Ordered mesoporous zirconium oxophosphate supported tungsten oxide solid acid catalysts: the improved Brønsted acidity for benzylation of anisole

Abstract

A series of WO₃ supported on ordered mesoporous zirconium oxophosphate (X wt% WO₃/M-ZrPO) solid acid catalysts with a WO₃ loading from 5 to 30 wt% were successfully synthesized, and their structure properties were characterized by X-ray diffraction (XRD), Raman spectroscopy, N₂-physisorption, transmission electron microscopy (TEM), UV-visible diffuse reflectance spectroscopy, Fourier transform infrared (FT-IR) spectroscopy, H₂ temperature-programmed reduction (H₂-TPR) and X-ray photoelectron spectroscopy (XPS). The catalytic performance of X wt% WO₃/M-ZrPO in liquid phase benzylation of anisole was studied and the relation between activity and states of tungsten species was investigated detailedly. The maximum catalytic activity was reached at a 20 wt% WO₃ loading, which possessed highly dispersed WO₃ species and the strongest Brønsted acidity. Meanwhile, the well dispersed WO₃ species strongly interacted with M-ZrPO, therefore, both sintering and leaching of WO₃ species were effectively restrained. Moreover, compared with the traditional zirconium phosphate synthesized from the sol–gel method (ZrP_sol–gel), the M-ZrPO with an abundant ordered mesostructure was propitious for improving the dispersion of WO₃ species and catalytic performance. In addition, the 20 wt% WO₃/M-ZrPO showed a markedly higher catalytic activity than H-ZSM5, H-Beta and 20 wt% WO₃/ZrP_sol–gel. Furthermore, the catalyst showed no discernible loss in activity or selectivity after five cycles.

---

1. Introduction

Supported tungsten oxide catalysts are a class of important heterogeneous catalysts and are widely utilized in many reactions, such as metathesis and isomerization of alkenes,^1–4 selective oxidation of hydrocarbon,^5–7 selective reduction of nitric oxide with ammonia,^8–11 hydrodesulfurization and hydrocracking of heavy fractions in petroleum chemistry.^12–15 For these applications, tungsten oxide-based catalysts have attracted vast attention in the field of solid acid catalysts.^16–18 Compared with traditional homogeneous liquid acid catalysts, novel heterogeneous solid acid catalysts should be encouraged due to the high reactivity, lack of corrosion, and being environmentally friendly, and easy to recover and reuse.^19,20 Owing to their excellent acidic property, stability and regenerability, the tungsten oxide-based catalysts, such as WO₃/Al₂O₃,^21,22 WO₃/TiO₂,²³ WO₃/SiO₂,²⁴ WO₃/ZrO₂,^25–27 tungsten oxide supported on metal sulfates and phosphates,^28–32 were widely employed as solid acid catalyst and exhibited superior catalytic performance in hydrolysis of cellulose, synthesis of biodiesel, Beckmann rearrangement of cyclohexanone oxime, isomerization of alkane and dehydration of alcohols. In addition, different influence factors, such as the dispersion, oxidation state, surface acidity and structure, largely influence the catalytic performance of tungsten based solid acid catalysts in these reactions.^6,26

Since the first mesoporous molecular sieve MCM-41 was successfully synthesized in 1992,^33,34 different kinds of mesoporous materials have been developed. Owing to the outstanding textural properties, mesoporous materials have been widely employed as catalysts and supports in the field of catalysis.^35,36 Compared with the traditional supports, mesoporous materials possess larger specific surface area, bigger pore volume and more ordered pore channels, which allow the active component highly dispersed on the support to provide more additional active sites for the reactant molecules resulting in improved catalytic performance.^37,38 Therefore, mesoporous materials, especially non-silica mesoporous materials like metal oxides and metal composite oxides, promise a series of potential carriers and have attracted growing interest in the field of catalysis. In addition, mesoporous materials supporting tungsten oxide have drawn considerable attention because of their diverse compositions that lead to potential applications in solid acid catalysis.^39,40 However, the reports about tungsten oxide supported on non-silica based ordered mesoporous materials are very few and have potential to be investigated.

In this paper, a series of WO₃ loaded on the ordered mesoporous zirconium oxophosphate (M-ZrPO) with different WO₃ loading was prepared through ultrasonic impregnation strategy and employed as solid acid catalyst in the Friedel–Crafts (FC) benzylation reactions. FC benzylation is one of the fundamental organic reactions for preparation of fine chemicals, dielectric fluids and pharmaceutical chemicals.⁴¹ Compared with traditional homogeneous catalysts, solid acid catalysts should be encouraged to utilize in the benzylation reactions.^42–44 To gain more insight into the possible relation between structure and catalytic properties, the present work systematically characterized X wt% WO₃/M-ZrPO by N₂-physisorption and TEM for the mesoporous properties of materials, XRD, Raman, UV-vis, FT-IR, H₂-TPR and XPS spectroscopy for the states of tungsten species in the materials as well as NH₃-TPD and pyridine-IR analyses for acidic property. In addition, the influences of tungsten species on Brønsted acidic property and catalytic activity of X wt% WO₃/M-ZrPO were discussed detailedly. Furthermore, the catalytic performance of 20 wt% WO₃/M-ZrPO was compared with conventional solid acids (H-Beta and H-ZSM5) and WO₃ loaded on the ZrP_sol–gel which was synthesized from sol–gel method.

2. Experimental section

2.1 Catalyst preparation

(EO)₁₀₆(PO)₇₀(EO)₁₀₆ triblock copolymer (Pluronic F127, M_av = 12 600, Sigma-Aldrich), zirconyl chloride octahydrate (ZrOCl₂·8H₂O, ≥99.0%, Tianjin Fengyue Reagent Company), trimethyl phosphate (PO(OCH₃)₃, ≥99.0%, Sinopharm Chemical Reagent Co. Ltd.), ammonium tungstate ((NH₄)₅H₅[H₂(WO₄)₆]·H₂O, ≥99.0%, Sinopharm Chemical Reagent Co. Ltd.), ethanol (≥99.7%, Sinopharm Chemical Reagent Co. Ltd.). All the reagents are A.R. grade and used as received without further purification.

Ordered mesoporous zirconium oxophosphate (M-ZrPO) was synthesized via improved one-pot evaporation-induced self-assemble (EISA) strategy as reported in previous report.⁴⁵ In a typical procedure of synthesizing M-ZrPO, 1.2 g of F127 (employed as structure-directing agents (SDAs)) were dissolved in 15 mL of anhydrous ethanol. As the SDAs were completely dissolved, 5 mmol of ZrOCl₂·8H₂O and 3.75 mmol of PO(OCH₃)₃ (the molar ratio of P/Zr was 0.75) were added into the above solution with vigorous stirring. The final solution was covered with PE film and stirred at least for 6 h. Finally, the transparent solution mixture was transferred to a Petri dish (d = 9 cm) to undergo the slow EISA process at 60 °C for 48 h, 100 °C for 24 h in a drying oven. The obtained colorless and transparent xerogel was calcined by slowly increasing temperature (1 °C min⁻¹ ramping rate) to 550 °C and kept at the final temperature for 5 h to remove the SDAs. The obtained template-free self-assembled mesoporous zirconium oxophosphate was denoted as M-ZrPO.

The amorphous zirconium phosphate synthesized from sol–gel method (ZrP_sol–gel) was as reported by Al-Qallaf et al.⁴⁶ H₃PO₄ and ZrOCl₂·8H₂O were employed as the source and the molar ratio of P/Zr (0.75) was kept consistent with the M-ZrPO.

The supported WO₃ solid acid catalysts were prepared via incipient wetness impregnation method assisted with 3 h ultrasound treatment with ammonium tungstate as the precursor of WO₃. After impregnation, the catalyst precursors were dried under irradiation of infrared lamp and then dried at 120 °C for 12 h. Finally, the catalyst precursors were calcined at 500 °C for 5 h. The obtained catalysts were denoted as X wt% WO₃/M-ZrPO and X wt% (X wt% = m_WO3/(m_WO3 + m_M-ZrPO) × 100%) stood for the WO₃ loading.

2.2 Characterization

Powder X-ray diffraction (XRD) measurements were performed using an X'Pert Pro Multipurpose diffractometer (PANalytical, Inc.) with Cu Kα radiation (0.15406 nm) at room temperature from 0.6° to 5.0° (small angle) and 10.0° to 80.0° (wide angle). Measurements were conducted using a voltage of 40 kV, current setting of 20 mA for small angle XRD (SXRD) and 40 mA for wide angle XRD (WXRD), step size of 0.02°, and count time of 4 s.

The Raman spectra were recorded on a Laboram 010 confocal Raman System (Horiba Jobin Yvon, France) equipped with 632 nm He–Ne laser.

The nitrogen adsorption and desorption isotherms at −196 °C were recorded on an Autosorb-iQ analyzer (Quantachrome Instruments U.S.). Prior to the tests, all the samples were pretreated at 200 °C for 2 h. The specific surface areas were calculated via the Brunauer–Emmett–Teller (BET) method in the relative pressure range of 0.05–0.3; the single-point pore volume was calculated from the adsorption isotherm at a relative pressure of 0.990; pore size distributions were calculated using adsorption branches of nitrogen adsorption–desorption isotherms by Barrett–Joyner–Halenda (BJH) method. The tungsten surface densities expressed as the number of W atoms per nanometer square area (W atoms nm⁻²) and were calculated using the equation: surface density of W = {[WO₃ loading (wt%)/100] × 6.023 × 10²³}/{[231.8 (formula weight of WO₃) × S_BET (m² g⁻¹) × 10¹⁸]}.⁴⁷

Transmission electron microscopy (TEM) images and energy-dispersive X-ray spectroscopy (EDX) measurements were performed on the TECNAI G² F20 high-resolution transmission electron microscopy under a working voltage of 200 kV.

UV-visible diffuse reflectance spectra were recorded on a PE Lambda 650S in the range of 190–800 nm.

Fourier transform infrared (FT-IR) spectra were recorded on KBr pellets by a Nexus 870 infrared spectrometer with the wave number from 4000 to 400 cm⁻¹.

X-ray photoelectron spectroscopy (XPS) analysis of the samples was performed on a Thermon Scientific ESCALAB250xi spectrometer. All binding energies were calibrated to the C_1s line (284.8 eV).

H₂ temperature-programmed reduction (H₂-TPR) measurements were performed on an AMI-100 unit (Zeton-Altamira instrument) employing H₂ gas as reducing agent. The samples (0.1 g) were loaded in a U-shaped quartz reactor. Prior to TPR measurements, the samples were pretreated at 200 °C for 1 h in flowing He gas (50 mL min⁻¹) to remove any moisture and other adsorbed impurities. After cooling the reactor to room temperature, a 5% H₂–He gas (50 mL min⁻¹) mixture was introduced. The catalyst was heated to 1100 °C at a rate of 20 °C min⁻¹ and the H₂ gas consumption was measured using an AMETEK (LC-D-200 Dycor AMETEK) mass spectrometer.

Temperature programmed desorption of ammonia (NH₃-TPD) was performed on the same instrument as reported in H₂-TPR. A typical experiment for the TPD measurement was as follows: the sample (0.1 g) was pretreated at 200 °C for 1 h in the flowing He gas to remove any moisture and absorbed impurities. After cooling to 100 °C, 10 mol% NH₃–He gas (50 mL min⁻¹) was introduced for 30 min. Physically absorbed NH₃ was then removed completely by desorbing in He gas at 100 °C (observed from the mass spectrometer signal). The TPD experiment was then carried out by raising the temperature of sample in a programmed manner (10 °C min⁻¹) to 700 °C.

The infrared spectra of adsorbed pyridine (pyridine-IR) were recorded on PE Frontier FT-IR spectrometer. The sample (∼15 mg) was pressed into a pellet (∼13 mm), activated at 400 °C for 1 h under vacuum, then cooled to room temperature and pyridine was introduced and adsorbed for 30 min. The sample was raised to a desired temperature and held for 30 min under vacuum, after which the spectra was recorded. The blank experiments were operated under the same conditions and used as the background to insure the accuracy of the infrared spectra of adsorbed pyridine. The Brønsted and Lewis acidity was determined by the method proposed by Emeis.⁴⁸

2.3 Catalytic reaction

Liquid phase Friedel–Crafts (FC) benzylation of anisole with benzyl alcohol was performed in a three-necked round bottom flask coupled with a reflux condenser in a temperature controlled oil bath as reported by Rao et al.³⁵ In a typical run, 0.2 g of X wt% WO₃/M-ZrPO was added to a mixture of anisole (20 mL) and benzyl alcohol (2 mL) with dodecane (0.4 mL) used as an internal standard for gas chromatography (GC) analysis. Nitrogen was introduced into the flask through one of the gas inlets. The reaction was carried out at 170 °C under reflux condition and with vigorous stirring for 2 h. The conversion of benzyl alcohol and selectivity of products were monitored by a gas chromatography (GC) instrument (Agilent-7890A; equipped with a flame ionization detector (FID) and HP-5 column (30 m × 0.32 mm × 0.25 μm)). Gas chromatography-mass spectroscopy (GC-MS) instrument (5975c vl MSD with triple-axis detector, GC Agilent-7890A) was employed to identify the reaction products. After completion of the reaction, the catalyst was separated by centrifugation and activated at 400 °C for 3 h for regeneration.

3. Results and discussion

3.1 Structure characterization of support M-ZrPO material

As our previous report, the M-ZrPO possessed ordered mesostructure.⁴⁵ Shortly, as shown in Fig. 1(1), obvious diffraction peaks in SXRD pattern could be clearly found, implying the existence of ordered mesopores. In addition, there were only two broad peaks in WXRD pattern, showing the pore wall structure was amorphous. Moreover, the typical H1-type hysteresis loop, narrow pore size distribution (Fig. 1(2)) and TEM images (Fig. 2) further confirmed the existence of ordered pores in M-ZrPO.

Fig. 1 Small-angle X-ray diffraction pattern (1) and wide-angle X-ray diffraction pattern (inset of (1)), isotherms (2) and pore size distribution (inset of (2)) of the support M-ZrPO.

Fig. 2 TEM images of the M-ZrPO.

3.2 Characterization of the as-prepared X wt% WO₃/M-ZrPO

3.2.1 XRD and Raman analysis. The WXRD patterns of X wt% WO₃/M-ZrPO solid acid catalysts are shown in Fig. 3. As shown in Fig. 3(1), the patterns exhibited similar profile to M-ZrPO, only presenting two broad peaks in the WXRD patterns as WO₃ loading was under 20 wt%. The absence of typical peaks of crystalline WO₃ implied that tungsten species existed as highly dispersed state on the M-ZrPO surface in these samples. However, the weak peaks at 22–25° (23.1°, 23.5°, 24.3° shown in Fig. 3(2)), 34°, 51° and 56°, which could be indexed to monoclinic WO₃ (JCPDS card no. 89-4476), began to appear on 25 wt% WO₃/M-ZrPO sample, suggesting that some of tungsten species congregated on the surface to form crystalline WO₃. Further increase WO₃ loading to 30 wt%, much enhanced intensity of diffraction peaks from crystalline WO₃ could be observed, illustrating that more tungsten species agglomerated and formed crystalline WO₃. Moreover, the color of X wt% WO₃/M-ZrPO changed from white to slight yellow at high WO₃ loading (≥25 wt%), which was a qualitative indication for the formation of crystalline WO₃.⁴⁹ However, the crystallite size of WO₃ was still difficult to be calculated by Scherrer equation due to the peak broadening in the WXRD patterns. This might imply that no obvious crystalline WO₃ particles existed in 25 wt% WO₃/M-ZrPO and 30 wt% WO₃/M-ZrPO.

Fig. 3 Wide-angle X-ray diffraction patterns of X wt% WO₃/M-ZrPO: (a) 5 wt% WO₃/M-ZrPO, (b) 10 wt% WO₃/M-ZrPO, (c) 15 wt% WO₃/M-ZrPO, (d) 20 wt% WO₃/M-ZrPO, (e) 25 wt% WO₃/M-ZrPO, (f) 30 wt% WO₃/M-ZrPO.

Moreover, Raman spectra were employed to further identify the major structural information concerning surface tungsten oxide species. As shown in Fig. 4, the bulk WO₃ showed obvious bands at 725 and 815 cm⁻¹, which were assigned to the W=O bending and stretching modes in the crystalline WO₃ species.^26,50 However, with WO₃ loading between 5 and 20 wt%, there were no distinct bands in the spectra, implying that no crystalline WO₃ existed in these samples. Further increasing WO₃ loading to 25 and 30 wt%, weak Raman bands could be observed at 725 and 815 cm⁻¹. This might be due to the agglomeration of microcrystalline WO₃ on the surface of the 25 wt% WO₃/M-ZrPO and 30 wt% WO₃/M-ZrPO. However, these bands were quite weak, implying the WO₃ species with low crystallinity existed in the two samples. Therefore, Raman results further confirmed what were previously observed by WXRD. Consequently, in the 5–20 wt% WO₃/M-ZrPO samples, the tungsten species were highly dispersed on the support, whereas, the content surpassed threshold of mono-dispersed state and crystalline WO₃ began to appear on the surface of M-ZrPO as WO₃ loading reached 25 and 30 wt%.⁵¹

Fig. 4 Raman spectra of the X wt% WO₃/M-ZrPO and bulk WO₃: (a) M-ZrPO, (b) 5 wt% WO₃/M-ZrPO, (c) 10 wt% WO₃/M-ZrPO, (d) 15 wt% WO₃/M-ZrPO, (e) 20 wt% WO₃/M-ZrPO, (f) 25 wt% WO₃/M-ZrPO, (g) 30 wt% WO₃/M-ZrPO, (h) bulk WO₃.

3.2.2 Nitrogen adsorption–desorption analysis. The nitrogen adsorption–desorption isotherms as well as pore size distributions of X wt% WO₃/M-ZrPO samples are displayed in Fig. 5. As shown in Fig. 5(1), all samples presented typical IV type isotherm with H1 shaped hysteresis loops, proving the successful preservation of mesostructure. Besides, all catalysts performed extremely narrow pore size distribution around 5.6 nm. In contrast with the M-ZrPO support, their average pore diameter did not change largely, indicating that the loaded WO₃ did not cause significant plugging of mesoporous pores even in 25 wt% WO₃/M-ZrPO and 30 wt% WO₃/M-ZrPO. Furthermore, the specific surface area, porosity and pore diameter of X wt% WO₃/M-ZrPO are summarized in Table 1. It could be observed that the specific surface area and pore volume decreased gradually with the increasing of WO₃ loading. This might be mainly due to that the loaded WO₃ species contributed little to the specific surface area and pore volume. Generally, all X wt% WO₃/M-ZrPO still preserved obvious mesostructure with large specific surface area, big pore volume and narrow pore size distribution even after a second high temperature calcination process, fully demonstrating the superior thermal stability of M-ZrPO. In addition, the W atom density (atom nm⁻²) calculated from WO₃ loading and specific surface area is shown in Table 1. Combined with the WXRD and Raman analyses, it was concluded that the WO₃ species were highly dispersed on the M-ZrPO when the W atom density was below 4.30 atom nm⁻² and crystalline WO₃ began to form at the W atom density of 5.95 atom nm⁻². However, for 20 wt% WO₃/ZrP_sol–gel, the specific surface area only showed 24.9 m² g⁻¹ and the W atom density reached 20.9 atom nm⁻². The high W atom density led to the formation of crystalline WO₃ in the surface of ZrP_sol–gel and this conclusion also could be proved by WXRD pattern (as shown in Fig. S1†). Therefore, M-ZrPO with high specific surface area was helpful for improving the dispersion of WO₃ species.

Fig. 5 Isotherms (1) and pore size distributions (2) of the X wt% WO₃/M-ZrPO: (a) 5 wt% WO₃/M-ZrPO, (b) 10 wt% WO₃/M-ZrPO, (c) 15 wt% WO₃/M-ZrPO, (d) 20 wt% WO₃/M-ZrPO, (e) 25 wt% WO₃/M-ZrPO, (f) 30 wt% WO₃/M-ZrPO.

Table 1 Textural properties of the X wt% WO₃/M-ZrPO, 20 wt% WO₃/ZrP_sol–gel and 20 wt% WO₃/M-ZrPO-used derived from nitrogen adsorption and desorption data

Sample	WO3 loading	Specific surface area (m^2 g^−1)	Pore size (nm)	Pore volume (cm^3 g^−1)	W atom density (atom nm^−2)
M-ZrPO	0	165.5	5.63	0.25	0
5 wt% WO3/M-ZrPO	5	157.2	5.61	0.21	0.83
10 wt% WO3/M-ZrPO	10	151.1	5.63	0.21	1.72
15 wt% WO3/M-ZrPO	15	128.8	5.62	0.19	3.03
20 wt% WO3/M-ZrPO	20	120.8	5.64	0.18	4.30
25 wt% WO3/M-ZrPO	25	109.1	5.65	0.17	5.95
30 wt% WO3/M-ZrPO	30	96.5	5.70	0.15	8.08
20 wt% WO3/ZrPsol–gel	20	24.9	—	—	20.9
20 wt% WO3/M-ZrPO-used	20	110.5	5.62	0.17	4.70

---

3.2.3 TEM analysis. The morphology analysis of X wt% WO₃/M-ZrPO catalysts was characterized by TEM. The TEM images of all X wt% WO₃/M-ZrPO samples are shown in Fig. 6. Compared with the M-ZrPO carrier, the obvious mesopores also could be observed, indicating the mesostructure still existed in X wt% WO₃/M-ZrPO samples. This agreed quite well with the characterization of N₂-physisorption. Besides, it was interesting that no apparent WO₃ particles were observed in 25 wt% WO₃/M-ZrPO and 30 wt% WO₃/M-ZrPO, which had weak diffraction peaks of crystalline WO₃ in the WXRD patterns. This might be due to that the diffraction peaks were weak and broad and the WO₃ existed as thin sheets with low crystallinity as reported by Yang et al.⁵² Moreover, as shown in Fig. 6g, the exclusive Zr, P and W peaks could be clearly observed in the EDX pattern of 20 wt% WO₃/M-ZrPO, illustrating that the W species were successfully introduced into the materials.

Fig. 6 TEM images of the X wt% WO₃/M-ZrPO: (a) 5 wt% WO₃/M-ZrPO, (b) 10 wt% WO₃/M-ZrPO, (c) 15 wt% WO₃/M-ZrPO, (d) 20 wt% WO₃/M-ZrPO, (e) 25 wt% WO₃/M-ZrPO, (f) 30 wt% WO₃/M-ZrPO, (g) EDX measurement of 20 wt% WO₃/M-ZrPO. (The scale in images (a–f) was 50 nm).

3.2.4 UV-vis spectra analysis. The UV-vis spectra were utilized to distinguish the states of tungsten species in X wt% WO₃/M-ZrPO and shown in Fig. 7.^6,53 The support M-ZrPO showed an obvious absorption peak at 205 nm, which was attributed to the Zr–O–P coordination. An additional weak peak at about 300 nm could be ascribed to Zr(IV) interacting with the phosphate counter anions in the framework. As the introduction of WO₃, a new peak at 260 nm, which could be assigned to the low-condensed oligomeric tungsten species, could be discovered and became more observable with the increasing of WO₃ loading.^6,29 For the samples with WO₃ loading below 20 wt%, the absence of peaks at 400 nm indicated that no crystalline WO₃ existed in these samples and the tungsten species were highly dispersed on the surface of M-ZrPO. However, the weak band at 400 nm began to be observed for 25 wt% WO₃/M-ZrPO, implying that the crystalline WO₃ appeared in the sample. Further increasing the WO₃ content to 30 wt%, the intensity of band at 400 nm became stronger due to the increasing quantity of crystalline WO₃ species on the surface. In short, the tungsten species existed as low-condensed oligomeric states in the samples with WO₃ loading below 20 wt% and crystalline WO₃ appeared in the 25 wt% WO₃/M-ZrPO and 30 wt% WO₃/M-ZrPO. The same conclusion also could be gotten from the WXRD and Raman characterizations.

Fig. 7 UV-vis diffuse reflectance spectra of the X wt% WO₃/M-ZrPO: (a) M-ZrPO, (b) 5 wt% WO₃/M-ZrPO, (c) 10 wt% WO₃/M-ZrPO, (d) 15 wt% WO₃/M-ZrPO, (e) 20 wt% WO₃/M-ZrPO, (f) 25 wt% WO₃/M-ZrPO, (g) 30 wt% WO₃/M-ZrPO.

3.2.5 FT-IR spectra analysis. The FT-IR spectra of X wt% WO₃/M-ZrPO are shown in Fig. 8. All the patterns of X wt% WO₃/M-ZrPO displayed obvious peaks at 3450 and 1630 cm⁻¹, which were assigned to the surface hydroxyl groups. The peak at 1070 cm⁻¹ was arised from the Zr–O–P network. Besides, a group of new peaks at 780, 630 and 470 cm⁻¹, which were characteristic of the stretching vibration mode of W–O–W,³² appeared and were very weak in 5–20 wt% WO₃/M-ZrPO samples, in which the WO₃ species existed as low-condensed oligomeric states. Further increasing WO₃ loading content to 25 and 30 wt%, these peaks became more recognizable. We deduced this might be explained by the appearance of crystalline WO₃ in 25 wt% WO₃/M-ZrPO and 30 wt% WO₃/M-ZrPO. These conclusions agreed quite well with the WXRD, Raman and UV-vis characterizations.

Fig. 8 FT-IR spectra of the X wt% WO₃/M-ZrPO: (a) M-ZrPO, (b) 5 wt% WO₃/M-ZrPO, (c) 10 wt% WO₃/M-ZrPO, (d) 15 wt% WO₃/M-ZrPO, (e) 20 wt% WO₃/M-ZrPO, (f) 25 wt% WO₃/M-ZrPO, (g) 30 wt% WO₃/M-ZrPO.

3.2.6 H₂-TPR analysis. The H₂-TPR profiles of X wt% WO₃/M-ZrPO and bulk WO₃ are shown in Fig. 9. The M-ZrPO showed no discernible reduction peaks at 550–1100 °C, whereas, the bulk WO₃ showed three main peaks at 790, 930 and 1000 °C. These three peaks could be assigned to the three-stepwise reduction of WO₃ to W (WO₃(VI) → W₂₀O₅₈(VI,V) → WO₂(IV) → W(0)).^54,55 However, the reduction of X wt% WO₃/M-ZrPO was strongly depended on the WO₃ loading. As reported in the literature, the reducibility of tungsten-based catalysts increased as the strength of interaction between metal oxide species and support decreased.^5,50,56 With the increasing of WO₃ loading from 5 to 30 wt%, a shift of reduction peaks to lower temperature was apparently observed, indicating that the interaction between WO₃ and support became weaker with the increasing of WO₃ loading. However, the reduction temperature was still much higher than bulk WO₃, illustrating the existence of intense interaction between WO₃ species and support. The strong interaction was in favor of reducing the sintering and leaching of tungsten species in the calcination and reaction process.⁶

Fig. 9 The H₂-TPR profiles of the X wt% WO₃/M-ZrPO: (a) M-ZrPO, (b) 5 wt% WO₃/M-ZrPO, (c) 10 wt% WO₃/M-ZrPO, (d) 15 wt% WO₃/M-ZrPO, (e) 20 wt% WO₃/M-ZrPO, (f) 25 wt% WO₃/M-ZrPO, (g) 30 wt% WO₃/M-ZrPO, (h) bulk WO₃.

3.2.7 XPS analysis. The high-resolution XPS analysis was employed to investigate the surface chemical composition as well as the oxidation state of tungsten species in X wt% WO₃/M-ZrPO. The W 4f XPS profiles of X wt% WO₃/M-ZrPO with different WO₃ loading are shown in Fig. 10(1). All XPS profiles in the W 4f region were similar to each other in shape and the intensity of W 4f region increased monotonously with the WO₃ loading. The core level spectra displayed two peaks corresponding to W 4f_7/2 and W 4f_5/2 with binding energies of 36.2 and 38.3 eV, which were related to the W(VI) oxidation state.⁵⁷ Therefore, it could be deduced that the tungsten species mainly existed as high oxidation state W(VI) in the materials. The high oxidation state W(VI) might account for the improvement of Brønsted acidic property with the increasing of tungsten content in X wt% WO₃/M-ZrPO.

Fig. 10 (1) High-resolution XPS spectra of the X wt% WO₃/M-ZrPO: (a) 5 wt% WO₃/M-ZrPO, (b) 10 wt% WO₃/M-ZrPO, (c) 15 wt% WO₃/M-ZrPO, (d) 20 wt% WO₃/M-ZrPO, (e) 25 wt% WO₃/M-ZrPO, (f) 30 wt% WO₃/M-ZrPO; (2) the relation between WO₃ loading and WO₃ content calculated from XPS spectra.

In addition, the correlation between surface WO₃ contents of gotten products calculated from XPS profiles and theoretical WO₃ contents as loaded in the preparation process is shown in Fig. 10(2). According to the comparison of surface WO₃ contents with theoretical values, it could be concluded that the WO₃ contents almost increased linearly with increasing of WO₃ loading. This might suggest that WO₃ species were dispersed uniformly on the surface of M-ZrPO even in the samples with high WO₃ loading (25 and 30 wt%).

3.2.8 Acidic property. The acidic property of X wt% WO₃/M-ZrPO was evaluated by NH₃-TPD. All the TPD profiles of X wt% WO₃/M-ZrPO are shown in Fig. 11. The characteristic TPD profile showed an obvious broad peak at 150–400 °C. This indicated that abundant acid sites with different strength existed in X wt% WO₃/M-ZrPO. Furthermore, the desorption peaks of NH₃ at 200 °C became larger with the introduction of WO₃ and reached the maximum at 20 wt% WO₃ loading, indicating the acidity was improved with the introduction of tungsten species. Further increasing WO₃ loading to 25 and 30 wt%, the desorption peaks decreased, showing that the acidity began to decrease with the appearance of crystalline WO₃ in 25 wt% WO₃/M-ZrPO and 30 wt% WO₃/M-ZrPO. In addition, the peak at 320 °C which might be the Lewis acidity caused by Zr(IV), became weak with the increasing of tungsten loading. This might be due to the loaded WO₃ species covered on the surface of M-ZrPO and reduced the Lewis acid sites caused by the Zr(IV).

Fig. 11 NH₃-TPD profiles of the X wt% WO₃/M-ZrPO: (a) M-ZrPO, (b) 5 wt% WO₃/M-ZrPO, (c) 10 wt% WO₃/M-ZrPO, (d) 15 wt% WO₃/M-ZrPO, (e) 20 wt% WO₃/M-ZrPO, (f) 25 wt% WO₃/M-ZrPO, (g) 30 wt% WO₃/M-ZrPO.

The pyridine-IR spectra recorded after adsorption of pyridine at room temperature and outgassed at 150 °C are shown in Fig. 12(1). The spectra showed bands at 1450, 1575 and 1610 cm⁻¹, ascribed to pyridine coordinately bonded to the Lewis acid sites. The samples showed bands at 1540 and 1640 cm⁻¹, which were due to protonated pyridine bonded to the Brønsted acid sites. The bands at 1490 cm⁻¹ were a combination between two separate bands at 1540 and 1450 cm⁻¹, indicating that both Brønsted and Lewis acid sites existed in the samples. When WO₃ loading increased from 0 to 20 wt%, the concentration of the Brønsted acid sites (bands at 1540 cm⁻¹) increased progressively. This might be owing to the increasing amounts of highly dispersed tungsten species on the surface of M-ZrPO. However, the amounts of Brønsted acid sites began to slightly decrease in 25 wt% WO₃/M-ZrPO and 30 wt% WO₃/M-ZrPO, possibly due to that the high content of WO₃ led to appearance of crystalline WO₃ in the two samples. Moreover, the spectra of X wt% WO₃/M-ZrPO outgassed at 250 °C are shown in Fig. 12(2). In the spectra, the band at 1540 cm⁻¹ disappeared in the sample without introduction of WO₃. However, the spectra of 20 wt% WO₃/M-ZrPO still showed observable band at 1540 cm⁻¹, illustrating that Brønsted acidity caused by the loaded tungsten species in 20 wt% WO₃/M-ZrPO was rather stronger than the Brønsted acidity in M-ZrPO. The Brønsted acid sites in 20 wt% WO₃/M-ZrPO gradually vanished with the desorption temperature reached 300 and 400 °C (as shown in Fig. S2†). Moreover, the Brønsted and Lewis acidity gotten from pyridine-IR spectra outgassed at 150 °C is shown in Table 2. The Brønsted acidity increased gradually and reached the maximum at 20 wt% WO₃ loading for 43.4 μmol g⁻¹. Further increasing WO₃ loading to 25 and 30 wt%, the Brønsted acidity began to decrease gradually. We deduced that the WO₃ dispersion state on surface of support could largely affect the Brønsted acidic properties. When the tungsten species were highly dispersed on support (WO₃ loading ≤20 wt%), the surface Brønsted acidity increased with the increasing of WO₃ loading. The loaded WO₃ species could be probably responsible for enhancing Brønsted acidity. Beyond the monolayer coverage, due to agglomeration of tungsten species and formation of crystalline WO₃ (as proved from XRD, Raman and UV-vis analyses), gradual loss of Brønsted acidity was observed in 25 wt% WO₃/M-ZrPO and 30 wt% WO₃/M-ZrPO (as shown in Scheme 1).⁴⁹ The strong Brønsted acid sites were essential to the benzylation of anisole with benzyl alcohol.

Fig. 12 IR spectra for pyridine adsorbed on the X wt% WO₃/M-ZrPO recorded at 150 °C (1) and 250 °C (2): (a) M-ZrPO, (b) 5 wt% WO₃/M-ZrPO, (c) 10 wt% WO₃/M-ZrPO, (d) 15 wt% WO₃/M-ZrPO, (e) 20 wt% WO₃/M-ZrPO, (f) 25 wt% WO₃/M-ZrPO, (g) 30 wt% WO₃/M-ZrPO.

Table 2 The acidity of X wt% WO₃/M-ZrPO gotten from pyridine-IR outgassed at 150 °C

Samples	Brønsted acidity (μmol g^−1)	Lewis acidity (μmol g^−1)	Total acidity (μmol g^−1)
M-ZrPO	20.0	83.0	103.0
5 wt% WO3/M-ZrPO	33.9	80.0	113.9
10 wt% WO3/M-ZrPO	40.2	88.5	128.7
15 wt% WO3/M-ZrPO	42.2	77.4	119.6
20 wt% WO3/M-ZrPO	43.4	84.3	127.7
25 wt% WO3/M-ZrPO	30.0	66.5	96.5
30 wt% WO3/M-ZrPO	25.0	63.0	88.0
20 wt% WO3/ZrPsol–gel	9.70	41.3	51.0

---

Scheme 1 The WO₃ state on the X wt% WO₃/M-ZrPO with increasing of WO₃ loading.

3.3 Catalytic performance

The catalytic performance of X wt% WO₃/M-ZrPO was investigated in liquid phase benzylation of anisole with benzyl alcohol. All the reaction schemes are presented in Scheme 2. In this reaction system, benzylanisoles are the products of benzylation of anisole with benzyl alcohol and dibenzyl ether is a product of the auto-etherification of benzyl alcohol. Meanwhile, dibenzyl ether also can react with anisole to produce benzylanisole.

Scheme 2 Friedel–Crafts benzylation reaction between anisole and benzyl alcohol.

To investigate the influence of WO₃ loading on catalytic activity, the catalytic performance of X wt% WO₃/M-ZrPO with different WO₃ loading was investigated in benzylation reaction and the results are shown in Fig. 13(1). As shown in Fig. 13(1), the sample without supporting WO₃ showed poor catalytic activity for benzylation reaction, and only showed 18.0% conversion of benzyl alcohol with 54.5% selectivity of benzylanisoles. However, with introduction of tungsten species, both the conversion and selectivity were largely improved and reached 76.3% and 74.8% at 5 wt% WO₃ loading. Further increase WO₃ loading, the conversion and selectivity were stepwise improved and reached the maximum at 20 wt% WO₃ loading with the 100% conversion of benzyl alcohol and 91.0% selectivity of benzylanisole. Therefore, the loaded WO₃ species successfully improved the catalytic activity of X wt% WO₃/M-ZrPO in benzylation reaction. In the literature, there is a general agreement that the benzylation of anisole with benzyl alcohol is catalyzed on the Brønsted acid sites.^58,59 Therefore, the improved catalytic activity might be due to the increasing of Brønsted acid sites with increasing of WO₃ loading. However, for 25 wt% WO₃/M-ZrPO and 30 wt% WO₃/M-ZrPO, the catalytic activity began to fall gradually. This might be on account of the active sites were covered by crystalline WO₃ and the number of Brønsted acidic sites decreased for excess WO₃ loading. Based on the characterization results presented above, it could be concluded that the proper content of tungsten oxide species and its highly dispersed state have a significant effect on the catalytic performance of X wt% WO₃/M-ZrPO solid acid catalysts. The samples with highly dispersed tungsten oxide species showed high conversion and selectivity for the benzylation reaction, whereas the appearance of WO₃ crystallites decreased the catalytic performance of X wt% WO₃/M-ZrPO.

Fig. 13 Friedel–Crafts benzylation reaction: (1) catalyzed by X wt% WO₃/M-ZrPO with different WO₃ loading; (2) catalyzed by 20 wt% WO₃/M-ZrPO at different times (BA: benzylanisoles).

The changing progress of benzyl alcohol conversion and benzylanisole yield was studied by analyzing the reaction products at different time intervals (Fig. 13(2)). Initially, the conversion of benzyl alcohol and yield of benzylanisole increased with time and the maximum benzyl alcohol conversion (100%) and benzylanisole yield (91.0%) were achieved at 120 min for 20 wt% WO₃/M-ZrPO. After 120 min, there were little changes in the curves. Moreover, when solid acid catalyst was employed, a crucial issue was possibility that the active sites might migrate from solid support to liquid phase and these leached species might become responsible for a significant part of the catalytic activity.^42,60 In order to determine if the tungsten species leaching from support, we investigated the yield of benzylanisoles with and without filtering 20 wt% WO₃/M-ZrPO solid acid catalyst. As shown in Fig. 13(2), the liquid phase was separated from the reaction system at 40 min. No further reaction was observed after filtering the solid acid catalyst, proving that the benzylation reaction was only being possible in the presence of 20 wt% WO₃/M-ZrPO solid acid catalyst and 20 wt% WO₃/M-ZrPO was a real heterogeneous catalyst. In addition, the possible leaching of WO₃ species into the reaction mixture was determined by the atomic adsorption spectroscopy (AAS) analysis. No detectable tungsten species could be found in the mixture, indicating that no obvious tungsten species leached to the mixture in the reaction process.

Moreover, in solid acid catalyzed reactions, both high activity and reusability are significant in consideration of the separation of products and recycling of the catalyst. In order to check the recyclability of the catalyst, we carried out five runs test over 20 wt% WO₃/M-ZrPO (shown in Fig. 14(1)). Compared with the fresh catalyst, no significant declines in conversion and selectivity were observed after five cycles, indicating that 20 wt% WO₃/M-ZrPO could be reused as solid acid catalyst in the benzylation reaction. Moreover, to investigate the stability and persistence of active tungsten species, 20 wt% WO₃/M-ZrPO catalyst recovered after five cycles was characterized by XPS analysis and the results showed that surface WO₃ content (20.6 wt%) changed little compared with fresh sample (20.1 wt%). All these results might be due to the presence of strong interaction between active tungsten species and M-ZrPO and the leaching of tungsten species were largely avoided in the reaction process. In addition, the catalyst regenerated after five cycles was characterized by N₂-physisorption and TEM. As shown in Fig. S3 and S4,† the mesostructure of 20 wt% WO₃/M-ZrPO still existed and the textural properties (shown in Table 1) of used catalyst changed little, indicating that the catalyst suffered little damage in the acid-catalyzed reaction. Moreover, Fig. 14(2) showed the activity of 20 wt% WO₃/M-ZrPO and control catalysts (H-Beta, H-ZSM5 and 20 wt% WO₃/ZrP_sol–gel) in the benzylation reaction. Compared with the control catalysts, 20 wt% WO₃/M-ZrPO showed a markedly higher catalytic activity. It was noteworthy that 20 wt% WO₃/ZrP_sol–gel only showed 41.7% conversion of benzyl alcohol with 76.9% selectivity of benzylanisoles. This might be due to that the poor dispersion of tungsten species on the ZrP_sol–gel and crystalline WO₃ appeared at a 20 wt% loading leaded to weak Brønsted acidity (as shown in Table 2). Therefore, the mesoporous structure was propitious for improving the dispersion of tungsten species on support to provide more additional active sites for the reactant molecules resulting in improved catalytic performance. Therefore, 20 wt% WO₃/M-ZrPO was an ideal solid acid catalyst in benzylation reaction due to the excellent catalytic activity and recyclability.

Fig. 14 Friedel–Crafts benzylation reaction: (1) the recyclability of 20 wt% WO₃/M-ZrPO; (2) catalyzed by different solid acid catalysts.

4. Conclusion

A series of WO₃ supported on ordered mesoporous zirconium oxophosphate (X wt% WO₃/M-ZrPO) solid acid catalysts with different WO₃ loading from 5 to 30 wt% were successfully synthesized and showed excellent catalytic performance in benzylation of anisole. Moreover, compared with traditional zirconium phosphate synthesized from sol–gel method (ZrP_sol–gel), the M-ZrPO with abundant ordered mesostructure was a better candidate to obtain highly dispersed tungsten species in the samples. The state of WO₃ largely influenced the catalytic performance of X wt% WO₃/M-ZrPO. When WO₃ species existed as highly dispersed state (≤20 wt%), the Brønsted acidic properties and catalytic performance was improved with the increasing of WO₃ loading. However, the Brønsted acidic properties and catalytic performance began to decrease with the formation of crystalline WO₃ (≥25 wt%). Moreover, due to the strong interaction between WO₃ and M-ZrPO, both sintering and leaching of tungsten species could be greatly restrained. Therefore, the X wt% WO₃/M-ZrPO catalysts showed excellent stability and reusability in the benzylation of anisole.

Acknowledgements

The authors sincerely acknowledge the financial support from the National Basic Research Program of PR China (no. 2011CB201404) and the National Natural Science Foundation of China (no. 21133011).

Notes and references

1. M. Hino and K. Arata, J. Chem. Soc., Chem. Commun., 1988, 1259–1260.
2. R. D. Wilson, D. G. Barton, C. D. Baertsch and E. Iglesia, J. Catal., 2000, 194, 175–187.
3. Y. Rezgui and M. Guemini, Appl. Catal., A, 2005, 282, 45–53.
4. Y. Rezgui and M. Guemini, Energy Fuels, 2007, 21, 602–609.
5. A. de Lucas, J. L. Valverde, P. Cañizares and L. Rodriguez, Appl. Catal., A, 1998, 172, 165–176.
6. X. L. Yang, W. L. Dai, R. Gao and K. Fan, J. Catal., 2007, 249, 278–288.
7. C. Hammond, J. Straus, M. Righettoni, S. E. Pratsinis and I. Hermans, ACS Catal., 2013, 3, 321–327.
8. H. Kamata, K. Takahashi and C. U. Ingemar Odenbrand, J. Catal., 1999, 185, 106–113.
9. W. Shan, F. Liu, H. He, X. Shi and C. Zhang, Appl. Catal., B, 2012, 115–116, 100–106.
10. F. Can, S. Berland, S. Royer, X. Courtois and D. Duprez, ACS Catal., 2013, 3, 1120–1132.
11. C. Liu, L. Chen, H. Chang, L. Ma, Y. Peng, H. Arandiyan and J. Li, Catal. Commun., 2013, 40, 145–148.
12. T. Kabe, W. Qian, A. Funato, Y. Okoshi and A. Ishihara, Phys. Chem. Chem. Phys., 1999, 1, 921–927.
13. A. Wang, X. Li, Y. Chen, D. Han, Y. Wang, Y. Hu and T. Kabe, Chem. Lett., 2001, 30, 474–475.
14. L. Coulier, G. Kishan, J. A. R. van Veen and J. W. Niemantsverdriet, J. Phys. Chem. B, 2002, 106, 5897–5906.
15. S. D. Kelly, N. Yang, G. E. Mickelson, N. Greenlay, E. Karapetrova, W. Sinkler and S. R. Bare, J. Catal., 2009, 263, 16–33.
16. D. Barton, S. Soled and E. Iglesia, Top. Catal., 1998, 6, 87–99.
17. W. Zhou, E. I. Ross-Medgaarden, W. V. Knowles, M. S. Wong, I. E. Wachs and C. J. Kiely, Nat. Chem., 2009, 1, 722–728.
18. N. Soultanidis, W. Zhou, A. C. Psarras, A. J. Gonzalez, E. F. Iliopoulou, C. J. Kiely, I. E. Wachs and M. S. Wong, J. Am. Chem. Soc., 2010, 132, 13462–13471.
19. N. Mizuno and M. Misono, Chem. Rev., 1998, 98, 199–218.
20. K. Tanabe and W. F. Hölderich, Appl. Catal., A, 1999, 181, 399–434.
21. V. M. Benitez and N. S. Fígoli, Catal. Commun., 2002, 3, 487–492.
22. X. Chen, G. Clet, K. Thomas and M. Houalla, J. Catal., 2010, 273, 236–244.
23. G. Lu, X. Li, Z. Qu, Q. Zhao, H. Li, Y. Shen and G. Chen, Chem. Eng. J., 2010, 159, 242–246.
24. J. E. Herrera, J. H. Kwak, J. Z. Hu, Y. Wang, C. H. F. Peden, J. Macht and E. Iglesia, J. Catal., 2006, 239, 200–211.
25. C. D. Baertsch, K. T. Komala, Y.-H. Chua and E. Iglesia, J. Catal., 2002, 205, 44–57.
26. N. R. Shiju, M. AnilKumar, W. F. Hoelderich and D. R. Brown, J. Phys. Chem. C, 2009, 113, 7735–7742.
27. R. Kourieh, S. Bennici, M. Marzo, A. Gervasini and A. Auroux, Catal. Commun., 2012, 19, 119–126.
28. S. Furuta, H. Matsuhashi and K. Arata, Catal. Commun., 2004, 5, 721–723.
29. K. N. Rao, A. Sridhar, A. F. Lee, S. J. Tavener, N. A. Young and K. Wilson, Green Chem., 2006, 8, 790–797.
30. M. K. Lam and K. T. Lee, Fuel Process. Technol., 2011, 92, 1639–1645.
31. G. Chandra Behera and K. M. Parida, Dalton Trans., 2012, 41, 1325–1331.
32. W. Xie and D. Yang, Bioresour. Technol., 2012, 119, 60–65.
33. J. S. Beck, J. C. Vartuli, W. J. Roth, M. E. Leonowicz, C. T. Kresge, K. D. Schmitt, C. T. W. Chu, D. H. Olson and E. W. Sheppard, J. Am. Chem. Soc., 1992, 114, 10834–10843.
34. C. T. Kresge, M. E. Leonowicz, W. J. Roth, J. C. Vartuli and J. S. Beck, Nature, 1992, 359, 710–712.
35. Y. Rao, M. Trudeau and D. Antonelli, J. Am. Chem. Soc., 2006, 128, 13996–13997.
36. J. Fan, Y. Dai, Y. Li, N. Zheng, J. Guo, X. Yan and G. D. Stucky, J. Am. Chem. Soc., 2009, 131, 15568–15569.
37. Q. Yuan, A. X. Yin, C. Luo, L. D. Sun, Y. W. Zhang, W. T. Duan, H. C. Liu and C. H. Yan, J. Am. Chem. Soc., 2008, 130, 3465–3472.
38. L. Xu, H. Zhao, H. Song and L. Chou, Int. J. Hydrogen Energy, 2012, 37, 7497–7511.
39. A. Bordoloi and S. B. Halligudi, J. Catal., 2008, 257, 283–290.
40. D. Hua, S. L. Chen, G. Yuan, Y. Wang, Q. Zhao, X. Wang and B. Fu, Microporous Mesoporous Mater., 2011, 143, 320–325.
41. G. A. Olah, Friedel-Crafts Chemistry, Wiley, New York, 1973.
42. F. Wang and W. Ueda, Chem.–Eur. J., 2009, 15, 742–753.
43. J. Dou and H. C. Zeng, J. Am. Chem. Soc., 2012, 134, 16235–16246.
44. C. R. Kumar, K. Jagadeeswaraiah, P. S. S. Prasad and N. Lingaiah, ChemCatChem, 2012, 4, 1360–1367.
45. Z. Miao, L. Xu, H. Song, H. Zhao and L. Chou, Catal. Sci. Technol., 2013, 3, 1942–1954.
46. F. A. H. Al-Qallaf, L. F. Hodson, R. A. W. Johnstone, J. Y. Liu, L. Lu and D. Whittaker, J. Mol. Catal. A: Chem., 2000, 152, 187–200.
47. A. S. Khder and A. I. Ahmed, Appl. Catal., A, 2009, 354, 153–160.
48. C. A. Emeis, J. Catal., 1993, 141, 347–354.
49. D. G. Barton, M. Shtein, R. D. Wilson, S. L. Soled and E. Iglesia, J. Phys. Chem. B, 1999, 103, 630–640.
50. A. Martínez, G. Prieto, M. A. Arribas, P. Concepción and J. F. Sánchez-Royo, J. Catal., 2007, 248, 288–302.
51. K. Song, H. Zhang, Y. Zhang, Y. Tang and K. Tang, J. Catal., 2013, 299, 119–128.
52. X. L. Yang, R. Gao, W. L. Dai and K. Fan, J. Phys. Chem. C, 2008, 112, 3819–3826.
53. M. S. Morey, J. D. Bryan, S. Schwarz and G. D. Stucky, Chem. Mater., 2000, 12, 3435–3444.
54. D. C. Vermaire and P. C. van Berge, J. Catal., 1989, 116, 309–317.
55. D. G. Barton, S. L. Soled, G. D. Meitzner, G. A. Fuentes and E. Iglesia, J. Catal., 1999, 181, 57–72.
56. A. Ramanathan, R. Maheswari, B. P. Grady, D. S. Moore, D. H. Barich and B. Subramaniam, Microporous Mesoporous Mater., 2013, 175, 43–49.
57. F. Di Gregorio and V. Keller, J. Catal., 2004, 225, 45–55.
58. M. H. C. de la Cruz, M. A. Abdel-Rehim, A. S. Rocha, J. F. C. da Silva, A. da Costa Faro Jr and E. R. Lachter, Catal. Commun., 2007, 8, 1650–1654.
59. T. Kitano, T. Shishido, K. Teramura and T. Tanaka, J. Phys. Chem. C, 2012, 116, 11615–11625.
60. N. T. S. Phan, K. K. A. Le and T. D. Phan, Appl. Catal., A, 2010, 382, 246–253.

---

Footnote

† Electronic supplementary information (ESI) available: The wide-angle X-ray diffraction of 20 wt% WO₃/ZrP_sol–gel, IR spectra for pyridine adsorbed on the 20 wt% WO₃/M-ZrPO recorded at different temperatures, isotherms and pore size distribution and TEM images of 20 wt% WO₃/M-ZrPO-used after five cycles. See DOI: 10.1039/c4ra02809k

---