Tunneling in C-H Oxidation Reactions by an Oxoiron(IV) Porphyrin Radical Cation: Direct Measurements of Very Large H/D Kinetic Isotope Effects

Abstract

Rate constants for oxidations of benzyl alcohol-d₀ and -d₇ by oxoiron(IV) tetramesitylporphyrin radical cation perchlorate in acetonitrile were measured in single-turnover kinetic studies. The kinetic isotope effect (k_H/k_D) increased from 28 at 23 °C to 360 at −30 °C due to extensive hydrogen atom tunneling that was analyzed in terms of a parabolic energy barrier to tunneling. Similarly large KIE values were found for oxidations of ethylbenzene-d₀ and -d₁₀ at room temperature. The large KIE values are a function of the porphyrin identity, and porphyrins containing electron-withdrawing groups display normal KIEs. KIEs found under catalytic turnover conditions are somewhat smaller than those obtained in single turnover reactions. The results should serve as benchmarks for computational studies of C-H oxidations by porphyrin and heme-iron-oxo systems.

Enzyme-catalyzed C-H oxidation reactions have been a subject of considerable research interest for decades due to the high energy demand of the reactions. Models of iron-containing enzymes include iron-porphyrin complexes that are based on heme enzymes such as the cytochrome P450 enzymes. These models function as oxidation catalysts when they are activated by sacrificial oxidants that give high-valent iron-oxo complexes.¹ For oxidations of C-H bonds, the transient catalyst oxidants usually have an iron atom in the formal +5 oxidation state. The observable highly oxidized models for the heme-containing enzymes are oxoiron(IV) porphyrin radical cations, examples of which were characterized as early as the 1980’s.²

Among the more intriguing features of enzyme-catalyzed C-H oxidations are very large H/D kinetic isotope effects (KIEs) ranging up to 100 found in, for example, the hydrogen-atom abstractions of lipoxygenase³ and insertions of oxygen into a C-H bond of methane by methane monooxygenase enzymes.⁴ Such large KIEs are ascribed to hydrogen atom tunneling reactions that can be optimized in the enzyme by controlling residue positions in the active site and can include vibrational or gating processes.⁵ We report here that very large H/D KIE’s due to tunneling can be observed in C-H hydroxylation reactions effected by an oxoiron(IV) porphyrin radical cation under single turnover conditions.^6–8 Quantifications of the tunneling reactions are expected to have ramifications for understanding enzyme-catalyzed oxidation reactions and to serve as benchmarks for computational studies of C-H oxidation reactions.

5,10,15,20-Tetramesitylporphyrin-iron(III) chloride and perchlorate complexes (TMP)Fe^III(X) were oxidized with m-chloroperoxybenzoic acid (mCPBA) to give the known oxoiron(IV) porphyrin radical cations (TMP)^+•Fe^IV(O)(Cl) and (TMP)^+•Fe^IV(O)(ClO₄).^2,9,10 The iron-oxo complexes reacted smoothly in pseudo-first-order processes with large excesses of substrates (benzyl alcohol or ethylbenzene) in acetonitrile in reactions with clear isosbestic points (Figure 1). The substrates are oxidized at the benzylic positions to give, respectively, benzaldeyde and 1-phenylethanol (see Supporting Information).⁹

Figure 1.

Time-resolved UV-vis spectrum for reaction of 2 (X = ClO₄) with 5.0 mM benzyl alcohol in CH₃CN at 23 °C; the time period is 5–50 s after mixing. The inset is an expansion of the region λ = 450–750 nm.

Variable temperature kinetic studies were conducted over the temperature range −30 °C to 23 °C. Rate constants were measured for a series of pseudo-first-order reactions at a given temperature, and the second-order rate constants were determined from Eq 1, where k_obs is the observed pseudo-first-order rate constant, k₀ is a background rate constant (ca. zero in these studies), k_ox is the second-order rate constant, and [sub] is the substrate concentration. The results are in Table 1.

$$k_{\text{obs}}=k_0+k_{\text{ox}}[\text{sub}]$$ (1)

Table 1.

Second-Order Rate Constants for Reactions of (TMP)^+•Fe^IV(O)(ClO₄ ⁻) in CH₃CN.

Substratea	Tempb	kobs(H)c,d	kobs(D)c,e	KIEf
BA	23	17.5 ± 1.4	0.62 ± 0.02	28 ± 2
	23	16.7 ± 1.6g	0.64 ± 0.04h
	10	11.8 ± 0.4	0.22 ± 0.02	54 ± 5
	0	8.8 ± 0.2	0.104 ± 0.006	85 ± 5
	−10	7.56 ± 0.15	0.040 ± 0.002	190 ± 10
	−20	4.20 ± 0.08	0.0179 ± 0.0009	235 ± 12
	−30	2.67 ± 0.06	0.0074 ± 0.0003	360 ± 20
EB	22i	1.64 ± 0.07	0.030 ± 0.001	55 ± 3
EBj	22i	4.7 ± 0.2	0.100 ± 0.008	47 ± 4

^aBA = benzyl alcohol; EB = ethylbenzene.

^bTemperature in °C.

^cSecond-order rate constants in units of M⁻¹ s⁻¹ with errors at 1σ.

^dRate constant for the undeuterated isotopomer.

^eRate constant for benzyl alcohol-d₇ or ethylbenzene-d₁₀.

^fObserved kinetic isotope effect with error at 1σ.

^gThe substrate was C₆D₅CH₂OH.

^hThe substrate was C₆H₅CD₂OH.

ⁱ± 0.5 °C.

^j(TMP)^+•Fe^IV(O)(Cl) was the oxidant; an increased reactivity of the complex with the "stronger" chloride ligand is known,^9b and a KIE > 20 for reaction of this oxidant with benzyl alcohol was reported.^9a

For oxidation of benzyl alcohol, a large KIE value was observed at ambient temperature, and tremendous KIE values were found at lower temperatures. The series of reactions with four isotopomeric benzyl alcohols at 23 °C confirmed that the KIE effect was from oxidation of the benzylic position. Specifically, C₆D₅CH₂OH reacted with the same rate constant as C₆H₅CH₂OH, and C₆H₅CD₂OH reacted with the same rate constant as C₆D₅CD₂OH. Studies with ethylbenzene demonstrated that the large KIE phenomenon was not specific to the substrate benzyl alcohol nor to the perchlorate counterion in the iron-oxo complex; it was a function of the porphyrin, however (see below).

The temperature dependent data for benzyl alcohol-d₀ and -d₇ require that tunneling is the major process in the oxidation reaction of the undeuterated isotopomer. Because we studied single turnover reactions, the KIEs must be ascribed to the single elementary reaction. For benzyl alcohol-d₇, the reaction appears to have conventional second-order entropic and enthalpic terms of log A = 8.6 and E_a = 12.0 kcal/mol. The reaction of undeuterated benzyl alcohol is the unusual case with an apparent Arrhenius preexponential term (log A = 4.9) that is not possible when compared to the value for the perdeuterated isotopomer as well as an impossibly small activation energy (E_a = 5.0 kcal/mol). Specifically, the E_a term for the "D" reaction is 7.0 kcal/mol greater than that for the "H" reaction, but the classical H/D KIE model requires tunneling contributions when ΔE_a > 1.2 kcal/mol.¹¹

The benzyl alcohol data can be analyzed by various tunneling models.¹² The parabolic energy barrier model described by Bell^12a is shown in Figure 2. An excellent fit was obtained for a chemical reaction with log A = 9.0, E_a(H) = 12.1 kcal/mol, and E_a(D) = 13.3 kcal/mol when combined with a tunneling reaction with a barrier half-width (a) of 0.43 Å. Tunneling is by far the major process for oxidation of C₆H₅CH₂OH; at room temperature, the tunneling reaction represents 93% of the total reaction. In the terminology of Bell,¹² Q = 13 at room temperature.

Figure 2.

Analysis of the kinetic data for oxidations of benzyl alcohol by (TMP)^+•Fe^IV(O)(ClO₄). The symbols are observed rate constants for reactions of benzyl alcohol-d₀ (red) and benzyl alcohol-d₇ (green). The black lines are the apparent Arrhenius functions. The colored lines are predicted rate constants using the parabolic energy barrier model of Bell (ref ^12a) with the parameters listed in the text.

A large tunneling component for the hydroxylation reactions of the benzylic positions we found likely also explains the origin of results reported for oxidations of cyclohexene by (TMP)^+•Fe^IV(O)(Cl).¹³ Widely different Arrhenius functions were found for the competing epoxidation and hydroxylation reactions of cyclohexene,¹³ where the allylic hydroxylation reaction had a remarkably small activation energy, apparently because hydroxylation involved a significant tunneling component. The authors noted that computational results for oxoiron(IV) porphyrin radical cations are not reliable because they do not account for the differences between entropic versus enthalpic control in the competing reactions.¹³ That conclusion may be qualified because the tunneling component expressed in the hydroxylation reaction by the TMP complex is not expressed to the same extent with other porphyrins.

The effect of porphyrin identity on the tunneling reaction can be evaluated in single-turnover (absolute kinetics) or catalytic (competition kinetics) reactions. In single-turnover studies, the KIEs for (Por)^+•Fe^IV(O)(X) oxidations change dramatically with the porphyrin macrocyclic ligand.10 The KIE for oxidations of ethylbenzene-d₀ and -d₁₀ by (TPFPP)^+•Fe^IV(O)(ClO₄), was only 4.4,⁹ much smaller than the value of 55 found for the TMP species. Similarly, the KIEs found in oxidations of benzyl alcohol-d₀ and -d₇ by three oxoiron(IV) porphyrins containing electron-withdrawing aryl groups were small (4.5 to 5.7).¹⁴

The same trends are found in catalytic processes. For a series of catalytic oxidations of benzyl alcohol-d₀ and -d₇ mixtures in CH₃CN at 0 °C with mCPBA as the sacrificial oxidant, we found k_H/k_D = 30 ± 3 for (TMP)Fe^III(ClO₄), k_H/k_D = 11 ± 2 for (TPP)Fe^IIICl, and k_H/k_D = 3.9 ± 0.8 for (TPFPP)Fe^IIICl. Similarly, the KIEs in catalytic oxidations of cyclohexane-d₀ and -d₁₂ by (TMP)Fe^III(Cl) are large,^7,15 whereas those for the cyclohexane oxidations catalyzed by halogen-containing aryl-substituted porphyrin-iron(III) complexes are small.¹⁶ The reduction of the apparent KIEs in the catalytic processes might result because a non-isotopically sensitive step is partially rate determining in a composite reaction or because a more reactive oxidant¹⁸ is formed in the catalytic cycle.

In either single turnover or catalytic processes, the qualitative trend is that the KIE values decrease with increasing electron-withdrawal by the porphyrin. The relationship appears to be complex, and not due solely to an increased chemical reactivity for the electron-withdrawing porphyrins.¹⁷ For example, the rate constants for oxidations of benzyl alcohol are similar for (TPFPP)Fe^IV(O)(X) (normal KIEs) and (TMP)^+•Fe^IV(O)(X) (large KIEs).^9,19 These results suggest that computational quantification of the phenomenon will be difficult.

In summary, hydroxylation reactions of benzylic C-H positions by (TMP)^+•Fe^IV(O)(X) occur mainly by tunneling at room temperature, resulting in large H/D KIEs similar to those found with some non-heme iron enzyme^3,4 and oxoiron(IV) reactions.⁸ Given the large number of heme-containing enzymes, such tunneling reactions might be important in enzyme-catalyzed hydroxylations where barrier widths for tunneling can be modulated by the enzyme.⁵ The tunneling results provided in this work should serve as a benchmark for computational studies of oxidations by heme-containing enzymes and their models.

Supplementary Material

Supp Info. Supporting Information Available.

Experimental description of kinetic studies and detailed results. This material is available free of charge via the internet at http://pubs.acs.org.