Huatian
Shi‡
,
Lin
Cheng‡
,
Yi
Pan
,
Chi-Keung
Mak
*,
Kai-Chung
Lau
* and
Tai-Chu
Lau
*
Department of Chemistry, City University of Hong Kong, Tat Chee Avenue, Kowloon Tong, Hong Kong, China. E-mail: bhtclau@cityu.edu.hk; kaichung@cityu.edu.hk; chikmak6@cityu.edu.hk
First published on 13th September 2022
The activation of metal-oxo species with Lewis acids is of current interest. In this work, the effects of a weak Brønsted acid such as CH3CO2H and a weak Lewis acid such as Ca2+ on C–H bond activation by KMnO4 have been investigated. Although MnO4− is rather non-basic (pKa of MnO3(OH) = −2.25), it can be activated by AcOH or Ca2+ to oxidize cyclohexane at room temperature to give cyclohexanone as the major product. A synergistic effect occurs when both AcOH and Ca2+ are present; the relative rates for the oxidation of cyclohexane by MnO4−/AcOH, MnO4−/Ca2+ and MnO4−/AcOH/Ca2+ are 1:
73
:
198. DFT calculations show that in the active intermediate of MnO4−/AcOH/Ca2+, MnO4− is H-bonded to 3 AcOH molecules, while Ca2+ is bonded to 3 AcOH molecules as well as to an oxo ligand of MnO4−. Our results also suggest that these synergistic activating effects of a weak Brønsted acid and a weak Lewis acid should be applicable to a variety of metal-oxo species.
Various metal-oxo species are readily activated by strong Lewis acids such as BF3, Sc(OTf)3 and other M3+ ions toward the oxidation of organic substrates.9–19 There are also a number of studies on the interaction of relatively weak Lewis acids such as Ca2+ and other group II ions with metal-oxo species, and as expected their activating effects are much smaller than those of Sc(OTf)3 or other M3+ ions.12,15,20–24
We recently reported the use of a weak Brønsted acid such as acetic acid in combination with a weak Lewis acid such as a group II ion to activate RuO4−.25 We found a remarkable cooperative activating effect of these two acids on RuO4− toward the oxidation of alkanes. The oxo ligands in RuO4− are basic and can be protonated by AcOH to generate [RuO3(OH)]OAc−, which readily abstracts H atoms from alkanes. In the presence of Ca2+, a more active intermediate is formed with [RuO3(OH)]OAc−, in which Ca(II) binds to both OAc− (which is H-bonded to OH) and an oxo ligand. In this work, we have carried out experimental and theoretical investigations on the use of AcOH and Ca2+ to activate MnO4− toward cyclohexane oxidation. MnO4− was chosen because of the relevance of Mn in the OEC of photosystem II. Remarkably, a similar synergistic activating effect of AcOH and Ca2+ is also observed for this oxoanion, although MnO4− is much less basic than RuO4− and is not protonated by AcOH. Our results suggest that the synergistic effects of a weak Brønsted and a weak Lewis acid should be applicable to a variety of metal-oxo species.
Kinetics studies were carried out at constant [KMnO4] (0.011 M) and [cyclohexane] (1.0 M) by monitoring the growth of cyclohexanone using a GC-FID (gas chromatograph-flame ionization detector). The increase in [cyclohexanone] follows pseudo-first-order kinetics (Fig. 1a), and the pseudo-first-order rate constant kobs at [AcOH] = 3.0 M was found to be 4.3 × 10−6 s−1. kobs increases with increasing [AcOH] (3–9 M), and the plot of kobsversus [AcOH]2 is linear (Fig. 1b), suggesting that under these conditions the active oxidizing species involves MnO4− binding to two molecules of AcOH. The reaction can be represented by eqn (1) and (2).
![]() | (1) |
![]() | (2) |
Under the condition that K1 ≪ 1, the rate-law is given by eqn (3).
d[C6H10O]/dt = kobs[MnO4−] = K1ka[AcOH]2[c-C6H12][MnO4−] | (3) |
From the slope of Fig. 1b, K1ka = (4.73 ± 0.19) × 10−7 M−2 s−1 at 23.0 °C.
The kinetic isotope effect for cyclohexane oxidation determined by competitive oxidation of an equimolar mixture of c-C6H12 and c-C6D12 was found to be 7.5 ± 0.1, indicating C–H bond cleavage in the rate-limiting step.
Analysis of the product solution by GC-FID and GC-MS revealed the formation of cyclohexanone (10.5%) and cyclohexanol (0.8%) in relatively low yields. The yields of products reached a maximum with just 1 equiv. of Ca2+.
Kinetics studies were also carried out at various [Ca(OTf)2] but at constant [KMnO4] (0.011 M) and [cyclohexane] (1.0 M). The increase in [cyclohexanone] follows pseudo-first-order kinetics (Fig. 3a), but saturation kinetics were observed on increasing [Ca(OTf)2] (Fig. 3b), and the plot of 1/kobsversus 1/[Ca(OTf)2] is linear. Kinetic studies were also conducted at various [cyclohexane] (Fig. 4a), while keeping both Ca(OTf)2 and KMnO4 at 0.011 M. The plot of kobsversus [cyclohexane] is linear with an intercept of (2.87 ± 3.3) × 10−4 s−1 (Fig. 4b), which should be due to the decomposition of KMnO4 in the absence of cyclohexane. An independent experiment showed that in the presence of 1 equiv. of Ca(OTf)2, the absorbance of KMnO4 (2.82 × 10−4 M) in CH3CN at around 527 nm gradually decreased with time with a rate constant of (9.5 ± 1.1) × 10−4 s−1 at 23 °C. Such a decomposition also accounts for the low yield of products. The kinetic results are consistent with a pre-equilibrium binding of Ca(II) to MnO4− to form an intermediate, which then oxidizes cyclohexane, as represented by eqn (4) and (5), and the rate law is shown in eqn (6).
![]() | (4) |
![]() | (5) |
![]() | (6) |
On the other hand, at constant [AcOH] (2.79 M) and [cyclohexane] (1.0 M), saturation kinetics were observed on increasing [Ca(OTf)2], and the plot of 1/kobsversus 1/[Ca(OTf)2] is linear (Fig. 6).
These results are consistent with the reaction scheme shown in eqn (7)–(9).
![]() | (7) |
![]() | (8) |
![]() | (9) |
For K3 ≪ 1, the rate law is
![]() | (10) |
The kinetic isotope effect for cyclohexane oxidation by the KMnO4/AcOH/Ca(OTf)2 system was determined to be 5.7 ± 0.1 by competitive oxidation of an equimolar mixture of c-C6H12 and c-C6D12, indicating that C–H bond cleavage is the rate limiting step.
A comparison of the activating effects of AcOH, Ca(OTf)2 and AcOH + Ca(OTf)2 on cyclohexane oxidation by KMnO4 is shown in Table 1. The rate constant (kobs) for the oxidation of cyclohexane (1.0 M) by KMnO4 (0.011 M) in the presence of AcOH (3.0 M) at 23 °C was found to be 5.73 × 10−6 s−1. When AcOH was replaced by 1 equiv. of Ca(OTf)2 (0.011 M), the rate constant (4.21 × 10−4 s−1) is 73 times faster, but the yield is much lower (10.5%). On the other hand, in the presence of both Ca(OTf)2 and AcOH, the rate (1.14 × 10−3 s−1) increases by 198 fold relative to AcOH alone, and the product yield is also substantially increased to 57%, indicating synergistic effects of AcOH and Ca2+ in the activation of MnO4− toward C–H bond activation.
3.0 M AcOH (275 equiv.) | 1 equiv. Ca(OTf)2 (0.011 M) | 1 equiv. Ca(OTf)2 + 3.0 M of AcOH | |
---|---|---|---|
a Reaction conditions: KMnO4, 0.011 M, cyclohexane, 1.0 M, T = 23 °C. b KMnO4 alone does not oxidize cyclohexane for >48 h at 23 °C. c The final oxidation state of the Mn product after oxidation was determined by iodometric titration. d % Yield was calculated based on KMnO4 acting as a 3-electron oxidant. Cyclohexanol is the minor product in all cases. | |||
k obs (relative rate) | 5.73 × 10−6 s−1 (1) | 4.21 × 10−4 s−1 (73) | 1.14 × 10−3 s−1 (198) |
Oxidation state of Mn productc | 4.1 ± 0.2 | 4.1 ± 0.2 | 4.0 ± 0.2 |
% Yield of cyclohexanoned | 38.8 | 10.5 | 57.3 |
Cyclohexanone/cyclohexanol | 18 ± 1 | 14 ± 1 | 7 ± 1 |
The rate constant for the oxidation of 1.0 M toluene by (nBu4N)MnO4 was reported to be 4.2 × 10−7 s−1 at 25.0 °C.26 As a comparison, we also investigated the oxidation of 1.0 M toluene by KMnO4 in the presence of 1 equiv. Ca(OTf)2 and 3.0 M AcOH, and the rate constant was found to be 5.2 × 10−3 s−1 at 23 °C (Fig. S3†). Based on these data, the oxidation of toluene by KMnO4/Ca2+/AcOH is over 4 orders of magnitude faster than that of MnO4− alone.
Based on CHN elemental analysis, as well as K, Mn and Ca analysis by ICP-AES, it was found that the brown solids obtained from MnO4−/Ca(OTf)2 (1) and MnO4−/Ca(OTf)2/c-C6H12 (2) have similar compositions and are probably a MnIV–μ(O)–μ(Ca) polymer with an empirical formula close to “Ca3Mn7O14(H2O)10(OTf)6(CH3CN)2” (see the ESI†). The reason for the conversion of MnVII to MnIV in the absence of c-C6H12 is not clear, presumably due to oxidation of the solvent.
We have carried out X-ray photoelectron spectroscopy (XPS) of 1 and 2 (Fig. S7 and S8†). In the Mn3s XPS spectra (Fig. S8†), the peak splitting of ΔE = 4.8 eV indicates that the oxidation state of Mn is +4.29
We have also determined the magnetic susceptibility of the solid samples of 1 and 2 using a magnetic balance. The μeff for each Mn was found to be 3.96 and 4.05 μB for 1 and 2, respectively, consistent with that of high spin Mn(IV) with S = 3/2.
No brown precipitate was observed when cyclohexane oxidation was carried out in the presence of AcOH, irrespective of whether Ca(OTf)2 was added. The Mn oxidation state of the brown solution determined by iodometric titration is again 4.0 ± 0.2. Presumably the brown solution contains colloidal MnO2, as observed in alkane oxidation by BF3/MnO4−.18
The potential energy surfaces (PESs) for all calculated reactions are summarized in Fig. S9.† In the c-C6H12 oxidation by MnO4− in CH3CN (Fig. 8), c-C6H12 and MnO4− first bind weakly together to form an intermediate [MnO4⋯C6H12]−, INT1(MnO4−). Hydrogen atom transfer (HAT) then occurs from c-C6H12 to MnO via a transition state, TS1(MnO4−), to give a second intermediate, INT2(MnO4−). The HAT step is rate-determining with a barrier height
of 22.7 kcal mol−1. Such a high barrier suggests that MnO4− will hardly oxidize c-C6H12 at room temperature. The carbon atom in the cyclohexyl radical of INT2(MnO4−) bears a −1.01 electron spin, consistent with the spin distribution for a HAT. In the next step, the cyclohexyl radical binds to another Mn
O to generate an alkoxo intermediate [MnO2(OH)(OC6H11)]−via TS2(MnO4−). Subsequently, proton transfer from Mn–OH to the alkoxide in INT3(MnO4−) occurs via TS3(MnO4−) to generate the cyclohexanol product. The reaction mechanism is similar to that of c-C6H12 oxidation by RuO4− found in our previous study.25
![]() | ||
Fig. 8 PES and structures for cyclohexane oxidation by [MnO4]− at the B3LYP-D3(BJ)/def2-SVPD level. The relative 298 K Gibbs free energies in acetonitrile are given in kcal mol−1. |
In the presence of an acetic acid molecule (Fig. S10†), AcOH and MnO4− first form INT1(AcOH), [MnO4(AcOH)⋯C6H12]−, through a hydrogen bond. The MnO that is H-bonded to AcOH is elongated to 1.611 Å, while the other three Mn
O bonds become more electrophilic and shorter with distances of 1.587–1.590 Å, compared with the bond length of 1.594 Å in MnO4−. In TS1(AcOH), HAT occurs from c-C6H12 to another free but shorter Mn
O bond (bond length = 1.588 Å). The
of TS1(AcOH) is 20.4 kcal mol−1, which is 2.3 kcal mol−1 lower than that of TS1(MnO4−). The subsequent steps are similar to those in c-C6H12 oxidation by MnO4−.
In the presence of three AcOH molecules, the AcOH molecules form hydrogen bonds with three MnO to give an intermediate, INT1(3AcOH) (Fig. 9). The bond lengths of the H-bonded Mn
O are slightly elongated from 1.591 to 1.601 Å. HAT then occurs from c-C6H12 to the free and more electrophilic Mn
O bond (bond length = 1.580 Å), with a
= 16.8 kcal mol−1. Compared with c-C6H12 oxidation by MnO4−, the
of HAT is further reduced by 5.9 kcal mol−1.
![]() | ||
Fig. 9 PES and structures for cyclohexane oxidation by [MnO4(AcOH)3]− at the B3LYP-D3(BJ)/def2-SVPD level. The relative 298 K Gibbs free energies in acetonitrile are given in kcal mol−1. |
In the presence of [Ca(OTf)]2 (Fig. 10), Ca(II) binds to two of the MnO bonds to form INT1(CaOTf), [MnO4(CaOTf)⋯C6H12]. The two Mn
O bonds that bind to Ca(II) in INT1(CaOTf) are longer (1.611 and 1.615 Å) than the unbound Mn
O bonds (1.576 Å) in INT1(MnO4−), as a result of the electron withdrawing effect of Ca(II). Subsequent HAT from c-C6H12 to an unbound Mn
O occurs with a
of 12.7 kcal mol−1via TS1(CaOTf), which is lower than that of MnO4− and MnO4−/3AcOH by 10.0 and 4.1 kcal mol−1, respectively; consistent with the experimental observation that the accelerating effect of Ca(II) is much higher than that of AcOH.
![]() | ||
Fig. 10 PES and structures for cyclohexane oxidation by [MnO4(CaOTf)] at the B3LYP-D3(BJ)/def2-SVPD level. The relative 298 K Gibbs free energies in acetonitrile are given in kcal mol−1. |
The activation barrier is further lowered when both [Ca(OTf)]+ and AcOH are present. In the presence of one molecule of [Ca(OTf)]+ and three molecules of AcOH, each AcOH is H-bonded to one MnO, while the [Ca(OTf)]+ is bonded to each AcOH and one Mn
O to form a relatively stable intermediate, INT1(CaOTf/3AcOH), as shown in Fig. 11. HAT then occurs from c-C6H12 to the free and shortest Mn
O bond with the lowest
of 9.7 kcal mol−1 (
= 11.5 kcal mol−1 for one Ca(OTf)+ and one AcOH, Fig. S11†), in agreement with the observed synergistic activating effects of AcOH and Ca(II). It should be noted that in this case HAT results in the direct formation of the alkoxo intermediate without going through a cyclohexyl radical intermediate.
![]() | ||
Fig. 11 PES and structures for cyclohexane oxidation by [MnO4(CaOTf)(AcOH)3] at the B3LYP-D3(BJ)/def2-SVPD level. The relative 298 K Gibbs free energies in acetonitrile are given in kcal mol−1. |
was also determined experimentally by performing kinetics studies at 20–50 °C. From the Arrhenius plot
was found to be 10.4 ± 0.6 kcal mol−1, in agreement with the calculated value (Fig. S12†).
We have also calculated the TS barrier heights for hydroxylation via a rebound mechanism and they are higher than those of the TS forming the alkoxo intermediate by 2–3.4 kcal mol−1 (Table S1†).
RuO4− | MnO4− | |
---|---|---|
No additive | 26.8 | 22.7 |
AcOH | 15.2 | 16.8 |
Ca2+ | 18.5 | 12.7 |
AcOH + Ca2+ | 10.8 | 9.7 |
The combination of a weak Brønsted and a weak Lewis acid is a potentially useful strategy for the activation of a metal-oxo species toward oxidation of various organic compounds, particularly for substrates that may contain functional groups that are sensitive to strong acids. It is possible that this synergistic activating effect on metal-oxo species may also occur in biological systems, where only relatively mild Brønsted acids and Lewis acids (metal ions) are present.
Footnotes |
† Electronic supplementary information (ESI) available. See https://doi.org/10.1039/d2sc03089f |
‡ These authors have equal contribution. |
This journal is © The Royal Society of Chemistry 2022 |