Aqueous-phase hydrodechlorination and further hydrogenation of chlorophenols to cyclohexanone in water over palladium nanoparticles modified dendritic mesoporous silica nanospheres catalyst

Abstract

Dendritic mesoporous silica nanospheres (DMSNs) have been synthesised in this work. The distance of each “arborization” is 7 nm, which plays the role of “pore”. Hydrodechlorination (HDC) of 4-chlorophenol (4-CP) as the target compound using Pd modified DMSNs as the catalyst (Pd/DMSNs) is carried out in aqueous sodium hydroxide solution under atmospheric H₂ pressure, fairly mild conditions for a potential application to treat industrial wastewater. Compared with some supported Pd catalysts, Pd/DMSNs exhibit an improved catalytic performance, owing to the specific dendritic structure, which can improve mass transfer, and increase the adsorption–desorption rate of compounds. In this work, both the dechlorination process and further hydrogenation process of 4-CP are studied under various conditions including different catalyst dosages and different temperatures. By analyzing the experimental results, it is clear that the influential factors mentioned above have a strong impact on the selectivity of the HDC experiment. In addition, 2-CP, 3-CP, and 2,4-DCP are also tested as target pollutants.

---

1. Introduction

As important commercial industrial raw materials, chlorophenols (CPs) have been widely employed in different domains, such as in the manufacture of herbicides, dyes and plant growth regulators, wood, paints, fibers, and leather preservatives, as well as disinfectants.^1,2 However, because they are highly toxic and poorly degradable in wastewater, soil and polluted groundwater, CPs have constituted a particular group of priority toxic chlorinated pollutants listed by the US EPA.³ Hence, how to convert CPs into less harmful, or even harmless, compounds has become a major environmental concern. In order to dispose of CPs, a lot of methods have been used, such as oxidative, photocatalytic degradation,⁴ thermal combustion, aerobic/anaerobic biodegradation,^5,6 catalytic reaction based on polymer membrane⁷ and reduction dechlorination.^8,9 In order to achieve efficient destruction/removal, a high temperature is necessary in the processes of thermal combustion and oxidative treatment of CPs; there is also the real possibility of the formation of highly toxic dioxins in these processes, which should not be overlooked. In addition, the application of a relatively environmentally friendly method, such as biological treatments, is also limited by the toxicity of CPs. These limitations are forcing researchers to develop a more environmentally friendly technique for the removal of CPs. When considering economic and reactive efficiency, catalytic hydrodechlorination (HDC) is now regarded as a promising technology, because it is suitable for a wide range of chlorinated compounds, requires only mild reaction conditions, and is able to facilitate raw material recycling compared with all the other alternatives.¹⁰

In the 1960s, the study of HDC of CPs was initiated, but the first comprehensive report of the liquid phase exhaustive HDC of CPs to phenol over Pd/C only appeared in 1992.¹¹ From then on, HDC of CPs has been extensively studied. In the literature, the research is mainly focused on the structure of the catalyst, reaction media, reaction conditions and sources of hydrogen.^12,13 Supported metal nanocatalysts have been largely studied, such as supported Pd,^13–17 Pt,^18,19 Rh,²⁰ Ni¹⁵ and some bimetallic alloy catalysts.²¹ By comparing the catalytic activity of these nanocatalysts, a Pd based catalyst is the best choice for HDC of CPs at room temperature and atmospheric pressure. In order to develop an efficient Pd based catalyst, multiple supported Pd catalysts have been employed in HDC of CPs, such as Pd/activated carbon (AC),^22–24 Pd/Al₂O₃ ²⁵ and Pd/zeolites.^26,27 As a support, Al₂O₃ shows a high mechanical resistance and a strong interaction with metal NPs, leading to enhanced metal dispersion. However, it is very sensitive to the HCl that is formed during the reaction, whereas activated carbons are much more resistant to this compound.²⁸ Thus, fibrous spherical dendritic mesoporous silica nanospheres (DMSNs) with a 7 nm pore structure may provide a better alternative. As a support, DMSNs possess features such as a high surface area and a dendritic mesoporous structure, which offer molecules an easy accessibility through its fibers (as opposed to the traditional use of pores). The dendritic structure can bind to metal NPs and the space between arborizations can improve mass transfer, and increase the adsorption–desorption rate of compounds. These characters make DMSNs an ideal support candidate to form noble metal-based catalysts. Considering its economic and environmentally friendly properties, high efficiency and facile synthesis, DMSNs supported Pd nanocatalyst (Pd/DMSNs) is designed and employed in HDC of CPs. In addition, the Pd/DMSNs nanocatalyst is a promising candidate for various Pd-based catalytic applications.

2. Experimental

2.1 Materials

Tetraethoxysilane (TOES), Pd(II) acetate, (3-aminopropyl)triethoxysilane (APTES) and hexadecyltrimethylammonium chloride (CTAC) were purchased from Aladdin Chemical Co. Ltd. 2-CP, 3-CP, 2,4-DCP and concentrated ammonia aqueous solution were purchased from Lanzhou Aihua Chemical Company. NaBH₄ was supplied by Sinopharm Chemical Reagent Co. Ltd. Organic solvents used were of analytical grade and did not require further purification.

2.2 Preparation of Pd/DMSNs (dendritic mesoporous silica nanospheres)

A DMSNs composite is reported by Dongyuan Zhao.²⁹ In this work, first generation DMSNs is used as the support, and a simple, green method (Scheme 1) has been used. First, the DMSNs are functionalized with APTES to obtain DMSNs–NH₂ nanocomposites. Second, 500 mg of the DMSNs–NH₂ nanocomposite were added to a 100 mL round-bottom flask with 108 mg of Pd(OAc)₂ and 50 mL of acetonitrile, then ultrasonically dispersed for 30 min followed by stirring for 12 h. Subsequently, the fresh NaBH₄ solution (0.3 M, 10 mL) is added dropwise into the abovementioned suspension. The product was isolated by centrifugation and washed several times with deionized water and ethanol, and then dried in vacuum overnight.

Scheme 1 Preparation of Pd/DMSNs nanocatalyst.

2.3 General procedure for the hydrodechlorination of CPs

HDC experiments were performed in a three-necked jacketed glass reactor equipped with a H₂ supply. A certain amount of catalyst is placed into the mixed solution of 30 mL solvent, 0.5 mmol of CP and a certain amount of base under a continuous flow of H₂ at 45 mL min⁻¹. The reaction was maintained for 270 min under vigorous stirring at a certain temperature. The results of the experiments were analysed by Gas Chromatography-Mass Spectroscopy (GC-MS).

2.4 General methods

Transmission electron microscopy (TEM) images were obtained on a Tecnai G2 F30, FEI, USA. The Brunauer–Emmett–Teller (BET) surface area and pore-size distribution were obtained by measuring N₂ adsorption isotherms at 77 K using a TriStar 3020 (Micromeritics). Powder X-ray diffraction (XRD) spectra were obtained by a Rigaku D/max-2400 diffractometer using Cu-Kα radiation in the 2θ range of 0–80°. X-ray photoelectron spectroscopy (XPS) was recorded on a PHI-5702 and the C1S line at 284.6 eV was used as the binding energy reference. The reaction conversion was estimated by using GC-MS (P.E. AutoSystem XL).

3. Results and discussion

3.1 Characterization

Fig. 1 shows the FT-IR spectra of DMSN and DMSN–NH₂. The adsorption peaks at 1091 cm⁻¹ and 806 cm⁻¹ correspond to the antisymmetric and symmetric stretching vibrations of the Si–O–Si bond in the oxygen–silica tetrahedron, respectively. The peak at 467 cm⁻¹ corresponds to Si–O stretching. The strong peak at 3419 cm⁻¹ shows a large number of Si–OH groups present, which proved to be advantageous in the modification of APTES on the DMSN surface by hydrogen bonds. The adsorption peak at 2986 cm⁻¹ corresponds to –CH stretching. In the FT-IR spectrum of DMSN–NH₂, the peak around 3419 cm⁻¹ represents the adsorption of –OH and –NH₂ groups. The nitrogen, hydrogen, and carbon contents are 1.29%, 1.57%, and 8.98%, measured by the elementary analysis, respectively. The FT-IR spectra and elementary analysis result reveal that the APTES was successfully grafted on the DMSN surface, thus enabling them to act as robust anchors for metal NPs.

Fig. 1 FT-IR spectra of DMSN and DMSN–NH₂.

SEM and TEM images of DMSNs and Pd/DMSNs are presented in Fig. 2. As illustrated in Fig. 2c, mesoporous silica nanospheres with a dendritic structure can be observed. The ordered dendritic fibers come from the center of the particle and are distributed uniformly in all directions. Besides, a representative pore dimension can be calculated by measuring the distance between two fibers in TEM micrographs, which is about 7 nm (Fig. 2c). According to the large-scale TEM image (Fig. 2b) of DMSNs, a size frequency curve can be obtained (inset of Fig. 2b). From this size frequency curve it can be concluded that the particle size of the prepared DMSNs is within the range 60–160 nm and the mean diameter size is within the range 90–130 nm. The TEM image of Pd NPs loaded on DMSNs catalyst is shown in Fig. 2d, which clearly shows that the well-dispersed Pd NPs load onto the internal surface of DMSNs and the calculated Pd NPs diameter is about 4 nm (Table 1).

Fig. 2 SEM image of DMSNs (a), TEM images of DMSNs (b and c) and Pd/DMSNs (d).

Table 1 Surface area, pore volume and pore size of DMSNs and Pd Pd/DMSNs

Sample	Surface area (m^2 g^−1)	Pore volume (cm^3 g^−1)	Pore size (nm)
DMSNs	664.7	1.30	7.80
Pd/DMSNs	566.7	0.88	6.02

---

N₂ adsorption–desorption isotherms for the DMSNs and Pd/DMSNs are given in Fig. 3. According to the IUPAC classification, the curves of DMSNs and Pd/DMSNs are type IV isotherms with a very sharp capillary condensation stepped at P/P₀ = 0.4–0.8 and H₂-type hysteresis loop, characterizing small-pore mesoporous materials with cylindrical channels. The pore size of DMSNs derived from BJH analysis on the desorption branch is 7.8 nm. The calculated BET surface area and pore volume of DMSNs are 664.7 m² g⁻¹ and 1.3 cm³ g⁻¹, respectively. Compared with DMSNs, the pore size of Pd/DMSNs reduced from 7.8 nm to 6.02 nm, and the pore volume values reduced from 1.3 cm³ g⁻¹ to 0.88 cm³ g⁻¹, which is due to the channels dispersed by metal NPs. Besides, the surface area of prepared samples reduced from 664.7 m² g⁻¹ to 566.7 m² g⁻¹, which confirms that Pd NPs are loaded on the DMSNs.

Fig. 3 Nitrogen adsorption–desorption isotherms and pore size distribution (inset) of DMSNs (a) and Pd/DMSNs (b).

XRD patterns of DMSNs and Pd/DMSNs are displayed in Fig. 4. DMSNs and Pd/DMSNs both showed a major diffraction peak at 2θ = 2.3°, which corresponds to plane (211) of mesoporous material. The broad peak between 20–30° corresponds to amorphous silica.³⁰ The XRD pattern of Pd NPs shows three characteristic diffraction peaks at 2θ = 40°, 46° and 68°, corresponding to (111), (200) and (220), respectively. Within the approximation of the Scherrer equation, one would expect a FWHM of 2.3° 2 theta for the (111) reflection of Pd with an average crystal size of 4.3 nm, which matches with the particle size of Pd NPs calculated from TEM images.

Fig. 4 Small-angle (a) and wide-angle (b) XRD patterns of DMSNs and Pd/DMSNs.

Fig. 5 presents XPS elemental survey scans of the surface of the Pd/DMSNs catalyst. Peaks corresponding to oxygen, silica and palladium are clearly observed. In the inset figure, Pd⁽⁰⁾ binding energy of Pd/DMSNs exhibits two sharp peaks centered at 334.3 eV and 340 eV, which are assigned to Pd 3d_5/2 and Pd 3d_3/2. In addition, no evidence proved the presence of palladium oxide, which indicates that Pd(OAc)₂ was completely reduced to Pd⁽⁰⁾ NPs.

Fig. 5 XPS spectrum of Pd/DMSN (inset: high resolution spectra of Pd 3d).

3.2 HDC of 4-CP

The catalytic activity of Pd/DMSNs nanocatalyst is investigated using the HDC of CPs. The HDC of CPs is negligible without catalyst or in the presence of pure DMSN at the same conditions, which shows that the presence of metal NPs is essential for high catalytic activity. The 4-CP HDC reaction mechanism is described as below: H₂ adsorbed on the active site of the Pd/DMSNs nanocatalyst is activated into two hydrogen atoms, which combine with 4-CP and then adsorbs onto the surface of the nanocatalyst. The C–Cl bond of 4-CP is attacked by the active hydrogen atoms to form phenol.²¹ Simultaneously, in the process of HDC of 4-CP, HCl is formed as a by-product, which can poison catalysts. In order to reduce HCl inhibition, NaOH is chosen as a base to neutralize HCl.²¹ With the addition of base, catalyst deactivation is mainly governed by HCl solubility/transport and the nature of the basic species in the catalyst matrix has reported higher HDC rates and enhanced catalyst stability.³¹

The HDC pathway is indicated in Scheme 2. As Diaz et al.¹⁸ established in a previous work, the route of 4-CP HDC proceeds through a set of series–parallel reactions where 4-CP gives rise to phenol followed by cyclohexanone (CYC), which is produced from phenol hydrogenation. Fig. 6 shows the time-dependent concentration of 4-CP and the concentration of the product in the HDC reaction when the Pd/DMSNs nanocatalyst is used. The investigation of the reaction conditions revealed that the maximum conversion of 4-CP in phenol occurs within 270 min. By analyzing the experimental results, it was clear that under low catalyst dosage (10 mg) and low temperature (20 °C), only phenol is produced as the dechlorination product. It should also be noted that when the conversion value of 4-CP is higher than 80%, a tiny amount of CYC begins to form. This is in agreement with other researchers' work under low Pd dosage and low temperature.^{21,30,32–34} When increasing the catalyst dosage from 10 mg to 20 mg and the reaction temperature from 20 °C to 40 °C or 65 °C, phenol and further hydrogenation products are detected in the process of HDC of 4-CP. This result indicates that the amount of activity phase (Pd) and the temperature have a strong effect on the selectivity of the experiment (Fig. 7).

Scheme 2 Schematic of the HDC pathway for CPs.

Fig. 6 Yield curves of HDC of 4-CP.

Fig. 7 Fitted kinetic rate constants of the 4-CP to phenol (left) and phenol to CYC (right).

According to the reaction pathway (Scheme 2), the following rate equations can be written based on the assumption of pseudo-first order kinetics:³⁵

The concentration–time curves for the catalysts at 10.01 wt% metal loading which was measured by inductively coupled plasma measurement is fitted to the above equations by a nonlinear regression programme that uses the Marquardt algorithm at the 95% probability level. The kinetic rate constants and the reaction rate constant per unit mass k^′_i = k_i/M_Pd are calculated and exhibited in Table 3.

In this study, the amount of catalyst dosage is doubled (from 10 mg to 20 mg) at 20 °C, the dechlorination process kinetic rate constant k₁ increases nearly two times (from 0.0135 min⁻¹ to 0.0259 min⁻¹) and CYC yield value increases nearly two times (from 2.42% to 4.87%), which indicates that in the reaction system, the relationship between the catalytic activity and catalyst dosage is linear. This result also suggests that the amount of catalyst dosage is not an influential factor of 4-CP HDC. Under the condition of 20 mg catalyst dosage, on increasing the reaction temperature from 20 °C to 40 °C, both the reaction rate of the dechlorination process and the further hydrogenation process increases. The reaction rate of the dechlorination process increases from ∼13.2 min⁻¹ g⁻¹ to 65 min⁻¹ g⁻¹, an increase of nearly four times, and the yield of CYC increases from 4.868% to 18.5%, an increase of almost four times. However, when increasing the reaction temperature from 40 °C to 65 °C, reaction rate k₁ reduces from 65 min⁻¹ g⁻¹ to 51 min⁻¹ g⁻¹, but k₂ increases from 0.35 min⁻¹ g⁻¹ to 0.6 min⁻¹ g⁻¹. This indicates that appropriately increasing the reaction temperature can speed up the dechlorination process of 4-CP HDC, but higher temperatures will restrain this process. On the contrary, further hydrogenation process does not show the same trend. According to this result, it can be concluded that by controlling the catalyst dosage and reaction conditions the selective degradation of CPs to phenol or CYC can be controlled.

In addition, a large amount of catalyst dosage (100 mg) was also studied at 65 °C. Under these conditions, 4-CP is completely converted into CYC in 270 min and a very small quantity of cyclohexanol is also detected. The HDC of 2-CP, 3-CP and 2,4-DCP are also tested and the results are depicted in Table 2. The HDC of 2,4-DCP can proceed in a stepwise and/or concerted fashion with 2-CP and 4-CP as partially dechlorinated products. Thus, over the same time frame, 67% of 2,4-DCP is converted.

Table 2 Yield of HDC of CPs catalyzed by Pd/DMSNs

Reactant	Temperature (°C)	Catalyst dosage (mg)	Phenol	CYC
			Time (min)/yield%	Time (min)/yield%
4-CP	20	10	120/100%	270/2.42%
4-CP	20	20	45/100%	270/4.87%
4-CP	40	20	30/100%	270/18.50%
4-CP	65	20	25/100%	270/29.00%
2-CP	20	10	120/100%
3-CP	20	10	120/100%
2,4-DCP	20	10	120/67%

---

The kinetic reaction rate of HDC catalyzed by Pd/Al₂O₃, and Pd/pillared clays catalysts are 3.33 min⁻¹ g⁻¹ and 7.6 min⁻¹ g⁻¹, respectively.^18,35 In comparison, Pd/DMSNs exhibited an excellent catalytic activity performance with a reaction rate of 13.2 min⁻¹ g⁻¹ at the same temperature. By comparing the rate constants for the dechlorination process and further hydrogenation process, the dechlorination proceeds significantly faster than the hydrogenation of the resulting primary product, phenol. Moreover, kinetic rate constants of dechlorination are much larger than further hydrogenation of the resulting secondary product, CYC.

According to the kinetic reaction rate of the dechlorination process k₁ and further hydrogenation process k₂ at different temperatures within the 20–65 °C range in Table 3, the corresponding values of apparent activation energy are calculated by the Arrhenius equation and a value of 61.5 kJ mol⁻¹ is obtained for the Pd/DMSNs catalyst. According to literature reports, the apparent activation energy of Pd on activated carbon (Pd/AC) ranges from 24.8 kJ mol⁻¹ to 47 kJ mol⁻¹.^25,36,37 In addition, apparent activation energy of the hydrogenation process is also calculated as 2.28 kJ mol⁻¹.

Table 3 Fitted kinetic rate constants k₁ for dechlorination process and k₂ for further hydrogenation

Catalyst dosage (mg)	Temperature (°C)	k1 (min^−1)	k1/MPd (min^−1 g^−1)	k2 (min^−1)	k2/MPd (min^−1 g^−1)	R^2
10	20	0.0135	13.5			0.98
20	20	0.0259	13.0			0.98
20	40	0.1300	65.0	0.0007	0.35	0.99
20	65	0.1020	51.0	0.0012	0.60	0.99

---

The circulation experiment is performed in a centrifuge tube with a H₂ supply. The catalyst is recovered by centrifugation and simple decantation of liquid products. The catalyst is then washed with deionized water, and used directly for the next cycle of the reaction without further purification. The recoverability and reusability are investigated by the HDC reaction of 4-CP and the results are summarized in Fig. 8. After 5 recycling times, the metal loading of catalyst of 9.89% is measured instead of 10.01% and the catalytic activity of Pd/MSNs shows a slight decrease. This result confirmed the high rate of recyclability of the Pd/DMSNs nanocatalyst and indicated that metal loss is an influencing factor for decreasing catalytic activity.

Fig. 8 HDC turnover rates of 4-CP HDC over recycled catalyst.

4. Conclusion

In conclusion, Pd/DMSNs catalyst was prepared and its catalytic activity for the HDC of CPs was investigated. The Pd nanoparticles are fairly active for 4-CP HDC and further phenol hydrogenation to CYC is observed, which agrees with the study for supported Pd catalysts. The dechlorination process and further hydrogenation of 4-CP can be well described by a simple pseudo-first-order rate equation. The reaction rate constant per unit mass k^′₁ is 13.2 min⁻¹ g⁻¹, which is much higher than other Pd supported catalysts under the same reaction temperature (20 °C). Apparent activation energy values within the range of 61.5 kJ mol⁻¹ are obtained for the disappearance of 4-CP, this is a little higher than the reported value for supported Pd catalyst, which indicates the important role of the supports on the reaction pathway. The Pd/DMSNs nanocatalyst acts as a relatively green, economical, and environmentally friendly catalyst, and as a promising candidate for various Pd based catalytic applications.

References

1. K. H. Buchel, Regul. Toxicol. Pharmacol., 1984, 4, 174–191.
2. A. I. Tsyganok, I. Yamanaka and K. Otsuka, Chemosphere, 1999, 39, 1819–1831.
3. Q. Wen, T. Yang, S. Wang, Y. Chen, L. Cong and Y. Qu, J. Hazard. Mater., 2013, 244–245, 743–749.
4. M. P. Ormad, J. L. Ovelleiro and J. Kiwi, Appl. Catal., B, 2001, 32, 157–166.
5. G. Buitron, M. E. Schoeb, L. Moreno-Andrade and J. A. Moreno, Water Res., 2005, 39, 1015–1024.
6. P. S. Majumder and S. K. Gupta, Bioresour. Technol., 2008, 99, 4169–4177.
7. D. Fritsch, K. Kuhr, K. Mackenzie and F. D. Kopinke, Catal. Today, 2003, 82, 105–118.
8. Y. Han, W. Li, M. Zhang and K. Tao, Chemosphere, 2008, 72, 53–58.
9. W. Zhang, X. Quan, J. Wang, Z. Zhang and S. Chen, Chemosphere, 2006, 65, 58–64.
10. G. S. Pozan and I. Boz, J. Hazard. Mater., 2006, 136, 917–921.
11. J. B. Hoke, E. N. Balko and G. A. Gramiccioni, Abstracts Of Papers Of the American Chemical Society, 1992, vol. 203, p. 200–ENVR.
12. H. M. Roy, C. M. Wai, T. Yuan, J. K. Kim and W. D. Marshall, Appl. Catal., A, 2004, 271, 137–143.
13. F. D. Kopinke, K. Mackenzie, R. Koehler and A. Georgi, Appl. Catal., A, 2004, 271, 119–128.
14. L. Calvo, M. A. Gilarranz, J. A. Casas, A. F. Mohedano and J. J. Rodriguez, Ind. Eng. Chem. Res., 2005, 44, 6661–6667.
15. L. Calvo, M. A. Gilarranz, J. A. Casas, A. F. Mohedano and J. J. Rodriguez, Appl. Catal., B, 2008, 78, 259–266.
16. M. A. Lillo-Ródenas, J. Juan-Juan, D. Cazorla-Amorós and A. Linares-Solano, Carbon, 2004, 42, 1371–1375.
17. G. Yuan and M. A. Keane, Appl. Catal., B, 2004, 52, 301–314.
18. E. Diaz, J. A. Casas, A. F. Mohedano, L. Calvo, M. A. Gilarranz and J. J. Rodriguez, Ind. Eng. Chem. Res., 2008, 47, 3840–3846.
19. J. Halasz, S. Meszaros and I. Hannus, React. Kinet. Catal. Lett., 2006, 87, 359–365.
20. G. Yuan and M. A. Keane, Catal. Commun., 2003, 4, 195–201.
21. Z. Dong, X. Le, C. Dong, W. Zhang, X. Li and J. Ma, Appl. Catal., B, 2015, 162, 372–380.
22. M. A. Keane, J. Chem. Technol. Biotechnol., 2005, 80, 1211–1222.
23. G. S. Pozan and I. Boz, J. Hazard. Mater., 2006, 136, 917–921.
24. G. Yuan and M. A. Keane, Chem. Eng. Sci., 2003, 58, 257–267.
25. G. Yuan and M. A. Keane, Catal. Today, 2003, 88, 27–36.
26. R. F. Howe, Appl. Catal., A, 2004, 271, 3–11.
27. A. Santos, P. Yustos, A. Quintanilla, S. Rodriguez and F. Garcia-Ochoa, Appl. Catal., B, 2002, 39, 97–113.
28. Z. M. de Pedro, E. Diaz, A. F. Mohedano, J. A. Casas and J. J. Rodriguez, Appl. Catal., B, 2011, 103, 128–135.
29. D. Shen, J. Yang, X. Li, L. Zhou, R. Zhang, W. Li, L. Chen, R. Wang, F. Zhang and D. Zhao, Nano Lett., 2014, 14, 923–932.
30. X. Le, Z. Dong, X. Li, W. Zhang, M. Le and J. Ma, Catal. Commun., 2015, 59, 21–25.
31. G. Yuan and M. A. Keane, Ind. Eng. Chem. Res., 2007, 46, 705–715.
32. X. Le, Z. Dong, W. Zhang, X. Li and J. Ma, J. Mol. Catal. A: Chem., 2014, 395, 58–65.
33. X. Le, Z. Dong, Y. Liu, Z. Jin, T.-D. Huy, M. Le and J. Ma, J. Mater. Chem. A, 2014, 2, 19696–19706.
34. H. Deng, G. Fan and Y. Wang, Synthesis And Reactivity In Inorganic Metal-Organic And Nano-Metal Chemistry, 2014, vol. 44, pp. 1306–1311.
35. C. B. Molina, A. H. Pizarro, J. A. Casas and J. J. Rodriguez, Appl. Catal., B, 2014, 148, 330–338.
36. Y. Shindler, Y. Matatov-Meytal and M. Sheintuch, Ind. Eng. Chem. Res., 2001, 40, 3301–3308.
37. J. A. C. Elena Díaz, Á. F. Mohedano, L. Calvo, M. A. Gilarranz and J. J. Rodríguez, Ind. Eng. Chem. Res., 2009, 48, 3351–3358.

---