Activation of Dioxygen by a TAML Activator in Reverse Micelles: Characterization of an Fe^IIIFe^IV Dimer and Associated Catalytic Chemistry

Abstract

Iron TAML activators of peroxides are functional catalase-peroxidase mimics. Switching from hydrogen peroxide (H₂O₂) to dioxygen (O₂) as the primary oxidant was achieved by using a system of reverse micelles of Aerosol OT (AOT) in n-octane. Hydrophilic TAML activators are localized in the aqueous microreactors of reverse micelles where water is present in much lower abundance than in bulk water. n-Octane serves as a proximate reservoir supplying O₂ to result in partial oxidation of Fe^III to Fe^IV-containing species, mostly the Fe^IIIFe^IV (major) and Fe^IVFe^IV (minor) dimers which coexist with the Fe^III TAML monomeric species. The speciation depends on the pH and the degree of hydration w₀, viz., the amount of water in the reverse micelles. The previously unknown Fe^IIIFe^IV dimer has been characterized by UV–vis, EPR, and Mössbauer spectroscopies. Reactive electron donors such as NADH, pinacyanol chloride, and hydroquinone undergo the TAML-catalyzed oxidation by O₂. The oxidation of NADH, studied in most detail, is much faster at the lowest degree of hydration w₀ (in “drier micelles”) and is accelerated by light through NADH photochemistry. Dyes that are more resistant to oxidation than pinacyanol chloride (Orange II, Safranine O) are not oxidized in the reverse micellar media. Despite the limitation of low reactivity, the new systems highlight an encouraging step in replacing TAML peroxidase-like chemistry with more attractive dioxygen-activation chemistry.

Graphical abstract

INTRODUCTION

TAML activators of peroxides (1, Chart 1) are a family of green oxidation catalysts that function similarly to catalase-peroxidase enzymes.^1–3 In the resting state, Fe^III occupies the cavity of the deprotonated tetraamido macrocyclic ligand and directs the oxidative power of hydrogen peroxide (H₂O₂) at different targets in aqueous solution. TAML activators are similar to peroxidases mechanistically in that Fe^IV4 and Fe^V5 relatives of enzymatic compounds II and I, respectively,⁶ have been detected and characterized. Superior activity of TAML activators has been demonstrated for H₂O₂ or organic peroxides catalysis in water.^7–11 While H₂O₂ is a desirable green oxidant, its concentrated solutions are potentially hazardous¹² and it comes at a real production and delivery cost. In contrast, dioxygen (O₂) is freely available and is the principal biochemical oxidizing agent. Therefore, our foremost strategic goal for TAML activator-based catalysis has been to master the use of O₂ as the primary oxidant.^13–15 TAML activators in the ferric state are readily oxidized by O₂ in weakly coordinating organic solvents such as CH₂Cl₂ to produce Fe^IV species.¹⁶ While iron(IV) TAML activators have been shown to oxidize some substrates in organic solvents, comparably efficient catalysis to the aqueous systems has yet to be observed.^2,16,17 TAML catalysis in water is characteristically conducted aerobically. Yet unless an oxidant such as hydrogen peroxide is added, no catalysis is observed, with the current suite of catalysts, bulk water appears thus far to be inimical to facile oxygen activation. Two conditions appear to be favorable for dioxygen activation. First, the TAML medium should be rich in O₂: under ambient conditions, this situation is best accommodated in organic solvents. Second, water should surround TAML activators for maintaining high catalytic activity. Recent studies have shown that some, but not too much, water is important for oxidations of ferric TAMLs by organic peroxides in organic solvents.^18–22 The amount of water should not be large because H₂O molecules compete with O₂ for coordination sites on Fe^III. Hydrogen bonding and/ or polarity effects may also play some role. In bulk water, iron(III) TAML activators are six-coordinate species with two axial water ligands,²³ which is consistent with the observed inactivity of oxygen activation where 55 M H₂O effectively blocks O₂ binding to Fe^III.

Chart 1.

Compounds Used in This Study Including the Catalysts 1a and 1b, the Known Homovalent μ-Oxo-Bridged Iron(IV) Dimer 2a and Heterovalent μ-Oxo(hydroxo)-Bridged Iron(IV/III) Dimer 3a Characterized in This Work, and the Substrates β-Nicotinamide Adenine Dinucleotide, Reduced Dipotassium Salt (NADH), Pinacyanol Chloride (PNC), and Hydroquinone (HQ)

The two above conditions are in principle attainable using reverse micelles.^24–26 Micellar media are nowadays common for catalytic applications.²⁷ A system of reverse micelles is composed of aqueous microparticles in organic media such as n-octane (Figure 1). At the level of the individual microreactors, the micellar wall is composed of surfactant molecules; the sodium salt of bis(2-ethylhexyl) sulfosuccinate (Aerosol OT or AOT) is commonly used.²⁸ The hydrophobic tails are directed at the bulk organic medium, whereas the hydrophilic heads comprise the encapsulating surface of the microreactor inner water cavity.²⁶ Chemical and enzymatic catalysis in reverse micellar systems is well documented.^24,25 TAML activators have excellent solubility properties for guaranteeing sequestration in reverse micellar microreactors; the n-octanol/water partition coefficient for 1a (Chart 1) is 0.036.^29,30 The size of the microreactor and the water content within is determined by the degree of hydration, i.e., the water-to-surfactant molar ratio in the bulk system: w₀ = [H₂O]/[AOT].²⁶ The solubility of O₂ in n-octane is 1 order of magnitude higher than in water,³¹ and therefore TAML-containing reverse micelles will be surrounded by an O₂-rich medium favorable for the oxidation of Fe^III. At the same time, Fe^III will be in contact with water, the amount of which is regulated by the degree of hydration w₀.

Figure 1.

Schematic drawing of the reverse micelle incorporating TAML activators.

In this work we show that oxidative catalysis is possible for reactive electron donors using 1 in reverse micelles of AOT in n-octane in the absence of an added oxidant such as H₂O₂ but in the presence of ambient O₂. Reverse micelles of AOT have been thoroughly investigated²⁸ and, importantly, have often been used as media for catalysis by peroxidase enzymes,^26,32 the biocatalysts closest to TAMLs.⁸ Here we demonstrate the following. (i) Dioxygen oxidizes 1a in the reverse micellar medium to multiple iron(IV)-containing species. (ii) These species coexist with the starting iron(III) TAML. (iii) The iron(IV) species are the earlier reported homovalent dimer Fe^IVFe^IV 2a and the less-oxidized, previously unknown heterovalent dimer Fe^IIIFe^IV 3a. (iv) Compound 3a in reverse micellar medium has been detected by EPR and characterized by UV–vis, EPR, and Mössbauer spectroscopies in glycerol-containing aqueous solutions at −20 °C. (v) TAML activators in this reverse micellar medium catalyze the oxidation by O₂ of reactive electron donors such as the dye pinacyanol chloride, NADH, and hydroquinone (Chart 1). (vi) The oxidation of NADH is a photocatalytic process, the rate of which is directly proportional to the flash frequency. (vii) The catalytic activity of TAML activators is a function of both the degree of hydration w₀ and the pH. (viii) The dyes Orange II and Safranin O, which are more resistant to oxidation than PNC, are not oxidized by O₂ in this system.

EXPERIMENTAL SECTION

Materials

All reagents, components of buffer solutions, and solvents were at least ACS reagent grade (Aldrich, Fisher, Acros, Fluka) and were used as received. TAML activators 1a,b were obtained from GreenOx Catalysts, Inc.⁵⁷Fe enriched 1a was synthesized as previously reported.^33,34 n-Octane (99%+, Acros) was used as received. Stock solutions of 1a (0.0010 and 0.027 M) and 1b (0.0010 M) were prepared in the 0.01 M buffered aqueous solutions using phosphate for pH 8 and 12 but carbonate for pH 10; all were stored in a fridge at 4 °C. Hydroquinone (1.0 mg, Acros, 99.5%) recrystallized from acetone was dissolved in a mixture of 0.1 M AOT in n-octane (15.3 mL) and 48 μL of 0.01 M phosphate (pH 8 or 12) or carbonate (pH 10) buffer. This allowed the preparation of the 6.0 × 10⁻⁴ M solution of HQ in reverse micelles with lower w₀. The solutions with larger w₀ were made by adding aqueous buffer to the stock solution. Stock solutions of H₂O₂ were prepared from 30% H₂O₂, and the concentration was determined by measuring the absorption at 230 nm (ε = 72.4 M⁻¹ cm⁻¹).³⁵ AOT (Aerosol OT, BioXtra 99%+, Aldrich) was either used as received or additionally purified by activated charcoal and dried in a vacuum oven for 40 h as recommended elsewhere.³⁶ Solutions of AOT in n-octane (0.1 M) were shown to be peroxide-free using the peroxide test strips (0.5–25 ppm, EM Quant).³⁷ The AOT solutions were also treated with aqueous KI³⁷ and no color change due to the formation of I₃⁻ was observed. Negative peroxide tests were obtained for AOT before and after the purification. Reverse micelles of various w₀ were prepared by adding a corresponding amount of the aqueous buffer of known pH to 0.1 M AOT in n-octane. For example, the sample with w₀ 7 and pH 8 was prepared by adding the buffer (25 μL, pH 8) to 1975 μL of 0.1 M AOT followed by vigorous shaking. Concentrations of all components reported throughout refer to the entire volume of the solution including n-octane, AOT, and buffer components.

NADH and NAD⁺ were purchased from Sigma-Aldrich and used as received. Their stock solutions were prepared daily using HPLC grade water. The extinction coefficients (ε) of NADH in reverse micelles were measured after adding 0.1 M AOT in n-octane to 2.0 mg NADH dissolved in 60 μL of pure water in a 5 mL volumetric flask. The ε₃₄₀ values of 4760, 4980, 5070, 5130, and 5200 (all ±20) M⁻¹ cm⁻¹ at w₀ 3, 7, 10, 15, and 25, respectively, were used in kinetic measurements. Pinacyanol chloride (Aldrich) was used without further purification. Orange II (Sigma-Aldrich) was recrystallized from 1:3 H₂O:EtOH. Safranin O (Acros) was used as received. Alcohol dehydrogenase from Saccharomyces cerevisiae was a Sigma-Aldrich preparation.

Methods

UV–vis measurements were performed at 25 °C in capped quartz cuvettes (1.0 cm) using either a photodiode array Agilent 8453 UV–vis spectrometer equipped with an automatic thermostated eight-cell positioner or a double beam Shimadzu UV 1800 instrument having the thermostated six-cell positioner. HPLC measurements were carried out³⁸ using a Shimadzu Prominence 2D HPLC instrument equipped with LC-20AB binary pump (method 1) or LC-20AD quaternary pump (method 2), SIL-20A autosampler, and SPD-M20A photodiode array detector (see Supporting Information (SI) for details).

EPR and Mössbauer Measurements

Solutions of 1a (0.15 mM) in the reverse micelles as prepared above were transferred into EPR tubes and frozen by liquid nitrogen. A solution of H₂O₂ was added to a solution of 1a (0.5–2.0 mM) in 0.1 M phosphate buffer (pH 11.8), transferred to EPR tubes, and frozen in liquid nitrogen 30 s after the addition of H₂O₂. Samples at low temperature (−20 °C) were prepared in the phosphate buffer containing 50% glycerol (v/v). The temperature was lowered by keeping the solution ∼1 cm above a dry ice/acetone bath. Samples for the Mössbauer spectroscopy were prepared by adding 1 equiv of H₂O₂ to a 5 mM solution of ⁵⁷Fe enriched 1a at −20 °C as above. When needed, the solutions were split for the Mössbauer cup and EPR studies and frozen simultaneously in liquid nitrogen 30 s after the addition of H₂O₂. X-band (9.66 GHz) EPR spectra were recorded on either a Bruker ESP 300 or a Bruker ELEXSYS-II E500 spectrometer equipped with an Oxford ESR-910 liquid helium cryostat. All experimental data were collected under nonsaturating microwave conditions. The quantification of signals was relative to a CuEDTA spin standard. The microwave frequency and the magnetic field were calibrated with a frequency counter and an NMR gaussmeter, respectively. The temperature was calibrated with resistors (CGR-1-1000) from LakeShore. A modulation frequency of 100 kHz and modulation amplitude of 1 mT was used for all spectra. The analysis of data utilized the standard spin Hamiltonians eqs 1 and 2, where the subscripts 3 and 4 refer to the Fe^III and Fe^IV iron sites:

$$H=-2J{\mathbf{S}}_3\cdot {\mathbf{S}}_4+{\sum }_{i=3,4}[\beta \mathbf{B}\cdot {\mathbf{g}}_{\mathit{i}}\cdot {\mathbf{S}}_{\mathit{i}}+{\mathbf{S}}_{\mathit{i}}\cdot {\mathbf{D}}_{\mathit{i}}\cdot {\mathbf{S}}_{\mathit{i}}+{\mathbf{S}}_{\mathit{i}}\cdot {\mathbf{A}}_{\mathit{i}}\cdot {\mathbf{I}}_{\mathit{i}}+H_{\mathit{ni}}]$$ (1)

$$H_{\mathit{ni}}=-{\beta }_n{g}_n\mathbf{B}\cdot {\mathbf{I}}_{\mathit{i}}+(eQ{V}_{\mathit{zzi}}/12)[3{I_{\mathit{zi}}}^2-I_i(I_i+1)+{\eta }_i({I_{\mathit{xi}}}^2-{I_{\mathit{yi}}}^2)]$$ (2)

Simulations of EPR spectra of 1a use the relevant parts of eq 1 for a single spin S = 3/2, or the electronic parts of eq 1 for 3a. Mössbauer spectra were recorded using two spectrometers with an available temperature range of 1.5 to 200 K and magnetic fields up to 8.0 T. Isomer shifts are reported relative to Fe metal at 298 K. Data analysis, spin quantification, and simulations of the EPR and Mössbauer spectra were performed with the software SpinCount written by one of the authors (M.P.H.).³⁹

RESULTS AND DISCUSSION

A Preliminary Comment on Speciation of Iron in Reverse Micelles

Chart 2 shows species that one might anticipate could coexist in the micellar media. Most of them are known. Fe^IIIOH₂ is the starting 1a. EPR silent diamagnetic d⁴ species Fe^IVFe^IV and Fe^IVO are made from 1a and t-BuOOH in water;⁴ they dominate at pH below and above 10, respectively. Fe^VO is produced from 1a and m-chloroperoxybenzoic acid at −40 °C in MeCN.⁵ The Fe^III superoxo complex Fe^IIIO₂ was made from 1a and KO₂ at 5 °C in MeCN.⁴⁰ By analyzing data from UV–vis, EPR, and Mössbauer spectra, we have been able to build a convincing case for the iron speciation of the reverse micellar media.

Chart 2. Iron Species That Could Coexist in the Reverse Micelles^a.

^aFe^IIIFe^IV was unknown prior to this work. Species in parentheses were obtained in MeCN.

Studies of 1a in the Reverse Micelles by UV–vis Spectroscopy

In nonprotic organic solvents, the TAML activator 1a is known to be oxidized by O₂ to the Fe^IVFe^IV dimer 2a.¹⁶ The oxidation does not occur in pure water. As introduced above, the overall water content in the system of AOT reverse micelles is much lower than in bulk water, which should increase the ratio of the O₂/H₂O exposure of 1a and favor its oxidation. The first evidence for such oxidation was obtained by recording the UV–vis spectra of 1a at variable pH (8–12) and degree of hydration w₀ (3–25). Complex 1a does not absorb light above 600 nm. Thus, the observed buildup of broad bands in the range of 600–1100 nm in the AOT reverse micelles (Figure 2) was a reliable signal that iron(III) oxidation occurs in the presence of O₂.^4,16 The spectra obtained at w₀ 3, 7, and 25 (pH 8) vary in the 600–1100 nm range, indicating that more than one species was formed and that the selectivity depends on w₀. The inset to Figure 2 shows that the species generated at w₀ 3 were more stable than those at w₀ 25.

Figure 2.

Spectra of 1a in the AOT reverse micelles in n-octane recorded 20 min after mixing all components. The spectrum of 1a in the aqueous buffer (bottom) is shown for comparison. Conditions: [1a] 1.36 × 10⁻⁴ M, pH 8, w₀ = 3, 7, and 25; 25 °C. Inset shows changes of absorbance at 750 nm with time for the three different values of w₀.

Spectra similar to those in Figure 2 were generated at pH 10 and 12 (see Figure S1 of SI). Oxidized iron species were produced in the AOT reverse micelles under all conditions studied but form slower at higher pH and higher w₀. Figures 2 and S1 (SI) allow for most of the known complexes to date to be labeled as likely or unlikely components of the reverse micellar media as explained next. The characteristic 360 nm band of 1a was present under all studied conditions in the reverse micellar media. The species Fe^IVFe^IV and Fe^IVO have been characterized previously⁴ as products of the reaction between 1a and t-BuOOH in water,⁴ dominating at pH below and above 10, respectively. Both are diamagnetic EPR silent d⁴ species, and their involvement in the speciation of the reverse micellar media is discussed in detail below. The UV–vis data provided evidence for the presence of Fe^IVFe^IV in the broad bands that can be observed at 800 and 1050 nm. The Fe^VO complex was originally produced from 1a and m-chloroperoxybenzoic acid at −40 °C in MeCN.⁵ Prior experience suggests that Fe^VO is unlikely to be stable under the conditions used herein.¹⁸ The Fe^VO complex in MeCN exhibits the characteristic UV–vis peak at 630 nm; there is no identifiable equivalent in Figure 2. Similarly, Fe^IIIO₂ made by Nam et al. from 1a and KO₂ at 5 °C in MeCN was not present.⁴⁰ The UV–vis spectra were ambiguous regarding the presence of mononuclear Fe^IVO. However, Figures 2 and S1 (SI) make it clear that one or more additional species were present, as signified by the broad bands at 440 and 760 nm; these bands are suggestive of at least one novel dimer.

EPR Spectroscopy of 1a in the Reverse Micelles

The EPR spectra of 1a (Fe^IIIOH₂) in water²³ and Fe^VO in MeCN⁵ have been characterized in earlier studies. The Fe^IIIFe^IV species of Chart 2 was previously unknown. The remaining species of Chart 2 are EPR silent. As stated above, the generation of Fe^IIIO₂ and Fe^VO from 1a and O₂ is improbable. The EPR spectrum of 1a in the reverse micelles (pH 12) displayed a signal from the Fe^IIIOH₂ S = 3/2 species (g = 4.0, 2.0)²³ and a new previously unreported S = 1/2 species with g = 1.99, 2.11, 2.14 (Figure 3E). Similar EPR spectra were also recorded at pH 8 and 10 (Figure S2, SI). Spin quantification of the S = 1/2 species indicated that 15% of the initial concentration of 1a converted to the S = 1/2 species. As shown below, the new species was assigned to heterovalent Fe^IIIFe^IV (Chart 2) or 3a (Chart 1) and was also generated from 1a and H₂O₂ in a buffered solution in the presence of glycerol.

Figure 3.

EPR spectra (ν = 9.651 GHz, T = 15 K) of 1a at pH 12 under different conditions. A: 1a; B: 1a + 1 equiv of H₂O₂; C: 1a in 50% glycerol; D: 1a + 1 equiv of H₂O₂ in 50% glycerol; E: 1a in reverse micelles (w₀ 10, pH 12). The spectral intensities are displayed for equal concentrations of 1a, with additional scale factors as displayed. The black line on C is a simulation for S = 3/2, D = +2.5 cm⁻¹, E/D = 0, g = 2.045, 2.05, 2.03, σ_g = 0.15, 0.15, 0.02). The species percentages are stated in the text. A small variable amount (∼2%) of 3a is generated in air without H₂O₂, which is presumed to be reaction of 1a with low levels of dioxygen.

EPR and Mössbauer Spectroscopy of 1a with H₂O₂

The EPR signal of 1a (Fe^IIIOH₂) in phosphate buffer (pH 11.8) was extremely broad (g = 4, Figure 3A) due to intermolecular interaction. The addition of 50% glycerol to the solution significantly sharpened the EPR signal (Figure 3C). The simulation in Figure 3C (black line) for a spin S = 3/2 system includes a broad distribution of g values which was indicative of some aggregation in the presence of glycerol but significantly less than without glycerol. As in the glycerol-free system,²³ the temperature dependence of the signal indicates that the spectrum originates from a ground spin doublet with a positive zero-field splitting with D = +2.5(5) cm⁻¹. The signal intensity was in quantitative agreement with the amount of iron added to the solution.

The addition of 1 equiv of H₂O₂ to 1a at room temperature and pH 11.8 generated a new EPR signal near g = 2 (Figure 3B) which was sharper in 50% glycerol (Figure 3D). The g values (1.992, 2.111 and 2.144) are indicative of a spin S = 1/2 system which matched the g values of the species observed in the reverse micelles without H₂O₂ (Figure 3E). In glycerol, the addition of H₂O₂ caused conversion of 60% of the initial amount of 1a into other complexes. The amount of iron that converted into the S = 1/2 species accounted for 30% of the initial iron concentration of 1a. The remainder (30% of the iron) converted to an EPR silent species, which is consistent with a previous report that peroxides convert Fe^III into an EPR-silent Fe^IV species.⁴ Similar EPR samples were generated with ⁵⁷Fe-enriched 1a and H₂O₂. Figure 4 shows an expanded scale of this species and S = 1/2 simulations. The broadening of the g = 1.992 peak (inset to Figure 4) was significant and was simulated with the addition of a hyperfine term for ⁵⁷Fe (I = 1/2) with A = 18 MHz. This value was in agreement with Mössbauer results shown below indicating that the S = 1/2 species was associated with iron.

Figure 4.

EPR spectra (red and green) of 3a under the same conditions as in Figure 3D. The inset shows an expanded view at g = 1.992 for ⁵⁶Fe and enriched ⁵⁷Fe samples. The dashed lines are S = 1/2 simulations with and without inclusion of a hyperfine term for I = 1/2 with A = 18 MHz.

TAML activators possess catalase-like activity⁸ which causes evolution of O₂ due to the disproportionation of H₂O₂ at millimolar concentrations of 1a. Therefore, the temperature was lowered to −20 °C (dry ice/acetone bath) prior to addition of H₂O₂ to slow the reaction. This increased the yield of the species, giving the S = 1/2 signal which was highest in the presence of 1 equiv of H₂O₂ but decreased at higher amounts of H₂O₂. The total iron concentration determined from quantification of the S = 1/2 and S = 3/2 EPR signals was significantly lower than total amount of iron introduced. However, as the initial concentration of 1a was lowered, the percent yield of 3a increased significantly (see Figure S3 of SI). At low 1a concentration, and assuming the S = 1/2 signal is associated with a dinuclear iron complex, nearly quantitative yield of the S = 1/2 species was observed based on spin quantification. Furthermore, the Mössbauer spectra (see Figure 5 below) did not show a significant amount of mono- or dinuclear iron(IV) species under such conditions.

Figure 5.

Mössbauer spectra with a parallel field of 45 mT at 4.2 K (A–C) and 140 K (D–F) of (A, D) 1a at pH 11.8 (0.1 M phosphate), 50% glycerol; (B, E) 1a + 1 equiv of H₂O₂, added at −20 °C; and (C, F) 3a, Fe^IIIFe^IV. The spectra of Fe^IIIFe^IV are subtractions of 45% of the spectra of 1a from the spectra with 1a + H₂O₂. At 140 K (F), the simulation (dashed line) is for two quadrupole doublets with equal area (solid black lines) with parameters given in the text.

These results indicated that the S = 1/2 species was a heterovalent Fe^IIIFe^IV dimer. The Mössbauer data (Figure 5) confirmed the oxidation states as Fe^III and Fe^IV. The microwave saturation of the S = 1/2 signal was measured over the temperature range of 2 to 46 K. The half-power saturation data as a function of temperature was fit to an Orbach relaxation process (see Figure S5, SI), which determined the value of the exchange interaction between the Fe^III and Fe^IV centers, J = −30 ± 10 cm⁻¹. There are a few characterized Fe^IIIFe^IV centers in model compounds^41–44 or proteins.^45,46 The exchange interactions for all of these previously characterized complexes were significantly larger in magnitude (J < −100 cm⁻¹). The exchange interaction for Fe^IVFe^IV species of TAML was also large (J < −60 cm⁻¹).⁴ The lower exchange value for 3a does not allow identification of a specific bridging species: the bridge could be oxo or hydroxo.

The spin quantification of 3a allowed a determination of the extinction coefficients of the UV–vis spectral characteristics at ambient conditions. The UV–vis sample was prepared by adding 1 equiv of H₂O₂ to 1a at pH 11.8 in 50% glycerol. The spectrum shown in Figure S4 (SI) did not change after 2 min needed for freezing EPR samples. The concentration of 3a determined from the quantification of the EPR signal was used to calculate the extinction coefficients at 760 and 930 nm of 2500 and 425 M⁻¹ cm⁻¹, respectively.

The 4.2 K Mössbauer spectrum of 1a enriched with ⁵⁷Fe at pH 11.8 in an applied magnetic field of 45 mT displayed a quadrupole doublet with δ = 0.14(4) mm s⁻¹ and ΔE_Q = 3.97(4) mm s⁻¹. The absence of a magnetic splitting pattern indicated that 1a was in the fast relaxation regime at 4.2 K. The Mössbauer spectrum in 50% glycerol (Figure 5A) was broad and remained broad at 140 K (Figure 5D), indicating an intermediate relaxation mode. The absence of a well resolved magnetic splitting pattern at 4.2 K was due to aggregation of 1a in the aqueous environment, which was consistent with the EPR data.

The Mössbauer spectra at 4.2 and 140 K after the addition of 1 equiv of H₂O₂ to 1a at pH 11.8 and −20 °C (50% glycerol) showed the partial loss of 1a and the appearance of a new species (Figure 5B,E). The spectra of the new species (Figure 5C,F) were obtained by subtracting the corresponding spectra of 1a (Figure 5A,C). The subtraction percentage (45% based on area) was determined by minimizing the features (indicated by arrows) attributed to 1a in the 4.2 K spectrum. In agreement, the EPR spectra of samples prepared from the same stock solution as the Mössbauer samples showed the S = 1/2 and S = 3/2 signals at the same percentages in iron (assuming two irons per S = 1/2 species). As shown in Figure 5F, the spectrum was fit with two doublets of equal area with parameters δ = 0.01 ± 4, ΔE_Q = 3.09 ± 5 mm s⁻¹ and δ = −0.11 ± 4, ΔE_Q = 3.33 ± 5 mm s⁻¹. These parameters could not be accurately determined from the spectra recorded at 4.2 K. The second-order Doppler correction would increase the isomer shift values by approximately +0.06 mm s⁻¹. Previously characterized mononuclear Fe^III macrocyclic complexes have δ and ΔE_Q in the range of 0.12 to 0.25 mm s⁻¹ and 3.6 to 4.19 mm s⁻¹, respectively.^47–49 The δ and ΔE_Q values for similar Fe^IV complexes are −0.19 to −0.03 mm s⁻¹ and 0.89 to 4.35 mm s⁻¹, respectively.^4,16,50,51 Including the correction, the isomer shifts of +0.07 mm s⁻¹ and −0.05 mm s⁻¹ from 3a are close to or in range of the isomer shifts from the previously characterized Fe^III and Fe^IV complexes, respectively.

The Mössbauer spectra of 3a recorded in the magnetic fields of 0.05, 3, and 8 T are shown in Figure 6. The sample was prepared as in Figure 5, and the spectra are displayed after subtraction of 45% of the corresponding spectra of 1a (based on area). Raw spectra are shown in Figure S6, SI. The spectra were simulated using eq 1 for an antiferromagnetic exchange of J = −30 cm⁻¹ between the Fe^III and Fe^IV sites as determined above from EPR spectroscopy. The agreement of Mössbauer spectra with simulations indicate that the spin states of the Fe^III and Fe^IV sites are S = 3/2 and 1, respectively. These spin states are the same as those observed for monomeric Fe^III or Fe^IV TAML complexes characterized previously.^4,23 The simulations use a set of parameters for D, E/D, and A-tensors given in the figure caption that are derived from least-squares fits of the data. The parameters set was not unique but was consistent with the spin state assignments.

Figure 6.

Mössbauer spectra (red lines) of 3a at pH 11.8, 50% glycerol, 1 equiv of H₂O₂ recorded at 4.2 K with an applied field of 8 T (A), 3 T (B), and 0.05 T (C). The spectra are after subtraction of 45% of the area the corresponding spectra of 1a. Simulations (black lines) are for Fe^III and Fe^IV sites: J = −30 cm⁻¹, S₃ = 1.5, D₃ = 3.2 cm⁻¹, E/D₃ = 0.18, A₃ = (−45, −120, 55) MHz, S₄ = 1, D₄ = 21 cm⁻¹, E/D₄ = 0.13, A₄ = (−112, −151, 110) MHz for the values of δ and ΔE_Q given in the text.

Speciation of Iron in Reverse Micelles

EPR and Mössbauer studies of 1a/H₂O₂ in 50% glycerol allowed identification of the Fe^IIIFe^IV (3a) species formed from 1a and O₂ in the reverse micelles. The plausible coexisting species are Fe^IIIOH₂, Fe^IIIFe^IV, Fe^IVFe^IV, and Fe^IVO (Chart 2). In water, Fe^IVFe^IV and Fe^IVO exist in a pH-controlled equilibrium.⁴ The former dominates at pH < 10, the latter at pH > 12. It is important to note that the effective pH in the AOT reverse micelles may differ from the pH of the aqueous buffer introduced (and indicated throughout). The acidity can be higher near the interface for negatively charged surfactants.⁵² The effective pH is usually lower by 1–3 units in the AOT reverse micelles than in the bulk water depending on w₀,^53,54 though the difference may be smaller than 1 unit when w₀ is high.²⁸ Therefore, Fe^IVO cannot be a dominating species and probably does not form at all. The absence of the peak around 435 nm in Figures 2 and S1, which is typical of Fe^IVO (ε = 2500 M⁻¹ cm⁻¹),⁴ gives additional evidence. Thus, the remaining species to be considered are Fe^IIIOH₂, Fe^IIIFe^IV, and Fe^IVFe^IV. The latter has a sharper band at 481 nm and a broader band at 856 nm, whereas the band at 365 nm is weaker as compared to that of Fe^IIIOH₂. The less oxidized Fe^IIIFe^IV has two bands at 440 and 760 nm. These bands allowed semiquantitative interpretation from the spectral data in the range of 760–930 nm where Fe^IIIOH₂ does not absorb. The estimations were obtained using eqs 3 and 4 that connect absorbances at 760 and 930 nm to concentrations of the iron dimers using the known values of ε₇₆₀ and ε₉₃₀ (2500 and 425 M⁻¹ cm⁻¹ for Fe^IIIFe^IV vs 4820 and 3056 M⁻¹ cm ⁻¹ for Fe^IVFe^IV, respectively). We assumed that the extinction coefficients were approximately independent of w₀ and pH.

$$A_{760}=2500[{\text{Fe}}^{\text{III}}{\text{Fe}}^{\text{IV}}]+4820[{\text{Fe}}^{\text{IV}}{\text{Fe}}^{\text{IV}}]$$ (3)

$$A_{930}=425[{\text{Fe}}^{\text{III}}{\text{Fe}}^{\text{IV}}]+3056[{\text{Fe}}^{\text{IV}}{\text{Fe}}^{\text{IV}}]$$ (4)

The amount of Fe^IIIOH₂ shown in Figure 7 was obtained by subtracting 2 × [Fe^IIIFe^IV] and 2 × [Fe^IVFe^IV] from the total iron in the system. Fe^IIIOH₂ dominates under almost all conditions tested, i.e., there is incomplete oxidation of Fe^III by O₂ in the reverse micellar media under the conditions studied. The more heterovalent dimer Fe^IIIFe^IV, the major Fe^IV-containing material, is formed at higher pH and moderate w₀, consistent with the fact that basic conditions stabilize Fe^IIIFe^IV in aqueous glycerol. The homovalent dimer is present at low w₀ and pH 8 because aprotic organic medium and this pH are both favorable for Fe^IVFe^IV.^4,16 The Fe^IV percentage drops at higher w₀ because “wet” micelles are closer to bulk water where 1a is not oxidized by O₂.¹⁶

Figure 7.

Estimated fractions of Fe^IIIOH₂, Fe^IIIFe^IV, and Fe^IVFe^IV (all derived from 1a) in the AOT reverse micelles at different pH and w₀. For other conditions, see caption to Figures 2 and S1 (SI).

The data in Figure 7 should be considered as qualitative because water inside reverse micelles is dissimilar to bulk water in terms of polarity, acidity, and microscopic viscosity.^24,54 As a result, the environment of 1a varies with the degree of hydration w₀ of the reverse micelles, and the ε values in reverse micelles and pure water may differ. Nevertheless, estimates in Figure 7 for the UV–vis data agree with the EPR results in Figures 3 and S2 (SI) because the spin quantitation gives the yields of Fe^IIIFe^IV of 13, 33, and 37% at pH 8, 10, and 12, respectively, at w₀ 10 and total iron of 1.5 × 10⁻⁴ M. Thus, 1a is oxidized in the O₂-rich system into Fe^IV species in the absence of peroxides. We will now describe the catalytic activation of O₂ by 1a in these reverse micelles.

Catalytic Oxidation of NADH in Reverse Micelles

The NADH/NAD⁺ couple is the essence of numerous biological processes^55–57 including those catalyzed by NAD⁺-dependent dehydrogenases, viz. enzymes that are widely used in organic synthesis.^56,58 The limitations of enzymatic processes are due to the cost of NAD⁺,⁵⁹ and therefore reliable systems for NAD⁺ regeneration are needed, as the current methods are still not adequately effective.⁶⁰ TAMLs catalyze the oxidation of NADH by enzymatically produced H₂O₂ under mild conditions,⁶¹ and therefore NADH was the first choice substrate.

Smooth conversion of NADH into NAD⁺ catalyzed by 1a/O₂ was monitored by UV–vis spectroscopy at 340 nm (maximum for NADH; NAD⁺ does not absorb) at w₀ in the range of 3 to 25 and pH 8, 10, and 12. At pH 10 the isosbestic point holds at 286 nm (Figure 8A). Control experiments in the reverse micelles in the absence of 1a showed negligible oxidation of NADH within the same time period (Figure S7, SI). The catalytic conversion of NADH into NAD⁺ rather than into smaller fragments was proven by HPLC³⁸ by comparing the elution times of the product and authentic NAD⁺. At pH 10 and w₀ 10, no NADH was observed after 9 h by UV–vis spectroscopy when the spectra were recorded every 30 s (see below for explanations). The 83% yield of NAD⁺ was calculated from the HPLC data. This corresponds to the turnover number (TON) of 88 and turnover frequency (TOF) of 0.003 s⁻¹ at [1a] = 2.46 × 10⁻⁶ M and [NADH] = 2.6 × 10⁻⁴ M. The control experiment in the absence of 1a showed 67% NADH and 27% NAD⁺ by HPLC after 9 h. Catalyst 1a converts NADH into the “enzymatically active” NAD⁺. Alcohol dehydrogenase (ADH) and ethanol rapidly reduce NAD⁺ formed to NADH,⁶⁰ and 95% NADH reappears almost instantly, much faster than the NADH oxidation (inset to Figure 8B). These results lead to the conclusion that the oxidation proceeds according to eq 5.

$$\text{NADH}+\frac{1}{2}{\mathrm{O}}_2+{\mathrm{H}}^{+}\stackrel{\text{TAML}}{\to }{\text{NAD}}^{+}+{\mathrm{H}}_2\mathrm{O}$$ (5)

Figure 8.

(A) Spectral changes that accompany 1a-catalyzed oxidation of NADH into NAD⁺ by O₂ and (B) regeneration of NADH by ADH/EtOH. Time interval between spectra shown in A is 20 min (scans were made every 30 s). ADH (from Saccharomyces cerevisiae, 0.7 mg, 210 units in 50 μL water) and 10 μL of EtOH were added to the reaction mixture after 3 h followed by vigorous shaking. Inset in B illustrates the time scale of TAML-catalyzed oxidation and the enzyme-catalyzed regeneration of NADH. Conditions: 0.1 M AOT, 0.01 M carbonate (pH 10), w₀ 7; [1a] 2.46 × 10⁻⁶ M, [NADH] 9.9 × 10⁻⁵ M.

Whenever the catalytic reduction of O₂ is being studied, it is important to be sure that adventitious oxidizing agents are not interfering with the reaction chemistry.⁶² A major concern is the presence of peroxidic impurities in the components of the reaction media. To rule out the possible involvement of peroxides, 1b, which is more reactive in oxidations by H₂O₂ in pure water than 1a,⁸ was used as the catalyst in the reverse micellar oxidation of NADH. Fluorinated TAMLs such as 1b react with O₂ in aprotic solvents much slower than their methylated counterparts.¹⁶ The data in Figure S8 (SI) show that 1b is 5-times less reactive than 1a in the reverse micelles at pH 10 and w₀ 3. This establishes that adventitious peroxides are unlikely major participants. Furthermore, colorimetric testing for peroxide with KI and peroxide test strips produce no color changes.³⁷ These results support the case that 1a does indeed launch an oxygenase-like process eq 5 in the reverse micelles and that H₂O₂ is not involved.

Next the ability of the iron(IV) complex 2a¹⁶ to oxidize NADH ([Fe^IV]:[NADH] = 1:1.05) was tested in reverse micelles in degassed media under argon to preclude or at least minimize autoxidation. The NADH absorbance around 340 nm decreased on addition of 2a, indicative of the oxidation of NADH (Figure 9). NAD⁺ was produced in 62 and 52% with respect to the total iron according to UV–vis and HPLC data, respectively. When ferric 1a was employed in place of 2a under otherwise identical conditions, the yield of NAD⁺ fell to 14% (HPLC). That any oxidation was observed at all suggested that O₂ was not excluded completely by the degassing procedure. These facts support the hypothesis that NADH can be oxidized by 2a and allow us to conclude that catalyst-based oxidations are competent to produce the observed chemistry in these reaction systems where one should always be on the lookout for adventitious free radical autoxidation.

Figure 9.

UV–vis spectra of NADH in the presence of 1a or 2a (1.5 mL methanol was added to 1.5 mL reaction mixture to quench the reaction). Conditions: w₀ 7, pH 12, [NADH] 2.1 × 10⁻⁴ M, [total iron] = 2.0 × 10⁻⁴ M. (a, b: spectra of NADH reacted with 1a or 2a in the absence of O₂; c, d: spectra of NADH reacted with 1a or 2a after exposure to air for 15 min; e: spectrum of NADH alone).

Comparison of Initial Rates of NADH Oxidation

NADH is slowly oxidized in reverse micelles in the absence of 1a as shown in Figure S7 (SI). TAML 1a accelerates the oxidation to such extent that the spontaneous process can be neglected in many cases. Nevertheless, the initial rates shown in 3D Figure 10 were corrected for the uncatalyzed reaction in the absence of 1a. The degree of hydration w₀ is a key factor that regulates the reactivity. Reaction 5 is strikingly faster in lower w₀ micelles. One explanation is that water impedes the binding of O₂ to the Fe^III center thus slowing the catalytic rate. However, the actual concentrations of iron species and/or NADH within a microreactor will be changing with w₀. The catalyst will certainly partition preferentially to the water, so concentration effects cannot be neglected and perhaps provide the dominant explanation.⁶³ Reaction 5 is much less sensitive to pH than to w₀ (Figure 10).

Figure 10.

Initial rates of NADH oxidation catalyzed by 1a in reverse micelles. Conditions: [1a] 2.5 × 10⁻⁶ M, [NADH] 5.1 × 10⁻⁵ M, the flash frequency is 30 s. See text for explanations.

Effect of Light on the Catalyzed NADH Oxidation in Reverse Micelles

TAML-catalyzed reaction 5 is accelerated by light. The effect was established by changing the scanning frequency when the NADH oxidation was monitored by a photodiode array UV–vis spectrometer. The process was noticeably faster when spectra were registered more frequently, i.e., when the flash frequency (F_f = inverse time between recording (TbR) of successive spectra) was higher (Figure 11). The slowest rate was registered when reaction 5 was followed using a double beam UV–vis spectrometer, but it increased linearly with increasing flash frequency when the photodiode array instrument was used (inset to Figure 11). The straight line has an insignificant positive intercept which may reflect a dark process. A double beam spectrometer emits much less light to the sample than a photodiode array instrument which utilizes the undispersed light beam in the 190–1100 nm spectral region. Reactions susceptible to “diode array acceleration” are known and have recently been reviewed.⁶⁴ The majority of these processes involve O₂. The mechanisms do not involve species in long-lived excited states and are usually complex requiring individual systematic studies.

Figure 11.

Kinetics of NADH oxidation registered by different UV–vis spectrometers applying different pulse frequencies. The top curve was obtained using a double beam instrument, and lower curves were produced using a photodiode array instrument. TbR is the time between recordings of successive spectra. Other conditions: pH 10, w₀ 10, [NADH] 5.14 × 10⁻⁵ M, [1a] 2.39 × 10⁻⁶ M.

The light accelerates the uncatalyzed oxidation of NADH in the reverse micelles in the absence of 1a (Figure S7, SI). Thus, the irradiation affects NADH. Experiments with pinacyanol chloride (Figure S9, SI) suggest that irradiation does not impact 1a directly. The bleaching rate of PNC was found to be practically insensitive to the flash frequency of the photodiode array spectrometer (Figure S9, SI); a minor effect is ascribed to the reported photodegradation of aggregated PNC in water,⁶⁵ which is also observed in the absence of 1a (inset to Figure S9, SI). The kinetics of absorbance growth at 750 nm has been studied to probe the formation of Fe^IIIFe^IV and/or Fe^IVFe^IV from 1a and O₂ in the reverse micelles as a function of the flash frequency at w₀ 15 and pH 12. Under such conditions the rates are lower and the measurements are more accurate. No light effect was registered (see Figure S10, SI), indicating that NADH is the photosensitive component of reaction 5.

Oxidation of Pinacyanol Chloride and Hydroquinone

In 1998, PNC was used to unveil the catalytic activity of 1a/H₂O₂.⁶⁶ PNC has limited solubility in aqueous buffer at pH 12 but dissolves readily in the AOT reverse micelles. As noted above, PNC decomposes slowly in the reverse micelles without 1a (Figure S9, SI); the process is faster in water at pH 10 due to the aggregation of PNC to form a dark solid.⁶⁷ The oxidation of PNC by O₂ in the reverse micelles is strongly catalyzed by 1a at pH 10 (Figures 12 and S9, SI). The bleaching of PNC was investigated in some detail at w₀ 3, 10, and 25 and pH 8, 10, and 12 (Table S4, SI). It is rather slow at pH 8 typical of lower activity of 1a/H₂O₂ in water at neutral/acidic pHs.³ At pH 10, the fastest bleaching was observed at w₀ 10, but it slowed down in the “drier micelles” at w₀ 3 (Figure S11, SI). The oxidation of Fe^III to Fe^IV and the oxidative bleaching of PNC by Fe^IV are affected oppositely by w₀. Lower w₀ values favor the former process but disfavor the latter. Under the optimal conditions (w₀/pH = 10/12) in the presence of just 1% 1a, TON and TOF equal 90 and 1.6 × 10⁻³ s⁻¹, respectively. The 1a-catalyzed oxidation of PNC by O₂ in the reverse micelles is less deep than by H₂O₂ in water. Figure 12 shows that a decrease of the main 600 nm peak is accompanied by a buildup of a smaller peak at 350 nm, which was not observed in the aqueous solution.⁶⁸

Figure 12.

Spectral changes that accompany 1a-catalyzed oxidation of PNC by O₂ in the reverse micelles. Conditions: [PNC] 4.5 × 10⁻⁵ M, [1a] 5.0 × 10⁻⁶ M, w₀ 10, pH 10, spectra shown with 10 min intervals.

The rapid oxidation of hydroquinone (λ_max 289 nm) to 1,4-benzoquinone (λ_max 247 nm)⁶⁹ is conveniently studied by UV–vis spectroscopy.⁶⁹ The process is very fast under basic conditions and is followed by a second oxidation step.⁷⁰ The spontaneous oxidation of HQ was found to slow in the reverse micelles prepared using a tiny amount of water (w₀ ∼1.6, see Experimental Section) and studied immediately after preparation. The oxidation of HQ is faster in the presence of 1a under all conditions tested. Notably, the HQ oxidation was faster in wet micelles with higher w₀ (Figure S13, SI), in contrast with the oxidation of NADH, which was found to be more rapid in micelles with low w₀, or PNC, which occurred most rapidly at intermediate w₀. This is probably because the effective pH inside the water pools decreases with w₀ to favor HQ stability. A secondary oxidation process is also visible for HQ at higher pH (pH 10 and 12, Figure S12, SI).

Limitations of Reverse Micelles as Media for TAML-Catalyzed Oxidation by O₂

TAML activators catalyze the bleaching of Orange II⁷ and Safranin O⁷¹ dyes by H₂O₂ in water. These two dyes are more difficult to decolorize than PNC.⁶⁸ Orange II has become widely used for assaying the catalytic activity of synthetic oxidation catalysts.^{7,11,30,72–74} Neither Orange II nor Safranin O were decolorized in the presence of 1a in the AOT reverse micelles even in the presence of equimolar amounts of 1a and Orange II or Safranine O in the reaction media.

On the Formation of the Heterovalent Dimer

The AOT reverse micelle microreactors have proven to be attractive and challenging media for TAML activator processes. By imposing limitations on the water content of the AOT microreactors to favor TAML interactions with O₂ in the adjacent dioxygen pool, the resting ferric species is easily converted into iron(IV) derivatives that can catalytically oxidize selected electron donors. These oxidations replace H₂O₂ as a primary oxidant of TAML catalysis with the far more technically attractive O₂. In the reverse micelles, O₂ converts iron(III) TAML 1a into the heterovalent Fe^IIIFe^IV dimer 3a via the simplified process of eq 6, assuming the μ-oxo bridged formulation. There are relatively few characterized Fe^IIIFe^IV centers in synthetic complexes^{41–44,75–77} and metalloproteins,^45,46,78,79 and this adds interest to the discovery of compound 3a. The reaction stoichiometry is rather unusual for iron–oxygen interactions^80,81 because one O₂ molecule formally requires eight iron(III) units for complete depositing its four electrons.

$$8{[{\text{Fe}}^{\text{III}}{\text{OH}}_2]}^{-}+{\mathrm{O}}_2\to 4{[{\text{Fe}}^{\text{IV}}{\text{OFe}}^{\text{IV}}]}^{3-}+6{\mathrm{H}}_2\mathrm{O}+4{\mathrm{H}}^{+}$$ (6)

$$4{[{\text{Fe}}^{\text{III}}{\text{OH}}_2]}^{-}+{\mathrm{O}}_2\to 2{[{\text{Fe}}^{\text{IV}}{\text{OFe}}^{\text{IV}}]}^{2-}+4{\mathrm{H}}_2\mathrm{O}$$ (7)

Equation 6 should be compared with the known process of eq 7 in which iron(III) TAML activators are rapidly oxidized by O₂ in aprotic organic solvents to form the homovalent Fe^IVFe^IV μ-oxo bridged dimer 2a.¹⁶ In this case, four iron(III) units are needed, which is of course, a more common scenario.⁸² Hydrogen peroxide as a 2e oxidant leads to 3a via eq 8, i.e., the iron(III) to H₂O₂ stoichiometry is 4:1 at pH 11.8 in 50% glycerol at lower temperatures.

$$4{[{\text{Fe}}^{\text{III}}{\text{OH}}_2]}^{-}+{\mathrm{H}}_2{\mathrm{O}}_2\to 2{[{\text{Fe}}^{\text{III}}{\text{OFe}}^{\text{IV}}]}^{3-}+4{\mathrm{H}}_2\mathrm{O}+2{\mathrm{H}}^{+}$$ (8)

It is intriguing that the yield of 3a produced in pure water is much lower than in the presence of glycerol; the dominant product in pure water is the homovalent dimer 2a.⁴ Explanations are conceivable. For example, glycerol may be more than a cosolvent⁸³ by behaving as a reducing agent^84,85 to favor the formation of the less oxidized heterovalent dimer 3a.

Reactivity Comparisons of TAML Species

The oxygenase-like activity of TAML activators in the reverse micelles established in this work is moderate; 1a/O₂ does not bleach what are facile targets in water, such as Orange II, but does catalyze the mild oxidation of reactive electron donors such as NADH; enzymatically active NAD⁺ is produced with TON of 88. The TOF of 0.003 is, however, lower than those recently reported for O₂ oxidations involving the iron(III) complex of meso-tetrakis(4-sulfonatophenyl)porphyrin (0.11 in the aqueous phosphate buffer of pH 7)⁸⁶ and 1a in the presence of glucose oxidase/D-glucose (ca. 0.03 at pH 7.5).⁶¹

The lower reactivity of 1a in these reverse micelles may have advantages. Typically, TAML activators in basic water catalyze multistep, deep oxidation of organic substrates by H₂O₂,¹⁷ fragmenting and nearly mineralizing persistent pollutants including polychlorophenols,⁸⁷ organophosphorus pesticides,⁸⁸ and many others.¹⁷ Here the mild process for NADH, even with light induction, is reminiscent of cellular processes where the NAD⁺ is available for catalytic cycling. Thus, the activity achieved may be of use for transformations of fragile molecules by O₂ that avoid destructive deep oxidation. We are further examining the potential.

The lower reactivity of TAML activators in these processes is adequately understood at a molecular level because there is a clear parallel between the reactivity and the iron oxidation state as shown in Chart 2 where the iron oxidation state in characterized TAML species increases from left to right (from 3+ to 5+); Fe^VO expresses a ca. 10⁴ fold rate advantage over Fe^IVFe^IV in organic media.^18,19 Monomers Fe^IVO are known to be more reactive than dimers Fe^IVFe^IV in aqueous solutions.⁸⁹ Thus, the dominating speciation of iron here as the least oxidized and oxidizing Fe^IIIFe^IV explains the observed reactivity picture.

CONCLUSIONS

By catalyzing the oxidation of NADH to NAD⁺ by O₂ in AOT reverse micelles with a TON of 88, peroxidase-mimicking TAML activators have been shown to be capable of oxygenase-mimicry under appropriately engineered conditions. The oxygen oxidations are less aggressive than hydrogen peroxide oxidations. Nevertheless, oxygen is special, making the findings important and suggesting that more research should be carried out to explore whether or not TAML activators can function effectively using affordable molecular oxygen in a wide range of oxidation processes.