C3N4-H5PMo10V2O40: a dual-catalysis system for reductant-free aerobic oxidation of benzene to phenol.

Hydroxylation of benzene is a widely studied atom economical and environmental benign reaction for producing phenol, aiming to replace the existing three-step cumene process. Aerobic oxidation of benzene with O2 is an ideal and dream process, but benzene and O2 are so inert that current systems either require expensive noble metal catalysts or wasteful sacrificial reducing agents; otherwise, phenol yields are extremely low. Here we report a dual-catalysis non-noble metal system by simultaneously using graphitic carbon nitride (C(3)N(4)) and Keggin-type polyoxometalate H(5)PMo(10)V(2)O(40) (PMoV(2)) as catalysts, showing an exceptional activity for reductant-free aerobic oxidation of benzene to phenol. The dual-catalysis mechanism results in an unusual route to create phenol, in which benzene is activated on the melem unit of C(3)N(4) and O2 by the V-O-V structure of PMoV(2). This system is simple, highly efficient and thus may lead the one-step production of phenol from benzene to a more practical pathway.

As an important commodity chemical, phenol is industrially produced by the three-step cumene process that suffers from a low one-pass yield ca. 5%, high energy cost and large amount of by-products12. Attempts to overcome these problems prompt the environmental benign one-step processes from benzene to phenol34567891011121314, in which the direct oxygenation of benzene to phenol by molecular oxygen (O2) is the most industrially important due to atom economy and economic superiority512. For decades, various catalysts or catalytic systems have been developed for this aerobic oxidation2, including low-temperature liquid-phase reactions591011, membrane process12 and high-temperature gas-phase catalyses151617. However, promoting its practical application remains great challenge because of the very low efficiency of the currently available catalytic systems2.

Benzene and O2 are all inert raw materials in low-temperature liquid-phase reactions. To oxidize benzene to phenol by O2, sacrificial reducing agents, i.e. H2, CO or ascorbic acid, are usually required to generate active oxygen species29. Otherwise, noble metal catalysts are needed1819202122, for example, palladium acetate [Pd(OAc)2] with a polyoxometalate (POM) can catalytically convert benzene to phenol in mild liquid-phase aerobic media18192021. The utilization of noble metals and/or wasteful reducing agents will largely increase the cost of the catalytic system2; therefore, it is very attractive to develop a reductant-free aerobic oxidation of benzene to phenol catalyzed by a non-noble metal catalyst, which is even referred to as a “dream oxidation” in chemical industry5.

Here we report a non-noble metal dual-catalysis system C3N4-H5PMo10V2O40 for efficient aerobic oxidation of benzene to phenol unaided by any reductant. Graphitic carbon nitride (C3N4) is a low price, insoluble and stable solid material23, and we use it as a heterogeneous catalyst for activating benzene chemically24. POMs are transition-metal oxygen anion clusters with structural diversity and have been widely used as acid, redox, and bifunctional catalysts2526, among which H5PMo10V2O40 (PMoV2) is a V-containing POM well recognized as an efficient homogeneous catalyst for organic oxidations with O227. We thus reason that the well dispersion of C3N4 in PMoV2 solution gives rise to a molecular-level contact between them, for which the creation of phenol may be possible from the immediate attack of PMoV2 catalyst to the already activated benzene ring on C3N4 surface. Indeed, the present results prove that the combination of C3N4 and PMoV2 converts benzene to phenol with a high phenol yield in a low-temperature liquid-phase aerobic system without any reductant. A dual-catalysis mechanism is proposed for understanding the highly efficient process.

Results

The major sample of C3N4 employed in this work is designated as C3N4(580), with the number in the parenthesis indicating the temperature of 580°C for heating melamine in C3N4 preparation. We first measured the single catalyst by using C3N4(580) or PMoV2 alone. Table 1 shows that neither former nor later alone was able to transform benzene in the absence of reductants (entries 1 and 2). On the contrary, a phenol yield of 2.1% was achieved in the dual-catalysis system containing both C3N4(580) and PMoV2 even with only a small amount of water solvent (2 mL) (Table 1, entry 3). The phenol yield reached 9.1% by changing the solvent to 50 vol.% aqueous solution of acetic acid (Table 1, entry 4), and arose to the maximum value of 13.6% using LiOAc as an effective additive (Table 1, entry 5)5182021. The above results were obtained at 4.5 h and 130°C optimized from our detailed investigations on various conditions (see Supplementary Fig. S6 online). Many results have been reported on the oxidation of benzene to phenol59101112151617, but reductant-free aerobic oxidation of benzene is still scarcely reported so far. Compared to the previous results under the reductant-free condition, the phenol yield of 13.6% over C3N4(580)-PMoV2 is more than three times higher than the yield of 3.7% over the nano-plate vanadium oxide catalyst at a longer reaction time (10 h) and a higher temperature (150°C)28, and even exceeds the yields on noble metal catalysts [e.g., the homogeneous Pd(OAc)2-PMoVx (X = 1, 2, 3) gives the phenol yield around 10%1821, which sharply drops to 3.4% when Pd(OAc)2 is immobilized on porous supports for recovering21]. Moreover, the turnover frequency (TOF) of our work 5.9 h−1 calculated by the definition mmol phenol/(mmol POM catalyst × h reaction time) is much higher than the POM-catalyzed systems with CO as the sacrificial reducing agent (1.5 h−1)5, or with ascorbic acid as the sacrificial reducing agent (0.86 h−1 and 2.0 h−1)1011, convincing that our reductant-free catalysis is even more active than those reductant-aided systems. Therefore, the present non-noble metal catalytic system C3N4(580)-PMoV2 shows a remarkably superior efficacy at the reductant-free condition.

Table 1

Aerobic oxidation of benzene over various catalysts. *Reaction conditions: C3N4 0.1 g; PMoV2 (PMo, PW, VOSO4, PMoV3, PMoV1 or CsPMoV2) 0.4 g; benzene 4 mL; solvent 25 mL; O2 2.0 MPa; 130°C; 4.5 h

Entry	Catalyst	Solvent	LiOAc (g)	Phenol Yield (%)
1	C3N4(520, 550 or 580)	acetic acid (50 vol.%)	0.6	0
2	PMoV2	acetic acid (50 vol.%)	0.6	0
3	C3N4(580)-PMoV2	water (2 mL)	0	2.1
4	C3N4(580)-PMoV2	acetic acid (50 vol.%)	0	9.1
5	C3N4(580)-PMoV2	acetic acid (50 vol.%)	0.6	13.6
6	melamine-PMoV2	acetic acid (50 vol.%)	0.6	0
7	melem-PMoV2	acetic acid (50 vol.%)	0.6	0
8	C3N4(520)-PMoV2	acetic acid (50 vol.%)	0.6	0.3
9	C3N4(550)-PMoV2	acetic acid (50 vol.%)	0.6	6.1
10	C3N4(580)-PMo	acetic acid (50 vol.%)	0.6	0
11	C3N4(580)-PW	acetic acid (50 vol.%)	0.6	0
12	C3N4(580)-VOSO4	acetic acid (50 vol.%)	0.6	0
13	C3N4(580)-PMoV3	acetic acid (50 vol.%)	0.6	9.5
14	C3N4(580)-PMoV1	acetic acid (50 vol.%)	0.6	0
15	C3N4(580)-CsPMoV2	acetic acid (50 vol.%)	0.6	0

Heating melamine in air at high temperatures has been a common approach for preparing C3N4, so the influence of heating temperatures for melamine on this reaction is investigated. The XRD patterns of Fig. 1a shows that heating melamine at 520°C and 550°C led to the formation of graphitic C3N4 products of C3N4(520) and C3N4(550), similar to C3N4(580), but the low heating temperature 400°C resulted in melem, an intermediate toward C3N42930. The non-C3N4-mediated systems of melamine-PMoV2 and melem-PMoV2 yielded no product (Table 1, entries 6 and 7). Though C3N4(520) and C3N4(550) were also inactive when used alone (Table 1, entry 1), their combination with PMoV2 gave phenol yields of 0.3% and 6.1%, respectively (Table 1, entries 8 and 9), much lower than 13.6% for C3N4(580)-PMoV2. The results prove that the C3N4 sample obtained at the optimal temperature of 580°C is more active and in favor of the high phenol yield.

Figure 1

(a) XRD patterns for the products by heating melamine at 400, 520, 550 and 580°C; (b) Magnification of the peak (002) in the 2θ range 26 ~ 29° for the C3N4 products obtained at 520, 550 and 580°C.

We further explored catalytic systems containing C3N4(580) and other POMs. With the V-free POMs, i.e. H3PMo12O40 (PMo) or H3PW12O40 (PW), to company C3N4(580), no phenol product appeared (Table 1, entries 10 and 11), suggesting that the V species should be indispensable. Nonetheless, C3N4(580) with the non-POM vanadium species VOSO4 caused an inactive system either (Table 1, entry 12); as a consequence, it is the V species in POM framework that is synergically active with C3N4 for this reaction. Moreover, when the other two less frequently used V-containing POMs (PMoV1 and PMoV3) were tested, the results show that C3N4(580)-PMoV3 exhibited comparable activity to C3N4(580)-PMoV2, but C3N4(580)-PMoV1 was definitely inactive (Table 1, entries 13 and 14), which means that not all the V species in POM framework can catalyze this reaction with C3N4.

Discussion

According to previous studies, the V species in V-POMs are well accepted as the catalytically active sites for versatile organic oxidations31. Particularly, for liquid-phase aerobic oxidations, PMoV2 takes a catalytic effect through Mars-van Krevelen-type mechanism, where the lattice oxygen of PMoV2 selectively oxygenates organic substrates via a valence variation between V5+ and V4+2732. Neumann and co-workers2732333435 have systematically studied series of PMoV2-catalyzed homogeneous oxidations, and based on the Mars-van Krevelen mechanism they propose that the isomers of PMoV2 with vanadium atoms in adjacent positions (i.e. V-O-V structure) are more likely to form bridge defects, favoring higher activity in oxygen-transfer reactions. Therefore, only PMoV2 and PMoV3 with the highly active V-O-V structure in their frameworks can allow the occurrence of oxygen transfer in hydroxylation of benzene to phenol, while lack of V-O-V is responsible for the inactivity of PMoV1.

Nonetheless, PMoV2 or PMoV3 alone cannot catalyze the reaction because of inertness of the substrate benzene, suggesting that C3N4 should play a key role. Recently, Goettmann et al.2436 conclude an unusual activation of aromatic rings via transferring electron density from the melem unit of C3N4 to arene based on reaction results plus DFT calculations. Besides, for the high-temperature gas-phase oxidation of benzene with O2 over copper exchanged HZSM5, a bifunctional catalytic mechanism has been reported: phenol is produced from the simultaneous activation of benzene and O2 on zeolitic acid and Cu metal sites, respectively1617. From above analyses, a dual-catalysis mechanistic pathway is proposed for understanding the catalytic performance of C3N4-PMoV2 in Fig. 2. Benzene is firstly catalytically activated by the melem unit of graphitic C3N4, forming a transitional intermediate of electron-enriched benzene ring. Immediately, the original oxidation state of PMoV2 with V5+ species, designated as PMoV2[ox], attacks the intermediate ring to produce phenol, wherein the lattice oxygen of a V-O-V structure in PMoV2[ox] moves into the benzene ring with the PMoV2[ox] thus being reduced to the V4+-containing PMoV2[red]. Finally, the catalytic cycle is closed with the resume of PMoV2[ox] after O2 re-oxidizes V4+ of PMoV2[red] into V5+ species.

Figure 2

Proposed mechanistic pathway for C3N4-PMoV2-catalyzed aerobic oxidation of benzene to phenol.

In the dual-catalysis mechanism above, the role of C3N4 is activating benzene according to the previous finding that the π-conjugated melem unit of C3N4 could transfer electron density to aromatic rings2436. It is further revealed that high temperatures for thermal condensation of melamine would enhance the π-conjugation by connecting more tri-s-triazine and extending the polymeric network of C3N437. The (002) diffraction peak of C3N4 is assigned to the interlayer distance of its graphitic structure30. In our case, as shown in the magnification of XRD patterns in Fig. 1b, the gradual shifting of the (002) peak to larger degrees along with the raise of heating temperatures means the shortening of the stacking distance and thus the stronger overlap of π orbital in C3N42930, indicating that the activation of benzene would be improved by a higher heating temperature up to 580°C. This accounts for the activity order C3N4(520)-PMoV2 < C3N4(550)-PMoV2 < C3N4(580)-PMoV2. On the other hand, melem-PMoV2 is inactive because melem itself has no graphitic characteristic of C3N430.

Also according to the mechanism in Fig. 2, the catalyst PMoV2 will remain in its reduced state PMoV2[red] as the reaction occurs in O2-deficient environment. Thus we conducted a separate run by introducing a much less amount of O2 (0.3 MPa) (see Supplementary Information) into the batch reactor. In this case, the recovered PMoV2 was green and exhibited an eight-line signal in ESR spectra (Fig. 3), index of the reduced state PMoV2[red]510, whereas the fresh and recovered PMoV2 from O2-sufficient condition were orange and ESR silent, denoting the oxidation state PMoV2[ox]. The above phenomena and comparisons strongly evidence our proposal that there exists V5+/V4+ switch during the reaction.

Figure 3

ESR spectra of (a) fresh PMoV2, (b) recycled PMoV2 from the O2-insufficient reaction, and (c) recycled PMoV2 from the O2-sufficient reaction, entry 1 of Table 1.

Moreover, the activation and oxidation of benzene should occur simultaneously in this mechanism. In order to reflect this point, the well-known heterogeneous Cs salt of PMoV2, CsPMoV210, was tried as a partner with C3N4(580). Though CsPMoV2 was as active as PMoV2 (see Supplementary Table S1 online)10 in the presence of the sacrificial reducing agent ascorbic acid, C3N4(580)-CsPMoV2 was inactive in our reaction system (Table 1, entry 15). The SEM image for CsPMoV2 (see Supplementary Fig. S5 online) shows a spherical morphology with spheres diameters being 800 ~ 900 nm. This bulk CsPMoV2 may not contact well with another solid surface of C3N4(580), hindering the simultaneous attachment of substrate with the dual-catalyst. In other words, the intimate and efficient contacts among C3N4, benzene and PMoV2 are essential for implementing the overall catalytic cycle, which further supports our mechanism.

Besides benzene, the simplest alkyl aromatic molecule toluene was also attempted as the substrate to further investigate the catalytic behavior of C3N4(580)-PMoV2 for aerobic oxidation of aromatic rings (see Supplementary Table S2 online). C3N4(580) alone was inert in this system, and yet, bare PMoV2 exclusively produced methyl-oxygenated compounds of benzaldehyde (7.7%) and benzyl alcohol (1.4%) due to the side chain oxidations. For reductant-free oxidations of alkyl aromatics, early studies reveal that oxidations of benzylic C-H bond are preferred rather than the aromatic ring93839. On the contrary, the dual-catalysis system C3N4(580)-PMoV2 resulted in a desirable yield of cresols (0.4%) due to the ring oxidation. This feature suggests that C3N4-PMoV2 should have enhanced the reactivity of the alkylated benzene ring, enabling occurrence of the ring oxygenation through the dual-catalysis mechanism in Fig. 1.

Catalytic reusability was first investigated by recycling C3N4(580) alone (Fig. 4). The phenol yield slowly decreased from 13.6% for the fresh catalyst to 12.7% for 1st, 9.8% for 2nd, and still kept at 6.2% for 3rd recycling. The XRD pattern for the last recycled C3N4(580) indicates a stable structural stability due to its identical diffraction peak to that of the fresh one (see Supplementary Fig. S1 online). Therefore, the above decrease of phenol yield can be ascribed to the tar deposition according to the gradually darkened color (inserted photos in Fig. 4) and variation of C content (see Supplementary Information) of C3N4(580) during the recycling process. In fact, tar is still an inevitable over-oxidation byproduct, because the main product phenol is more reactive than the substrate benzene4618. Even so, when C3N4(580), PMoV2 and LiOAc were simultaneously recovered (see Methods), the phenol yield was 10.3% and 6.5% for the 1st and 2nd recycling, and still 2.1% for the 3rd recycling (Fig. 4).

Figure 4

Phenol yields and C3N4 recovery rate during the recycling test; insertion: C3N4 photo for each run.

All the above results demonstrate that the dual-catalysis non-noble metal system C3N4-PMoV2 provides a high phenol yield of 13.6% in reductant-free aerobic oxidation of benzene. A dual-catalysis mechanism involving cooperative activations of benzene on melem unit of C3N4 and O2 by V-O-V structure of PMoV2 is demonstrated for interpreting catalytic results. The present dual-catalysis process appears to be simpler, much more efficient and cost-effective when compared with the currently available catalytic systems, paving a promising step towards practical application of hydroxylation of benzene to phenol by molecular oxygen.

Methods

Materials and general methods

All chemicals were analytical grade and used as received. H3PMo12O40 (PMo) and H3PW12O40 (PW), purchased commercially, were dried before used. XRD patterns were collected on the Bruker D8 Advance powder diffractometer using Ni-filtered Cu Kα radiation source at 40 kV and 20 mA, from 5 to 50° with a scan rate of 0.2° S−1, and before measurements the samples were dried at 100°C for 2 h. Elemental analyses were performed on a CHN elemental analyzer (FlashEA 1112). BET surface areas were calculated from the sorption isotherms measured at the temperature of liquid nitrogen using a Micromeritics ASAP2010 analyzer; the samples were degassed at 300°C to a vacuum of 10−3 Torr before analysis. FT-IR spectra were recorded on a Nicolet 360 FT-IR instrument (KBr discs) in the 4,000–400 cm−1 region. ESR spectra were recorded on a Bruker EMX-10/12 spectrometer at X-band. The measurements were done at −110°C in a frozen solution provided by a liquid/gas nitrogen temperature regulation system controlled by a thermocouple located at the bottom of the microwave cavity within a Dewar insert.

Preparation of catalysts

Graphitic carbon nitride (C3N4)

The procedure for the synthesis of C3N4(580) is similar to the previous reports4041. Melamine was transferred into a crucible and heated in a muffle furnace under air at a rate of 15°C/min to reach the temperature of 580°C and kept at 580°C for 4 h, then the resulting yellow sample was cooled to room temperature in the oven. Melem, C3N4(520) and C3N4(550) were prepared by the similar method at 400°C, 520°C and 550°C, respectively.

H5PMo10V2O40 (PMoV2)

The Keggin-structured double V-containing POM was prepared according to the procedure described in our previous report42. The detail of the preparation of PMoV2 procedure is as the following. MoO3 (16.59 g) and V2O5 (2.1 g) were added to deionized water (250 mL). The mixture was heated up to the reflux temperature under vigorously stirring with a water-cooled condenser, then at 120°C the 85 wt% aqueous solution of H3PO4 (1.33 g) was added drop-wise to the reaction mixture. When a clear orange-red solution appeared, it was cooled to room temperature. The orange-red powder PMoV2 was obtained by evaporation of the solution to dryness, followed with re-crystallizing for purification.

Cs was prepared according to the literature10, with Cs2CO3 instead of CsNO3. FT-IR spectrum for CsPMoV2 is presented in Fig. S4b.

Catalytic tests

The hydroxylation of benzene was carried out in 100 ml stainless steel autoclave equipped with a mechanical stirrer and an automatic temperature controller. In a typical test, 0.1 g C3N4, 0.4 g PMoV2, 0.6 g LiOAc, and 4.0 mL benzene were added into 25 mL of the aqueous solution of acetic acid (50 vol%) successively. After the system was charged with 2.0 MPa O2 at room temperature, the hydroxylation reaction was conducted at 130°C K for 4.5 h with vigorous stirring. After the reaction, 1, 4-dioxane was added into the product mixture as an internal standard for product analysis. The mixture was analyzed by a gas chromatograph (GC) with a FID and a capillary column (SE-54; 30 m × 0.32 mm × 0.25 μm). Yield of phenol was calculated as mmol phenol/mmol initial benzene. Catechol, hydroquinone and benzoquinone were not detected by our GC analysis, so the tar that cannot be detected by the GC technique was the over-oxidation product.

Recycling of the catalyst system

After the reaction, the reaction mixture was centrifuged and the solid C3N4(580) was recovered, followed by washing with acetic acid and dried in vacuum, and then reused in the next run. After the solid C3N4(580) was separated by centrifuging, water was added into the left liquid phase followed by extraction with isopropyl ether. The combined aqueous extracts were filtered and concentrated by evaporation under reduced pressure. The resulting solid mixture containing used PMoV2 and LiOAc was obtained.

Author Contributions

Z.Y.L. and J.W. conceived and designed the experiments. Z.Y.L. performed all the experiments and analyzed all the data. G.J.C. and W.L.G. performed catalysts characterization. Z.Y.L., Y.Z. and J.W. co-wrote the paper. All the authors discussed the results and commented on the manuscript.