Steam Reforming of Bio-Ethanol to Produce ... - Semantic Scholar

catalysts Article

Steam Reforming of Bio-Ethanol to Produce Hydrogen over Co/CeO2 Catalysts Derived from Ce1´xCox O2ý Precursors Yanyong Liu *, Kazuhisa Murata and Megumu Inaba Research Institute of Energy Frontier, National Institute of Advanced Industrial Science and Technology, AIST, Tsukuba West, Onogawa 16-1, Tsukuba, Ibaraki 305-8569, Japan; [email protected] (K.M.); [email protected] (M.I.) * Correspondence: [email protected]; Tel.: +81-29-861-4826; Fax: +81-29-855-0815 Academic Editors: Rafael Luque, Sudipta De and Alina M. Balu Received: 13 November 2015; Accepted: 27 January 2016; Published: 5 February 2016

Abstract: A series of Ce1´x Cox O2ý precursors were prepared by homogeneous precipitation using urea as a precipitant. The Co/CeO2 catalysts obtained from the Ce1´x Cox O2ý precursors were used for the steam reforming of ethanol to produce hydrogen. Co ions could enter the CeO2 lattices to form Ce1´x Cox O2ý mixed oxides at x ď 0.2 using the homogeneous precipitation (hp) method. CeO2 was an excellent support for Co metal in the steam reforming of ethanol because a strong interaction between support and metal (SISM) exists in the Co/CeO2 catalysts. Because Co/CeO2 (hp) prepared by homogeneous precipitation possessed a high BET surface area and small Co metal particles, Co/CeO2 (hp) showed a higher ethanol conversion than the Co/CeO2 catalysts prepared using the co-precipitation (cp) method and the impregnation (im) method. The selectivity of CO2 over Co/CeO2 (hp) increased with increasing reaction temperature at from 573 to 673 K, and decreased with increasing reaction temperature above 673 K due to the increase of CO formation. The carbonaceous deposits formed on the catalyst surface during the reaction caused a slow deactivation in the steam reforming of ethanol over Co/CeO2 (hp). The catalytic activity of the used catalysts could be regenerated by an oxidation-reduction treatment, calcined in air at 723 K and then reduced by H2 at 673 K. Keywords: bio-ethanol; steam reforming; hydrogen; urea hydrolysis; homogeneous precipitation; Co/CeO2 catalyst; Ce1´x Cox O2ý mixed oxide; solid-phase crystallization

1. Introduction Biomass-derived ethanol (so-called as bio-ethanol) is a renewable fuel because it can be produced by fermentation of biomass, such as starch, sugar, crop wastes, and so on. The concentration of ethanol is 12–15 wt. % in bio-ethanol by the fermentation of biomass. In using bio-ethanol as an additive to gasoline, high energy intensive and rather expensive processes have been used for recovering highly concentrated ethanol (>99.5 wt. %) due to the co-boiling of ethanol and water. On the other hand, hydrogen is an efficient and clean energy and is predicted to become a major source of energy in the future. Hydrogen can be produced by the steam reforming of ethanol (Equation (1)) and this reaction can be used in fuel cell systems [1–3]. C2 H5 OH ` 3H2 O Ñ 2CO2 ` 6H2

(1)

As shown in Equation (1), the molar ratio of ethanol to water is three and thus the concentration of ethanol in the mixed solution is about 46 wt. %. Because a 46 wt. % ethanol aqueous is easy to be

Catalysts 2016, 6, 26; doi:10.3390/catal6020026

www.mdpi.com/journal/catalysts

Catalysts 2016, 6, 26 Catalysts 2016, 6, 26

2 of 14 2 of 14

solution of ethanol) to produce hydrogen has been proposed to be an effective method for the obtained by the simple distillation of bio-ethanol, the steam reforming of bio-ethanol (aqueous solution utilization of bio-ethanol fuel. of ethanol) to produce hydrogen has been proposed to be an effective method for the utilization of Base metals (such as Co, Ni, and Cu) and noble metals (such as Rh, Au, and Ir) have been used bio-ethanol fuel. as the active metal catalysts for ethanol steam [3–25]. Among them, Co is an attractive active metal Base metals (such as Co, Ni, and Cu) and noble metals (such as Rh, Au, and Ir) have been due to its high catalytic performance and low price. The oxide support plays a significant role in the used as the active metal catalysts for ethanol steam [3–25]. Among them, Co is an attractive active performance of metal catalysts for ethanol steam. CeO2 is an excellent support for the metal catalysts metal due to its high catalytic performance and low price. The oxide support plays a significant in the steam reforming of ethanol. Moreover, the synthesis method greatly influences the catalytic role in the performance of metal catalysts for ethanol steam. CeO2 is an excellent support for the performance of a metal-supported catalyst. Impregnation and co-precipitation are the methods metal catalysts in the steam reforming of ethanol. Moreover, the synthesis method greatly influences usually to be used for preparing CeO2-supported metal catalysts in the steam reforming of ethanol. the catalytic performance of a metal-supported catalyst. Impregnation and co-precipitation are the On the other hand, the homogeneous precipitation method by urea hydrolysis has been reported as methods usually to be used for preparing CeO2 -supported metal catalysts in the steam reforming an excellent method for preparing Cu/ZnAlO catalyst in the steam reforming of methanol and the of ethanol. On the other hand, the homogeneous precipitation method by urea hydrolysis has been gas-shift reaction [26,27]. Moreover, the solid-phase crystallization (SPC) method using multireported as an excellent method for preparing Cu/ZnAlO catalyst in the steam reforming of methanol component precursors is an effective method for synthesizing highly active metal-supported and the gas-shift reaction [26,27]. Moreover, the solid-phase crystallization (SPC) method using catalysts [28]. We have used the concept of SPC method to prepare active catalysts for some important multi-component precursors is an effective method for synthesizing highly active metal-supported reactions [29–38]. In this study, we synthesized the Co/CeO2 catalysts derived from Ce1−xCoxO2−y catalysts [28]. We have used the concept of SPC method to prepare active catalysts for some important precursors by the homogeneous precipitation for the steam reforming of ethanol to produce reactions [29–38]. In this study, we synthesized the Co/CeO2 catalysts derived from Ce1´x Cox O2ý hydrogen. precursors by the homogeneous precipitation for the steam reforming of ethanol to produce hydrogen. 2. 2. Results Resultsand andDiscussion Discussion 2.1. 2.1. Catalyst Catalyst Characterization Characterization Figure shows the the X-ray X-raydiffraction diffractionpatterns patternsofofCeCe CoxxO O22−y homogeneous Figure 11 shows prepared by by homogeneous ý prepared 1´1−x x Co precipitation calcination at 723 K K for for 33 h. h. The CeO22 support precipitation after after calcination at 723 The CeO support showed showed four four main main reflections reflections corresponding (111), (200), (200), (220) (220) and and (311) (311) crystallographic crystallographic planes, planes, implying implying that that the the CeO CeO22 corresponding to to the the (111), support had a fluorite structure with cubic (fcc) cells. A Co 3 O 4 phase ((111) crystallographic plane) support had a fluorite structure with cubic (fcc) cells. A Co3 O4 phase ((111) crystallographic plane) appeared Co xOx2−y oxides when x was larger thanthan 0.2, 0.2, andand the appeared at at 36.9° 36.9˝ in in the the patterns patterns of ofthe theCe Ce1−x O2mixed oxides when x was larger ý mixed 1´ x Co intensity of the 4 peak increased with increasing Co content in the samples. the intensity of Co the3O Co O peak increased with increasing Co content in the samples. 3 4

Figure 1. O2−y preparedby by homogeneous homogeneous precipitation precipitation after after 2ýprepared Figure 1. X-ray X-ray diffraction diffractionpatterns patternsofofCeCe 1−xCo CoxxO 1´x calcination at at 723 723 K K for for 33 h. h. calcination

Catalysts 2016, 6, 26

3 of 14


3 of 14

Figure 2 shows the lattice constant of CeO2 in the Ce1−xCoxO2−y prepared by homogeneous precipitation after calcination at 723 K for 3 h. The lattice constants of the CeO2-based materials were Figure 2 shows the lattice constant of CeO2 in the Ce1´x Cox O2ý prepared by homogeneous calculated based on the four main reflections in the XRD patterns. The lattice constant of CeO2 was precipitation after calcination at 723 K for 3 h. The lattice constants of the CeO2 -based materials were 0.5410 nm without Coon addition. The reflections lattice constant of patterns. CeO2 decreased when Co was was added, but calculated based the four main in the XRD The lattice constant of CeO 2 almost remained nm)constant when xofwas than 0.2 in the Ce1−xCoxO2−y mixed 0.5410 nm constant without Co(around addition.0.5354 The lattice CeOlarger 2 decreased when Co was added, but almost remained constant (around 0.5354 nm) when x was larger than 0.2 thethe Ce1ĆeO oxides. These results indicated that the smaller Co ions could enterin in to form a 2ý mixed x Cox2Olattice oxides. These results indicated that the smaller Co ions could enter in the CeO lattice to form a a part 2 homogeneous Ce1−xCoxO2−y solid solution at x ≤ 0.2. For the Ce1−xCoxO2−y samples with x > 0.2, homogeneous Ce1´x Cox O2ý solid solution at x ď 0.2. For the Ce1´x Cox O2ý samples with x > 0.2, a of Co ions entered in the CeO2 lattice to form a solid solution and the remaining Co ions formed large part of Co ions entered in the CeO2 lattice to form a solid solution and the remaining Co ions formed Co3O4 particles surface ofsurface CeO2. of CeO . large Co on O the particles on the 3

4

2

2. Lattice constant of CeO theCe Ce1−x CoxxO O2−y prepared byby homogeneous precipitation after 2ýprepared 2 in 1´xCo Figure 2.Figure Lattice constant of CeO 2 in the homogeneous precipitation after calcination at 723 K for 3 h. calcination at 723 K for 3 h.

1 summarizes the physical properties of various Co-supported catalysts with 7.7 wt. % Table 1 Table summarizes the physical properties of various Co-supported catalysts with 7.7 wt. % Co Co loading. All samples were reduced in H2 at 673 K for 10 h. The actual Co loadings, which were loading. measured All samples were reduced in H2 at 673 K for 10 h. The actual Co loadings, which were by ICP elemental analysis, were very similar to the Co loadings designed in the catalysts. measured by ICP elemental analysis, were very similar to the Co loadings designed in the catalysts. Table 1. Physical characters of various Co-supported catalysts with 7.7 wt. % Co loading.

Table 1. Physical characters of various Co-supported catalystsa with 7.7 wt. % Co loading. a 2 ´1 Sample

BET Surface Area (m ¨ g

)

Co Particle Size (nm)

135 (m ·g ) 7.6 Size (nm) SampleCo/CeO2 (hp)BET Surface Area Co Particle Co/CeO2 (cp) 102 9.5 Co/CeO2 (hp) 135 7.6 Co/CeO2 (im) 88 17.7 14.3 Co/CeO2 Co/Al (cp) 2 O3 (im) 102 147 9.5 a Obtained by CO adsorption. Co/CeO2 (im) 88 17.7 Co/Al2O3 (im) 147 14.3 2

−1

Co Dispersion Degree (%)

a

12.6 Co Dispersion Degree (%) a 10.1 12.6 5.4 6.7 10.1 5.4 6.7

The 7.7 wt. % Co/CeO2 (hp) catalyst obtained by reducing the Ce0.8 Co0.2 O2ý precursor showed a Obtained by CO adsorption. a BET surface area of 135 m2 ¨ g´1 , which was higher than those of Co/CeO2 (cp) (102 m2 ¨ g´1 ) and 2 ´ 1 (im) (88 m ¨ g2 (hp) ). Cocatalyst ions wereobtained doped in the the Ce0.80.2Co O2ý precursor, 2 positions TheCo/CeO 7.7 wt.2 % Co/CeO by CeO reducing theinCe 0.8Co O0.2 2−y precursor showed a which caused a decrease in the crystallization degree of CeO -based support. Further, the decrease of 2 BET surface area of 135 m2·g−1, which was higher than those of Co/CeO2 (cp) (102 m2·g−1) and Co/CeO2 the crystallization degree could cause an increase of the BET surface area in Co/CeO2 (hp). The Co (im) (88 particles’ m2·g−1). size Comeasured ions were doped in the positions in for theCo/CeO Ce0.8Co 0.2O2−y precursor, which by CO absorption wasCeO 7.6, 29.5, and 17.7 nm 2 (hp), Co/CeO2 (cp), caused aand decrease the respectively. crystallization degree of CeO2-based support.may Further, the decrease of the Co/CeOin The use of multi-component precursors yield well-dispersed 2 (im), metal particles on the surface of supports after calcination and reduction; this property, known as The Co crystallization degree could cause an increase of the BET surface area in Co/CeO 2 (hp).

particles’ size measured by CO absorption was 7.6, 9.5, and 17.7 nm for Co/CeO2 (hp), Co/CeO2 (cp), and Co/CeO2 (im), respectively. The use of multi-component precursors may yield well-dispersed metal particles on the surface of supports after calcination and reduction; this property, known as solid-phase crystallization (SPC), is important in the preparation of highly active metal-supported


4 of 14

solid-phase crystallization (SPC), is important in the preparation of highly active metal-supported catalysts [27–33]. Using the homogeneous precipitation, Co ions could enter in the CeO2 lattice to form a Ce1´x Cox O2ý solid solution after calcination in air at 723 K. On the other hand, Co ions could not Catalysts 26 2 lattice using the impregnation, which caused the large Co particles to be formed 4 of in 14 enter in2016, the 6, CeO Co/CeO2 (im). Figure shows the the H H2-TPR profiles of various Co-based samples after calcination at 723 K for Figure 33 shows 2 -TPR profiles of various Co-based samples after calcination at 723 K for 33 h. The Co loading was 7.7 wt. % 2-TPR is a useful tool to trace the reduction of h. The Co loading was 7.7 wt. %for foreach eachsample. sample.HH 2 -TPR is a useful tool to trace the reduction the oxide phases. The Co 3O4 particles formed on the support surface upon calcination are very of the oxide phases. The Co3 O4 particles formed on the support surface upon calcination are very difficult to be be reduced reduced to to active active Co Co00 species species in in the the H H2 flow [11,39]. In Figure 3, all samples showed the difficult to 2 flow [11,39]. In Figure 3, all samples showed the reduction peaks at low temperature (with the maximum about513 513K)K) assigning reduction reduction peaks at low temperature (with the maximum atatabout assigning to to thethe reduction of 3+ 2+ 2+ → Co0. of Co → Co and at high temperature (from 563 to 723 K) assigning to the reduction of Co 3+ 2+ 2+ Co Ñ Co and at high temperature (from 563 to 723 K) assigning to the reduction of Co Ñ Co0 . 2+ → Co0 contained two unseparated peaks for each sample due to the Moreover, the reduction reduction of of Co Co2+ Moreover, the Ñ Co0 contained two unseparated peaks for each sample due to the 2+ species absorbed on the supports [11,39,40]. The reduction stepwise reduction steps of various Co2+ stepwise reduction steps of various Co species absorbed on the supports [11,39,40]. The reduction temperature of Co 3O4 on various catalysts followed the order of Co/CeO2 (hp) < Co/CeO2 (cp) < temperature of Co3 O4 on various catalysts followed the order of Co/CeO2 (hp) < Co/CeO2 (cp) < Co/CeO Co/Al 2O3 (im). The catalysts were pretreated in H2 flow at 673 K (for 10 h) in this study Co/CeO2 2(im) (im)< < Co/Al 2 O3 (im). The catalysts were pretreated in H2 flow at 673 K (for 10 h) in this in order to avoid the severe sintering of Co particles at at high reduction study in order to avoid the severe sintering of Co particles high reductiontemperature. temperature.As As shown shown in in 2+ → Co0 (H) (strong absorbed Co2+ to Co0) in the H2-TPR profiles Figure 3, the reduction peak of Co 2+ 0 2+ 0 Figure 3, the reduction peak of Co Ñ Co (H) (strong absorbed Co to Co ) in the H2 -TPR profiles 2+ species in of Co/Al2OO3 (im) and Co/CeO2 (im) reached over 673 K, which means that some Co2+ of Co/Al species in 2 3 (im) and Co/CeO2 (im) reached over 673 K, which means that some Co 0 Co/Al 2O3 (im) and Co/CeO2 (im) could not be reduced to active Co 0species under the pretreatment Co/Al2 O3 (im) and Co/CeO2 (im) could not be reduced to active Co species under the pretreatment condition in this this study study (H (H2 reduction at 673 K). condition in 2 reduction at 673 K).

Figure H22-TPR -TPR profiles profiles of of various various Co-based Co-based samples samples after after calcination calcination at at 723 723 K K for for 33 h. h. Figure 3. 3. H

Figure shows the the TEM TEM image image of of 7.7 7.7 wt. wt. % % Co/CeO Co/CeO22 (hp) Figure 44 shows (hp) catalyst. catalyst. The The sample sample was was pre-reduced pre-reduced in H22 at for 10 10 h. h. The nm, which which is is in H at 673 673 K K for The particle particle size size of of CeO CeO22 from from the the TEM TEM image image is is about about 10 10 nm, consistent the size size of of the the CeO CeO22 particles particles calculated calculated from from the the XRD XRD pattern. pattern. Moreover, Moreover, the the CeO CeO22 consistent with with the particles are are distributed distributed uniformly, uniformly, which which is is an an advantage advantage of of the the homogeneous particles homogeneous precipitation precipitation method method for catalysts. Because for preparing preparing catalysts. Because the the size size of of the the Co Co particles particles (7.6 (7.6 nm nm from from CO CO adsorption) adsorption) is is similar similar to to the size of of the the CeO CeO22 particles (about 10 10 nm nm from from XRD XRD pattern) pattern) in in the the 7.7 7.7 wt. wt. % % Co/CeO Co/CeO22 (hp) the size particles (about (hp) catalyst, catalyst, it is difficult difficult to to distinguish distinguish Co Co particles particles from from CeO CeO22 particles particles in in the the TEM TEM image. image. it is


5 of 14


5 of 14

Figure 4. TEM image of 7.7 wt. % Co/CeO2 (hp) catalyst.

Figure 4. TEM image of 7.7 wt. % Co/CeO2 (hp) catalyst.

2.2. Catalyst Activity

2.2. Catalyst Activity

Table 2 summarizes the reaction results over various catalysts for the steam reforming of ethanol

Table the showed reactiona results over various catalysts the steam reforming of ethanol at 673 2K.summarizes Co/Al2O3 (imp) low conversion of 25.5% at 673 for K. A relatively large amount of C 2 H 4 was formed over Co/Al 2 O 3 (im) because the acidity of Al 2 O 3 support caused the dehydration of at 673 K. Co/Al2 O3 (imp) showed a low conversion of 25.5% at 673 K. A relatively large amount of Co/CeO 2 (im) showed an ethanol conversion (40.3%) much higher than that over Co/Al2O3 C2 H4ethanol. was formed over Co/Al 2 O3 (im) because the acidity of Al2 O3 support caused the dehydration of (im) (25.5%) prepared by the same impregnation method. This indicates thatthan CeOthat 2 is an effective ethanol. Co/CeO2 (im) showed an ethanol conversion (40.3%) much higher over Co/Al2 O3 support for Co metal in the steam reforming of ethanol. The CeO2-supported catalysts possess strong (im) (25.5%) prepared by the same impregnation method. This indicates that CeO2 is an effective interaction between support and metal (SISM) during the reaction [41]. The SISM effect gave the support for Co metal in the steam reforming of ethanol. The CeO2 -supported catalysts possess strong CeO2-supported metal catalysts high activities for many reactions [29–43]. The synthesis method interaction andperformance metal (SISM) during2 catalysts the reaction The SISM effect gave the greatly between influencessupport the catalytic of Co/CeO for the[41]. steam reforming of ethanol. CeO2The -supported catalysts higha activities many which reactions Thethan synthesis method Co/CeO2 metal (hp) catalyst showed conversionfor of 85.9%, was [29–43]. much higher those over greatly influences the catalytic performance of Co/CeO catalysts for the steam reforming of ethanol. Co/CeO2 (cp) (68.4%) and Co/CeO2 (im) (40.3%) with the 2 same Co loading under the same reaction conditions. Use catalyst of multi-component precursors of may give which well-dispersed metal particles the over The Co/CeO showed a conversion 85.9%, was much higher thanon those 2 (hp) surface of supports calcination and reduction, which is same called Co as solid-phase crystallization Co/CeO (68.4%) after and Co/CeO (40.3%) with the loading under the same(SPC) reaction 2 (cp) 2 (im) methodUse for preparing highly activeprecursors metal-supported catalysts [27–33]. Formetal Co/CeO 2 (hp) catalyst, Co conditions. of multi-component may give well-dispersed particles on the surface ions were introduced into the CeO 2 framework to form Ce0.8Co0.2O2−y precursor after calcination. of supports after calcination and reduction, which is called as solid-phase crystallization (SPC) method Therefore, after reducing Ce0.8Co0.2O2−y solid solution in H2, Co metal particles uniformly formed on for preparing highly active metal-supported catalysts [27–33]. For Co/CeO2 (hp) catalyst, Co ions were the CeO2 surface and strongly interact with the CeO2 support. As shown in Table 2, Co/CeO2 (hp) introduced into the CeO2 framework to form Ce0.8 Co0.2 O2ý precursor after calcination. Therefore, obtained the largest ethanol conversion and the largest selectivity to CO2 among various catalysts, after which reducing Ce0.8 Cothe solution in H2 , Co metal particles uniformly formed on the CeO2 y solid 0.2 O meant that H22 ´ yield of Co/CeO 2 (hp) was much higher than other catalysts. surface and strongly interact with the CeO2 support. As shown in Table 2, Co/CeO2 (hp) obtained the Table 2. Reaction and results over variousselectivity catalysts for to theCO steam reforming of ethanol at 673 K awhich . largest ethanol conversion the largest various catalysts, meant 2 among that the H2 yield of Conversion Co/CeOof higher than other catalysts. 2 (hp) was much Selectivity to C-products (%) H2 Yield (Molar Ratio of Catalyst b

Ethanol (%) CO2 CO CH4 C2H4 OC c Co/Al 2O3 (im) 25.5 19.5 13.9reforming 3.3 Table 2. Reaction results over54.9 various catalysts for8.3 the steam Co/CeO2 (im) 40.3 70.6 15.2 4.4 3.7 4.8 Co/CeO2 (cp) 68.4 73.3 14.6 3.7 2.8 4.3 Selectivity to C-products (%) Conversion of Co/CeO 2 (hp) 85.9 76.2 12.1 3.5 3.1 4.2 b Catalyst Ethanol (%) CO2 CO CH4 C2 H4 OC c a

Catalyst amount: 1 g; C2H5OH: 17 mL·min−1; H2O: 50 mL·min−1; N2: 17 mL·min−1.

25.5 54.9 19.5 Co/Al2 O3 (im) 3CHO, CH3COCH3, and so on. compounds, including CH40.3 70.6 15.2 Co/CeO 2 (im) 68.4 73.3 14.6 Co/CeO2 (cp) 85.9 76.2 12.1 Co/CeO2 (hp) a

8.3 4.4 3.7 3.5

13.9 3.7 2.8 3.1

b

[H2]out/[EtOH]in)

of ethanol at1.1673 K a .

2.1 3.6 H2 Yield (Molar 4.7 Ratio of [H2 ]out /[EtOH]in )

Co loading: 7.7 wt. %.

c

OC: oxygenated

3.3 4.8 4.3 4.2

Catalyst amount: 1 g; C2 H5 OH: 17 mL¨ min´1 ; H2 O: 50 mL¨ min´1 ; N2 : 17 mL¨ min´1 . 7.7 wt. %. c OC: oxygenated compounds, including CH3 CHO, CH3 COCH3 , and so on.

1.1 2.1 3.6 4.7 b

Co loading:


6 of 14

The main C-containing products of the steam reforming of ethanol are CO2 , CO, C2 H4 , CH4 , CH3 CHO, and CH3 COCH3 . CO2 is formed by the steam reforming of ethanol (Equation (1)) and Catalysts 2016,shift 6, 26 reaction (Equation (2)). There are several ways to form CO: the ethanol 6 of 14 the water-gas steam reforming to syngas (Equation (3)); the ethanol cracking (Equation (4)), and the reverse water-gas The main C-containing products of the steam reforming of ethanol are CO2, CO, C2H4, CH4, shift CH reaction (Equation (5)). CH is formed by the ethanol cracking (Equation (4)). C2 Hand formed 4 is the 3CHO, and CH3COCH3. CO42 is formed by the steam reforming of ethanol (Equation (1)) by the ethanol shift dehydration (Equation(2)). (6)).There Acetaldehyde the main oxygenated compound water-gas reaction (Equation are severalisways to form CO: the ethanol steamand it is formed by the ethanol dehydrogenation (Equation (7)). Acetone is formed by the acetaldehyde reforming to syngas (Equation (3)); the ethanol cracking (Equation (4)), and the reverse water-gas shift reaction(Equation (Equation (8)). (5)). CH 4 is formedvalue by theof ethanol cracking (Equation H4 /[EtOH] is formedinby decomposition The largest H2 yield (molar ratio of(4)). [HC ) is six 2 ]2out ethanol dehydration (Equation and (6)). Acetaldehyde theCO main oxygenated compound and it is whenthe both the C the selectivityis to 100% in the steam reforming of 2 H5 OH conversion 2 are formed by the(1)). ethanol dehydrogenation (Equation (7)). Acetone is formed by the acetaldehyde ethanol (Equation The H yield decreases when the side-reactions (Equations (2)–(8)) occur during 2 decomposition (Equation (8)). The largest value of H2 yield (molar ratio of [H2]out/[EtOH]in) is six when the reaction. both the C2H5OH conversion and the selectivity to CO2 are 100% in the steam reforming of ethanol CO ` H2 O Ñ CO2 ` H2 (2) (Equation (1)). The H2 yield decreases when the side-reactions (Equations (2)–(8)) occur during the reaction. C2 H5 OH ` H2 O Ñ 2CO ` 4H2

(3)

H2CH O → CO 2 + H2 C2 H5CO OH+ Ñ 4 ` CO ` H2

(4)

(2)

C2H5OH + H2O → 2CO + 4H2

(3)

C2H5OH → CH4 + CO + H2

(4)

CO2 + H2 → CO + H2O

(5)

C2 H5 OH Ñ CH3 CHO ` H2 C2H5OH → C2H4 + H2O

(6)

(7)

2CH3 CHO `C2H Ñ→ CH `2 CO2 ` 2H2 3 COCH H25O OH CH 3CHO +3H

(7)

(8)

CO2 ` H2 Ñ CO ` H2 O

C2 H5 OH Ñ C2 H4 ` H2 O

(5) (6)

2CH3CHO + HCo/Al 2O → CH 3 + CO2 +the 2H2Co/CeO2 (im) catalyst(8) As shown in Table 2, compared to the (im) catalyst, formed 2 O33COCH a less amount of C H because the basicity of CeO support suppressed the ethanol dehydration 2 4 2 As shown in Table 2, compared to the Co/Al2O3 (im) catalyst, the Co/CeO2 (im) catalyst formed (Equation (6)). Moreover, because high suppressed oxygen storage capacity, the ethanol 2 support a less amount of C2H4 because theCeO basicity of CeOhas 2 support the ethanol dehydration decomposition (Equation (4)) was suppressed in the has Co/CeO catalysts by transferring mobile oxygen (Equation (6)). Moreover, because CeO2 support high 2oxygen storage capacity, the ethanol decomposition (4))Hence, was suppressed in theof Co/CeO 2 catalysts by transferring mobile oxygen species to Co from(Equation CeO2 [24]. the amounts CO and CH4 formed over Co/CeO (im) were 2 species to Co from CeO Hence, the amounts of CO and CH 4 formed over Co/CeO2 (im) were less than those formed over2 [24]. Co/Al O (im) during the reaction. 2 3 less than5 those formed over of Co/Al O3 (im) during reaction.(hp) catalyst for the steam reforming of Figure shows the effect Co2loading in thethe Co/CeO 2 Figure 5 shows the effect of Co loading in the Co/CeO2 (hp) catalyst for the steam reforming of ethanol at 673 K. The reaction conditions are the same as those in Table 2. The ethanol conversion ethanol at 673 K. The reaction conditions are the same as those in Table 2. The ethanol conversion increased with increasing Co loading and showed the maximum value at a 17.4 wt. % Co loading increased with increasing Co loading and showed the maximum value at a 17.4 wt. % Co loading (precursor: Ce Ce Co 0.4 O2ý ). However, the ethanol conversion did not show an obvious increase at (precursor:0.6 0.6Co0.4O2−y). However, the ethanol conversion did not show an obvious increase at aboveabove 7.7 wt. % Co loading inin the (hp)catalyst. catalyst. 7.7 wt. % Co loading theCo/CeO Co/CeO22 (hp)

Figure 5. Effect of Co loading in the Co/CeO2 (hp) catalyst for the steam reforming of ethanol at 673 K.

Figure 5. Effect of Co loading in the Co/CeO2 (hp) catalyst for the steam reforming of ethanol at 673 K.


7 of 14 7 of 14

The turnover frequency (TOF) was defined as the number of converted ethanol molecules per The turnover frequency defined the number of converted molecules per Co surface site per second. In (TOF) Figurewas 5, the TOF as value was calculated from ethanol the number of metallic surface site per second. Figure 5, the TOF wasofcalculated from number of at metallic CoCo atoms (measured by CO In chemisorption) andvalue the rate the convert C2the H5 OH atoms 673 K Co over atoms (measured by CO chemisorption) and the rate of the convert C2H5OH atoms at 673 K over the the Co/CeO2 (hp) catalyst. The TOF values decreased with increasing Co loading in the Co/CeO2 Co/CeO2 (hp) catalyst. The TOF values decreased with increasing Co loading in the Co/CeO2 (hp) (hp) catalyst. The TOF value of 7.7 wt. % Co/CeO2 (hp) (precursor: Ce0.8 Co0.2 O2ý ) was just slightly catalyst. The TOF value of 7.7 wt. % Co/CeO2 (hp) (precursor: Ce0.8Co0.2O2−y) was just slightly lower lower that of 3.6 wt. % Co/CeO2 (hp) (precursor: Ce0.9 Co0.1 O2ý ). However, the TOF value of that of 3.6 wt. % Co/CeO2 (hp) (precursor: Ce0.9Co0.1O2−y). However, the TOF value of 12.2 wt. % 12.2 wt. % Co/CeO2 (hp) (precursor: Ce0.7 Co0.3 O2ý ) greatly decreased as comparison with that of Co/CeO2 (hp) (precursor: Ce0.7Co0.3O2−y) greatly decreased as comparison with that of 7.7 wt. % 7.7 wt. % Co/CeO2 (hp). There was a limitation of the Co capacity entered in the CeO2 lattice in Co/CeO2 (hp). There was a limitation of the Co capacity entered in the CeO2 lattice in Ce1−xCoxO2−y. Ce1´x Cox O2ý . The Co ions which did not dissolve into the CeO2 lattices formed large Co particles The Co ions which did not dissolve into the CeO2 lattices formed large Co particles after reduction after and activity of large Colow particles lowreforming for the steam reforming of ethanol. The andreduction the activity of the large Co particles was for thewas steam of ethanol. The Co ions could Conot ions could not the completely CeO2 lattice inOthe Ce0.7 Co0.3 which O2ý precursor, which dissolve into dissolve the CeO2into lattice incompletely the Ce0.7Co0.3 2−y precursor, caused a large caused a large decrease in the TOF value over the 12.2 wt. % Co/CeO (hp) catalyst. decrease in the TOF value over the 12.2 wt. % Co/CeO2 (hp) catalyst. 2 Figure 6 shows the the effect of reaction temperature on theonethanol conversion over various catalysts. Figure 6 shows effect of reaction temperature the ethanol conversion over various The reactionThe conditions the same asthe those in as Table 2. in ForTable the reaction without a without catalyst,athe ethanol catalysts. reaction are conditions are same those 2. For the reaction catalyst, conversion was very low ( Co/CeO (im) > Co/Al (im). 2By usingBy theusing Co/CeO derived of Co/CeO 2 (hp) > Co/CeO 2 (cp) Co/CeO 2 (im) >3Co/Al O3 (im). the Co/CeO 2 (hp) catalyst 2 (cp) 2> 2O 2 (hp) catalyst from Ce1´xfrom Cox OCe a high ethanol conversion had been achieved at aachieved low temperature derived CoxO2−y precursor, a high ethanol conversion had been at a lowfor 2ý1−xprecursor, for theof steam reforming of ethanol. thetemperature steam reforming ethanol.

Figure6.6.Effect Effectofofreaction reactiontemperature temperature on on the the ethanol Figure ethanol conversion conversionover overvarious variouscatalysts. catalysts.

Figure 7 shows the effect of reaction temperature on the selectivity to C-containing products Figure 7 shows the 2effect reaction temperature on the to C-containing products over over 7.7 wt. % Co/CeO (hp). of The reaction conditions are the selectivity same as those in Table 2. The selectivity 7.7towt. % Co/CeO (hp). The reaction conditions are the same as those in Table 2. The selectivity CH4 did not 2show a large change at various reaction temperatures. Both the selectivity toto CH a large change various reaction temperatures. the selectivity to oxygenated 4 did not show oxygenated compounds and the at selectivity to C2H4 decreased with Both increasing reaction temperature. compounds and the selectivity to C H decreased with increasing reaction temperature. On the other 4 increased with increasing reaction temperature, which On the other hand, the selectivity to2 CO meant hand, the selectivity to CO increased with increasing reaction temperature, which meant that the that the selectivity to CO2 had a maximum value of 673 K. selectivity to CO2 had a maximum value of 673 K.


8 of 14 8 of 14


8 of 14

Figure 7. Effect on selectivity to over 7.7 wt. % Co/CeO 2 (hp). Figure 7. of reaction Effect oftemperature reaction temperature onC-containing selectivity products to C-containing products over 7.7 wt. % Co/CeO (hp). 2 Figure 7. Effect of reaction temperature on selectivity to C-containing products over 7.7 wt. % Co/CeO2 (hp).

Figure 8 shows the effect of reaction temperature on H2 yield over 7.7 wt. % Co/CeO2 (hp). The reaction conditions same as thosetemperature in Table 2. Because the ethanol increased Figure showsare thethe effect of reaction reaction temperature onHH22yield yield over7.7 7.7conversion wt.% %Co/CeO Co/CeO The Figure 88 shows the effect of on over wt. (hp).with The 22 (hp). increasing reaction temperature, the H 2 yield increased with increasing the reaction temperature at a reaction Table 2. 2. Because Because the the ethanol ethanol conversion conversion increased reaction conditions conditions are are the the same same as as those those in in Table increased with with reaction temperature below 723 K.the However, selectivity CO greatlythe increased the ethanol increasing reaction the H22 yield increased with reaction temperature at increasing reaction temperature, temperature, H yieldthe increased withtoincreasing increasing the reactionwhen temperature at aa conversion reached to near 100% at a reaction temperature above 723 K. The CO formation decreased reaction the selectivity reaction temperature temperature below below 723 723 K. K. However, However, the selectivity to to CO CO greatly greatly increased increased when when the the ethanol ethanol the selectivity to COto 2, near and thus thetemperature H2 yield for above the reaction. Hence, the maximum value conversion reached 100% at 723 CO decreased conversion reached to near 100%decreased at aa reaction reaction temperature above 723 K. K. The The CO formation formation decreased of 2 yield wasto 723decreased K over 7.7the wt. 2 (hp). the selectivity CO2,, and andatthus thus decreased theH% H2Co/CeO yieldfor for thereaction. reaction.Hence, Hence, the maximum value theHselectivity toobserved CO yield the the maximum value of 2

2

of 2 yield was observed 723 K over Co/CeO 2 (hp). H2Hyield was observed at at 723 K over 7.77.7 wt.wt. %% Co/CeO 2 (hp).

Figure 8. Effect of reaction temperature on H2 yield over 7.7 wt. % Co/CeO2 (hp). Figure 8. 8. Effect Effect of of reaction reaction temperature temperatureon onH H2 yield yield over over 7.7 7.7 wt. wt. % % Co/CeO Co/CeO2 (hp). Figure 2 2 (hp).

2.3. Catalyst Deactivation and Regeneration

2.3. Catalyst Regeneration Figure 9Deactivation shows theand catalyst deactivation and regeneration for ethanol steam reforming over 7.7 wt. % Co/CeO 2 (hp) at 673 K. The reaction conditions are the same as thosesteam in Table 2. Figure 9a Figure 9 shows the catalyst deactivation and regeneration for ethanol reforming over shows the time course of ethanol steam reforming over fresh catalyst at 673 K. The catalyst showed 7.7 wt. % Co/CeO2 (hp) at 673 K. The reaction conditions are the same as those in Table 2. Figure 9a shows the time course of ethanol steam reforming over fresh catalyst at 673 K. The catalyst showed


9 of 14

2.3. Catalyst Deactivation and Regeneration Figure 9 shows the catalyst deactivation and regeneration for ethanol steam reforming over 7.7 wt. % Co/CeO2 (hp) at 673 K. The reaction conditions are the same as those in Table 2. Figure 9a shows the time of ethanol steam reforming over fresh catalyst at 673 K. The catalyst Catalysts 2016, 6,course 26 9 ofshowed 14 the initial conversions of 85.9% and it decreased to 70.4% after reaction at 673 K for 120 h. As for the the initial conversions of are 85.9% andpossible it decreased to 70.4% reactionCo at oxidation, 673 K for 120 h. As for the catalyst deactivation, there three reasons: Coafter sintering, and carbonaceous catalyst deactivation, there are three possible reasons: Co sintering, Co oxidation, and carbonaceous deposits. Because the CO absorption measurement indicated that the size of the Co particles did not deposits. Because the CO absorption measurement indicated that the size of the Co particles did not change after reaction at 673 K for 120 h, it is unlikely that the deactivation would be caused by the change after reaction at 673 K for 120 h, it is unlikely that the deactivation would be caused by the sintering of the Co. The deactivated catalyst (after 120 h on stream at 673 K) was reduced by H2 at sintering of the Co. The deactivated catalyst (after 120 h on stream at 673 K) was reduced by H2 at 673 K673 forK10 and then used again reformingofofethanol ethanol at 673 K. The course forh10 h and then used againfor forthe thesteam steam reforming at 673 K. The timetime course was was shown in Figure 9b. The catalytic activity did not recover and the initial conversion was about shown in Figure 9b. The catalytic activity did not recover and the initial conversion was about 70%. 70%. This This indicates thatthat thethe oxidation reasonfor forthe thecatalyst catalyst deactivation. Moreover, indicates oxidationofofCo Coisisnot not the the reason deactivation. Moreover, the the deactivated catalyst (after streamatat673 673 was regenerated by calcination airK at deactivated catalyst (after120 120hhon on stream K)K) was regenerated by calcination in air atin 723 for723 K and then then reduction at 673 K for 10 h. The time course was shown in Figure 9c. The for 33hhand reductionininH2Hflow flow at 673 K for 10 h. The time course was shown in Figure 9c. 2 catalytic activity recovered to its initial conversion of about 85%. These results indicated that the The catalytic activity recovered to its initial conversion of about 85%. These results indicated that carbonaceous deposits on on thethe catalyst surface caused the catalyst deactivation during the the carbonaceous depositsformed formed catalyst surface caused the catalyst deactivation during reaction. the reaction.

Figure 9. Catalyst deactivation and ethanolsteam steamreforming reforming over wtCo/CeO % Co/CeO Figure 9. Catalyst deactivation andregeneration regeneration for ethanol over 7.7 7.7 wt % 2 2 (hp) at K. 673(a) K.Fresh (a) Fresh catalyst; used catalyst(after (after reaction reaction at (hp) at 673 catalyst; (b)(b) thethe used catalyst at 673 673KKfor for120 120h)h)was wasreduced reduced by 67310 K h; for(c) 10the h; (c) the catalyst used catalyst reaction at 673 K for was calcinedininair airatat723 K H2Katfor H2 atby 673 used (after(after reaction at 673 K for 120120 h) h) was calcined 723 K for 3 h and then reduced by H 2 at 673 K for 10 h. for 3 h and then reduced by H2 at 673 K for 10 h.

The amount of carbonaceous deposits on the used catalyst was calculated by a temperature programmed oxidation (TPO) method [31,32]. After testing the fresh 7.7 wt. % Co/CeO2 (hp) catalyst at 673 K for 120 h, the reactor was filled with nitrogen and cooled to 373 K. Then, an air flow of 1.5 L·h−1 was introduced in the reactor. Finally, the TPO experiment was performed by heating the


10 of 14

The amount of carbonaceous deposits on the used catalyst was calculated by a temperature programmed oxidation (TPO) method [31,32]. After testing the fresh 7.7 wt. % Co/CeO2 (hp) catalyst at 673 K for 120 h, the reactor was filled with nitrogen and cooled to 373 K. Then, an air flow of 26 10 of 14the 1.5Catalysts L¨ h´12016, was6,introduced in the reactor. Finally, the TPO experiment was performed by heating reactor from 373 to 923 K with a rate of 2.5 K¨ min´1 in the air flow. The off-gases were analysed reactor from 373 to 923 K with a rate of 2.5 K·min−1 in the air flow. The off-gases were analysed by by GC-TCD as usual. During the TPO treatment, the carbonaceous deposits formed on the catalyst GC-TCD as usual. During the TPO treatment, the carbonaceous deposits formed on the catalyst surface were oxidized to CO2 by oxygen in the air flow. The amounts of formed CO2 were recorded by surface were oxidized to CO2 by oxygen in the air flow. The amounts of formed CO2 were recorded TCD at various temperatures. by TCD at various temperatures. Figure 10 shows the CO2 formation from carbonaceous deposits over the 7.7 wt. % Co/CeO2 (hp) Figure 10 shows the CO2 formation from carbonaceous deposits over the 7.7 wt. %Ćo/CeO 2 (hp) 1 ) by increasing catalyst after reaction treated in in air air (1.5 (1.5L·h L¨ h −1) by catalyst after reactionatat673 673KKfor for120 120h.h.The Theused used catalyst catalyst was was treated increasing thethe temperature min´−11.. CO CO22formation formationwas wasobserved observed above 373 and showed two peaks temperatureatat2.5 2.5K¨K·min atat above 373 KK and showed two peaks with the maximum values at 523 and 673 K. An integration of the rate gave the amount of total , with the maximum values at 523 and 673 K. An integration of the rate gave the amount of total COCO 2, 2 from which the amount of carbonaceous deposits formed on the 7.7 wt. % Co/CeO (hp) catalyst was 2 from which the amount of carbonaceous deposits formed on the 7.7 wt. % Co/CeO2 (hp) catalyst was calculated asas follows: on the the surface surfaceof of7.7 7.7wt. wt.%%Co/CeO Co/CeO after calculated follows:2.1 2.1wt. wt.%%carbonaceous carbonaceous deposits on 2 (hp) after 2 (hp) 120120 h on stream h on streamatat673 673K.K.

Figure 10. CO2 formation from carbonaceous deposits over the 7.7 wt. % Co/CeO2 (hp) catalyst after Figure 10. CO2 formation from carbonaceous deposits over the 7.7 wt. % Co/CeO2 (hp) catalyst after reaction at 673 K for 120 h. reaction at 673 K for 120 h.

There are two peaks with the highest rates of CO2 formation at 523 K and 673 K in Figure 10, There are two peaks highest rates deposits of CO2 formation K andsurface 673 K during in Figure implying there were two with typesthe of carbonaceous formed on at the523 catalyst the 10, implying typesthat of carbonaceous deposits formed on the catalyst can surface duringtothe reaction.there It haswere beentwo reported the ethylene and oxygen-containing molecules contribute the coke formation for the that steam of ethanol [13]. Moreover, the formation of carbon reaction. It has been reported thereforming ethylene and oxygen-containing molecules can contribute to the filaments on the havereforming been confirmed in the[13]. steam reforming offormation ethanol [13]. coke formation forcatalyst the steam of ethanol Moreover, the of carbon filaments carbonaceous mayin bethe mainly formed by theofBoudouard reaction (Equation (9)) and on the The catalyst have beendeposits confirmed steam reforming ethanol [13]. theThe polymerization of by-product ethylene (Equation (10)). carbonaceous deposits may be mainly formed by the Boudouard reaction (Equation (9)) and the polymerization of by-product ethylene 2CO (Equation → CO2(10)). +C (9) nC2HÑ 4 → (CH2)2n (10) (9) 2CO CO2 ` C In general, the oxidation temperature of (CH2)2n polymers is lower than that of carbon filaments. nC2 H4 Ñ pCH 2 q2n Hence, we think the peak at 523 K corresponds to the carbonaceous deposits formed from the(10) ethylene polymerization (Equation (10)) and the )peak at 673 K corresponds to the carbonaceous In general, the oxidation temperature of (CH 2 2n polymers is lower than that of carbon filaments. deposits formed from the Boudouard reaction (Equation (9)), respectively. Hence, we think the peak at 523 K corresponds to the carbonaceous deposits formed from the ethylene 3. Experimental Section 3.1. Catalyst Preparation Co/CeO2 (hp) was obtained through a precursor of Ce1−xCoxO2−y mixed oxide by homogeneous precipitation using urea as a precipitant [26,27]. Urea was added into a mixed aqueous solution of


11 of 14

polymerization (Equation (10)) and the peak at 673 K corresponds to the carbonaceous deposits formed from the Boudouard reaction (Equation (9)), respectively. 3. Experimental Section 3.1. Catalyst Preparation Co/CeO2 (hp) was obtained through a precursor of Ce1´x Cox O2ý mixed oxide by homogeneous precipitation using urea as a precipitant [26,27]. Urea was added into a mixed aqueous solution of Ce(NO3 )3 and Co(NO3 )2 at room temperature to form a homogeneous solution. Then, the solution was heated at 363 K and hence the urea was hydrolyzed slowly. During the hydrolysis of urea, hydroxide ions were generated and a precipitate was formed slowly. After heating the solution at 363 K for 10 h, the formed Ce1´x Cox O2ý precipitate was filtered, washed with distilled water, dried in air at 373 K for 24 h and finally calcined in air at 723 K for 3 h. Co/CeO2 (cp) was prepared by a co-precipitation method using NaOH as a precipitant [30]. A mixed aqueous solution of Ce(NO3 )3 and Co(NO3 )2 was co-precipitated by adding 2 M NaOH solution till pH = 10 at room temperature. After being aged at 363 K for 1 h, the precipitate was filtered, washed with distilled water, dried at 373 K for 24 h, and finally calcined in air at 723 K for 3 h. CeO2 , which was used as a support in the impregnation Co/CeO2 (im) catalyst, was prepared by the precipitation of Ce(NO3 )3 in an aqueous solution using 2 M NaOH solution as a precipitant. After being aged at 363 K for 1 h, the precipitate was filtered, washed with distilled water, dried at 373 K for 24 h, and finally calcined in air at 723 K for 3 h. Co/CeO2 (im) was prepared by an impregnation of the CeO2 support with an aqueous solution of Co(NO3 )2 . The obtained solid sample was dried at 373 K for 24 h and calcined in air at 723 K for 3 h. Co/Al2 O3 (im) was prepared by an impregnation of the Al2 O3 support (JRC-ALO-4, 167 m2 ¨ g´1 ) with an aqueous solution of Co(NO3 )2 . The obtained solid sample was dried at 373 K for 24 h and calcined in air at 723 K for 3 h. 3.2. Catalyst Characterization Powder X-ray diffraction (XRD) patterns were measured using a MAC Science MXP-18 diffractometer (Tokyo, Japan) with Cu Kα radiation operated at 40 kV and 50 mA. Inductively coupled plasma (ICP) analyses were measured by a Thermo Jarrel Ash IRIS/AP (Yokohama, Japan). BET surface areas of the samples were measured using a BELCAT-B automatic instrument (Osaka, Japan). Transmission electron microscopy (TEM) images were obtained using a JEOL JEM 2010FX electron microscope (Tokyo, Japan) equipped with a Hitachi/Keves H-8100/Delta IV EDS (Tokyo, Japan) operating at 200 kV. The ex situ treated samples were supported on Mo grids for the observations. The Co particle sizes of the Co supported catalysts were measured by a CO adsorption method using a flow technique at room temperature [39,40,42,43]. A 0.3 g portion of sample was reduced in a flow H2 at 673 K for 10 min and then used for the measurements. The particle size of Co (D) was calculated by the relations of S = αNS1 and D = 6/ρS, where S the metal surface area of Co (m2 ¨ g´1 Co), α is the amount of CO adsorbed (mol¨ g´1 Co), N the Avogadro’s number, S1 is the cross-sectional area of a CO molecule (13 Å), and ρ is the specific gravity of Co. Hydrogen temperature-programmed reduction (H2 -TPR) was recorded using a BELCAT-B automatic instrument equipped with a TCD and a mass spectrometer. The sample (ca. 0.15 g) was flushed with Ar at 423 K for 1 h in a quartz tubular reactor to drive away water and then cooled down to 373 K. Then, 10% H2 /Ar was switched on at a flow rate of 30 mL¨ min´1 . The temperature was raised at a rate of 10 K¨ min´1 from 373 to 873 K and was held at the final temperature for 30 min. The H2 consumption was detected by TCD and was recorded automatically by a PC.


12 of 14

3.3. Catalytic Reaction The catalytic reaction was conducted under atmospheric pressure using a fixed-bed flow system equipped with an 8 mm φ quartz tubular reactor. An ethanol aqueous solution was fed from a tank by a high-pressured micro-pump (Shimadzu LC-10, Kyoto, Japan) to the reactor set at the reaction temperature. At the same time, nitrogen was also introduced into the reactor from a gas cylinder as a carry gas. The inlet gas contained 20% ethanol, 60% steam and 20% nitrogen. During reaction, the flow rates of ethanol, steam and nitrogen were 17, 50, and 17 mL¨ min´1 , respectively. The catalyst amount was one gram for the reaction, and thus the total gas hourly space velocity (GHSV) was 5000 mL¨ h´1 ¨ g´1 . The steam reforming of ethanol was mainly carried out at 673 K over various catalysts in this study. Calcination and reduction of a Co-supported catalyst at a high temperature must cause a severe sintering of Co particles and a remarkable decrease of BET surface area. We calcined the catalysts at 723 K for 3 h in this study in order to maintain relatively large BET surface areas. Moreover, we reduced the catalysts in H2 flow at a low temperature (673 K, the same as reaction temperature) for a long time (10 h) in this study in order to avoid the sintering of Co particles. The reaction was started after the catalyst (1 g, 60 meshes) was pretreated in a H2 flow (50 mL¨ min´1 ) at 673 K for 10 h. The products were analyzed using three on-line gas chromatographs. H2 , N2 , CO, and CO2 were analyzed by a TCD (Shimadzu, Kyoto, Japan)with a New Carbon-ST column (Shinwa Chem. Ind. Ltd., Kyoto, Japan), hydrocarbons were analyzed by an FID (Shimadzu, Kyoto, Japan) with a CP-Al2 O3 /KCl capillary column (Agilent Technologies Inc., Santa Clara, CA, USA), and oxygenated compounds were analyzed by an FID with a Stablewax capillary column (Restek Co., Bellefonte, PA, USA). The ethanol conversion was reported as a percent conversion: CEtOH = ([EtOH]in ´ [EtOH]out )/[EtOH]in ˆ 100. The selectivity of each carbon-containing product was reported as a percent selectivity: Sproduct = [product]/Σ[product] ˆ 100. The yield of hydrogen was reported as a molar ratio of the formed H2 to the fed C2 H5 OH: YH2 = [H2 ]out /[EtOH]in . The analytical results satisfied the carbon balance adequately (