Introduction

Phenol, as one of the most important chemicals, has been widely used in phenol resins, related pharmaceuticals, dyes, antioxidants, and bisphenols [1,2,3,4]. Traditional industrial production (above 90%) of phenol utilizes a three-step process with high energy consumption and low yield, inevitably resulting in some environmental problems [5]. Direct hydroxylation of benzene to phenol is recognized as the most practical method in chemical industry. However, the Csp2–H bonds are kinetically inert and thermodynamically strong [6], and, at the same time, the hydroxy group of phenol is easily over-oxidized in the reaction. Therefore, the production of phenol via direct activation of a C–H bond from benzene has been a topic of potential interest [7,8,9,10,11] and a challenging assignment [12,13,14].

In the last two decades, remarkable achievement has been witnessed in one-step hydroxylation of benzene to phenol with O2 [15,16,17,18,19,20,21] and N2O in gaseous phase [22,23,24]. Meanwhile, many scientists have focused on direct hydroxylation of benzene to phenol in liquid phase with H2O2 as the environmentally benign oxidant; this process is inexpensive, readily available, can be handled concisely, and only produces water as by-product [25,26,27,28].

Heteropoly acids with a tunable structure at the molecular level are used in catalytic reaction and in designing compounds exhibiting special characteristics such as facilely modified acid–base properties and redox potential. There have some achievements in hydroxylation of benzene to phenol [29]. Vanadium is an efficient catalytic element; the catalytically active center of V-containing Keggin-type heteropoly acids contain vanadium(IV) or vanadium(V) and are capable of mutual transformation through electron transfer. During this process, the Csp2–H bond is activated and the direct result is that benzene can be oxidized by oxidants [30]. For the same reaction, catalyst state, and post-treatment of the heterogeneous catalysis, the catalytic reaction usually takes place on a solid surface, and the substrate needs to be sufficiently diffused to the active site and react adequately at the boundary layer. The products also need to be desorbed from the surface by the reverse process, which relates to the pore structure, such as pore size, volume, distribution, and the micro-environmental state of the surface. These features can affect the catalytic activity and selectivity significantly and are generally different from the homogeneous reaction. Immobilization of heteropolyacids with catalytically active compounds is an effective and easy heterogeneous catalytic strategy.

Recently, a dual-catalysis non-noble metal system has received great attention [31]; the system uses graphitic carbon nitride (g-C3N4) and Keggin-type polyoxometallate H5PMo10V2O40 as catalyst and LiOAc as an effective additive. In this catalytic system, effective catalytic components are composed of several substances including g-C3N4, PMoV2, and LiOAc. However, the yield of hydroxylation of benzene to phenol is 13.6% using that catalyst. In this work, we have prepared a reductant-free supported polyoxometallate heterogeneous catalyst system to oxidize benzene to phenol, leading to a higher conversion rate and selectivity than those of the reported catalysts. Heteropoly acid compounds were synthesized according to reported literatures [32] and we selected three kinds of carriers including carbon nitride (g-C3N4), montmorillonite, and activated carbon to support the heteropoly acid salt as catalyst. Direct hydroxylation of benzene to phenol is then reported.

Meanwhile, some of the reaction conditions were optimized. We focused on the catalytic performance of CN-PMoV2 because g-C3N4 has a strong binding charge capacity conducive to the stable transition of charged intermediates to the product as compared to the other two carriers.

Experimental

Chemicals

All solvents and reagents of analytical reagent (AR) grade were purchased commercially from Aladdin and used without further purification. Phosphomolybdic acid, sodium metavanadate, molybdenum trioxide, phosphoric acid (85%, by mass, aqueous solution), melamine, montmorillonite, activated carbon, benzene, acetonitrile, acetic acid, hydrogen peroxide (30%, by mass, aqueous solution), and tap water were deionized before use.

Preparation of the Na2H3PMo10V2O40 ·xH2O

Vanadium-substituted molybdophosphorus heteropoly acid sodium salt Na2H3PMo10V2O40 ·xH2O (PMoV2) was prepared according to the reported methods [33]. Briefly, 14.40 g (0.10 mol) of MoO3 and 2.44 g of NaVO3 (0.02 mol) were added into a 500-mL three-necked flask. Then, 250 mL of deionized water was added and the resulting suspension was stirred at reflux. Then, 1.15 g (85%, by mass) of phosphoric acid dissolved in 10.00 mL of water was added dropwise to the previous suspension. After 24 h of vigorous stirring, an orange-red clear solution was obtained. Finally, water was distilled off at 100 °C under reduced pressure and an orange-red solid powder was obtained. The PMoV2 was purified by recrystallization with water.

Preparation of the C3N4

According to the literature [34], the mesoporous carbon nitride (C3N4) carrier was prepared by thermal polymerization of melamine. Typically, 5.00 g of melamine was added to a crucible with a cover, the temperature was raised to 550 °C at a heating rate of 3 °C/s in a muffle furnace and maintained for 4 h. After sufficient reaction and cooling to room temperature, C3N4 powder was obtained by grinding the light-yellow block solid.

Preparation of the Supported Catalyst

Preparation of CN-PMoV2 involves dissolving 1 g of PMoV2 in 25 mL of deionized water, adding 5 g of C3N4, ultrasound for 15 min, then, reflux for 24 h in an oil bath. After full load, filtration, and washing, the light-yellow-green product was obtained. The powder was calcined in a crucible at 550 °C for 2 h, giving CN-PMoV2. The M-PMoV2 was prepared using the same method (PMoV2 load on montmorillonite). It was not required to roast for the preparation of C-PMoV2 which is different from that of the former.

Characterization methods

FT-IR spectra were determined on a Nicolet NEXUS 670 FT-IR spectrophotometer using KBr discs in the 400–4000-cm−1 region under atmospheric conditions. Inductively coupled plasma (ICP) was measured with an Optima 5300 DV ICP-optical emission spectrometry system and the heterogeneous samples were preprocessed by acid digestion with hydrochloric acid. The reaction products of oxidation were monitored using a GC-7890II gas chromatograph (GC) with an OV-1701 column (50 m × 0.25 mm × 0.25 μm) in a flame ionization detector (FID). Thermogravimetric analysis (TGA) measurements were determined with a NETZSCH STA 449 thermal analyzer using an α-Al2O3 crucible. Scanning electron microscopy (SEM) images were recorded with an FEI QUANTA 200 microscope. The pore diameter and specific surface area of the materials were measured with a Beishide 3H-2000BET-A tester at 77 K.

Catalytic test

The hydroxylation of benzene to phenol is as follows: catalyst (CN-PMoV2 0.10 g) was added to a 25-mL round-bottom flask, with 5.00 mL of acetic acid and 5.00 mL of acetonitrile as solvent, 1 mL of benzene (12.2 mmol), and 3.4 mL of H2O2 (30%, 34 mmol) added successively. After mixing the reaction substrate at room temperature, the reaction was conducted at 80 °C for 20 h with vigorous stirring in an oil bath. After completion of the reaction, the mixture was filtered to remove the solid catalyst for the next reaction after cleaning and drying. Simultaneously, the liquid mixture was analyzed by GC with an FID and a capillary column (SE-54; 50 m × 0.25 mm × 0.25 μm). The temperature of the GC column oven was keep at 60 °C for 2 min and then increased to 180 °C at a rate of 8 °C/min; injection temperature and detection temperature were set at 240 and 180 °C, respectively.

Results and discussion

Characterization of the catalyst

Heteropoly acid with Keggin structure appears on four obvious characteristics peaks in the range of 1100–700 cm−1. The FT-IR spectra of Na2H3PMo10V2O40 is presented in Fig. 1a. The four characteristic bands for the Keggin structure of Na2H3PMo10V2O40 appear at 1060 (P–Oa), 960 (Mo–Od), 868 (Mo–Ob–Mo), and 792 (Mo–Oc–Mo) cm−1, respectively [35].

Fig. 1
figure 1

The FT-IR spectra and thermogravimetric analysis of Na2H3PMo10V2O40

Full size image

TGA of Na2H3PMo10V2O40 was performed in nitrogen atmosphere at a scanning of 12 °C/min between 40 and 800 °C and shown in Fig. 1b. The resulting heteropoly acid has four weight-loss processes, corresponding to loss of adsorbed water, crystal water, combined with water, and structured water, respectively. A small amount of quality reduction of PMoV2 before 100 °C resulted from the loss of the adsorbed water [36]; 100–200 °C loss of the crystal water; 200–350 °C and 350–500°C loss of the combined with water and structured water, respectively.

The metal element content of the catalyst was quantitatively analyzed by ICP-atomic emission spectroscopy (ICP-AES). The results are shown in Table 1. It was found that the molar ratio of Mo-to-P was 9.91 and the molar ratio of V-to-P was 2.01, which was close to stoichiometric ratios of 10 and 2; that is to say, the synthetic sodium phosphomolybdate has a higher purity.

Table 1 Elemental analysis of the catalyst
Full size table

To obtain textural properties of the g-C3N4 support (Fig. 2a inset) and the as-prepared CN-PMoV2 (Fig. 2b inset) catalyst, the isotherms and the Barrett–Joyner–Halenda (BJH) pore size distributions were recorded. The nitrogen adsorption–desorption isotherm curves reveal a surface area of 22.896 m2/g and type IV adsorption isotherms, revealing the mesoporous structure of the g-C3N4. The BJH desorption average pore diameter is 19.575 nm. After the loading of PMoV2, the isotherm shape and the pore diameter of the CN-PMoV2 were slightly smaller compared with that of the g-C3N4, indicating that the PMoV2 did not block the pore distribution and Brunauer–Emmett–Teller (BET) surface area made a big increase, up to 244.330 m2/g, indicating adsorption on the surface successfully and increasing the BET surface area after introduction of PMoV2.

Fig. 2
figure 2

Pore size and particle size distribution characterizations of the catalyst

Full size image

In the microstructure of g-C3N4 and CN-PMoV2, based on SEM analysis (Fig. 3), the carrier adsorbs the heteropolyacid salt catalyst as the active site on the surface evenly, in which outward expansion makes the specific surface area much larger than the carrier.

Fig. 3
figure 3

Scanning electron microscopy of g-C3N4 and CN-PMoV2

Full size image

Catalytic performance of direct hydroxylation of benzene to phenol

Heterogeneous catalysis for the aerobic oxidation of benzene to phenol by the supported polyoxometallate catalyst without a reducing agent was investigated. At first, the carried material vanadium-substituted phosphomolybdic heteropolyacids on the catalytic reaction were studied with results shown in Table 2. No phenol was obtained when we conducted the reaction without catalyst (entry 1) or with the V-free phosphomolybdic acid (entry 2).

Table 2 Performance of catalysts for the hydroxylation of benzene to phenol with H2O2
Full size table

As shown in Table 2, no phenol was obtained when no vanadium-substituted heteropoly acid compound was used (entry 2). The hydroxylation of benzene to phenol with hydrogen peroxide by PMoV2 in a homogeneous reaction produced less than 5.0% phenol with conversion of benzene at 10.1% (entry 3). This demonstrated that V is indispensable to catalyze benzene hydroxylation to phenol with H2O2. Similar yields (29.0–30.5%) of phenol were obtained using C-PMoV2 (entry 5) and M-PMoV2 in heterogeneous reactions (entry 6). In addition, it can be seen that CN-PMoV2 provided a satisfactory conversion rate of 38.9% and the selectivity was over 99.0% using H2O2 as the oxidant (entry 4) compared to its latter two (entries 5, 6). The results showed that the role of vanadium is indispensable in the catalytic process, in which the catalytic performance is far superior to vanadium-free heteropoly compounds. Then, the loading process greatly improves the catalytic activity and the selectivity of phenol.

Oxygen, t-butylhydroperoxide (TBHP), and H2O2 were chosen as oxidants in our study (entries 1, 2, and 3, respectively). Among the three oxidants, hydrogen peroxide is the most efficient one compared to other two oxidants, oxygen, or TBHP (Table 3, entries 1, 2, and 3). Further optimization of different solvents, it was found the mixed solvent of acetic acid and acetonitrile is the most efficient compared to the individual acetic acid or acetonitrile (Table 3, entries 3, 4, 5), which maybe contributed to the fact strong acidity conditions can accelerate the decomposition of H2O2 and thus reduce the yield of phenol, while weak acid environment can increase the yield. In addition, no phenol is detected in the solvent-free reaction [37] (entry 5).

Table 3 Oxidizing agent and solvent for the hydroxylation of benzene to phenol
Full size table

Reaction temperature

The reaction temperature was studied and the results are shown in Fig. 4a. As we can see from Fig. 4a, the yield of phenol is extremely low when the temperature is below 60 °C, while the yield of phenol increases rapidly from 17.4 to 38.8% when the temperature increases from 60 to 80 °C. Therefore, we choose 80 °C as an optimum temperature for this catalytic system.

Fig. 4
figure 4

a Reaction temperature, b reaction time, c catalyst amount, and d hydrogen peroxide amount for hydroxylation of benzene to phenol with H2O2

Full size image

Reaction time

The optimum reaction time was studied and the results are shown in Fig. 4b: The conversion of benzene to phenol was significantly low in the first 6 h with 15.1% of yield. In the second 6 h, the yield continues to increase, while after 20 h, the yield increased to 38.8% dramatically. Then, the yield starts to decrease, indicating that the phenol was oxidized after 20 h.

Catalyst amount

We investigated the catalyst amount for the reaction and the results are shown in Fig. 4c. From Fig. 4c., we found the yields of phenol increased gradually from 14.9 to 38.8% when the amount of catalyst was increased from 50 to 100 mg. However, the yield of phenol showed a downward trend as the amount of catalyst continued to increase (more than 100 mg). Therefore, it was found 100 mg is the best amount of catalyst in our exploration, probably due to the fact H2O2 can be quickly decomposed when adding more catalyst.

Hydrogen peroxide amount

The effect of the amount of hydrogen peroxide on the yield of phenol is shown in Fig. 4d. As we can see from Fig. 4d, the yield of phenol increased rapidly from around 5.1 to 38.8% as the molar ratio of hydrogen peroxide to benzene was increased from 1 to 3 eq, while the yield was significantly decreased when the amount of hydrogen peroxide was increased from 3 to 4 eq. Possible reasons are as follows: the catalyst accelerates the decomposition of hydrogen peroxide when catalyzing the hydroxylation of benzene to phenol; the direct result is that only part of the hydrogen peroxide added to the reaction system is involved in the hydroxylation of benzene to phenol. In addition, too much hydrogen peroxide further oxidizes phenol to other byproducts and leads to a decrease in the overall yield of phenol. Our exploration indicates the yield of phenol is maximized when the molar ratio of hydrogen peroxide to benzene is three.

Catalyst reusability

The reusability of CN-PMoV2 catalyst was tested and the results showed that it can be easily recovered by direct filtration. Further, the catalytic capacity of recovered CN-PMoV2 catalyst was studied. Expectedly, we found that this catalyst can be repeatedly used up to 3 times during the hydroxylation reaction of benzene to phenol with H2O2, while the selectivity of hydroxylation to phenol is still up to 99%. Although the conversion of benzene to phenol gradually decreases during the repeated use (three times), it is still a good catalyst during the reused process. The FT-IR spectra of initial and recycled CN-PMoV2 are shown in Fig.  5 with lines 1, 2, 3, and 4 expressing the first, second, third, and fourth use, respectively. Based on the above experiment results, we speculate that PMoV2 adsorption on the surface of g-C3N4 is not very strong, and that acidic, heated conditions or strong stirring facilitate the shedding of catalytically active substances.

Fig. 5
figure 5

FT-IR spectra of initial and recycled catalyst

Full size image

Possible mechanism

The reaction mechanism for this reaction is suggested based on reported procedure [28, 36]. In this reaction system, benzene and hydrogen peroxide are respectively contacted with the heteropoly compound, which is adsorbed on the surface of the supporter. During this process, the benzene structure is activated to cause cleavage of the C–H bond; at the same time, pentavalent vanadium is reduced to tetravalent valence. In a corresponding process, tetravalent vanadium is oxidized to pentavalent vanadium by hydrogen peroxide under acidic conditions. Cross-conversion of tetravalent vanadium and pentavalent vanadium through electron transfer achieved sustained catalytic effect [30, 38] (Fig. 6).

Fig. 6
figure 6

A possible catalysis mechanism by CN-PMoV2

Full size image

Conclusions

In conclusion, the vanadium-substituted phosphomolybdic heteropoly acid sodium salt immobilized on C3N4 (CN-PMoV2) provided a satisfactory catalytic effect for hydroxylation of benzene to phenol. Under optimal reaction conditions [benzene (1.00 mL, 11.22 mmol), hydrogen peroxide (30 wt%, 3.40 mL, 33.90 mmol), acetonitrile (5.00 mL), acetic acid (5.00 mL), catalyst (100 mg), 80 °C, 20 h], the yield of phenol is more than 38.0% and the selectivity is up to 99.1%. As a heterogeneous catalytic system in this subject, the catalyst is easily recycled and still exhibits some certain catalytic effect even after three repeated uses.