Silica supported copper and cerium oxide catalysts for ethyl acetate oxidation

Journal of Colloid and Interface Science 404 (2013) 155–160

Contents lists available at SciVerse ScienceDirect

Journal of Colloid and Interface Science www.elsevier.com/locate/jcis

Silica supported copper and cerium oxide catalysts for ethyl acetate oxidation Tanya Tsoncheva a,⇑, Gloria Issa a, Teresa Blasco b, Patricia Concepcion b, Momtchil Dimitrov a, Selene Hernández b, Daniela Kovacheva c, Genoveva Atanasova c, José M. López Nieto b a b c

Institute of Organic Chemistry with Centre of Phytochemistry, Bulgarian Academy of Sciences, Sofia, Bulgaria Instituto de Tecnologia Química (UPV-CSIC), Campus Universidad Politécnica Valencia, 46022 Valencia, Spain Institute of General and Inorganic Chemistry, Bulgarian Academy of Sciences, Sofia, Bulgaria

a r t i c l e

i n f o

Article history: Received 20 February 2013 Accepted 1 May 2013 Available online 14 May 2013 Keywords: Copper ceria catalysts Ethyl acetate oxidation Active sites

a b s t r a c t The formation of active sites in the silica supported copper and cerium oxide bi-component catalysts for total oxidation of ethyl acetate was studied by Nitrogen physisorption, XRD, XPS, UV–Vis, Raman, FTIR of adsorbed CO spectroscopies and TPR. It was found that the interaction between the copper oxide nanoparticles and the supported on the silica ceria ones is realized with the formation of interface layer of penetrated into ceria lattice copper ions in different oxidative state. This type of interaction improves the dispersion of copper oxide particles and provides higher accessibility of the reactants to the copper active sites even at low copper amount. Ó 2013 Elsevier Inc. All rights reserved.

1. Introduction Recently, copper cerium oxide binary system have been intensively studied in various reactions, such as combustion of CO and methane, SO2 reduction, hydrogen production by steam reforming of methanol, reduction of NO, decomposition of H2O2, wet oxidation of phenol, preference CO oxidation in excess of hydrogen, water gas shift reaction, and as additives in the three way catalysts for purification of automobile exhaust gas in order to reduce the cost of noble metals [1–13]. Although copper–cerium materials have been considered as effective oxidation catalysts, their performance in volatile organic compounds (VOCs) oxidation has been scarcely studied. Delimaris and Ioannides [14] and Hu et al. [4] demonstrated an optimum in copper content to achieve excellent catalytic activity in ethanol and ethyl acetate combustion, while no significant effect of copper addition to ceria was observed at low temperatures for toluene oxidation. Up to now, most the investigations were focused on the behavior in bulk, non-supported materials, where copper oxide was doped with ceria or vice versa. It was found that the ‘‘intimate’’ contact between different metal oxide particles depends on the amount of both components as well as on the preparation and activation method used [15,16]. The formation of surface oxygen vacancies and facile Ce4+/Ce3+ and Cu2+/Cu1+ redox transitions as well as the increased copper oxide dispersion has been assumed as crucial points for the unique high catalytic activity of these materials [4–6,8,11]. In our ⇑ Corresponding author. E-mail address: [email protected] (T. Tsoncheva). 0021-9797/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jcis.2013.05.005

previous study, [17] we reported that ceria doping with small amount of copper by co-precipitation of nitrate precursors leads to significant incorporation of copper ions into the ceria lattice, which decreases their reducibility and catalytic activity in VOCs oxidation. It was assumed that the increase in the Cu/Ce ratio promotes the segregation of finely dispersed, easily reducible, and highly active copper oxide crystallites on ceria surface. Despite the large number of papers focused on the optimization of the intimate contact between copper and ceria ([3] and refs. therein), information on the effect of support is negligible [2,11,13]. Rao et al. [2] assumed higher interaction between different oxide species, which provided higher activity in CO oxidation, when they were supported on SiO2 and ZrO2 as compared to the Al2O3 support. We demonstrated that copper ceria interaction could be restricted by the porous structure if mesoporous silica type SBA-15 was used as a support [17]. The aim of the current paper is to study the catalytic behavior of bi-component copper and cerium materials supported on conventional silica, where the role of pore structure is minimized. In order to clarify the formation of the catalytic active sites, data for selected bulk materials were used as reference. 2. Experimental Copper and cerium supported on silica (Cabosil M5) materials with total metal content of 6 wt.% were prepared by impregnation of 1 g of silica with 0.5 M Cu(NO3)22H2O and/or 0.25 M Ce(NO3)36H2O aqueous solution in appropriate ratio, followed by drying at ambient temperature over night. Bulk analogs were obtained by precipitation of the corresponding nitrate precursors with a 2 M

T. Tsoncheva et al. / Journal of Colloid and Interface Science 404 (2013) 155–160

aqueous solution of KOH at a pH of 10.5. The powder samples were calcined at 773 K for 2 h and denoted as mCunCe/SiO2 and mCunCeOx for the supported and bulk materials, respectively, where m/n corresponds to the ratio between the amount of different metals in wt.%. Nitrogen physisorption data were obtained in a Micromeritics ASAP 2000 instrument. Powder X-ray diffraction patterns were collected on a Bruker D8 Advance diffractometer with Cu Ka radiation. The average metal oxide crystallite size was evaluated according the Scherrer equation. The UV–Vis spectra were recorded on a Jasco V-650 UV–Vis spectrophotometer. X-ray photoelectron spectra were collected using a SPECS spectrometer with a 150 MCD-9 detector and using a nonmonochromatic Al Ka (1486.6 eV) X-ray source. Spectra were recorded using analyzer pass energy of 30 eV, an X-ray power of 200 W, and under an operating pressure of 10–9 mbar. During data processing of the XPS spectra, binding energy (BE) values were referenced to C1s peak (284.5 eV). Spectra treatment has been performed using the CASA software. Selected XPS measurements were done in the UHV chamber of ESCALAB-Mk II (VG Scientific) electron spectrometer. Raman spectra were recorded using a Renishaw system 1000, equipped with argon ion laser (532 nm). IR spectra of adsorbed CO were recorded at 97 K with a Nexus 8700 FTIR spectrometer using a DTGS detector and acquiring at 4 cm1 resolution. An IR cell allowing in situ treatments in controlled atmospheres and temperatures from 97 K to 773 K has been connected to a vacuum system with gas dosing facility. For IR studies, the samples were pressed into self-supported wafers and treated at 423 K in vacuum (105 mbar) for 1 h. After activation, the samples were cooled down to 97 K under dynamic vacuum conditions followed by CO dosing at increasing pressure (0.4–8.5 mbar). IR spectra were recorded after each dosage. In order to evaluate the stability of the IR bands, the samples have been submitted to dynamic vacuum conditions after maxima CO dosing and the evolution of the intensity of each IR band has been recorded with time. The TPR/TG (temperature-programed reduction/thermogravimetric) analyses were performed in a Setaram TG92 instrument. Typically, 40 mg of the sample was placed in a microbalance crucible and heated in a flow of 50 vol% H2 in Ar (100 cm3 min1) up to 773 K at 5 K min1 and a final hold-up of 1 h. The weight loss during the reduction was calculated on the base of TPR-TG curves, and the values were normalized to one and the same catalyst weight (40 mg). The oxidation of ethyl acetate was performed in a flow type reactor with 1.21 mol% ethylacetate in air and WHSV of 300 h1, using 0.03 g of catalyst. GC analyses were done on a HP 5890 apparatus using carbon-based calibration. 3. Results and discussion 3.1. Structural characterization of the initial materials XRD patterns of all silica supported materials are presented in Fig. 1. Well defined reflections typical of CeO2 face centered cubic fluorite phase (PDF files 34-0394) with average crystallite size of about 5 nm are registered for Ce/SiO2 (Fig. 1 and Table S1). The broadening of these reflections for all bi-component materials indicates an increase in ceria dispersion (Table S1). The sharp reflections observed for all copper containing materials reveal the presence of CuO phase with monoclinic tenorite structure (PDF 48-1548) and particle size about 20–30 nm. The preservation of ceria unit cell parameters after copper addition (Table S1) does not indicate formation of solid solution [1,11]. In order to verify this assumption, Raman spectra were collected (Fig. 2a). The spectra of all supported copper containing samples present three main peaks at around 275, 330 and 619 cm1, typical of CuO phase [1]. The intense peak at 466 cm1

* CuO o CeO2

o o

o

o

o

o

2Cu4Ce/SiO2

o

o

3Cu3Ce/SiO2

Ce/SiO2

o * *

Intensity,a.u.

156

o

o

*

*

*

* o

4Cu2Ce/SiO2

* * *

10 20

30

40 50

Cu/SiO2 60 70

80 90

2 Theta,deg Fig. 1. XRD patterns of silica supported materials.

in ceria containing materials is assigned to F2g mode of CeO2 fluorite structure [8]. We would like to stress on the broadening and slight red shift of the position of the main CeO2 peak for all bi-component samples. A simultaneous increase in the intensity of the peak at ca. 610 cm1 is also observed in CeO2/SiO2 sample. Generally, this peak is assigned to the formation of oxygen vacancies due to the replacement of Ce4+ by metal ion with different valence [1,8–10]. Accordingly, the observed effects could be assigned to ceria dispersion increase and/or to interaction between different metal oxide nanoparticles with creation of oxygen vacancies. The UV–Vis spectroscopy is sensitive to the changes in metal ions environment and has been used to obtain more information (Fig. 2b). The peak at around 350 nm in Ce/SiO2 spectrum is typical of O2 ? Ce4+ CT transitions [2]. The observed blue shift of this peak for Ce/SiO2 in comparison with bulk CeO2 indicates the higher dispersion of cerium in supported materials. The absorption in the 240–320 nm and 600–800 nm regions in the spectrum of Cu/SiO2 is related to O2 ? Cu2+ CT and d–d transitions, respectively, of crystalline CuO [1]. The observed changes in the 350–500 nm region for all bi-component materials confirm the assumption done above of the existence of strong interaction between metal ions and/or to the increase in metal oxides dispersion. More precise information for the state of metal ions was obtained by FTIR spectra of adsorbed CO (Fig. 3). The IR spectra of CO adsorption on both 4Cu2Ce/SiO2 and 2Cu4Ce/SiO2 samples shows an intense IR band at 2157 cm1 together with bands at 2123 cm1 and 2133–2135 cm1. The band at 2157 cm1 is related to CO interacting with silanol groups in accordance with the corresponding shift in the hydroxyl region [18], whereas the IR bands at 2123 cm1 and 2135 cm1 can be assigned to Ce3+ ions in different coordination environments [19–21], in accordance with the high sensitivity of CO as probe molecule [22]. The latter IR band (2135 cm1) can also be associated with physisorbed CO, but we can neglect this assignation taken into account the high stability of the band during evacuation. Comparing the IR spectra of both 4Cu2Ce/SiO2 and 2Cu4Ce/SiO2 samples, higher amount of Ce3+ ions can be deduced in the 4Cu2Ce/SiO2 sample. On the other hand, a blue asymmetry is observed on the IR band of the 4Cu2Ce/SiO2 sample at low CO coverage (shoulder at 2172 cm1 in spectra c) which is associated with very small amount of Cu+ ions [23]. Thus, we can conclude in the coexistence of reduced copper and cerium species which should be related to an interaction between both metals [5,12].

157

T. Tsoncheva et al. / Journal of Colloid and Interface Science 404 (2013) 155–160

b

a

Ce/SiO2

Ce/SiO2

2Cu4Ce/SiO2

3Cu3Ce/SiO 2

Absorption, a.u.

Intensity, a.u.

2Cu4Ce/SiO2

4Cu2Ce/SiO2

Cu/SiO 2

4Cu2Ce/SiO2 Cu/SiO 2

dash line-bulk samples 500

200

1000

400

Raman shift, cm-1

600

800

Wavelength,nm

Fig. 2. Raman (a) and UV–Vis (b) spectra of copper cerium samples: bulk (dashed lines) SiO2-supported catalyst (solid lines).

2157

Absorbance (a.u.)

Absorbance (a.u.)

2157

2135 2123

2133

2172 f

f

a

a

2200

2150

2100

Wavenumber (cm-1)

2123

2200

2150

2100

Wavenumber (cm-1 )

Fig. 3. FTIR of adsorbed CO at 97 K on selected copper cerium samples: (a) 4Cu2Ce/SiO2; and (b) 2Cu4Ce/SiO2. Prior the experiment, the samples were pre-treated at 423 K in vacuum (105 mbar) for 1 h, cooled down to under dynamic vacuum conditions followed by CO dosing at increasing pressure (0.4–8.5 mbar).

The XPS spectra of the Cu2p3/2 core level are shown in Fig. 4a and Table 1 represents data for the surface composition of catalysts. The shake-up peak at 943 eV observed in all samples is characteristic of the presence of copper ions in the 2+ oxidation state [4,12], while the asymmetry and the shift of the peak maxima of the Cu2p3/2 main peak reveal the contribution of copper ions in an oxidation state bellow 2+ [5,24]. However, the CuL3M45M45 auger peak is not well resolved, making difficult the identification of copper ions in different oxidative state (see Fig. 4b). Nevertheless, we assume the co-existence of Cu2+/Cu+ ions as more realistic for the studied samples. On the other hand, a shift of the Cu binding energy to higher values (+0.4 eV) is observed in the ceria containing samples, which could be associated with either electronic effects or a higher dis-

persion of the copper ions due to the ceria. The deconvolution of the Cu2p3/2 main peak shows the presence of two components with BE 934.6–934.2 eV (related to Cu2+) and 932.8–933.2 eV (related to Cu+). Quantitative data of the XPS analysis reveals a higher contribution of Cu+ species in the 4Cu2Ce/SiO2 sample (Table 1), which agrees with the results obtained from FTIR analyses (see above). Moreover, a well defined tendency of surface copper concentration increase is observed for all bi-component materials, while this tendency is not clearly seen for the cerium ions. The experimental Cu/Cu + Ce ratio is close to the theoretic one for the sample with lower copper content, while it significantly exceeds the expected value for the 4Cu2Ce/SiO2 sample. The Ce3d photoemission spectra for the selected bi-component samples are shown in Fig. 4c. The Ce4+ has been fitted with six

158

T. Tsoncheva et al. / Journal of Colloid and Interface Science 404 (2013) 155–160

b

a

Cu2p3/2

4Cu2Ce/SiO2 ν2

CPS

CPS

CPS

943 2Cu4Ce/SiO2 934.6 933.3

4Cu2Ce/SiO2

c

Ce3d

4Cu2Ce/SiO2

ν1

2Cu4Ce/SiO2

ν0

u0

u1

Cu/SiO2 Cu/SiO2

ν∗0

Cu AES 948

944

940

936

932

910

BE (eV)

915

920

925

930

920

910

KE (eV)

900

890

880

BE (eV)

Fig. 4. XPS spectra of the Cu2p3/2 core level (a), CuL3M45M45 AES auger peak (b) and Ce3d photoemission spectra (only the Ce3d5/2 component has been labeled for clarity) for selected samples. Binding energy (BE) values were referenced to C1s peak (284.5 eV).

Table 1 XPS data for selected copper and cerium oxide supported on SiO2 modifications. Samples

Cu2p3/2 BE (eV)

Ce/SiO2

Cu+/Cu2+ ratio

Cu (at%)

–

0.4

Ce (at%)

Cu/Ce + Cuexp (wt.%)

Cu/Ce + Cutheor (wt.%)

2Cu4Ce/SiO2

934.6 933.3

Cu2+ Cu+

0.3

0.3

0.5

27.2

33.3

4Cu2Ce/SiO2

934.5 933.2

Cu2+ Cu+

2.1

0.4

0.2

92.2

66.7

Cu/SiO2

934.2 932.8

Cu2+ Cu+

0.9

0.2

–

peaks showing binding energies in agreement with the literature [25,26]. m0 = 882.8 eV, m1 = 889.7 eV, m2 = 898.0 eV, for Ce3d5/2 and m00 ¼ 901:4 eV, m01 ¼ 907:5 eV, m02 ¼ 916:5 eV for the Ce3d3/2 component. On the other hand, Ce3+ has been fitted with four components at u1 = 881.2 eV, u0 = 884.7 eV for the Ce3d5/2 and u01 ¼ 898:3 eV and u00 ¼ 903:7 eV for the Ce3d3/2 component. In the 2Cu4Ce/SiO2 sample, a peak labeled as m0 is observed associated with differential charging of the ceria ions, which seems to indicate different environments of cerium in the sample. Quantitative data give a Ce3+/Ce4+ ratio of 0.3 in the 4Cu2Ce/SiO2 sample and 0.1 in the 2Cu4Ce/SiO2 sample, indicating a higher contribution of Ce3+ ions in the former sample, which is in agreement with the FTIR-CO spectra (see above). On the other hand, the surface Ce/ Si ratio is higher for 2Cu4Ce/SiO2 revealing an enrichment of ceria on the surface (Table 1). Fig. 5 presents the TPR-DTG profiles of the samples treated in hydrogen. In case of silica supported materials (Fig. 5a), the reduction of Ce/SiO2 corresponds to about 14% reduction transition of Ce4+ to Ce3+. For Cu/SiO2, the observed DTG effect in this range corresponds to about 88% one step reduction of Cu2+ to Cu0, indicating the presence of copper species in strong interaction with the silica. For the bi-component materials, the main reduction effect is shifted to lower temperatures, but a decrease in the overall reduction degree (calculated on the base of Ce4+ ? Ce3+ and Cu2+ ? Cu0 transitions) to 60–70% is also registered. We assign this observation to the presence of copper ions in different environment: hardly reducible, which are in strong interaction with ceria matrix, probably involved in a metal oxide interface layer; and easily reducible, in highly dispersed copper oxide species [7,17]. Reduction of Cu+ ions is not excluded as well. This assumption is also supported with the reduction behavior of bulk analogs (Fig. 5b), which strongly depend on the Cu/Ce ratio. Here, extremely low

reduction degree (about 45%) combined with DTG shift to higher temperature is observed for the sample with the lowest Cu/Ce ratio, while just the opposite effect is found for 4Cu2Ce. The catalysts have been tested in the ethyl acetate oxidation in the temperature interval 520–700 K. Fig. 6 shows comparatively the variation of ethyl acetate conversion (Fig. 6a) and the selectivity to CO2 (Fig. 6b) with the copper content for both bulk and SiO2 based catalysts at a reaction temperature of 570 K. CO2 is the main reaction product at high reaction temperature and/or high ethyl acetate conversion, whereas ethanol (EtOH), acetaldehyde (AA), and acetic acid (AcAc) have been also detected as by-products at low temperatures and low ethyl acetate conversions (Fig. S1). Note the extremely low activity for both mono-component copper samples. At the same time, for the bulk and supported ceria materials, about 60% ethyl acetate conversion with relatively high selectivity to ethanol is observed. This result is not surprising taking into account that ethyl acetate oxidation is a stepwise process, including hydrolysis to ethanol and acetic acid and their further oxidation via redox mechanism [5]. Obviously, the first step is facilitated by the presence of acid sites, which are in high amount in pure ceria materials [1]. All silica supported bi-component samples reveal higher catalytic activity and CO2 selectivity in comparison with the corresponding mono-component ones. For the bulk samples, this relation is not so simple and 2Cu4Ce exhibits extremely low catalytic activity and CO2 selectivity. Here, a good correlation between the reducibility (Fig. 5) and the catalytic activity and selectivity to CO2 (Fig. 6) is found. The nitrogen physisorption measurements of ceria and copper bulk materials reveal much higher BET surface area for CeO2 (89 m2 g1) compared to CuO (4 m2 g1) and intermediate values for the bi-component bulk materials (35 and 55 m2 g1 for 2Cu4Ce

159

T. Tsoncheva et al. / Journal of Colloid and Interface Science 404 (2013) 155–160

a

b

CeO2 /SiO2

CeO2 2Cu4Ce

2Cu4Ce/SiO2 4Cu2Ce

3Cu3Ce/SiO2

CuO

DTG

DTG

4Cu2Ce/SiO2

CuO/SiO2

400

500

600

400

700

Temperature, K

500

600

700

Temperature, K

Fig. 5. TPR-DTG profiles in hydrogen for supported (a) and bulk (b) materials. 40 mg sample, 100 cm3 min1 flow of H2 in Ar (1:1), heating rate of 5 K min1.

a

100

80

80

60

60

40

40

20

20

20

0

0

0 0

20

40

60

80

100

Cu Content, wt.%

b

100

Specific activity,%/m 2 (*20)

80

CO2 Selectivity, mol%

Conversion,%

100

60

40

0

20

40

60

80

100

Cu Content, wt.%

Fig. 6. Ethyl acetate conversion (a) and CO2 selectivity (b) for bulk (d) and SiO2-supported (j) catalysts. For comparison, the specific activity over bulk catalysts (s) is also included. Experimental conditions: 0.030 g of catalyst (3-fold diluted with crash glass), ethyl acetate (1.21 mol%) in air, WHSV of 300 h1, reaction temperature (570 K), 40 min time on stream.

and 4Cu2Ce, respectively). In order to ignore the effect of surface area, we calculated the specific catalytic activity (SA) for the bulk samples as conversion per unit surface area (Fig. 6a). Surprisingly, it was about 5-fold higher for CuO in comparison with CeO2. So, we could expect that the catalytic behavior of bi-component materials depends mainly on the state of copper species. We would like also to stress that with the exception of 2Cu4Ce sample, the experimental activity of bi-component materials are 2–3 times higher than those calculated considering them as mechanical mixtures of mono-component materials. The XPS, FTIR, and Raman analyses demonstrate that the close contact between different metal oxide particles creates an interface layer, where copper ions are penetrated into the ceria lattice with simultaneous formation of oxygen defects. The experiments reveal that the copper ions from this interface in the bulk materials are hardly reducible (Fig. 5) and probably possess low catalytic activity in ethyl acetate combustion

via Mars van Krevelen mechanism [17]. However, the formation of this interface seems to facilitate the stabilization of finely dispersed, easily reducible, and highly active CuO nanoparticles (Figs. 2 and 5). The mode of metal oxides interaction could be controlled both by the Cu/Ce ratio and by their deposition on the silica support. In bulk materials, ‘‘encapsulation’’ of copper particles into ceria matrix is predominantly realized for the samples with lower copper content, while formation of well stabilized ‘‘shell’’ of copper oxide phase around the finely dispersed ceria particles is more probable with Cu/Ce ratio increase. This leads to different exposure of copper to the reactants and, in the case of 4Cu2Ce, provides higher reducibility and catalytic activity (Fig. 5). The contact between different metal oxide phases leads to the formation of interface layer where copper ions in different oxidative state are partially incorporated into the ceria lattice. The deposition of metal oxides on silica support seems to change the mode of interaction.

160

T. Tsoncheva et al. / Journal of Colloid and Interface Science 404 (2013) 155–160

From physicochemical analyses, we assume formation of several types of metal species: (i) finely dispersed CeO2 particles; (ii) large CuO particles on the silica matrix; (iii) copper ions penetrated into ceria matrix; and (iv) finely dispersed CuO particles in close contact with ceria. The observed clear tendency of surface copper concentrations increase (Table 1) urge the authors to believe that the interaction between metal oxide particles realizes via ‘‘sandwich’’ model of CuO nanoparticles over the supported on the silica ceria ones. The stabilization of these ‘‘sandwich’’ type particles realizes by creation of highly defective interface layer of penetrated into ceria lattice copper ions in different oxidative state. The increase in ceria dispersion, which occurs after deposition on silica support, promotes the formation of these ‘‘sandwich’’ type particles even at low copper content and ensures better reducibility and exposure of the active copper species to the reactants. 4. Conclusion Despite that ceria exhibits its own catalytic activity in ethyl acetate oxidation, the state of copper provides the main effect in the catalytic behavior of bi-component samples. The penetrated in ceria lattice copper ions are hardly reducible and possess low activity in ethyl acetate oxidation. However, they facilitate the stabilization of highly dispersed copper oxide particles over the ceria surface, which improves the catalytic activity. The relative ratio between the copper ions in different environment and oxidative state determines the catalytic behavior which reflects in complicated interpretation of the catalytic behavior of the bi-component materials. The role of silica is predominantly on the improving of the dispersion of the supported on it ceria particles and thus regulating their intimate contact with copper. Acknowledgments The paper is published with the support of Project No. BG051PO001-3.3-05/0001. Financial support of BAS, FNI of Bulgarian Ministry of Education, Project LHTC/Rbnaq01/8, DGICYT in Spain (Project CTQ2012-37925-C03-01) and bilateral project Bulgarian–Spain Inter-academic Exchange Agreement (Project 2009BG0002) are also acknowledged.

Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jcis.2013.05.005. References [1] J.L. Ayastuy, A. Gurbani, M.P. Gonzalez-Marcos, M.A. Gutierrez-Ortiz, Int. J. Hydrogen Energy 35 (2010) 1232–1244. [2] K.N. Rao, P. Bharali, G. Thrimurthulu, B.M. Reddy, Catal. Commun. 11 (2010) 863–866. [3] A. Pintar, J. Batista, S. Hocˇevar, J. Colloid Interface Sci. 307 (2007) 145–157. [4] C. Hu, Q. Zhu, Z. Jiang, L. Chen, R. Wu, Chem. Eng. J. 152 (2009) 583–590. [5] C. Hu, Chem. Eng. J. 168 (2011) 1185–1192. [6] S.M. Saqer, D.I. Kondarides, X.E. Verykios, Appl. Catal. B: Environ. 103 (2011) 275–286. [7] T. Caputo, L. Lisi, R. Pirone, G. Russo, Appl. Catal. A: Gen. 348 (2008) 42–53. [8] J. Chen, Y. Zhan, J. Zhu, C. Chen, X. Lin, Q. Zheng, Appl. Catal. A: Gen. 377 (2010) 121–127. [9] B.Z. Matovic, D.M. Bucevac, M. Rosica, B.M. Babica, Z.D. Dohcevic-Mitrovic, M.B. Radovic, Z.V. Popovic, Ceram. Int. 37 (2011) 3161–3165. [10] A.C. Ferreira, A.M. Ferraria, A.M.B. Rego, A.P. Goncalvesa, A.V. Girãoc, R. Correiad, T.A. Gasche, J.B. Branco, J. Mol. Catal. A: Chem. 320 (2010) 47–55. [11] U. Menon, H. Poelman, V. Bliznuk, V.V. Galvita, D. Poelman, G.B. Marin, J. Catal. 295 (2012) 91–103. [12] J. Zhu, Q. Gao, Z. Chen, Appl. Catal. B: Environ. 81 (2008) 236–243. [13] G. Aguila, F. Gracia, P. Araya, Appl. Catal. A: Gen. 343 (2008) 16–24. [14] D. Delimaris, T. Ioannides, Appl. Catal. B: Environ. 89 (2009) 295–302. [15] S. Hocevar, J. Batista, J. Levec, J. Catal. 184 (1999) 39–48. [16] R. Prasad, G. Rattan, Bull. Chem. React. Eng. Catal. 5 (2010) 7–30. [17] T. Tsoncheva, G. Issa, T. Blasco, M. Dimitrov, M. Popova, S. Hernández, D. Kovacheva, G. Atanasova, J.M. López Nieto, Appl. Catal. A: Gen. 453 (2013) 1– 12. [18] A. Gallo, T. Tsoncheva, M. Marelli, M. Mihaylov, M. Dimitrov, V. Dal Santo, K. Hadjiivanov, Appl. Catal. B: Environ. 126 (2012) 161–171. [19] C. Binet, M. Daturi, J.-C. Lavalley, Catal. Today 50 (1999) 207–225. [20] F. Bozon-Verduraz, A. Bensalem, J. Chem. Soc. Faraday Trans. 90 (1994) 653– 657. [21] P. Concepción, A. Corma, J. Silvestre-Albero, V. Franco, J.Y. Chang-Ching, J. Am. Chem. Soc. 126 (2004) 5523–5532. [22] M.I. Zaki, H. Knozinger, Spectrochim. Acta 43A (1987) 1455–1459. [23] R. Kefirov, A. Penkova, K. Hadjiivanov, S. Dzwigaj, M. Che, Microporous Mesoporous Mater. 116 (2008) 180. [24] G. Moretti, G. Fierro, M. Lo Jacono, P. Porta, Surf. Interface Anal. 14 (1989) 325– 336. [25] F. Larachi, J. Pierre, A. Adnot, A. Bernis, Appl. Surf. Sci. 195 (2002) 236–250. [26] M. Romeo, K. Bak, J. El Fallah, F. Le Normand, L. Hilaire, Surf. Interface Anal. 20 (1993) 508–512.