Вы находитесь на странице: 1из 9

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 7 6 9 0 e7 6 9 8

Available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/he

Ethanol steam reforming and water gas shift over Co/ZnO catalytic honeycombs doped with Fe, Ni, Cu, Cr and Na
Albert Casanovas a, Maria Roig a, Carla de Leitenburg b, Alessandro Trovarelli b, Jordi Llorca a,c,*
a

`cniques Energe `tiques, Universitat Polite `cnica de Catalunya, Diagonal 647, Ed. ETSEIB, 08028 Barcelona, Spain Institut de Te ` di Udine, via del Cotonicio 108, 33100 Udine, Italy Dipartimento di Scienze e Tecnologie Chimiche, Universita c `cnica de Catalunya, Pasqual i Vila 1-15, 08028 Barcelona, Spain Centre for Research in NanoEngineering, Universitat Polite
b

article info
Article history: Received 17 February 2010 Received in revised form 7 May 2010 Accepted 24 May 2010 Available online 25 June 2010 Keywords: Ethanol steam reforming Water gas shift Hydrogen Catalytic honeycomb Cobalt catalyst

abstract
The effect of Fe, Ni, Cu, Cr, and Na (1%) addition over ZnO-supported Co (10%) honeycomb catalysts in the steam reforming of ethanol (ESR) and water gas shift reaction (WGS) for the production of hydrogen was studied. HRTEM and EEL spectroscopy revealed the formation of metal alloys between Co and Fe, Ni, Cu, and Cr. Catalysts promoted with Fe and Cr performed better in the ESR, and the sample promoted with Fe showed high activity for WGS at low temperature. As deduced from TPR and oxidation pulse experiments, alloy particles in catalysts promoted with Fe and Cr exhibited a rapid and higher degree of redox exchange between reduced and oxidized Co, which may explain the better catalytic performance. 2010 Professor T. Nejat Veziroglu. Published by Elsevier Ltd. All rights reserved.

1.

Introduction

The search for an active and selective catalyst for the generation of hydrogen through ethanol steam reforming (equation (1)) at low temperature constitutes an active research area because ethanol is a renewable fuel with low toxicity and high energy density that is also easy to handle and distribute [1e3]. Moreover, a bioethanol-to-H2 system has the advantage of being CO2 neutral. C2 H5 OH 3H2 O/6H2 2CO2 (1)

There are numerous studies that demonstrate the feasibility of generating hydrogen from ethanolewater mixtures through catalytic steam reforming, either with powder catalysts [4,5]

and over catalytic walls [6e10]. Among all catalysts tested so far, those based on cobalt exhibit the highest activity and selectivity towards hydrogen at low temperature [11e27]. Concerning the support, acidic supports should be avoided since they favor massive carbon deposition through ethanol dehydration into ethylene, whereas supports with basic and redox characteristics are preferred, such as ZnO [28]. Thus, much work has been carried out over the Co/ZnO system. It has been demonstrated by in situ magnetic studies coupled to reaction tests and by in situ diffuse reectance infrared spectroscopy under real operation that the simultaneous presence of metallic cobalt and cobalt oxide is required for the progress of the reaction [29e31]. Two steps of the reaction have been identied. First, ethanol dehydrogenates into acetaldehyde

` cniques Energe ` tiques, Universitat Polite ` cnica de Catalunya, Av. Diagonal 647, Ed. ETSEIB, 08028 * Corresponding author. Institut de Te Barcelona, Spain. Tel.: 34 93 401 17 08; fax: 34 93 401 71 49. E-mail address: jordi.llorca@upc.edu (J. Llorca). 0360-3199/$ e see front matter 2010 Professor T. Nejat Veziroglu. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.ijhydene.2010.05.099

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 7 6 9 0 e7 6 9 8

7691

and hydrogen (equation (2)) over cobalt oxide (Co3O4). Hydrogen partly reduces the surface of cobalt particles into metallic cobalt and then, the second step, the reforming of acetaldehyde into the nal products H2 and CO2, takes place (equation (3)) with the participation of the water gas shift reaction (equation (4)). C2 H5 OH/CH3 CHO H2 CH3 CHO 3H2 O/5H2 2CO2 H2 O CO4H2 CO2 C2 H5 OH/CH4 CO H2 (2) (3) (4) (5)

673 K. This procedure was repeated several times in order to obtain the desired weight gain (10e12% w/w).

2.2.

Characterization

In order to favor the redox exchange between metallic cobalt and cobalt oxide under reaction conditions, several cobalt alloy formulations have been attempted. Alloying cobalt with the more electronegative rst-row transition metals Ni and Cu did not improve the performance for the ethanol steam reforming reaction [32]. Also, alloying cobalt with noble metals favors the formation of methane through ethanol decomposition at low temperature (equation (5)). In contrast, alloying cobalt with the less electronegative rst-row transition metals Fe and Mn showed to be positive for the steam reforming of ethanol in terms of both catalytic activity and selectivity towards hydrogen [33e35]. In this work, we extend the study of the steam reforming of ethanol (ESR) over ZnO-supported, Co-based honeycomb catalysts by evaluating the effect of Fe, Ni, Cu, Cr, and Na addition. We also report the performance of these catalysts for the water gas shift reaction (WGS), which participates in the reaction scheme (4). It is well known that honeycomb structures are preferred for the generation of hydrogen in industrial environments and for mobile applications, such as fuel cell powered vehicles equipped with internal reformers [8e10]. They are robust, easy to scale up and replace, and offer homogeneous ow distribution patterns with low pressure drop.

Mechanical stability of the catalyst coatings in honeycomb samples was evaluated by direct exposure to mechanical vibration. The vibration frequency was raised progressively from 20 to 50 Hz at a xed acceleration value of 2 G, and at 50 Hz the acceleration was progressively increased from 2 to 10 G. Weight loss was monitored after 30 min at each frequency and acceleration, and after 3 h under the most severe vibration conditions (50 Hz, 10 G). G levels were controlled directly on the vibration test board with a Bru el & Kjaer 4370 accelerometer. Scanning electron microscopy was accomplished using a JEOL JSM 6400 instrument at an acceleration voltage of 20 kV. X-ray diffraction proles (XRD) were collected at a step width of 0.02 degrees and by counting 10 s at each step with a Siemens D-500 instrument equipped with a Cu target and a graphite monochromator. High-resolution transmission electron microscopy (HRTEM) was conducted at 200 kV with a JEOL JEM 2010F microscope equipped with a eld emission gun and an EELS detector. Samples were dispersed in alcohol and deposited on grids with holey carbon lms. Temperature programmed reduction (TPR) was carried out with a Micromeritics AutoChem II 2920 instrument using a H2/Ar mixture (5% H2) at 10 K min1 and a TCD detector. Oxidation experiments were carried out at 723 K with 30 consecutive 0.05 mL oxygen pulses (1 pulse/min). Surface area measurements (BET) were performed with a Micromeritics TriStar 3000 apparatus. Chemical composition was obtained by optical emission spectroscopy with inductively-coupled plasma (ICP-OES, PerkineElmer Optima apparatus).

2.3.

Catalytic tests

2.
2.1.

Experimental method
Preparation of powder catalysts and honeycombs

Co/ZnO catalysts (10% wt. Co) doped with Fe, Ni, Cu, Cr, and Na (1% wt.) were prepared by incipient wetness co-impregnation from nitrate aqueous solutions over ZnO (Kadox 15). The resulting solids were dried at 383 K overnight and calcined in air at 673 K for 6 h. These catalysts were labeled as Co(Fe)/ZnO, Co(Ni)/ZnO, Co(Cu)/ZnO, Co(Cr)/ZnO, and Co(Na)/ZnO. For comparative purposes, a monometallic cobalt catalyst supported on ZnO, Co/ZnO, was prepared in a similar way. For the preparation of honeycomb catalysts, 400 cpsi (cells per square inch) cordierite monolith cylinders with a diameter of 2 cm and a length of 2 cm were used. They were obtained by cutting larger monolith pieces with a diamond saw. Honeycomb catalysts were prepared by the washcoating method from vigorously agitated suspensions of powdered catalysts in deionized water (w5% w/w). After each immersion, honeycombs were dried at 373 K under continuous rotation and calcined at

Ethanol steam reforming was carried out over honeycomb catalysts at atmospheric pressure and 473e773 K in a tubular reactor at a total ow of 80 mL min1. C2H5OH (0.33 mL min1) and H2O were fed separately at a C2H5OH:H2O molar ratio of 1:6 (S/C 3) and the mixture was balanced with He. The efuent of the reactor was monitored on line with an MKS Cirrus mass spectrometer or an Agilent micro-GC. Samples were rst pretreated inside the reactor with a H2:N2 mixture (50 mL min1, 10% H2) at 723 K for 4 h, the temperature was lowered to 473 K under N2, and then the reaction mixture was introduced at 473 K. Monoliths operated under isothermal conditions as deduced from temperature monitoring inside their channels, located either in contact with the reactor wall or at the center of the reactor. Unaltered conversion level was measured at various owrates at reactor inlet while keeping constant the ratio between the weight of catalyst and the molar ow rate of ethanol at reactor inlet, thus ensuring the absence of external mass transfer limitations. The water gas shift reaction was carried out at atmospheric pressure in the 473e673 K temperature range using a CO:H2:H2O:N2 1:2:6:14 molar mixture (total ow 50 mL min1). Water was provided with a syringe pump and vaporized before entering the reactant stream. Analysis of products was performed with a Varian micro-GC.

7692

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 7 6 9 0 e7 6 9 8

3.

Results and discussion

3.1. Redox and structural characterization of promoted Co/ZnO catalysts


Table 1 compiles the catalysts prepared along with their surface area (BET method) and chemical composition. The surface area values recorded for all samples were similar and about ca. 10e12 m2 g1. In order to get insight into the redox characteristics of the cobalt catalysts promoted with different metals, detailed temperature programmed reduction (TPR) and oxygen pulse experiments were carried out to study the redox exchange between oxidized and reduced cobalt, which has been demonstrated to be the clue for the production of hydrogen by ethanol steam reforming over Co-based systems [29e35]. Fig. 1 shows the TPR proles of all samples after oxidation at 673 K. All of them exhibit two main hydrogen consumption peaks, which correspond to the well known two-step transformation of cobalt spinel into metallic cobalt: Co3O4 / CoO / Co [8]. However, upon addition of only 1% w/w promoter, the TPR proles for the different promoted catalysts differ signicantly. For the bare Co/ZnO sample (Fig. 1a), the two peaks are centered at about 560 and 690 K, whereas for the promoted samples the rst hydrogen uptake (Co3O4 / CoO) occurs clearly at a lower temperature, ca. 520e540 K (Fig. 1bef). In contrast, the temperature of the second hydrogen uptake and the relative intensity between the two hydrogen consumption peaks vary considerably among the different samples. The Co(Cr)/ZnO and Co(Na)/ZnO samples exhibit a CoO / Co transformation similar to that of Co/ZnO, both in temperature and relative intensity, whereas for the Co(Ni)/ZnO catalyst the amount of hydrogen consumed in the second peak is clearly larger with respect to the rst hydrogen uptake. On the other hand, the position of the second hydrogen consumption peak is shifted towards a higher reduction temperature in the Co(Fe)/ZnO catalyst and towards a much lower temperature in the Co(Cu)/ZnO sample. In this case, the area of the second hydrogen uptake is rather small compared to the other promoted samples. Finally, an additional hydrogen uptake at low temperature, ca. 500 K, is particularly observed in the sample promoted with Cu (Fig. 1d). It is concluded that the reduction behavior of the promoted samples does not follow a clear trend in electron donation taking into account the electron afnity of the promoters, and the variations in the TPR proles may be due to more complex effects, such as alloying.

Fig. 1 e Temperature programmed reduction proles of catalysts Co/ZnO (a), Co(Fe)/ZnO (b), Co(Ni)/ZnO (c), Co(Cu)/ ZnO (d), Co(Cr)/ZnO (e), and Co(Na)/ZnO (f).

Table 1 e Chemical analysis, surface area, and mean cobalt particle size of catalysts determined by X-ray diffraction and transmission electron microscopy. Catalyst
Co/ZnO Co(Fe)/ZnO Co(Ni)/ZnO Co(Cu)/ZnO Co(Cr)/ZnO Co(Na)/ZnO

w/w Co (%)
9.8 10.1 9.7 10.2 9.9 9.4

w/w M (%)
e 0.98 1.05 1.11 0.97 0.89

BET (m2 g1)


11.3 10.9 12.2 12.8 11.4 11.8

dxRD (nm)
17.1 16.2 19.1 19.8 18.3 18.5

dTEM (nm)
15.6 14.1 17.4 18.1 16.7 16.2

In order to get insight into the amount of cobalt that can reversibly participate in a redox exchange, several alternate TPR and oxygen pulse experiments were carried out over samples Co/ZnO, Co(Fe)/ZnO and Co(Cr)/ZnO. Three TPR proles and two oxidation experiments using oxygen pulses (OP) at 723 K were alternated: TPR1 / OP1 / TPR2 / OP2 / TPR3. The amount of oxygen uptake on a metal basis recorded over the Co(Fe)/ZnO and Co(Cr)/ZnO samples in OP1 was signicantly higher than that of sample Co/ZnO. Taking into account the metal loading of the different samples (Table 1), the extent of reoxidation for catalysts Co(Fe)/ZnO and Co(Cr)/ZnO was about 90 and 85%, respectively, whereas for the Co/ZnO sample the degree of reoxidation was much lower, about 70%. In addition, the dynamics of the oxygen uptake differed considerably between Co/ZnO and the samples promoted with Fe and Cr. In the promoted samples, the transition between complete oxygen uptake and non-oxygen uptake was fast (3e4 pulses), whereas the transition in the Co/ZnO catalyst was signicantly slower (9e10 pulses). It can be concluded that the redox dynamics between oxidized and reduced cobalt is clearly

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 7 6 9 0 e7 6 9 8

7693

enhanced in the presence of these promoters. The TPR proles recorded over catalysts Co/ZnO and Co(Fe)/ZnO before and after each oxygen pulse experiment are reported in Fig. 2. The TPR proles recorded after the oxygen pulse experiments, TPR2 and TPR3 (b-c), differed from TPR1 (a). In both samples, the area of the second hydrogen consumption peak increased considerably with respect to the low temperature hydrogen uptake. In addition, in the Co(Fe)/ZnO sample, extra hydrogen uptake signals appeared at ca. 570 and 710 K after reoxidation, suggesting the formation of new active sites upon redox cycling. Similar results were obtained with the Co(Cr)/ZnO sample. The structure of all catalysts prepared in this work have been characterized in detail. After calcination at 673 K, all catalysts showed by X-ray diffraction (XRD) the characteristic peaks of the Co3O4 spinel phase. In addition to ZnO peaks, no other signals appeared in the diffraction patterns due to the low promoter loading (1% w/w). In addition, no differences were observed in the position and shape of the spinel diffraction peaks. After reduction at 673 K, peaks due to metallic cobalt (fcc) and/or cobalt-metal alloys appeared in the XRD patterns of all catalysts and no cobalt oxide phases (Co3O4, CoO) were observed. Table 1 compiles the mean particle size of cobalt/cobalt alloy particles as deduced from the Scherrer equation. Particle size of about 16e20 nm were calculated in all cases, indicating that the incorporation of promoter had no signicant effect on particle size. Since XRD did not provide information about the phases where the promoters were present, a detailed microstructural study was carried out on the reduced samples by combined high-resolution transmission electron microscopy and energy electronloss spectroscopy (HRTEM-EELS) in order to determine if cobalt and promoter entities were in contact, or occurred as separate phases. This is important for elucidating the role of promoters in the catalytic behavior of these catalysts with respect to Co/ZnO in ESR and WGS reactions. More than 150 metal particles covering various parts of the samples were used for size distribution measurements and more than 40 EEL spectra were recorded for each sample.

Fig. 3 shows representative bright eld and lattice fringe TEM images of all samples along with EEL spectra recorded over individual metal particles. As a general rule, well dispersed metal particles over ZnO support were present in all catalysts, with similar particle size distribution centered at about 14e18 nm, which is well in accordance with values calculated from XRD (Table 1). The dispersion of cobalt particles over the support in the Co/ZnO catalyst is well exemplied in Fig. 3a. The insets show an HRTEM image and an EEL spectrum recorded over one individual cobalt particle. The particle shows lattice fringes according to a single metallic Co crystallite oriented along the [012] direction. As expected, the EEL spectrum showed peaks characteristic of cobalt at 769.7 and 816.4 eV corresponding to Co L3 edge. A similar analysis performed over Co/ZnO from a different preparation batch showed similar results, thus indicating that the preparation method used in this work is fully reproducible. Fig. 3b corresponds to the Co(Fe)/ZnO sample. Again, a good dispersion of metal particles is encountered. The inset in Fig. 3b shows an HRTEM image of a single particle oriented along the [001] crystallographic direction. An accurate analysis of lattice fringes in individual metal particles did not reveal any signicant modication of the cobalt fcc structure. However, it should be taken into account that the lattice parameter of metallic cobalt is not expected to change signicantly with the addition of 1% w/w iron, and that such modication would be under the accuracy of the HRTEM images in such small particles. In contrast, EELS showed to be very useful for determining the occurrence of cobalt alloying with iron. In all single particles analyzed, both Co and Fe peaks (at 683.8 eV) were identied in the EEL spectra (see inset), with an approximate Co:Fe atomic ratio of 9:1, which corresponds well with the cobalt and iron content of the sample. Therefore, the Co(Fe)/ZnO sample is constituted by CoeFe alloy particles well dispersed over ZnO. Similarly, the EEL spectra recorded over individual particles in samples Co (Ni)/ZnO, Co(Cu)/ZnO, and Co(Cr)/ZnO (Fig. 3cee) showed the simultaneous occurrence of Co and promoter peaks with an

Fig. 2 e Temperature programmed reduction proles of catalysts Co/ZnO and Co(Fe)/ZnO recorded over the fresh samples (a), and after consecutive oxygen pulse experiments (b,c).

7694

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 7 6 9 0 e7 6 9 8

Fig. 3 e Transmission electron microscopy images and EEL spectra corresponding to catalysts Co/ZnO (a), Co(Fe)/ZnO (b), Co(Ni)/ZnO (c), Co(Cu)/ZnO (d), Co(Cr)/ZnO (e), and Co(Na)/ZnO (f).

approximate Co:M atomic ratio of 9:1, indicating that alloying between Co and the promoter occurred in these samples, too. As discussed above, lattice fringe analysis could not distinguish between pure Co particles and those alloyed with the promoters, but served to unambiguously identify the cobaltbased particles. The alloy particles were sometimes covered by a layer of Co3O4 (Fig. 3ced), with distinctive lattice spacing at 4.67 and 2.86  A corresponding to (111) and (220) planes, respectively. Finally, the sample promoted with sodium,

Co(Na)/ZnO (Fig. 3f), was virtually identical to the nonpromoted Co/ZnO catalyst, indicating that, in this case, sodium is likely atomically dispersed and not incorporated into the cobalt structure. No signals corresponding to the NaeK edge at 1050e1150 eV appeared in the EEL spectra. From the microstructural characterization results it is concluded that the different redox behavior exhibited by the samples is likely related to the formation of cobalt alloys and not to a particle size effect.

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 7 6 9 0 e7 6 9 8

7695

3.2.

Catalytic tests over honeycomb catalysts

Catalytic honeycombs were imaged by scanning electron microscopy (SEM) in frontal and transverse views as well as by confocal imaging and a good catalyst coating homogeneity was observed in all cases. The mean catalytic layer thickness was about 150 mm. Mechanical stability of the catalytically active phase in honeycomb catalysts is a critical issue for practical application purposes because coating loss and banking up should be completely avoided. The weight loss of the catalytic coatings in all honeycombs was less than 5% after 5 h of exposure to mechanical vibration (up to 50 Hz and 10 G), which means an excellent adherence. Catalytic honeycombs were tested for the ethanol steam reforming reaction (ESR) with a steam to carbon ratio of S/C 3 and W/FEtOH 42 min gcat mol1 (volume hourly space velocity, VHSV 2500 h1). Fig. 4 shows the evolution of ethanol conversion as well as the distribution of products at different reaction temperatures for all samples. For the sake of clarity, only the main products H2, CO2, CO, CH4, and C2H4O (acetaldehyde) have been plotted. Other products identied were ethylene and ethane in trace amounts (<0.5%) and dimethylketone up to 2% on a dry basis. The data reported for each catalyst at each temperature corresponds, at least, to ten different measurements and their uncertainty lie within the symbols used in the graphs. The general trend in the

performance of all catalyst is similar; at low reaction temperatures (573e673 K) ethanol transforms into a mixture of acetaldehyde, hydrogen, carbon oxides, and methane, whereas at higher temperatures (673e773 K) acetaldehyde is reformed and the only products of the reaction are hydrogen, carbon oxides, and methane. These results are in accordance with the reaction scheme discussed in the introduction section of three consecutive reactions. First ethanol dehydrogenates into acetaldehyde and hydrogen (equation (2)), and then acetaldehyde reacts with steam to yield hydrogen and carbon oxides (equation (3)) followed by the water gas shift reaction (equation (4)). The presence of methane in the reactor outlet is likely produced by the decomposition of ethanol (equation (5)) because cobalt catalysts are not active for methanation under ESR conditions at low to moderate temperatures [8]. However, the degree of acetaldehyde steam reforming, the amount of methane and the relative CO2/CO ratio at the reactor outlet is signicantly different for the different honeycomb catalysts tested. The extent of acetaldehyde reforming follows the trend: Co(Fe)/ZnO w Co(Cr)/ ZnO > Co(Na)/ZnO w Co(Ni)/ZnO > Co/ZnO > Co(Cu)/ZnO. It is observed that the incorporation of Cu promoter in the catalyst results in a strong enhancement of ethanol dehydrogenation, but the capability for the ulterior reforming of acetaldehyde with steam is inhibited. On the other hand, catalysts promoted with Cr, Fe, and Na are more active and

Fig. 4 e Catalytic performance of honeycombs Co/ZnO (a), Co(Fe)/ZnO (b), Co(Ni)/ZnO (c), Co(Cu)/ZnO (d), Co(Cr)/ZnO (e), and Co(Na)/ZnO (f) in the ethanol steam reforming. S/C [ 3, 0.33 mL C2H5OH minL1, VHSV [ 2500 hL1. Ethanol conversion C. Selectivity to H2 B, CO2 6, CO >, CH4 *, and acetaldehyde ,.

7696

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 7 6 9 0 e7 6 9 8

selective towards the reforming products, H2 and CO2, with respect to the bare Co/ZnO catalytic honeycomb. Promotion by Ni results in larger amounts of methane among the reaction products, specially at high temperature. From the TPR and OP results discussed in Section 3.1, it appears that catalysts that perform better in ESR are also those that exhibit a higher amount of cobalt which is able to participate in redox exchange mechanisms. This correlation was rst proposed in [35] for Co/ZnO catalysts promoted with Mn. Figure 5 shows the results attained at 673 and 773 K in a two-dimensional plot, where the amount of hydrogen obtained on a molar basis with respect to ethanol in the reactor inlet is plotted against the amount of carbon dioxide. From the stoichiometry of the steam reforming of ethanol (equation (1)), the expected molar ratio H2/CO2 is 3. This is indicated in the graph as a dashed line (the reforming line). Other competitive routes for ethanol transformation (ethanol decomposition, dehydration, etc.) result in deviations from

the reforming line, making this type of graph very useful, since the position of the different catalysts serve as a measure of both their activity and selectivity to the reforming products H2 and CO2. At low temperature (673 K, Fig. 5a), the honeycomb catalyst containing only ZnO-supported cobalt, Co/ZnO, was active for ethanol transformation, but it plots to the right side of the reforming line, thus indicating that ethanol reforming is accompanied by ethanol decomposition, originating H2/CO2 < 3. At this temperature, honeycomb catalyst Co(Ni)/ZnO plots even worse in the graph, according to a larger ethanol decomposition to methane. In contrast, sample Co(Cu)/ZnO exhibits a similar hydrogen yield but a much lower CO2/C2H5OH ratio. It plots to the left side of the reforming line (H2/CO2 > 3), which means a higher dehydrogenation activity with respect to the other catalytic honeycombs. Honeycomb promoted with Na and particularly honeycombs promoted with Cr and Fe progressively show higher hydrogen yields and plot closer to the reforming line, indicating that these samples are more active and selective for ESR with respect to the Co/ZnO catalytic honeycomb. As expected from thermodynamics (ESR is an endothermic process), at high temperature (773 K, Fig. 5b) all samples exhibit a better performance towards the reforming products. However, and in spite that all samples cluster around the reforming line at high H2/C2H5OH and CO2/C2H5OH values, catalytic honeycombs Co(Fe)/ZnO and Co(Cr)/ZnO still perform better than the non-promoted Co/ZnO sample. Over Co(Fe)/ZnO and Co(Cr)/ZnO catalytic honeycombs, complete ethanol transformation was attained at ca. 673 K and acetaldehyde transformed almost completely at 723 K (0.2% C2H4O for Co(Fe)/ZnO and 2.9% C2H4O for Co(Cr)/ZnO on a dry basis). At this temperature, the CO2/CO molar ratio was 5.3 and 3.8 for Co(Fe)/ZnO and Co(Cr)/ZnO, respectively, whereas in both samples CO2/CH4 molar ratios greater than 10 were recorded. Finally, catalysts were tested in the water gas shift reaction (WGS, equation (4)) under conditions simulating the outlet of an ethanol steam reformer (CO:H2:H2O 1:2:6). Table 2 shows the catalytic activity of all samples in terms of CO conversion temperature, both for CO conversion values of 25% (T25) and 50% (T50). As expected from the ESR results reported above, no methane was formed in the WGS catalytic tests. The values of T50 for the different samples did not substantially differ, and temperatures in the range 540e550 K were recorded. In contrast, catalyst Co(Fe)/ZnO showed higher activity for WGS

Table 2 e Catalytic performance of catalysts in the water gas shift reaction. CO:H2:H2O:N2 [ 1:2:6:14, GHSV [ 15,000 hL1. Catalyst
Co/ZnO Co(Fe)/ZnO Co(Ni)/ZnO Co(Cu)/ZnO Co(Cr)/ZnO Co(Na)/ZnO

T25 (K)
526 512 526 534 530 523

T50 (K)
542 549 542 547 545 540

Fig. 5 e Yields of H2 and CO2 in the ethanol steam reforming at 673 K (a) and 773 K (b). S/C [ 3, 0.33 mL C2H5OH minL1, VHSV [ 2500 hL1.

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 7 6 9 0 e7 6 9 8

7697

at low temperature, with T25 512 K vs. 526 K for Co/ZnO, suggesting that the low amount of CO obtained under ESR conditions over the Co(Fe)/ZnO catalytic honeycomb at low temperature was related to its WGS activity.

4.

Conclusions

Co/ZnO honeycomb catalysts promoted with Fe, Ni, Cu, Cr, and Na containing 10% w/w Co and M/Co w 0.1 are effective for hydrogen production at low temperature from ethanol steam reforming and water gas shift reaction. The presence of Fe facilitates the redox exchange between reduced and oxidized Co, which has a clear positive effect on both reactions. Also, incorporation of Cr and Na promoters results in a better catalytic performance towards ethanol steam reforming. In contrast, a negative effect in encountered with honeycomb samples doped with Ni and Cu, since the presence of Ni favors ethanol decomposition at low temperature and Cu promotes the formation of acetaldehyde. Alloying of Co with Fe, Ni, Cu, and Cr has been evidenced by high-resolution TEM and electron energy loss spectroscopy.

Acknowledgements
This work was supported by grant CTQ2009-12520. A.T. and C.d.L. thanks regione Friuli Venezia Giulia and MIUR for nancial support. J.L. is grateful to ICREA Academia Program.

references

[1] Deluga GA, Salge JR, Schmidt LD, Verykios XE. Renewable hydrogen from ethanol by autothermal reforming. Science 2004;303:993. [2] Frusteri F, Freni S. Bio-ethanol, a suitable fuel to produce hydrogen for a molten carbonate fuel cell. J Power Sources 2007;173:200. [3] Idriss H, Scott M, Llorca J, Chan SC, Chiu W, Sheng PY, et al. A Phenomenological Study of the Metal-Oxide Interface: The Role of Catalysis in Hydrogen Production from Renewable Resources. ChemSusChem 2008;1:905. [4] Haryanto A, Fernando S, Murali N, Adhikari S. Current status of hydrogen production techniques by steam reforming of ethanol: A review. Energy Fuels 2005;19:2098. [5] Vaidya PD, Rodrigues AE. Insight into steam reforming of ethanol to produce hydrogen for fuel cells. Chem Eng J 2006; 117:39. [6] Casanovas A, Saint-Gerons M, Griffon F, Llorca J. Autothermal generation of hydrogen from ethanol in a microreactor. Int J Hydrogen En 2008;33:1827. [7] Llorca J, Casanovas A, Trifonov T, Rodr guez A, Alcubilla R. First use of macroporous silicon loaded with catalyst lm for a chemical reaction: A microreformer for producing hydrogen from ethanol steam reforming. J Catal 2008;255: 228. [8] Casanovas A, De Leitenburg C, Trovarelli A, Llorca J. Catalytic monoliths for ethanol steam reforming. Catal Today 2008; 138:187.

[9] Dom nguez M, Taboada E, Molins E, Llorca J. Co-SiO2 aerogelcoated catalytic walls for the generation of hydrogen. Catal Today 2008;138:193. pez E, Llorca J. [10] Casanovas A, Dom nguez M, Ledesma C, Lo Catalytic walls and micro-devices for generating hydrogen by low temperature steam reforming of ethanol. Catal Today 2009;143:32. [11] Llorca J, Homs N, Sales J, Ram rez de la Piscina P. Efcient production of hydrogen over supported cobalt catalysts from ethanol steam reforming. J Catal 2002;209:306. [12] Llorca J, Ram rez de la Piscina P, Dalmon JA, Sales J, Homs N. CO-free hydrogen from steam-reforming of bioethanol over ZnO-supported cobalt catalysts - Effect of the metallic precursor. Appl Catal 2003;B 43:355. [13] Tuti S, Pepe F. On the catalytic activity of cobalt oxide for the steam reforming of ethanol. Catal Lett 2008;122:196. [14] Galetti AE, Gomez MF, Arrua LA, Marchi AJ, Abello MC. Study of CuCoZnAl oxide as catalyst for the hydrogen production from ethanol reforming. Catal Commun 2008;9:1201. [15] Vargas JC, Libs S, Roger AC, Kiennemann A. Study of Ce-ZrCo uorite-type oxide as catalysts for hydrogen production by steam reforming of bioethanol. Cat Today 2005;107:417. [16] Llorca J, Homs N, Sales J, Fierro JLG, Ram rez de la Piscina P. Effect of sodium addition on the performance of Co-ZnObased catalysts for hydrogen production from bioethanol. J Catal 2004;323:470. [17] Bichon P, Haugom G, Venvik HJ, Colmen A, Blekkan EA. Steam reforming of ethanol over supported Co and Ni catalysts. Top Catal 2008;49:38. [18] Lin SSY, Kim DH, Ha SY. Hydrogen production from ethanol steam reforming over supported cobalt catalysts. Catal Lett 2008;122:295. [19] Freni S, Cavallaro S, Mondello N, Spadaro L, Frustreri F. Production of hydrogen for MC fuel cell by steam reforming of ethanol over MgO supported Ni and Co catalysts. Catal Commun 2003;4:259. [20] Sun J, X-Qiu P, Wu F, Zhu W-T. H2 from steam reforming of ethanol at low temperature over Ni/Y2O3, Ni/La2O3 and Ni/ Al2O3 catalysts for fuel-cell application. Int J Hydrogen Energy 2005;30:437. [21] Marin o F, Baronetti G, Jobbagy M, Laborde M. Cu-Ni-K/ gamma-Al2O3 supported catalysts for ethanol steam reforming Formation of hydrotalcite-type compounds as a result of metal-support interaction. Appl Catal 2003;A 238: 41. [22] Batista MS, Santos RKS, Assaf EM, Assaf JM, Ticianelli EA. High efciency steam reforming of ethanol by cobalt-based catalysts. J Power Sources 2004;134:27. [23] Song H, Zhang L, Watson RB, Braden D, Ozkan US. Investigation of bio-ethanol steam reforming over cobaltbased catalysts. Catal Today 2007;129:346. [24] Wang H, Ye JL, Liu Y, Li YD, Qin YN. Steam reforming of ethanol over Co3O4/CeO2 catalysts prepared by different methods. Catal Today 2007;129:305. [25] Benito M, Padilla R, Rodr guez L, Sanz JL, Daza L. Zirconia supported catalysts for bioethanol steam reforming: Effect of active phase and zirconia structure. J Power Sources 2007; 169:167. [26] Kaddouri A, Mazzocchia C. A study of the inuence of the synthesis conditions upon the catalytic properties of Co/SiO2 or Co/Al2O3 catalysts used for ethanol steam reforming. Catal Commun 2004;5:339. [27] Haga F, Nakajima T, Miya H, Mishima S. Catalytic properties of supported cobalt catalysts for steam reforming of ethanol. Catal Lett 1997;48:223. [28] Llorca J, Ram rez de la Piscina P, Sales J, Homs N. Direct production of hydrogen from ethanolic aqueous solutions over oxide catalysts. Chem Commun; 2001:641.

7698

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 5 ( 2 0 1 0 ) 7 6 9 0 e7 6 9 8

[29] Llorca J, Dalmon JA, Ram rez de la Piscina P, Homs N. In situ magnetic characterisation of supported cobalt catalysts under steam-reforming of ethanol. Appl Catal 2003;A 243: 261. [30] Llorca J, Homs N, Ram rez de la Piscina P. In situ DRIFTmass spectrometry study of the ethanol steam-reforming reaction over carbonyl-derived Co/ZnO catalysts. J Catal 2004;227:556. [31] Llorca J, Ram rez de la Piscina P, Dalmon JA, Homs N. Transformation of CO3O4 during ethanol steam-re-forming. Activation process for hydrogen production. Chem Mater 2004;16:3573. [32] Homs N, Llorca J, Ram rez de la Piscina P. Low-temperature steam-reforming of ethanol over ZnO-supported Ni and Cu

catalysts - The effect of nickel and copper addition to ZnOsupported cobalt-based catalysts. Catal Today 2006;116:361. J, [33] Torres JA, Llorca J, Casanovas A, Dom nguez M, Salvado D. Steam reforming of ethanol at moderate Montane temperature: Multifactorial design analysis of Ni/La2O3Al2O3, and Fe- and Mn-promoted Co/ZnO catalysts. J Power Sources 2007;169:158. D. [34] Nedyalkova R, Casanovas A, Llorca J, Montane Electrophoretic deposition of Co-Me/ZnO (Me Mn,Fe) ethanol steam reforming catalysts on stainless steel plates. Int J Hydrogen En 2009;34:2591. [35] Casanovas A, de Leitenburg C, Trovarelli A, Llorca J. Ethanol steam reforming and water gas shift reaction over Co-Mn/ ZnO catalysts. Chem Eng J 2009;154:267.

Вам также может понравиться