A theoretical study on the formation mechanism of peroxovanadate and the origin of its high activity for oxidation of sulfide

Abstract

Polyoxovanadates (POVs) show great promise in catalytic oxidative desulphurisation, an industrially important yet challenging process, in which peroxovanadate generated by the oxidation of POVs is an essential active species. In such reaction, H₂O₂ is generally used as an oxidant because it is environmentally friendly. However, the formation mechanism of peroxovanadate is complicated, and the origin of its high reactive activity in sulfide oxidation remains unclear. Herein, we carefully investigate the generation mechanism of peroxovanadate active species formed from the reaction between POVs and H₂O₂ by density functional theory (DFT) calculations. The results clearly show that POV is oxidized by H₂O₂ to afford the highly reactive oxidant peroxovanadate intermediate via a proton transfer pathway rather than a radical pathway. Interestingly, the oxidation of VOSO₄ by H₂O₂ affording a peroxovanadate intermediate differs from that of POVs, which likely occurs via a HO–OH bond cleavage (radical) pathway. On the other hand, the high activity of peroxovanadates in sulfide oxidation originates from the lower-energy empty orbitals of the oxygen atoms compared to those in POVs. These findings may provide a new perspective for the application of POV-based materials in the catalytic field.

---

Rong-Lin Zhong

Rong-Lin Zhong was born and raised in Chongqing, China. He completed his PhD degree in 2015 with Prof. Zhong-Min Su at Northeast Normal University. Then, he joined in Jilin University as a lecturer and was promoted as an associate professor in 2022. During 2016–2019, he worked as a post-doctor in Kyoto University with the supervision of Prof. Shigeyoshi Sakaki. During 2021–2022, he was named a Hong Kong scholar and joined Prof. Fuk Yee Kwong's group in The Chinese University of Hong Kong. His research interests focus on the development and application of theoretical methods for simulating multi-component transition metal and polyoxometalate catalytic reactions.

---

Introduction

In recent years, acid rain has caused significant environmental damage, primarily due to the release of sulfur oxides (SO_x) from the combustion of sulphur-containing fuels.¹ Hence, it is imperative to reduce the sulfur content of fuels. Oxidative desulphurisation (ODS) has attracted considerable attention as a promising method for converting SO_x into easily removable sulfone and sulfoxide compounds.² It is well known that sulfoxide or sulfone compounds are critical in various fields, including synthesis chemistry, pharmaceuticals, pesticides, and materials science.^3–7 Hydrogen peroxide (H₂O₂), one of the most environmentally friendly oxidants, serves as an ideal reservoir for more reactive oxygen species such as superoxide radicals (˙O₂⁻), hydroperoxide radicals (˙O₂H) and hydroxyl radicals (˙OH).⁸ Consequently, H₂O₂ has recently been widely used in the catalytic oxidative desulfurization of sulphur-containing matrices to produce sulfoxides or sulfones. Over the past decade, various catalytic materials were developed for sulfur-containing substrate oxidation, including transition-metal catalysts (e.g., Mn,⁹ Cu,¹⁰ and Co¹¹), organic molecular catalysts,¹² photocatalysts,¹³ and electrocatalysts.¹⁴ However, various limitations may be associated with these catalysts, such as the use of environmentally unfriendly solvents, high catalyst loadings, long reaction times, reliance on noble metals, and poor catalyst reusability. Therefore, it is urgent to develop an economical, environmentally friendly, sustainable, and stable catalyst for sulfur-containing substrate oxidation.

Polyoxometalates (POMs) are widely recognized as environmentally friendly catalysts.^15–20 POMs are a large class of metal (e.g., Mo^VI, W^VI, V^V, Nb^V, and Ta^V) oxides and represent a tremendous range of crystalline inorganic clusters with unparalleled physical and chemical properties.²¹ It is also noteworthy that the properties of POMs, including solubility, redox ability, acidity, magnetism, and thermal/chemical stability, can be tuned by modifying the constituent atoms, organic bridging components, and structural dimensions (such as POM-based chains, layers, and frameworks).^22–29 Previously, some classic POMs have been extensively investigated as homogeneous catalysts in the selective oxidation of various sulfides.^30–32 On the surface of POMs, numerous oxygen atoms act as proton receptors.^33–36 Therefore, POMs and their derivatives can facilitate proton transfer to promote catalytic activation. Polyoxovanadates (POVs), a significant branch of POMs have variable coordination geometries of V atoms ({VO₄} tetrahedral, {VO₅} tetragonal cone, and {VO₆} octahedral), which contribute to their structural diversity. Furthermore, the versatile oxidation states of vanadium (III, IV, and V) contribute to their unique redox properties, making them promising candidates for catalyzing oxidation reactions of small organic molecules.³⁷ It is worthy of note that POVs have attracted considerable attention due to their efficient catalytic activity and potential applications in biomedicine and nano-magnetism.^38,39

POVs are widely employed as catalysts or catalyst precursors for various oxidative reactions involving H₂O₂, ^tBuOOH (TBHP) or O₂ under mild conditions.⁴⁰ Recently, several experimental studies have reported on the catalytic activity of POVs in oxidative desulphurization reactions.^41,42 Xu's group synthesized four methoxy and organic amine-functionalized windmill-shaped {V₈} clusters for the oxidation of diphenyl thioether, resulting in diphenyl sulfoxide.¹⁵ Niu and coworkers reported a high-nuclearity POV-based aliphatic carboxylate derivative, K₆H[V₁₇^VV₁₂^IV(OH)₄O₆₀(OOC(CH₂)₄COO)₈]·nH₂O,⁴¹ which also demonstrated catalytic activity for the oxidation of diphenyl thioether using TBHP as an oxidant. Recently, our group synthesized a series of POV-based metal–organic polyhedra with highly efficient catalytic activity for the oxidation of sulphur-containing substrates to sulfoxides and over-oxidized sulfones,^43–48 showing good performance under milder reaction conditions. For instance, the reaction conditions of the {V₅O₉Cl}-based metal–organic polyhedral catalytic system are moderate, with methanol (CH₃OH) as the solvent. Most notably, tert-butyl hydroperoxide (TBHP), an environment-friendly oxidant, was used instead of more aggressive chemicals, as shown in Fig. 1(a).⁴⁸ The metal–organic polyhedral [(V₅O₉Cl)₃(TDA)₆]⁶⁻ consists of three {V₅O₉Cl} units connected by a six thiophene dicarboxylic bridge (TDA), where each {V₅O₉Cl} unit contains four V^IV atoms and one V^V atom, five terminal oxygen atoms (O_t) and four bridge oxygen atoms (O_b) (Fig. 1(b)). According to previous investigations,^49–51 in this system, the oxidation of a POV unit by TBHP likely generates a V-peroxo intermediate as active species, by which a sulfur containing substrate is oxidized to sulfoxides or sulfones.

Fig. 1 (a) POV-based catalytic system for thioether oxidation. (b) Structure of {V₅O₉Cl}, with the following color code: V (dark gray), O (red), Cl (green) and C (light gray). (c) Conversion rate for oxidation of phenylmethyl sulfide by different catalysts.

The experimental results provide an environmentally friendly strategy for oxidative desulphurization. However, further theoretical studies are needed to better understand the formation mechanism of POV and the origin of its high activity in sulfide oxidation. Specifically, three questions remain regarding the experimental findings: (i) during the oxidation of {V₅O₉Cl}, both H–OOH and HO–OH bond cleavage of H₂O₂ might occur (Scheme 1). Which is the more favorable pathway for the oxidation process catalyzed by POV clusters? (ii) What role does the solvent (methanol) molecule play in the catalytic reaction? Vanadium oxovanadium sulphate/hydrochloride (VOSO₄·nH₂O/VOCl₂·nH₂O) exhibits some catalytic activity, but its catalytic performance is significantly enhanced when converted to the cluster form, as shown in Fig. 1(c). (iii) What is the key factor highlighting the advantages of the cluster? Therefore, it is crucial to disclose the origin of the high catalytic performance of the POV-based cluster. Answering these questions will provide the fundamental understanding and valuable information for the development of more excellent catalysts.

Scheme 1 Two possible mechanisms for H₂O₂ activation.

In this work, the mechanism of desulfurization oxidation catalyzed by [(V₅O₉Cl)₃(TDA)₆]⁶⁻ and the key factors dominating the catalytic performance of {V₅O₉Cl} cluster were theoretically investigated by using density functional theory (DFT) calculations. The entire catalytic cycle consists of three elementary steps: activation of H₂O₂, generation of peroxovanadate, and oxygen atom transfer (OAT) from peroxovanadate to thioether. In this case, the preliminary two steps occur via the proton transfer pathway, facilitated by a hydrogen bonding network formed between the solvent molecule (methanol), H₂O₂ and POV cluster, which decreases the energy barrier. Differences in reactive behavior of VOSO₄ and POV with H₂O₂ are mainly due to their different coordination environments of vanadium atoms, which lead to different pathways of H₂O₂ activation (via O–O or H–O bond cleavage). Furthermore, the study reveals that the molecular cage effect, formed by the six TDAs, results in a slight decrease in the energy barrier. The results clearly disclose that the high reactivity of peroxovanadate for sulfide oxidation originates from the lower energy of its empty orbitals compared to that in polyoxovanadates. These findings provide new perspectives on the application of POV-based catalytic materials in improving the environment.

Computational details and models

All DFT calculations were performed using the Gaussian 16 program.⁵² The (U)B3LYP functional^53–56 with Grimme's semi-empirical dispersion potential D3(BJ) was used for geometry optimization.⁵⁷ For the basis sets, valence electrons of V were described using the SDD basis sets⁵⁸ with effective core potentials representing their core electrons, while the 6-31G(d) basis set was used for all other atoms, referred to as BS-I. The functional and basis set testing results indicate that the B3LYP-D3(BJ) combined with BS-I efficiently reproduce the crystal structure of the {V₅O₉Cl}-based metal–organic polyhedra, as shown in Table S1 (ESI).†⁵⁹ To confirm whether each optimized stationary point corresponds to an energy minimum or a transition state and to evaluate the zero-point vibrational energy as well as thermal corrections, frequency calculations were conducted using the same level of theory as for the geometry optimizations. Single-point energy calculations were performed with the (U)B3LYP-D3(BJ) functional and a mixed basis set of SDD for V and 6-311++G(d,p) for the other atoms, see Fig. S2 and Table S2 in the ESI† for the quality of this basis set. Solvent effects were considered by the continuum solvation model SMD⁶⁰ developed by Truhlar's group. In this work, discussions were based on the Gibbs free energy. Thermal corrections and entropy contributions to the Gibbs energy were evaluated at 298.15 K and 1 atm, with the translational entropy in solution corrected by the method of Whiteside et al.⁶¹ The 3D diagrams of key intermediates and transition states were generated using CYLview⁶² and VESTA.⁶³ The interaction region indicator (IRI) isosurface map and independent gradient model based on Hirshfeld partition (IGMH) isosurface map were created through combination of the Multiwfn⁶⁴ wave function analysis program and VMD⁶⁵ visualization program to investigate chemical bonds and weak interactions in the system.^66,67

Under experimental conditions, [(V₅O₉Cl)₃(TDA)₆] was employed as the catalyst for the H₂O₂ oxidative desulfurization reaction,⁵⁹ with the three V₅O₉Cl clusters serving as the active units. Therefore, we explore the reaction by a representative cluster model V₅O₉Cl(TDA)₄ to reduce the complexity and reduce the computational cost, as shown in Fig. 1(b). According to the crystal structure of [(V₅O₉Cl)₃(TDA)₆]⁶⁻, it can be inferred that one V^V and four V^IV atoms exist in the model cluster [V₅O₉Cl(TDA)₄]²⁻. Furthermore, as shown in Fig. S4–S7,† the spin and SOMOs indicate that the oxidative state of four peripheral V atoms is V^IV and the center one is in V^V.

Results and discussion

A comparative study was carried out to reveal the unique catalytic properties of [V₅O₉Cl(TDA)₄]²⁻. First, we will calculate the reaction potential energy surface for the direct oxidation of thioethers by H₂O₂ and then compare it with that for the thioether oxidation catalyzed by VOSO₄, aiming to underscore the significance of peroxovanadate species. Next, we will explore the catalytic mechanism of VOSO₄ to elucidate the characteristic differences between its catalytic oxidation and that of POV. Then, we will discuss the active species, reaction mechanism, and rate-determining steps of the [V₅O₉Cl(TDA)₄]²⁻ catalyzed desulfurization oxidation reaction, as a clear understanding of these issues is necessary for comprehending POV catalysts. As mentioned above, the solvent methanol also plays a crucial role in the reaction. We will analyze the solvent effect of methanol by considering the methanol molecule into the reaction model. Lastly, the shape effect of molecular cages will be explored. The ultimate goal of this study is to gain a comprehensive understanding of the reaction mechanism of {V₅O₉Cl}-based metal–organic polyhedra catalyzed oxidation of sulfide and to provide new insights for the application of POV-based materials in the catalytic field.

Reaction mechanism of [V₅O₉Cl(TDA)₄]²⁻-catalyzed desulfurization oxidation

Prior to a theoretical investigation of the catalytic mechanism, we calculated the energy barrier for the direct oxidation of sulfides by H₂O₂. As shown in Fig. S8,† the formation of the S–O bond is facilitated though the transition state TS_ox, where the energy barrier is 29.1 kcal mol⁻¹, to generate a methyl phenyl sulfone. Such a high barrier is not consistent with experimental results that this reaction occurs at room temperature. As shown in Fig. 1(c), VOSO₄ could also catalyze the oxidation of sulfides by H₂O₂ with a relatively lower yield. This catalytic cycle involves three elementary steps: HO–OH bond cleavage of H₂O₂, generation of peroxovanadium sulphate and oxygen atom transfer, as shown in Fig. 2. Among these, the rate-determining step is the generation of peroxovanadium sulphate, with an activation energy (ΔG°^‡) of 23.0 kcal mol⁻¹. Peroxovanadium sulphate is a highly oxidative intermediate, and the energy barrier for sulfides oxidation is only 4.8 kcal mol⁻¹, significantly lower than that (29.1 kcal mol⁻¹) of direct oxidation by H₂O₂. These results clearly show that the peroxovanadium intermediate is the key active species for the oxidation of sulfides, which is crucial for the subsequent investigation into the mechanism of the [V₅O₉Cl(TDA)₄]²⁻-catalyzed desulfurization oxidation reaction. This proposal is consistent with the Raman spectrum of 30% H₂O₂-treated NaVO₃, (Fig. S9†) where the characteristic peak at 879 cm⁻¹ corresponds to O–O stretching vibrations of peroxovanadate species.^49,68

Fig. 2 Gibbs energy profile (in kcal mol⁻¹) of VOSO₄ catalyzed methyl phenyl sulfoxide oxidation. The red dashed line represents the transition state part.

However, VOSO₄ may encounter the challenge of competitive decomposition of H₂O₂, as discussed in detail in Fig. S10.† Computational results suggest that VOSO₄ shows catalytic activity for H₂O₂ decomposition, hence only a little portion of H₂O₂ is involved in the oxidation process of sulfide substrate, resulting in the lower conversion rate of sulfides.

Herein, we aim to briefly discuss the result that the activation energy of H₂O₂ activation by VOSO₄ occurs via HO–OH bond cleavage, with this value being significantly lower than that for H–OOH bond cleavage. In VOSO₄, the V atom has an unsaturated coordination environment and H₂O₂ interacts with VOSO₄ to form intermediate II, exhibiting a considerable binding energy (−7.3 kcal mol⁻¹). The oxidation of V^VI to V^V thus occurs via a concerted transition state TSIA, exothermically (−12.7 kcal mol⁻¹) generating intermediate III with two hydroxyl groups bonding to the V^V atom, a step with an energy barrier of 11.2 kcal mol⁻¹. In addition, the path (TSIB and TSIC) via the H–OOH bond cleavage has higher energy barriers of 26.5 kcal mol⁻¹ and 14.6 kcal mol⁻¹. Thus, VOSO₄ activation of H₂O₂ is via HO–OH bond cleavage. As illustrated in Fig. S11,† the frontier molecular orbital analysis shows that the interaction between the d orbital of V and the σ* orbital of the O–O bond occurs in a way that the p orbital of one OH group forms a bonding interaction with V, while the other one does not. Nevertheless, the inherent instability of the free hydroxyl group and unsaturated coordination of the V atom lead to the formation of intermediate IIIA, which converts to peroxovanadium sulphate though the dissociation of the H₂O molecule.

According to the above theoretical studies on VOSO₄ catalyzed oxidation of sulfides, the formation of peroxovanadium is important, and two plausible reaction mechanisms seem to exist for the oxidation process where POV is converted into the peroxovanadate active species, as shown in Scheme 2. In the first mechanism, H₂O₂ is activated via an O–O bond cleavage, forming a hydroxyl-containing V intermediate (V–OH) and a ˙OH free radical. This intermediate then undergoes hydrogen atom transfer dehydration, converting POV-based V₅ clusters into the peroxovanadate intermediate. In the other mechanism, H₂O₂ is activated via a H–OOH bond cleavage, generating a peroxy V intermediate (V–OOH). The peroxovanadate intermediate is subsequently generated via proton transfer, which leads to the cleavage of the O–O bond and dehydration from the V–OOH intermediate. Next, the sulfide is oxidized by the peroxovanadate intermediate to form sulfoxide, with the oxidation proceeding through an oxygen atom transfer pathway.

Scheme 2 Reaction mechanism of POV-catalyzed methyl phenyl sulfoxide oxidation. The organic ligands of the catalyst are omitted for clarity.

The Gibbs energy profile of H₂O₂ activated though the O–O bond cleavage is shown in Fig. 3. H₂O₂ approaches [V₅O₉Cl(TDA)₄]²⁻ (1) to form the adduct [(V₅O₉Cl)(TDA)₄⋯H₂O₂]^2– (2A) with a considerable binding energy (−4.7 kcal mol⁻¹) due to the formation of two hydrogen bonds (O_b⋯H and O_t1⋯H) with a distance of 1.90 Å. Subsequently, the transformation from 2A to 2B is endothermic (9.4 kcal mol⁻¹) because it involves a deformation of the V₅O₉ cluster to ensure the reaction with H₂O₂. Specifically, the two V–O_b bonds of V₅O₉ elongate from 1.85 Å to 2.02 Å. It is important to note that the energy barriers will be particularly high (49.7 kcal mol⁻¹), if the V₅O₉ cluster does not undergo deformation and is directly activated viaTS1A. Next, H₂O₂ coordinates to the V atom in 2B because the V–O_h1 distance is 4.30 Å and the O_h1–O_h2 distance is 1.46 Å, which are almost the same as those in the equilibrium structure. From 2B, the O–O activation of H₂O₂ proceeds viaTS1B, forming hydroxyl-containing intermediate 3B and a ˙OH free radical. The V–O_h1 decreases to 1.95 Å in TS1B, and O_h1–O_h2 elongates to 1.77 Å in TS1B. The V–O_h1 distance further decreases to 1.75 Å in 3B accompanied by the elongation of the O_h1–O_h2 bond (4.47 Å). The value of ΔG°^‡ in this elementary step is 35.1 kcal mol⁻¹ with a somewhat endothermic reaction energy (ΔG° = 2.2 kcal mol⁻¹). After the dissociation of V–O_b, a similar pathway to the activation of H₂O₂ by VOSO₄ is considered, and the barrier is 44.6 kcal mol⁻¹ (TS1C). This seems difficult to be overcome at room temperature, and thus we omit the discussion on this pathway. 3B undergoes hydrogen atom transfer dehydration, converting POV-based V₅ clusters into peroxovanadate intermediates 4viaTS2B. The O_h2–H₁ distance decreases from 3.60 Å in 3B to 1.14 Å in TS2B, and O_h1–H₁ is elongated from 1.00 Å in 3B to 1.23 Å in TS2B. The ΔG°^‡ is 10.6 kcal mol⁻¹ and the ΔG° value is −4.6 kcal mol⁻¹. The rebound of the two V–O_b easily occurs from 4 to peroxovanadate intermediate 5 with a ΔG° value of −5.4 kcal mol⁻¹. Interestingly, this peroxovanadate intermediate is highly active because the barrier of oxygen atom transfer to thioanisole viaTS3 is only 15.6 kcal mol⁻¹ with a high exothermic reaction energy (ΔG° = −41.6 kcal mol⁻¹). The O_h1–S distance decreases from 3.92 Å in 6 to 2.03 Å in TS3 and that of O_h1–V is elongated from 1.79 Å to 1.91 Å in TS3. The results confirm that the generation of peroxovanadate intermediate 5 is the key for oxidation of thioanisole. However, the rate determining step for generation of a peroxovanadate intermediate is the cleavage of the O–O bond with a relatively high activation energy (35.1 kcal mol⁻¹), which is even higher than that (29.1 kcal mol⁻¹) of the direct activation by hydrogen peroxide. Therefore, this pathway is unlikely to proceed, as the activation energy is too high to be overcome at room temperature (experimental conditions). In the next section, another pathway (H–OOH bond cleavage) for generation of the peroxovanadate intermediate needs to be discussed.

Fig. 3 Gibbs energy profile (in kcal mol⁻¹) of H₂O₂ oxidising methyl phenyl sulfoxide via HO–OH bond cleavage. The bond length is given in Å. The red dashed line represents the bond formation and activation in the transition states.

Proton transfer can occur at both O_t and O_b sites during H–OOH bond cleavage, as shown in Fig. 4. Starting from intermediate 2A, the H–OOH bond cleavage proceeds viaTS1F, and proton transfer from H₂O₂ to the O_b atom forms a peroxy-containing intermediate 3F. The O_b–H₁ distance decreases from 1.90 Å in 2 to 1.26 Å in TS1F, while the distance of O_h1–V is shortened from 3.66 Å in 2 to 2.12 Å in TS1F, and O_h1–H₁ is elongated from 0.98 Å to 1.14 Å in TS1F. The distance of V–O_h1 further decreases to 1.94 Å in 3F accompanied by elongation of the O_h1–H₁ bond (2.20 Å). The activation energy of this elementary step is 26.4 kcal mol⁻¹ with an endothermic reaction energy (ΔG° = 23.8 kcal mol⁻¹). The peroxovanadate intermediate is then generated from 3FviaTS2F. The O_h2–H₁ distance decreases from 1.66 Å in 3F to 1.06 Å in TS2F, and the O_b–H₁ distance is elongated to 1.43 Å in TS2F. The value of ΔG°^‡ is 13.3 kcal mol⁻¹, and the ΔG° value is −22.2 kcal mol⁻¹. TS2F is the transition state with the highest energy, so the ΔG°^‡ to complete the catalytic cycle is the energy difference between TS2F and 2, which is 37.1 kcal mol⁻¹. On the other hand, the pathway of proton transfer from H₂O₂ to O_t atoms is also considered, and the barrier is 41.9 kcal mol⁻¹ (TS2D). It is worth noting that rather large ΔG°^‡ values for both pathways are also inconsistent with the reaction temperature (298.15 K). The results of these studies indicate that it is not sufficient to consider only the interaction between the {V₅O₉Cl} cluster and H₂O₂. However, a comparison of the energy barriers shows that the cleavage of the H–OOH bond is more likely to occur. In this case, the proton is transferred from O_h of H₂O₂ to the O_b site, not to the O_t site. Therefore, proton transfer plays a crucial role, and solvents and counterions (H₂Me₂N⁺) may influence this process, which will be further discussed in the following section.

Fig. 4 The Gibbs activation energy (in kcal mol⁻¹) and associated structures for activating H₂O₂ through H–OOH bond cleavage. The bond distances are shown, and its unit is in Å. The red dashed line represents bond formation and activation in the transition states.

We further considered the potential influence of electrostatic interactions in the reaction process and introduced two counterions (H₂Me₂N⁺) for the simulations. However, as shown in Fig. S12–S14,† for the above reaction pathways, the rate determining step and the important geometric changes of the overall reaction process with the addition of (NH₂Me₂)⁺ are almost the same as those without (NH₂Me₂)⁺, and there is little difference in activation energy. Consequently, it can be concluded that a negligible influence might be induced by the counterions in the reaction.

Effects of solvent

The above results clearly show that proton transfer is a key elementary step for H₂O₂ activation, while direct proton transfer from H₂O₂ to polyvanadium is unlikely due to its high activation energy. Notably, the surface of the polyvanadium metal oxides contains multiple O atoms that might form hydrogen bonds with a protic solvent. Therefore, we investigate the possibility of proton transfer assisted by methanol, as the extensive hydrogen bonding network in methanol may help stabilize the transition state of proton transfer.

The Gibbs activation energy of the catalytic cycle with methanol is shown in Fig. 5. In this case, the methanol molecule primarily influences two stages of proton transfer. Upon the approach of H₂O₂ and CH₃OH to 2-1, the reaction is initiated, resulting in the formation of the adduct 2-2. Intermediate 2-2 is stabilized by –7.0 kcal mol⁻¹ though a triple hydrogen bond: the O_b⋯H₁ bond between 2-1 and H₂O₂ with a distance of 1.90 Å, O_c⋯H₂ bond between CH₃OH and H₂O₂ with a distance of 1.69 Å, and a hydrogen bond between H₂O₂ and H₂Me₂N⁺ with a distance of 1.71 Å. From 2-2, the H–OOH bond cleavage proceeds via2-TS1 to form intermediate 2-3, in which proton transfer from H₂O₂ to the O_b atom is similar to TS1F. It is noteworthy that the distance between O_t2 and hydrogen on the hydroxyl group of methanol (H_c) in 2-TS1 is 1.83 Å, indicating that the methanol molecule forms a hydrogen bonding interaction with the terminal oxygen atom (O_t) on the cluster surface. Simultaneously, the distance between O_c and H₁ in 2-3 is 1.62 Å, and the distance between O_t2 and H_c in 2-3 is 1.79 Å. These distances indicate that the methanol molecule, surface oxygen atoms, and hydrogen of H₂O₂ form a hydrogen bonding network, stabilizing the proton-transfer transition state. The value of ΔG°^‡ in this step is 23.3 kcal mol⁻¹, which is moderately lower than that (26.2 kcal mol⁻¹) in the absence of methanol molecule. From 2-3, the second step of proton transfer occurs via2-TS2, where H₁ is transferred from O_b to O_h1, resulting in peroxovanadate species. The O_b–H₁ distance increases from 1.01 Å in 2-3 to 1.46 Å in 2-TS2, while O_h2–H₁ is shortened from 1.54 Å in 2-3 to 1.04 Å in 2-TS2. Furthermore, in 2-TS2, the O_t2–H_c distance is 1.54 Å, and O_c–H₂ is 1.16 Å, suggesting a complicated hydrogen bonding network within the structure. Interestingly, the activation energy of this rate-determining step is 25.0 kcal mol⁻¹, which is significantly lower than the 35.8 kcal mol⁻¹ observed in the absence of methanol molecule. Such an activation energy is consistent with the experimental results that reaction occurs at room temperature. The results clearly show that methanol molecules play a significant role in facilitating the proton transfer elementary step. With the assistance of methanol, oxygen atoms on the cluster surface and H₂O₂ form a hydrogen bonding network that stabilizes the transition state for proton transfer processes for generating peroxovanadate intermediate. Subsequently, the peroxovanadate intermediate 2-5 is formed, with the dissociation of H₂O. As discussed above, a rapid oxygen atom transfer from the peroxovanadate intermediate 2-5 to thioanisole then occurs, completing the oxidation process. In this process, the charge transfer from the lone pair (p-orbital) of sulfur to the antibonding orbital (σ*) of the O–O bond is crucial for the oxidation of sulfide. The details are presented in Fig. S16 of the ESI.†

Fig. 5 Gibbs energy profile (in kcal mol⁻¹) of H₂O₂ activation via H–OOH bond cleavage (methanol participation). The bond length is given in Å. Long dashed lines represent the transition state, while short dashed lines indicate hydrogen bonding interactions.

Notably, POV also may undergo a competing H₂O₂ decomposition like VOSO₄, as illustrated in Fig. S17.† However, computational results show that the energy barrier (37.5 kcal mol⁻¹) for H₂O₂ decomposition is particularly high to be overcome when POVs act as catalysts. In contrast, VOSO₄ mediates the same process with a substantially lower barrier of 11.4 kcal mol⁻¹. This side reaction pathway exhibits a relatively lower activation energy compared to sulfide substrate oxidation (23.0 kcal mol⁻¹). The marked catalytic activity of VOSO₄ toward H₂O₂ decomposition consequently diverts a significant fraction of H₂O₂ from the intended sulfide oxidation process, thereby depressing the overall yield (Fig. 1(c)). Conversely, POV intermediates show negligible catalytic activity for H₂O₂ decomposition. Therefore, the conversion rate is not always linearly dependent on the barrier of the rate-determining step.

In this work, the transition state of proton transfer is stabilized by the methanol molecule (about 13.4 kcal mol⁻¹). To clearly illustrate the stabilization of the hydrogen-bonding network, the weak interaction analysis of the proton-transfer transition state 2-TS2 was conducted using an independent gradient model based on Hirshfeld partition (IGMH) method, as shown in Fig. 6(b). As shown in Fig. 6(b), there is a large green isosurface between the methyl group of methanol and the oxygen atom on the surface of POV, while a blue isosurface exists between the hydroxyl group of methanol and the hydrogen atom of H₂O₂ as well as the terminal oxygen atom of POV. These results indicate that CH₃OH simultaneously forms hydrogen bonds with POVs (O_cH_c–O_t2, 1.54 Å) and H₂O₂ (O_h1H₁–O_c, 1.16 Å), promoting proton transfer from POVs to H₂O₂. Also, a weak interaction exists between the methyl group of methanol and the surface oxygen atoms of POV. This result is consistent with the IRI isosurface maps shown in Fig. 6(c). Therefore, the transition state of the rate-determining step is stabilized by multiple hydrogen bonds.

Fig. 6 Crucial bond lengths in the geometric structure (a), the isosurface maps of IGMH analysis (b) and IRI isosurface maps (c) of 2-TS2. Binding energy is calculated between the methanol molecule and 1-TS2F in 2-TS2. Green color isosurface represents van der Waals interaction, red color isosurface represents steric hindrance, and blue isosurface represents a relatively strong weak interaction.

This process significantly differs from the activation mechanism of H₂O₂ in the VOSO₄-catalyzed system, which involves the cleavage of the H–O bond and the transfer of protons to the bridging oxygen sites of the clusters. In the case of {V₅O₉Cl}, the vanadium centers are coordinatively saturated. If the reaction is via the radical pathway, one of the V–O bonds needs to be broken to stabilize the radical intermediate. Such a large distortion of {V₅O₉Cl} will result in significant energy increase (24.5 kcal mol⁻¹ as shown in Fig. S18†), and the pathway is energetically unfavorable (the activation energy is 35.1 kcal mol⁻¹). Meanwhile, the surface of {V₅O₉Cl} is rich in oxygen atoms, which can form an extensive hydrogen bonding network with H₂O₂ and solvent methanol molecules, efficiently facilitating the proton transfer and thus driving the protonation pathway. In contrast, the oxygen atoms in VOSO₄ cannot form an extensive hydrogen bonding network to stabilize the transition state of proton transfer. Furthermore, the vanadium center is coordinatively unsaturated in the VOSO₄-catalyzed system. The available coordination site in VOSO₄ can stabilize hydroxyl radicals generated from O–O bond cleavage of H₂O₂, thus facilitating the radical pathway. In this context, {V₅O₉Cl} selectively cleaves H–OOH bonds of H₂O₂ through a proton transfer pathway assisted by multiple hydrogen bonds, while VOSO₄ cleaves the HO–OH bond via a hydroxyl radical pathway caused by its the coordinatively saturated V center.

The shape effect on catalytic activity of molecular cages

The Gibbs energy profile of the catalytic cycle of polyoxovanadate-catalyzed sulfide oxidation reaction is shown in Fig. 7. As discussed above, the catalytic cycle consists of three elementary steps, activation of H–OOH, generation of peroxovanadate, and oxygen atom transfer. The rate-determining step is the generation of peroxovanadium like that in the [V₅O₉Cl(TDA)₄]²⁻-catalyzed sulfide oxidation, with the ΔG°^‡ value of 24.5 kcal mol⁻¹, as shown in Fig. 7, which is slightly lower than that (25.0 kcal mol⁻¹) catalyzed by [V₅O₉Cl(TDA)₄]²⁻. It should be noted that in intermediate 3-2, the stabilization energy of the hydrogen bond network is −5.3 kcal mol⁻¹, which is 1.7 kcal mol⁻¹ less stable than that in intermediate 2-2 (−7.0 kcal mol⁻¹). This difference arises from the fact that the {V₅O₉Cl} unit in the polyoxovanadate structure is more restricted by the ligand compared to the [V₅O₉Cl(TDA)₄]²⁻ model, preventing it from relaxing into a similar stable structure. However, such a relatively unstable hydrogen-bonded precursor may result in a lower activation energy barrier for H₂O₂. This is because the activation energy is defined as the difference between the highest-energy transition state in the catalytic cycle and the precursor. In the case of 3-TS2, the restriction of the polyoxovanadate shape only makes the hydrogen bonding interaction 1.2 kcal mol⁻¹ less stable. Therefore, the influence of the polyoxovanadate shape on catalytic activity can be attributed to the formation of a relatively less stable hydrogen-bonded precursor, which facilitates a lower activation energy barrier for the reaction.

Fig. 7 Gibbs energy profile (in kcal mol⁻¹) of (NH₂Me₂)₆[(V₅O₉Cl)₃(TDA)₆] catalyzed methyl phenyl sulfoxide oxidation. The red dashed line represents the transition state part.

Origin of the high activity of peroxovanadate

The discussions above show that peroxovanadate (2-5) plays a key role in the catalytic oxidation of sulfide. To better understand the origin of its high catalytic activity, its electronic properties are analyzed. Specifically, peroxovanadate exhibits a high degree of similarity to the spin distribution and SOMO of intermediate 2-1, where the unpaired electrons are mainly localized on the surrounding V^IV atoms, as shown in Fig. S20.† In order to cleave the peroxy bond via oxygen atom transfer, it is necessary for electrons to transfer from the sulfur atom to the empty orbital of the oxygen atom. For cluster 2-1, the lowest-energy empty orbital of the oxygen is the LUMO+2 orbital, with an energy level of −1.52 eV. In contrast, in peroxovanadate, the empty orbitals of the oxygen atoms correspond to its LUMO. It is worthy of note that the energy level of this empty orbital (−2.93 eV) is significantly lower than that of the corresponding orbital in 2-1, as shown in Fig. 8. In order to evaluate the stabilization in the transition state by such a CT interaction, we employed linear combination of fragment MO analysis on a model (TS-mo), as shown in Fig. S16(b).† The results show the amount of charge transfer from the lone pair (p) orbital of sulfur to the antibonding orbital (σ*) of the O–O bond is 0.291 e. Such a CT interaction stabilize the model (TS-mo) by −11.4 kcal mol⁻¹, which is only slightly more negative than that in 2-TS3. These results clearly show that the charge transfer from the lone pair (p-orbital) of sulfur to the antibonding orbital (σ*) of the O–O bond is crucial for the oxidation of sulfide. Therefore, from the perspective of electron transfer, a lower energy of the unoccupied molecular orbitals of the oxygen atom enhances the transfer of electrons from the sulfide to the catalyst. In other words, the superior catalytic activity of peroxovanadate over polyoxovanadate likely arises from the significantly lower energy levels of the unoccupied molecular orbitals of the oxygen atoms in peroxovanadate, facilitating more efficient electron transfer and bond cleavage.

Fig. 8 The empty orbitals of the oxygen atoms in 2-5 and 2-1.

Conclusions

Taking [(V₅O₉Cl)₃(TDA)₆]⁶⁻ as a model reaction, we theoretically investigated the catalytic mechanism of {V₅O₉Cl}-based metal–organic polyhedra in the oxidation of sulfides. Our findings reveal that the oxidation of polyoxovanadate by H₂O₂ to form peroxovanadate is a critical step in its catalytic oxidation of sulfides. The formation of peroxovanadate is facilitated by solvent-assisted proton transfer, which involves the cleavage of the H–OOH bond in H₂O₂. Both the geometric structure and IRI isosurface analyses indicate the importance of the hydrogen-bonding network in stabilizing the key proton transfer transition state. Furthermore, we compared the H₂O₂ activation mechanisms in two different systems. In the VOSO₄ system, the activation energy of H₂O₂ is lower via the cleavage of the HO–OH bond. In contrast, in the {V₅O₉Cl} system, the vanadium atoms have a stable five-coordinated lattice structure. {V₅O₉Cl} selectively cleaves the H–OOH bonds in H₂O₂via a proton transfer pathway, rather than breaking the HO–OH bond via a hydroxyl radical pathway. This is due to the lack of vacant coordination sites on V atoms in {V₅O₉Cl} for the latter mechanism, while oxygen atoms in the cluster can act as proton acceptors. In addition, the shape effect on the catalytic activity of molecular cages originates from the fact that polyoxovanadate forms a relatively less stable hydrogen-bonded precursor. Moreover, compared to polyoxovanadate, the empty orbitals of oxygen atoms in peroxovanadate are at a lower energy level, resulting in its higher reactivity. The present work is expected to provide effective fundamental insights that will guide both experimental and computational approaches in the development of highly efficient and environmentally benign POV-based catalysts.

Data availability

The authors confirm that the data supporting the findings of this study are available within the article and its ESI.†

Author contributions

Wei-Xuan Shu: calculation, data analysis, writing – original draft. Lin-Yan Bao: calculation, data analysis. Li-Li Wang: data analysis and validation. Xiao-Xia You: data analysis. Xin Ma: data analysis. Ya Wang: experiment. Zhong-Min Su: conceptualization, supervision. Rong-Lin Zhong: methodology, conceptualization, writing – review & editing.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was financially supported by the National Key R&D Program of China (2021YFA1501600), the National Natural Science Foundation of China (22473047 & 92461310), and the Fundamental Research Funds for the Central Universities; the Super-computing facilities were provided by the Hefei Advanced Computing Center.

Notes and references

1. X. Feng, Q. Liu, S. Wang, L. Cen and H. Li, J. Hazard. Mater., 2021, 420, 126612.
2. A. Rajendran, T.-Y. Cui, H.-X. Fan, Z.-F. Yang, J. Feng and W.-Y. Li, J. Mater. Chem. A, 2020, 8, 2246–2285.
3. J. S. Scott, A. M. Birch, K. J. Brocklehurst, A. Broo, H. S. Brown, R. J. Butlin, D. S. Clarke, O. Davidsson, A. Ertan, K. Goldberg, S. D. Groombridge, J. A. Hudson, D. Laber, A. G. Leach, P. A. Macfaul, D. McKerrecher, A. Pickup, P. Schofield, P. H. Svensson, P. Sorme and J. Teague, J. Med. Chem., 2012, 55, 5361–5379.
4. W. M. Xu, F. F. Han, M. He, D. Y. Hu, J. He, S. Yang and B. A. Song, J. Agric. Food Chem., 2012, 60, 1036–1041.
5. B. Zhang, A. M. Wassermann, M. Vogt and J. Bajorath, J. Chem. Inf. Model., 2012, 52, 3138–3143.
6. W. Zheng, X. Chen, F. Chen, Z. He and Q. Zeng, Chem. Rec., 2021, 21, 396–416.
7. P. G. McCaw, N. M. Buckley, S. G. Collins and A. R. Maguire, Eur. J. Org Chem., 2016, 2016, 1630–1650.
8. C. Ling, X. Liu, H. Li, X. Wang, H. Gu, K. Wei, M. Li, Y. Shi, H. Ben, G. Zhan, C. Liang, W. Shen, Y. Li, J. Zhao and L. Zhang, Angew. Chem., Int. Ed., 2022, 61, e202200670.
9. L. Wang, J. Wang, P. Xu, Q. Li, W. Zhang and S. Cui, Appl. Catal., A, 2015, 508, 52–60.
10. Y. Wang, J. Deng, J. Chen, F. Cao, Y. Hou, Y. Yang, X. Deng, J. Yang, L. Wu, X. Shao, T. Shi and Z. Wang, ACS Catal., 2020, 10, 2707–2712.
11. J. Huo, L. Jin, C. Chen, D. Chen, Z. Xu, C. D. Wilfred, Q. Xu and J. Lu, ACS Appl. Mater. Interfaces, 2023, 15, 43976–43984.
12. F. Wang, W. Xiang, Y. Xie, L. Huai, L. Zhang and X. Zhang, Sci. Adv., 2024, 10, eadq2768.
13. C. Ye, Y. Zhang, A. Ding, Y. Hu and H. Guo, Sci. Rep., 2018, 8, 2205.
14. J. Wu, T. Ye, Y. Wang, P. Yang, Q. Wang, W. Kuang, X. Chen, G. Duan, L. Yu, Z. Jin, J. Qin and Y. Lei, ACS Nano, 2022, 16, 15734–15759.
15. J. P. Cao, Y. S. Xue, N. F. Li, J. J. Gong, R. K. Kang and Y. Xu, J. Am. Chem. Soc., 2019, 141, 19487–19497.
16. K. Y. Monakhov, W. Bensch and P. Kogerler, Chem. Soc. Rev., 2015, 44, 8443–8483.
17. M. P. Santoni, G. L. Ganga, V. M. Nardo, M. Natali, F. Puntoriero, F. Scandola and S. Campagna, J. Am. Chem. Soc., 2014, 136, 8189–8192.
18. X. Zhao, Y. Duan, F. Yang, W. Wei, Y. Xu and C. Hu, Inorg. Chem., 2017, 56, 14506–14512.
19. Z. Li, X. Yi, Q. Wang, Y. Li, D. Li, R. Palkovits, A. K. Beine, C. Liu and X. Wang, ACS Catal., 2023, 13, 4575–4586.
20. Y.-J. Chu, X.-M. Chen and C.-G. Liu, Inorg. Chem. Front., 2019, 6, 3482–3492.
21. L. Vila-Nadal, K. Peuntinger, C. Busche, J. Yan, D. Luders, D. L. Long, J. M. Poblet, D. M. Guldi and L. Cronin, Angew. Chem., Int. Ed., 2013, 52, 9695–9699.
22. N. Mizuno and K. Kamata, Coord. Chem. Rev., 2011, 255, 2358–2370.
23. A. Proust, B. Matt, R. Villanneau, G. Guillemot, P. Gouzerh and G. Izzet, Chem. Soc. Rev., 2012, 41, 7605–7622.
24. J. M. Clemente-Juan, E. Coronado and A. Gaita-Arino, Chem. Soc. Rev., 2012, 41, 7464–7478.
25. D. L. Long, E. Burkholder and L. Cronin, Chem. Soc. Rev., 2007, 36, 105–121.
26. J. Cabezas-Giménez, V. Lillo, J. L. Núñez-Rico, M. N. Corella-Ochoa, J. Jover, J. R. Galán-Mascarós and A. Vidal-Ferran, Chem.–Eur. J., 2021, 27, 16956–16965.
27. Y. F. Song and R. Tsunashima, Chem. Soc. Rev., 2012, 41, 7384–7402.
28. J. Jover, C. K. Brozek, M. Dincă and N. López, Chem. Mater., 2019, 31, 8875–8885.
29. Q. Zhao, Y. Zeng, Z. Jiang, Z. Huang, D.-L. Long, L. Cronin and W. Xuan, Angew. Chem., Int. Ed., 2025, 64, e202421132.
30. T. A. G. Duarte, S. M. G. Pires, I. C. M. S. Santos, M. M. Q. Simões, M. G. P. M. Neves, A. M. V. Cavaleiro and J. A. S. Cavaleiro, Catal. Sci. Technol., 2016, 6, 3271–3278.
31. C. Jahier, S. S. Mal, R. Al-Oweini, U. Kortz and S. Nlate, Polyhedron, 2013, 57, 57–63.
32. M. Muñoz, M. A. Gallo, A. Gutiérrez-Alejandre, D. Gazzoli and C. I. Cabello, Appl. Catal. B Environ., 2017, 219, 683–692.
33. Z. L. Lang, G. C. Yang, N. N. Ma, S. Z. Wen, L. K. Yan, W. Guan and Z. M. Su, Dalton Trans., 2013, 42, 10617–10625.
34. X. F. Su, L. K. Yan and Z. M. Su, Inorg. Chem., 2019, 58, 15751–15757.
35. X. F. Su, W. Guan, L. K. Yan and Z. M. Su, J. Catal., 2020, 381, 402–407.
36. S. Piccinin, A. Sartorel, G. Aquilanti, A. Goldoni, M. Bonchio and S. Fabris, Proc. Natl. Acad. Sci. U.S.A., 2013, 110, 4917–4922.
37. J. Li, D. Zhang, Y. Chi and C. Hu, Polyoxometalates, 2022, 1, 9140012.
38. M. Aureliano, N. I. Gumerova, G. Sciortino, E. Garribba, A. Rompel and D. C. Crans, Coord. Chem. Rev., 2021, 447, 214143.
39. J. Wang, X. Liu, Z. Du and Y. Xu, Dalton Trans., 2021, 50, 7871–7886.
40. M. Sutradhar, L. M. D. R. Martins, M. F. C. Guedes da Silva and A. J. L. Pombeiro, Coord. Chem. Rev., 2015, 301–302, 200–239.
41. K. Wang, Y. Niu, D. Zhao, Y. Zhao, P. Ma, D. Zhang, J. Wang and J. Niu, Inorg. Chem., 2017, 56, 14053–14059.
42. T.-Y. Dang, R.-H. Li, H.-R. Tian, Q. Wang, Y. Lu and S.-X. Liu, Inorg. Chem. Front., 2021, 8, 4367–4375.
43. Y. Gong, Y. Zhang, C. Qin, C. Sun, X. Wang and Z. Su, Angew. Chem., Int. Ed., 2018, 58, 780–784.
44. N. Xu, H. Gan, C. Qin, X. Wang and Z. Su, Angew. Chem., Int. Ed., 2019, 58, 4649–4653.
45. Y. Zhang, H. Gan, C. Qin, X. Wang, Z. Su and M. J. Zaworotko, J. Am. Chem. Soc., 2018, 140, 17365–17368.
46. Y.-T. Zhang, S.-B. Li, X.-L. Wang, Y.-R. Gong, K.-Z. Shao and Z.-M. Su, Dalton Trans., 2016, 45, 14898–14901.
47. Y.-T. Zhang, X.-L. Wang, S.-B. Li, Y.-R. Gong, B.-Q. Song, K.-Z. Shao and Z.-M. Su, Chem. Commun., 2016, 52, 9632–9635.
48. H.-M. Gan, C. Qin, L. Zhao, C. Sun, X.-L. Wang and Z.-M. Su, Cryst. Growth Des., 2021, 21, 1028–1034.
49. H.-R. Tian, Z. Zhang, T.-Y. Dang, S.-M. Liu, Y. Lu and S.-X. Liu, Inorg. Chem., 2021, 60, 840–845.
50. C. Valeria, C. Alessia, F. Barbara, L. Giulia and Z. Cristiano, Coord. Chem. Rev., 2011, 255, 2165–2177.
51. B. Olga and C. Valeria, J. Inorg. Biochem., 2005, 99, 1549–1557.
52. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, G. A. Petersson, H. Nakatsuji, X. Li, M. Caricato, A. V. Marenich, J. Bloino, B. G. Janesko, R. Gomperts, B. Mennucci, H. P. Hratchian, J. V. Ortiz, A. F. Izmaylov, J. L. Sonnenberg, Williams, F. Ding, F. Lipparini, F. Egidi, J. Goings, B. Peng, A. Petrone, T. Henderson, D. Ranasinghe, V. G. Zakrzewski, J. Gao, N. Rega, G. Zheng, W. Liang, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, K. Throssell, J. A. Montgomery Jr, J. E. Peralta, F. Ogliaro, M. J. Bearpark, J. J. Heyd, E. N. Brothers, K. N. Kudin, V. N. Staroverov, T. A. Keith, R. Kobayashi, J. Normand, K. Raghavachari, A. P. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, J. M. Millam, M. Klene, C. Adamo, R. Cammi, J. W. Ochterski, R. L. Martin, K. Morokuma, O. Farkas, J. B. Foresman and D. J. Fox, Gaussian 16, Revision C.01, Gaussian, Inc., Wallingford, CT, 2019.
53. C. Lee, W. Yang and R. G. Parr, Phys. Rev. B Condens., 1988, 37, 785–789.
54. A. D. Becke, J. Chem. Phys., 1993, 98, 1372–1377.
55. P. J. Stephens, F. J. Devlin, C. F. Chabalowski and M. J. Frisch, J. Phys. Chem., 1994, 98, 11623–11627.
56. K. S. Thanthiriwatte, E. G. Hohenstein, L. A. Burns and C. D. Sherrill, J. Chem. Theory Comput., 2011, 7, 88–96.
57. S. Grimme, J. Antony, S. Ehrlich and H. Krieg, J. Chem. Phys., 2010, 132, 154104.
58. D. Andrae, U. Haeussermann, M. Dolg, H. Stoll and H. Preuss, Theor. Chim. Acta, 1990, 77, 123–141.
59. Y. Zhang, X. Wang, S. Li, B. Song, K. Shao and Z. Su, Inorg. Chem., 2016, 55, 8770–8775.
60. A. V. Marenich, C. J. Cramer and D. G. Truhlar, J. Phys. Chem. B, 2009, 113, 6378–6396.
61. M. Mammen, E. I. Shakhnovich, J. M. Deutch and G. M. Whitesides, J. Org. Chem., 1998, 63, 3821–3830.
62. C. Y. Legault, CYLview20, Université de Sherbrooke, 2020, http://www,cylview.org, accessed August 2016. 2023.
63. K. Momma and F. Izumi, J. Appl. Crystallogr., 2008, 41, 653–658.
64. T. Lu and F. Chen, J. Comput. Chem., 2012, 33, 580–592.
65. W. Humphrey, A. Dalke and K. Schulten, J. Mol. Graph., 1996, 14, 33–38.
66. T. Lu and Q. Chen, Chem. Methods, 2021, 1, 231–239.
67. T. Lu and Q. Chen, J. Comput. Chem., 2022, 43, 539–555.
68. M. N. Bhattacharjee, M. K. Chaudhuri and N. S. Islam, Inorg. Chem., 1989, 28, 2420–2423.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5ta02513c

---