ISSN 0036-0244, Russian Journal of Physical Chemistry A, 2018, Vol. 92, No. 3, pp. 449–455. © Pleiades Publishing, Ltd., 2018.

CHEMICAL KINETICS AND CATALYSIS

Effect of Silica Particle Size on Texture, Structure, and Catalytic Performance of Cu/SiO2 Catalysts for Glycerol Hydrogenolysis1 Ye Tong Qi*, Chen Hong Zhe, and Xiang Ning Department of Chemical Engineering and Food Processing, Hefei University of Technology, Hefei, Anhui Province, 230009 China *e-mail: [email protected] Received November 1, 2016

Abstract—The influences of carrier particle sizes of Cu/SiO2 catalysts for hydrogenolysis of glycerol were studied use mono-dispersed silica as models. Catalysts were prepared by precipitation method with the average size of the mono-dispersed silica supports varying of 10, 20, and 90 nm. Characterization of the catalysts show that the physical properties such as pore volume and BET surface area of the catalysts were largely affected by the carrier particle size of silica. However, the copper dispersion of the three samples were similar. XPS patterns show a difference in the chemical states of copper species, small carrier particle size induced formation of copper phyllosilicate, which benefits on the stability of copper species in reaction. The overall activity in the reaction of glycerol hydrogenolysis shows a correlation with the carrier particle size. The small carrier particles prevent the copper species from aggregation thus such catalysts exhibit good catalytic activity and stability. Keywords: Cu/SiO2 catalyst, carrier particle size, glycerol hydrogenolysis, 1,2-propanediol, metal-support interaction DOI: 10.1134/S0036024418030366

INTRODUCTION Catalytic conversion of renewable biomass to fuels and commodity chemicals becomes more and more important because of uncertain supply of fossil fuels and global warming problems nowadays [1, 2]. As the byproduct of biodiesel production from vegetable oils, glycerol is one of the top 12 building block chemicals that can be converted to marketable products [3]. Many efforts have been put toward to the transformation of glycerol by various catalytic processes, such as oxidation [4, 5], hydrogenolysis [6, 7], reforming [8, 9], dehydration [10, 11], and esterification [12, 13]. Among them, one of the most promising approaches is to produce 1,2-propanediol (1,2-PDO) by selective hydrogenolysis of glycerol. This process provides an environmentally friendly and economically competitive route for the synthesis of 1,2-PDO from biomassderived glycerol instead of petroleum-based propylene oxide currently used. The selective hydrogenolysis of glycerol to 1,2PDO is a catalytic reaction and heterogeneous catalyst such as supported noble metal catalysts have been extensively studied [14–16]. Although the noble metals such as Ru, Rh and Pt exhibited high reactivity, they often promoted excessive C–C cleavage resulting in a poor selectivity to 1,2-PDO. As is well known, 1 The article is published in the original.

glycerol hydrogenolysis to propylene glycol is proposed to proceed via C–O bond breaking (dehydration–hydrogenation or dehydrogenation–dehydration–hydrogenation) [17–19]. Thus the Cu-based catalysts have attract more attention recently for the high efficiency of C–O bond cleavage and inactive for C–C bond scission [20]. Dasari et al. tested various commercial copper catalysts and found copper-chromite to be the most effective catalyst [14]. However, the toxic chromium limited its wide application for unresolved environmental issues. More recently, various supports such as SiO2, MgO, ZnO, Al2O3 and zeolite have been used to Cu-based catalysts [7, 21, 22]. Among these Cu-based catalysts, Cu/SiO2 has been extensively studied and attracted considerable attention for its low cost and good selectivity to 1,2-PDO. As to Cu-based catalysts, highly dispersed and stable Cu nanoparticles has been supposed to be a dominant factor in achieving excellent catalytic performance for glycerol hydrogenolysis. Zhu et al. investigated the promoting effect of boron oxide on Cu/SiO2 catalyst [23]. Addition of B2O3 to Cu/SiO2 catalyst pronouncedly enhanced the activity, 1,2-PDO selectivity, and stability for glycerol hydrogenolysis via stabilizing effect of B2O3 on Cu nanoparticles and strong interaction between copper and boron species. Vasiliadou et al. studied the bimetallic Ru–Cu catalyst supported on silica [20], and showed

449

450

YE TONG QI et al.

90 nm

90 nm

90 nm

Fig. 1. TEM images of SiO2 powders: (a) 10, (b) 20, and (c) 90 nm.

a synergetic effect between the two metals. Besides promoting effect of additives, the catalyst preparation methods have also been studied to improve the Cu dispersion hence to higher activity and selectivity. Bienholz and co-workers effectively converted glycerol to 1,2-PDO over Cu/SiO2 catalysts prepared by incipient wetness method and ion-exchange technique [24]. With the N2O chemisorption analysis, they found a linear relationship between the specific copper surface area and the catalytic activity for all of the catalysts. Huang et al. investigated the effect of precipitation manner on the Cu/SiO2 catalysts prepared by homogeneous and heterogeneous deposition–precipitation methods [25], and found that the texture, structure and composition of the catalysts were largely affected by the precipitation manner. Unfortunately, little attention has been paid to studying the influence of silica particle size on Cu/SiO2 catalyst performance in glycerol hydrogenolysis, although it is well known that the catalyst activity is strongly affected by support particles. Present work aims to elucidating the differences in the structure and catalytic performance of the Cu/SiO2 catalysts prepared with different particle size of commercial carriers. The catalysts used in this work were prepared via the simple precipitation and drying method instead of other complex synthesis techniques for focusing on the size effects of silica. Catalysts were characterized by various techniques including BET, XRD, XPS, H2-TPR, N2O chemisorption and Raman spectra. The relationship between the catalytic performance in glycerol hydrogenolysis and the structure of the Cu/SiO2 catalysts was also discussed. EXPERIMENTAL Catalyst Preparation The SiO2 powders were purchased from Xuan Cheng Jing Rui New Material Co., Ltd. without further treatment. The typical TEM images of the three silica supports are shown in Fig. 1, the average particle sizes of the carriers are calculated as 10, 20, and 90 nm, respectively. Catalysts were prepared by precipitation method and the calculated CuO contents are 10 wt % for all of the catalysts. The SiO2 powder

was added to the solution of Cu(NO3)2 · 3H2O (A.R.) (1 mol L–1). After stirring at room temperature for 4 h, Cu(NO3)2 solution was precipitated by adding an aqueous solution of NaOH (4 mol L–1) till pH >9. The precipitate was left to age in the mother liquor for 3 h at about 333 K. The resulted catalyst precursor was washed with distilled water until pH 7, then dried at 373 K for 12 h, and calcined at 723 K for 4 h in air. The as-synthesized catalysts were denoted as Cu/SiO2(x), in which x denotes the average particle size (nm) of SiO2 support. Catalyst Characterization Nitrogen adsorption experiments for pore size distribution, pore volume, and surface area measurements were conducted on a COULTER SA 3100 analyzer. All samples were calcined at 673 K under vacuum before the measurements. The X-ray diffraction (XRD) patterns were measured on an X’pert Pro Philips diffractometer with a CuKα radiation. (λ = 0.154 nm). The measurement conditions were in the range of 2θ = 10°–80°, step counting time 5 s, and step size 0.017° at 298 K. The surface elements and their states were analyzed by X-ray photoelectron spectroscopy (XPS). The XPS measurements were performed on an ESCALAB-250 (Thermo-VG Scientific, USA) spectrometer with AlKα (1486.6 eV) radiation source. Raman spectra were recorded on a LabRam HR Evolution with a 532 nm single frequency laser. The microstructures of the SiO2 supports were investigated using a JEOL-2010 scanning electron microscope. The reducibility of the calcined catalysts was determined by temperature programmed reduction (TPR) at a heating rate of 10 K min–1 under a flow of 10 vol % H2/Ar. The copper dispersion and particle size were determined by the dissociative N2O adsorption method. The experiments were performed using the same apparatus as for the TPR. The catalysts were first reduced at 623 K under 10% H2/Ar for 2 h. After cooling to 323 K in a He flow, the reduced samples were exposed to a N2O flow for 0.5 h at 323 K. Finally, the samples were cooled to room temperature to start another TPR run with 10% H2/Ar at a ramping rate of 10 K min–1 to 623 K. The average copper particle size and dispersion were calculated by assuming 1.4 × 1019 copper atoms per m2 and a molar stoichiometry N2O/Cus = 0.5 [26], where the symbol Cus means the copper atoms on the surface. The dispersion of copper was calculated by the following equation:

Dispersion, % =

Cu surface × 100. Cu total

(1)

A CuO sample was used as reference to quantify H2 consumption.

RUSSIAN JOURNAL OF PHYSICAL CHEMISTRY A

Vol. 92

No. 3

2018

EFFECT OF SILICA PARTICLE SIZE ON TEXTURE, STRUCTURE

Catalytic Test The hydrogenolysis of glycerol was carried out in a 50 mL stainless steel autoclave at a stirring speed of 400 rpm. Prior to the reaction, the catalysts were prereduced in H2 stream at 623 K for 3 h. In a typical experiment, 0.5 g of catalyst and 20 mL of aqueous glycerol solution (40 wt %) were loaded into the reactor. The autoclave was purged four times with hydrogen to remove air. Then the reactor was heated to 493 K, and the H2 pressure was increased to about 6.0 MPa and maintained during the reaction. The products were analyzed by gas chromatography on an instrument equipped with a flame ionization detector (FID) connected to a PEG-20M capillary column (for the liquid products) and a TCD connected to a Porapak Q packed column (for the gas products). The performance of hydrogenolysis was evaluated by the conversion of glycerol (C, %) and the selectivity of products (S, %), according to the following equations:

Conversion, % moles of glycerol reacted = × 100, initial moles of glycerol

=

∑

Selectivity, % moles of product formed × 100. moles of glycerol converted

Table 1. BET data for silica supports Sample

SBET, m2 g–1

d, nm

V, cm3 g–1

SiO2(10) SiO2(20) SiO2(90)

324.8 124.5 40.3

10.4 15.5 33.6

0.84 0.48 0.34

d is pore size; V is pore volume.

Table 2. The texture parameters of catalysts Parameter SBET, m2 g–1 V, cm3 g–1 d, nm Cu dispersion, % dCu(I), nm dCu(II), nm

Cu/SiO2(10) Cu/SiO2(20) Cu/SiO2(90) 274 1.16 16.2 76.3 1.31 11.6

161 0.79 23.3 68.8 1.45 16.9

69 0.31 26.1 71 1.41 19.1

dCu(I) and dCu(II) are calculated by the N2O method for reduced catalysts and from XRD data based on Sherrer equation for used catalysts, respectively.

(2)

(3)

Repeated runs showed that data variation was in the range of ±10% (relative value). RESULTS AND DISCUSSION Results of Characterization Table 1 lists the BET results of the silica supports with different particle sizes. As can be seen, with increasing of the particle size from 10 to 90 nm, the BET area of the supports gradually decreased from 324.8 to 40.3 m2 g–1, along with the pore volume decreased from 0.84 to 0.34 cm3 g–1 and pore size increased from 10.4 to 33.6 nm. This can be easily explained by the close packing model. The spheroid silica particles are packed together to form the pores. The loading of copper species on the silica supports may affect the texture parameters in two ways. Firstly, the copper particles may penetrate into small pores leading to an increase in average pore size and a decrease in surface area and pore volume. Secondly, the supported copper particles may form some new pores with an increase in pore volume and surface area. As shown in Table 2, after loading of copper on the silica supports, the average pore size of Cu/SiO2(10) and Cu/SiO2(20) samples were increased, which was mainly ascribed to the first factor of the copper particles. However, the pore volume RUSSIAN JOURNAL OF PHYSICAL CHEMISTRY A

451

increased from 0.84 to 1.16 cm3 g–1 for Cu/SiO2(10) sample and 0.48 to 0.79 cm3 g–1 for Cu/SiO2(20) could be explained in terms of the second factor of the copper particles. Both of the two factors have great influence on the BET surface area of the samples. The surface area of SiO2(10) sample was mostly due to the micro pores plugging), thus the first factor (decreased its surface area after loading of CuO. For the SiO2(20) and SiO2(90) samples, the second factor is dominant providing increased surface areas. The data on Cu dispersion of fresh catalysts as measured from dissociative N2O adsorption method are summarized in Table 2. As can be seen, the copper dispersion of the catalysts seems not dependent on the carrier particle size or BET surface area, and all of them are in the range of 68–77%. Accordingly, the calculated average copper particle sizes (dCu) have similar values in the range of 1.3–1.5 nm. However, such small particles can’t be detected by XRD method [27], which is consistent with the XRD results presented below. XRD patterns for the fresh calcined catalysts are presented in Fig. 2a. For all of the catalysts, the broad and diffuse diffraction peak centered at around 21.7° was assigned to amorphous silica [23]. Diffraction peaks of crystalline phase of CuO were not observed, indicating that copper species have good dispersion. After reduction, as Fig. 2b shows, diffraction peak of copper species was still not found on the sample of Cu/SiO2(10) and Cu/SiO2(20). However, two weak and broad diffraction peaks centered at 36.4° and 43.3° on Cu/SiO2(90) sample appeared, attributed to Cu2O and metallic copper, respectively [28], indicat-

Vol. 92

No. 3

2018

452

YE TONG QI et al.

SiO2

(a)

(a)

Cu2− 2p3/2

a 20

40

60

Intensity, a.u.

Intensity, a.u.

b

b

a

80 2θ, deg 930

SiO2

(b)

c

Satellite

c

940

950 960 Binding energy, eV

Cu2O Cu

(b)

Intensity, a.u.

f e d

20

40

60

Intensity, a.u.

c

b

a

80 2θ, deg

Fig. 2. Typical XRD patterns of fresh (a–c) and reduced (d–f) Cu/SiO2(x) catalysts, x = 10 (a, d), 20 (b, e), and 90 nm (c, f).

ing a slight agglomeration of copper species on the surface. The chemical states of the Cu species were investigated by XPS analysis (Fig. 3a). The intense and broad photoelectron peak at somewhat above 929.0 eV along with the presence of the characteristic shakeup satellite peaks suggests that the copper oxidation state is +2 in all catalysts [29]. However, the shapes between 929.0 and 938.6 eV of the XPS spectrum are quite different from each other. As shown in Fig. 3b, there has only one peak at 932.7 eV for the sample of Cu/SiO2(90), while another peak centered at 935.0 eV for the Cu/SiO2(20) and Cu/SiO2(10) appeared. Van Der Grift et al. suggested that the Eb of copper phyllosilicate is 2.0 eV higher than copper (II) oxide [30], thus we attributed the latter peak to the formation of copper phyllosilicate. By fitting the peaks, the peak area ratios are estimated to be 1.47 : 1 for Cu/SiO2(20) and 1 : 2.48 for Cu/SiO2(10), respectively. The variation of the Cu 2p3/2 Eb values is an indication of the different chemical environments of the copper species in calcined Cu/SiO2 samples. Results above showed that small carrier particles induced formation of more copper phyllosilicate species hence stronger metal–support interaction. The different copper-silica interactions were also demonstrated by the Raman spectra of fresh calcined

930

932

934 936 Binding energy, eV

Fig. 3. XPS spectra of fresh calcined catalysts Cu/SiO2(x), x = 10 (a), 20 (b), and 90 nm (c).

catalysts and bulk CuO which compiled in Fig. 4. The bulk CuO presented three characteristic bands at 290, 338, and 625 cm–1. However, these characteristic bands were well preserved on the Cu/SiO2 catalysts but underwent obvious shift toward lower wavenumbers with the decreasing of the carrier particle size. Specifically, the vibration band in Ag mode of CuO ranged from 290 cm–1 for Cu/SiO2(90) to 278 cm–1 for Cu/SiO2(10), reflecting a strong metal-support interaction derived from the formation of copper phyllosilicate on catalysts surface [31]. This metal-support interaction, however, stabilizes the copper species in hydrogenolysis reaction. In order to investigate the influence of carrier particle size on the reducibility of the copper species on Cu/SiO2 catalysts. Temperature-programmed reduction (TPR) measurements were carried out and reported in Fig. 5. As can be seen, all of the catalyst samples displayed a narrow and almost symmetrical hydrogen consumption peak, indicating that the particle size distribution was homogeneous. The reduction temperatures lower than for bulk CuO indicate good dispersion of copper species on the surface. However, the peak temperature increases gradually with

RUSSIAN JOURNAL OF PHYSICAL CHEMISTRY A

Vol. 92

No. 3

2018

EFFECT OF SILICA PARTICLE SIZE ON TEXTURE, STRUCTURE

RUSSIAN JOURNAL OF PHYSICAL CHEMISTRY A

c b a 200

400

600

800 ν, cm−1

Intensity, a.u.

Fig. 4. Raman spectra of (a) bulk CuO and (b–d) fresh Cu/SiO2(x) catalysts, x = 90 (b), 20 (c), and 10 nm (d).

Catalytic Performance of the Catalysts

d c b a 200

250

300

T, °C

80 60

80

Cu/SiO2(10) Cu/SiO2(20) Cu/SiO2(90)

60

40

40 Cu/SiO2(10) Cu/SiO2(20)

20

Cu/SiO2(90)

4

8 12 16 Reaction time, h

20

1,2-PDO selectivity, %

Fig. 5. Typical H2-TPR profiles: (a) bulk CuO and (b–d) Cu/SiO2(x) catalysts, x = 10 (b), 20 (c), and 90 nm (d).

Conversion, %

The catalysts were tested at 493 K and 6 MPa hydrogen pressure using an aqueous solution of glycerol (40 wt %) as the reactant for various reaction times. Before testing, the catalysts were pre-reduced at 623 K for 2 h under a stream of 10 vol % H2/N2. Figure 6 summarizes the results of glycerol hydrogenolysis over the Cu/SiO2 catalysts. As can be seen, the 1,2-PDO selectivity was not affected seriously by the silica supports. However, the glycerol conversion depends greatly on the silica supports used. After 20 h reaction, the glycerol conversion reached 44.7 and 34.2% on Cu/SiO2(20) and Cu/SiO2(90) catalysts, respectively, while the Cu/SiO2(10) catalyst presented the highest glycerol conversion of 57.4%. It is well known that the activity of Cu/SiO2 catalyst in glycerol hydrogenolysis reaction depends on the copper dispersion, and the deactivation behavior was attributed mainly to the aggregation of active species [33]. The used catalysts were characterized to explain their behavior in glycerol hydrogenolysis. Figure 7 shows the XRD patterns obtained with the Cu/SiO2 catalysts after 20 h reaction at 493 K, showing that copper particles are mainly in the metallic state together with minor formation of Cu2O [31]. Furthermore, with the carrier particle size decreasing, the intensity of diffraction peaks for Cu0 declined notably, and the peak width broadened gradually. The crystallite sizes of Cu0 were estimated by Scherrer’s equation and shown in Table 2. As can be seen, the Cu crystallite size increases with increase in the carrier particle size and in the order of Cu/SiO2(10) < Cu/SiO2(20) < Cu/SiO2(90). As is well known, hydrogenolysis of

d

Intensity, a.u.

decreasing the carrier particle size. The Cu/SiO2(10) sample exhibits most high reduction temperature of 542 K, while Cu/SiO2(90) catalyst shows about 13 K lower than it. This shift in reduction temperature is clearly unrelated to the copper dispersion, when considering of the above results of dissociative N2O adsorption method. Thus the trend should be ascribed to the different copper-support interactions and formation of copper phyllosilicate due to carrier particle size. The results confirmed that dispersed and surface interacted Cu2+ species are more difficult to reduce, as Guerreiro et al. suggested [32]. It is worth to note that the fresh calcined Cu/SiO2 catalysts presented different colors from each other though they have nearly the same contents of CuO. Differ from the dark gray of Cu/SiO2(20) and light gray of Cu/SiO2(90), the Cu/SiO2(10) exhibited blue color. As Zhu et al. suggested [23], the black color related to CuO species while the blue color due to the partial of formation cupric phyllosilicate on the surface, which accordance with the observation from XPS analyses.

453

20

Fig. 6. Performance of glycerol hydrogenolysis for various catalysts. Reaction conditions: 40% aqueous solution of glycerol 20 g, reduced catalyst 0.5 g, T = 493 K, P = 6.0 MPa.

glycerol proceeds via a sequence of reaction steps [14, 17, 34], namely the initial dehydration of glycerol to acetol followed by hydrogenation of the latter to pro-

Vol. 92

No. 3

2018

454

YE TONG QI et al.

Cu

(a)

Cu2O

c

Cu

Cu

b

Cu2− 2p3/2 c

Intensity, a.u.

Intensity, a.u.

SrO2

b a

a 20

40

60

80 2θ, deg

930

The alterations of the atomic states on the catalyst’s surfaces after hydrogenolysis reaction were also investigated by the XPS measurements. As can be seen from Fig. 8a, in all cases the characteristic shakeup satellite lines of Cu2+ disappeared, indicate that most of the surface copper species became reduced (Cu0 and/or Cu+) [25]. Furthermore, the two overlapping Cu LMM Auger kinetic energy peaks centered at about 917.8 and 915.7 eV were observed in Fig. 8b, shows the coexistence of both Cu0 and Cu+ on the surface. As the peaks at 917.8 eV are more prominent in the three samples, the Cu0 species were the major species for all of the three catalysts. From the above discussion, we can draw a conclusion that different average particle sizes of silica supports seems had no obvious influence on the chemical states of the reduced copper species in the used catalysts as the XPS results showed. But, it induced different aggregation levels of copper species in the reaction. The stability of Cu/SiO2(x) catalysts in hydrogenolysis of glycerol were examined under the conditions described in the experimental section, every catalyst was recycled three times. As can be seen from Fig. 9, the Cu/SiO2(10) catalyst showed excellent recyclability after used three times while the activity of Cu/SiO2(20) decreased slightly after each run. However, the Cu/SiO2(90) catalyst exhibited the worst reusability and lost nearly 50% of its activity after the third time used, which could ascribed to the aggregation of copper particles in the reaction as the XRD results shown.

950 960 Binding energy, eV

(b)

c

Intensity, a.u.

Fig. 7. Typical XRD patterns for used catalysts: (a) Cu/SiO2(10), (b) Cu/SiO2(20), (c) Cu/SiO2(90).

panediol. According to Lemonidou [20], the smaller copper particles more effectively catalyze the first dehydration reaction which is regarded as the rate determining step [35]. Thus the high conversion on Cu/SiO2(10) might be due to the presence of well dispersed Cu particles, while the low glycerol conversion of Cu/SiO2(90) catalyst might be due to the aggregation of Cu particles.

940

b a

905

910

915 920 Kinetic energy, eV

Fig. 8. Cu 2p XPS spectra (a) and Cu LMM Auger electron spectra (b) for used catalysts: (a) Cu/SiO2(10), (b) Cu/SiO2(20), (c) Cu/SiO2(90).

CONCLUSION The influence of carrier particle sizes on the catalysts for hydrogenolysis of glycerol were studied using mono-dispersed silica as models. The texture of catalysts apparently depends on the silica supports. With the decrease in carrier particle size, the pore volume and BET surface area increased significantly. Cu loading may influence the BET surface area in two ways, providing decreased surface area of the SiO2(10) and increased surface area of the SiO2(20) and SiO2(90). However, the carrier particle size seems to have no influence on the copper dispersion in the fresh calcined catalysts. The average copper particle sizes calculated by dissociative N2O adsorption method are all in the range of 1.3–1.5 nm. XPS and Raman spectra show the small carrier particles are in favor of formation of copper phyllosilicate and strong metal-support interaction, which considered as benefit on the stability of copper species in catalytic reaction. In other words, although the carrier particle size has no influence on the dispersion of copper species on the fresh calcined catalysts, it has influence on the stability of copper species under reaction conditions. The small carrier particles prevent the copper species from aggregation providing high catalyst activity and stability. The glycerol conversion mainly

RUSSIAN JOURNAL OF PHYSICAL CHEMISTRY A

Vol. 92

No. 3

2018

EFFECT OF SILICA PARTICLE SIZE ON TEXTURE, STRUCTURE

Glycerol conversion, %

Run 1

Run 2

Run 3

60

40

20

0 Cu/SiO2(10) Cu/SiO2(20) Cu/SiO2(90) Reaction cycles Fig. 9. Reaction cycles of glycerol hydrogenolysis over various catalysts. Reaction conditions: 40% aqueous solution of glycerol 20 g, reduced catalyst 0.5 g, T = 493 K, P = 6.0 MPa.

depends on the copper particle size. The important role of small copper particles on the catalytic performance will be further investigated in our future work. REFERENCES 1. G. W. Huber, S. Iborra, and A. Corma, Chem. Rev. 37, 4044 (2006). 2. T. Q. Ye, L. X.Yuan, Y. Q. Chen, et al., Catal. Lett. 127, 323 (2009). 3. T. Werpy, G. Petersen, A. Aden, et al., Top Value Added Chemicals from Biomass (U.S. Dep. of Energy, 2004), Vol. 1. 4. R. Nie, D. Liang, L. Shen, et al., Appl. Cata.l B: Environ. 127, 212 (2012). 5. L. E. Chinchilla, C. M. Olmos, A. Villa, et al., Catal. Today 253, 178 (2015). 6. S. X. Xia, Z. Yuan, L. N. Wang, et al., Bioresource Tech. 104, 814 (2012). 7. C. C. Wang, H. Jiang, C. L. Chen, et al., Chem. Eng. J. 264, 344 (2015). 8. S. M. D. Rezende, C. A. Franchini, M. L. Dieuzeide, et al., Chem. Eng. J. 272, 108 (2015). 9. T. Pairojpiriyakul, E. Croiset, K. Kiatkittipong, et al., Int. J. Hydrogen Energy 39, 14739 (2014). 10. J. A. Cecilia, C. Garcia-sancho, J. M. Meridarobles, et al., Catal. Today 254, 43 (2015). 11. G. S. Kumae, Y. Wee, I. Lee, et al., Chem. Eng. J. 276, 283 (2015).

RUSSIAN JOURNAL OF PHYSICAL CHEMISTRY A

455

12. A. Corma, S. Iborra, and A. Velty, Chem. Rev. 107, 2411 (2007). 13. S. Kale, S. B. Umbarkar, M. K. Dongare, et al., Appl. Catal. A: Gen. 490, 10 (2015). 14. M. A. Dasari, P. P. Kiatsimkul, W. R. Sutterlin, et al., Appl. Catal. A: Gen. 281, 225 (2005). 15. Y. Kusunoki, T. Miyazawa, K. Kunimori, et al., Catal. Commun. 6, 645 (2005). 16. M. L. Barbell, G. F. Santori, and N. N. Nichio, Bioresource Tech. 111, 500 (2012). 17. T. Miyazawa, S. Koso, K. Kunimori, et al., Appl. Catal. A: Gen. 318, 244 (2007). 18. I. Furikado, T. Miyazawa, S. Koso, et al., Green Chem. 9, 582 (2007). 19. C. Montassier, J. C. Menezo, L. C. Hoang, et al., J. Mol. Catal. 70, 99 (1991). 20. E. S. Vasilidou and A. A. Lemonidou, Appl. Catal. A: Gen. 396, 177 (2011). 21. C. J. Yue, L. P. Gu, Y. Su, et al., React. Kinet. Mech. Catal. 111, 633 (2014). 22. Y. H. Feng, H. B. Yin, L. Q. Shen, et al., Chem. Eng. Tech. 36, 73 (2013). 23. S. H. Zhu, X. Q. Gao, Y, L. Zhu, et al., J. Catal. 303, 70 (2013). 24. A. Bienholz, H. Hofmann, P. Claus, Appl. Catal. A: Gen. 391, 153 (2011). 25. Z. W. Huang, F. Cui, J. J. Xue, et al., Catal. Today 183, 42 (2012). 26. S. X. Xia, R. F. Nie, X. Y. Lu, et al., J. Catal. 296, 1 (2012). 27. J. H. Guo, G. Y. Xu, Z. Han, et al., ACS Sustain Chem. Eng. 2, 2259 (2014). 28. L. Ma, D. H. He, and Z. P. Li, Catal. Commun. 9, 2489 (2008). 29. Z. W. Huang, F. Cui, H. X. Kang, et al., Chem. Mater. 20, 5090 (2008). 30. D. G. C. J. G. Van, P. A. Elberse, A. Muldr, et al., Appl. Catal. 59, 275 (1990). 31. S. H. Zhu, X. Q. Gao, Y. L. Zhu, et al., Catal. Sci. Tech. 5, 1169 (2015). 32. E. D. Guerreiro, O. F. Gorriz, G. Larsen, et al., Appl. Catal. A: Gen. 204, 33 (2000). 33. E. S. Vasiliadou, T. M. Eggenhuisen, P. Munnik, et al., Appl. Catal. B: Environ. 145, 108 (2014). 34. E. P. Maris and R. J. J. Davis, Catal. 249, 328 (2007). 35. Z. M. Zhou, X. Li, T. Y. Zeng, et al., Chin. J. Chem. Eng. 18, 384 (2010).

Vol. 92

No. 3

2018