Mechanistic Investigation into Olefin Epoxidation with H₂O₂ Catalyzed by Aqua‐Coordinated Sandwich‐Type Polyoxometalates: Role of the Noble Metal and Active Oxygen Position

Abstract

Aqua‐coordinated sandwich‐type polyoxometalates (POMs), {[WZnTM₂(H₂O)₂](ZnW₉O₃₄)₂}ⁿ⁻ (TM=Rh^III, Pd^II, and Pt^II), catalyze olefin epoxidation with hydrogen peroxide and have been well established, and they present an advance toward the utilization of olefins. To elucidate the epoxidation mechanism, we systematically performed density functional calculations. The reaction proceeds through a two‐step mechanism: activation of H₂O₂ and oxygen transfer. The aqua‐coordinated complexes show two distinct H₂O₂ activation pathways: “two‐step” and “concerted”. The concerted processes are more facile and proceed with similar and rate‐determining energy barriers at the Rh‐, Pd‐, and Pt‐containing transition states, which agrees well with the experimental results. Next, the resulting TM−OH−(μ‐OOH) intermediate transfers an O atom to olefin to form an epoxide. The higher reactivity of the Rh‐containing POM is attributed to more interactions between the Rh and hydroperoxo unit. We also calculated all active oxygen positions to locate the most favorable pathway. The higher reactivity of the two‐metal‐bonded oxygen position is predominantly ascribed to its lower stereoscopic hindrance. Furthermore, the presence of one and two explicit water solvent molecules significantly reduces the energy barriers, making these sandwich POMs very efficient for the olefin epoxidation with H₂O₂.

1.  Introduction

The olefin epoxidation with hydrogen peroxide has been extensively studied for several decades and still is a challenge for complicated polymetallic complexes.1, 2, 3, 4, 5 As epoxides are a useful raw material, they have a common application in the synthesis of functionalized organic compounds, such as surfactants, medicine, paints, and so forth.1, 2, 3, 4, 5 Hydrogen peroxide, as an active oxygen‐rich donor, has a relatively cheap price and forms water as it's only by product.6, 7, 8, 9 However, the energy of the O−O bond has reached approximately 50 kcal mol⁻¹, thus the search for an environmentally friendly catalyst has received significant attention.10, 11

Polyoxometalates (POMs), as abundant molecular‐based inorganic materials, are potentially interesting catalysts for the epoxidation of olefins.12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40 Particular attention has been paid to transition‐metal‐substituted polyoxotungstates, which have shown significant catalytic activity and selectivity, such as Venturello‐type anions, {PO₄[WO(O₂)₂]₄}³⁻,15, 16, 17 Keggin‐type18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29 and Tourné sandwich‐type polyoxotungstates.30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40 In the literature, the low‐valent noble transition‐metal‐substituted sandwich‐type polyoxotungstates, {[WZnTM₂(H₂O)₂](ZnW₉O₃₄)₂}ⁿ⁻ (TM=Rh^III, Pd^II, and Pt^II), referred to as 1 Rh, 1 Pd, and 1 Pt (Figure 1), have been reported to be highly active and selective catalysts with a good solvolytic stability for olefin epoxidation with H₂O₂.39

Figure 1.

Polyhedral/ball‐and‐stick representation of [(WZnTM₂(H₂O)₂)(ZnW₉O₃₄)₂]ⁿ⁻ (TM=Rh^III, Pd^II, and Pt^II).

According to previous reports on the epoxidation by POMs, the catalytic reaction proceeded through a two‐step mechanism: 1) activation of H₂O₂ and 2) oxygen transfer; in addition, three distinct H₂O₂ activation mechanism have been reported.24, 25, 26, 27, 28, 29 Musaev and co‐workers reported a divanadium‐substituted POM, [γ‐1,2‐H₂SiV₂W₁₀O₄₀]⁴⁻, which contained a core of [OV−(η‐OH)₂−VO].24 They proposed a peroxo‐type mechanism, that is, the heterolytic cleavage of H₂O₂ forms a [OV−(η‐OOH)(−OH)−VO] peroxo intermediate, and then the oxygen transfer is performed favorably through a ‘water‐assisted’ pathway (Scheme 1).

Scheme 1.

Epoxidation mechanism based on the divanadium‐substituted polyoxometalate, [γ‐1,2‐H₂SiV₂W₁₀O₄₀]⁴⁻.24

For the second type, several computational studies have reported the lacunary‐substituted POM, [γ‐H₂SiW₁₀O₃₆]⁴⁻,25 as well as mono‐ and dititanium‐substituted POMs [PTi(OH)W₁₁O₃₉]⁴⁻ and [Ti₂(OH)₂As₂W₁₉O₆₇(H₂O)]⁸⁻.26, 27, 28, 29 A two‐step mechanism has been proposed. First, the catalytic reaction starts with the heterolytic cleavage of H₂O₂, forming a TM−OOH or TM−OO(μ‐OH) intermediate, and then the alpha‐ or beta‐oxygen atom of the intermediates is transferred to the olefin to form an epoxide (Scheme 2).

Scheme 2.

Epoxidation mechanism based on the mono‐ and ditransition‐metal‐substituted polyoxometlates, [PTM(OH)W₁₁O₃₉]⁴⁻ and [TM₂(OH)₂As₂W₁₉O₆₇(H₂O)]⁸⁻, TM=Ti, V, Zr, Nb, Mo, W and Re.26, 27, 28, 29

For the third aqua‐coordinated POM type, Neumann et al. have studied Rh‐, Pd‐, and Pt‐substituted sandwich‐type POMs {[WZnTM₂(H₂O)₂](ZnW₉O₃₄)₂}ⁿ⁻ (TM=Rh^III, Pd^II, and Pt^II).39 They indicated that the activation of H₂O₂ leads to the formation of a stable TM−OH−(μ‐OOH) hydroperoxo‐tungstate intermediate, and subsequently determined the reactivity and selectivity. Interestingly, they observed a similar reactivity for each of these three catalysts. Furthermore, they also found that the bridge‐oxygen atom of the TM−O−W unit (Figure 2) is the active position, and the three‐metal‐bonded (TM, W, and Zn) oxygen atom (O_3M) is more favorable, owing to the increased electrophilicity. However, because the O_3M atom is located inside the TM−O−W triad, the electronic effect will compete with the higher steric hindrance. We know of no reports documenting the epoxidation mechanism of olefin with H₂O₂ catalyzed by aqua‐coordinated sandwich‐type POMs. Thus, we initially hypothesize a mechanism (Scheme 3): step 1, H₂O₂ activation to form the TM−OH−(μ‐OOH) intermediate; step 2, oxygen transfer from the intermediate to the olefin to form an epoxide.

Figure 2.

Four isomers of intermediate 3 Rh with selected bond lengths (in Å). Energy is given in kcal mol⁻¹.

Scheme 3.

Proposed epoxidation mechanism based on {[WZnTM₂(H₂O)₂](ZnW₉O₃₄)₂}ⁿ⁻, TM=Rh(III), Pd(II) and Pt(II).

The main goal of the present paper is to analyze the viability of the epoxidation mechanism of an olefin catalyzed by aqua‐coordinated POMs, {[WZnTM₂(H₂O)₂](ZnW₉O₃₄)₂}ⁿ⁻ (TM=Rh^III, Pd^II and Pt^II), including analyses of the roles of the substituted noble metal, active oxygen position, and solvent. Computational analyses of the catalytic mechanism with larger POM systems, such as those reported in this study, remain a challenge for computational chemistry and will be very helpful for evaluating and improving epoxidation catalysts.

2.  Results and Discussion

2.1.  Structures

The reactants, intermediates, transition states, and products are fully optimized by using the polarizable continuum model (PCM) with water as the solvent; the effect of explicit water molecules will be discussed in the following section. In Rh^III‐containing local minima, the rhodium center is η ⁶‐coordinated, owing to the d⁶ electronic configuration of Rh^III, and all of the Rh−O bonds are approximately 2.1 Å in length. However, in most Pd^II‐ and Pt^II‐containing local minima, owing to the d⁸ electronic configuration of Pd^II and Pt^II, the metal centers are η ⁵‐coordinated and the TM−O_2M(Zn) and TM−O_3M(W) bond lengths reach more than about 2.7 Å (see the Supporting Information). Herein, we use a Rh‐containing POM as the example, and cyclohexene (C₆H₁₀) as the olefin model.

Neumann et al. indicated that the O_3M position is more favorable than the O_2M position upon H₂O₂ activation, owing to its higher electrophilicity.39 However, as mentioned above, there is a competition between stereoscopic and electronic effects. To obtain the most facile pathway, we optimized all four intermediates. In Figure 2, around each Rh^III center, four kinds of oxygen atom: O_2M(W), O_2M(Zn), O_3M(W), and O_3M(Zn) can be attacked by H₂O₂ to form four similar intermediates.

The energy of 3 Rh−O_3M is higher than that of 3 Rh−O_2M by about 9–22 kcal mol⁻¹ (Figure 2), which will lead to a much higher overall energy barrier at the O_3M position in the next oxygen‐transfer step. Similar results are obtained when the central metal is Pd^II or Pt^II (Figures S1 and S2). Thus, O_3M is not expected to be the active position, that is, the stereoscopic effect will play a more important role. In the following sections, we initially use the O_2M(Zn) to discuss the epoxidation mechanism, and the remaining active oxygen positions will subsequently be studied.

2.2.  General View of the Energy Profile and Mechanism

As proposed in Scheme 3, this epoxidation reaction can be performed through a two‐step mechanism: activation of H₂O₂ (step 1) and oxygen transfer (step 2). For step 1, two possible methods can form the TM−OH−(μ‐OOH) intermediate, that is, step 1 a, dissociation of the water ligand and subsequent cleavage of the HO−OH bond, or step 1 b, cleavage of the HO−OH bond accompanied with H abstraction and OH addition. In step 2, the TM−OH−(μ‐OOH) intermediate transfers an oxygen atom to the olefin to form an epoxide and to regenerate the catalyst. The calculated potential energy profile is exhibited in Figure 3, showing the electronic energy and free energy. In the following sections, we will describe each step in detail.

Figure 3.

Potential energy profile for olefin epoxidation with Rh‐containing POM. Red dashed lines correspond to the less favorable path. Numbers given in bold and plain are calculated relative to the O_2M(Zn) and O_2M(W) positions, respectively. Electronic energy is given in kcal mol⁻¹ (values in parentheses correspond to free energy).

2.3.  Activation of H₂O₂

As mentioned above, there are two methods that can activate H₂O₂ (Scheme 3). In step 1 a, starting from reactant 1 Rh, coordinated H₂O is released to form 2 Rh. Owing to the similar geometries of 1 Rh, 1 Pd, and 1 Pt, we only softly scanned the dissociation of the water ligand for 1 Rh (Figure S3). It is shown that this step is an endothermic process, by 16.9 kcal mol⁻¹, and proceeds without an energy barrier. Next, cleavage of the HO−OH bond simultaneously generated two OH moieties bound to Rh^III and the O_2M atom, respectively, leads to the formation of 3 Rh. It proceeds with a 25.6 kcal mol⁻¹ energy barrier at TS1Rh (Figure 4), whereas the overall barrier height reaches 42.1 kcal mol⁻¹. A similar behavior is observed when the central transition metal is Pd^II (ΔE=27.8 kcal mol⁻¹, overall ΔE=39.5 kcal mol⁻¹) (Figure S4). Surprisingly, TS1Pt has a much lower energy barrier (8.6 kcal mol⁻¹) (Figure S5). The reason can be seen from the reactive center of TS1Pt, as one increased intramolecular hydrogen bond is found between the proton of the water ligand and the bridge‐oxygen atom of subunit [ZnW₉O₃₄]^[12− (Figure S6), which reduces the energy barrier of TS1Pt. Nevertheless, the overall energy barrier still reaches 31.2 kcal mol⁻¹. In an attempt to reduce the energy barrier of step 1 a, we tried many times to locate similar TS1Rh or TS1Pd species with more intramolecular hydrogen bonds, but it always leads to the same results. Alternatively, to avoid the artificial effect, we investigated the role of explicit water solvent, which is discussed in the following section. Furthermore, we also tried to find a TM−OO‐ or TM−OOH‐type transition state, for example, the substitution of H₂O by H₂O₂. However, many efforts led to failure, owing to the much higher steric hindrance. Therefore, step 1 a is not a favorable pathway for activating H₂O₂.

Figure 4.

Optimized structure of TS1Rh (distances in Å).

Step 1 b is a concerted process (Scheme 3), including the cleavage of a HO−OH bond accompanied by H abstraction and OH addition with two OH moieties, which leads to the formation of 3 Rh and a H₂O molecule. For simplicity of discussion, denote the process: POM−Rh−O¹H¹H²(μ‐O²) + H³O³O⁴H⁴ → POM−Rh−O¹H¹−(μ‐O²O³H³) + H²O⁴H⁴.

Step 1 b is calculated to be endothermic by about 2 kcal mol⁻¹. A stable six‐membered‐ring transition state, TS2Rh (Figure 5), is located with an energy barrier of 25.6 kcal mol⁻¹, and the free energy barrier is 32.3 kcal mol⁻¹. All of the calculated important bond parameters (O³−O⁴=1.87 Å, O²−O³=1.82 Å, and H²−O⁴=1.07 Å) correspond to the cleavage of the HO−OH bond and the formation of 3 Rh and a water molecule. The imaginary normal vibration modes of TS2Rh also show the corresponding scratching vibration of O³−O⁴, O²−O³, and H²−O⁴ bonds.

Figure 5.

Optimized structure of TS2Rh (distances in Å).

Furthermore, Pd‐ and Pt‐containing POMs were also studied. Table 1 collects the electronic and free energy barriers as well as some selected bond parameters. It is shown that the energy barriers of these three transition states are about 26 (ΔE) or 32 kcal mol⁻¹ (ΔG) at the O_2M(Zn) position. Furthermore, as shown in Figures 3, S4, and S5, the energy barrier of step 1 b is higher than that of step 2 by about 2–14 kcal mol⁻¹ (overall ΔE) or 2–9 kcal mol⁻¹ (overall ΔG). Therefore, from an energy point of view, step 1 b is the rate‐determining step, which is in good agreement with the experimental observation, that is, the cleavage of H₂O₂ determines the reactivity.39

Table 1.

Selected bond parameters and energy barriers for TS2Rh, TS2Pd, and TS2Pt.

	TS2Rh		TS2Pd		TS2Pt
	Zn[a]	W	Zn	W	Zn	W
d(O2−O3) [Å]	1.82	1.82	1.84	1.83	1.80	1.86
d(O3−O4) [Å]	1.87	1.86	1.86	1.85	1.85	1.85
d(H2−O4) [Å]	1.07	1.07	1.06	1.09	1.11	1.07
ΔE [kcal mol−1]	25.6	25.5	25.8	30.5	25.5	35.0
ΔG [kcal mol−1]	32.3	31.0	32.7	32.1	35.3	40.8

[a] Zn or W corresponds the O_2M(Zn) or O_2M(W) position, respectively.

2.4.  Oxygen Transfer

This step starts with attack of the olefin on intermediate 3 Rh, and follows the O‐atom transfer to form an epoxide and the catalyst, that is, POM−Rh−O¹H¹−(μ‐O²O³H²) + C₆H₁₀ → POM−RhO¹H¹H²−(μ‐O²) + C₆H₁₀O³. It can be seen that step 2 is typical beta‐oxygen transfer from the bridging hydroperoxo moiety to the olefin. Recently, Poblet and co‐workers have proposed that the mechanism of Ti−OOH through beta‐oxygen has a lower energy barrier for dititanium‐substituted sandwich‐type POMs.28, 29 As could be expected, this process will proceed with a slightly lower energy barrier; the calculated energy barrier was 16.6 (ΔE) or 29.7 kcal mol⁻¹ (ΔG) at the located transition state, TS3Rh (Figure 6).

Figure 6.

Optimized structure of TS3Rh (distances in Å).

In TS3Rh, the O²−O³ bond length is elongated to 1.78 Å, and the H² to O¹ distance decrease from 1.49 to 1.47 Å. It can be seen that the cleavage of the O²−O³ bond accompanies the O³ atom transfers to the olefin, and simultaneously the H² proton transfers to the O¹ atom, which leads to the formation of an epoxide and the catalyst. To verify the located transition state, a frequency calculation was performed. The only imaginary normal vibration of TS3Rh shows that the O³ atom oscillates between the O² atom and two carbon atoms, and the H² atom oscillates between O³ and O¹ atoms, indicating that TS3Rh connects with 3 Rh and the products, that is, the epoxide and the catalyst. This step was calculated to be exothermic by more than 50 kcal mol⁻¹.

The oxygen‐transfer step is performed through nucleophilic attack of the olefin. Thus, the electrons will be transferred from the π_C=C orbital of the olefin to σ*(O²−O³) of the TM−OOH unit. Natural bond orbital (NBO) calculation results (Table 2) show a natural population analysis (NPA) electronic density transfer of about 0.1 e from the olefin to hydroperoxo anions.45 Previous DFT studies have found that the lower the energy of the σ*(O²−O³) orbital, the lower the energy barrier for the alpha‐oxygen‐transfer process.41, 42, 43, 44 Surprisingly, as for the beta‐oxygen‐transfer process, we find that the energy barriers of TS3s do not follow the latter tend at the O_2M−Zn position, that is, TS3Rh always has slightly higher energy barriers. To check this, the electric configuration of the central metal was analyzed. In comparison with Pd^II {[Kr]4d⁸} and Pt^II {[Xe]4f¹⁴5d⁸}, Rh^III {[Kr]4d⁶} has more empty d orbitals, which helps alpha‐O² atom transfer the p‐electron to the beta‐O³ atom, and it induces O²−O³ bond polarization toward the O² atom. As a result, in 3 Rh, the beta‐O³ atom becomes more positively charged in comparison with 3 Pd and 3 Pt (Table 2). Thus, the decreased charge of distal O³ will increase the electrophilicity of hydroperoxo group and the activity of 3 Rh. These reasons also reveal that, in comparison to the cleavage of the O−O bond in the H₂O₂ activation step, the polarization of the O²−O³ bond slightly reduces the energy barriers in the epoxidation step. Besides, the lower energy barrier of TS3Rh also agrees well with experimental observation of a slightly higher yield of epoxide when the reaction is catalyzed by a Rh‐containing POM.39

Table 2.

NBO analysis for 3 Rh, 3 Pd, and 3 Pt and oxygen‐transfer energy barriers.

	3 Rh		3 Pd		3 Pt
	Zn[a]	W	Zn	W	Zn	W
d(O2−O3) [Å]	1.78	1.78	1.78	1.78	1.78	1.79
σ*(O2−O3) [eV]	2.91	3.02	2.81	2.58	2.61	2.48
q NPA(O2) [au]	−0.358	−0.367	−0.397	−0.400	−0.399	−0.397
q NPA(O3) [au]	−0.498	−0.497	−0.523	−0.518	−0.522	−0.516
Δq NPA(O2−O3) [au]	0.14	0.13	0.126	0.118	0.123	0.119
ΔE [b] [kcal mol−1]	16.6	17.3	27.0	21.4	22.7	15.5
ΔG [b] [kcal mol−1]	29.7	30.1	40.5	30.9	36.8	27.5

[a] Zn or W represents the O_2M(Zn) or O_2M(W) position, respectively. [b] ΔE and ΔG were calculated relative to the 3 TM intermediate.

2.5.  Role of the Active Oxygen Position

Around each noble‐metal center, four active oxygen atoms can be attacked by H₂O₂ (Figure 2). However, many efforts have been made to locate a transition state at the O_3M position either for step 1 or step 2, but failed because of the higher stereoscopic hindrance. Besides, as mentioned above, the energy values of intermediates 3 TM−O_3M are higher those of 3 TM−O_2M. As a result, only O_2M positions are available. As step 1 a is unfavorable, we only recalculate step 1 b and step 2. All corresponding geometries were fully optimized at the same level, and details are given in the Supporting Information. In step 1 b, the optimized geometries of transition states at the O_2M−W position are similar as those at the O_2M−Zn position (Figure S7). The free‐energy barrier of TS2−Rh−O_2MW and TS2−Pd−O_2MW were slightly reduced by about 1.0 kcal mol⁻¹; however, the barrier of TS2−Pt−O_2MW increased by about 5 kcal mol⁻¹. Thus, from an the energy point of view, for the Pt‐containing POMs, O_2M−W is not a facile position for olefin attack.

For step 2, the order of the energy barriers of all transition states at the O_2M−W position agrees well with the order of energy for the σ*(O²−O³) orbital (Table 2). For example, the order of energies of the σ*(O²−O³) orbital (O_2M−Zn, 2.91 vs. O_2M−W, 3.02 eV) corresponds well with the order of the energy barriers (TS3Rh−O_2MZn, 16.6 vs. TS3Rh−O_2MW, 17.3 kcal mol⁻¹). However, the overall energy barrier of TS3Pt−O_2MW increases by about 4 kcal mol⁻¹ more than that of TS3Pt−O_2MZn, owing to the energy discrepancy between 3 Pt−O_2MW and 3 Pt−O_2MZn. Therefore, the O_2M−W position is favored for Rh‐ and Pd‐containing POMs, and the process is not facile for the Pt‐containing POM. Nevertheless, from an energy perspective, the step 1 b is still the rate‐determining step.

2.6.  Water‐Assisted Process

As mentioned above, previous theoretical studies proposed that the presence of explicit solvent can significantly reduce the energy barrier caused by intermolecular hydrogen bonds.24, 25 However, as there are 40 water molecules per unit of reactant,39 it is very expensive to calculate this for all solvents. To identify the position of protonation, an NBO calculation was performed.45 It is shown that the four bridge‐oxygen atoms around the reactive center have a slightly higher negative charge distribution (e.g. O_b, −0.71 au vs. O_t, 0.65 au in 1 Rh). Therefore, we calculated the solvent effects by adding one/two explicit water molecules in the vicinity of the reactive center. The energy values are listed in Table 3, and the optimized geometries are given in the Supporting Information (Figures S9, S10, and S11).

Table 3.

Calculated overall energy barriers of the water‐assisted pathways.

	TS2Rh	TS2Pd	TS2Pt	TS3Rh	TS3Pd	TS3Pt
ΔE 1 water [a] [kcal mol−1]	20.6	21.1	20.1	14.1	18.9	10.7
ΔE 2 waters [a] [kcal mol−1]	20.1	20.5	20.0	13.7	18.4	10.4

[a] All electronic energies were calculated relative to reactants 1 TM+H₂O₂.

The inclusion of the first water solvent located above the reactive center (Supporting Information) reduces the overall energy barriers of TS2s or TS3s by about 5 kcal mol⁻¹. However, the addition of the second water solvent only reduced the barriers by about 1 kcal mol⁻¹. Thus, the remaining water solvents may have a similar effect in decreasing the barrier height of the rate‐determining step.

2.7.  Comparison with Experiments

Neumann and co‐workers investigated the olefin epoxidation by using noble‐metal‐substituted sandwich‐type POMs with H₂O₂.39 They indicated that the cleavage of the HO−OH bond in initial ligation or interaction of H₂O₂ with substituted transition metals formed the TM−OH−(μ‐OOH) intermediate, and that it determines the reactivity and selectivity. In particular, Rh^III‐, Pd^II‐, and Pt^II‐substituted POM catalysts possessed similar reactivities. The geometric character of the TM−OH−(μ‐OOH) intermediate is of great importance, as it enables us to propose two possible pathways for the activation of H₂O₂. The values of the calculated energy barriers in the H₂O₂ activation step are not only close to each other, but are also higher than the energy barriers of the oxygen‐transfer step, which is in good agreement with the experimental observations.39 Next, the TM−OH−(μ‐OOH) intermediate transfers an oxygen atom to the olefin to form an epoxide and the catalyst. We find that the interaction between the noble metal and the peroxo unit plays an important role in lowering the energy barrier and the yield of the epoxide, which is in agreement with experimental results.39 Furthermore, Neumann and co‐workers indicated that four active oxygen sites can be attacked by H₂O₂, and that the three‐metal‐bonded O_3M position will be more reactive, owing to its relatively higher electrophilicity. However, on the basis of our calculation results, owing to the higher stereoscopic hindrances of the O_3M position and the higher energy values of the TM−OH−(μ‐O_3MOH) intermediates, the O_2M position is more favorable than the O_3M position. Therefore, future work should focus on decorating the central noble‐metal atom and the coordination environment of the catalytic center.

3.  Conclusions

The cyclohexene epoxidation mechanism has been investigated by using noble‐metal‐substituted sandwich‐type POMs {[WZnTM₂(H₂O)₂](ZnW₉O₃₄)₂}ⁿ⁻ (TM=Rh^III, Pd^II, and Pt^II) with H₂O₂ at the B3LYP level. We achieved some important results, which are detailed as follows:

1. Two steps are proposed to describe the mechanism of olefin epoxidation: activation of H₂O₂ (step 1) and oxygen transfer (step 2). We found that aqua‐coordinated POM has two novel pathways for the activation of H₂O₂ to form the TM−OH−(μ‐OOH) intermediate. Step 1 a, the dissociation of water molecule, and then the cleavage of HO−OH bond, or step 1 b, cleavage of the HO−OH bond accompanied with H abstraction and OH addition. Step 1 b is more facile, but it is also the rate‐limiting pathway, and the similar energy barriers of the three catalysts are in good agreement with experimental results.

2. When we compare the three POMs, the Rh‐containing POM becomes more competitive, because of the increased interactions between Rh^III and the peroxo unit.

3. Through the analysis of all active oxygen positions, we found that the O_2M position is more favorable than the O_3M position. O_2M−W is preferred for Rh‐ and Pd‐containing POMs, whereas the O_2M−Zn position is slightly favored for the Pt‐containing POM, suggesting that the future design of catalysts should focus on improving the coordination environment of reactive center.

4. When one and two water solvent molecules are present in the reactive system, the increased hydrogen bonding significantly reduces the energy barriers.

Computational Details

All calculations were carried out by using the Gaussian 09 A.02 software package and NBO program (version 3.1).45, 46 The geometries were fully optimized at the B3LYP level without any symmetry constraints.47, 48, 49 For Rh, Pd, Pt, Zn, and W atoms, the LANL2DZ pseudopotential was added.50 The 6–31 g (d,p) basis set was used for the reaction center, that is, H₂O₂ and C₆H₁₀, as well as C, H, O atoms that directly link with the central metals.51, 52, 53 For the rest of the atoms, the 6–31g basis set was used.51, 52, 53 Normal‐mode frequency analysis was performed at the same level to characterize the nature of each stationary point. The local minima possess all real frequencies. The transition states were identified by the only one imaginary frequency and the normal modes that correspond to the expected reaction path. H₂O solvent was included in all corresponding calculations by using the polarized continuum model (PCM).54 A very tight convergence (10⁻⁸ au) criterion was employed in all calculations. To evaluate the effect of explicit solvent, all reactants were fully optimized; whereas, in the transition states (O_2M−Zn position), the central reaction coordinates including TM−O, O−O, O−C, and O−H from the corresponding transition states were kept frozen.

Supporting information

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supplementary

C. Ci, H. Liu, L. Yan, Z. Su, ChemistryOpen 2016, 5, 470.