1 Introduction

Photocatalysis has been widely applied in various areas, such as carbon dioxide reduction [1,2,3], water splitting [4, 5], and pollutant degradation [6,7,8]. In recent years, photocatalysis has become popular in organic synthesis [9, 10], which includes selective partial oxidations, reduction reactions [11], coupling reactions [12, 13], and fuel production [14]. The synthesis of benzaldehyde (BAD) by photocatalytic oxidation of toluene has the advantages of mild reaction conditions, green reaction process, atomic economy and high target product selectivity, which has attracted high attention from researchers [15, 16]. As far as we know, the C–H bond activation is the most critical step in the photocatalytic oxidation reaction of toluene [17, 18]. Unfortunately, it is rather tricky to activate the C(sp3)–H bonds, because of the high bond dissociation energy (85–105 kcal mol−1) and chemical inertness [19]. Therefore, it is imperative to develop high-efficiency heterogeneous photocatalysts to activate the notoriously inert C–H bonds.

In general, the photocatalytic reaction process for traditional semiconductor involves three steps: (1) the semiconductor absorbs photons and is excited to produce photogenerated electron–hole pairs; (2) the separation and transfer of photogenerated electron–hole pairs; (3) the photogenerated electrons and holes participate in surface reduction and oxidation reactions, respectively [20,21,22]. In recent years, a variety of photocatalysts have been reported successively, which include TiO2-based photocatalysts [18, 23], metal oxide photocatalysts [24, 25], bismuth-series semiconductor photocatalysts [26], metal sulfide [27] and non-metallic polymers [28]. Although multifarious photocatalysts have been studied, the catalytic performance of them is not ideal due to the low light absorption and the severe carrier recombination. However, Zavahir et al. [29] efficiently realized the photocatalytic selective oxidation of aliphatic alcohols to aldehydes or ketones by V6O13 catalyst. The V6O13 catalyst and aliphatic alcohols would form V6O13–alkoxide, which could be excited by visible light to activate the C–H bond of α–C [29]. This reaction mechanism is different from conventional semiconductor photocatalytic mechanism and common for homogeneous systems containing metal complexes [30]; also, it can be achieved in heterogeneous systems by anchoring these active metal sites on the supports, and the resulting catalysts can be termed single-atom site catalysts [31,32,33].

We are inspired by the work of Zavahir's group [29] to speculate that the V6O13 catalyst also have the potential of toluene C(sp3)–H activation. However, it is unknown whether the catalyst V6O13 could efficiently achieve photocatalytic oxidation of toluene and there is a similar photocatalytic mechanism for toluene by catalyst V6O13. Therefore, we present a detailed theoretical study of the photocatalytic activity and mechanism of V6O13 clusters for toluene, using density functional theory (DFT) calculations to further expand the application of the catalyst V6O13 in various fields. The results show that V6O13 catalyst can effectively activate toluene C(sp3)–H bond into benzyl. Moreover, The V6O13–toluene complex has excellent light absorption in the range from 200 to 800 nm, which is critical for the activation of toluene C–H bond.

2 Calculation Methods

All the calculations were performed by the density functional theory (DFT) with Gaussian 09 package [34]. The V6O13 neutral clusters model is simulated [29] and the optimized structure is shown in Fig. 1. The atomic coordinates for the calculated structure of V6O13 clusters is shown in Table S1. The Gibbs free energy (E) of reactants, intermediates, transition states (TS) and products in the acetonitrile solvent was calculated [35, 36]. The complete geometry optimization and vibration analysis were carried out at the B3LYP-D3(BJ)/6-31G (d) level of theory. All geometries of reactants, products, intermediates and transition states (TS) were optimized without any symmetry restriction. To improve the accuracy of the results, Grimme’s DFT-D3 correction as a dispersion corrected method was taken into account. The stationary point (no imaginary frequency) and transition states (only single imaginary frequency) were identified by vibrational analyses [35, 36]. Moreover, the right transition states are confirmed by the intrinsic reaction coordinate (IRC) [50], and the SMD solvation model in acetonitrile was used to consider the influence of the solvation effect for the reaction [37,38,39,40]. The high precision single-point energy were calculated using B3LYP-D3(BJ)/def2-TZVP calculation level on the above optimized structures.

Fig. 1
figure 1

The optimized structures of V6O13 cluster and toluene molecular. a The face view of V6O13 cluster; b the side view of V6O13 cluster; c the top view of V6O13 cluster; and d the optimized diagram of toluene molecular structure. In the V6O13 cluster structure, red balls represent oxygen atoms, gray balls represent vanadium atoms and the number represent atomic sequence; in toluene structure, gray balls represent carbon atoms and white balls represent hydrogen atoms. Structure from Zavahir [29]

Full size image

The optical absorption properties of V6O13 cluster and V6O13–toluene complex were calculated using B3LYP-D3(BJ)/6-31G(d) calculation level in the time-dependent DFT (TDDFT) framework [29, 41, 42]. The light absorption spectra were plotted by the Multiwfn 3.8 software [43]. The adsorption energy Eadp of toluene, benzyl alcohol (BA) and benzaldehyde (BAD) was calculated as follows:

$$E_{adp}=E_{tot}-(E_{adb}+E_{pure})$$
(1)

where Eadp is the adsorption energy. For Toluene (or BA, BAD) adsorbed on the V6O13 cluster, Etot is the total system energy of Toluene (or BA, BAD) adsorbed on V6O13 clusters, Eadb is the energy of Toluene (or BA, BAD), and Epure is the energy of V6O13 clusters.

3 Results and Discussion

3.1 Activation of Toluene C(sp3)–H Bond

As we all know, the cleavage of C(sp3)–H bond is the first and decisive step of toluene activation. However, the dissociation energy of C(sp3)–H bond is as high as 85–105 kcal mol−1 [17,18,19]. Hence, reducing the activation energy of C(sp3)–H bond is the key for the reaction. Figure 2 shows the Gibbs free energy profile of toluene C(sp3)–H bond activation by the V6O13 clusters (left) and the hydroxyl radical ⋅OH (right). The two activation processes involve transition states TS0 (left), TS1 (right) and the transition states TS0, TS1 connect the corresponding initial and final state structures, respectively. The reaction energy barrier for TS0 (V6O13 clusters activating toluene) is 14.2 kcal mol−1, and the energy of the system is reduced by 2.2 kcal mol−1. The activation reaction energy barrier for TS1 (⋅OH) is only 0.6 kcal mol−1, and the system energy decreases by 30.7 kcal mol−1. The results show that the sp3 C–H bond of toluene is more likely to be dissociated by ⋅OH radical than V6O13 catalyst, because the activation process for ⋅OH has a lower energy barrier and more energy is released from the system. As shown in Fig. 2, the catalysis mechanism of V6O13 photocatalyst (TS0) is different from the widely studied V2O5 photocatalyst [29]. The V2O5 photocatalyst takes V5+ as the active center and reacts by semiconductor mechanism [48]. However, the V6O13 photocatalyst adsorbs toluene molecule to form the V⋯C bond, Which could be excited by irritation. And the V⋯C bond length decreases to 2.060 Å from 2.381 Å. Then the O atom of V6O13 catalyst would capture the H atom of toluene. Finally the resultant benzyl would bind with active species (O2, ⋅O2−, ⋅OOH and ⋅OH) and the V6O13 catalyst enters the next cycle. By the above analysis, the activation mechanism of the C–H bond of V6O13–toluene complex is similar to that of the C–H bond activation of V6O13–alkoxide α-C proposed by Zavahir et al. [29]. From the energy point of view, the formation potential of V6O13–alkoxide (9.7 kcal mol−1) is similar to that of V6O13–toluene complex (14.2 kcal mol−1), which means that the formation of both V6O13–alkoxide and V6O13–toluene complex is easy under mild conditions. We suspect that the activation mode of C(sp3)–H bond by V6O13 catalysts is related to the property for V6O13 material undergoing the insulator-to-metal transition at the temperature of – 123 °C [49]. More importantly, in the experimental part they achieved 100% alcohol conversion and excellent 3-hexanone selectivity (96%) by photocatalytic oxidation of 3-hexanol with V6O13 clusters as an example [29]. Therefore, in our work, we prefer to introduce this similar mechanism into the activation of toluene C–H bond to achieve efficient toluene conversion and BAD selectivity. For the TS1 (⋅OH), the adsorbed ⋅OH in the vicinity of the V6O13 photocatalyst would capture the H atom of toluene to generate benzyl and H2O molecule. Moreover, the superoxide radical anion ⋅O2−, the hydroperoxyl radicals ⋅OOH and molecular oxygen O2 also were compared. It should be pointed out that, for ⋅O2−, ⋅OOH and O2, the reaction energy are shown to be endothermic (ΔG > 0), which means it is unlikely to take place under mild reaction condition.

Fig. 2
figure 2

Gibbs free energy profile of C–H bond activation process from toluene to benzyl radical by V6O13 cluster catalyst (left) and ⋅OH (right) radical. Bond distances are in Å

Full size image

3.2 UV–Vis Spectrum

In order to further explore the photocatalysis mechanism of V6O13 catalyst, we also calculated the optical absorption properties of V6O13 cluster and V6O13–toluene. As shown in Fig. 3, we plotted the optical absorption spectrum in the range from UV to visible light by calculating the excited states of V6O13 cluster and V6O13–toluene. First, we import the calculated jump energies and oscillator intensities of each electronic excited state into Multiwfn 3.8 software [43]. Second, we enter the 11/3/2 function in Multiwfn 3.8 software in turn to output the spectral curve, and finally import the output file into Origin software for plotting. The results show that the V6O13–toluene complex has excellent light absorption both in the UV and visible region. Moreover, it is clear that the V6O13–toluene has more excited states and stronger light absorption compared to the individual V6O13 cluster. It is therefore likely that visible light excites the V6O13–toluene and cleavage of the C−H bond of toluene in the light-excited state of V6O13–toluene would be much easier than that in the unexcited state. Based on the above discussions, the irradiation is an important reason for the activation of toluene, and the excellent light absorption of V6O13–toluene complex in visible region from 400 to 800 nm increases the possibility of V6O13 photocatalyst application in industry.

Fig. 3
figure 3

The oscillator strength for the allowed excited states of individual V6O13 cluster (black square) and V6O13–toluene complex (red pentagram star) in the UV–Vis light wavelength range was simulated (left axle), and the light absorption spectrum of the two structures are plotted (right axle)

Full size image

3.3 The Possible Reaction Path of Benzyl Oxidation to BA and BAD

After conversion of toluene to benzyl, the benzyl quickly binds with active species and continues subsequent reactions, the specific oxidation path is shown in Scheme 1a. The solid line arrows in the profile represent thermodynamically favorable paths, and the red and blue numbers represent the energy barrier and the reaction Gibbs free energy (kcal mol−1), respectively. We considered four active species, such as O2, ⋅O2−, ⋅OOH and ⋅OH. The reaction of benzyl with the O2, ⋅O2−, ⋅OOH and ⋅OH active species is exothermic by 17.8, 21.0, 30.5, and 32.7 kcal mol−1, respectively, where the reaction is a barrier-free reaction.

Scheme 1
scheme 1

a Reaction mechanism for toluene oxidation to BA and BAD. b Reaction mechanism for BA oxidation to BAD and BAD oxidation to benzoic acid. Solid lines arrows represent the thermodynamically possible paths. The red and blue numbers represent the barrier and the reaction Gibbs free energy (kcal mol−1), respectively

Full size image

From the thermodynamically favorable path, the benzyl would react with O2 and ⋅O2− to generate C6H5CH2OO⋅ and C6H5CH2OO− intermediates [44], and then bind with benzyl and proton [46, 47] to generate the TS2 (C6H5CH2OO–C6H5CH2) transition state and TS3 (C6H5CH2OOH) transition state [44], respectively. The transition state TS2 generates BA and BAD through an intramolecular dissociation reaction, whereas the transition state TS3 generates BAD and H2O molecule by a dehydration reaction. For the ⋅OOH, the reaction rule is similar to the ⋅O2−, which could generates BAD and H2O molecule by the transition state TS3. For the ⋅OH, it would react with benzyl to generate BA and the specific oxidation path of BA would be detailedly discussed in Sect. 3.4.

The Gibbs free energy profile of the dissociation process of TS2 (C6H5CH2OO–CH2C6H5), as shown in Fig. 4. For the transition state TS2, the CHOO section of TS2 could form a four-element ring configuration, in which the O⋯O bond and the C⋯H bond break, and then the O atom at the far end snatches the H atom to generate BA. Subsequently, the C⋯O single bond becomes the C=O double bond to generate BAD, and the whole process needs to overcome the energy barrier of 34.8 kcal mol−1 and releases the energy of 62.4 kcal mol−1. As shown in Fig. 5, the transition state TS3 (C6H5CH2OOH) involves two possible dehydration processes, the intramolecular dehydration and the intermolecular dehydration with a water molecule participation. For the first configuration (left), the C6H5CH2OOH forms a four-element ring inside the molecule, and then remove a molecule of water to generate BAD. For the second configuration (right), the C6H5CH2OOH forms a six-element ring with a molecule of water, and eventually removes two molecules of water to form BAD. The first dehydration process requires crossing the energy barrier of 49.0 kcal mol−1 and releases the energy of 67.6 kcal mol−1, whereas the second dehydration process only requires overcoming the energy barrier of 35.0 kcal mol−1 and releases more energy (93.7 kcal mol−1). From the perspective of energy, the participation of water molecule promotes the reaction to a large extent. From the perspective of structure, it is likely that formation of the quaternary ring configuration causes the large molecular distortion. Hence, the participation of water molecule decreases the distortion of the molecular configuration to a certain extent, and the formation of six-element ring reduces the intramolecular tension, which is conducive to the reaction. This acceleration effect for promoting the production of BAD has been presented in experimental works. In addition, da Silva et al. [44] also calculated the energy barrier of the dehydration process for C6H5CH2OOH (without water molecule and catalyst participation) is 41.1 kcal mol−1, which is increased by 6.1 kcal mol−1 than our calculation. We attribute the reduction of the barrier to the participation of V6O13 catalyst and water molecule. We admit that this dehydration reaction is not easy to take place at room temperature. However, more importantly, we pay more attention to the water-induced effects in photocatalytic oxidation reactions in kinetics and thermodynamics in our study.

Fig. 4
figure 4

Gibbs free energy profile of the dissociation process of TS2 (C6H5CH2OO–CH2C6H5). Bond distances are in Å

Full size image
Fig. 5
figure 5

Gibbs free energy profile of the C6H5CH2OOH dehydration process. The intramolecular dehydration is shown on the left and the intermolecular dehydration with a H2O molecule on the right. Bond distances are in Å

Full size image

3.4 The Possible Reaction Path of BA Oxidation to BAD

The resultant BA would continue to be oxidized to BAD, and the specific oxidation path is shown in Scheme. 1b. The calculation results show that BA can be activated by ⋅OOH radical (TS4) and ⋅OH radical (TS5) to form C6H5CHOH⋅ intermediate. From BA to C6H5CHOH⋅ by ⋅OOH radical, the process involves a free energy barrier of 19.6 kcal mol−1 and an exothermicity of 31.9 kcal mol−1. From BA to C6H5CHOH⋅ by ⋅OH radical, the free energy barrier is calculated to be 1.9 kcal mol−1 and it is exothermic by 42.9 kcal mol−1. The specific reaction process is shown in Fig. S1. Based on the above discussions, the oxidation of BA by ⋅OOH and ⋅OH to C6H5CHOH⋅ intermediate is favorable in thermodynamics and kinetics. Since the energy barrier of BA oxidation by ⋅OH is lower and the more system energy has been released than ⋅OOH, we believe that it is the main path of BA oxidation to C6H5CHOH⋅ intermediate. The resulting C6H5CHOH⋅ intermediate binds to the ⋅OH radical to generate the transition state TS6 (C6H5CHOHOH), which can remove a water molecule to generate BAD. The transition state TS6 (C6H5CHOHOH) [45] and the transition state TS3 (C6H5CH2OOH) have the same formation mechanism, and both of them involve two possible dehydration processes, one is intramolecular dehydration, and the other is intermolecular dehydration with a water molecule participation. The formation barriers of two configurations for TS6 are 26.3 and 19.5 kcal mol−1 and release the energy of 2.6 and 27.6 kcal mol−1, respectively, the specific dehydration process is shown in Fig. S2. The intramolecular dehydration process is through the formation of a four-element ring remove a molecule of water. The intermolecular dehydration process forms a six-element ring through C6H5CHOHOH and a molecule of water, which finally takes off two molecules of water to generate BAD. The dehydration processes of TS3 and TS6 indicate that the presence of water molecules is a positive effect for the whole reaction. We demonstrate theoretically that water molecules and reactants could form a six-member ring to reduce the formation barrier of transition states and the molecular configuration distortion. Although water has a promoting effect on the reaction, the amount of water added is also worth studying. Some researchers found that when the water is added for more than a certain amount, the conversion rate of reactant is reduced [51,52,53]. We suspect there are two possible reasons. First, in the dehydration reaction, although water can act as a reactant, it is also a product. When excessive water is added, the reaction would proceed in opposite directions. The second reason is related to V6O13 catalyst. Some studies have shown that the water molecule can react with V6O13 catalyst, in which the V–O bond of V6O13 catalyst is broken, the H2O molecule would split into ⋅OH and ⋅H, and ⋅OH would connect with V atom, while ⋅H would connect with O atom [54]. This reaction could deactivate our catalyst, while adding excessive water will aggravate the reaction, which is harmful for the photocatalytic reaction.

3.5 The Process of BAD Oxidation to Benzoic Acid (BAC)

As shown in Scheme 1b, the BAD may be further oxidized. Similar to toluene and BA, BAD is more likely to be oxidized by the ⋅OH radical to produce intermediate C6H5CO⋅. As is illustrated in Fig. S3, the H atom on the aldehyde group (–CHO) of BAD is abstracted through the transition state TS7. For the process of converting BAD to intermediate C6H5CO⋅, the free energy barrier is calculated to be 3.9 kcal mol−1, and the system is exothermic by 31.2 kcal mol−1. The intermediate C6H5CO⋅ then readily reacts with ⋅OH radical to form BAC, which is exothermic by 40.7 kcal mol−1. Base on the above analysis, the ⋅OH radical has a two-sided effect for the whole reaction. On the one hand the ⋅OH radical can activate toluene, BA and BAD, on the other hand, the ⋅OH radical could cause the further oxidation of BAD to form BAC. Therefore, attention should be paid to monitoring the number of ⋅OH radicals in the experiment to improve the conversion rate of toluene and the selectivity of BAD.

3.6 Adsorption Energy of Toluene, BA and BAD on V6O13 Clusters

In addition to the calculations of the reaction path, we also calculated the adsorption energy of toluene, BA and BAD for V6O13 catalyst, respectively, − 26.6 kcal mol−1, + 22.6 kcal mol−1 and + 4.7 kcal mol−1, as shown in Fig. S4. From the perspective of energy, the adsorption energy of BA and BAD on V6O13 catalyst is positive, which indicate that the adsorption processes of BA and BAD on V6O13 catalyst too difficult to occur in the mild condition. That is to say, BA and BAD can not be directly activated by V6O13 catalyst. In contrast, the adsorption energy of toluene on the V6O13 catalyst is negative, indicating that the adsorption process is spontaneous, and toluene can be directly activated by V6O13 catalyst. At the same time, it also shows that resultant BA and BAD can be quickly desorbed from the catalyst to release the active site, which is beneficial for the reaction.

4 Conclusion

In summary, our work demonstrates theoretically through DFT calculations that the catalyst V6O13 can efficiently activate toluene C(sp3)–H bonds with the activation energy is 14.2 kcal mol−1 and the activation mechanism is similar to the C–H bond of aliphatic alcohol by V6O13 clusters, making it a good candidate material for photocatalyst. Toluene and V6O13 catalyst form the V6O13–toluene complex through chemical adsorption, which can be excited by light to effectively activate the toluene C(sp3)–H bond into benzyl. The V6O13–toluene complex has strong light absorption in the range from 200 to 800 nm. Therefore, for the V6O13 catalyst, the excellent absorption is the key of the toluene oxidation. Moreover, we found that water can directly participate in the dehydration process to reduce the barrier of TS3 (C6H5CH2OOH) and TS6 (C6H5CHOHOH) from 49.0 to 35.0 kcal mol−1 and from 26.3 to 19.5 kcal mol−1, respectively. Remarkably, the ⋅OH radical could rapidly oxidize BAD to BAC. Therefore, in order to get the maximum selectivity of BAD, the excess of ⋅OH radical should be avoided. As shown in Scheme 2, there are three possible photocatalytic oxidation paths of toluene into BAD or BAC on V6O13 catalyst. In a word, it is hoped that our study may provide new theoretical insights for the photocatalytic fields and theoretical guidance for photocatalytic selective oxidation of toluene into BAD.

Scheme 2
scheme 2

The photocatalytic oxidation mechanism of toluene into BAD or BAC on V6O13 cluster

Full size image