Although one can not rule out some sort of impurity in the CD₃CN as the source of deuterium, it is unlikely that D₂O from “wet” CD₃CN is responsible, given that the ¹⁸O-label in the spectrum of Figure 3c is not “washed out” via exchange with water (H₂¹⁶O) contamination in the CH₃CN. It is also unlikely that D₂O, as opposed to H₂O, would be a contaminant in the CD₃CN. Deuterium labeled 2 was also obtained from a reaction between 1 and O₂ in CD₂Cl₂. Additional evidence to support the incorporation of deuterium from CD₃CN is shown by the vibrational spectrum (Figure S-9) of crystalline 2 (synthesized in CH₃CN), which displays a sharp feature at 3610 cm⁻¹ that shifts to 2730 cm⁻¹ when 2 is synthesized in CD₃CN. The energies and observed isotopic shift (ν_O-H/ν_O-D = 1.32 vs. theoretical = 1.35) for these vibrational features are consistent with O-H and O-D stretches, respectively.

Scheme 2.

Figure 3.

ESI-MS of the cleaved monomeric product observed when dicationic {[Fe^III(O^Me2N₄(6-Me-DPEN))]₂(μ-OH)₂²⁺ (2), obtained from (a) the reaction between 1 and ¹⁶O₂ in CH₃CN, (b) the reaction between 1 and ¹⁶O₂ in CD₃CN, and (c) the reaction between 1 and ¹⁸O₂ in CH₃CN, is placed in the mass spectrometer.

Implications Regarding the Involvement of a Reactive Intermediate

Deuterium incorporation (from CD₃CN) suggests that an unobserved intermediate capable of abstracting a H-atom from CH₃CN forms in the dioxygen reaction (Scheme 2) en route to 2. Given the high C–H bond dissociation energy (BDE= 97 kcal/mol) of acetonitrile, this indicates that this intermediate is a potent oxidant, possibly a high-valent iron oxo. Consistent with this, iodosylbenzene (PhIO) also reacts with 1 in CD₃CN (Scheme 3) to afford the deuterated Fe(III)₂(μ-OD)₂ derivative of 2 (93–96 % yield), as verified by FT-IR (ν_O-D = 2727 cm⁻¹; Figure S-11) and ESI-MS (m/z= 427.2; Figure S-10). Intermediates are not spectroscopically observed in either reaction (O₂ and PhIO) even at low-temperatures (−80 °C), indicating that this intermediate has a very short life-time, likely due to its highly reactive nature. Iodosylbenzene (PhIO) would be expected to afford a Fe(IV)=O upon its reaction with Fe(II) (Scheme 3).^65–69 High-valent Fe(IV)=O species have been shown to be highly reactive and capable of abstracting H-atoms, especially in the presence of anionic ligands.^68–73 Hydrogen atom abstraction from acetonitrile has been previously observed during the O₂-promoted oxidation of [Mn(II)1^cyp]^1- to [Mn(III)1^cyp(OH)]^1- via a high-valent Mn-oxo intermediate.⁷⁴ High-valent iron oxo compounds [Fe(IV)(O)BnTPEN)]²⁺,⁷⁵ [Fe(IV)(PY5)(O)]²⁺,⁷⁶ and aqueous Fe(IV)O²⁺ have also been shown to abstract hydrogen atoms from CH₃CN.⁷⁷

Scheme 3.

Reactivity of Dihydroxo-Bridged 2 with Substrates Containing Weaker C-H Bonds

The reactivity of the dihydroxo-bridged product 2 towards stoichiometric amounts of substrates containing relatively weak C-H bonds was explored under anaerobic conditions in CH₃CN at room temperature. Under these conditions, 9,10-dihydroanthracene (C-H BDFE = 76.0 kcal·mol⁻¹)⁷⁸ was oxidized to anthracene in 23(3) % yield (Figures S-12, S-13). Complex 2 was unreactive, however, with organic compounds containing stronger C-H bonds, such as toluene and cyclohexane (C-H BDFE = 87 kcal·mol⁻¹ and 99 kcal·mol⁻¹, respectively).⁷⁸ This reactivity is similar to other Fe(III)-OH species, such as lipoxygenase^79,80 and [Fe^III(PY5)(OH)]²⁺,⁷⁶ which have been found to abstract hydrogen atoms from substrates containing relatively weak C-H bonds.

Conclusions

In conclusion, we have described the synthesis and characterization of a new coordinatively unsaturated Fe(II) complex (1) that reacts with O₂ to form an Fe(III)₂(μ-OH)₂ diamond core (2) in high yield. Isotopic labeling studies suggest that an unobserved intermediate formed during the conversion of 1 to 2 is capable of oxidizing CH₃CN, presumeably via H^• atom transfer. Iodosylbenzene (PhIO, 1.0 eq) also oxidizes 1 to afford 2 in high yield in CH₃CN. Although low temperature intermediates were not observed during reactions between 1 and either O₂ or PhIO, an unobserved high-valent Fe(IV)=O intermediate is proposed as a common intermediate in both of these reactions that is capable of, and responsible for, the observed solvent oxidation. Reactivity studies have shown that hydroxo-bridged 2 is itself also capable of oxidizing the relatively weak C-H bonds of DHA at room temperature. Efforts towards further exploring this reactivity, as well as comparing the properties and reactivity of thiolate-ligated analogues, are currently underway.