Reductive cyclization of 2-nitro-2′-hydroxy-5′-methylazobenzene to benzotriazole over K-doped Pd/γ-Al₂O₃

Abstract

A series of Pd/γ-Al₂O₃ catalysts modified by potassium salts were prepared and evaluated in the reductive cyclization of 2-nitro-2′-hydroxy-5′-methylazobenzene without additional base. These solid base-hydrogenation bifunctional catalysts were characterized and the results demonstrated that potassium salts could have an important impact on the properties and catalytic performance of Pd/γ-Al₂O₃.

---

As one of the most important ultraviolet absorbers, 2-(2′-hydroxy-5′-methylphenyl)benzotriazole (BTA) is widely employed for polymer materials.^1,2 Currently in industry, the typical synthetic method of BTA is the reductive cyclization of 2-nitro-2′-hydroxy-5′-methylazobenzene (NAB) with zinc dust in the presence of sodium hydroxide.³ Unfortunately, the resulting zinc sludge and lye after reaction cause severe environmental pollution. Among all the reduction methods,^3–9 catalytic hydrogenation of NAB to BTA is almost an ideal one for the synthesis of BTA because the only byproduct is water, which can fully solve the serious environment pollution caused by the production of BTA. However, majority of the related works were restricted to the theoretical discussion, and no significant improvement in the catalytic hydrogenation of NAB to BTA until the continuous synthesis of BTA over Pd/γ-Al₂O₃ was reported.¹⁰ The catalytic hydrogenation of NAB to BTA needs to be conducted with extra base, which could provide alkaline environment, such as triethylamine. Otherwise, without base, the yields of amino byproducts shoot up.¹¹ The heavy use of base inevitably leads to troublesome separation procedures and serious environmental contaminations. From the perspective of green chemistry and environmental protection, there is yet room for the further study on the catalytic hydrogenation of NAB to BTA.

As presented in our previous works, Pd/γ-Al₂O₃ exhibited excellent performance for the reductive cyclization of NAB to BTA, but required organic amines to provide alkaline condition.¹⁰ Solid base is proposed as the effective substitute of the homogeneous base catalyst with the consideration that it possesses high activity in various reactions under mild conditions and minimizes the production of pollutants.^12,13 It is generally known that hydrogenation reaction proceeds at the micro-environmental sites provided by the catalyst. If there exist alkaline sites on the surface of the hydrogenation catalyst would render it certain alkaline property besides hydrogenation catalysis, thus the reductive cyclization is possibly effectively catalyzed by this kind of solid base-hydrogenation bifunctional catalysts without additional base. However, compared with solid acid-hydrogenation catalysts, solid base-hydrogenation bifunctional catalysts have been rarely reported so far.¹³

In this work, three solid base-hydrogenation bifunctional catalysts were designed, prepared and evaluated in the reductive cyclization of NAB. 1% Pd/γ-Al₂O₃ catalysts modified with potassium salts (KNO₃, K₂CO₃, and KOH) were prepared by impregnation technique. To investigate the structure–activity relationship, these catalysts were characterized by Fourier transform infrared spectroscopy (FT-IR), X-ray diffraction (XRD), Brunauer–Emmer–Teller (BET) surface area measurement, transmission electron microscopy (TEM), scanning transmission electron microscopy with elemental mapping (STEM-elemental mapping), temperature programmed desorption of carbon dioxide (CO₂-TPD), and H₂-temperature programmed reduction (H₂-TPR). Subsequently, these K-doped catalysts were employed in the reductive cyclization of NAB to BTA and the catalytic performance was evaluated.

2-Nitro-2-hydroxy-5-methylazobenzene (≥90%, determined by HPLC) was prepared as the customary procedures of diazotization and coupling.³ Pd/γ-Al₂O₃ (the content of Pd in the Pd/γ-Al₂O₃ is 1 wt%) catalyst was prepared by incipient wetness impregnation.¹⁰ K-doped 1 wt% Pd/γ-Al₂O₃ catalysts (the mass ratio of potassium and Pd/γ-Al₂O₃ is 1 : 10) were prepared by impregnation of the pre-reduced Pd/γ-Al₂O₃ catalyst in an aqueous solution of the corresponding potassium salt (KNO₃, K₂CO₃, or KOH). The catalyst is designated as Pd/γ-Al₂O₃–X, with X denoting the kind of potassium salt. For example, Pd/γ-Al₂O₃–KNO₃ means the K in Pd–K/γ-Al₂O₃ catalyst coming from KNO₃.

FT-IR spectra of γ-Al₂O₃, Pd/γ-Al₂O₃, Pd/γ-Al₂O₃–KNO₃, Pd/γ-Al₂O₃–K₂CO₃, and Pd/γ-Al₂O₃–KOH were measured and described in Fig. 1. There was no obvious change in the spectra of γ-Al₂O₃ and Pd/γ-Al₂O₃, even among the spectra of K modified samples, which indicated that the main structures of Al₂O₃ in these catalysts were similar. The two peaks at 3443 cm⁻¹ and 1629 cm⁻¹ were attributed to the O–H stretching vibration and the O–H bending vibration of physical adsorbed water.^14–16 Water might come from air during the testing and the preparation process of the catalyst. In addition, the band at 3443 cm⁻¹ could be partly attributed to the stretching vibrations of Al–O–K bonds.^16–19 Compared with the spectra of γ-Al₂O₃ and Pd/γ-Al₂O₃, there were two new bands at 1527 cm⁻¹ and 1400 cm⁻¹ in the curves of Pd/γ-Al₂O₃–KNO₃, Pd/γ-Al₂O₃–K₂CO₃, and Pd/γ-Al₂O₃–KOH. These two bands were caused by CO₃²⁻ ions weakly bounded to K⁺ on the surface of the catalyst, which came from K₂O·CO₂ species. The current assumption is that K₂O·CO₂ species on K-doped materials come from the adsorption of ambient CO₂ on the basic sites of K₂O which was formed during the calcination procedure.^{14,16,20–22} It has been confirmed that K₂O·CO₂ was the catalytically active sites for some reactions.¹⁶ According to previous reports, the two bands appeared at 1121 cm⁻¹ and 783 cm⁻¹ were attributed to AlO₄ units, as well as the band at 586 cm⁻¹ was ascribed to AlO₆ units, indicating the presence of Al–O–Al framework.^16,19,25 Moreover, compared with γ-Al₂O₃ and Pd/γ-Al₂O₃, there were some small changes in the shoulders of three K-doped catalysts at 1121 cm⁻¹ and 783 cm⁻¹, which were caused by the effect of K on Al–O bonds (AlO₄ units).^16,19

Fig. 1 FT-IR spectra of (a) γ-Al₂O₃, (b) Pd/γ-Al₂O₃, (c) Pd/γ-Al₂O₃–KNO₃, (d) Pd/γ-Al₂O₃–K₂CO₃, and (e) Pd/γ-Al₂O₃–KOH.

Fig. 2 shows the XRD curves of Pd/γ-Al₂O₃, Pd/γ-Al₂O₃–KNO₃, Pd/γ-Al₂O₃–K₂CO₃, and Pd/γ-Al₂O₃–KOH. The peaks appeared at 37°, 46°, and 67° related to the 111, 400, and 440 plans (JCPDS 00-029-0063), respectively, were attributed to γ-Al₂O₃.^10,23,24 The diffraction peak at 40° (JCPDS 00-005-0681) cannot be assigned to palladium (111) phase categorically because it was covered by γ-Al₂O₃ diffraction peaks and its intensity was too weak due to the low loading of palladium.^23,25 After modifying by potassium salts, the intensities of γ-Al₂O₃ typical diffraction peaks (67°, 46°, and 37°) decreased sharply, demonstrating the formation of new phase with no change in positions is observed.^16,26,29 Pd/γ-Al₂O₃–KNO₃, Pd/γ-Al₂O₃–K₂CO₃, and Pd/γ-Al₂O₃–KOH exhibited new diffraction peaks at 2θ value of 21° and 29°, which were ascribed to orthorhombic α-KAlO₂ species (JCPDS 01-089-8451). It could be speculated that the reaction between potassium salts and γ-Al₂O₃ generated a new phase of Al–O–K during the calcination.^16,27 In addition, the new phase of K₂O (2θ = 31°) could be observed clearly in the XRD patterns of all the three K-doped catalysts (JCPDS 00-027-0431).^19,20,28 The above analyses agreed well with the FT-IR results. Both K₂O species and Al–O–K sites on the surface of Al₂O₃ have been proved to be the main active sites for some reactions.^19,29 Moreover, the base sites were probably obtained through the decomposition of potassium salts and the interaction between potassium salts and γ-Al₂O₃ during the preparation of these three catalysts. It was also found that γ-Al₂O₃ could enhance, at least in part, the decomposition of potassium salts because some potassium salts, such as K₂CO₃, were difficult to be decomposed even above 1000 °C.¹⁹ It was worth noting that the intensities of diffraction peaks at 2θ = 40° in the XRD pattern of Pd/γ-Al₂O₃–KNO₃ was much stronger than the other two K-doped catalysts, while it was the weakest in the XRD pattern of Pd/γ-Al₂O₃–KOH. According to the above results, it was speculated that palladium species might distribute differently on different K-doped catalysts.

Fig. 2 XRD patterns of (a) Pd/γ-Al₂O₃, (b) Pd/γ-Al₂O₃–KNO₃, (c) Pd/γ-Al₂O₃–K₂CO₃, (d) Pd/γ-Al₂O₃–KOH. (#), Al–O–K; (*), K₂O; (◊), γ-Al₂O₃.

Previous reports suggested that the dispersion of palladium crystals on the support materials might affect their catalytic performance.^10,30 Thus, TEM studies of Pd/γ-Al₂O₃–KNO₃, Pd/γ-Al₂O₃–K₂CO₃, and Pd/γ-Al₂O₃–KOH were carried out to visually confirm the dispersions and particle sizes of the palladium phase. As shown in Fig. 3a–c, palladium particles were visible as black dots on all the catalysts. According to particle size distribution histograms, it could be found that palladium particles of Pd/γ-Al₂O₃–KNO₃ were much larger than those of the other two catalysts and the average palladium diameter was 9.31 nm. The smallest palladium particles were found in Fig. 3c, and the average palladium diameter of Pd/γ-Al₂O₃–KOH was just 3.60 nm. Pd/γ-Al₂O₃–K₂CO₃ had an average palladium diameter of 4.68 nm, which was smaller than Pd/γ-Al₂O₃–KNO₃ but larger than Pd/γ-Al₂O₃–KOH (Fig. 3b). These results indicated that the modification of Pd/γ-Al₂O₃ with potassium salts presented a significant effect on the distribution of palladium phase on the surface of the catalyst, which could further affected the catalytic performance. Fig. 3d–f shows the STEM images of Pd/γ-Al₂O₃–KNO₃, Pd/γ-Al₂O₃–K₂CO₃, and Pd/γ-Al₂O₃–KOH, with elemental mapping (Al, O, Pd and K). Among all the K-doped catalysts, the distribution of K was similar with that of Pd. Pd and K components almost appeared in the same position, indicating that K could control the particle size of Pd. It also could be found that K from different potassium salts resulted in a different effect on Pd distribution. Based on the results of BET surface area measurements (ESI†) and earlier studies, it was obvious that the catalytic performance of these modified catalysts remarkably depended on the kind of potassium salt.³¹ TEM and STEM-elemental mapping characterizations were in accordance with the results of XRD.

Fig. 3 TEM images and Pd particle size distributions of (a) Pd/γ-Al₂O₃–KNO₃, (b) Pd/γ-Al₂O₃–K₂CO₃, and (c) Pd/γ-Al₂O₃–KOH; STEM images of (d) Pd/γ-Al₂O₃–KNO₃, (e) Pd/γ-Al₂O₃–K₂CO₃, and (f) Pd/γ-Al₂O₃–KOH, with elemental mapping (Al, O, Pd, K).

CO₂-TPD was employed to probe the alkaline information of Pd/γ-Al₂O₃, Pd/γ-Al₂O₃–KNO₃, Pd/γ-Al₂O₃–K₂CO₃, and Pd/γ-Al₂O₃–KOH. The profiles are described in Fig. 4. It was easy to understand the CO₂-TPD signal of Pd/γ-Al₂O₃ was rather weak compared with Pd/γ-Al₂O₃–KNO₃, Pd/γ-Al₂O₃–K₂CO₃ and Pd/γ-Al₂O₃–KOH due to the acidic nature of Pd/γ-Al₂O₃.³² By contrast, Pd/γ-Al₂O₃–KNO₃, Pd/γ-Al₂O₃–K₂CO₃ and Pd/γ-Al₂O₃–KOH displayed remarkable and broad desorption peaks. As previously reported, the peaks below 200 °C in CO₂-TPD profiles are assigned to the desorption of CO₂ from weak basic sites, the peaks between 200 and 400 °C are ascribed to the desorption of CO₂ from moderate basic sites, and the peaks above 400 °C are attributed to the desorption of CO₂ from strong basic sites.^16,33 The formation of multiple basic sites was possibly caused by new phases Al–O–K, K₂O·CO₂ and K₂O species formed during the calcination.^16,19,34 This conclusion was consistent with FT-IR and XRD results. Meanwhile, it was found that the spectra of these three K-doped catalysts were similar but not identical. Pd/γ-Al₂O₃–K₂CO₃ exhibited the largest ratio of weak basic sites among these three catalysts. The basic sites distribution of Pd/γ-Al₂O₃–KOH was similar as Pd/γ-Al₂O₃–KNO₃, while the ratio of strong basic sites of Pd/γ-Al₂O₃–KOH was even higher (Table S2†). Taking TEM, STEM-elemental mapping and CO₂-TPD results into consideration, a conclusion might be draw that the introduction of potassium salts into Pd/γ-Al₂O₃ affected not only the palladium species distribution of Pd/γ-Al₂O₃, but also the basic property of the catalyst as well.

Fig. 4 CO₂-TPD profiles of (a) Pd/γ-Al₂O₃, (b) Pd/γ-Al₂O₃–KNO₃, (c) Pd/γ-Al₂O₃–K₂CO₃, and (d) Pd/γ-Al₂O₃–KOH.

H₂-TPR experiments are generally used to investigate the reducibility of the species on a given support. By comparing the TPR results of undoped and doped samples, it is possible to determine the effect of doped metals.³⁵ The H₂-TPR profiles for Pd/γ-Al₂O₃, Pd/γ-Al₂O₃–KNO₃, Pd/γ-Al₂O₃–K₂CO₃, and Pd/γ-Al₂O₃–KOH are shown in Fig. 5. There were two reduction peaks in the TPR profile of Pd/γ-Al₂O₃. The peak below 200 °C was assigned to the reduction of PdO to metallic palladium, while another peak at higher temperature could be attributed to the reduction of the species formed between PdO and γ-Al₂O₃.¹⁰ Notably, TPR profiles of all K-doped palladium catalysts presented a remarkable enhancement of hydrogen adsorption and the two reduction peaks distinctly shifted to higher temperature, indicating that the potassium introduction lowered the reducibility of palladium precursor. The reasons that cause these phenomena lie in two aspects: on one hand, palladium dispersed on alkali metal modified supports can significantly improve the reduction of the doped species, and the reduction is assisted by a hydrogen spillover from palladium metal to additives. On the other hand, the strong interaction between palladium oxide and K-doped support can obviously increase the reduction temperature of palladium precursors. Simultaneously, decoration effect and alloy formation can also increase the reduction temperature of palladium precursors during the reduction process.³⁶ It was reported that the above analyses were true in not only palladium catalysts but also some other noble metal catalysts.³⁷ Moreover, it was found that a slight shift of reduction peaks in the Pd/γ-Al₂O₃–KNO₃ profile towards higher temperature in comparison with Pd/γ-Al₂O₃–K₂CO₃ and Pd/γ-Al₂O₃–KOH. The increase of reduction temperature meant that there existed bigger particle size of palladium precursor on the surface of Pd/γ-Al₂O₃–KNO₃.¹⁰ Similarly, the palladium particle size of Pd/γ-Al₂O₃–KOH was the smallest among these three K-doped catalysts. The above results agreed well with the XRD, TEM and STEM-elemental mapping characterizations.

Fig. 5 H₂-TPR curves of precursors of (a) Pd/γ-Al₂O₃, (b) Pd/γ-Al₂O₃–KNO₃, (c) Pd/γ-Al₂O₃–K₂CO₃, and (d) Pd/γ-Al₂O₃–KOH.

The reductive cyclization of NAB is quite complicated because azo and nitro groups coexist in the molecule. Our previous work has confirmed that besides BTA, 2-amino-p-cresol, o-phenylenediamine (AC), 2-(2′-hydroxy-5′-methylphenyl)benzo-triazole N-oxide (NO) and tetrahydro-2-(2′-hydroxy-5′-methylphenyl)benzotriazole (THB) were detected in this reaction mixture and the reaction pathway is proposed as Scheme 1.¹⁰ In order to evaluate the catalytic performance of these K-doped catalysts and the effect of potassium adulteration, Pd/γ-Al₂O₃–KNO₃, Pd/γ-Al₂O₃–K₂CO₃, and Pd/γ-Al₂O₃–KOH, together with Pd/γ-Al₂O₃ were employed for the reductive cyclization of NAB without additional base and the results are summarized in Table 1.

Scheme 1 The reaction pathway for the reductive cyclization of NAB.

Table 1 Reductive cyclization of NAB over several catalysts

Catalyst	Conversion^c (%)	Selectivity (%)
		AC	NO	BTA	THB	Others^d
a Reaction conditions: molar ratio of NAB/triethylamine: 1 : 2; temperature: 60 °C; hydrogen pressure: 2.5 MPa; liquid hourly space velocity (LHSV): 0.23 h^−1.b Reaction conditions: temperature: 60 °C; hydrogen pressure: 2.5 MPa; liquid hourly space velocity (LHSV): 0.23 h^−1. Each data point is an average of three or more runs.c The conversion of NAB.d Including contaminations in raw material, errors of measurement instrument and trace impurities produced in the reaction.
Pd/γ-Al2O3^a	100	3.00	1.39	93.24	—	2.37
Pd/γ-Al2O3^b	99.51	25.39	9.64	59.08	1.08	4.32
Pd/γ-Al2O3–KNO3^b	100	4.78	6.54	84.79	0.59	3.30
Pd/γ-Al2O3–K2CO3^b	99.93	4.14	8.57	22.11	62.20	2.91
Pd/γ-Al2O3–KOH^b	100	1.58	2.54	3.96	89.31	2.61

---

As shown in Table 1, there was a marked decline in the selectivity of BTA over Pd/γ-Al₂O₃ without additional base while AC selectivity showed a fairly noticeable growth.¹⁰ So it can be found that alkaline condition is essential to the hydrogenation of NAB to BTA.¹¹ Regardless of the catalyst used, the conversion of NAB was satisfied (>99%), but the catalytic performances of K-doped catalysts were both similar and distinctly different from each other. In contrast with Pd/γ-Al₂O₃, the selectivity of AC over Pd/γ-Al₂O₃–KNO₃, Pd/γ-Al₂O₃–K₂CO₃, or Pd/γ-Al₂O₃–KOH decreased sharply. It was possibly attributed to the basicity of catalyst endowed by the modification with potassium salts. According to CO₂-TPD profiles (Fig. 4), it can be found that K-doped catalysts could offer alkaline condition for the hydrogenation reaction, and lead to the decrease of AC selectivity, while it is hard to find the regular pattern of the basic strength on the activities for the catalysts.¹⁰ It is also known that byproduct AC formed by the breakage of azo group is the main reason why benzotriazole ring is unable to form during the reaction process. Thus, the introduction of potassium salts is of considerable significance. However, except for Pd/γ-Al₂O₃–KNO₃, the selectivity of BTA over Pd/γ-Al₂O₃–K₂CO₃ and Pd/γ-Al₂O₃–KOH were 22.11% and 3.96%, respectively, while THB selectivities were 62.20% and 89.31%, respectively. These data suggested that, for Pd/γ-Al₂O₃–K₂CO₃ and Pd/γ-Al₂O₃–KOH, the reductive cyclization of NAB not only yielded BTA, but also further hydrogenated to THB. This is most probably because the distinct palladium particles sizes among these three catalysts.¹⁰ The average palladium diameter of Pd/γ-Al₂O₃–KNO₃ was much larger than that of Pd/γ-Al₂O₃–K₂CO₃ and Pd/γ-Al₂O₃–KOH. Hence, the hydrogenation activity of Pd/γ-Al₂O₃–KNO₃ might be lower than the other two, which could efficiently prevent the further hydrogenation and 84.79% of BTA selectivity was obtained. In addition, we found that, when the reaction temperature dropped to 40 °C, the selectivity of THB over Pd/γ-Al₂O₃–KOH was still higher than that of BTA, even with the decrease of NAB conversion (NAB conversion 91.08%, NO selectivity 68.67%, BTA selectivity 9.89%, and THB selectivity 18.14%). So for the K-doped palladium catalysts, the appropriate size of palladium particles is the key factor for the high yield of BTA. Through the above analysis, a speculation can be made that different potassium salts bring different impacts on the properties and catalytic performance of Pd/γ-Al₂O₃, which profoundly influence their application.

In summary, 1% Pd/γ-Al₂O₃ catalysts modified by potassium salts (KNO₃, K₂CO₃, and KOH) were prepared and characterized by FT-IR, XRD, BET surface area measurement, TEM, STEM-elemental mapping, CO₂-TPD, as well as H₂-TPR. A novel and green method for reductive cyclization of NAB to BTA without additional base was successfully established by virtue of the solid base-hydrogenation bifunctional catalyst. The results revealed that the introduction of potassium salts (KNO₃, K₂CO₃, and KOH) affects the properties of Pd/γ-Al₂O₃ distinctly, such as morphology, basicity, reducibility, and so on. During the evaluation of catalytic performance, K-doped catalysts could effectively hinder the formation of AC and exhibited different BTA selectivity. These were probably attributed to the basicity of catalysts endowed with potassium salts and the distribution of palladium phase controlled by K on the surface of Pd/γ-Al₂O₃–KNO₃, Pd/γ-Al₂O₃–K₂CO₃ and Pd/γ-Al₂O₃–KOH. Among these three K-doped catalysts, Pd/γ-Al₂O₃–KNO₃ exhibited the best catalytic performance and about 85% yield of BTA was obtained. It is obvious that solid base-hydrogenation bifunctional catalysts play similar or even better role as the concerted catalysis of Pd/γ-Al₂O₃ and base in the synthesis of BTA from NAB. These results are of considerable interest in terms of the practical and alternative route to BTA production.

Acknowledgements

Financial support was provided by the National Natural Science Foundation of China (Grant No. 21476163). The authors are grateful to Prof. Ming Meng and Dr Kui Ma in CO₂-TPD and H₂-TPR characterizations.

Notes and references

1. J. C. Crawford, Prog. Polym. Sci., 1999, 24, 7–43.
2. J. Catalán, J. L. G. de Paz, M. R. Torres and J. D. Tornero, J. Chem. Soc., Faraday Trans., 1997, 93, 1691–1696.
3. A. G. Koutsimpelis, C. G. Screttas and O. Igglessi-Markopoulou, Heterocycles, 2005, 65, 1393–1401.
4. H. L. White, U. S. Pat., 4041044, 1977.
5. B. H. Kim, S. K. Kim, Y. S. Lee, Y. M. Jun, W. Baik and B. M. Lee, Tetrahedron Lett., 1997, 38, 8303–8306.
6. J. Rody and A. C. Rochat, U. S. Pat., 4001266, 1977.
7. K. B. Park and I. W. Lee, Germany Patent 3731860, 1988.
8. S. Tanimoto and T. Kamano, Synthesis, 1986, 8, 647–649.
9. S. Kintopf and U. Kress, U. S. Pat., 4230867, 1980.
10. B. Wang, W. Fang, L. Si, Y. Li, X. Yan, L. Chen and S. Wang, ACS Sustainable Chem. Eng., 2015, 3, 1890–1896.
11. W. F. Whitmore and A. J. Revukas, J. Am. Chem. Soc., 1940, 62, 1687–1693.
12. Y. Sun, L. Sun, T. Li and X. Liu, J. Phys. Chem. C, 2010, 114, 18988–18995.
13. Y. Ono, J. Catal., 2003, 216, 406–415.
14. W. Xie, H. Peng and L. Chen, Appl. Catal., A, 2006, 300, 67–74.
15. S. U. Rege and R. T. Yang, Chem. Eng. Sci., 2001, 56, 3781–3796.
16. X. Li, D. Yu, W. Zhang, Z. Li, X. Zhang and H. Huang, Appl. Catal., A, 2013, 455, 1–7.
17. B. W. Krupay and Y. Amenomiya, J. Catal., 1981, 67, 362–370.
18. Y. Amenomiya and G. Pleizier, J. Catal., 1982, 76, 345–353.
19. H. Liu, L. Su, F. Liu, C. Li and U. U. Solomon, Appl. Catal., B, 2011, 106, 550–558.
20. W. Xie and H. Li, J. Mol. Catal. A: Chem., 2006, 255, 1–9.
21. C. Murugan, H. C. Bajaj and R. V. Jasra, Catal. Lett., 2010, 137, 224–231.
22. Y. Amenomiya, J. Catal., 1978, 55, 205–212.
23. A. Maione, F. André and P. Ruiz, Appl. Catal., B, 2007, 75, 59–70.
24. J. Zhu, Y. Wang, Y. Chun and X. Wang, J. Chem. Soc., Faraday Trans., 1998, 94, 1163–1169.
25. S. Hinokuma, K. Murakami, K. Uemura, M. Matsuda, K. Ikeue, N. Tsukahara and M. Machida, Top. Catal., 2009, 52, 2018–2111.
26. E. Rashtizadeh, F. Farzaneh and M. Ghandi, Fuel, 2010, 89, 3393–3398.
27. S. L. G. Husheer, J. G. Thompson and A. Melnitchenko, J. Solid State Chem., 1999, 147, 624–630.
28. M. S. Kotwal, P. S. Niphadkar, S. S. Deshpande, V. V. Bokade and P. N. Joshi, Fuel, 2009, 88, 1773–1778.
29. K. Noiroj, P. Intarapong, A. Luengnaruemitchai and S. Jai-In, Renewable Energy, 2009, 34, 1145–1150.
30. Y. Liu, S. Wang, D. Gao, T. Sun, C. Zhang and S. Wang, Fuel Process. Technol., 2013, 11, 55–61.
31. R. Pellegrini, G. Leofanti, G. Agostini, E. Groppo, M. Rivallan and C. Lamberti, Langmuir, 2009, 25, 6476–6485.
32. W. Wang, Y. Li, X. Zhang, W. Fang and Y. Yang, Catal. Commun., 2015, 69, 104–108.
33. D. Yanmin, C. Xingquan, Z. Chunxiang and Z. Tiansheng, J. Mol. Catal. A: Chem., 2010, 331, 125–129.
34. P. Sun, D. Yu, K. Fu, M. Gu, Y. Wang, H. Huang and H. Ying, Catal. Commun., 2009, 10, 1345–1349.
35. R. Pellegrini, G. Leofanti, G. Agostini, L. Bertinetti, S. Bertinetti, E. Groppo, A. Zecchina and C. Lamberti, J. Catal., 2009, 267, 40–49.
36. C. Shi and P. Zhang, Appl. Catal., B, 2012, 115–116, 190–200.
37. M. G. Musolino, C. V. Caia, C. Busacca, F. Mauriello and R. Pietropaolo, Appl. Catal., A, 2009, 357, 106–113.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra26800a

---