Cerium promoted V-g-C3N4 as highly efficient heterogeneous catalysts for the direct benzene hydroxylation

A series of Cex-V-g-C3N4 catalysts with different cerium content were synthesized by a facile co-assembly method. Compared with pure V-g-C3N4 catalyst, the addition of cerium facilitated the high dispersion of vanadium species as well as the benzene adsorption ability of the corresponding catalysts. Also, the existence of cerium promoted the partial reduction of vanadium species, which improved the redox property of vanadium species as the active centres. The Cex-V-g-C3N4 catalysts showed considerably improved activity in the benzene hydroxylation reaction compared with V-g-C3N4 catalyst. Among the catalysts studied, Ce0.07-0.07 V-g-C3N4 exhibited the best catalytic activity with a benzene conversion of 33.7% and a phenol yield of 32.3% with good structural and catalytic stability, while only 24.7% of benzene conversion and phenol yield of 24.2% were obtained over 0.07 V-g-C3N4.


Introduction
Phenol, as an important chemical intermediate in industry, is widely employed in the synthesis of aniline, resins, plastics, bactericides and agrochemicals [1]. However, the current phenol production is based upon the three-step cumene process, which has some inevitable disadvantages, e.g. the by-product acetone with low market demand, the complex synthesis steps and the high energy consumption [2]. From the view of green chemistry, the direct hydroxylation of benzene to phenol has attracted great interest in the past few decades.
2018 The Authors. Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original author and source are credited.

Synthesis
Melamine, vanadylacetylacetonate (C 10 H 14 O 5 V) and Ce(NO 3 ) 3 ·6H 2 O were all purchased from Aladdin Industrial Corporation and used without further purification. g-C 3 N 4 was prepared by the direct calcination of melamine at 550°C for 2 h under nitrogen atmosphere [26]. 0.07 V-g-C 3 N 4 sample was obtained as follows: 0.42 g of C 10 H 14 O 5 V and 2.50 g of melamine were mixed with 50 ml of ethanol.
After stirring vigorously at 50°C for 1 h, the solution was dried at 60°C overnight. The resulting solid sample was calcined in N 2 from room temperature to 550°C with a heating rate of 2°C min −1 and kept at 550°C for another 2 h. After cooling to room temperature, the product was collected and denoted as 0.07 V-g-C 3 N 4 , where 0.07 represented the theoretical weight content of vanadium. 0.05 V-g-C 3 N 4 and 0.10 V-g-C 3 N 4 catalysts were prepared according to the same synthesis procedure as for 0.07 V-g-C 3 N 4 by adding desired amount of C 10 H 14 O 5 V.
As for 0.07Ce-g-C 3 N 4 , the synthesis method was also similar to that of 0.07 V-g-C 3 N 4 , except for adding 0.23 g of Ce(NO 3 ) 3 ·6H 2 O rather than 0.42 g of C 10

Benzene hydroxylation reaction
The direct benzene hydroxylation reaction was conducted as follows. Typically, 1 ml of benzene, 10 ml of 80 wt% acetic acid and 40 mg of catalyst were added into a 25 ml three-necked flask connected with a reflux condenser. After heating to 70°C, 3.5 ml of 30 wt% H 2 O 2 was added dropwise in 30 min with vigorously stirring. The reaction mixture was stirred for another 4 h. After reaction, the catalyst was separated by centrifugation and the content of liquid products was analysed immediately by gas chromatography using toluene as the internal standard.

Material characterization
Elemental analysis was performed with a Thermo Elemental IRIS Intrepid inductively coupled plasmaatomic emission spectrometer (ICP-AES). Transmission electron microscopic (TEM) images were acquired with a FEI Tecnai G 2 F20 S-Twin field-emission transmission electron microscope operated at 200 kV. Elemental mapping was conducted using a Philips XL 30 microscope with energy dispersive X-ray spectrometer operated at 30 kV. Fourier transform infrared (FT-IR) spectra were recorded with a Nicolet iS10 infrared instrument using KBr discs. X-ray diffraction (XRD) patterns were recorded on a Bruker D8 Advances X-ray diffractometer using Cu-Kα radiation with a voltage of 40 kV and a current of 40 mA. X-ray photoelectron spectra (XPS) were recorded with a Perkin-Elmer PHI 5000C ESCA system equipped with a dual X-ray source by using Mg Kα (1253.6 eV) anode and a hemispherical energy analyser. Specific surface area results were obtained at 77 K using a Micromeritics Tristar 3000 apparatus. The benzene adsorption was measured using a Hiden intelligent gravimetric analyser. All samples were degassed under a vacuum of less than 10 −3 Pa at 300°C for 6 h prior to the adsorption measurement.

Results and discussion
The results of benzene hydroxylation reaction over all catalysts studied are shown in table 1 and  electronic supplementary material, table S1. By optimizing the vanadium content in the catalysts, the best catalytic activity in the benzene hydroxylation reaction was obtained over 0.07 V-g-C 3 N 4 catalyst with 7 wt% of vanadium content. With the introduction of cerium into the 0.07 V-g-C 3 N 4 catalyst (Ce/V = 0.05), the benzene conversion is improved slightly from 24.7% to 25.3% (entry 2, table 1). With the further rise of cerium content, the benzene conversion increases remarkably to the highest 33.7% over Ce 0.07 -0.07 V-g-C 3 N 4 (entry 3, table 1) and then decreases to 29.4% over Ce 0.10 -0.07 V-g-C 3 N 4 (entry 4, table 1), while the phenol selectivity remains almost constant with the variation of cerium content. Moreover, the carbon balance value shown in table 1 are greater than 99% over the three Ce-containing catalysts, indicating a high carbon yield with negligible side reactions. The TOF value of 17.1 h −1 based on the vanadium content also indicates that Ce 0.07 -0.07 V-g-C 3 N 4 is the most active catalyst. Thus, we deduce that a proper molar ratio of Ce/V possibly facilitates both the redox properties of V species and the efficient decomposition of H 2 O 2 , leading to the remarkable catalytic activity in the benzene conversion reaction [27].
In order to elucidate the synergistic effects between V and Ce species, we selected g-C 3 N 4 , 0.07 Vg-C 3 N 4 and Ce 0.07 -0.07 V-g-C 3 N 4 as the representative catalysts to carry out detailed characterizations. Generally, the specific surface area of a catalyst has a certain influence on the catalytic activity. As shown in table 2, the specific surface area increases slightly after the introduction of metal species compared with that of the non-modified g-C 3 N 4 catalyst (entries 1-3) and Ce 0.07 -0.07 V-g-C 3 N 4 has the highest specific surface area. However, the quite different catalytic performances between 0.07 V-g-C 3 N 4 and Ce 0.07 -0.07 V-g-C 3 N 4 demonstrate that the specific surface area is not the main reason for the different catalytic activity.
The morphologies of the catalysts were studied by TEM. As shown in figure 1a-d, all the samples exhibit the typical large lamellar structure, indicating that the structure of g-C 3 N 4 support remains after the incorporation of metal species. As for 0.07 V-g-C 3 N 4 and Ce 0.07 -0.07 V-g-C 3 N 4 catalysts (figure 1bd), no obvious metal particles are observed over both catalysts, implying the high dispersion of metal species. Also, the element-mapping images of 0.07 V-g-C 3 N 4 and Ce 0.07 -0.07 V-g-C 3 N 4 catalysts (figure 1e-g) further demonstrate that both V and Ce species are highly dispersed over the g-C 3 N 4 support.
However, it is worth noting that the vanadium dispersion of 0.07 V-g-C 3 N 4 is slightly poorer than that of Ce 0.07 -0.07 V-g-C 3 N 4 . This implies that the existence of cerium may facilitate the high dispersion of vanadium, leading to the high catalytic activity [28]    The graphitic stacking structures of the catalysts were also confirmed by XRD patterns. For pure g-C 3 N 4 (electronic supplementary material, figure S1a), two distinct diffractions are observed at ca 13.2°and 27.5°, corresponding to (100) diffraction of in-planar repeating motifs of tris-s-triazine units and (002) diffraction of interlayer stacking aromatic systems, respectively [29]. After the addition of metal species, both corresponding (100) and (002) diffraction peaks become broader, suggesting that the ordered structure of g-C 3 N 4 support decreases slightly. In addition, the (002) peaks of 0.07 V-g-C 3 N 4 and Ce 0.07 -0.07 V-g-C 3 N 4 catalysts both shift to low angle slightly compared with that of g-C 3 N 4 , indicating that the metal oxides were incorporated into the g-C 3 N 4 sheets successfully [30]. Furthermore, no distinct peak originating from either vanadium or cerium species can be observed, which implies that the metal species are in non-crystallized state or dispersed well on the g-C 3 N 4 layers. These results are consistent with the results of TEM.   The FT-IR spectra of g-C 3 N 4 , 0.07 V-g-C 3 N 4 and Ce 0.07 -0.07 V-g-C 3 N 4 catalysts are shown in figure 2. As for g-C 3 N 4 (figure 2a), the major bands between 1200 and 1650 cm −1 are attributed to the stretching modes of CN heterocycles, while the band at 804 cm −1 corresponds to the stretching mode of triazine units (C 6 N 7 ). The broad bands in the range of 3000-3400 cm −1 can be ascribed to the stretching of N−H bonds in both uncondensed amino groups and adsorbed water molecules [31]. For 0.07 V-g-C 3 N 4 and Ce 0.07 -0.07 V-g-C 3 N 4 (figure 2b,c), their spectra are similar to that of g-C 3 N 4 except for a small additional band at 2157 cm −1 corresponding to the disturbance of conjugated N=C−N units after metal doping [32,33].
In order to investigate the surface chemical composition of the catalysts, XPS measurement was carried out. As shown in electronic supplementary material, figure S2, C and N species, which refer to the peaks at binding energies of 288.0 (C 1s) and 400.0 (N 1s), are the main elements in all catalysts. By subtracting the peak area of the contaminant carbon C 1s at 284.6 eV, the peak area ratios of C 1s/N 1s of g-C 3 N 4 , 0.07 V-g-C 3 N 4 and Ce 0.07 -0.07 V-g-C 3 N 4 catalysts are 0.89, 0.83 and 0.73, respectively. As the theoretical molar ratio of C/N in g-C 3 N 4 is 3/4, the correction factor between peak area ratio of C/N and molar ratio of C/N should be 1.187. Based on the correction factor, the C/N molar ratio of 0.07 V-g-C 3 N 4 and Ce 0.07 -0.07 V-g-C 3 N 4 can be calculated as 3/4.3 and 3/4.9, respectively, which implies the formation of N-rich carbon nitride by the incorporation of metal species. The reason for   the C/N molar ratio variation may be caused by the increase of -NH groups on the g-C 3 N 4 support surface as the metal species destroy the ordered structure of g-C 3 N 4 partially. As reported previously, in preparation of C−N materials, the addition of metal species restrained the decomposition of nitrogen species and accelerated the decomposition of carbon species, leading to high N content in the samples [34,35]. As shown in figure 3a, the C 1s spectrum of g-C 3 N 4 can be deconvoluted into three peaks with binding energies of 284.6, 285.7 and 287.5 eV, corresponding to the graphitic carbon (C−C), C−O and sp 2 hybridized carbon (N−C=N), respectively [36]. After the incorporation of metal species, the peak intensities of both C−O and N−C=N increase, while that of C−C decreases. These imply that the rigid structural regularity of g-C 3 N 4 was partially broken and more defects were produced on the surface of g-C 3 N 4 , generating more sites for metal species anchoring. In figure 3b, the N 1s spectrum of g-C 3 N 4 shows three main peaks at 398.2, 399.3 and 400.6 eV, assigned to triazine nitrogen (C=N−C), tertiary nitrogen (N−(C) 3 ) and amino function group (N−H), respectively [37]. It is worth noting that both the C 1s and N 1s peak positions of 0.07 V-g-C 3 N 4 and Ce 0.07 -0.07 V-g-C 3 N 4 are slightly shifted to high binding energy regions compared with those of g-C 3 N 4 . Since the graphite analogue CN matrix is believed to stabilize the metal species with the 'coordination nest' consisting of C and N atoms [38], it is reasonable to deduce that a strong interaction between metal species and g-C 3 N 4 support may exist, which is reflected in the XPS study. In addition, as shown in the V 2p 3/2 XPS spectra (figure 4), 0.07 V-g-C 3 N 4 merely contains V 5+ and V 4+ , while V 3+ species appears in Ce 0.07 -0.07 V-g-C 3 N 4 . Also, the content of low oxidation state V (V 4+ , V 3+ ) species increases after the addition of cerium ( figure 4b-d). These may be ascribed to the reduction of partial V 5+ species by the Ce 3+ species during calcination [39]. In fact, similar results were reported for other Ce-V catalysts [40,41]. As listed in table 3, the (V 4+ +V 3+ )/V 5+ ratio (calculated by the corresponding area ratio) increases gradually from 0.45 to 1.15 with the increase of cerium content, corresponding to the rise of low oxidation state vanadium species. It seems that Ce 3+ /Ce 4+ species play an important role in maintaining the content of V 4+ and V 3+ species. During the oxidation of benzene to phenol over vanadium-containing catalyst using H 2 O 2 as an oxidant, the low valent vanadium species can be oxidized to V 5+ -O-O• or V 4+ -O-O• radicals which perform as main active centres for the conversion of benzene to phenol [42]. In electronic supplementary material, figure S5, the V 2p 3/2 XPS measurements of both 0.07 V-g-C 3 N 4 and Ce 0.07 -0.07 V-g-C 3 N 4 catalysts were conducted after four recycles. The recovered Ce 0.07 -0.07 V-g-C 3 N 4 catalyst shows little change in the ratios of (V 4+ +V 3+ )/V 5+ and V 3+ /V 4+ relative to the fresh one, indicating that the excellent redox ability of Ce 3+ /Ce 4+ improves the redox cycles of V 5+ /V 4+ and V 4+ /V 3+ . Comparing with Ce 0.07 -0.07 V-g-C 3 N 4 , the recovered Ce 0.07 -0.07 V-g-C 3 N 4 catalyst has small amount of Ce 4+ species with the appearance of Ce 4+ fingerprint (electronic supplementary material, figure S6) [43]. The existence of Ce 3+ /Ce 4+ is confirmed, which promotes the formation of low state vanadium. However, •OH radicals produced by V 4+ lead to overoxidation of benzene [44]. It is believed that V 4+ species are more reactive than V 5+ ones during the benzene conversion [45]. Thus a proper molar ratio of (V 4+ +V 3+ )/V 5+ may be beneficial to achieve a good conversion of benzene and a high selectivity of phenol. Therefore, in the present case, it is reasonable to indicate that the addition of proper cerium amount improves the redox capacity of active vanadium species along with the efficient decomposition of hydrogen peroxide. The adsorption properties of g-C 3 N 4 , 0.07 V-g-C 3 N 4 and Ce 0.07 -0.07 V-g-C 3 N 4 catalysts were evaluated by studying the gravimetric uptake of benzene. In figure 5a, pure g-C 3 N 4 exhibits an evident adsorption of benzene, probably attributed to the strong π-π interactions between benzene molecules and g-C 3 N 4 [28]. Interestingly, Ce 0.07 -0.07 V-g-C 3 N 4 exhibits the highest benzene adsorption ability among all the catalysts. As reported earlier [46,47]   (e.g. amino or hydroxyl groups) in activated carbon, the adsorption affinity for the nonpolar molecules, e.g. benzene, was enhanced. Therefore, combining with the results of XPS, it is reasonable to deduce that the increase of nitrogen-containing groups in g-C 3 N 4 support after doping metal species can promote the benzene adsorption ability of the corresponding catalyst effectively, leading to the excellent benzene conversion. The reusability of Ce 0.07 -0.07 V-g-C 3 N 4 catalyst was investigated as it showed the best catalytic activity in the target reaction (figure 5b). After each reaction, the Ce 0.07 -0.07 V-g-C 3 N 4 catalyst was separated, washed and dried for the next fresh reaction. After four recycles, the catalyst retained high activity without marked loss in both benzene conversion and phenol selectivity. The vanadium content of both the last recycled catalyst and the reaction solution was measured by ICP-AES, and no leaching of vanadium happened (entry 9, electronic supplementary material, table S1). In table 3, as expected, the (V 4+ + V 3+ )/V 5+ peak area ratio of recovered Ce 0.07 -0.07 V-g-C 3 N 4 is 0.85, much higher than that of recovered 0.07 V-g-C 3 N 4 . These further demonstrate that the existence of Ce species facilitates the formation of V 4+ and V 3+ species. Moreover, XRD and FT-IR (electronic supplementary material, figures S3 and S4) measurements were conducted for the recovered Ce 0.07 -0.07 V-g-C 3 N 4 from the fourth run. The structural properties of the recovered catalyst are nearly identical to those of fresh one, indicating the strong interactions between metal species and g-C 3 N 4 support which accounts for high stability of Ce 0.07 -0.07 V-g-C 3 N 4 in the direct benzene hydroxylation. The excellent stability is because of abundant defects of g-C 3 N 4 after cerium modification that improve the dispersion and stability of V species. Although the role of V 3+ species remains to be further studied, the work gives an insight into the design of efficient catalysts for the titled reaction in future.
A possible reaction mechanism of benzene hydroxylation is proposed based on literature work [22]. Benzene is chemically absorbed onto the surface of g-C 3 N 4 , and then the surface dispersed V 4+ species are oxidized by H 2 O 2 to produce V 5+ -O-O• radicals and H 2 O. The main active V 5+ -O-O• radicals are rapidly activated and react with absorbed benzene to gain the target phenol and V 5+ is reduced to V 4+ . At the same stage, the redox V 5+ /V 4+ is supported by Ce 3+ /Ce 4+ to accelerate the reaction into the next recycle.

Conclusion
In this work, we developed a facile method to synthesize cerium-doped V-g-C 3 N 4 catalysts for the direct oxidation of benzene to phenol. The dispersion of vanadium species was distinctly improved after the addition of cerium species. Among the catalysts studied, Ce 0.07 -0.07 V-g-C 3 N 4 showed excellent catalytic performance with good reusability in the titled reaction, ascribed to the enhanced vanadium redox property, improved benzene adsorption ability and strong interactions between metal species and g-C 3 N 4 support.