Sustainable Highly Selective Toluene Oxidation to Benzaldehyde

: Thanks to the well-recognized role of benzaldehyde in industry, nowadays the research of new and sustainable approaches to selectively synthesize such an interesting product is receiving great attention from the chemists’ community. In this paper, a V-based catalytic biphasic system is adopted to perform toluene oxidation to benzaldehyde. Importantly, to pursue sustainability, organic solvents have been avoided, so toluene is used as substrate and co-solvent, together with water. Also, the use of hydrophobic ionic liquids has been explored. To perform oxidation, NH 4 VO 3 catalyst, H 2 O 2 , and a safe and inexpensive co-catalyst are used. Among the tested co-catalysts, KF and O 2 were found to be the best choice, to guarantee good yields, in mild reaction conditions. In fact, with such a sustainable method, up to 30% of benzaldehyde can be obtained at 60 ◦ C and, more interestingly, the oxidative system can be recharged, raising-up the yield. The entire process results highly selective, since no traces of benzyl alcohol or benzoic acid are detected. Hence, it constitutes a very appealing synthetic route, even suitable to be easily scaled-up at an industrial level.


Introduction
Toluene oxidation is currently one of the most important industrial processes: through oxidation, such simple aromatic hydrocarbon can be converted in high added value products, namely benzyl alcohol, benzaldehyde, benzoic acid, and benzoates. The importance of such process resides in toluene oxidation products that are essential molecules, playing a central role in our everyday life. In fact, they are widely used in the manufacture of perfumes, dyes, plasticizers, flame-retardants, preservatives, pesticides, as well as pharmaceuticals [1]. Among the others, benzaldehyde is the most diffused one [2,3], being extensively applied also in food and fragrance industries, as an additive for aroma compositions (due to its bitter almond scent) and as starting material for aliphatic fragrances [4].
Benzaldehyde is a natural product, so it can be easily obtained by extraction or distillation from different botanical sources, like almond, and peach and apricot kernel. However, natural sources are not sufficient to satisfy the market demand, therefore, it is necessary to synthetize benzaldehyde on a large scale, so that its production can exceed 90,000 tons per year globally [5]. Accordingly, in 1887 Étard proposed the first synthetic route to produce benzaldehyde [6]. Thanks to the peculiar high yield and selectivity, such process has been normally used over the years. However, it must be pointed out that Étard reaction occurred in the presence of highly toxic reagents (i.e., CrO 2 Cl 2 ) and solvents (CCl 4 , CS 2 , or CHCl 3 ). Hence, through the years, that process has been entirely replaced. Nowadays the industrial methodology to produce benzaldehyde involves liquid-phase chlorination of toluene, followed by hydrolysis in the presence of a base [7]. Nevertheless, In this work, toluene oxidation is presented, exploiting the same experimental conditions, but avoiding the use of chlorinated solvents, with the aim to raise-up the sustainability of the process. Hence, toluene is used as substrate and co-solvent at the same time. Additionally, KF is adopted as inorganic salt to replace KBr, in order to avoid formation of brominated products (Scheme 1). As a matter of fact, the use of such a green system may constitute an appealing approach to selectively achieve toluene oxidation. Scheme 1. Toluene oxidation in a biphasic system.

Results and Discussion
Toluene oxidation reaction was initially explored by varying NH4VO3 and H2O2 amounts, inorganic salt nature and quantity, and temperature. As well, it was performed either under O2 or N2 atmosphere. Importantly, in such conditions, toluene oxidation occurs at the interphase between the aqueous media and the organic solvent, hence a powerful stirring (≥1000-rpm) in a large flask is required.
Product analysis was carried out after the complete H2O2 consumption (usually 4 or 24 h, as checked with the iodide starch paper test). Interestingly, in all experiments, chemospecific formation of benzaldehyde was detected by GC-MS analysis. In fact, no overoxidation products were found. It is important to underline that, being toluene used in large excess, yields have been calculated with respect to H2O2, i.e., the limiting reagent in the mixture. Results are summarized in Table 1. Toluene oxidation performed with 10% of NH4VO3 with respect to H2O2 selectively led to the formation of benzaldehyde, but yields were not satisfactory (entries 1,2). Raising-up the amount of the vanadium salt to 50% (entries 3,4) increased significantly benzaldehyde yield. Because of the concomitant vanadium catalysed H2O2 decomposition, stepwise addition of the oxidant was performed not achieving better performances. Conversely, working with equimolar amount of NH4VO3 and H2O2 led to a drastic drop of the yield (entry 5). Likely, in these conditions, hydrogen peroxide degradation was faster than toluene oxidation. So reagents concentration in solution was decreased: 15% of benzaldehyde after only 4 h was detected working with 0.05 mmol of both H2O2 and NH4VO3. Furthermore, the reaction with 50% of vanadium with respect to H2O2 showed a decrease of the yield (entry 7). To establish the role of the inorganic salt, a blank experiment in the

Results and Discussion
Toluene oxidation reaction was initially explored by varying NH 4 VO 3 and H 2 O 2 amounts, inorganic salt nature and quantity, and temperature. As well, it was performed either under O 2 or N 2 atmosphere. Importantly, in such conditions, toluene oxidation occurs at the interphase between the aqueous media and the organic solvent, hence a powerful stirring (≥1000-rpm) in a large flask is required.
Product analysis was carried out after the complete H 2 O 2 consumption (usually 4 or 24 h, as checked with the iodide starch paper test). Interestingly, in all experiments, chemospecific formation of benzaldehyde was detected by GC-MS analysis. In fact, no overoxidation products were found. It is important to underline that, being toluene used in large excess, yields have been calculated with respect to H 2 O 2 , i.e., the limiting reagent in the mixture. Results are summarized in Table 1. Toluene oxidation performed with 10% of NH 4 VO 3 with respect to H 2 O 2 selectively led to the formation of benzaldehyde, but yields were not satisfactory (entries 1,2). Raisingup the amount of the vanadium salt to 50% (entries 3,4) increased significantly benzaldehyde yield. Because of the concomitant vanadium catalysed H 2 O 2 decomposition, stepwise addition of the oxidant was performed not achieving better performances. Conversely, working with equimolar amount of NH 4 VO 3 and H 2 O 2 led to a drastic drop of the yield (entry 5). Likely, in these conditions, hydrogen peroxide degradation was faster than toluene oxidation. So reagents concentration in solution was decreased: 15% of benzaldehyde after only 4 h was detected working with 0.05 mmol of both H 2 O 2 and NH 4 VO 3 . Furthermore, the reaction with 50% of vanadium with respect to H 2 O 2 showed a decrease of the yield (entry 7). To establish the role of the inorganic salt, a blank experiment in the absence of KF was performed but neither benzaldehyde nor other products were detected (entry 8).
Despite the promising yield accomplished using toluene as solvent and substrate (15%, Table 1 entry 6), its conversion was still too low to be considered acceptable in a sustainable modern process. Therefore, inspired by previous results with similar systems [37], the reaction was performed adding a stoichiometric amount of the substrate with respect to H 2 O 2 , in water ( Table 2). Yet, performing reactions with a stoichiometric amount of toluene in a large flask, with a condenser on the top, caused substrate loss: at the end of the reaction, only 1% of benzaldehyde was detected in the mixture, while toluene almost completely disappeared (entries 1-3). Hence, reaction was repeated in a closed vessel achieving 12% of benzaldehyde after 48 h (entry 4). Although yield was encouraging, result was not acceptable yet. Therefore, a classical two-phase system was exploited [30]: toluene was dissolved in CH 2 Cl 2 (entry 5) or cyclohexane (entry 6), a more sustainable alternative to the chlorinated solvent. In both cases, no reaction was observed. Inspired by the excellent results achieved in previous studies [32,38,39], a hydrophobic ionic liquid (i.e., [pmim]Tf 2 N 1-propyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide) was adopted as co-solvent. In fact, in the last years ionic liquids became a valid alternative to the conventional molecular solvents, due to their unique properties such as non-volatility, high thermal stability and reusability [40]. 0.5 mmol of toluene were dissolved in 1 mL of [pmim]Tf 2 N; 10 mL of an aqueous solution containing NH 4 VO 3 were then added, with H 2 O 2 and KF. After 24 h, 24% of benzaldehyde was detected in the mixture, together with traces of the benzyl alcohol.
The best compromise, hitherto, in terms of yield and selectivity remains the one achieved using toluene as substrate-solvent, where 15% yield of benzaldehyde was obtained after only 4 h (Table 1, entry 6). Therefore, such conditions have been adopted to investigate the process by changing inorganic salt, atmosphere composition and temperature.
Importantly, the blank experiment without KF ( Table 1, entry 8) showed that such inorganic salt is pivotal to obtain benzaldehyde. To confirm such evidence, other different inorganic salts were screened, with the aim to increase yield and to understand the salt role in the reaction mechanism. However, KF remained the best choice, since no improvement was observed varying the nature of cation or anion (Table 3). Notably, with KBr, a slight increase of benzaldehyde yield was observed, probably because of the presence of Br − ions in solution, which hamper the radical degradation of H 2 O 2 usually occurring in the presence of vanadium (entry 8) [41,42]. Such effect results in an increasing amount of H 2 O 2 available for toluene oxidation. However, with KBr, also bromination products were obtained in good yields [27]. Therefore, KF is the best compromise between yield and selectivity. KF concentration in the reaction mixture was then varied (Table 4). Results showed that an equimolar amount of NH 4 VO 3 and KF was necessary to attractively form benzaldehyde. In particular, working with 0.025 mmol of both (entry 3), 19% yield of benzaldehyde was obtained after 24 h.
In an attempt to further improve yield and to understand reaction mechanism, toluene oxidation was carried out also under O 2 or N 2 atmosphere (Table 5). Interestingly, under O 2 atmosphere, a significant increase of the yield was observed. Indeed, up to 33% of benzaldehyde was obtained (entry 1). To note, working in the absence of KF, but using 0.05 mmol of both NH 4 VO 3 and H 2 O 2 in the presence of dioxygen led to a remarkable yield (40%, entry 2). If 50% of NH 4 VO 3 was used with respect to H 2 O 2 , 30% of benzaldehyde was obtained (entry 3). Conversely, reactions performed with 20% of NH 4 VO 3 with respect to the oxidant (entry 4) or in the absence of H 2 O 2 (entry 5) were not successful. Under N 2 atmosphere, using the standard conditions reported in Table 1, entry 6, a decrease in benzaldehyde conversion was observed, obtaining 12% of benzaldehyde.
Thus likely, in this process, hydrogen peroxide is the actual oxidant, being involved in the formation of the vanadium(V)-peroxido complex [43], while O 2 and KF have a promoting effect in the oxidation of toluene.
Attempting to increase benzaldehyde yield, temperature and reaction time were changed ( Figure 1, Table S1).
Working with an equimolar amount of H 2 O 2 and NH 4 VO 3 , in the presence of KF led to 22% yield after 72 h at 25 • C. Increasing the temperature to 40 • C, produced a 28% yield in 72 h, while at 60 • C the same amount of benzaldehyde was obtained after only 24 h. Consequently, in order to increase the catalyticity, reactions were carried out at 60 • C, decreasing the amount of the V(V)-salt.
To note, an appreciable yield of 28% was obtained using 20% of NH 4 VO 3 under dioxygen at 60 • C ( Table 6, entry 3). The same result was attained in the presence of KF (Table 6, entry 4). To recap, the best performances were obtained with 50% NH 4 VO 3 at 25 • C ( Table 5, entry 3), or 20% NH 4 VO 3 at 60 • C ( Thus likely, in this process, hydrogen peroxide is the actual oxidant, being involved in the formation of the vanadium(V)-peroxido complex [43], while O2 and KF have a promoting effect in the oxidation of toluene.
Attempting to increase benzaldehyde yield, temperature and reaction time were changed ( Figure 1, Table S1). Working with an equimolar amount of H2O2 and NH4VO3, in the presence of KF led to 22% yield after 72 h at 25 °C. Increasing the temperature to 40 °C, produced a 28% yield in 72 h, while at 60 °C the same amount of benzaldehyde was obtained after only 24 h. Consequently, in order to increase the catalyticity, reactions were carried out at 60 °C, decreasing the amount of the V(V)-salt.  To note, an appreciable yield of 28% was obtained using 20% of NH4VO3 under dioxygen at 60 °C (Table 6, entry 3). The same result was attained in the presence of KF (Table 6, entry 4). To recap, the best performances were obtained with 50% NH4VO3 at 25 °C (Table 5, entry 3), or 20% NH4VO3 at 60 °C (Table 6, entry 3) both under O2 atmosphere.
It is important to stress out that, after 24 h, H2O2 and HClO4 total consumption is usually observed. Therefore, the possibility to recharge H2O2 and HClO4, thus restoring the V(V)-monoperoxido complex in solution was explored, taking advantage from the two-phase system (water/toluene). The recyclability assays were carried out using the optimized reaction conditions reported (Table 5, entry 3 and Table 6, entry 3). Four additions have been performed every 24 h; at each round, benzaldehyde quantification has been carried out. Regarding the system at 25 °C ( Figure 2, Table S2) a significant increase in benzaldehyde formation was observed after the second addition, i.e., round 2.  It is important to stress out that, after 24 h, H 2 O 2 and HClO 4 total consumption is usually observed. Therefore, the possibility to recharge H 2 O 2 and HClO 4 , thus restoring the V(V)-monoperoxido complex in solution was explored, taking advantage from the two-phase system (water/toluene). The recyclability assays were carried out using the optimized reaction conditions reported (Table 5, entry 3 and Table 6, entry 3). Four additions have been performed every 24 h; at each round, benzaldehyde quantification has been carried out. Regarding the system at 25 • C ( Figure 2, Table S2) a significant increase in benzaldehyde formation was observed after the second addition, i.e., round 2.
Indeed, up to 25 × 10 −3 mmol of benzaldehyde were obtained, with a round yield of 30% and an overall yield of 25%. The third and fourth additions did not cause an appreciable improvement in benzaldehyde production. In fact, throughout third and fourth round, a plateau was reached, and consequently a decrease of the overall yield was observed. On the other hand, the reaction performed with 20% of NH 4 VO 3 at 60 • C continuously produced benzaldehyde after each oxidant addition. (Figure 3, Table S3)  Indeed, up to 25 × 10 −3 mmol of benzaldehyde were obtained, with a round yield of 30% and an overall yield of 25%. The third and fourth additions did not cause an appreciable improvement in benzaldehyde production. In fact, throughout third and fourth round, a plateau was reached, and consequently a decrease of the overall yield was observed. On the other hand, the reaction performed with 20% of NH4VO3 at 60 °C continuously produced benzaldehyde after each oxidant addition. (Figure 3, Table S3) Remarkably, at the end of the fourth round, 74 × 10 −3 mmol of benzaldehyde were detected in solution, with a constant round yield and an overall yield of 37%. Such an exciting result opens the possibility to continuously add H2O2 to the solution, likely fully converting toluene.
From a mechanistic point of view, in the adopted experimental conditions, the Vcatalysed toluene oxidation likely occurs through a radical pathway. Therefore, a plausible catalytic cycle has been proposed ( Figure 4).  Indeed, up to 25 × 10 −3 mmol of benzaldehyde were obtained, with a round yield of 30% and an overall yield of 25%. The third and fourth additions did not cause an appreciable improvement in benzaldehyde production. In fact, throughout third and fourth round, a plateau was reached, and consequently a decrease of the overall yield was observed. On the other hand, the reaction performed with 20% of NH4VO3 at 60 °C continuously produced benzaldehyde after each oxidant addition. (Figure 3, Table S3) Remarkably, at the end of the fourth round, 74 × 10 −3 mmol of benzaldehyde were detected in solution, with a constant round yield and an overall yield of 37%. Such an exciting result opens the possibility to continuously add H2O2 to the solution, likely fully converting toluene.
From a mechanistic point of view, in the adopted experimental conditions, the Vcatalysed toluene oxidation likely occurs through a radical pathway. Therefore, a plausible catalytic cycle has been proposed ( Figure 4). Remarkably, at the end of the fourth round, 74 × 10 −3 mmol of benzaldehyde were detected in solution, with a constant round yield and an overall yield of 37%. Such an exciting result opens the possibility to continuously add H 2 O 2 to the solution, likely fully converting toluene.
From a mechanistic point of view, in the adopted experimental conditions, the Vcatalysed toluene oxidation likely occurs through a radical pathway. Therefore, a plausible catalytic cycle has been proposed ( Figure 4).  In fact, it is already well established that the dissolution of NH4VO3 in acid water (pH = 1) containing H2O2 leads to the formation of a vanadium (V) peroxido-complex (1), which is a very effective oxidant [34,43,44]. In the first step of the cycle, the homolytic cleavage of the O-O bond possibly occurs, forming the radical V-complex (2), that can abstract a H  from toluene in solution, thus forming vanadium complex (3) and benzyl radical. The latter can react in two different pathways: with complex (3) itself, to form Vcomplex (4) in a non-productive way, or it can react with O2 in solution, to obtain the benzyl peroxy radical. The coupling of two benzyl peroxy radicals forms benzaldehyde, benzyl alcohol and O2 through a well-established mechanism [45]. The formed by-product, namely benzyl alcohol, is then immediately oxidized to benzaldehyde through a similar radical pathway, passing through H-abstraction from the OH group.
As a matter of fact, due to the higher reactivity of benzyl alcohol with respect to toluene, it has never been detected as by-product, since it is readily converted to benzaldehyde. As a verification, benzyl alcohol oxidation in the optimised experimental conditions (H2O2 = 0.05 mmol; NH4VO3 = 0.01 mmol; T = 60 °C; O2) selectively produces benzaldehyde.
Notably, in the presence of KF, fluoride coordination to vanadium likely occurs, leading to V-fluorooxoperoxido complex [46,47]. UV-vis absorption spectrum of V-monoperoxido complex in aqueous solution shows appreciable variations in the presence of KF, possibly supporting fluoride coordination to vanadium ( Figure S1). However, due to the reported instability, such V-species have not been isolated and further characterized, therefore, no experimental verification about their structure in solution have been obtained. For this reason, such complexes have not been included in further mechanistic studies.
To better explicate the mechanism, DFT calculations have been carried out, using Gaussian 16 rev. A.03. Geometry optimization of reagents, intermediates and transition states (TS) has been performed in vacuum using B3LYP functional. 6-31G(d) basis set was used for H, C, and O atoms, while LanL2DZ basis set was adopted for vanadium [48]. The relative energy diagram vs. reaction coordinate is represented in Figure 5. In fact, it is already well established that the dissolution of NH 4 VO 3 in acid water (pH = 1) containing H 2 O 2 leads to the formation of a vanadium (V) peroxido-complex (1), which is a very effective oxidant [34,43,44]. In the first step of the cycle, the homolytic cleavage of the O-O bond possibly occurs, forming the radical V-complex (2), that can abstract a H • from toluene in solution, thus forming vanadium complex (3) and benzyl radical. The latter can react in two different pathways: with complex (3) itself, to form V-complex (4) in a non-productive way, or it can react with O 2 in solution, to obtain the benzyl peroxy radical. The coupling of two benzyl peroxy radicals forms benzaldehyde, benzyl alcohol and O 2 through a well-established mechanism [45]. The formed by-product, namely benzyl alcohol, is then immediately oxidized to benzaldehyde through a similar radical pathway, passing through H-abstraction from the OH group.
As a matter of fact, due to the higher reactivity of benzyl alcohol with respect to toluene, it has never been detected as by-product, since it is readily converted to benzaldehyde. As a verification, benzyl alcohol oxidation in the optimised experimental conditions (H 2 O 2 = 0.05 mmol; NH 4 VO 3 = 0.01 mmol; T = 60 • C; O 2 ) selectively produces benzaldehyde.
Notably, in the presence of KF, fluoride coordination to vanadium likely occurs, leading to V-fluorooxoperoxido complex [46,47]. UV-vis absorption spectrum of V-monoperoxido complex in aqueous solution shows appreciable variations in the presence of KF, possibly supporting fluoride coordination to vanadium ( Figure S1). However, due to the reported instability, such V-species have not been isolated and further characterized, therefore, no experimental verification about their structure in solution have been obtained. For this reason, such complexes have not been included in further mechanistic studies.
To better explicate the mechanism, DFT calculations have been carried out, using Gaussian 16 rev. A.03. Geometry optimization of reagents, intermediates and transition states (TS) has been performed in vacuum using B3LYP functional. 6-31G(d) basis set was used for H, C, and O atoms, while LanL2DZ basis set was adopted for vanadium [48]. The relative energy diagram vs. reaction coordinate is represented in Figure 5. In the adopted experimental conditions, tetrahedral V-monoperoxido complex (1) is the catalytically active species in solution. Importantly, the positively charged complex (1) has been modelled with one water molecule coordinated to the V-center, being this species the most abundant one experimentally observed [43]. V=O, V-O and O-O distances have been calculated and they resulted in line with previously published data (calculated distances = 1.56 Å, 1.77 Å and 1.42 Å, respectively) [49]. Calculations clearly indicate that the first step is the rate determining one, since the activation energy to form complex (2) is 46.4 kcal•mol −1 ( Figure 5). In fact, in the first step, the homolytic cleavage of O-O bond occurs, leading to the vanadium diradical species (2) (calculated O-O distance = 2.89 Å), 14.4 kcal•mol -1 higher in energy than (1). At this stage, H-abstraction from toluene is predictable, to form benzyl radical and tetrahedral V-radical complex (3), thus gaining a significant stabilization of the system (relative energy = −95.5 kcal•mol −1 ). Here, the activation energy barrier resulted about 5.3 kcal•mol −1 .
Afterwards, two different routes have been hypothesized: (i) benzyl radical and Vradical complex (3) coupling, leading to V-complex (4); (ii) reaction between benzyl radical and molecular dioxygen in solution, to form benzyl peroxy radical.
In this context, DFT calculations have been carried out to predict the energetically favoured pathway. However, a direct comparison between the two possible routes is not appropriate, since very different species are involved.
Nevertheless, experimental data underline the key role of O2 for the successful outcome of the reaction: experiments performed under N2 atmosphere led to a remarkable yield drop-off, while under O2 a sensible improvement of benzaldehyde yield was detected.
As a matter of fact, DFT calculations revealed that benzyl radical reaction with dioxygen is energetically favored, being the activation energy barrier of about 0.6 kcal•mol −1 ( Figure 6). Consequently, benzyl peroxy radical formation is highly probable. Moreover, it can be speculated that reactions performed under O2 atmosphere are even more favored because of statistic reasons. In the adopted experimental conditions, tetrahedral V-monoperoxido complex (1) is the catalytically active species in solution. Importantly, the positively charged complex (1) has been modelled with one water molecule coordinated to the V-center, being this species the most abundant one experimentally observed [43]. V=O, V-O and O-O distances have been calculated and they resulted in line with previously published data (calculated distances = 1.56 Å, 1.77 Å and 1.42 Å, respectively) [49]. Calculations clearly indicate that the first step is the rate determining one, since the activation energy to form complex (2) is 46.4 kcal·mol −1 ( Figure 5). In fact, in the first step, the homolytic cleavage of O-O bond occurs, leading to the vanadium diradical species (2) (calculated O-O distance = 2.89 Å), 14.4 kcal·mol −1 higher in energy than (1). At this stage, H-abstraction from toluene is predictable, to form benzyl radical and tetrahedral V-radical complex (3), thus gaining a significant stabilization of the system (relative energy = −95.5 kcal·mol −1 ). Here, the activation energy barrier resulted about 5.3 kcal·mol −1 .
Afterwards, two different routes have been hypothesized: (i) benzyl radical and Vradical complex (3) coupling, leading to V-complex (4); (ii) reaction between benzyl radical and molecular dioxygen in solution, to form benzyl peroxy radical.
In this context, DFT calculations have been carried out to predict the energetically favoured pathway. However, a direct comparison between the two possible routes is not appropriate, since very different species are involved.
Nevertheless, experimental data underline the key role of O 2 for the successful outcome of the reaction: experiments performed under N 2 atmosphere led to a remarkable yield drop-off, while under O 2 a sensible improvement of benzaldehyde yield was detected.
As a matter of fact, DFT calculations revealed that benzyl radical reaction with dioxygen is energetically favored, being the activation energy barrier of about 0.6 kcal·mol −1 ( Figure 6). Consequently, benzyl peroxy radical formation is highly probable. Moreover, it can be speculated that reactions performed under O 2 atmosphere are even more favored because of statistic reasons. Catalysts 2021, 11, x FOR PEER REVIEW 10 of 15 Conversely, coupling between benzyl radical and V-complex (3) appears energetically disfavoured and it would lead to complex (4), which cannot produce benzaldehyde.
To conclude, reaction of benzyl radical with O2 is pivotal in the process and it readily occurs in solution. Indeed, coupling between two benzyl peroxy radicals leads to a very unstable species that immediately disproportionates forming benzaldehyde, O2, and benzyl alcohol [45]. The latter, being more reactive than toluene, is easily converted to the aldehyde, as experimentally demonstrated.

Materials and Methods
All commercial reagents and solvents were purchased from Sigma Aldrich/Merck Life Science (KGaA, Darmstadt, Germany), with the highest degree of purity, and they were used without any further purification. UV-vis absorption spectra were recorded with a UV-Vis 2450 Shimadzu (Kyoto, Japan) spectrophotometer. GC-MS analyses have been performed with a Shimadzu (Kyoto, Japan) GCMS QP2010 Ultra system. Table 1 and Table 4 In a 100 mL Erlenmeyer flask equipped with a condenser, ammonium vanadate, and KF (0.25 or 0.05 mmol) were dissolved in 10 mL of distilled water. 10 mL of toluene (containing 2 × 10 −3 M of decane as the internal standard) were added to the aqueous phase. pH was adjusted to 1 with 90 μL of 10.8 M perchloric acid. Afterwards, H2O2 was added to the mixture. The aqueous and organic solvent were kept under vigorous stirring (1000 rpm) at 25 °C. At selected time, an aliquot of 100 μL of the organic phase was taken, diluted with 100 μL of PhCH3 and analyzed by GC-MS. Table 2 (Entries 1-4) In a 100 mL Erlenmeyer flask equipped with a condenser (or with a cap, when specified), ammonium vanadate and 14.5 mg of KF (0.25 mmol) were dissolved in 10 mL of distilled water. Then, toluene was added to the aqueous solution, together with 90 μL of 10.8 M perchloric acid and H2O2. The mixture was stirred at 1000 rpm, at 25 °C. At selected time, reaction mixture was extracted with two portions of 5 mL of diethyl ether (containing decane 1 × 10 −3 M as the internal standard). Organic phase was dried with Na2SO4 and 1 μL of this solution was analyzed by GC-MS. Table 2 (Entries 5, 6) Conversely, coupling between benzyl radical and V-complex (3) appears energetically disfavoured and it would lead to complex (4), which cannot produce benzaldehyde.

Reactions Reported in
To conclude, reaction of benzyl radical with O 2 is pivotal in the process and it readily occurs in solution. Indeed, coupling between two benzyl peroxy radicals leads to a very unstable species that immediately disproportionates forming benzaldehyde, O 2 , and benzyl alcohol [45]. The latter, being more reactive than toluene, is easily converted to the aldehyde, as experimentally demonstrated.

Materials and Methods
All commercial reagents and solvents were purchased from Sigma Aldrich/Merck Life Science (KGaA, Darmstadt, Germany), with the highest degree of purity, and they were used without any further purification. UV-vis absorption spectra were recorded with a UV-Vis 2450 Shimadzu (Kyoto, Japan) spectrophotometer. GC-MS analyses have been performed with a Shimadzu (Kyoto, Japan) GCMS QP2010 Ultra system.

Reactions in Biphasic System (PhCH 3 :H 2 O 1:1 v/v)
In a 100 mL Erlenmeyer flask equipped with a condenser, ammonium vanadate, and the proper amount of KF were dissolved in 10 mL of distilled water. 10 mL of toluene (containing 2 × 10 −3 M of decane as the internal standard) were added to the aqueous phase. pH was adjusted to 1 with 90 µL of 10.8 M perchloric acid. Afterwards, H 2 O 2 was added to the mixture. The aqueous and organic solvent were kept under vigorous stirring (1000 rpm) at 25 • C. At selected time, an aliquot of 100 µL of the organic phase was taken, diluted with 100 µL of PhCH 3 and analyzed by GC-MS. Results are reported in Table 1 and  Table 4.

Reactions in H 2 O with a Stoichiometric Amount of Toluene
In a 100 mL Erlenmeyer flask equipped with a condenser (or with a cap, when specified), ammonium vanadate and 14.5 mg of KF (0.25 mmol) were dissolved in 10 mL of distilled water. Then, the proper amount of toluene was added to the aqueous solution, together with 90 µL of 10.8 M perchloric acid and H 2 O 2. The mixture was stirred at 1000 rpm, at 25 • C. At selected time, reaction mixture was extracted with two portions of 5 mL of diethyl ether (containing decane 1 × 10 −3 M as the internal standard). Organic phase was dried with Na 2 SO 4 and 1 µL of this solution was analyzed by GC-MS. Results are reported in Table 2 (entries 1-4).

Reactions in Biphasic System (Organic Solvent:H 2 O 1:1 v/v)
In a 100 mL Erlenmeyer flask equipped with a condenser, ammonium vanadate and 14.5 mg of KF (0.25 mmol) were dissolved in 10 mL of distilled water. Then, 10 mL of the organic solvent (dichloromethane or cyclohexane) containing decane as internal standard were added to the mixture. Afterwards, toluene was added, together with 90 µL of 10.8 M perchloric acid and the proper amount of H 2 O 2 . The reaction was kept under stirring at 1000 rpm, at 25 • C. At selected time, 100 µL of the organic phase were diluted with 100 µL of the same solvent and analyzed by GC-MS. Results are reported in Table 2 (entries 5, 6).  Table 2 (entry 7).

Reactions in Biphasic System (PhCH 3 :H 2 O 1:1 v/v): Inorganic Salt Screening
In a 100 mL Erlenmeyer flask equipped with a condenser, ammonium vanadate and 0.25 mmol of selected halogenated inorganic salt were dissolved in 10 mL of distilled water. 10 mL of toluene (containing 2 × 10 −3 M decane as internal standard) were added to the aqueous phase. pH was adjusted to 1 with 90 µL of 10.8 M perchloric acid. Afterwards, H 2 O 2 was added to the mixture. The mixture was kept under vigorous stirring (1000 rpm) at 25 • C. At selected time, an aliquot of 100 µL of the organic phase was taken, diluted with 100 µL of PhCH 3 and analyzed by GC-MS. Results are reported in Table 3.  Table 5.  Table S1.  In a 50 mL two-neck round bottom flask, ammonium vanadate and KF were dissolved in 10 mL of distilled water. 10 mL of toluene (containing 2 × 10 −3 M decane as internal standard) were added to the aqueous phase. pH was adjusted to 1 with 90 µL of 10.8 M perchloric acid. Afterwards, H 2 O 2 (0.05 mmol) was added to the mixture. The mixture was bubbled with O 2 for 30 minutes and kept under vigorous stirring (1000 rpm) at 60 • C under O 2 atmosphere. At selected time, an aliquot of 100 µL of the organic phase was taken, diluted with 100 µL of PhCH 3 and analyzed by GC-MS. Reactions Reported in Table 6 (entries 1 and 3).

DFT Calculations
DFT calculations have been performed using Gaussian 16 rev. A.03 (Gaussian, Inc., Wallingford CT, 2016) [50]. Geometry optimization of reagents, intermediates and transition states (TS) has been performed in the vacuum with B3LYP functional. 6-31G(d) basis set was used for H, C, and O atoms, while LanL2DZ basis set was adopted for vanadium. Transition state optimization was performed with TS and QTS3 option.

Conclusions
In this paper, a sustainable approach for the selective toluene oxidation to benzaldehyde has been presented. The oxidizing system includes a cheap and safe vanadium catalyst dissolved in an acid aqueous solution, and a benign oxidant, such as hydrogen peroxide. In order to develop a sustainable process, organic solvents have been avoided, so toluene is used as substrate and co-solvent at the same time. To note, the addition of KF or O 2 atmosphere is pivotal to guarantee the success of the reaction.
The best results have been obtained performing the reaction at 25 • C or 60 • C, using 50% or 20% of the vanadium catalyst with respect to hydrogen peroxide, under O 2 atmosphere. In such conditions, about 30% of benzaldehyde was formed in 24 h. Despite such a result being in line with previously reported papers on toluene oxidation, here the most relevant feature is that the reaction is very selective, having never detected benzyl alcohol or benzoic acid. Moreover, the catalytic system is composed by safe and non-toxic materials, so hazardous reagents and solvents have been avoided, thus making it highly environmentally friendly.
Importantly, recycle of the reaction mixture was successfully achieved at least for four runs, with preservation of yield and selectivity. Notably, reactions performed at 60 • C showed a continuous production of benzaldehyde, with a constant round yield. At the end of the fourth round, 74 × 10 −3 mmol of benzaldehyde were obtained, with an overall yield of 37%. Such result opens the possibility to explore the application of flow-technique-i.e., continuous addition of reagents-thus allowing full conversion of toluene to benzaldehyde.
DFT calculations allowed to draw the energetic diagram of the reaction pathway. Results highlighted the fundamental role of O 2 , since, if not involved, reactions proceed through a non-productive route, that would not form benzaldehyde.
To conclude, considering the widespread interest in benzaldehyde production, the use of safe reagents and mild conditions, the remarkable selectivity, together with the opportunity to sensibly increase toluene conversion through subsequent batch processes, make the proposed system an appealing alternative to the currently adopted techniques related with the production of benzaldehyde.  Figure S1: UV-vis absorption spectra of V-monoperoxo complex in the presence of KF; Optimized cartesian coordinates (in Angstroms).
Funding: This research was funded by University of Rome Tor Vergata, grant HYPHOTOCAT project, "Beyond borders" and "B-LIGZYMES" project (RISE H2020).

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.