Biogas reforming over La-NiMgAl catalysts derived from hydrotalcite-like structure: Influence of calcination temperature

Catalysis Communications 12 (2011) 961–967

Contents lists available at ScienceDirect

Catalysis Communications j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / c a t c o m

Short Communication

Biogas reforming over La-NiMgAl catalysts derived from hydrotalcite-like structure: Influence of calcination temperature A. Serrano-Lotina a,⁎, L. Rodríguez a, G. Muñoz a, A.J. Martin b, M.A. Folgado b, L. Daza a,b a b

Instituto de Catálisis y Petroleoquímica (CSIC), C/Marie Curie 2, Campus Cantoblanco, 28049 Madrid, Spain Ciemat, Av. Complutense 22, 28040 Madrid, Spain

a r t i c l e

i n f o

Article history: Received 23 November 2010 Received in revised form 15 February 2011 Accepted 20 February 2011 Available online 26 February 2011 Keywords: Biogas Reforming Hydrogen Hydrotalcite Lanthanum

a b s t r a c t Hydrotalcite-like compound with general formula [M(II)1 − xM(III)x(OH)2]x+(An−x/n) · mH2O, where An− is the compensation anion, has been used as precursor of active catalysts for biogas reforming. This precursor was calcined at six different temperatures between 250 and 750 °C and the resulting catalysts were tested in order to evaluate the influence of the calcination temperature on the catalytic activity and stability. XRD characterization showed that from 250 °C the hydrotalcite structure is no longer detected, leading to Mg(Ni, Al)O solid solutions, where no peaks related to lanthanum appear. An increase on the calcination temperature increased the grain size and cell parameter value. 50 h-catalytic tests were carried out at 700 °C, CH4:CO2 molar ratio of 1:1 and a mass/feed alimentation ratio (W/F) of 0.4 mg min cm− 3. Used catalysts were characterized by temperature programmed oxidation (TPO), scanning electron microscopy (SEM) and Raman spectroscopy in order to obtain information about coke deposition. Catalytic tests highlighted the great influence of calcination temperature over catalytic activity and stability, having found that, as a general trend, calcination temperatures below 750 °C decrease both the stability and catalytic activity, with the exception of the catalyst calcined at 550 °C, where a higher activity was achieved but with a comparatively low stability. © 2011 Elsevier B.V. All rights reserved.

1. Introduction Biogas is a renewable gas that constitutes an alternative fuel that can reduce fossil fuel dependency and emissions of greenhouse gasses. There are different technologies to convert the chemical energy of biogas into electricity, the most common being gas turbines and internal combustion engines [1]. However, these processes generate pollutants as NOx, SOx and dioxins, while their efficiencies are low, around 35%, and limited by the Carnot cycle [2]. A process commonly used to obtain hydrogen from biogas is the steam reforming [3–5]. It requires the evaporation of great quantities of water, which is an energy demanding step [6] as well as the CO2 removal from the inlet stream. Although a cheaper alternative to this process is CO2 reforming of methane, an important drawback is also present: catalyst deactivation mainly produced by carbon formation. Many different catalysts have been studied so far, indicating that noble metal-based catalysts are less sensitive to coking than those based on Ni [7]. However, considering their high cost and limited availability, Ni-based catalysts seem to be more appropriated. Different supports and promoters have been tested in order to

⁎ Corresponding author. Tel.: + 34 91 5854793; fax: + 34 91 5854760. E-mail address: [email protected] (A. Serrano-Lotina). 1566-7367/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.catcom.2011.02.014

minimize coke formation [8–14]. Hydrotalcites have been studied as precursors which after calcination lead to catalysts with good activity. Since small and highly dispersed particles of the active metal are formed, a hindered sinterization process may be expected. They also show high surface areas and basic properties which improve CO2 chemisorption and consequently the resistance to coke formation [15–17]. In addition, in a previous work [18], after calcination at 750 °C, an increase in stability and a decrease in activity were reported when lanthanum was added. Lanthanides can also strengthen CO2 adsorption on the support, which hinders the formation of deposited carbon via reverse disproportionation [12]. They also favor metal dispersion and the oxycarbonate on the La2O3 support might be considered as a dynamic oxygen pool, favoring coke removal [19]. Different calcination temperatures have been reported in the literature. It is known that high calcination temperatures tend to strengthen the interaction inside Mg(Ni,Al)O solid solution, leading to a considerable amount of nickel diffusing from the catalyst surface to the bulk, where it becomes irreducible and therefore ineffective for catalysis [20]. Lucredio et al. [21] and Daza et al. [22] calcined at 500 °C, while Olafsen et al. [23] calcined at 750 °C. However, PerezLopez et al. [24] reported that calcination temperature had little influence over catalytic activity. The aim of this work is to study whether calcination temperature affects catalytic activity and stability and the subsequent determination of an adequate range of calcination temperatures.

962

A. Serrano-Lotina et al. / Catalysis Communications 12 (2011) 961–967

2. Experimental

3. Results and Discussion

2.1. Precursor and Catalysts Preparation

3.1. Catalysts Characterization

A precursor denoted as LaHT2 where number 2 indicates nominal Mg/Al molar ratio, was prepared by co-precipitation at pH = 8 and 60 °C, under continuous stirring by dropwise adding an aqueous solution containing Mg(NO3)2·6H2 O (Panreac, 98% assay), Al (NO3)3·9H2O (Panreac, 98.0–102.0% assay), Ni(NO3)2·6H2O (Panreac, 99% assay) and La(NO3)2·6H2O (Panreac, 99,0% assay) to a saturated aqueous solution containing NaHCO3 (Panreac, 99.7–100.3% assay) at pH = 8, followed by aging for 90 min at 60 °C. pH was adjusted with a mixture of NaHCO3 and NaOH solution (Riedel-de-Haën, 99% assay). Special care of filtering and precipitate washing was taken in order to eliminate Na+ ions completely. It was dried overnight at 110 °C and calcined at 250, 350, 450, 550, 650 and 750 °C. The calcination was performed at a rate of 5 °C min− 1 and maintained for 2 h. After calcination, catalysts denoted as LaHT2-x were obtained. Letter x represents the calcination temperature. The final measured Mg/Al molar ratio was 1.7 whereas Ni and La contents are 2.1 and 1.1%, respectively.

Ni and La nominal contents measured by ICP-MS are summarized in Table 1. It can be seen that they increase below a calcination temperature of 550 °C, which agrees with TPO experiments over the precursor [18], where almost all H2O and CO2 were eliminated after 550 °C. The higher the amount of H2O and CO2 on the sample is the less the relative nickel content the sample contains. XRD characterization of the calcined catalysts (Fig. 1) showed that from 250 °C, hydrotalcite structure is no longer present, leading to diffraction bands at 37, 43, 63, 75 and 79°. These bands can be identified as MgO phase (JCPDS 00-045-0946), but if cell parameters a are calculated (Fig. 2) values lower than the MgO cell parameter (4.21 Å) are obtained, which can be explained by Mg(Ni,Al)O formation, since cell parameter decreases as Al3+ ionic radius is smaller than Mg2+ and Ni2+ ones. Cell parameter a wasffi calculated by plotting the cell parameter a (calculated as pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2 θ cos2 θ d h2 + k2 + l2 ) vs. cos (the Nelson–Riley factor) for peaks at senθ + θ 43, 63 and 79°. This procedure minimizes error sources like absorption of the X-ray beam, displacement of rotation axis or inaccurate determination of camera constants [25]. As above mentioned, from 250 °C hydrotalcite structure is lost. According to Vaccari et al. [26], hydrotalcite decomposes from 300 °C. Hibino et al. [27] established 350 °C as the decomposition temperature. Constantino et al. [28] reported that at temperatures below 250 °C hydrotalcite structure is retained. These differences among reported results may be a consequence of the differences on thermal stabilities due to the synthesis method or as Constantino et al. [28] proposed, due to the rehydration and reconstitution of hydrotalcite structure. Since we detected this reconstitution process in our laboratory, DRX measurements were carried out immediately after calcination. No peak related to lanthanum can be seen in Fig. 1, probably due to its low concentration (1.1–1.6%) or high dispersion. XRD pattern of LaHT2-250 catalyst shows intermediate features between hydrotalcite and Mg(Ni,Al)O solid solution structures, since hydrotalcite peaks at 11 and 24° are still present, while peaks at 35 and 60° show higher intensities than peak at 43°. From 350 °C, the higher the calcination temperature is, the sharper the peaks are obtained, suggesting an increase in grain size. An increase of cell parameter between 450 and 650 °C is observed (Fig. 2), an approach to pure MgO cell parameter as the calcination temperature rises. This could be a result of Al3+ cations leaving the lattice, since the presence of Al3+ ions in the Ni–Mg–O cubic lattice is expected to reduce this value. However, no presence of possible crystalline phases of Al2O3 was detected. Preliminary results obtained by XPS characterization (not shown), show that surface and bulk chemical compositions differ. The surface is enriched on Al, which supports the hypothesis of Al3+ leaving Ni–Mg–O cubic lattice and migrating to the surface. By TPR experiments (not shown), an appropriate reduction temperature was chosen for each catalyst. TPR showed two reduction peaks, one at high temperature and one below 700 °C. We established the reduction temperature at the end of the first peak, since we have observed that reducing the nickel that is more strongly interacting with the support decreases the catalytic stability. Since the applied

2.2. Catalysts Characterization Catalyst composition was determined after acid digestion by an ICP-MS Elan 6000 Perkin-Elmer Sciex equipped with an autosampler AS 91. X-ray diffraction of the calcined catalysts was performed by an X-ray diffractometer (XPERT-PRO, PANanalytical) using Cu Kα radiation (λ = 0.154 nm). Temperature programmed reduction (TPR) was carried out, though not shown here, in order to establish the appropriate reduction temperature. Carbon deposition was determined by temperature programmed oxidation (TPO) over post-reaction catalysts with a Mettler-Toledo TGA/SDTA 851 thermo-balance and STAR 8.10 software coupled to a mass spectrometer detector Pfeiffer Thermostar. TPO tests were performed between 25 and 950 °C (10 °C min − 1 ) using a mixture of O 2 /N 2 10/ 40 mL N min− 1. CO2 desorption in the mass spectrometer detector (m/z = 44) was used to determine coke gasification temperatures. SEM measurements were performed on a Hitachi S-3000N Scanning Electron Microscope coupled to an INCAx-sight energy-dispersive spectrometer, using a Sputter Caoter SC502 to pre-treat the samples. Raman spectroscopy of used catalysts was performed on a PerkinElmer Raman 400F, using 100 mW laser power of 785 nm excitation, a CCD detector and with a spectral resolution of 4 cm− 1 and a total exposure time of 20 s. 2.3. Catalytic Testing Catalytic tests were carried out in a Microactivity Reference PID Eng&Tech equipment. They were performed in a tubular fixed-bed quartz reactor at 700 °C and at CH4:CO2 molar ratio of 1:1. In order to validate the comparison with previous reported results [18], the same conditions were tested, briefly, a catalyst mass of 40 mg with a particle size between 0.5 and 0.42 mm, chosen to avoid excessive pressure drop, a mass/feed alimentation ratio (W/F) of 0.4 mg min cm− 3 and 50 h duration. The testing protocol includes catalyst heating up to the reduction temperature in N2 (100 mL min− 1), followed by reduction in H2 (100 mL min− 1) for 1 h. LaHT2-250 and LaHT2-350 were reduced at 600 °C, while 650 °C was chosen for the rest. After reduction, the reactor was heated up to 700 °C in N2 (100 mL min− 1) followed by the start of the catalytic test by feeding CH4 and CO2. Reaction products were analyzed with an Agilent chromatograph 6890N connected in line, equipped with a TCD detector and Chromosob 102 and Porapak P5 Q columns. CH4, N2, H2, and CO2 gasses were fed from Praxair gas bottles with a purity of 99.5 for CH4 and 99.999% for the rest.

Table 1 Ni and La nominal contents determined by ICP-MS for hydrotalcite-based catalysts calcined from 250 to 750 °C. Catalyst

Ni nominal content/%

La nominal content/%

LaHT2-250 LaHT2-350 LaHT2-450 LaHT2-550 LaHT2-650 LaHT2-750

2.3 2.7 2.8 3.1 3.0 3.0

1.3 1.5 1.6 1.6 1.6 1.6

A. Serrano-Lotina et al. / Catalysis Communications 12 (2011) 961–967

2θ / o 10

20

30

40

50

60

200

Mg(Ni, Al)O

70

80

90

963

LaHT2-250

LaHT2-350

LaHT2-450

LaHT2-550 LaHT2-750

LaHT2-650

LaHT2-750

Time /h

220 222

0

10

20

0

10

20

30

40

50

30

40

50

50

CH4 conversion /%

LaHT2-650

LaHT2-550

I / a.u

LaHT2-450

LaHT2-350

40

30

20

10 60 55

HT structure 003 006

10

20

012

30

110

40

50

60

CO2 conversion /%

LaHT2-250

LaHT2

70

80

90

50 45 40 35 30 25

2θ / o

20

Fig. 1. XRD patterns of catalysts based on hydrotalcite calcined at temperatures between 250 and 750 °C.

reduction temperatures are lower than the temperature required to reduce all Ni species, Ni reduction is partial.

Time /h Fig. 3. CH4 and CO2 conversion vs. time for catalysts based on hydrotalcite calcined at temperatures between 250 and 750 °C. Test conditions: CO2/CH4 = 1, 700 °C and W/ F = 0.4 mg min cm− 3.

3.2. Catalytic Tests In order to compare the activity of the catalysts, they were tested in the conditions detailed in the Experimental section. In Fig. 3 it can be observed that CO2 conversion (Fig. 3b) was higher than CH4 conversion (Fig. 3a) in all cases. As it has been previously reported [18], this is mainly due to the reverse water–gas-shift reaction, (CO2 + H2 ⇆ H2O + CO). The presence of H2O (6–7%) in the product stream (not shown) confirmed this fact. An activation period of 5 h is detected in all catalysts with the exception of LaHT2-450 and

LaHT2-550. This initial activation period may imply the existence of a strong tendency of those catalysts to form the stable surface structure only under the working reaction, as was observed by Verykios et al. [19]. The highest conversions were obtained for the LaHT2-550 catalyst, followed by LaHT2-650, LaHT2-750, LaHT2-450, LaHT2-350 and finally LaHT2-250. Catalytic activity (Fig. 4) was calculated from these results as the moles of CH4 converted per hour and per Ni nominal content. Note that Ni nominal content is different in each catalyst. This fact explains that LaHT2-450 catalyst was the less active

4.25

a parameter /Å

aMgO 4.20

4.15

4.10

4.05

50

Catalyst activity / molCH4·h-1·gNi

-1

4.30

40

30

20

10

0

4.00 250

350

450

550

650

750

o

Calcination temperature / C Fig. 2. Plot of cell parameter a vs. calcination temperature for catalysts in Fig. 1.

250

350

450

550

650

750

Calcination temperature /oC Fig. 4. Catalytic activity vs. calcination temperature for catalysts in Fig. 3.

964

A. Serrano-Lotina et al. / Catalysis Communications 12 (2011) 961–967

catalyst according to Fig. 4, instead of the third less active as Fig. 3 shows. This is due to lower Ni content of LaHT2-250 and LaHT2-350 than LaHT2-450, showing less methane conversion. This catalytic activity was calculated after the activation period, approximately 10 h after the beginning of the reaction. According to the catalytic results, the influence of calcination temperature over catalytic activity can be asserted. An increase on the calcination temperature may decrease the activity, since a higher interaction inside Mg(Ni, Al)O solid solution is expected. However, it is not the case. Surface characterization experiments, like determination of surface composition, basicity and Ni dispersion are under development, in order to clarify this point. Better stabilities are observed on LaHT2-350, LaHT2-650 and LaHT2750 (Fig. 3). Deactivation rates, kd, (Fig. 5a) were calculated according to Hardiman et al. procedure [29], a useful procedure in order to calculate deactivation rates over coupled reaction–deactivation processes. The activity decay, a, is governed by the formula: − da = kd am , where a is dt XCH4 , being XCH4,0 the methane conversion right after the activation XCH4;0 period, kd is the deactivation rate and m is an experimental exponent. When methane conversion is represented vs. time (Fig. 3), a cuasi lineal decay law is observed (after the initial activation period). It implies that the empirical exponent m is well approximated by 0. The slopes of the evolution of methane conversion in Fig. 3 are therefore kd·XCH4,0, meaning that the division of measured slopes by methane conversions after the activation period yields to the corresponding deactivation rates. Catalytic activities, deactivation rates and coke formation rates are summarized in Table 2. Deactivation rates were higher on LaHT2-250, LaHT2-450 and LaHT2-550. Catalysts with low deactivation rates are desired in order to develop stable catalysts for long term operations. Since catalyst deactivation frequently occurs due to carbon deposition, coke formation rate was calculated and shown in Fig. 5b from TPO postreaction tests. These are average values calculated from the weight loss from thermogravimetric results (not shown), caused by CO2 formation divided by the duration of the test. It can be observed that an increase on the calcination temperature leads to a lower carbon formation. The highest coke formation rate was found on LaHT2-350 catalyst. However, the deactivation rate was negative, indicating that the activation period was still in process. This may suggest that CO2 desorption on this catalyst is not only due to coke deposits gasification but also to CO2 interaction with La or CO2 still present in the structure. Fig. 6 represents typical responses of the CO2 signal (m/z = 44) used for catalysts during TPO experiments. At temperatures under

Table 2 Summary of catalytic activities (Fig. 4), deactivation rates (Fig. 5a) and carbon formation rates results (Fig. 5b) obtained by catalytic tests and TPO experiments. Catalyst

Catalytic activity/ 1 molCH4 h− 1 g− Ni

kd/h− 1

Coke formation 1 −1 rate/gC g− cat h

LaHT2-250 LaHT2-350 LaHT2-450 LaHT2-550 LaHT2-650 LaHT2-750

34.2 29.1 23.9 48.4 34.6 34.6

1 · 10− 3 − 9 · 10− 4 1 · 10− 3 9 · 10− 4 1 · 10− 4 3 · 10− 4

0.05 0.06 0.04 0.03 0.01 0

400 °C, different peaks appear and it may be ascribed either to the desorption of physisorbed CO2, consequence of the basicity of the catalysts [26], or chemisorbed CO2 over La2O3 [12] together with Cα gasification [30]. Cα corresponds to the active carbon species responsible for the formation of synthesis gas. More intense peaks are observed between 400 and 675 °C which correspond to Cβ or filamentous coke. Gasification peaks above 675 °C are associated to Cγ or graphitic coke. Cβ and Cγ are less reactive species, normally responsible for catalyst deactivation since they tend to accumulate on the active phase [30]. Therefore, this coke species are undesirable. On LaHT2-350 (the catalysts with the higher coke formation rate but with the lowest deactivation rate), the less reactive carbon species were also detected which are more prone to provoke catalyst deactivation. However, a remarkable stability during the 50 h test was found (Fig. 3). On LaHT2-550, a peak at 500 °C was identified, which can be indicative of nickel carbide formation during reaction [31]. It may be responsible for catalyst deactivation as active sites are partially blocked. In LaHT2-

Temperature / oC 100

LaHT2-250

400

500

600

700

Cβ

Cα

800

900

Cγ

LaHT2-350

LaHT2-450

a I / a.u

-3

kd/ h

-1

1x10

300

m/z =44(CO2)

-3

2x10

200

0

LaHT2-550

-3

-1x10

-3

-2x10 -1

C formation ratio /gC·gcat ·h

-1

LaHT2-650

0.08

b 0.06 LaHT2-750 0.04 0.02 0.00

100 250

350

450

550

650

750

o

Calcination temperature / C Fig. 5. a) Deactivation rate and b) coke formation ratio vs. calcination temperature for catalysts in Fig. 3.

200

300

400

500

600

700

800

900

o Temperature / C

Fig. 6. MS signal (m/z = 44) in TPO characterization tests (post reaction) for catalysts in Fig. 3. Cα denotes active carbon, leading to the formation of synthesis gas. Cβ is filamentous carbon and Cγ is graphitic carbon.

A. Serrano-Lotina et al. / Catalysis Communications 12 (2011) 961–967

965

Fig. 7. SEM characterization of fresh (left) and used (right) catalysts based on hydrotalcite calcined at temperatures between 250 and 750 °C: a) and b) LaHT2-250, c) and d) LaHT2350, e) and f) LaHT2-450, g) and h) LaHT2-550, i) and j) LaHT2-650, k) and l) LaHT2-750.

966

A. Serrano-Lotina et al. / Catalysis Communications 12 (2011) 961–967

Fig. 7. (continued).

650, the slight asymmetry of the peak at 600 °C must be ascribed to nickel carbide gasification [31]. On LaHT2-250 and LaHT2-450 catalysts, gasification of graphitic carbon is more intense than gasification of the other species, contrary to what happened to the rest of the catalysts. High deactivation rates were observed in both catalysts (10− 3 h− 1 in both cases), which may indicate a correlation between the formation of this coke and catalyst deactivation. It is observed from SEM images (Fig. 7), that the sample in which a higher amount of coke appeared was LaHT2-550 (Fig. 7h), being also the catalyst with the highest deactivation rate (9 · 104 h− 1). None of the mentioned types of coke was observed on LaHT2-350 catalyst (Fig. 7d) and LaHT2-250 (Fig. 7b), despite being the catalysts with the 1 −1 highest coke formation rate (0.05 and 0.06 gC g− , respectively). cat h LaHT2-450 and LaHT2-650 catalysts also showed filamentous carbon (Fig. 7f and j) but in less quantity than LaHT2-550 catalyst. SEM images of LaHT2-750 (Fig. 7l), published previously [18], did not show any type of coke as was expected by its low deactivation and carbon deposition ratio. Raman characterization (Fig. 8) was carried out in order to help clarify contradictory results. Despite some evidence of fluorescence, only in the LaHT2-550 catalyst bands G and D (1590 and 1290 cm− 1, respectively) ascribed to coke, were observed, suggesting that there is very little coke on LaHT2-250 and LaHT2-350 surfaces. G band is ascribed to E2g carbon−carbon in-plane stretching vibration and is labeled ‘G’ for ‘graphite’. D band is ascribed to A1g symmetry and labeled ‘D’ for ‘disorder’ [32,33]. Contradictory results observed by TPO and SEM seem therefore to be due to the presence of CO2 chemisorbed over the catalyst and then interfering on TPO analysis. Further characterization of the carbonaceous species is ongoing. Meanwhile, it can be extracted that calcination 300000

LaHT2-250

280000

I /a.u.

260000 240000

LaHT2-350

200000 1290 cm-1

LaHT2-550

180000

1590 cm-1

160000 1000

1200

1400

1600

υ/ cm-1 Fig. 8. Raman spectroscopy characterization of LaHT2-250, LaHT2-350 and LaHT2-550 used as catalysts.

temperatures below 750 °C decrease both the stability of the catalysts, and, with the exception of LaHT2-550, the catalytic activity. 4. Conclusions A hydrotalcite-like precursor was calcined at different temperatures and characterized by XRD. It is observed that from 250 °C hydrotalcite structure is lost, leading to Mg(Ni,Al)O solid solutions. The increase of the calcination temperature increases grain size and cell parameter. Based on catalytic tests and post-reaction characterization we can assert the great importance of calcination temperature election in order to obtain good catalytic activities and stabilities. 750 °C provides good activity and stability during reaction tests, while temperatures below 750 °C provoke a decrease of the stability of the catalysts and, with the exception of 550 °C, a decrease also on the catalytic activity. Since coke formation determinations showed contradictory results, surface characterization is being performed in order to explain the correlation between calcination temperature and catalytic activity and stability and further characterization of carbonaceous species, especially in LaHT2250 and LaHT-350 catalysts. Acknowledgements Financial support from Comunidad de Madrid (DIVERCEL-CM, S2009/ENE-1475) and CDTI (CENIT 2007-1039) is gratefully acknowledged. The authors thank Dr. M.A. Bañares, R. López-Medina and E. M. Mikolajska from Instituto de Catálisis (CSIC) for their help with the Raman analysis and for their helpful discussions. References [1] S. Coehlo, S. Velazquez, S. Orlando, O. Martins, V. Pecora, F. Abreu, Energy generation by a renewable source — sewage biogas, R106-World Climate & Energy Event, 17-18th November 2006, Rio de Janeiro, RJ, Brasil, 2006. [2] D. Deublin, A. Steinhouser, Biogas from Waste and Renewable Resources. An Introduction, Wiley-Vch, 2008, p. 368. [3] P. Kolbitsch, C. Pfeifer, H. Hofbauer, Fuel 87 (2008) 701. [4] A. Effendi, K. Hellgardt, Z. Zhang, T. Yoshida, Fuel 84 (2005) 869. [5] T. Sato, T. Suzuki, M. Aketa, Y. Ishiyama, K. Mimura, N. Itoh, Chem. Eng. Sci. 65 (2010) 451. [6] N. Muradov, F. Smith, A. T-Raissi, Int. J. Hydrogen Energy 33 (2008) 2023. [7] J.R. Rostrup-Nielsen, Stud. Surf. Sci. Catal. 81 (1994) 25. [8] M.C.J. Bradford, M.A. Vannice, Catal. Rev. –Sci. Eng. 41 (1999) 1. [9] E. Ruckenstein, Y.H. Hu, Appl. Catal. A 133 (1995) 149. [10] S. Wang, G.Q. Lu, Appl. Catal. A 169 (1998) 271. [11] F. Pompeo, N.N. Nichio, M.G. Gonzalez, M. Montes, Catal. Today 107–108 (2005) 856. [12] J.Z. Luo, Z.L. Yu, C.F. Ng, C.T. Au, J. Catal. 194 (2000) 198. [13] M. Benito, S. García, P. Ferreira-Aparicio, L. García Serrano, L. Daza, J. Power Sources 169 (2007) 177. [14] A.F. Fonseca, G. Jerkiewickz, E.M. Assaf, Appl. Catal. A 333 (2007) 90. [15] A. Bhattacharyya, V.W. Chang, D.J. Schumacher, Appl. Clay Sci. 13 (1998) 317. [16] T. Shishido, M. Sukenobu, H. Morioka, R. Furukawa, H. Shirahase, K. Takehira, Catal. Lett. 73 (2001) 1. [17] F. Basile, G. Fornasari, E. Poluzzi, A. Vaccari, Appl. Clay Sci. 13 (1998) 329. [18] A. Serrano-Lotina, L. Rodríguez, G. Muñoz, L. Daza, J. Power Sources 196 (2011) 4404.

A. Serrano-Lotina et al. / Catalysis Communications 12 (2011) 961–967 [19] X.E. Verykios, Int. J. Hydrogen Energy 28 (2003) 1045. [20] F. Arena, F. Frusteri, A. Parmaliana, L. Plysanova, A.N. Shmakov, J. Chem. Soc., Faraday Trans. 92 (1996) 469. [21] A.F. Lucredio, G.T. Filho, E.M. Assaf, Appl. Surf. Sci. 255 (2009) 5851. [22] C.E. Daza, J. Gallego, J.A. Moreno, F. Mondragón, S. Moreno, R. Molina, Catal. Today 133–135 (2008) 357. [23] A. Olafsen, C. Daniel, Y. Schuurman, L.B. Raberg, U. Olsbye, C. Mirodatos, Catal. Today 115 (2006) 179. [24] O.W. Perez-Lopez, A. Senger, N.R. Morcilio, M.A. Lansarin, Appl. Catal. A 303 (2006) 234. [25] J.B. Nelson, D.P. Riley, Proc. Phys. Soc. (London) 57 (1945) 160.

967

[26] F. Cavani, F. Trifirò, A. Vaccari, Catal. Today 11 (1991) 173. [27] T. Hibino, Y. Yamashita, K. Kosuge, A. Tsunashima, Clays Clay Miner. 43 (1995) 427. [28] V.R.L. Constantino, T.J. Pinnavaia, Inorg. Chem. 34 (1995) 883. [29] K.M. Hardiman, F.J. Trujillo, A.A. Adesina, Chem. Eng. Process. 44 (2005) 987–992. [30] J. Guo, H. Lou, X. Zheng, Carbon 45 (2007) 1314. [31] A. Djaidja, S. Libs, A. Kiennemann, A. Barama, Catal. Today 113 (2006) 194. [32] F.D. Green, C.A. Johnson, K.M. Thomas, Fuel 62 (1983) 1013. [33] A. Iglesias-Jueza, A.M. Bealea, K. Maaijena, T.C. Wengb, P. Glatzelb, B.M. Weckhuysen, J. Catal. 276 (2010) 268.