- Email: [email protected]
Promoting effect of small amount of Fe addition onto Co catalyst supported on α-Al2O3 for steam reforming of ethanol
Applied Catalysis A: General 383 (2010) 96–101
Contents lists available at ScienceDirect
Applied Catalysis A: General journal homepage: www.elsevier.com/locate/apcata
Promoting effect of small amount of Fe addition onto Co catalyst supported on ␣-Al2 O3 for steam reforming of ethanol A. Kazama, Y. Sekine ∗ , K. Oyama, M. Matsukata, E. Kikuchi Applied Chemistry, Waseda University, 65-301, Okubo, Shinjuku, Tokyo, 169-8555, Japan
a r t i c l e
i n f o
Article history: Received 27 January 2010 Received in revised form 20 May 2010 Accepted 20 May 2010 Available online 27 May 2010 Keywords: Steam reforming of ethanol Promotion effect of iron Cobalt catalyst
a b s t r a c t We examined the promotion effect of loading small amounts of Fe onto various Co catalysts for steam reforming of ethanol. Among these catalysts, catalysts supported on SrTiO3 and ␣-Al2 O3 showed remarkable effects of iron loading onto cobalt catalyst. The Fe-loaded Co/␣-Al2 O3 showed higher yield of hydrogen, low coke deposition on the catalysts, and high activity for steam reforming of acetaldehyde, which was an intermediate of the steam reforming of ethanol. Characterization of the catalyst revealed that Fe and Co metal coexisted on the catalyst support. Synergetic effects of these two metals were observed. © 2010 Elsevier B.V. All rights reserved.
1. Introduction Recently, biomass has attracted much attention as a renewable energy resource [1,2]. Among many utilizations of biomass, utilization of biomass-derived-ethanol has been conducted actively [3,4]. Ethanol obtained through fermentation of saccharide (sugar) includes much water, so the water must be removed completely before the utilization as a gasoline-substitute. Large amounts of energy are required for rectification of ethanol. A useful process is therefore steam reforming of ethanol [5,6], which is applicable to hydrogen production directly with no rectification; the reaction is endothermic without loss of calorific value. The resultant hydrogen is available for various applications such as fuel cells and hydrogen combustion. For those reasons, the obtained biomassderived hydrogen can be regarded as a high-efficiency, clean energy resource [7–9]. To date, various transition metal catalysts such as Rh, Ni, and Co have been investigated for steam reforming of ethanol. Some reports show that noble metals such as Rh exhibit high activity and high stability [10,11]. Nevertheless, it is undesirable to use noble metals because of their high cost. On the other hand, many reports have described that base metals such as Ni [12–15] and Co [15–25] exhibit high activity for steam reforming of ethanol. The Ni-based catalysts show high ethanol conversion at low temperatures [14], but Ni-based catalysts generate more byproduct CH4 than Co-based catalysts [15]. Therefore, Ni-based catalysts are
deemed unsuitable for aspiring to a high H2 yield. The Co-based catalysts, with lower generation of undesired byproduct CH4 [16], have been researched actively among various transition metal catalysts, but their deactivation by sintering of the catalyst or carbon deposition on Co-based catalysts presents a serious obstacle to their wider use [17,18]. Some investigations have revealed that application of a basic oxide such as MgO [13,18], CeO2 [26], ZnO [19] or perovskite-type oxide [27] is a promising method for suppressing carbon deposition on the catalyst. We have reported that Co-based catalysts supported on perovskite-type oxides (SrTiO3 ) show high activity [20]. Regarding the second metal addition on Co-based catalysts, carbon deposition on catalysts was depressed by sodium addition [22], and Ru addition on Co/␥-Al2 O3 was effective among some noble metals (Pt, Pd, Ru, and Ir) [23]. The Fe supplementation of a Co-based catalyst, Co3 O4 [24] and Co/ZnO [25] has been investigated by other researchers; we also have found that addition of small amounts of Fe to Co/SrTiO3 promotes its catalytic activity [21]. Therefore, we investigated the effects of Fe addition on Co-based catalyst with various supports by controlling many factors and parameters such as reaction temperature and contact time (=W/F), etc. In addition, the physical structure of Fe-loaded Co catalyst was characterized to investigate the interaction of Co and Fe. Moreover, a comparison of the catalytic nature between Fe/Co/␣Al2 O3 and physically mixed catalyst (Fe/␣-Al2 O3 and Co/␣-Al2 O3 ) was conducted.
2. Experimental ∗ Corresponding author. Tel.: +81 3 5286 3114; fax: +81 3 5286 3114. E-mail address: [email protected] (Y. Sekine). 0926-860X/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.apcata.2010.05.031
As catalyst supports, SiO2 (JRC-SIO-5), MgO (Ube 1000A), SrTiO3 (synthesis method explained in Ref. [21]) of perovskite oxide and ␣-
A. Kazama et al. / Applied Catalysis A: General 383 (2010) 96–101
Al2 O3 (JRC-ALO-1; calcined at 1573 K for 3 h) were used. In addition, 5 wt% of Co and/or 0.22 wt% of Fe as an active metal was supported by an impregnation method. As precursors of these supported metals, Co(NO3 )2 ·6H2 O and Fe(C5 H7 O2 )3 were used. Distilled water was used as a solvent to support Co; acetone was used for Fe. First, a mixture of catalyst-support and liquid solution of the precursor was heated on the stirrer and evaporated to dryness. The resultant powder was dried in an oven at 393 K for 20 h in air. Then, it was calcined in a muffle furnace at 823 K for 3 h. Then Fe addition onto Co-based catalysts was conducted using a sequential impregnation method, and 0.22 wt% Fe was supported using the above-referenced method. Activity tests were conducted in a continuous flow reactor with a fixed bed of catalyst under atmospheric pressure. A quartz tube (8 mm inner diameter) was used as a reactor. The catalyst powder was pressed at 60 kN for 15 min, crushed, and sieved to obtain a particle size of 250–500 m. Then the sieved catalyst was charged in the catalyst bed with dilution of SiC to 10-mm height. The temperature inside the reactor was controlled using a heater with a thermocouple in the catalyst bed. The reactor was heated with 10 K min−1 up to the reduction temperature (873 K) in H2 flow (50 mL min−1 ). Then it was held for 1 h. After the reactor was cooled to the reaction temperature with purging by N2 , an aqueous solution of ethanol (steam to carbon ratio: S/C = 5.0) was fed to the reactor using a syringe pump through a vaporizer. The flow rate of ethanol was 7.7 × 10−4 mol min−1 , and W/F was changed between 0.08 and 5.2 g-cat h mol−1 by changing the catalyst amount. As an internal standard gas, N2 (20 mL min−1 ) was fed simultaneously, and 80 mL min−1 of Ar was fed simultaneously to the reactor as a carrier gas. Qualitative and quantitative analyses of gas components were conducted using a gas chromatograph (GC). Then H2 and N2 were analyzed using a thermal conductivity detector (TCD)-GC (GC-8A; Shimadzu Corp.). Subsequently, CO, CO2 and CH4 were analyzed using a flame ionization detector (FID)-GC (GC-8A; Shimadzu Corp.) with a methanizer (Ru/Al2 O3 ). In addition, hydrocarbon compounds (CH4 , C2 and C3), liquid products (such as CH3 CHO) and the unreacted ethanol were analyzed using a FID-GC (GC-14B; Shimadzu Corp.) with a six-way valve. Ethanol conversion and H2 yield were calculated according to Eqs. (1) and (2). The H2 yield would be maximum (200%) from Eq. (2) if all the ethanol was reformed to CO2 and H2 with a water gas shift reaction.
ethanol conversion (%) =
out Fcarbon mole of all product in Fcarbon mole of C
(1)
2 H5 OH
H2 yield (%) =
FHout × 100 2
FCin H 2
5 OH
×3
(2)
The amount of deposited carbon on the catalyst was analyzed using temperature-programmed oxidation (TPO). The catalyst was heated 10 K min−1 up to 1173 K with 10% O2 /N2 gas flow, and the quantities of generated CO and CO2 were measured using an infrared gas analyzer (Shimadzu CGT-7000). Finally, the amount of carbon was calculated using these values. The Co particle size was measured using transmission electron microscopy (TEM, JEM-1011; JEOL). The average diameter of Co particles was measured using TEM and then calculated. The physical structures of Co and Fe on Fe/Co/␣-Al2 O3 were observed using scanning transmission electron microscopy (STEM, HF-2210; Hitachi Ltd.) and energy dispersive X-ray spectrometry (EDX, Genesis4000; EDAX Inc.).
97
3. Results and discussion 3.1. Catalytic activities of various Fe-loaded catalysts We previously reported that catalytic activity over Co/SrTiO3 catalyst was promoted by Fe addition [21]. First, we investigated effects of support among Fe/Co-based catalysts. The ethanol conversion, H2 yield and amount of carbon were compared on Fe/Co-based catalysts with various supports, SiO2 , MgO, SrTiO3 , ␣Al2 O3 . Table 1 presents results of catalytic activity tests and amount of carbon after 185 min of reaction. Here, the optimized amount of Fe loading (=0.22 wt%) was chosen from the previous result over Fe/Co/SrTiO3 catalyst [21]. From Table 1, improvement of H2 yield and suppression of carbon deposition were not observed by Fe addition for catalysts of Co/SiO2 and Co/MgO. On the other hand, the promoting effects of Fe on ethanol conversion and H2 yield, and the suppression of carbon deposition were observed for catalysts of Co/SrTiO3 and Co/␣-Al2 O3 . Although the Co/␣-Al2 O3 catalyst showed less ethanol conversion and H2 yield than Co/SrTiO3 catalyst, Feloaded Co/␣-Al2 O3 catalyst showed higher activity and stability than Fe/Co/SrTiO3 catalyst. From this result, we concluded that Fe addition on Co/␣-Al2 O3 was effective for steam reforming of ethanol. Therefore, effects of Fe addition on Co/␣-Al2 O3 catalyst were considered by changing amounts of Fe loading and the calcination time. Table 2 presents results of activity tests. As shown in Table 2, all Fe-loaded Co/␣-Al2 O3 catalysts showed higher activity and less selectivity to CH3 CHO than that of Co/␣-Al2 O3 catalyst. Some reports have described that steam reforming of ethanol proceeds via CH3 CHO as an intermediate [5,21]. Therefore, it was considered that steam reforming of CH3 CHO was promoted by Fe addition from results of Table 2 and that the H2 yield was improved. Based on these results, we chose 0.22 wt% Fe/5 wt% Co/␣-Al2 O3 (calcined for 3 h after Fe loading) as the best catalyst for additional investigations. 3.2. Selectivity to products over Fe-loaded catalyst To investigate the effective temperature window on these catalysts, we examined the effect of reaction temperature over catalysts of Co/␣-Al2 O3 and Fe/Co/␣-Al2 O3 . Activity tests were conducted at each reaction temperature (753, 783, 823, and 873 K). Fig. 1a presents ethanol conversion and Fig. 1b shows the H2 yield at each reaction temperature. Based on the results portrayed in Fig. 1a and b, ethanol conversion and H2 yield were promoted by Fe addition at each temperature. Furthermore, the improvement of activity by Fe addition was observed dominantly at 783 and 823 K. The apparent activation energy was calculated from Fig. 1a. The value for Co/␣-Al2 O3 catalyst was 67.3 kJ mol−1 ; that for Fe/Co/␣-Al2 O3 catalyst was 53.9 kJ mol−1 . From these values, it was confirmed that the apparent activation energy was reduced by Fe addition. Subsequently, catalytic activity over catalysts of Co/␣-Al2 O3 and Fe/Co/␣-Al2 O3 was compared by changing the contact time (=W/F) to examine the reaction passage over these catalysts. Fig. 2 shows carbon-based yields of products on each W/F after 185 min of reaction. Results confirmed that Fe/Co/␣-Al2 O3 catalyst promoted steam reforming and that larger amounts of CO, CO2 —of course including hydrogen—were obtained than with Co/␣-Al2 O3 catalyst, as shown in Fig. 2. Furthermore, Fe/Co/␣-Al2 O3 catalyst showed a lower yield of the reaction intermediate, CH3 CHO, than with Co/␣-Al2 O3 catalyst by comparison between high W/F. These results also demonstrated that Fe addition promoted the steam reforming activity of CH3 CHO. In addition, the yield of byproducts such as CH4 and carbon over Fe/Co/␣-Al2 O3 catalyst was lower
98
A. Kazama et al. / Applied Catalysis A: General 383 (2010) 96–101
Table 1 Effect of Fe loading on Co catalysts supported on various supports. Support
Conversion/% Co
a
SiO2 MgOb SrTiO3 b ␣-Al2 O3 b
Deposited carbon/mg g-cat−1
H2 yield/% Fe/Co
Co
10 min
185 min
10 min
185 min
77.1 72.4 74.4 66.9
62.3 48.3 71.2 57.9
97.3 60.2 89.4 95.7
70.9 43.6 76.3 89.6
10 min 62 95 106 97
Fe/Co 185 min 63 49 101 92
Co
Fe/Co
10 min
185 min
185 min
185 min
65 75 131 168
54 44 103 149
186 47 331 350
156 41 85 51
Loading amount of Co was 5 wt%, reaction temperature = 823 K, W/F = 2.6, S/C = 5.0. a Fe loading = 0.11 wt%, calcination condition = 673 K, 1 h. b Fe loading = 0.22 wt%, calcination condition = 823 K, 3 h.
Table 2 Effect of Fe loading on Co/␣-Al2 O3 catalysts with various conditions. Catalyst
Calcination condition
5 wt%Co/␣-Al2 O3 0.11 wt%Fe/Co/␣-Al2 O3 0.22 wt%Fe/Co/␣-Al2 O3 0.22 wt%Fe/Co/␣-Al2 O3
– 823 K, 1 h 823 K, 1 h 823 K, 3 h
Ethanol conv./%
H2 yield/%
10 min
185 min
10 min
185 min
CO
CO2
CH4
CH3 CHO
66.9 91.5 93.9 95.7
57.9 90.5 90.0 89.6
97 152 154 168
92 126 124 149
12.2 21.9 24.6 21.7
68.4 68.4 66.7 70.7
4.7 2.8 2.7 2.9
14.3 6.5 5.5 4.5
Selectivity at 10 min./%
Deposited carbon after 185 mina /mg g-cat−1
350–419 63–110 67–124 51–87
Reaction temperature = 823 K, W/F = 2.6, S/C = 5.0. a Carbon amount was measured twice using TPO.
Fig. 1. Effect of reaction temperature on ethanol conversion (a) and H2 yield (b) over each catalyst: W/F = 0.65, S/C = 5.0, reaction at 10 min.
Fig. 2. Product yield over Co/␣-Al2 O3 (a) or Fe/Co/␣-Al2 O3 through 185 min reaction: reaction temperature = 823 K, S/C = 5.0, W/F = 0.08–5.2, reaction time = 185 min.
A. Kazama et al. / Applied Catalysis A: General 383 (2010) 96–101
Fig. 3. Amount of carbon deposited on catalysts at each W/F: reaction temperature = 823 K, S/C = 5.0, reaction time = 185 min.
than that over Co/␣-Al2 O3 catalyst. Fig. 3 shows the amount of deposited carbon over catalyst after 185 min of reaction on each W/F. The amount of carbon over Fe/Co/␣-Al2 O3 catalyst was less than that over Co/␣-Al2 O3 catalyst at high W/F conditions. Furthermore, the amount of carbon over Fe/Co/␣-Al2 O3 catalyst was not increased to exceed W/F = 1.3 g h mol−1 . The Fe addition promoted steam reforming without promoting the formation of byproducts such as carbon and methane. 3.3. Stability of catalysts To compare the stability of these catalysts of Co/␣-Al2 O3 and Fe/Co/␣-Al2 O3 , we conducted activity tests with reduced amount
99
Fig. 4. Catalytic activity over Co/␣-Al2 O3 and Fe/Co/␣-Al2 O3 : reaction temperature = 823 K, W/F = 0.65; S/C = 5.0; 䊉, ethanol conversion on Co/␣-Al2 O3 ; , H2 yield on Co/␣-Al2 O3 ; , ethanol conversion on Fe/Co/␣-Al2 O3 ; ♦, H2 yield on Fe/Co/␣Al2 O3 .
of catalysts. Fig. 4 presents ethanol conversion and H2 yield over these catalysts with time on stream. From Fig. 4, Fe-loaded catalyst showed higher ethanol conversion and H2 yield through 185 min of reaction. The particle size of Co measured using TEM (modal diameter) on Co/␣-Al2 O3 catalyst was 19 nm; that of Fe/Co/␣-Al2 O3 catalyst was 17 nm before the reaction (after calcination). Therefore, the difference of the catalytic activities was not derived from the metallic surface area of active metal (=Co). Consequently, the promotion effect of Fe addition was unrelated with the Co particle size. We were able to observe slight deactivation of the catalyst for these two catalysts as shown in Fig. 4 and Table 2. Regarding the deactivation of the catalysts, we also measured the modal diameter of Co particle over Fe/Co/␣-Al2 O3 catalyst at each step (after calcination, after reduction and after the reaction) by TEM obser-
Fig. 5. STEM image and EDX spectra of Co/␣-Al2 O3 and Fe/Co/␣-Al2 O3 catalyst after reaction.
100
A. Kazama et al. / Applied Catalysis A: General 383 (2010) 96–101
Fig. 6. Catalytic activity over each catalyst: (a) Co/␣-Al2 O3 (reaction temperature = 823 K, W/F = 0.65, S/C = 5.0); (b) Fe/␣-Al2 O3 (reaction temperature = 823 K, W/F = 2.6, S/C = 5.0); (c) physical mixture of Fe/␣-Al2 O3 and Co/␣-Al2 O3 (reaction temperature = 823 K, W/F = 0.65, S/C = 5.0); (d) Fe/Co/␣-Al2 O3 (reaction temperature = 823 K, W/F = 0.65, S/C = 5.0).
vation. The value after calcination was 17 nm; that after reduction was also 17 nm. That after the reaction was 19 nm. We considered that these values were almost the same and that little influence of agglomeration of Co-metal was exerted on the deactivation of the catalyst over Fe/Co/␣-Al2 O3 .
3.4. Structure of Fe-loaded catalyst To evaluate the promotion effect of Fe loading, we investigated the physical structure of Fe-loaded Co/␣-Al2 O3 catalyst using STEM-EDX. In addition, Fe/Co/␣-Al2 O3 catalysts were used after
A. Kazama et al. / Applied Catalysis A: General 383 (2010) 96–101
reduction and after the reaction. Fig. 5 shows the STEM image and EDX spectra of Co/␣-Al2 O3 and Fe/Co/␣-Al2 O3 catalysts after reaction. First, on EDX measurement, the peak of ␣-Al2 O3 or that with a very weak peak of Co, was detected from the points uncovered by particles of Co and Fe (not shown). On the other hand, the peaks of Co and Fe were measured simultaneously from the points of Co particles on Fe/Co/␣-Al2 O3 catalyst as shown in Fig. 5. The calculated weight ratio between Co and Fe was 20–40:1. This value was close to the actual weight ratio (23:1), suggesting that Co and Fe coexisted over ␣-Al2 O3 . Furthermore, from this result, we considered that Fe was supported on Co particles selectively using the sequential impregnation method.
101
reforming of ethanol. Results show that Fe addition on Co/␣-Al2 O3 catalyst was the most effective among these catalysts. Observations indicated that Fe addition promoted steam reforming of CH3 CHO while suppressing formation of byproducts such as CH4 and carbon. Characterization of Fe/Co/␣-Al2 O3 catalyst by STEM-EDX revealed that sequentially impregnated Fe was supported selectively on Co particles. Physically mixed catalyst (mixed Co/␣-Al2 O3 catalyst with Fe/␣-Al2 O3 catalyst) did not show high activity such as that of Fe/Co/␣-Al2 O3 . Therefore, we considered that Fe supported on Co particles affected the promotion of catalytic activity of steam reforming of ethanol and CH3 CHO. References
3.5. Interaction of Co and Fe To investigate the interaction between Co and Fe on catalytic activity, we prepared physically mixed catalyst (mixed Co/␣Al2 O3 and Fe/␣-Al2 O3 ), and compared its catalytic activity to that of Fe/Co/␣-Al2 O3 catalyst. This physically mixed catalyst was intended to produce isolated active sites of Co and Fe on the support. Fig. 6 portrays their catalytic activities and selectivity to products with time on stream over catalysts of (a) 5 wt%Co/␣-Al2 O3 , (b) 0.22 wt%Fe/␣-Al2 O3 , (c) physically mixed catalyst of (a) and (b), and (d) 0.22 wt%Fe/5 wt%Co/␣-Al2 O3 . From Fig. 6, Fe/␣-Al2 O3 catalyst showed low activity of dehydrogenation of ethanol and less activity of steam reforming of ethanol. Comparison between Co/␣-Al2 O3 catalyst and physically mixed catalyst shows that the physically mixed catalyst had higher ethanol conversion than Co/␣Al2 O3 catalyst. However, comparing the conversion and selectivity to products, the respective catalytic activities of steam reforming of CH3 CHO on these catalysts were almost the same. The Fe/␣-Al2 O3 catalyst showed only activity of dehydrogenation of ethanol, so we considered that the physically mixed catalyst showed no synergetic effect of Co and Fe. On the other hand, activity of steam reforming of CH3 CHO was clearly promoted by Fe addition on Fe/Co/␣-Al2 O3 . Based on this result and on measurements of EDX, we concluded that interaction and promotion of catalytic activity were expressed by the close contact of Co and Fe. 4. Conclusion Effects of Fe addition on Co-based catalyst were investigated using several supports (SiO2 , MgO, SrTiO3 , ␣-Al2 O3 ) for steam
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27]
P. McKendry, Bioresour. Technol. 83 (2002) 37–46. P. McKendry, Bioresour. Technol. 83 (2002) 47–54. H.V. Blottnitz, M.A. Curran, J. Cleaner Prod. 15 (2007) 607–619. I.C. Macedo, J.E.A. Seabra, J.E.A.R. Silva, Biomass Bioenergy 32 (2008) 582–595. A. Haryanto, S. Fernando, N. Murali, S. Adhikari, Energy Fuels 19 (2005) 2098–2106. M. Ni, D.Y.C. Leung, M.K.H. Leung, Int. J. Hydrogen Energy 32 (2007) 3238–3247. L.F. Brown, Int. J. Hydrogen Energy 26 (2001) 381–397. F. Frusteri, S. Freni, J. Power Sources 173 (2007) 200–209. M. Momirlan, T.N. Veziroglu, Int. J. Hydrogen Energy 30 (2005) 795–802. D.K. Liguras, D.I. Kondarides, X.E. Verykios, Appl. Catal. B: Environ. 43 (2003) 345–354. J.P. Breen, R. Burch, H.M. Coleman, Appl. Catal. B: Environ. 39 (2002) 65–74. N. Homs, J. Llorca, P.R. de la Piscina, Catal. Today 116 (2006) 361–366. S. Freni, S. Cavallaro, N. Mondello, L. Spadaro, F. Frusteri, J. Power Sources 108 (2002) 53–57. J. Suna, X.P. Qiu, F. Wu, W.T. Zhua, Int. J. Hydrogen Energy 30 (2005) 437–445. F. Haga, T. Nakajima, K. Yamashita, S. Mishima, S. Suzuki, Nippon Kagaku Kaishi 1 (1997) 33–36. F. Haga, T. Nakajima, H. Miya, S. Mishima, Catal. Lett. 48 (1997) 223–227. F. Frusteri, S. Freni, L. Spadaro, V. Chiodo, G. Bonura, S. Donato, S. Cavallaro, Catal. Commun. 5 (2004) 611–615. S. Cavallaro, N. Mondello, S. Freni, J. Power Sources 102 (2001) 198–204. J. Llorca, N. Homs, J. Sales, P.R. de la Piscina, J. Catal. 209 (2002) 306–317. K. Urasaki, K. Tokunaga, Y. Sekine, M. Matsukata, E. Kikuchi, Catal. Commun. 9 (2008) 600–604. Y. Sekine, A. Kazama, Y. Izutsu, M. Matsukata, E. Kikuchi, Catal. Lett. 132 (2009) 329–334. J. Llorca, N. Homs, J. Sales, J.L.G. Fierro, P.R. de la Piscina, J. Catal. 222 (2004) 470–480. L.P.R. Profeti, E.A. Ticianelli, E.M. Assaf, J. Power Sources 175 (2008) 482–489. ˜ O’Shea, R. Nafria, P.R. de la Piscina, N. Homs, Int. J. Hydrogen V.A. de la Pena Energy 33 (2008) 3601–3606. J.A. Torres, J. Llorca, A. Casanovas, M. Domínguez, J. Salvadó, D. Montané, J. Power Sources 169 (2007) 158–166. W. Cai, B. Zhang, Y. Li, Y. Xu, W. Shen, Catal. Commun. 8 (2007) 1588–1594. K. Urasaki, Y. Sekine, S. Kawabe, E. Kikuchi, M. Matsukata, Appl. Catal. A: Gen. 286 (2005) 23–29.
Contents lists available at ScienceDirect
Applied Catalysis A: General journal homepage: www.elsevier.com/locate/apcata
Promoting effect of small amount of Fe addition onto Co catalyst supported on ␣-Al2 O3 for steam reforming of ethanol A. Kazama, Y. Sekine ∗ , K. Oyama, M. Matsukata, E. Kikuchi Applied Chemistry, Waseda University, 65-301, Okubo, Shinjuku, Tokyo, 169-8555, Japan
a r t i c l e
i n f o
Article history: Received 27 January 2010 Received in revised form 20 May 2010 Accepted 20 May 2010 Available online 27 May 2010 Keywords: Steam reforming of ethanol Promotion effect of iron Cobalt catalyst
a b s t r a c t We examined the promotion effect of loading small amounts of Fe onto various Co catalysts for steam reforming of ethanol. Among these catalysts, catalysts supported on SrTiO3 and ␣-Al2 O3 showed remarkable effects of iron loading onto cobalt catalyst. The Fe-loaded Co/␣-Al2 O3 showed higher yield of hydrogen, low coke deposition on the catalysts, and high activity for steam reforming of acetaldehyde, which was an intermediate of the steam reforming of ethanol. Characterization of the catalyst revealed that Fe and Co metal coexisted on the catalyst support. Synergetic effects of these two metals were observed. © 2010 Elsevier B.V. All rights reserved.
1. Introduction Recently, biomass has attracted much attention as a renewable energy resource [1,2]. Among many utilizations of biomass, utilization of biomass-derived-ethanol has been conducted actively [3,4]. Ethanol obtained through fermentation of saccharide (sugar) includes much water, so the water must be removed completely before the utilization as a gasoline-substitute. Large amounts of energy are required for rectification of ethanol. A useful process is therefore steam reforming of ethanol [5,6], which is applicable to hydrogen production directly with no rectification; the reaction is endothermic without loss of calorific value. The resultant hydrogen is available for various applications such as fuel cells and hydrogen combustion. For those reasons, the obtained biomassderived hydrogen can be regarded as a high-efficiency, clean energy resource [7–9]. To date, various transition metal catalysts such as Rh, Ni, and Co have been investigated for steam reforming of ethanol. Some reports show that noble metals such as Rh exhibit high activity and high stability [10,11]. Nevertheless, it is undesirable to use noble metals because of their high cost. On the other hand, many reports have described that base metals such as Ni [12–15] and Co [15–25] exhibit high activity for steam reforming of ethanol. The Ni-based catalysts show high ethanol conversion at low temperatures [14], but Ni-based catalysts generate more byproduct CH4 than Co-based catalysts [15]. Therefore, Ni-based catalysts are
deemed unsuitable for aspiring to a high H2 yield. The Co-based catalysts, with lower generation of undesired byproduct CH4 [16], have been researched actively among various transition metal catalysts, but their deactivation by sintering of the catalyst or carbon deposition on Co-based catalysts presents a serious obstacle to their wider use [17,18]. Some investigations have revealed that application of a basic oxide such as MgO [13,18], CeO2 [26], ZnO [19] or perovskite-type oxide [27] is a promising method for suppressing carbon deposition on the catalyst. We have reported that Co-based catalysts supported on perovskite-type oxides (SrTiO3 ) show high activity [20]. Regarding the second metal addition on Co-based catalysts, carbon deposition on catalysts was depressed by sodium addition [22], and Ru addition on Co/␥-Al2 O3 was effective among some noble metals (Pt, Pd, Ru, and Ir) [23]. The Fe supplementation of a Co-based catalyst, Co3 O4 [24] and Co/ZnO [25] has been investigated by other researchers; we also have found that addition of small amounts of Fe to Co/SrTiO3 promotes its catalytic activity [21]. Therefore, we investigated the effects of Fe addition on Co-based catalyst with various supports by controlling many factors and parameters such as reaction temperature and contact time (=W/F), etc. In addition, the physical structure of Fe-loaded Co catalyst was characterized to investigate the interaction of Co and Fe. Moreover, a comparison of the catalytic nature between Fe/Co/␣Al2 O3 and physically mixed catalyst (Fe/␣-Al2 O3 and Co/␣-Al2 O3 ) was conducted.
2. Experimental ∗ Corresponding author. Tel.: +81 3 5286 3114; fax: +81 3 5286 3114. E-mail address: [email protected] (Y. Sekine). 0926-860X/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.apcata.2010.05.031
As catalyst supports, SiO2 (JRC-SIO-5), MgO (Ube 1000A), SrTiO3 (synthesis method explained in Ref. [21]) of perovskite oxide and ␣-
A. Kazama et al. / Applied Catalysis A: General 383 (2010) 96–101
Al2 O3 (JRC-ALO-1; calcined at 1573 K for 3 h) were used. In addition, 5 wt% of Co and/or 0.22 wt% of Fe as an active metal was supported by an impregnation method. As precursors of these supported metals, Co(NO3 )2 ·6H2 O and Fe(C5 H7 O2 )3 were used. Distilled water was used as a solvent to support Co; acetone was used for Fe. First, a mixture of catalyst-support and liquid solution of the precursor was heated on the stirrer and evaporated to dryness. The resultant powder was dried in an oven at 393 K for 20 h in air. Then, it was calcined in a muffle furnace at 823 K for 3 h. Then Fe addition onto Co-based catalysts was conducted using a sequential impregnation method, and 0.22 wt% Fe was supported using the above-referenced method. Activity tests were conducted in a continuous flow reactor with a fixed bed of catalyst under atmospheric pressure. A quartz tube (8 mm inner diameter) was used as a reactor. The catalyst powder was pressed at 60 kN for 15 min, crushed, and sieved to obtain a particle size of 250–500 m. Then the sieved catalyst was charged in the catalyst bed with dilution of SiC to 10-mm height. The temperature inside the reactor was controlled using a heater with a thermocouple in the catalyst bed. The reactor was heated with 10 K min−1 up to the reduction temperature (873 K) in H2 flow (50 mL min−1 ). Then it was held for 1 h. After the reactor was cooled to the reaction temperature with purging by N2 , an aqueous solution of ethanol (steam to carbon ratio: S/C = 5.0) was fed to the reactor using a syringe pump through a vaporizer. The flow rate of ethanol was 7.7 × 10−4 mol min−1 , and W/F was changed between 0.08 and 5.2 g-cat h mol−1 by changing the catalyst amount. As an internal standard gas, N2 (20 mL min−1 ) was fed simultaneously, and 80 mL min−1 of Ar was fed simultaneously to the reactor as a carrier gas. Qualitative and quantitative analyses of gas components were conducted using a gas chromatograph (GC). Then H2 and N2 were analyzed using a thermal conductivity detector (TCD)-GC (GC-8A; Shimadzu Corp.). Subsequently, CO, CO2 and CH4 were analyzed using a flame ionization detector (FID)-GC (GC-8A; Shimadzu Corp.) with a methanizer (Ru/Al2 O3 ). In addition, hydrocarbon compounds (CH4 , C2 and C3), liquid products (such as CH3 CHO) and the unreacted ethanol were analyzed using a FID-GC (GC-14B; Shimadzu Corp.) with a six-way valve. Ethanol conversion and H2 yield were calculated according to Eqs. (1) and (2). The H2 yield would be maximum (200%) from Eq. (2) if all the ethanol was reformed to CO2 and H2 with a water gas shift reaction.
ethanol conversion (%) =
out Fcarbon mole of all product in Fcarbon mole of C
(1)
2 H5 OH
H2 yield (%) =
FHout × 100 2
FCin H 2
5 OH
×3
(2)
The amount of deposited carbon on the catalyst was analyzed using temperature-programmed oxidation (TPO). The catalyst was heated 10 K min−1 up to 1173 K with 10% O2 /N2 gas flow, and the quantities of generated CO and CO2 were measured using an infrared gas analyzer (Shimadzu CGT-7000). Finally, the amount of carbon was calculated using these values. The Co particle size was measured using transmission electron microscopy (TEM, JEM-1011; JEOL). The average diameter of Co particles was measured using TEM and then calculated. The physical structures of Co and Fe on Fe/Co/␣-Al2 O3 were observed using scanning transmission electron microscopy (STEM, HF-2210; Hitachi Ltd.) and energy dispersive X-ray spectrometry (EDX, Genesis4000; EDAX Inc.).
97
3. Results and discussion 3.1. Catalytic activities of various Fe-loaded catalysts We previously reported that catalytic activity over Co/SrTiO3 catalyst was promoted by Fe addition [21]. First, we investigated effects of support among Fe/Co-based catalysts. The ethanol conversion, H2 yield and amount of carbon were compared on Fe/Co-based catalysts with various supports, SiO2 , MgO, SrTiO3 , ␣Al2 O3 . Table 1 presents results of catalytic activity tests and amount of carbon after 185 min of reaction. Here, the optimized amount of Fe loading (=0.22 wt%) was chosen from the previous result over Fe/Co/SrTiO3 catalyst [21]. From Table 1, improvement of H2 yield and suppression of carbon deposition were not observed by Fe addition for catalysts of Co/SiO2 and Co/MgO. On the other hand, the promoting effects of Fe on ethanol conversion and H2 yield, and the suppression of carbon deposition were observed for catalysts of Co/SrTiO3 and Co/␣-Al2 O3 . Although the Co/␣-Al2 O3 catalyst showed less ethanol conversion and H2 yield than Co/SrTiO3 catalyst, Feloaded Co/␣-Al2 O3 catalyst showed higher activity and stability than Fe/Co/SrTiO3 catalyst. From this result, we concluded that Fe addition on Co/␣-Al2 O3 was effective for steam reforming of ethanol. Therefore, effects of Fe addition on Co/␣-Al2 O3 catalyst were considered by changing amounts of Fe loading and the calcination time. Table 2 presents results of activity tests. As shown in Table 2, all Fe-loaded Co/␣-Al2 O3 catalysts showed higher activity and less selectivity to CH3 CHO than that of Co/␣-Al2 O3 catalyst. Some reports have described that steam reforming of ethanol proceeds via CH3 CHO as an intermediate [5,21]. Therefore, it was considered that steam reforming of CH3 CHO was promoted by Fe addition from results of Table 2 and that the H2 yield was improved. Based on these results, we chose 0.22 wt% Fe/5 wt% Co/␣-Al2 O3 (calcined for 3 h after Fe loading) as the best catalyst for additional investigations. 3.2. Selectivity to products over Fe-loaded catalyst To investigate the effective temperature window on these catalysts, we examined the effect of reaction temperature over catalysts of Co/␣-Al2 O3 and Fe/Co/␣-Al2 O3 . Activity tests were conducted at each reaction temperature (753, 783, 823, and 873 K). Fig. 1a presents ethanol conversion and Fig. 1b shows the H2 yield at each reaction temperature. Based on the results portrayed in Fig. 1a and b, ethanol conversion and H2 yield were promoted by Fe addition at each temperature. Furthermore, the improvement of activity by Fe addition was observed dominantly at 783 and 823 K. The apparent activation energy was calculated from Fig. 1a. The value for Co/␣-Al2 O3 catalyst was 67.3 kJ mol−1 ; that for Fe/Co/␣-Al2 O3 catalyst was 53.9 kJ mol−1 . From these values, it was confirmed that the apparent activation energy was reduced by Fe addition. Subsequently, catalytic activity over catalysts of Co/␣-Al2 O3 and Fe/Co/␣-Al2 O3 was compared by changing the contact time (=W/F) to examine the reaction passage over these catalysts. Fig. 2 shows carbon-based yields of products on each W/F after 185 min of reaction. Results confirmed that Fe/Co/␣-Al2 O3 catalyst promoted steam reforming and that larger amounts of CO, CO2 —of course including hydrogen—were obtained than with Co/␣-Al2 O3 catalyst, as shown in Fig. 2. Furthermore, Fe/Co/␣-Al2 O3 catalyst showed a lower yield of the reaction intermediate, CH3 CHO, than with Co/␣-Al2 O3 catalyst by comparison between high W/F. These results also demonstrated that Fe addition promoted the steam reforming activity of CH3 CHO. In addition, the yield of byproducts such as CH4 and carbon over Fe/Co/␣-Al2 O3 catalyst was lower
98
A. Kazama et al. / Applied Catalysis A: General 383 (2010) 96–101
Table 1 Effect of Fe loading on Co catalysts supported on various supports. Support
Conversion/% Co
a
SiO2 MgOb SrTiO3 b ␣-Al2 O3 b
Deposited carbon/mg g-cat−1
H2 yield/% Fe/Co
Co
10 min
185 min
10 min
185 min
77.1 72.4 74.4 66.9
62.3 48.3 71.2 57.9
97.3 60.2 89.4 95.7
70.9 43.6 76.3 89.6
10 min 62 95 106 97
Fe/Co 185 min 63 49 101 92
Co
Fe/Co
10 min
185 min
185 min
185 min
65 75 131 168
54 44 103 149
186 47 331 350
156 41 85 51
Loading amount of Co was 5 wt%, reaction temperature = 823 K, W/F = 2.6, S/C = 5.0. a Fe loading = 0.11 wt%, calcination condition = 673 K, 1 h. b Fe loading = 0.22 wt%, calcination condition = 823 K, 3 h.
Table 2 Effect of Fe loading on Co/␣-Al2 O3 catalysts with various conditions. Catalyst
Calcination condition
5 wt%Co/␣-Al2 O3 0.11 wt%Fe/Co/␣-Al2 O3 0.22 wt%Fe/Co/␣-Al2 O3 0.22 wt%Fe/Co/␣-Al2 O3
– 823 K, 1 h 823 K, 1 h 823 K, 3 h
Ethanol conv./%
H2 yield/%
10 min
185 min
10 min
185 min
CO
CO2
CH4
CH3 CHO
66.9 91.5 93.9 95.7
57.9 90.5 90.0 89.6
97 152 154 168
92 126 124 149
12.2 21.9 24.6 21.7
68.4 68.4 66.7 70.7
4.7 2.8 2.7 2.9
14.3 6.5 5.5 4.5
Selectivity at 10 min./%
Deposited carbon after 185 mina /mg g-cat−1
350–419 63–110 67–124 51–87
Reaction temperature = 823 K, W/F = 2.6, S/C = 5.0. a Carbon amount was measured twice using TPO.
Fig. 1. Effect of reaction temperature on ethanol conversion (a) and H2 yield (b) over each catalyst: W/F = 0.65, S/C = 5.0, reaction at 10 min.
Fig. 2. Product yield over Co/␣-Al2 O3 (a) or Fe/Co/␣-Al2 O3 through 185 min reaction: reaction temperature = 823 K, S/C = 5.0, W/F = 0.08–5.2, reaction time = 185 min.
A. Kazama et al. / Applied Catalysis A: General 383 (2010) 96–101
Fig. 3. Amount of carbon deposited on catalysts at each W/F: reaction temperature = 823 K, S/C = 5.0, reaction time = 185 min.
than that over Co/␣-Al2 O3 catalyst. Fig. 3 shows the amount of deposited carbon over catalyst after 185 min of reaction on each W/F. The amount of carbon over Fe/Co/␣-Al2 O3 catalyst was less than that over Co/␣-Al2 O3 catalyst at high W/F conditions. Furthermore, the amount of carbon over Fe/Co/␣-Al2 O3 catalyst was not increased to exceed W/F = 1.3 g h mol−1 . The Fe addition promoted steam reforming without promoting the formation of byproducts such as carbon and methane. 3.3. Stability of catalysts To compare the stability of these catalysts of Co/␣-Al2 O3 and Fe/Co/␣-Al2 O3 , we conducted activity tests with reduced amount
99
Fig. 4. Catalytic activity over Co/␣-Al2 O3 and Fe/Co/␣-Al2 O3 : reaction temperature = 823 K, W/F = 0.65; S/C = 5.0; 䊉, ethanol conversion on Co/␣-Al2 O3 ; , H2 yield on Co/␣-Al2 O3 ; , ethanol conversion on Fe/Co/␣-Al2 O3 ; ♦, H2 yield on Fe/Co/␣Al2 O3 .
of catalysts. Fig. 4 presents ethanol conversion and H2 yield over these catalysts with time on stream. From Fig. 4, Fe-loaded catalyst showed higher ethanol conversion and H2 yield through 185 min of reaction. The particle size of Co measured using TEM (modal diameter) on Co/␣-Al2 O3 catalyst was 19 nm; that of Fe/Co/␣-Al2 O3 catalyst was 17 nm before the reaction (after calcination). Therefore, the difference of the catalytic activities was not derived from the metallic surface area of active metal (=Co). Consequently, the promotion effect of Fe addition was unrelated with the Co particle size. We were able to observe slight deactivation of the catalyst for these two catalysts as shown in Fig. 4 and Table 2. Regarding the deactivation of the catalysts, we also measured the modal diameter of Co particle over Fe/Co/␣-Al2 O3 catalyst at each step (after calcination, after reduction and after the reaction) by TEM obser-
Fig. 5. STEM image and EDX spectra of Co/␣-Al2 O3 and Fe/Co/␣-Al2 O3 catalyst after reaction.
100
A. Kazama et al. / Applied Catalysis A: General 383 (2010) 96–101
Fig. 6. Catalytic activity over each catalyst: (a) Co/␣-Al2 O3 (reaction temperature = 823 K, W/F = 0.65, S/C = 5.0); (b) Fe/␣-Al2 O3 (reaction temperature = 823 K, W/F = 2.6, S/C = 5.0); (c) physical mixture of Fe/␣-Al2 O3 and Co/␣-Al2 O3 (reaction temperature = 823 K, W/F = 0.65, S/C = 5.0); (d) Fe/Co/␣-Al2 O3 (reaction temperature = 823 K, W/F = 0.65, S/C = 5.0).
vation. The value after calcination was 17 nm; that after reduction was also 17 nm. That after the reaction was 19 nm. We considered that these values were almost the same and that little influence of agglomeration of Co-metal was exerted on the deactivation of the catalyst over Fe/Co/␣-Al2 O3 .
3.4. Structure of Fe-loaded catalyst To evaluate the promotion effect of Fe loading, we investigated the physical structure of Fe-loaded Co/␣-Al2 O3 catalyst using STEM-EDX. In addition, Fe/Co/␣-Al2 O3 catalysts were used after
A. Kazama et al. / Applied Catalysis A: General 383 (2010) 96–101
reduction and after the reaction. Fig. 5 shows the STEM image and EDX spectra of Co/␣-Al2 O3 and Fe/Co/␣-Al2 O3 catalysts after reaction. First, on EDX measurement, the peak of ␣-Al2 O3 or that with a very weak peak of Co, was detected from the points uncovered by particles of Co and Fe (not shown). On the other hand, the peaks of Co and Fe were measured simultaneously from the points of Co particles on Fe/Co/␣-Al2 O3 catalyst as shown in Fig. 5. The calculated weight ratio between Co and Fe was 20–40:1. This value was close to the actual weight ratio (23:1), suggesting that Co and Fe coexisted over ␣-Al2 O3 . Furthermore, from this result, we considered that Fe was supported on Co particles selectively using the sequential impregnation method.
101
reforming of ethanol. Results show that Fe addition on Co/␣-Al2 O3 catalyst was the most effective among these catalysts. Observations indicated that Fe addition promoted steam reforming of CH3 CHO while suppressing formation of byproducts such as CH4 and carbon. Characterization of Fe/Co/␣-Al2 O3 catalyst by STEM-EDX revealed that sequentially impregnated Fe was supported selectively on Co particles. Physically mixed catalyst (mixed Co/␣-Al2 O3 catalyst with Fe/␣-Al2 O3 catalyst) did not show high activity such as that of Fe/Co/␣-Al2 O3 . Therefore, we considered that Fe supported on Co particles affected the promotion of catalytic activity of steam reforming of ethanol and CH3 CHO. References
3.5. Interaction of Co and Fe To investigate the interaction between Co and Fe on catalytic activity, we prepared physically mixed catalyst (mixed Co/␣Al2 O3 and Fe/␣-Al2 O3 ), and compared its catalytic activity to that of Fe/Co/␣-Al2 O3 catalyst. This physically mixed catalyst was intended to produce isolated active sites of Co and Fe on the support. Fig. 6 portrays their catalytic activities and selectivity to products with time on stream over catalysts of (a) 5 wt%Co/␣-Al2 O3 , (b) 0.22 wt%Fe/␣-Al2 O3 , (c) physically mixed catalyst of (a) and (b), and (d) 0.22 wt%Fe/5 wt%Co/␣-Al2 O3 . From Fig. 6, Fe/␣-Al2 O3 catalyst showed low activity of dehydrogenation of ethanol and less activity of steam reforming of ethanol. Comparison between Co/␣-Al2 O3 catalyst and physically mixed catalyst shows that the physically mixed catalyst had higher ethanol conversion than Co/␣Al2 O3 catalyst. However, comparing the conversion and selectivity to products, the respective catalytic activities of steam reforming of CH3 CHO on these catalysts were almost the same. The Fe/␣-Al2 O3 catalyst showed only activity of dehydrogenation of ethanol, so we considered that the physically mixed catalyst showed no synergetic effect of Co and Fe. On the other hand, activity of steam reforming of CH3 CHO was clearly promoted by Fe addition on Fe/Co/␣-Al2 O3 . Based on this result and on measurements of EDX, we concluded that interaction and promotion of catalytic activity were expressed by the close contact of Co and Fe. 4. Conclusion Effects of Fe addition on Co-based catalyst were investigated using several supports (SiO2 , MgO, SrTiO3 , ␣-Al2 O3 ) for steam
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27]
P. McKendry, Bioresour. Technol. 83 (2002) 37–46. P. McKendry, Bioresour. Technol. 83 (2002) 47–54. H.V. Blottnitz, M.A. Curran, J. Cleaner Prod. 15 (2007) 607–619. I.C. Macedo, J.E.A. Seabra, J.E.A.R. Silva, Biomass Bioenergy 32 (2008) 582–595. A. Haryanto, S. Fernando, N. Murali, S. Adhikari, Energy Fuels 19 (2005) 2098–2106. M. Ni, D.Y.C. Leung, M.K.H. Leung, Int. J. Hydrogen Energy 32 (2007) 3238–3247. L.F. Brown, Int. J. Hydrogen Energy 26 (2001) 381–397. F. Frusteri, S. Freni, J. Power Sources 173 (2007) 200–209. M. Momirlan, T.N. Veziroglu, Int. J. Hydrogen Energy 30 (2005) 795–802. D.K. Liguras, D.I. Kondarides, X.E. Verykios, Appl. Catal. B: Environ. 43 (2003) 345–354. J.P. Breen, R. Burch, H.M. Coleman, Appl. Catal. B: Environ. 39 (2002) 65–74. N. Homs, J. Llorca, P.R. de la Piscina, Catal. Today 116 (2006) 361–366. S. Freni, S. Cavallaro, N. Mondello, L. Spadaro, F. Frusteri, J. Power Sources 108 (2002) 53–57. J. Suna, X.P. Qiu, F. Wu, W.T. Zhua, Int. J. Hydrogen Energy 30 (2005) 437–445. F. Haga, T. Nakajima, K. Yamashita, S. Mishima, S. Suzuki, Nippon Kagaku Kaishi 1 (1997) 33–36. F. Haga, T. Nakajima, H. Miya, S. Mishima, Catal. Lett. 48 (1997) 223–227. F. Frusteri, S. Freni, L. Spadaro, V. Chiodo, G. Bonura, S. Donato, S. Cavallaro, Catal. Commun. 5 (2004) 611–615. S. Cavallaro, N. Mondello, S. Freni, J. Power Sources 102 (2001) 198–204. J. Llorca, N. Homs, J. Sales, P.R. de la Piscina, J. Catal. 209 (2002) 306–317. K. Urasaki, K. Tokunaga, Y. Sekine, M. Matsukata, E. Kikuchi, Catal. Commun. 9 (2008) 600–604. Y. Sekine, A. Kazama, Y. Izutsu, M. Matsukata, E. Kikuchi, Catal. Lett. 132 (2009) 329–334. J. Llorca, N. Homs, J. Sales, J.L.G. Fierro, P.R. de la Piscina, J. Catal. 222 (2004) 470–480. L.P.R. Profeti, E.A. Ticianelli, E.M. Assaf, J. Power Sources 175 (2008) 482–489. ˜ O’Shea, R. Nafria, P.R. de la Piscina, N. Homs, Int. J. Hydrogen V.A. de la Pena Energy 33 (2008) 3601–3606. J.A. Torres, J. Llorca, A. Casanovas, M. Domínguez, J. Salvadó, D. Montané, J. Power Sources 169 (2007) 158–166. W. Cai, B. Zhang, Y. Li, Y. Xu, W. Shen, Catal. Commun. 8 (2007) 1588–1594. K. Urasaki, Y. Sekine, S. Kawabe, E. Kikuchi, M. Matsukata, Appl. Catal. A: Gen. 286 (2005) 23–29.