Effect of Brønsted Acid on the Reactivity and Selectivity of the Oxoiron(V) Intermediates in C-H and C=C Oxidation Reactions

Abstract

The effect of HClO₄ on the reactivity and selectivity of the catalyst systems 1,2/H₂O₂/AcOH, based on nonheme iron complexes of the PDP families, [(^Me2OMePDP)Fe^III(μ-OH)₂Fe^III(^MeOMe2PDP)](OTf)₄ (1) and [(^NMe2PDP)Fe^III(μ-OH)₂Fe^III(^NMe2PDP](OTf)₄ (2), toward oxidation of benzylideneacetone (bna), adamantane (ada), and (3aR)-(+)-sclareolide (S) has been studied. Adding HClO₄ (2–10 equiv. vs. Fe) has been found to result in the simultaneous improvement of the observed catalytic efficiency (i.e., product yields) and the oxidation regio- or enantioselectivity. At the same time, HClO₄ causes a threefold increase of the second-order rate constant for the reaction of the key oxygen-transferring intermediate [(^Me2OMePDP)Fe^V=O(OAc)]²⁺ (1a), with cyclohexane at −70 °C. The effect of strong Brønsted acid on the catalytic reactivity is discussed in terms of the reversible protonation of the Fe=O moiety of the parent perferryl intermediates.

1. Introduction

High-valent iron-oxo complexes are generally accepted to be the key intermediates of metalloenzyme-mediated oxidations in living nature, as well as in some bioinspired model catalyst systems based on iron complexes and hydrogen peroxide [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22].

For nonheme, iron-containing enzymes, both iron(IV)- and iron(V)-oxo species have been proposed as active intermediates [11,12,13,14,15,16,17,18,19,20,21,22,23]. However, while the involvement of the iron(IV)-oxo intermediates in enzymatic oxidations is firmly established, there are still no experimental data witnessing the participation of the nonheme iron(V)-oxo intermediates in these oxidations [18,19].

In contrast to enzymatic systems, for model nonheme iron-based systems, there have been several examples of spectroscopically characterized iron(V)-oxo intermediates relevant to catalytic transformations (Figure 1) [24,25,26,27,28,29,30,31,32,33]. The comparison of the reactivities of nonheme Fe^IV=O and Fe^V=O species bearing the same macrocyclic ligand bTAML (Figure 1) has shown that even after correcting for the pH difference, the second-order rate constant for benzyl alcohol oxidation by Fe(V)=O at pH 7 is 2500 times higher than that for Fe(IV)=O at pH = 12 [34]. The reactivity studies of Fe^V=O and Fe^IV=O species supported by the same tetradentate N-donor PyNMe₃ ligand (Figure 1) have shown that Fe^V=O accomplishes hydrogen atom transfer (HAT) from C-H groups 4 to 5 orders of magnitude faster than Fe^IV=O [35].

It was established that Brønsted and Lewis acids are able to enhance the oxidizing power of nonheme Fe^IV=O complexes [36,37,38,39,40,41,42,43,44]. It was also shown that acids trigger O-O bond heterolysis of nonheme Fe^III(OOH) precursor species to facilitate the formation of the active intermediates capable of hydroxylating strong C-H bonds [45,46]. However, those intermediates (presumably Fe^V=O) have not been observed spectroscopically.

It was shown previously that the catalyst system 1/H₂O₂/AcOH, where 1 is the [(^Me2OMePDP)Fe^III(μ-OH)₂Fe^III(^MeOMe2PDP)](OTf)₄ complex (^Me2OMePDP is 3,5-Me₂,4-OMe substituted (S,S)-PDP ligand) and AA is acetic acid, exhibits the oxoiron(V) intermediate 1a (g₁ = 2.07, g₂ = 2.01, g₃ = 1.96) [47]. This intermediate can be assigned to the low-spin [(^Me2OMePDP)Fe^V=O(OAc)]²⁺ species on the basis of the identity of its EPR spectrum to that of the Fe^V=O intermediate observed in the catalyst system [(MeO-PyNMe₃)Fe^II(CF₃SO₃)₂]/CPCA (CPCA-cyclohexyl peroxycarboxylic acid) (Figure 1). The assignment of the Fe^V oxidation state of the latter intermediate was supported by Mössbauer spectroscopy [32]. In the figure above, 1a directly oxidizes alkenes [47], alkanes [48] and arenes [49] at −70 °C.

The system 2/H₂O₂/AA (2 is [(^NMe2PDP)Fe^III(μ-OH)₂Fe^III(^NMe2PDP](OTf)₄, complex (^NMe2PDP is 4-NMe₂ substituted (S,S)-PDP ligand) displays the EPR spectrum of the high-spin intermediate 2a (g₁ = 4.30, g₂ = 3.69, g₃ = 1.96). In the figure above, 2a, which reacts with cyclohexene and cyclohexane at −40 °C, has been assigned to the high-spin (S = 3/2) oxoiron(V) species [(^NMe2PDP)Fe^V=O(OAc)]²⁺ (Figure 1) [50,51,52].

Figure 1. Spectroscopically characterized oxoiron(V) complexes [24,26,31,32,47,51].

Figure 1. Spectroscopically characterized oxoiron(V) complexes [24,26,31,32,47,51].

In this work, we scrutinized the effect of Brønsted acid (HClO₄) on the oxidation of C=C and C-H groups of various organic substrates (Figure 2) by the catalyst systems 1/H₂O₂/AA and 2/H₂O₂/AA. The observed correlations between the stability and reactivity of 1a and 2a and the chemo-, regio-, and stereoselectivities of the corresponding catalyst systems are discussed.

2. Results

2.1. Effect of HClO₄ on the Catalyst Systems 1/H₂O₂/AA and 2/H₂O₂/AA: EPR Data

As was shown previously [47], the putative oxoiron(V) intermediate [(^Me2OMePDP)Fe^V=O(OAc)]²⁺ (1a) can be detected by EPR in the catalyst system 1/H₂O₂/AA ([1] = 0.02 M) at low temperature (Figure 3a).

The intermediate 1a is unstable and decays with the first-order rate constant k₁ = (1.8 ± 0.2) × 10⁻³ s⁻¹ at −70 °C. This decay of 1a is accelerated by the addition of cyclohexane. The corresponding second-order rate constant k₂ for the reaction of 1a with cyclohexane at −70 °C is (2 ± 0.4) × 10⁻³ M⁻¹ s⁻¹ [48]. Adding 2 equiv. (vs. Fe) of HClO₄ to the catalyst system 1/H₂O₂/AA at −70 °C reduces the stability of 1a, resulting in a threefold increase of the rate constant of 1a self-decay (k₁ = (6 ± 0.5) × 10⁻³ s⁻¹). It is worth noting that, while affecting the stability of 1a, adding HClO₄ does not visibly change the parameters of its EPR spectrum, which is evidence that HClO₄ does not interfere with the first coordination sphere of the central atom. Adding 5 equiv. (vs. Fe) of HClO₄ to the catalyst system 1/AcOOH/AA/C₆H₁₂ (1:3:10:10) at −70 °C increases by factor of 2.5 the second-order rate constant k₂ of the reaction of 1a with cyclohexane, demonstrating the positive effect of HClO₄ on the C-H oxidation reactivity of 1a.

The catalyst system 1/H₂O₂/HClO₄ = 1:3:2, containing no acetic acid, exhibits a novel intermediate 1x. The EPR spectrum of 1x (g₁ = 2.26, g₂ = 2.04, g₃ = 1.83, Figure 3b) markedly differs from that of 1a (g₁ = 2.07, g₂ = 2.01, g₃ = 1.96, Figure 3a). The intermediate 1x is more stable than 1a and decays with the self-decay rate constant k = (1 ± 0.2) × 10⁻³ s⁻¹ only at −40 °C, while 1a decays with a comparable rate at −70 °C [48]. The EPR parameters of 1x (g₁ = 2.26, g₂ = 2.04, g₃ = 1.83) are rather close to those of activated bleomycin (g₁ = 2.26, g₂ = 2.17, g₃ = 1.94) [49], which is a well-established ferric hydroperoxo complex. The intermediate 1x could be assigned to a ferric hydroperoxo complex [(^Me2OMePDP)Fe^III(OOH)(CH₃CN)]²⁺; In the presence of acetic acid, 1x would rapidly exchange its CH₃CN with AcO⁻ and further give 1a. In agreement with its ferric hydroperoxo nature, 1x is inert toward cyclohexane at −40 °C (whereas 1a reacts with this substrate even at −70 °C). These data suggest that additives of HClO₄ can affect the catalytic properties of the system 1/H₂O₂/AA by enhancing the reactivity of the oxidizing intermediate [(^Me2OMePDP)Fe^V=O(OAc)]²⁺ (1a). The contribution of the intermediate [(^Me2OMePDP)Fe^III(OOH)(CH₃CN)]²⁺ (1x) to the catalytic reaction should not be significant, since 1x cannot compete with 1a in aliphatic C-H oxidations.

Attempts to detect reactive iron-oxygen intermediates in the catalyst system 2/H₂O₂/HClO₄ using EPR spectroscopy were unsuccessful, whereas intermediate 2a could be readily observed in the catalyst system 2/H₂O₂/AA at −40 °C [53,54]. Previously, it was established that the active species of the catalyst system 2/H₂O₂/AA = 1:3:10 was the high-spin (S = 3/2) iron-oxygen intermediate with a proposed structure [(^NMe2PDP)Fe^V=O(OAc)]²⁺ (2a) (g₁ = 4.36, g₂ = 3.69, g₃ = 1.96) (Figure 4a) [50,51,52,53,54]. The intermediate 2a decays with the rate constant k = (2 ± 0.2) × 10⁻³ s⁻¹ at −40 °C [52]. Like in the case of 1a, the presence of HClO₄ does not alter the EPR spectrum of 2a (Figure 4b). However, adding HClO₄ (2 equiv. vs. Fe) accelerates the self-decay of 2a (k = (5 ± 0.5) × 10⁻³ s⁻¹ at −40 °C, ([HClO₄] = 0.08 M). The high-spin intermediate 2a is less reactive toward C=C and C-H groups than the low-spin intermediate 1a: while 1a reacts with cyclohexene and cyclohexane at −70 °C [48], 2a displays comparable reactivity toward these substrates only at −40 °C [52,53].

2.2. Enantioselective Oxidation of Benzylideneacetone by the Catalyst Systems 1/H₂O₂/AA and 2/H₂O₂/AA

The catalyst systems 1/H₂O₂/HClO₄, 1/H₂O₂/AA, and 1/H₂O₂/AA/HClO₄ have been compared in the enantioselective oxidation of benzylideneacetone (bna) (

Table 1. The effect of HClO₄ on the asymmetric epoxidation of benzylideneacetone with H₂O₂ in the presence of complexes 1 and 2 ^a.

Entry	Catalyst	Carboxylic Acid	Strong Acid	Epoxide Yield (%)	ee (%)
1	1	—	HClO4	52	34
2	1	AA	—	61	35
3	1	AA	HClO4	89	36
4	1	EHA	—	44	63
5	1	EHA	HClO4	13	66
6	2	—	HClO4	12	46
7 b	2	AA	—	11	59
8	2	AA	HClO4	5	51
9	2	EHA	—	58	85
10	2	EHA	HClO4	2	— c

^a At 0 °C, [catalyst]:[substrate]:[H₂O₂]:[acetic acid]:[HClO₄] = 1 μmol:100 μmol:200 μmol:55 μmol:10 μmol; oxidant was added by a syringe pump over 30 min, and the mixture was stirred for an additional 2.5 h, followed by NMR and HPLC analysis. ^b From ref. [54]. ^c Not measured.

). It can be seen that the enantioselectivities of these catalyst systems are rather close (34–36% ee, entries 1–3, Table 1). This result has been rather unexpected because different intermediates were detected in the catalyst systems 1/H₂O₂/HClO₄ and 1/H₂O₂/AA (1x and 1a, respectively). Apparently, 1x rapidly converts into the intermediate [(^Me2OMePDP)Fe^V=O(S)]²⁺ (1a′, where S = CH₃CN or H₂O) under the conditions of the catalytic experiment (0 °C). We believe that intermediates [(^Me2OMePDP)Fe^V=O(OAc)]²⁺ (1a) and 1a′ have close structures and therefore, should display close epoxide yields and enantioselectivities in bna oxidation (Table 1).

It is worth noting that the yield of the oxidation product is reasonably higher for the 1/H₂O₂/AA/HClO₄ system than for the systems 1/H₂O₂/HClO₄ and 1/H₂O₂/AA (89% vs. 52–61%, entries 1 and 2 vs. entry 3, Table 1). The likely reason of this difference is the HClO₄-induced reactivity enhancement of the intermediate 1a, operating in the system 1/H₂O₂/AA/HClO₄.

Replacing acetic acid with 2-ethylhexanoic acid (EHA) leads to a larger difference between the enantioselectivities of the systems 1/H₂O₂/HClO₄ and 1/H₂O₂/EHA/(HClO₄) (Table 1, entries 4 and 5 vs. 1), thus clearly witnessing that carboxylic acid is present in the structure of the active perferryl species at the step of enantioselective oxygen transfer in both systems (1/H₂O₂/EHA and 1/H₂O₂/EHA/HClO₄). The unexpectedly low epoxide yield in the latter case (Table 1, entry 5) may reflect the possibility that the effect of HClO₄ on the intermediate [(^Me2OMePDP)Fe^V=O(OC(O)C₇H₁₅)]²⁺ (1a^EHA) [55] is more destabilizing than activating towards selective epoxidation, thus resulting in rapid catalyst degradation after only a few catalytic turnovers.

The epoxide yield in the catalyst system 2/H₂O₂/AA/bna is much lower than that in the catalyst system 1/H₂O₂/AA/bna (11% vs. 61%, entry 7, Table 1 vs. entry 2, Table 1), which most likely reflects a higher contribution of unproductive H₂O₂ decomposition in the case of 2/H₂O₂/AA/bna, operating via the less reactive intermediate 2a. The lower reactivity of 2a compared to 1a is in line with the higher bna epoxidation enantioselectivity of the system 2/H₂O₂/AA, compared to the 1/H₂O₂/AA system (59% ee vs. 35% ee, entry 7, Table 1 vs. entry 2, Table 1). Adding HClO₄ deteriorates the bna epoxidation enantioselectivity (from 59% to 51% ee) and reduces the epoxide yield (cf. entry 8 vs. 7, Table 1) at the same time. If one uses EHA instead of AA, the drop of the epoxide yield is more pronounced (from 58% to only 2%, cf. entry 10 vs. 9, Table 1). This negative effect is even more significant than in the system 1/H₂O₂/EHA/(HClO₄)/bna (cf. entries 4 and 5 of Table 1).

Overall, adding HClO₄ (2 equiv. vs. Fe) to the system 1/H₂O₂/AA/bna has been shown to improve the epoxide yield and at the same time, the epoxidation enantioselectivity, whereas for the system 2/H₂O₂/AA/bna, this trend is not the case.

2.3. Regioselective Oxidation of Adamantane by the Catalyst Systems 1/H₂O₂/AA and 2/H₂O₂/AA

The systems discussed in the previous section have been compared in the regioselective oxidation of adamantane (ada) (

Table 2. The effect of HClO₄ on the regioselective oxidation of adamantane with H₂O₂ in the presence of complexes 1 and 2 ^a.

Entry	Catalyst	Carboxylic Acid	Strong Acid	Conversion (%)	1-ol:2-ol:2-one	3°/2° b
1	1	—	HClO4	17.5	15.0:1.8:0.7	18.0
2 c	1	AA	—	59.2	55.1:1.7:2.4	40.0
3	1	AA	HClO4	72.1	67.3:0.5:4.3	42.1
4	2	—	HClO4	9.8	8.1:1.3:0.4	14.3
5 c	2	AA	—	48.4	43.3:4.0:1.1	25.4
6	2	AA	HClO4	57.9	53.1:1.5:2.6	38.8

^a At 5 °C, [catalyst]:[substrate]:[H₂O₂]:[acetic acid] ]:[catalyst]:[HClO₄] = 0.5 μmol:100 μmol:200 μmol:110 μmol:3 μmol; oxidant was added by a syringe pump over 30 min, and the mixture was stirred for an additional 5 h, followed by GC analysis. ^b 3°/2° = 3 × [1-adamantanol]/([2-adamantanol] + [2-adamantanone]). ^c From ref. [53].

). Like with bna epoxidation, the system 1/H₂O₂/AA/HClO₄ demonstrates higher conversion in ada oxidation than the systems 1/H₂O₂/AA and 1/H₂O₂/HClO₄ (72.1% vs. 59.2 and 17.5%, entry 3 vs. entries 2 and 1, Table 2). Noticeably, the adamantane oxidation regioselectivity (reflected by normalized 3°/2° ratio) is higher for the catalyst systems 1/H₂O₂/AA and 1/H₂O₂/AA/HClO₄ (40–42 vs. 18, entries 3 and 2 vs. entry 1, Table 2), which behavior is similar to the progressive bna epoxidation enantioselectivity enhancement when passing from the system 1/H₂O₂/HClO₄ to 1/H₂O₂/AA and further to 1/H₂O₂/AA/HClO₄ (Table 1).

The catalyst system 2/H₂O₂/HClO₄ oxidizes adamantane with low conversion (9.8%, entry 4, Table 2), much lower than for the system 2/H₂O₂/AA (48.4%, entry 5, Table 2). Like for the system 1/H₂O₂/AA (cf. Table 2), adding HClO₄ to the 2/H₂O₂/AA noticeably improves the conversion of adamantane oxidation (from 48.4% to 57.9%, entries 5 and 6, Table 2). The oxidation regioselectivity (3°/2° ratio) also increases when passing from the system 2/H₂O₂/HClO₄ (3°/2° = 14.3) to the system 2/H₂O₂/AA (3°/2° = 25.4) and further to 2/H₂O₂/AA/HClO₄ (3°/2° = 38.8) (entries 4–6, Table 2).

In fact, the above data demonstrate that HClO₄ positively affects both the C-H oxidation regioselectivity and the product yield in the systems 1 or 2/H₂O₂/AA/HClO₄/adamantane (Table 2), which holds considerable practical promise. Interestingly, this increase of regioselectivity is accompanied by the reactivity enhancement of the corresponding perferryl intermediates towards C-H oxidation (of cyclohexane, see Section 2.1). Even though it seems counterintuitive, this apparent violation of the “reactivity-selectivity” principle in C-H oxidation is not unusual for catalyst systems based on biomimetic catalysts of the Fe(PDP) family [50,51,53].

2.4. Chemo- and Regioselective Oxidation of (+)-Sclareolide by the Catalyst Systems 1/H₂O₂/AA and 2/H₂O₂/AA

Previously, (3aR)-(+)-sclareolide (S) was identified as, so far, the only substrate for which a uniform reactivity-selectivity correlation in C-H hydroxylations, mediated by iron-based catalysts of the Fe(PDP) family, has been documented [50,51]. The structures of the previously identified products of S oxidation are shown in Figure 5.

To probe the effect of HClO₄ on the chemo- and regioselectivity of the catalyst systems 1/H₂O₂ and 1/H₂O₂/AA in the oxidation of S, the following conditions were used: 1/H₂O₂/HClO₄/S = 1:350:6:300; 1/H₂O₂/AA/S = 1:350:250:300; and 1/H₂O₂/AA/HClO₄/S = 1:350:250:6:300. The results are shown in

Table 3. The effect of HClO₄ on the product distribution of oxidation of (3aR)-(+)-sclareolide with H₂O₂ in the presence of complexes 1 and 2.

Entry	Cat.	Conv of S (%)	Products Distribution (%)										Regioseletivity (%)
			S3=O	S2=O	S1=O	S2(eq)-OH	S3(eq)-OH	S3(ax)-OH	S1(eq)-OH	S1(ax)-OH	S7(eq)-OH	S2(eq)-OH−3=O	C1	C2	C3	C7
1 a	1	9.6	24.0	18.3	15.9	27.7	3.0	3.2	3.8	—	3.8	—	19.7	46.0	30.5	3.8
2 b	1	20.6	24.3	19.8	12.6	31.1	2.4	2.4	3.0	0.5	3.0	0.9	16.1	50.9	29.1	3.0
3 c	1	50	28.9	37.6	12.9	18.9	—	0.9	—	—	0.8	—	12.9	56.5	29.8	0.8
4 d	2	9.4	19.1	10.6	4.3	53.2	3.2	3.2	2.1	1.0	3.3	—	7.4	63.8	25.5	3.3
5 e	2	29	30.3	24.8	7.7	34.8	—	—	—	—	2.4	—	7.7	59.6	30.3	2.4

Reaction conditions: ^a 1/H₂O₂/HClO₄/S = 1:350:6:300; ^b 1/H₂O₂/AA/S = 1:350:250:300; ^c 1/H₂O₂/AA/HClO₄/S = 1:350:250:6:300; ^d 2/H₂O₂/AA/S = 1:350:250:300; ^e 2/H₂O₂/AA/HClO₄/S = 2:350:250:6:300. Solvent CH₃CN; the oxidant was added by a syringe pump over 30 min at 0 °C, and the mixture was stirred for an additional 2.5 h at 0 °C, followed by a workup and a NMR analysis.

. The addition of HClO₄ to the catalyst system 1/H₂O₂/AA dramatically improves the conversion of S (from 20.6 to 50%, entries 2 and 3, Table 3), which is accompanied by an increase of C2 regioselectivity (from 50.9 to 56.5%, entries 2 and 3, Table 3). At the same time, the C2 hydroxylation chemoselectivity decreases (cf. S_2(eq)-OH/S_2=O ratio drops from 1.57, entry 2, to 0.50, entry 3). Apparently, this drop is caused simply by a more pronounced ketonization of the C2 methylenic group at the much higher conversion. The catalyst system 1/H₂O₂/HClO₄/S = 1:350:6:300 shows poor conversion, which observation is similar to that for the system 1/H₂O₂/HClO₄/ada (cf. entry 1 of Table 2).

The catalyst system 2/H₂O₂/AA/S = 1:350:250:300 demonstrates high regioselectivity towards the C2-methylenic site of S (63.8%, entry 4, Table 3). However, in contrast to the system 1/H₂O₂/AA/S (see above), adding HClO₄ to the sample 2/H₂O₂/AA/HClO₄/S = 1:350:250:6:300 results in a minor but noticeable drop of the C2 regioselectivity (59.6%, entry 5, vs. 63.8%, entry 4 of Table 3). At the same time, the C3 selectivity increases by ca. 5 percent points. In the case of 1/H₂O₂/AA/S (see above), adding HClO₄ significantly improves the substrate conversion (29% vs. 9.4%, entries 5 and 4, Table 3).

3. Discussion

It is particularly interesting to rationalize the molecular mechanism of the simultaneous positive effect of HClO₄ on (1) the C-H oxidation reactivity, (2) product yield, and (3) oxidation regioselectivity of nonheme catalyst systems based on Fe complexes of the PDP family. Previously, it was established that Brønsted acids (such as HOTf and HClO₄) and Lewis acids (such as Sc³⁺) are able to enhance the oxidizing power of nonheme Fe^IV=O complexes [36,37,38,39,40,41,42,43,44]. In some cases, this reactivity enhancement was discussed in terms of acid-triggered O-O bond heterolysis of a nonheme Fe^III(OOH) species, to facilitate formation of the active intermediates capable of hydroxylating strong C-H bonds [45,46]. However, in our case, this explanation is unsuitable, since the addition of HClO₄ to the intentionally generated key perferryl intermediates [(^Me2OMePDP)Fe^V=O(OAc)]²⁺ (1a) and [(^NMe2PDP)Fe^V=O(OAc)]²⁺ (2a) has been shown to increase their self-decay rates and reactivity to cyclohexane (see Section 2.1).

Alternatively, it has been shown that strong acid (HOTf) can protonate the chelating ligand of the nonheme complex [(TAML)Fe^V=O]⁻ at remote positions, thus increasing the electrophilicity of the nonheme iron(V) oxo complex and enhancing its reactivity in oxygen transfer (OT) and electron transfer (ET) reactions [56]. Strictly speaking, accepting such an explanation in our case would have left the question open, how the increased electrophilicity (and hence reactivity) fits together with increased oxidation regio- and enantioselectivity. Nevertheless, we have considered the hypothesis that HClO₄ protonates one of the remote NMe₂ groups of the active species [(^NMe2PDP)Fe^V=O(OAc)]²⁺ (2a), and calculated the protonated state by DFT at the B3LYP/def2-TZVPP (for Fe)/6-311G(d) (for other atoms) level of theory (see Supporting Information for details).

One can see that NMe₂ mono-protonation is a moderately endergonic process (Figure 6), which in principle, might account for the enhanced electrophilicity of [(^HNMe2PDP)Fe^V=O(OAc)]³⁺ (2a^NH+). However, the calculated Gibbs energies suggest that the equilibrium constant for its formation

$${{\displaystyle \mathrm{K}}}_{\mathrm{eq}}=\frac{\left[\left({}^{\mathrm{HNMe}2}\mathrm{PDP}\right){{\displaystyle \mathrm{Fe}}}^{\mathrm{V}}={\displaystyle \mathrm{O}}{{\displaystyle \left(\mathrm{OAc}\right)}}^{3+}\right]\left[{\displaystyle {{\displaystyle \mathrm{ClO}}}_4^{-}}\right]}{\left[\left({}^{\mathrm{HNMe}2}\mathrm{PDP}\right){{\displaystyle \mathrm{Fe}}}^{\mathrm{V}}={\displaystyle \mathrm{O}}{{\displaystyle \left(\mathrm{OAc}\right)}}^{2+}\right]\left[{{\displaystyle \mathrm{HClO}}}_4\right]}$$

(1)

should not exceed 1 × 10⁻² (at T = 273 K), which apparently rules out significant contribution of 2a^NH+ to the catalytic reaction. Furthermore, as mentioned above, this hypothesis does not bring additional understanding to the selectivity enhancement upon protonation.

On the other hand, one could consider protonation of the terminal oxygen atom of [(^NMe2PDP)Fe^V=O(OAc)]²⁺ (2a) with HClO₄ to form [(^NMe2PDP)Fe^IV-OH(OAc)]³⁺. Related models were invoked to explain the modulation of the catalytic properties of the ferryl complex[(N₄Py)Fe^IV(O)]²⁺ by the additives of HOTf and HClO₄ [40,41]. Such protonation has been found to be exergonic (for both the S = 3/2 and S = 1/2) spin states (Figure 6), thus suggesting that adding HClO₄ should convert 2a into predominantly [(^NMe2PDP)Fe^IV-OH(OAc)]³⁺ (Supplementary Materials Figure S2). The latter species, owing to being protonated, may be expected to possess higher electrophilicity than the parent intermediate [(^NMe2PDP)Fe^V=O(OAc)]²⁺ (2a), yet perhaps be incapable of breaking C-H bonds directly. We believe that [(^NMe2PDP)Fe^IV-OH(OAc)]³⁺ (2a^OH+) could actually be considered as a reservoir of the active species. Being in fast dynamic equilibrium with the parent perferryl intermediate 2a, the protonated reservoir species 2a^OH+ effectively “stabilizes” (in terms of decreasing free energy) the parent intermediate 2a. This stabilization, in accordance with the Hammond−Leffler principle, should lead to more product-like transition states, thus resulting in higher oxidation selectivity.

In Supplementary Materials Figure S2, selected partial bond orders and spin densities for the intermediates ⁴[(^NMe2PDP)Fe^IV-OH(OAc)]³⁺ (⁴2a^OH+) and ²[(^NMe2PDP)Fe^IV-OH(OAc)]³⁺ (²2a^OH+) are presented and compared with those of the parent intermediate [(^NMe2PDP)Fe^V=O(OAc)]²⁺ (2a) [50]. The electronic structures of the protonated-at-terminal oxygen intermediates ⁴2a^OH+ are best described as iron(IV)-hydroxo complexes, with the S = 1 at the Fe^IV center, coupled with the S = 1/2 located at the ligand, ferromagnetically in the case of ⁴2a^OH+ or antiferromagnetically in the case of ²2a^OH+_.

Previously, it was reported that the protonation of the [(N₄Py)Fe^IV(O)]²⁺ complex, with HClO₄ at the terminal Fe=O, eliminates the primary KIE for C-H bond breaking (mesitylene/d₁₂-mesitylene: k_H/k_D drops from 31 to 1.0), which was interpreted as a changeover of the mechanism from direct hydrogen atom transfer (HAT) to proton-coupled electron transfer (PCET) [40]. We have compared competitive oxidations of cyclohexane/d₁₂-cyclohexane in the absence and in the presence of HClO₄ and witnessed similar KIE values in all cases, which is characteristic of metal-mediated, rate-limiting C-H bond breaking (Supplementary Materials Table S1). This result indicates that in the case of the intermediates 1a and 2a, adding HClO₄ does not lead to a mechanism changeover, thus additionally corroborating our proposal of the role of their protonated forms, 1a^OH+ and 2a^OH+, as reservoirs of the true active species.

4. Materials and Methods

4.1. Materials

All chemicals and solvents were purchased from Aldrich, Acros Organics or Alfa Aesar and were used without additional purification. Iron complexes [(^Me2OMePDP)₂Fe^III₂(μ-OH)₂](OTf)₄ (1) [47] and [(^NMe2PDP)₂Fe^III₂(μ-OH)₂](OTf)₄ (2) [52] were prepared as described.

4.2. Instrumentation

EPR spectra (−196 °C) were measured in 3 mm quartz tubes on a CMS 8400 EPR spectrometer at 9.4–9.5 GHz, a modulation frequency of 100 kHz, and a modulation amplitude of 5 G. Experiments were conducted in a quartz finger Dewar filled with liquid nitrogen. EPR signals were quantified by double integration, with a frozen solution of chromium(III) acetylacetonate as a standard at −196 °C.

NMR spectra were measured on a Bruker Avance 400 NMR spectrometer at 400.13 MHz (¹H) and 100.61 MHz (¹³C) in 5 mm and 10 mm cylindrical glass tubes at room temperature, using CDCl₃ as the solvent.

The epoxide yield and enantiomeric excess values were measured on a Bruker Avance 400 NMR spectrometer and Shimadzu LC-20 chromatograph (with Chiralcel OB-H chiral stationary phase), respectively, as previously reported [57]. Experimental uncertainty of ee measurements did not exceed ± 1%. The yields of adamantane oxidation products were determined using a DB-WAX capillary column [30 m × 0.25 mm × 0.25μm, He carrier gas] on an Agilent 6890N gas chromatograph with a flame-ionization detector, with an uncertainty of 2% [58].

4.3. Sample Preparation for EPR Measurements

Using a gas-tight microsyringe connected with a polyethylene capillary, an appropriate amount of H₂O₂ in 0.05 mL of a 1.8:1 CH₂Cl₂/CH₃CN mixture was added to the solution (0.15 mL) of the ferric complex and acetic acid at −70 °C directly in a quartz EPR tube (d = 3 mm). After stirring for 3 min with a polyethylene capillary at the required temperature, the sample was frozen by immersion in liquid nitrogen, and the EPR spectrum was measured at −196 °C.

4.4. General Catalytic Oxidation Procedure

The organic substrate and carboxylic acid were added to the solution of the iron complex in CH₃CN (0.4 mL), and the mixture was thermostated at 0 °C. Then, a calculated amount of 30% H₂O₂ was injected by a syringe pump over 30 min upon stirring at 0 °C. The resulting mixture was stirred for 2.5–5h at 0–5 °C (see Table 1, Table 2 and Table 3 for details). Iron was removed by short-column purification (eluent: acetone). Volatiles were evaporated, and the reaction mixtures were analyzed by NMR and HPLC (for Table 1, refs. [54,57] for details), GC (for Table 2, ref. [53] for details), or NMR (for Table 3, ref. [51] for details). In the absence of either the catalyst or H₂O₂, the oxidations did not occur.

5. Conclusions

The effect of strong Brønsted (HClO₄) acid on the reactivity and selectivity of the catalyst systems 1 and 2/H₂O₂/AcOH, based on nonheme iron complexes of the PDP family, [(^Me2OMePDP)Fe^III(μ-OH)₂Fe^III(^MeOMe2PDP)](OTf)₄ (1) and [(^NMe2PDP)Fe^III(μ-OH)₂Fe^III(^NMe2PDP)](OTf)₄ (2), toward benzylideneacetone (bna), adamantane (ada), and (3aR)-(+)-sclareolide (S) has been examined. Adding HClO₄ to the catalyst systems 1 and 2/H₂O₂/AcOH results in an enhancement of the catalytic efficiencies (turnover numbers or product yields) of these systems in C-H oxidation reactions and at the same time, increases the oxidation regioselectivity.

For the system 1/H₂O₂/AcOH, a similar, simultaneous increase of catalytic efficiency and (enantio)selectivity upon addition of HClO₄ has been documented in the asymmetric epoxidation of bna. EPR spectroscopic monitoring has witnessed a threefold increase of reactivity (second-order rate constant) of the intermediate [(^Me2OMePDP)Fe^V=O(OAc)]²⁺ (1a) towards cyclohexane at −70 °C upon the addition of HClO₄. At the same time, the spectral parameters of 1a did not noticeably change, thus witnessing that HClO₄ does not affect the first coordination sphere of the central atom. DFT modeling corroborates the hypothesis that the positive effect of strong acid could be due to the protonation of the terminal Fe=O moiety of the parent perferryl intermediates to form species of the type [(L)Fe^IV-OH(OAc)]³⁺, which should be more electron-deficient than the parent perferryl species [(L)Fe^V=O(OAc)]²⁺, yet may be not capable of breaking strong aliphatic C-H bonds directly. Instead, they could play the role of the reservoir of the active species, thus effectively stabilizing the active perferryl intermediates. This stabilization would lead to more product-like transition states, thus ensuring higher oxidation selectivity, whereas the increased electrophilicity of [(L)Fe^IV-OH(OAc)]³⁺ leads to a higher apparent reactivity toward C(sp³)-H groups of aliphatic substrates.

We believe that the above approach to the improvement of the catalytic reactivity of catalyst systems, based on the complexes of the PDP family, holds considerable promise for the future, and the proposed model of the effect of HClO₄ would contribute to disclosing the molecular mechanisms of the action of strong Brønsted acids on the catalytic reactivity of nonheme systems, thus extending the existing mechanistic landscape. Related studies, focused on establishing the peculiarities of the effect of Lewis acids on similar iron-based catalyst systems, are underway in our laboratory.