CeO2-Nb2O5 catalysts for preferential CO oxidation and total combustion of toluene

CeO2-Nb2O5 catalysts for preferential CO oxidation and total combustion of toluene

Applied Catalysis A: General 502 (2015) 129–137 Contents lists available at ScienceDirect Applied Catalysis A: General journal homepage: www.elsevie...

1MB Sizes 0 Downloads 31 Views

Applied Catalysis A: General 502 (2015) 129–137

Contents lists available at ScienceDirect

Applied Catalysis A: General journal homepage: www.elsevier.com/locate/apcata

High performance of Cu/CeO2 -Nb2 O5 catalysts for preferential CO oxidation and total combustion of toluene Erika de Oliveira Jardim, Soledad Rico-Francés, Zinab Abdelouahab-Reddam, Fernando Coloma, Joaquín Silvestre-Albero, Antonio Sepúlveda-Escribano, Enrique V. Ramos-Fernandez ∗ Laboratorio de Materiales Avanzados, Departamento de Química Inorgánica-Instituto Universitario de Materiales, Universidad de Alicante, Ap. 99, E-03080 Alicante, Spain

a r t i c l e

i n f o

Article history: Received 1 April 2015 Received in revised form 20 May 2015 Accepted 29 May 2015 Available online 4 June 2015 Keywords: Niobium addition Cuo-ceo2 CO oxidation Toluene oxidation Copper catalysts

a b s t r a c t Copper-based catalysts supported on niobium-doped ceria have been prepared and tested in the preferential oxidation of CO in excess of H2 (PROX) and in total oxidation of toluene. Supports and catalysts have been characterized by several techniques: N2 adsorption, ICP-OES, XRF, XRD, Raman Spectroscopy, SEM, TEM, H2 -TPR and XPS, and their catalytic performance has been measured in PROX, with an ideal gas mixture (CO, O2 and H2 ) with or without CO2 and H2 O, and in total oxidation of toluene. The effects of the copper loading and the amount of niobium in the supports have been evaluated. Remarkably, the addition of niobia to the catalysts may improve the catalytic performance in total oxidation of toluene. It allows us to prepare cheaper catalysts (niobia it is far cheaper than ceria) with improved catalytic performance. © 2015 Elsevier B.V. All rights reserved.

1. Introduction The preferential oxidation of CO in hydrogen-rich gas mixtures (PROX) has gained importance due to its potential application in fuel cell energy systems. On the other hand, volatile organic compounds (VOCs) are toxic and contribute significantly to the formation of the photochemical smog, which has remarkable impact to the air quality; therefore, the research on the removal of VOCs has attracted increasing interests during the last years. In this way, Cu-based catalysts have generated a lot of interest to be used in PROX and VOCs reactions due to their easy availability, low cost and high activity and selectivity [1–3]. In these cases, among the most studied catalysts for these reactions are those based on the CuOx –CeO2 system. The advantages to use this type of catalysts are associated with the fact that the presence of copper oxide on ceria could promote the oxygen storage capacity (OSC), the thermal stability of cerium oxide [4,5] and its redox properties (between the metal and the support at interfacial sites [1,6–9]). It means that the redox properties of copper oxide, in general, depend on a more or less extent of its interaction with the cerium oxide.

∗ Corresponding author. Tel.: +34 96 590 9350. E-mail address: [email protected] (E.V. Ramos-Fernandez). http://dx.doi.org/10.1016/j.apcata.2015.05.033 0926-860X/© 2015 Elsevier B.V. All rights reserved.

Furthermore, there are some factors that could influence the catalytic properties of these materials, like: (i) the synthesis method, (ii) the calcination temperature and (iii) the metallic loading [10–12]. In this sense, Liu et al. [11], have investigated the influence of different preparation methods for CuO–CeO2 catalysts (co-precipitation, quelate effect, citric acid method and critic phase), and the authors observed that the “quelate effect” method offered the best results. This observation could be related with the formation of defects in the cerium oxide and, consequently, with the synergic effect between the Cu(I)/Cu(II) and Ce(III)/Ce(IV). Martínez-Arias et al. [12] suggested a model for the redox process on copper-ceria catalysts. They proposed that the first step involves the reduction of copper oxide, which might occur at interface positions with the support, favoring the reduction of the latter. In this case, the re-oxidation is generally easier. In a second step, the reduction of copper oxide is extended to ceria positions distant from the interface. In this case, the re-oxidation is takes place firstly on the ceria surface situated far from the interface, while copper oxidation happens next, and it is softened a little bit due to the stabilization of a copper reduced state (Cu(I)) during the redox cycle. One of the main problems of CuOx /CeO2 catalysts is the volatile price of ceria. Thus, prices for rare earths had been flat for most of the 1990s into mid-2000s, but then they have increased signif-

130

E.d.O. Jardim et al. / Applied Catalysis A: General 502 (2015) 129–137

icantly from the early 2010s due to a combination of constraints placed on mining, processing and exporting levels by countries producers (mainly China), and a continued strong global demand for ceria products. Ceria reached a peak level in mid-2011 before declining through early 2012. Still, today the price of ceria doubles the price in 2007. It is clear that catalysts having less content in ceria and maintaining or improving their catalytic activity are needed [13]. Niobium oxide is an abudant and cheap oxide that is mainly produced in Brazil, and its production is totally decoupled from the production of rare earths. Furthermore, niobium oxide possesses several properties that make it promising for catalytic applications. Bearing this in mind, niobia based materials are effective catalysts in selective oxidation reactions due to its redox properties [14]. Furthermore, niobia-doped ceria materials have shown a good tolerance to carbon deposition and excellent properties as solid oxide fuel cell (SOFC) anodes [14,15]. Previously, in our work group, we have studied the influence of Nb addition to Pt/CeO2 catalysts in the PROX reaction [16]. The results have shown that it favors CO oxidation, although the selectivity towards CO2 was still a little low, about 50%. With the idea of preparing catalysts with lower cost and higher activity and selectivity, the aim of this work is to study the influence of copper properties and the addition of niobium oxide to cerium oxide support for two important reactions, the preferential CO oxidation reaction in the presence of hydrogen and the total oxidation of toluene as representative of VOCs removal reactions.

2. Experimental Section 2.1. Preparation of ceria and niobium-doped ceria supports Cerium oxide support was prepared by the homogeneous precipitation method, according with the procedure reported in reference [19]. 13 g of Ce(NO3 )3 · 9H2 O (Aldrich, 99%) and 8 g of urea (Fluka, 98%) were dissolved in 400 ml of ultra-pure water. The mixture was heated at 363 K with constant stirring for 11 h. Finally, between 12 ammonia solution (Panreac, 30%) was added drop-wise to ensure complete precipitation. The solid was separated by filtration, washed with ultrapure water, and finally dried at 383 K overnight. The power sample was calcined in air at 773 K for 4 h, using a rate of 3 K·min−1 . A series of niobium-doped ceria supports was prepared following the same process showed above. An appropriate amount of cerium nitrate (Aldrich, 99%) and ammonium niobate (V) oxalate hydrate (Aldrich, 99.99%), together with urea (Fluka, 98%), were dissolved in 400 ml of ultra-pure water to obtain different supports of composition xCeO2 -(1 − x)Nb2 O5 , with x = 1, 0.9, 0.7 and 0.4. The mixture was heated at 363 K with constant stirring during 11 h. Finally, ammonia solution (Panreac, 30%) was added dropwise to ensure complete precipitation. The solid was separated by filtration, washed with ultrapure water, and finally dried at 383 K overnight. The power samples were calcined in air at 773 K for 4 h, using a rate of 3 K · min−1 .

2.2. Preparation of the CuO catalysts Different Cu/xCeO2 -(1 − x)Nb2 O5 catalysts were prepared by the wetness impregnation method. The supports, xCeO2 -(1 − x)Nb2 O5 (x = 1, 0.9, 0.7 and 0.4), were impregnated with an aqueous solution of Cu(NO3 )2 ·3H2 O (Panreac, 99.999%) with the appropriate concentration to achieve 4 and 8 wt.% Cu loading. Then, catalysts were dried at 383 K overnight and calcined in air at 773 K for 4 h, with a heating rate of 3 K min−1 .

2.3. Materials characterization The textural properties of the supports were characterized by nitrogen adsorption measurements at 77 K, which were performed in a house developed fully automated manometric equipment. Prior to the adsorption experiments, samples were degassed under vacuum (10−4 Pa) at 523 K for 4 h. The BET surface area was estimated after application of the BET equation. The actual metal loading of the different catalysts was determined by ICP in a PerkinElmer device (Optimal 3000). For this purpose, the metal was extracted from the catalysts by refluxing them in aqua regia for 8 h. XRF patterns were performed on a X-ray sequential spectrometer Philips Magix Pro equipped with a rhodium X-ray tube and beryllium window was used. X-Ray powder diffraction patterns were recorded on a Bruker D8-Advance with Göebel mirror and a Kristalloflex K 760–80 F Xray generation system, coupled with a Cu cathode and a Ni filter. Spectra were registered between 20 and 80◦ (2) with a step of 0.05◦ and a time per step of 3 s. Temperature-programmed reduction (TPR-H2 ) with H2 measurements were carried out on calcined catalysts in a U-shaped quartz cell using a 5% H2 /He gas flow of 50 ml min−1 , with a heating rate of 10 K min−1 . Samples were treated with flowing He at 423 K for 1 h before the TPR run. Hydrogen consumption was followed by on-line mass spectrometry and calibrated by carrying out the reduction of CuO and assuming that it is completely reduced to metallic copper. Raman spectra were recorded on a FT-Raman Bruker RFS/100 spectrometer with coupled microscope. Spectra were recorded at room temperature (274 K, 85% relative humidity) between 4000 and 400 cm−1 with a spectral resolution of 4 cm−1 . Raman spectra were obtained using different conditions depending on the type of sample used. In the case of samples with a very intense Raman effect, a laser power of 100 mW and 32 scans was used, whereas for other samples (for example, those with a larger amount of niobium) 200 mW were used and between 128 and 256 scans collected. Calibration was carried out with a silicon single crystal at 520.7 ± 2 cm−1 . The frequency of the exciting Raman source was 1064 nm. TEM images were obtained on a JEOL electron microscope (model JEM-2010) working at 200 kV. It was equipped with an INCA Energy TEM 100 analytical system and a SIS MegaView II camera. Samples for analysis were suspended in ethanol and placed on copper grids with a holey-carbon film support. SEM micrographs were obtained on a scanning electron microscope Hitachi S3000N, which is equipped with a Bruker XFlash 3001 X-ray detector for microanalysis (EDS) and mapping. X-ray photoelectron spectroscopy was performed with a KAlpha spectrometer (Thermo Scientific). All spectra were collected using Al-K␣ radiation (1486.6 eV), monochromatized by a twin crystal monochromator, yielding a focused X-ray spot with a diameter of 400 ␮m, at 3 mA × 12 kV. The alpha hemispherical analyzer was operated in the constant energy mode with survey scan pass energies of 200 eV to measure the whole energy band and 50 eV in a narrow scan to selectively measure the particular elements. Charge compensation was achieved with the system flood gun that provides low energy electrons and low energy argon ions from a single source. The powder samples were pressed into small Inox cylinders, mounted on the sample holder and placed in the vacuum chamber. Before recording the spectrum, the samples were maintained in the analysis chamber until a residual pressure of ca. 5 × 10−7 N m−2 was reached. The quantitative analysis were estimated by calculating the integral of each peak, after subtracting the S-shaped background, and by fitting the experimental curve to a combination of Lorentzian (30%) and Gaussian (70%) lines.

E.d.O. Jardim et al. / Applied Catalysis A: General 502 (2015) 129–137

131

Table 1 Chemical composition and textural characterization of Cu/xCeO2 -(1-x)Nb2 O5 catalysts. Catalyst

SBET , m2 g−1 a

CeO2 , wt.%

Nb2 O5 , wt.%

Nb/Ce Atomic ratio

Cu loading, wt.%

4Cu/0.4CeO2 -0.6Nb2 O5 4Cu/0.7CeO2 -0.3Nb2 O5 4Cu/0.9CeO2 -0.1Nb2 O5 4Cu/CeO2

140 115 85 110

45 72 89 100

55 28 11 0

1.52 0.48 0.15 0

3.9 3.9 3.9 3.9

a

BET surface area for the supports.

2.4. Catalytic behavior PROX reaction was studied in the temperature range 313–473 K at atmospheric pressure, using a reaction mixture containing 20% H2 , 2% CO, 2% O2 , 0–5% CO2 , 0–5% H2 O and He as balance (total flow: 50 ml min−1 and GHSV = 17.000 h−1 ). For the determination of the catalytic behavior, catalysts were placed in a U-shaped quartz reactor. Before any catalytic measurement, the catalysts were oxidized in-situ under flowing oxygen (10%, 50 ml min−1 ) at 313 K (heating rate of 5 K min−1 ). Reaction products were analyzed by online gas chromatography (TCD and FID), using a Plot/Q and a Molesieve capillary columns to separate the reactants and the reaction products. Only CO2 and H2 O were detected as products. The CO conversion (XCO) and the selectivity (SCO2 ) towards CO2 were calculated using the following equations:



XCO =

[CO]in − [CO]out [CO]in

 SCO2 = 0.5



× 100

[CO]in − [CO]out [O2 ]in − [O2 ]out

 × 100

The experiments of total oxidation of toluene were carried out in a U-shape quartz reactor operating in continuous mode, using 150 mg of catalyst diluted in SiC to a total volume of 1.5 ml . Catalytic activity was evaluated in the temperature range 323–523 K under atmospheric pressure, using a reaction mixture containing 1000 ppm of toluene in an air flow of 100 ml min−1 . The reactant mixture was prepared by flowing a part of the air stream through a thermostabilized saturator containing toluene and then mixing with air to obtain a concentration of toluene of 1000 ppm in air flow of 100 ml/min. The toluene vapour pressure in the saturator was calculated from the Antoine equation. The concentration of toluene and the reaction products were analysed with an on-line gas chromatograph (Agilent 6890N) equipped with a flame ionization detector and HP-Plot/Q (30 m × 0.53 mm) column. An on-line IR detector (Sensotran IR) was also employed for the quantitative analysis of CO2 . All gas lines of the reaction system were heated to 383 K in order to avoid toluene and water adsorption and condensation on tube walls. 3. Results and discussion 3.1. Catalyst characterization Table 1 reports the chemical composition of the prepared catalysts and their BET surface areas. It can be seen that the composition of the oxides covers a wide range, from pure cerium oxide to 40 wt.% CeO2 . Samples will be labelled with the nominal composition, which in all cases differs only slightly from the actual values. Additionally, it is important to note that values obtained for copper content were very similar to the nominal ones. It is interesting to observe the influence of the niobium content in the textural properties. The surface area of the pure ceria support is 110 m2 g−1 ; passing through a minimum for a composition x = 0.9 (85 m2 g−1 ) and then it slightly increased with a further increase in the niobium

Fig. 1. X-ray difractograms of the catalysts 4Cu/xCeO2 -(1−x)Nb2 O5 (x = 1, 0.9, 0.7 and 0.4).

content. This increase of surface area can be explained in basis to two factors: (i) the incorporation of niobium into the ceria lattice could favor the formation of a mixed oxide with modified crystallinity, (ii) but it can also be due to a promoting effect of niobia hindering the formation of large ceria nanocrystals. The XRD patterns of Cu/xCeO2 -(1−x)Nb2 O5 catalysts are presented in Fig. 1. The XRD profiles show six diffraction lines corresponding to the fluorite structure of ceria, which are assigned to the (1 1 1), (2 0 0), (2 2 0), (3 1 1), (2 2 2) and (4 0 0) diffraction planes (face centred cubic structure) [17–20]. The niobium incorporation produces both a progressive decrease in the peak intensity and a broadening of the diffraction peaks. This observation may be attributed either to a decrease of the ceria content or, more likely, to a decrease in the crystal size, in close agreement with textural properties described above. Only the characteristic peaks of CeO2 are observed, suggesting the formation of a solid solution or the aggregation of small niobia particles [16]. Only for higher niobium contents (4Cu/0.4CeO2 -0.6Nb2 O5 catalyst) additional peaks attributed to the hexagonal structure of niobium oxide were observed [21–24]. Furthermore, no diffraction peaks for copper species could be identified in the XRD patterns of catalysts (2 = 35.5◦ y 38.7◦ ) [25,26] could be observed. This suggests either the presence of amorphous copper oxide or, more probably, a good superficial dispersion of CuO [27,28], in agreement with the TEM images (Fig. S1); except for catalyst 4Cu/0.4CeO2 -0.6Nb2 O5 , that presented large CuO particles (approximately 100 nm) agglomerated in specific locations on the support’s surface. However, these large particles do not show any peak in XRD, probably because they were composed of amorphous CuO or agglomeration of tiny nanoparticles. The determination of solid solutions formation is far for being trivial, since the crystallinity is not good enough. For that reason other techniques have been used to elucidate this point. The Raman spectra for the supports and the catalyst are shown in Fig. 2(a) and (b), respectively. The Raman spectrum of pure ceria exhibits a main peak at 462 cm−1 , which is assigned to the F2g mode

132

E.d.O. Jardim et al. / Applied Catalysis A: General 502 (2015) 129–137

Fig. 2. Raman spectra of (a) the supports and (b) the Cu/xCeO2 -(1−x)Nb2 O5 (x = 1, 0.9, 0.7 and 0.4) catalysts.

due to symmetrical stretching of Ce O vibrational unit in eightfold coordination [16,29–31]. A decrease of the intensity of this band with an increase of the niobia content was evident for the supports, Fig. 2 (a), and can be due to the framework deformation after niobium incorporation. Furthermore, this decrease is accompanied by a shift of this peak to approximately 463 cm−1 for supports with x = 0.9 and x = 0.7, although no shift was observed for the support with x = 0.4. This shift can be attributed to the lattice distortion due to the incorporation of a smaller cataion such as Nb(V). It can then be concluded that, in supports with low niobia contents, the substitution of some Ce(IV) cations by Nb(V) takes place to some extent for the solids with x = 0.9 and x = 0.7. It can be concluded that no substitution is produced in sample with x = 0.4, since no shift is observed in the Raman spectra. Additionally, it was observed the presence of a band centered between 540 and 600 cm−1 , which is attributed to the LO vibration mode of CeO2 due to the relaxation of its symmetry [32] and the generation of oxygen vacancies [31,33]. Previous studies have shown that while oxide vacancies can be introduced into ceria by doping with oxides of metals with lower oxidation state, oxygen vacancies in CeO2 can be removed by doping with oxides of higher valences, such as Nb(V) [34–36]. In this case, this band is not very clear and, after the amplification of

the spectra, a small band can be observed only on the pure support, what could indicate that the OSC is not favored in these materials. In the case Nb2 O5 content, x = 0.4, a band appears at 706 cm−1 , which is characteristic of the edge sharing vibration of octahedral Nb2 O5 phase. This band, together with the decrease in intensity of the main band, allows us to conclude that Nb2 O5 appears as a segregated phase in this support. With regard to the Raman spectra of catalysts, Fig. 2(b), new bands at about 310 cm−1 and 800 cm−1 appear in the niobium containing catalysts. Some authors [37,38] observed the presence of a band at 831 cm−1 in a CuO/CeO2 catalyst previously treated with oxygen. This band was related with the formation of peroxide species (O2 2− ) in ceria. However, in the case of the catalysts studied in this work, as oxygen pretreatment has not been carried out before the Raman experiments, this possibility was rejected. The peak at 831 cm−1 might be assigned to isolated entities of NbO4 tetrahedra. The peak at 310 cm−1 is ascribed to multiple distortions of isolated octahedron of Nb species [39–41]. These new species point out that calcination after copper incorporation produces a Niobium surface segregation. The degree of segregation is different depending on the niobia content. Thus, for the sample with higher niobia content we could detect polymeric species (Raman peak at

E.d.O. Jardim et al. / Applied Catalysis A: General 502 (2015) 129–137

133

Table 2 XPS results obtained from the spectra. Catalyst

Nb/Ce Atomic ratio

Cu/(Ce + Nb + Cu) Atomic ratio

Cu2p3/2 BE (eV)

Cu LVV Kinetic Energy (eV)

CuShake-up Area /Cu2p3/2Area

Ce(III) %

4Cu/0.4CeO2 -0.6Nb2 O5

4.6

0.08

917

0.24

41.2

4Cu/0.7CeO2 -0.3Nb2 O5

2.1

0.13

917.5

0.61

43.6

4Cu/0.9CeO2 -0.1Nb2 O5

1

0.12

916.8

0.27

37.8

4Cu/CeO2

-

0.09

932.2 934.2 932.1 934 932.6 934.4 932.2 934

916.2

0.18

38.8

Intensity, a.u.

x=0.4 x=0.7

x=0.9

x=1

965 Fig. 3. Reduction profiles at programmed temperature for the copper catalysts Cu/xCeO2 -(1−x)Nb2 O5 (x = 1, 0.9, 0.7 and 0.4).

706 cm−1 ). It means that cluster of Nb2 O5 may be being formed. For the other samples, only monomeric species were detected, so it means that molecularly dispersed niobia is formed at the surface. This surface niobia segregation can be also confirmed by XPS where we see that the Nb/Ce atomic ratio found by XPS is large that the bulk values (Tables 1 and 2). It is also important to remark that the Nb incorporation into the ceria lattice is still taking place for all the samples except for the one with the highest niobia content. Looking at the peak centered at 462 cm−1 we can see the characteristic shifting due to Nb incorporation. At the same time, the characteristic bands of CuO at 290 cm−1 , 340 cm−1 and 630 cm−1 [37,38,42] were not observed, this suggesting a well dispersion of CuO on the catalyst surface, in agreement with the XRD results. Chemical mapping was used to analyze the distribution of cerium and niobium on the surface of supports. (Fig. S2). SEM micrographs show the change in the morphology of supports with increasing of niobium content, i.e. the catalyst shows thin and needle shaped particles. It is also interesting to observe the decrease in the particle size after niobium incorporation, in close agreement with the textural properties described above. Furthermore, from the SEM-chemical mapping analysis of cerium and niobium it was possible to observe a good distribution of both metals in both evaluated materials (catalysts with x = 0.9 and x = 0.7 catalysts). However, in the case of high niobium content (x = 0.4) it was possible to observe areas richer in Nb2 O5 . This fact confirms the results obtained by XRD and Raman suggesting the presence of two segregated oxide phases exists on this support. The pure CeO2 support (Fig. S3) exhibits reduction peaks in experiments at 760 and 1050 K [16], which are attributed to surface and bulk reduction processes, respectively [43–45]. The H2 -TPR profiles of the copper catalysts are shown in Fig. 3. Only features appearing up to a reduction temperature of 650 K are shown, as this is the temperature range where the most interesting peaks appear. It is known that a synergic interaction between copper oxide and

960

955

950

945

940

935

930

925

Binding energy, eV Fig. 4. XPS Cu 2p3/2 spectra of the calcined Cu/xCeO2 -(1−x)Nb2 O5 (x = 1, 0.9, 0.7 and 0.4) catalysts with 4 wt.% Cu.

cerium oxide favors the reduction of the copper species, which are on the cerium surface. Generally, the reduction profile of copper oxide can be associated with different Cu O species: (i) the presence of highly dispersed Cu(II) species which could hardly interact with the support; (ii) weak magnetic interaction between Cu(II) ions which are close; (iii) small bi- and tri-dimensional clusters without regular structure; (iv) big tridimensional aggregates with identical properties than pure copper oxide; (v) copper species with a different oxidation state [10,45,46]. The CuO/CeO2 catalyst presented two overlapping reductions peaks at 487 and 524 K, which is an indication of the existence of more than one copper oxide species in this catalyst. Studies described in literature demonstrate that reduction of pure CuO takes place at 623 K [10]. Pure CeO2 , on the other hand, starts to get reduced at temperatures higher than 653–673 K. Thus, ceria promotes the reduction of copper oxide surface species. In this way, the peak at low temperature (487 K) is related to highly dispersed copper oxides species, whereas the high temperature peak might be assigned to larger copper oxide particles. The addition of Niobia to the support has two effects; first, the low temperature peak shifts toward lower temperature and a second a new peak above 550 K appears. The shifting of the first peak towards lower temperatures might be related to the enhanced interaction of copper species with the doped support. The appearance of a third peak at higher temperatures, is assigned to the reduction of copper oxide species in closed proximity with Niobia. It is well-reported that the reduction of copper species supported on niobia happens at higher temperatures. The surface composition of the catalysts was analyzed by XPS after calcination at 723 K. The X-ray photoelectronic spectra of the Cu 2p3/2 transition for the different catalysts are shown in Fig. 4. The morphology of the spectra is indicative of the surface heterogeneity of these samples. Binding energies higher than 933.5 eV for the Cu 2p3/2 peak and the presence of shake-up peaks are char-

134

E.d.O. Jardim et al. / Applied Catalysis A: General 502 (2015) 129–137

acteristic of CuO, while lower binding energies (around 932.2 eV) and the absence of shake-up peaks are characteristic of Cu2 O or metallic Cu [47]. In this sense, at least two copper species can be observed, Cu(II) and Cu(I). Binding energies around 932.5 and 934 eV and kinetic energies of the LVV Auger transition at around 917 eV are found for all the catalysts (Table 2). So, the presence of both Cu species is widely demonstrated. Furthermore, quantification of each contribution are calculated for spectra in Fig. 4 using Lorentzian–Gaussian (30:70) functions. Taking into account these values, the contributions of shake-up peaks change in the different catalysts. Furthermore, when shake-up peaks in Fig. 4 are analyzed, it can be noted that their contributions change in the different catalysts. This can be quantified by Cu Shake-upArea /Cu2p3/2Area ratio (Table 2); the higher ratio corresponds to the Cu/0.7CeO2 0.7Nb2 O5 catalyst, what is indicative of a higher Cu(II) proportion. On the other hand, Cu dispersion can be estimated through the atomic Cu/(Cu + Ce + Nb) ratio. Again, sample 4Cu/0.7CeO2 0.3Nb2 O5 presents the best Cu surface distribution (Table 2). These results are in agreement with those obtained by H2 -TPR, where the heterogeneity of the surface and a synergic interaction between copper oxide and cerium oxide was observed which favors the reduction of the former species on the cerium-niobium oxide surface. The oxidation state of Ce has been studied through spectra of Ce 3d transition. Values for the Ce(III) percentage presented in each sample surface have been calculated according to the method used in other studies [48–50] (Table 2); the highest Ce(III) content corresponds to the sample with highest Cu(II) content, the 4Cu/0.7CeO2 -0.3Nb2 O5 catalyst. So, this is the catalyst with the highest Cu dispersion, the highest Ce(III) content and the highest Cu(II) content, probably as a consequence of an “optimum” niobium content. The formation of defects in cerium oxide is favored and, consequently, a synergic effect between Cu(I)/Cu(II) and Ce(III)/Ce(IV) is produced. Since the only difference lies in the percentage of niobium oxide, it is obvious to highlight the importance of the presence of this component in the obtained results. In terms of Nb/Ce atomic ratio (Table 2), the catalysts show values much higher than the bulk atomic ratio, 0.12, 0.38 and 1.22 for 4Cu/0.9CeO2 -0.1Nb2 O5 , Cu/0.7CeO2 -0.3Nb2 O5 and 4Cu/0.4CeO2 0.6Nb2 O5 , respectively. This behavior can be explained on the basis of a higher surface content of niobium oxide. It is well reported that the dopant tends to segregate towards the surface during the calcination and reduction treatments [51,52]. The substitution of Nb(V) by Ce(IV) in the niobium oxide network adds a negative charge, which may make up for an oxygen vacancy trapping one electron. However, the substitution of Ce(IV) by Nb(V) in the ceria network brings a positive charge, which may be compensated by the formation of Ce(III) ions. Zhang et al. [53] showed the increase of Ce(III) concentration in a double layer CeO2 /Nb2 O5 thin film. It has been justified that the diffusion of Nb(V) from the CeO2 /Nb2 O5 interphase takes place following this reaction, Eq. (1): Ce4+ O2 + xNb2 5+ O5 → Ce1−2x 4+ Ce2x 3+ Nb2x 5+ O2+4x + x/2O2

(1)

In this case, the Nb(V) ion diffuses into the CeO2 thin film to replace Ce(IV). The same amount of Ce(IV) in CeO2 films will be reduced to Ce(III) to maintain the electroneutrality, whereas the concentration of Ce(III) ions in the CeO2 /Nb2 O5 thin film will be increased compared with the CeO2 thin film. At this point, the characterization results suggest that Nb incorporation in the ceria lattice is favored, and a solid solution was formed for low niobium concentrations. The oxygen content and its chemical environment have been studied considering the values of the binding energies for the O 1 s transition. Contrary to what one might expect, no big differ-

ences have been found in the four samples in this sense. For all catalyst, the O 1 s spectrum can be deconvoluted into three peaks at 529.5 ± 0.2, 531.5 ± 0.2 and 533.0 ± 0.2 eV. The lowest binding energy peak can be associated to lattice oxygen species, O2− [54–56], whereas the peak at intermediates binding energies is assigned to less electron-rich oxygen species. Studies described in the literature have assigned the bands appearing at intermediates O2 2− /O1− [57,58], or related to hydroxyl species, −OH [59] and probably also from carbonate species, CO3 2− [60] (the C 1 s spectra also showed a small component at 289.0 eV). All these species have been deconvoluted in a single peak, since the difference in binding energies is very small and all these species are prompt to participate in the reaction. Finally, the component at the highest binding energies is assigned to adsorbed molecular water [59,61]. The total oxygen surface percentages are more or less the same in all the samples (42.6 ± 0.5%), and there are no large differences or trends between binding energies for O1 s and OLAT , Osurf and OADS percentages in the four catalysts involved. The opposite was the case with the platinum catalysts prepared with the same supports [16]. In this series, a displacement of the O 1 s bands to higher binding energies with the increase of the niobium content could be observed, but not in the Cu catalysts used in this paper, this suggesting an effect of copper interacting strongly with Ce-Nb oxide mixed. 3.2. Catalytic activity 3.2.1. Influence of the CeO2 /Nb2 O5 ratio on the CO oxidation performance With the aim of understanding of the effect of Nb addition to the support, the series of catalysts on supported on the xCeO2 (1−x)Nb2 O5 materials was tested in the preferential oxidation of CO. The catalytic activity and selectivity results are shown in Fig. 5(a) and (b), respectively. As it can be appreciated, the catalysts present a high degree of CO conversion irrespective of the Ce/Nb ratio, with values comparable with those achieved with the pure ceria-supported catalyst. Practically all the catalysts present 100% CO conversion with selectivity values above 96% at 413 K. A loss in selectivity with the increase of the reaction temperature is clearly observed, which means that at high reaction temperatures hydrogen oxidation is favored. Zou et al. [3] observed that the pre-treatment of the catalysts with 20% O2 favors CO conversion, due to the formation of Cu(I)/Cu(II) and CuO activated species on the catalyst surface. Martínez-Arias et al. [12] established a model of the redox processes occurring on CuO/CeO2 catalysts, in which the main active sites ar located at the CuOx -CeOx interface; the first reduced species is copper oxide at interface positions, which promotes ceria reduction at the same area. Then, the reduction of the copper oxide component is extended to ceria positions far away from the interface. Finally, the re-oxidation continues from these latter positions toward the interfacial ones. Depending on the reduction degree, the complete re-oxidation up to Cu(II) can be difficult; if this is the case, copper is stabilized in the Cu(I) state. The same authors [62] showed the presence of the Cu(I)–CO carbonyl band by using DRIFTS in experiments under CO-PROX conditions. This band might be correlated with the CO oxidation activity and provided an estimation of the reduction degree of dispersed oxide species under reaction conditions. In our case, all the copper catalysts supported over niobiumcerium oxide show the same CO conversion profile, even when the support is pure ceria. This trend is in accordance with the characterization results obtained with H2 -TPR, XRD, TEM and Raman spectroscopy, but is not with XPS results. In fact, it is interesting to note the case of 4Cu/0.4CeO2 -0.6Nb2 O5 , where a different profile in the selectivity to CO2 formation with low CO2 content has been observed. The catalyst was less selective at high temperatures, and

E.d.O. Jardim et al. / Applied Catalysis A: General 502 (2015) 129–137

100

135

100

80

80

60

60

40

4Cu/0.4CeO2-0.6Nb2O5

X CO, %

XCO, %

(a)

4Cu/0.7CeO2-0.3Nb2O5

20

40

Ideal model CO2

20

4Cu/0.9CeO2-0.1Nb2O5

H2O

4Cu/CeO2 0 300

320

340

360

380

400

420

CO2 y H2O 440

460

0 300

480

320

340

360

Temperature, K

380

400

420

440

460

480

Temperature, K 100

100

(b) 80

60

S CO2, %

SCO2, %

80

40

40

Ideal model CO2

4Cu/0.4CeO2-0.6Nb2O5 4Cu/0.7CeO2-0.3Nb2O5

20

20

H2O

4Cu/0.9CeO2-0.1Nb2O5 4Cu/CeO2

0 300

60

320

340

360

380

400

420

440

460

480

Temperature, K Fig. 5. CO Conversion (a) and CO2 Selectivity (b) for the serie 4Cu/xCeO2 (1−x)Nb2 O5 (x = 1, 0.9, 0.7 and 0.4), feed composition 20% H2 , 2% CO, 2% O2 and He as balance.

this can be related with the presence of segregated niobium oxide, as it was observed in the XRD studies, Raman spectroscopy and SEM images. However, the surface heterogeneity observed by XPS is not reflected in the catalytic results. This high activity is related with the Cu(I)/Cu(II) ratio, but although this ratio is different for the four samples, no differences in activity have been found, probably because the values become more alike under reaction conditions. It is well known that CuOx /CeO2 system acts as a redox buffer for PROX reaction where the Cu(I)/Cu(II) and Ce(IV)/Ce(III) pairs work synergistically. It is logical to think that the atomic ratio of these redox pairs gets accommodate to the feed composition. In any case, the very good performance ceria-supported copper catalysts in the PROX reaction has been largely demonstrated [10]. In the present work, at least the same very good results in terms of activity and selectivity have been obtained for catalysts in which some portion of ceria has been replaced by cheaper niobium oxide. So, a significant lowering of the catalysts prizes has been achieved. 3.2.2. Influence of the presence of CO2 and H2 O in the feed of CO oxidation performance. The majority of the studies about CO preferential oxidation reaction in the presence of H2 have been performed using gas mixtures consisting of CO, O2 , H2 and an inert component, He or N2 [27,63–68]. Nevertheless, this model does not simulate completely the actual conditions of a reforming stream. Additionally to the

0 300

CO2 y H2O 320

340

360

380

400

420

440

460

480

Temperature, K Fig. 6. CO conversion (a) and CO2 selectivity (b) in different feed conditions for the Cu/0.7CeO2 -0.3Nb2 O5 catalyst, feed composition 20% H2 , 2% CO, 2% O2 , 0–5% CO2 , 0–5% H2 O and He as balance.

mentioned gases, there is a certain amount of CO2 and H2 O present in the feed composition. With this in mind, the effect of the CO2 and H2 O addition on the feed composition for the Cu/0.7CeO2 -0.3Nb2 O5 catalyst was studied. The results of CO conversion and selectivity towards CO2 are shown in the Fig. 6(a) and (b). As it can be seen, the presence of CO2 in the feed stream produced generated a decrease in catalytic activity and CO2 selectivity when it is compared with those obtained under ideal feed conditions. This negative effect may be due to the competitive adsorption of CO (and H2 ) and CO2 on the catalyst surface. The same behavior has been reported in the literature for some gold and copper catalysts [27,63,65–69]. Zou et al. [68] studied the CO conversion in CuZrCeO catalysts using DRIFTS. The authors observed that CO2 present in the feed composition blocks the adsorption sites for CO and H2 , and decreases the CO conversion ability. Marino et al. [25] have also proposed that the decrease in CO conversion and selectivity after the addition of CO2 may be related to the competitive adsorption of CO2 in active sites of copper and/or to the inhibition of oxygen mobility by the formation superficial carbonates on the cerium oxide support. Gamarra et al. [70] also observed, using operando-DRIFTS, the deactivation of CuO–CeO2 catalysts caused by the presence of CO2 in the feed composition. According to these authors, this fact was related to the formation of superficial carbonates and, as a consequence, the redox properties of the catalyst were reduced. Water addition, in the feed stream, affects negatively the catalytic activity. It has been observed in the present work that the

136

E.d.O. Jardim et al. / Applied Catalysis A: General 502 (2015) 129–137

100

X TOLUENE, %

80

60

4Cu/0.4CeO2-0.6Nb2O5 4Cu/0.7CeO2-0.3Nb2O5

40

4Cu/0.9CeO2-0.1Nb2O5 4Cu/CeO2

20

0 300

350

400

450

500

550

600

650

Temperature, K Fig. 7. Toluene conversion as a function of reaction temperature for the Cu/xCeO2 (1−x) Nb2 O5 catalysts.

presence of water produces a decrease in the catalytic activity of Cu/0.7CeO2 -0.3Nb2 O5 . This deactivation can be related to a blockage on the active surface sites by water molecules, in such a way that the access of reactant molecules is limited [66,70,71]. In this sense, Park et al. [69] have proposed that water block the active sites of CuO–CeO2 /Al2 O3 catalysts at low temperatures. In the case of the simultaneous presence of CO2 and H2 O in the feed stream, a negative effect in the catalytic activity is observed. Again, it can be assigned to their adsorption on the active sites, blocking then to the reactant molecules. Interestingly, the decrease of the catalytic activity of Cu/0.7CeO2 -0.3Nb2 O5 is similar to that which occurs when only CO2 is added. Thus, it can be possible that water does not adversely affect the catalytic activity in the presence of CO2 . Ayastuy et al. [64] studied Pt/CeO2 catalysts and founded that the activity was greatly affected by the presence of both CO2 and H2 O in the system, and this effect depended on the temperature used. At low temperatures the catalytic activity decreased, but at higher temperatures the effect was negligible. Njagi et al. [2] observed a decrease of the catalytic activity of copper and manganese oxides catalysts, at temperatures below 323 K, due to the adsorption of these molecules on the active sites. However, a competition for the active sites between the CO and these gases was observed at higher temperatures (more than 348 K). Gamarra et al. [70] have suggested that the negative effect of CO2 and H2 O in the feed is due to formation of specific carbonates in the interfacial sites and the blockage of active sites by adsorbed water molecules. Monyanon et al. [66] also found a negative effect of the presence of CO2 and H2 O in the activity, especially at low temperatures. 3.2.3. Influence of the CeO2 /Nb2 O5 ratio on the total oxidation of toluene The evolution of toluene conversion versus the reaction temperature for the catalysts is shown in Fig. 7. The reaction products were only CO2 and water. Large differences in catalytic performance have been found for the studied catalysts. In this way, catalysts CuOx /CeO2 and Cu/0.4CeO2 -0.6Nb2 O5 present almost no activity at reaction temperatures lower than 498 K. However, Cu/0.9CeO2 0.1Nb2 O5 and Cu/0.7CeO2 -0.3Nb2 O5 are catalytically active from 323 K (the lowest reaction temperature tested). This behavior could be related to the Cu(II)/Cu(I) ratio. Thus, catalysts which have higher Cu(II) proportion are those that present better catalytic results at lower temperatures. This, together with the greater dispersion of copper oxide and a relatively high percentage of Ce(III), makes Cu/0.9CeO2 -0.1Nb2 O5 and Cu/0.7CeO2 -0.3Nb2 O5 the most suitable

catalysts to work in the low temperature range, probably due to the increased interaction of copper with the surface of the Ce Nb mixed oxide. By increasing the reaction temperature above 550 K the catalysts exhibit a similar behavior, yielding practically the same conversion values, probably because the Cu(II)/Cu(I) ratio is more similar in all cases. Indeed, the XPS spectra of the four samples post-reaction were obtained and, in all cases, the Cu(II) percentage increased to values close to 80% irrespective of the of the amount of Nb thereof (and better dispersion) and highest Ce(III)% values (Fig. S3). It can be seen in Fig. 7 that Cu/0.7CeO2 -0.3Nb2 O5 exhibits the optimum activity at lower temperatures for toluene oxidation. As mentioned above, this performance can be related to different facts: (i) the presence of a CuOx –CeOy optimal interface, (ii) the proper Cu(I)/Cu(II) ratio and, (iii) the optimum amount of Nb at the support. Niobia places an important role for this reaction. Thus, when the loading of Nb is low or medium, Nb promotes the formation of Cu(II) and it is incorporated in the ceria lattice, enhancing in this way the redox properties of the support. Furthermore, niobium oxide is considered an acid oxide and it might lead to a strong adsorption of toluene, producing a synergetic effect between redox and acid sites that improve the catalyst performance in the total oxidation reaction. When the concentration of Nb is too high, niobia forms a segregated phase which is probably not in close contact with the redox couple CuOx /CeO2 , This poor contact or interaction between niobia and ceria makes that the synergetic effect acid – redox site does not happen, and the catalytic activity is not enhanced. 4. Conclusion We have prepared a series of Cu/CeO2 -Nb2 O5 catalysts with different amounts of Nb2 O5 . Their catalytic behavior in the preferential oxidation of CO and the total oxidation of toluene was evaluated. When we look up the results for PROX reaction it was possible to observe that the good redox properties of CuOx /CeO2 were not affected by the presence of niobium oxide. However, when the catalysts were tested in the total oxidation of toluene, niobia strongly improved the catalytic performance of the prepared catalysts. This behavior was ascribed to the synergetic effect of redox (CuOx /CeO2 ) and acid sites (Nb2 O5 ). We have incorporated up to 30 wt.% niobia in the catalysts and the good catalytic performance in PROX has been preserved; furthermore, the catalyst behavior in total oxidation of toluene has been improved. It makes possible to produce cheaper catalysts, with less dependence of ceria supply. Acknowledgments The authors gratefully acknowledge the financial support from MINECO (Spain, Project MAT2010-21147)8/346) and Generalitat Valenciana (PROMETEOII/2014/004). EVRF thanks the MINECO for his Ramón y Cajal Fellow RYC-2012-11427. EOJ thanks the CNPq – Brazil for her grant. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.apcata.2015.05. 033 References [1] A. Martínez-Arias, D. Gamarra, M. Fernández-García, A. Hornés, P. Bera, Z. Koppány, Z. Schay, Catal. Today 143 (2009) 211–217. [2] E.C. Njagi, H.C. Genuino, C.K. Kingondu, C.H. Chen, D. Horvath, S.L. Suib, Int. J. Hydrogen Energ. 36 (2011) 6768–6779. [3] H. Zou, S. Chen, W. Lin, J. Nat. Gas Chem. 17 (2008) 208–211. [4] S. Kacimi, J. Barbier Jr., R. Taha, D. Duprez, Catal. Lett. 22 (1993) 343–350.

E.d.O. Jardim et al. / Applied Catalysis A: General 502 (2015) 129–137 [5] L. Kundakovic, M. Flytzani-Stephanopoulos, Appl. Catal. A Gen. 171 (1998) 13–29. [6] A. Martínez-Arias, M. Fernández-García, J. Soria, J.C. Conesa, J. Catal. 182 (1999) 367–377. [7] A. Martínez-Arias, M. Fernández-García, O. Gálvez, J.M. Coronado, J.A. Anderson, J.C. Conesa, J. Soria, G. Munuera, J. Catal. 195 (2000) 207–216. [8] G. Sedmak, S. Hocevar, J. Levec, J. Catal. 213 (2003) 135–150. [9] J.B. Wang, S.C. Lin, T.J. Huang, Appl. Catal. A Gen. 232 (2002) 107–120. ˜ D. Duprez, Top. Catal. 51 (2008) [10] N. Bion, F. Epron, M. Moreno, F. Marino, 76–88. [11] Z. Liu, R. Zhou, X. Zheng, J. Mol. Catal. A Chem. 267 (2007) 137–142. [12] A. Martínez-Arias, A.B. Hungría, M. Fernández-García, J.C. Conesa, G. Munuera, J. Phys. Chem. B 108 (2004) 17983–17991. [13] 2015, http://www.arultd.com/rare-earths/pricing [14] E. Ramírez-Cabrera, A. Atkinson, D. Chadwick, Solid State Ionics 136–137 (2000) 825–831. [15] K. Yashiro, T. Suzuki, A. Kaimai, H. Matsumoto, Y. Nigara, T. Kawada, J. Mizusaki, J. Sfeir, J. Van herle, Solid State Ionics 175 (2004) 341–344. [16] E.O. Jardim, S. Rico-Francés, F. Coloma, J.A. Anderson, E.V. Ramos-Fernandez, J. Silvestre-Albero, A. Sepúlveda-Escribano, Appl. Catal. A Gen. 492 (2015) 201–211. [17] I. Dobrosz-Gómez, I. Kocemba, J.M. Rynkowski, App. Catal. B Environ. 83 (2008) 240–255. [18] H.S. Roh, H.S. Potdar, K.W. Jun, S. Han, J.W. Kim, Catal. Lett. 93 (2004) 203–207. [19] J. Silvestre-Albero, F. Coloma, A. Sepúlveda-Escribano, F. Rodríguez-Reinoso, Appl. Catal. A Gen. 304 (2006) 159–167. [20] M. Thammachart, V. Meeyoo, T. Risksomboon, S. Osuwan, Catal. Today 68 (2001) 53–61. [21] S.H. Chai, H.P. Wang, Y. Liang, B.Q. Xu, J. Catal. 250 (2007) 342–349. [22] I. Nowak, M. Ziolek, Chem. Rev. 99 (1999) 3603–3624. [23] M. Paulis, M. Martín, D.B. Soria, A. Díaz, J.A. Odriozola, M. Montes, Appl. Catal. A Gen. 180 (1999) 411–420. [24] M. Casapu, A. Bernhard, D. Peitz, M. Mehring, M. Elsener, O. Kröcher, App. Catal. B Environ. 103 (2011) 79–84. ˜ C. Descorme, D. Duprez, App. Catal. B Environ. 58 (2005) 175–183. [25] F. Marino, [26] F.A.C. Garcia, J.C.M. Silva, J.L. de Macedo, J.A. Dias, S.C.L. Dias, G.N.R. Filho, Micropor. Mesopor. Mat. 113 (2008) 562–574. [27] E. Moretti, M. Lenarda, L. Storaro, A. Talon, T. Montanari, G. Busca, E. Rodríguez-Castellón, A. Jiménez-López, M. Turco, G. Bagnasco, R. Frattini, Appl. Catal. A Gen. 335 (2008) 46–55. [28] A. Razeghi, A. Khodadadi, H. Ziaei-Azad, Y. Mortazavi, Chem. Eng. J. 164 (2010) 214–220. [29] R.J. Cava, R.S. Roth, T. Negas, H.S. Parker, D.B. Minor, J. Solid State Chem. 40 (1981) 318–329. [30] B.M. Reddy, A. Khan, Y. Yamada, T. Kobayashi, S. Loridant, J.C. Volta, J. Phys. Chem. B 107 (2003) 11475–11484. [31] Q. Wu, J. Chen, J. Zhang, Fuel Process. Technol. 89 (2008) 993–999. [32] X.M. Lin, L.P. Li, G.S. Li, W.H. Su, Mater. Chem. Phys. 69 (2001) 236–240. [33] S.T. Aruna, V.K. William Grips, V. Ezhil Selvi, K.S. Rajam, J. Appl. Electrochem. 37 (2007) 991–1000. [34] N. Laosiripojana, W. Sutthisripok, P. Kim-Lohsoontorn, S. Assabumrungrat, Int. J. Hydrogen Energy 35 (2010) 6747–6756. [35] M. Mogensen, N.M. Sammes, G.A. Tompsett, Solid State Ionics 129 (2000) 63–94.

137

[36] E. Ramírez-Cabrera, A. Atkinson, D. Chadwick, App. Catal. B Environ. 36 (2002) 193–206. [37] G. Águila, F. Gracia, P. Araya, Appl. Catal. A Gen. 343 (2008) 16–24. [38] J. Guzman, S. Carrettin, J.C. Fierro-Gonzalez, Y. Hao, B.C. Gates, A. Corma, Angew. Chem. Int. Ed. 44 (2005) 4778–4781. [39] J.M. Jehng, I.E. Wachs, J. Raman Spectrosc. 22 (1991) 83–89. [40] L.J. Burcham, J. Datka, I.E. Wachs, J. Phys. Chem. B 103 (1999) 6015–6024. [41] J.M. Jehng, I.E. Wachs, Chem. Mater. 3 (1991) 100–107. [42] C. Hu, Q. Zhu, Z. Jiang, Powder Technol. 194 (2009) 109–114. [43] S. Ricote, G. Jacobs, M. Milling, Y. Ji, P.M. Patterson, B.H. Davis, Appl. Catal. A Gen. 303 (2006) 35–47. [44] P. Fornasiero, G. Balducci, R. Di Monte, J. Kaspar, V. Sergo, G. Gubitosa, A. Ferrero, M. Graziani, J. Catal. 164 (1996) 173–183. [45] S. Xu, D. Sun, H. Liu, X. Wang, X. Yan, Catal. Commun. 12 (2011) 514–518. [46] Z. Wu, H. Zhu, Z. Qin, H. Wang, L. Huang, J. Wang, App. Catal. B Environ. 98 (2010) 204–212. [47] Perkin-Elmer Corporation, 1979. [48] A. Laachir, V. Perrichon, A. Badri, J. Lamotte, E. Catherine, J.C. Lavalley, J. El Fallah, L. Hilaire, F. Le Normand, E. Quemere, G.N. Sauvion, O. Touret, J. Chem. Soc. Faraday Transac. 87 (1991) 1601–1609. [49] M. Daturi, C. Binet, J.C. Lavalley, A. Galtayries, R. Sporken, Phys. Chem. Chem. Phys. 1 (1999) 5717–5724. [50] G. Dong, J. Wang, Y. Gao, S. Chen, Catal. Lett. 58 (1999) 37–41. [51] E.V. Ramos-Fernandez, N.R. Shiju, G. Rothenberg, RSC Adv. 4 (2014) 16456–16463. [52] I. Atribak, A. Bueno-López, A. García-García, J. Mol. Catal. A Chem. 300 (2009) 103–110. [53] Z. Zhang, X. Du, W. Wang, J. Vac. Sci. Technol. A 18 (2000) 2928–2931. [54] A.F. Carley, M.W. Roberts, A.K. Santra, J. Phys. Chem. B 101 (1997) 9978–9983. [55] J.L.G. Fierro, L.G. Tejuca, App. Surf. Sci. 27 (1987) 453–457. [56] G.U. Kulkarni, C.N.R. Rao, M.W. Roberts, J. Phys. Chem. 99 (1995) 3310–3316. [57] L. Richter, S.D. Bader, M.B. Brodsky, Phys. Rev. B 22 (1980) 3059–3064. [58] T. Iwaki, M. Komuro, K. Hirosawa, M. Miura, J. Catal. 39 (1975) 324–333. [59] J.L.G. Fierro, Catal. Today 8 (1990) 153–174. [60] T.H. Fleisch, R.F. Hicks, A.T. Bell, J. Catal. 87 (1984) 398–413. ˜ Appl. Surf. Sci. 31 (1988) [61] L. González Tejuca, A.T. Bell, J.L.G. Fierro, M.A. Pena, 301–316. [62] D. Gamarra, C. Belver, M. Fernández-García, A. Martínez-Arias, J. Am. Chem. Soc. 129 (2007) 12064–12065. [63] G. Avgouropoulos, M. Manzoli, F. Boccuzzi, T. Tabakova, J. Papavasiliou, T. Ioannides, V. Idakiev, J. Catal. 256 (2008) 237–247. [64] J.L. Ayastuy, A. Gil-Rodríguez, M.P. González-Marcos, M.A. Gutiérrez-Ortiz, Int. J. Hydrogen Energy 31 (2006) 2231–2242. [65] O.H. Laguna, E.M. Ngassa, S. Oraá, A. Álvarez, M.I. Domínguez, F. Romero-Sarria, G. Arzamendi, L.M. Gandía, M.A. Centeno, J.A. Odriozola, Catal. Today 180 (2012) 105–110. [66] S. Monyanon, S. Pongstabodee, A. Luengnaruemitchai, J. Power Sources 163 (2006) 547–554. [67] M.M. Schubert, A. Venugopal, M.J. Kahlich, V. Plzak, R.J. Behm, J. Catal. 222 (2004) 32–40. [68] H. Zou, S. Chen, Z. Liu, W. Lin, Int. J. Hydrogen Energy 34 (2009) 9324–9333. [69] J.W. Park, J.H. Jeong, W.L. Yoon, Y.W. Rhee, J. Power Sources 132 (2004) 18–28. [70] D. Gamarra, A. Martínez-Arias, J. Catal. 263 (2009) 189–195. [71] M. Daté, M. Haruta, J. Catal. 201 (2001) 221–224.