DOI: 10.1002/cssc.201403433
Full Papers
Sustainable Hydrogen Production by Ethanol Steam Reforming using a Partially Reduced Copper–Nickel Oxide Catalyst Li-Chung Chen, Hongkui Cheng, Chih-Wei Chiang, and Shawn D. Lin*[a] Hydrogen production through the use of renewable raw materials and renewable energy is crucial for advancing its applications as an energy carrier. In this study, we fabricated a solid oxide solution of Cu and Ni within a confined pore space, followed by a partial reduction, to produce a highly efficient catalyst for ethanol steam reforming (ESR). At 300 8C, EtOH is completely converted, a H2 yield of approximately 5 mol per mol is achieved, and CO2 is the main carbon-containing product. This
demonstrates that H2 production from bioethanol is an efficient and sustainable approach. Such a highly efficient ESR catalyst is attributed to the ability of the metal–oxide interface to facilitate the transformation of CHx adspecies from acetaldehyde decomposition into methoxy-like adspecies, which are reformed readily to produce H2 and consequently reduce CH4 formation.
Introduction Hydrogen is an anticipated clean energy carrier. To achieve sustainability, renewable raw materials and renewable energy should be adopted for the production of hydrogen. Water and bio-derived compounds are suitable raw materials, and solar energy is a reasonable choice as an energy supply. Thus, solarassisted water splitting for hydrogen production is an attractive route, but its efficiency remains extremely low. By contrast, thermal energy at T 300 8C can be obtained from either solar energy or waste heat in industrial processes. The reforming, especially the steam reforming, of H-containing bio-derived compounds under such temperature conditions could be another viable route for sustainable hydrogen production. Methanol steam reforming (MSR) has recently been examined extensively for hydrogen production, and it can produce hydrogen efficiently at temperatures below 300 8C through the use of Cu-containing catalysts.[1, 2] Methanol is inexpensive, but it is synthesized chiefly from syngas, which may not be generated from renewable raw materials. The steam reforming of bioethanol, the ethanol obtained from biomaterials, can achieve a closed CO2 cycle for H2 production. However, current ethanol steam reforming (ESR) catalysts typically require high reaction temperatures of 450 8C and above.[3–5] The high reaction temperatures are typically attributed to the required breaking of the C¢C bonds; however, the drawbacks of these high temperatures include high energy consumption and an increase in side reactions such as cracking and coke formation.[6–9] The challenge in efficient ESR at 300 8C and below is the design of an appropriate catalyst. In this paper, we demonstrate that cat-
[a] Dr. L.-C. Chen, Dr. H. Cheng, C.-W. Chiang, Prof. S. D. Lin Department of Chemical Engineering National Taiwan University of Science and Technology 43, Keelung Road Sec. 4, Taipei, 106, Taiwan (ROC) E-mail: [email protected]
ChemSusChem 2015, 8, 1787 – 1793
alysts with a synergistic metal–oxide interface can overcome this challenge. ESR is widely acknowledged to proceed first through dehydrogenation to form an acetaldehyde intermediate, the C¢C bond of which is then broken to form ¢CHx (methyl or methylene) and ¢CHO (formyl) adspecies.[9–14] The ¢CHx species on the metal surface either desorb through recombination with H atoms to form methane or are further dehydrogenated to form coke.[15] Neither process is desirable because methane formation decreases the hydrogen yield and coke formation deactivates the catalyst. A suitable approach for improving the ESR efficiency is to transform the ¢CHx species from acetaldehyde decomposition to methoxy-like adspecies through interactions with O-containing surface species. Song and Ozkan reported that the O mobility of a CeO2 support decreased coke formation over Co catalysts during ESR;[16] therefore, the mobile O atoms of CeO2 may transform the ¢CHx intermediates. Dickens and Stair reported that methyl radical adsorption on NiO can lead to the formation of alkoxy species.[17] We previously proposed that the interface between a CuNi alloy and NiO can promote the acetaldehyde steam reforming (ASR) pathway in ESR.[18] Cu is an efficient methanol steam reforming catalyst, and intermediates such as methoxy and formyl species can be reformed efficiently over the Cu surface. Thus, the Cu–NiO interface may enhance ESR efficiency by suppressing the formation of methane and coke. In this study, we prepared a CuO– NiO mixed oxide by using a solid template method. Thereafter, the partial reduction of the catalyst led to the CuNi–NiO interface, which catalyzes ESR efficiently at 300 8C.
Results The as-prepared (Cu–NiO)SC catalysts (SC denotes preparation by the space-confinement method) did not produce a CuO dif1787
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers served Ni metal phase. Furthermore, the Ni (200) diffraction peaks for (10Cu–NiO)SC and 10 % Cu/NiOSC at 2 q 458 were observed at lower 2 q values than those of the completely reduced (10Cu–NiO)SC. This indicates the formation of the CuNi alloy phase under these partial-reduction conditions. The TPR with H2 (Figure 2) shows that the reduction of NiOSC occurred with a peak temperature of approximately 330 8C.
Figure 2. TPR analyses of NiOSC, 10 % Cu/NiOSC, (5Cu–NiO)SC, and (10Cu– NiO)SC catalysts.
Figure 1. XRD patterns of (a) the calcined and (b) the 250 8C-reduced NiOSC, 10 % Cu/NiOSC, (5Cu–NiO)SC, and (10Cu–NiO)SC catalysts. The completely reduced (10Cu–NiO)SC included in (b) as a reference was pretreated through H2 reduction at 500 8C.
fraction signal in XRD analysis (Figure 1 a), whereas the CuO phase was identified in the sample prepared by the impregnation method (i.e., 10 % Cu/NiOSC). The XRD results clearly revealed the diffraction peaks attributable to the NiO phase. The NiO diffraction peaks of the (Cu–NiO)SC catalysts shifted to lower 2 q values compared with those of NiOSC and Cu/NiOSC. This indicates a mixed oxide phase with Cu incorporated into the NiO framework. After the reduction of the catalyst by H2 at 250 8C, the diffraction peaks of the NiO phase remained, and all of the (Cu–NiO)SC, Cu/NiOSC, and NiOSC, samples exhibited similar 2 q diffraction-peak positions for the NiO phase (Figure 1 b). This indicates that the Cu2++ ions in the as-prepared (Cu–NiO)SC can be reduced in the NiO framework under these reduction conditions. However, no diffraction peaks of the Cu0 metal phase could be identified in any of the samples, whereas diffraction peaks similar to those of the Ni0 metal phase were observed for (10Cu–NiO)SC and 10 %Cu/NiOSC but not for (5Cu– NiO)SC and NiOSC. The absence of the Cu phase may be caused by the similarity of its diffraction-peak positions with those of the NiO phase, its highly dispersed state, or the formation of an alloy with Ni. No Ni metal phase formed from the reduction of NiOSC ; therefore, the mild reduction conditions were insufficient for the reduction of NiO. However, the presence of the Ni metal phase in (Cu–NiO)SC and 10 % Cu/NiOSC can be attributed to Cu-enhanced NiO reduction. The following temperature-programmed reduction (TPR) results indicate that the reduced Cu can promote the reduction of NiO, which results in the obChemSusChem 2015, 8, 1787 – 1793
www.chemsuschem.org
The reduction of NiO shifted to a lower temperature in the presence of CuO. The reduction of CuO to Cu occurs before the reduction of NiO.[9] The shift in the NiO reduction temperature is attributed to the initial reduction of CuO to Cu0, which then promotes the reduction of NiO. The extent of the shift in the NiO reduction temperature is possibly related to the Cu– NiO interface perimeter. The Cu-catalyzed NiO reduction may also result in the CuNi alloy phase (Figure 1 b) in close contact with NiO, because the phase diagram for Cu and Ni indicates alloy formation for all compositions. The catalysts reduced at 250 8C were only partially reduced, and the NiO phase was retained as an active support for the CuNi alloy phase. The partially reduced catalysts were analyzed again by TPR, and the extent of reduction was calculated by using the ratio of the hydrogen consumption of the partially reduced catalyst to that of the as-prepared catalysts. By assuming that the Cu2++ ions were completely reduced before the reduction of the Ni2++ ions, the compositions of the metal phases in the partially reduced catalysts can be estimated. Furthermore, the exposed metal surfaces of the partially reduced catalysts were analyzed through N2O adsorption with a subsequent TPR to quantify the adsorbed O. The results of these analyses are listed Table 1. The (Cu¢NiO)SC and 10 % Cu/NiOSC catalysts reduced at 250 8C had Ni-rich bimetallic nanoparticles dispersed on NiO. The metal dispersion of the (Cu¢NiO)SC reduced at 250 8C is significantly higher than that of 10 % Cu/NiOSC, which indicates a more intimate metal–NiO contact. The ESR over the catalysts reduced at 250 8C was compared under the temperature-programmed reaction mode, as shown in Figure 3, wherein the observed reaction products contained only H2, acetaldehyde (AA), CO, CO2, and CH4. The changes in EtOH conversion versus temperature are shown in Figure 3 a.
1788
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers achieved nearly 100 % EtOH conversion at T 300 8C under the MS/M[b] SA[c] PV[d] Composition UptakeN O Sample xRED[a] conditions used in this study; [mmol g¢1] [%] [m2 g¢1] [cm3 g¢1] [%] this conversion is higher than those over (5Cu–NiO)SC or 10 % NiOsc 0 NiO nil – 64.5 0.17 (5Cu–NiO)sc 34.6 29.6 % Cu17Ni83/NiO 74.0 52.2 – – Cu/NiOSC. The completely re54.7 48.3 % Cu21Ni79/NiO 78.7 34.5 63.5 0.23 (10Cu–NiO)sc duced (10Cu–NiO)SC catalyst had 10 % Cu/NiOsc 56.5 50.1 % Cu21Ni79/NiO 13.3 3.1 52.0 0.16 a substantially lower EtOH con[a] Extent of reduction. [b] Ratio of surface metal to total metal. [d] Brunauer–Emmett–Teller surface area. version (not shown) and lower [d] Pore volume. hydrogen yield compared with those of the partially reduced (Cu–NiO)SC. The calculated ASR selectivity, defined as the ratio of ASR to the sum of ASR and acetaldehyde decomposition (AD) during ESR,[9] is shown in Figure 3 b. The (Cu–NiO)SC catalysts exhibited a superior ASR selectivity than that of 10 % Cu/NiOSC. At 300 8C, at which the conversion of EtOH was complete, the produced methane comprised less than 10 % of the carbon selectivity over (10Cu– NiO)SC, as shown in Figure 3 c. This carbon selectivity to CH4 is significantly lower than those over 10 % Cu/NiOSC and 10 % CuNi/SiO2.[9, 18] The calculated ASR selectivity during ESR decreased as the amount of Cu in (Cu–NiO)SC decreased. This suggests that the CuNi–NiO interface promoted ASR selectivity over AD during ESR. The highly efficient ESR performance over (10Cu–NiO)SC was verified in constant-temperature reaction mode at 300 8C for 24 h of on-stream time, as shown in Figure 4, and the decay in EtOH conversion was less than 5 % over 24 h. The hydrogen yield was maintained constantly at approximately 5. The carbon selectivity to CH4 was constantly maintained at approximately 6 %, and the extent of coking, calculated from the missing carbon in the effluent, was less than 5 % of the carbon selectivity. As shown in Figure 4 c, the spent catalyst contained a CuNi alloy phase, a NiO phase, and a new phase, which can be attributed to hexagonal close packed (hcp) Ni and/or a Ni3C phase. We cannot unequivocally identify the new phase at present. However, the presence of CuNi alloy and NiO phases indicates that the catalyst maintained partly the same structure after 24 h of reaction as that at the beginning of reaction. A comparison of ESR, ASR, and MSR over (10Cu–NiO)SC under the temperature-programmed reaction mode is shown in Figure 5. Both ESR and ASR resulted in the products H2, CO, CO2, and CH4 at high conversions, and CH4 had a carbon selectivity of less than 10 %. Ethanol conversion started at 150 8C under the conditions of this study, chiefly through dehydrogenation, as indicated by a hydrogen yield of approximately 1 (Figure 5 c), and its major product was acetaldehyde. The hydrogen yield and water conversion during ESR increased simultaneously with the reaction temperature to approximately 250 8C. This corresponds to the increased acetaldehyde converFigure 3. (a) EtOH conversion and (b) calculated ASR selectivity over (Cu– NiO)SC and 10 % Cu/NiOSC catalysts;(c) the distribution of carbon-containing sion in ASR and suggests that the acetaldehyde intermediate products over the (Cu–NiO)SC catalyst during temperature-programmed ESR in ESR reacts with water to generate hydrogen. The conver¢1 at S/E = 6 and WHSV = 2 h . All catalysts were reduced in-line with H2 at sion–temperature curve of ASR over (10Cu–NiO)SC exhibited 250 8C before the ESR. a similar trend to that for ESR, except the curve shifted to a temperature higher than that for ESR. In addition, the hydrogen yield during ASR was smaller by 1 than that during ESR. The NiOSC catalyst exhibited a negligible activity, attributable Such differences can be attributed to the required dehydroto the absence of a reduced metal phase. The (10Cu–NiO)SC Table 1. The properties of the different catalysts after partial reduction at 250 8C. 2
ChemSusChem 2015, 8, 1787 – 1793
www.chemsuschem.org
1789
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers
Figure 4. (a) EtOH conversion and H2 yield and (b) the distribution of carbon-containing products over the (10Cu–NiO)SC catalyst during the 24 h ESR test; (c) XRD pattern of the spent catalyst after the 24 h ESR test. The ESR stability test was performed at 300 8C, S/E = 6, and WHSV = 2 h¢1.
Figure 5. The results of the temperature-programmed ESR, ASR, and MSR over the (10Cu–NiO)SC catalyst. All three reactions were conducted at S/C = 3 and WHSV = 2 h¢1.
genation of EtOH to AA during ESR. MSR over (10Cu–NiO)SC exhibited low conversion at temperatures below 400 8C. This suggests that the low CH4 yield during ESR or ASR cannot be attributed to MSR and, consequently, the low CH4 selectivity indicates that the methanation reaction did not proceed significantly. This supports our claim of a high selectivity for ASR compared with that for acetaldehyde decomposition to CO and CH4 during ESR over (10Cu–NiO)SC, as is consistent with the calculated ASR selectivity shown in Figure 3 b. The MSR over (10Cu–NiO)SC exhibited a low methane conversion but had a hydrogen yield of approximately 4, which is the stoichiometric yield of MSR to H2 and CO2. This indicates that the (10Cu–NiO)SC catalyst can convert adspecies from CH4 to H2 and CO2 effectively. ChemSusChem 2015, 8, 1787 – 1793
www.chemsuschem.org
We previously proposed that ¢CHx adspecies from acetaldehyde decomposition near the CuNi–NiO interface could lead to methoxy-like species and, consequently, an improved hydrogen yield of ESR.[18] The C¢O bond order is higher than that of C¢M, which suggests that methoxy-like adspecies would be more stable than CHx species in the presence of active oxygen sites. Methoxy reforming has been observed in methanol steam reforming over Cu catalysts, wherein methoxy species are the key intermediates.[1, 2] The formation of the methoxy intermediate over (10Cu–NiO)SC from adsorbed acetaldehyde was demonstrated through diffuse-reflectance infrared Fourier transform spectroscopy (DRIFTS), wherein acetaldehyde was adsorbed at room temperature over the (10Cu–NiO)SC reduced
1790
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers
Figure 6. DRIFTS spectra during the stepwise TPD of adsorbed acetaldehyde over the (10Cu–NiO)SC catalyst reduced at 250 8C. The difference spectra are indicated by 300–200 8C, 250–200 8C, and 300–250 8C. The acetaldehyde adsorption was conducted at room temperature.
at 250 8C, followed by stepwise temperature ramping under an inert gas. No significant changes were observed in the absorbance spectrum until 200 8C, whereas CO2 became evident at 250 and 300 8C (Figure 6). The difference spectra in Figure 6 clearly indicate the disappearance of acetaldehyde (by the negative peaks for nC=O at n˜ = 1727 cm¢1, nO¢C¢O at n˜ = 1441 cm¢1, dC¢H at n˜ = 1380 cm¢1, and nC¢C at n˜ = 1147 cm¢1) and methoxy formation (by the positive peak for nC¢O at n˜ = 1018 cm¢1). These assignments are consistent with those reported for methanol adsorption over NiO,[19] acetaldehyde adsorption over Ni/MgO,[20] and EtOH adsorption over mixed oxide.[21] The DRIFTS results indicate the conversion of adsorbed acetaldehyde to methoxy-like species and CO2 formation. A conversion temperature of approximately 250 8C is consistent with the onset temperature observed in ASR as well as the observed acetaldehyde conversion during ESR (Figure 6).
Discussion We demonstrated that the partially reduced (Cu–NiO)SC catalysts exhibited an unprecedented high efficiency for ESR with a hydrogen yield of approximately 5 at 300 8C, at which EtOH was converted completely. The high efficiency of the (Cu– NiO)SC ESR catalyst can be explained by the transformation of the ¢CHx adspecies from acetaldehyde decomposition to methoxy-like adspecies, the subsequent reformation of which to CO, CO2, and H2 occurs readily over the Cu surface. The completely reduced (Cu–NiO)SC catalyst had a substantially lower activity and lower hydrogen yield compared with those of the partially reduced (Cu–NiO)SC ; therefore, NiO plays an active role in this transformation of ¢CHx species to methoxy-like species. Dickens and Stair reported the formation of alkoxy species from methyl radicals adsorbed on NiO.[17] Guo and Zaera found that methyl moieties over Ni(110) produced formaldehyde in the presence of surface oxygen,[22] whereas CH4 was formed in the presence of surface hydrogen.[23] These results are consistent with our observation of methoxy-like species from adsorbed acetaldehyde (Figure 6). ChemSusChem 2015, 8, 1787 – 1793
www.chemsuschem.org
NiO can clearly provide its surface oxygen atoms for the transformation of ¢CHx species to methoxy-like adspecies. Oxygen vacancies on NiO can be generated under mild conditions such as 280 8C under H2.[24] To maintain a steady catalytic reaction, a dynamic balance in the oxygen transfer of NiO is required. Hidayat et al. showed that H2O can suppress the reduction of NiO by H2,[25] which suggests that H2O can replenish the consumed oxygen on NiO. This provides an explanation for the stable ESR performance of (10 %Cu–NiO)SC over 24 h at 300 8C (Figure 4). Furthermore, the presence of oxygen vacancies can promote water gas shift (WGS) over mixed oxides.[26] This explains the high CO2/CO ratio during ESR over (Cu–NiO)SC (Figure 4) and also suggests the activation of H2O for the reaction. This indicates that steam can replenish active oxygen over NiO during ESR to enhance not only the transformation of ¢CHx adspecies but also WGS to convert CO. Our results indicate that the metal–NiO interface provides a synergy that results in highly efficient ESR. The metal–oxide interface has long been proposed to provide a special catalytic effect.[27–30] Several recent studies have involved engineering the metal–oxide interface for an enhanced catalytic effect. Yamada et al. demonstrated that a tandem reaction scheme can occur through the combination of two metal–oxide interfaces around a metal particle.[31] Chen et al. reported that the (Fe,Ni)OH–Pt interface has a superior CO oxidation activity through the oxygen activation of a coordinately unsaturated metal site (oxygen vacancies).[32] Schalow et al. reported an improved CO oxidation activity of Pd in the presence of an Fe3O4 interface, and an oxygen storage and oxidation mechanism was proposed.[33] An et al. demonstrated that the CO oxidation activity of a Pt interface with various oxides can be orders of magnitude higher than those of oxide or Pt alone.[34] These results have all suggested that the oxygen atoms of an oxide near its interface with a metal can participate in catalytic reactions and that the oxygen vacancies (or coordinately unsaturated metal sites) of an oxide can activate oxygen. For steam reforming, the oxygen vacancies of oxides would require activating water for reactions. Reducible oxides have been used as supports to enhance ESR performance over supported metal catalysts. Ni and CuNi catalysts for ESR have shown improved performances with oxides such as CeO2 compared with those with nonreducible oxide supports.[35–39] However, none of these previous studies have achieved a H2 yield as high as that in the present study. Low-temperature ESR is considered a viable route for H2 production. MariÇo et al. proposed that ESR at a temperature below 400 8C is a practical approach, even if CH4 constitutes some of the production gas.[40] Palma and coworkers have also emphasized that reducing the reaction temperature of ESR can lead to efficient H2 production.[6, 41,42] A comparison of several recent studies on low-temperature ESR with our results is presented ion Table 2. Noble metals have frequently been used in these previous reports.[6, 41–44] Although the different steam/ethanol (S/E) ratios and space velocities (SVs) may render the comparison difficult, this study clearly provides a high H2 yield, likely as a consequence of the low CH4 selectivity. The results of this study suggest that the oxygen atoms of NiO near its interface with reduced CuNi can
1791
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers Table 2. Comparison of low-temperature ESR over different catalysts. Catalyst
T [8C]
S/E[a]
GHSV[b] [h¢1]
XEtOH[c] [%]
YH2[d] [mol]
SCO2/CH4[e] [mol]
Ref.
PtNi/CeO2 Pt/CeO2 RhCo/CeO2 RhNi/CeO2 Co hydrotalcite Ni/La2O3 CuNi/K–Al2O3 (10Cu–NiO)SC
340 300 300 350 400 300 300 300
3 3 10 4 4 3 2.5 6
750 75 ~ 300 4400 74 ~ 4500 – ~ 180
~ 100 ~ 100 ~ 100 ~ 50 66 ~ 95 95 ~ 100
1.2 ~2 4.3 1.2 1.2 ~2 1.3 ~5
~1 ~ 0.8 1.8 ~ 0.5 13 ~ 0.5 nil ~ 13
[6] [41] [43] [44] [45] [46] [40] this study
[a] Steam/EtOH molar ratio. [b] Gas hourly space velocity based on EtOH feed rate. [c] EtOH conversion. [d] Number of mol of hydrogen formed per mol of EtOH reacted. [e] Number of mol of CO2 formed per mol of CH4 formed.
participate in the transformation of ¢CHx adspecies, and steam must have replenished the oxygen atoms. This type of catalytic effect of the metal–oxide interface would be magnified through an intimate contact of the metal with the oxide. In this study, such an interface was achieved by first preparing a solid oxide solution within a confined pore space, followed by a partial reduction of the mixed oxide. An appropriate design of the metal–oxide interface is clearly a viable strategy for the development of more effective catalysts.
Conclusions Ethanol steam reforming (ESR) can be conducted effectively with an H2 yield of approximately 5 at 300 8C, at which EtOH is converted completely, by using a partially reduced CuNi–NiO catalyst. The catalysts prepared from the partial reduction of the mixed oxide CuO–NiO have higher activities and higher H2 yields than those prepared through the impregnation of Cu on NiO. The high H2 yield was accompanied by low CH4 and CO selectivity. CO2 was the major C-containing product. The low CH4 yield cannot be explained by the high methane reforming activity. The catalyst was also effective for acetaldehyde steam reforming. Infrared spectroscopy analysis indicated that adsorbed acetaldehyde can lead to the formation of methoxy-like adspecies. Thus, the proposed ESR mechanism over such catalysts involves EtOH dehydrogenation to acetaldehyde, followed by acetaldehyde decomposition to form ¢CHx and ¢CHO adspecies; the ¢CHx adspecies can be transformed into methoxy adspecies. Both ¢CHO and methoxy adspecies can be reformed readily to produce H2 and CO2. Thus, ESR can become an effective and sustainable approach for H2 production.
(xCu–NiO)SC, in which x represents a target composition of x % Cu in the format of Cu/NiO (reduced Cu over a NiO support), and NiOSC denotes NiO prepared by the same method. For comparison, a 10 % Cu/NiOSC catalyst was prepared by the incipient-wetness impregnation method to load Cu(NO3)2 on NiOSC, followed by calcination at 550 8C. Temperature-programmed reduction (TPR) analyses revealed that all of the prepared Cu–NiO catalysts can be reduced completely at 350 8C by H2. The partial reduction of all of the prepared catalysts was conducted at 250 8C for 1 h under a H2 flow.
Catalyst characterization
TPR analysis was performed at a heating rate of 5 8C min¢1 by using 10 % H2 in N2 and a thermal conductivity detector. The H2O formed during the TPR was trapped by using a molecular sieve column installed before the detector. Typically, the catalyst was purged by using He at 50 8C until the TCD stabilized, the flow was then switched to H2/N2, and the temperature was ramped. The extent of reduction of the partially reduced catalysts was evaluated through the comparison of the H2 consumption in the TPR after an in-line reduction with that of the as-prepared catalyst. For metal dispersion analysis, the catalyst after the in-line reduction was cooled to 25 8C under He and then subjected to a N2O flow at 25 8C for 30 min. The catalyst bed was then purged with He, and the H2 consumption in a subsequent TPR was used to analyze the adsorbed oxygen (Oad) quantitatively. The metal dispersion was calculated by assuming an Oad/MS ratio of 1:2 (MS represents the Cu or Ni surface atoms). The obtained metal dispersion approximates the calculated dispersion of Cu (or Ni) particles on the basis of the average size obtained from XRD and TEM analysis[9] . XRD analysis was performed by using a commercial instrument with CuKa radiation (Shimadzu XRD-6000). The diffraction patterns were recorded at room temperature at a scan rate of 0.58 min¢1. The infrared spectra were recorded with a Nicolet Magna-IR 550 spectrometer with a mercury cadmium telluride detector and a diffuse-reflectance FTIR (DRIFT) accessory (Thermo Spectra-Tech). The spectra were recorded at a resolution of 4 cm¢1 through the accumulation of 64 scans. The reference spectra were obtained after the samples were pretreated and cooled to the experimental temperature under a N2 flow. Acetaldehyde vapor was carried into the system by a N2 flow through a bubbler for 1 h, and then the cell was purged with N2 until the spectra stabilized. A stepwise temperature-programmed desorption (sTPD) process was then conducted under a N2 flow to 300 8C, and the spectra were recorded at each predetermined temperature. The high-purity (99.995 %) N2, H2, and He gases used in this study were obtained from San-Fu Gas Co. Ltd. and all were passed through drying and deoxygenation columns before use.
Experimental Section
Steam reforming reaction tests
Catalyst preparation
The ESR evaluation was conducted under the temperature-programmed reaction mode at a heating rate of 1 8C min¢1, H2O/EtOH (S/E) = 6, and a weighted hourly space velocity (WHSV; mass of EtOH/mass of catalyst) of 2 gEtOH gcat.¢1 h¢1 for a comparison of the preliminary activity. The detailed procedure has been reported previously.[9, 18] The EtOH conversion was calculated from (EtOHIN¢EtOHOUT)/EtOHIN, and the H2 yield was calculated by dividing the number mol of H2 produced by the number of mol of re-
We prepared the Cu–NiO catalysts by first forming their mixed oxides inside the pores of SBA-15. Cu(NO3)2 and Ni(NO3)2 precursors were used to fill the pores of SBA-15 by employing ethanol and hexane alternatively as the solvents, followed by calcination at 550 8C for 3 h before SBA-15 was removed with NaOH(aq). The catalysts prepared by this space-confinement method are denoted as ChemSusChem 2015, 8, 1787 – 1793
www.chemsuschem.org
1792
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers acted EtOH. The steady-state ESR performance was also examined under isothermal conditions with the same steam-to-carbon (S/C) ratio and WHSV. Acetaldehyde steam reforming (ASR) and methane steam reforming (MSR) were also examined under the same S/C and WHSV conditions.
Acknowledgements The authors gratefully acknowledge the financial support of Ministry of Education, Taiwan, under Top University Projects, and partial support from Ministry of Science and Technology, Taiwan. Keywords: alcohols · biomass · heterogeneous catalysis · hydrogen · interfaces [1] D. R. Palo, R. A. Dagle, J. D. Holladay, Chem. Rev. 2007, 107, 3992 – 4001. [2] S. S, H. Silva, L. Brand¼o, J. M. Sousa, A. Mendes, Appl. Catal. B 2010, 99, 43 – 57. [3] M. Ni, D. Y. C. Leung, M. K. H. Leung, Int. J. Hydrogen Energy 2007, 32, 3238 – 3247. [4] P. D. Vaidya, A. E. Rodrigues, Chem. Eng. J. 2006, 117, 39 – 49. [5] G. Wu, C. Zhang, S. Li, Z. Huang, S. Yan, S. Wang, X. Ma, J. Gong, Energy Environ. Sci. 2012, 5, 8942 – 8948. [6] V. Palma, F. Castaldo, P. Ciambelli, G. Iaquaniello, Appl. Catal. B 2014, 145, 73 – 84. [7] A. Iulianelli, A. Basile, Int. J. Hydrogen Energy 2010, 35, 3170 – 3177. [8] K. Vasudeva, N. Mitra, P. Umasankar, S. C. Dhingra, Int. J. Hydrogen Energy 1996, 21, 13 – 18. [9] L.-C. Chen, S. D. Lin, Appl. Catal. B 2011, 106, 639 – 649. [10] J.-H. Wang, C. S. Lee, M. C. Lin, J. Phys. Chem. C 2009, 113, 6681 – 6688. [11] A. F. Lee, D. E. Gawthrope, N. J. Hart, K. Wilson, Surf. Sci. 2004, 548, 200 – 208. [12] K. H. Lim, Z. X. Chen, K. M. Neyman, N. Rçsch, J. Phys. Chem. B 2006, 110, 14890 – 14897. [13] Z. Zhong, H. Ang, C. Choong, L. Chen, L. Huang, J. Lin, Phys. Chem. Chem. Phys. 2009, 11, 872 – 880. [14] J. R. Roy, M. A. Lalibert¦, S. Lavoie, M. Castonguay, P. H. McBreen, Surf. Sci. 2005, 578, 43 – 56. [15] L. V. Mattos, G. Jacobs, B. H. Davis, F. B. Noronha, Chem. Rev. 2012, 112, 4094 – 4123. [16] H. Song, U. S. Ozkan, J. Catal. 2009, 261, 66 – 74. [17] K. A. Dickens, P. C. Stair, Langmuir 1998, 14, 1444 – 1450. [18] L.-C. Chen, S. D. Lin, Appl. Catal. B 2014, 148 – 149, 509 – 519. [19] St. G. Christoskova, M. Stoyanova, N. Danova, O. Argirov, Appl. Catal. A 1998, 173, 101 – 105. [20] C. Resinia, S. Cavallaro, F. Frusteri, S. Freni, G. Busca, React. Kinet. Catal. Lett. 2007, 90, 117 – 126.
ChemSusChem 2015, 8, 1787 – 1793
www.chemsuschem.org
[21] J. V. Ochoa, C. Trevisanut, J.-M. M. Millet, G. Busca, F. Cavani, J. Phys. Chem. C 2013, 117, 23908 – 23918. [22] H. Guo, F. Zaera, J. Phys. Chem. B 2004, 108, 16226 – 16232. [23] H. Guo, F. Zaera, J. Phys. Chem. B 2004, 108, 16220 – 16225. [24] J. A. Rodriguez, J. C. Hanson, A. I. Frenkel, J. Y. Kim, M. Perez, J. Am. Chem. Soc. 2002, 124, 346 – 354. [25] T. Hidayat, M. A. Rhamdhani, E. Jak, P. C. Hayes, Metall. Mater. Trans. B 2009, 40, 1 – 16. [26] J. A. Rodriguez, J. C. Hanson, W. Wen, X. Wang, J. L. Brito, A. MartnezArias, M. Fernndez-Garca, Catal. Today 2009, 145, 188 – 194. [27] S. J. Tauster, S. C. Fung, R. T. K. Baker, J. A. Horsley, Science 1981, 211, 1121 – 1125. [28] S. J. Tauster, Acc. Chem. Res. 1987, 20, 389 – 394. [29] M. A. Vannice, Top. Catal. 1997, 4, 241 – 248. [30] M. A. Vannice, J. Mol. Catal. 1990, 59, 165 – 177. [31] Y. Yamada, C.-K. Tsung, W. Huang, Z. Huo, S. E. Habas, T. Soejima, C. E. Aliaga, G. A. Somorjai, P. Yang, Nat. Chem. 2011, 3, 372 – 376. [32] G. Chen, Y. Zhao, G. Fu, P. N. Duchesne, L. Gu, Y. Zheng, X. Weng, M. Chen, P. Zhang, C.-W. Pao, J.-F. Lee, N. Zheng, Science 2014, 344, 495 – 499. [33] T. Schalow, M. Laurin, B. Brandt, S. Schauermann, S. Guimond, H. Kuhlenbeck, D. E. Starr, S. K. Shaikhutdinov, J. Libuda, H.-J. Freund, Angew. Chem. Int. Ed. 2005, 44, 7601 – 7605; Angew. Chem. 2005, 117, 7773 – 7777. [34] K. An, S. Alayoglu, N. Musselwhite, S. Plamthottam, G. Melaet, A. E. Lindeman, G. A. Somorjai, J. Am. Chem. Soc. 2013, 135, 16689 – 16696. [35] W. Xu, Z. Liu, A. C. Johnston-Peck, S. D. Senanayake, G. Zhou, D. Stacchiola, E. A. Stach, J. A. Rodriguez, ACS Catal. 2013, 3, 975 – 984. [36] P. Biswas, D. Kunzru, Int. J. Hydrogen Energy 2007, 32, 969 – 980. [37] Q. Liu, Z. Liu, X. Zhou, C. Li, J. Ding, J. Rare Earths 2011, 29, 872 – 877. [38] A. C. Furtado, C. G. Alonso, M. P. Cant¼o, N. R. C. Fernandes-Machado, Int. J. Hydrogen Energy 2009, 34, 7189 – 7196. [39] G. Bonura, C. Cannilla, F. Frusteri, Appl. Catal. B 2012, 121 – 122, 135 – 147. [40] F. MariÇo, G. Baronetti, M. Jobbagy, M. Laborde, Appl. Catal. A 2003, 238, 41 – 54. [41] P. Ciambelli, V. Palma, A. Ruggiero, Appl. Catal. B 2010, 96, 18 – 27. [42] V. Palma, F. Castaldo, P. Ciambelli, G. Iaquaniello, G. Capitani, Int. J. Hydrogen Energy 2013, 38, 6633 – 6645. [43] L. Huang, C. Choong, L. Chen, Z. Wang, Z. Zhong, C. Campos-Cuerva, J. Lin, ChemCatChem 2013, 5, 220 – 234. [44] J. Kugai, S. Velu, C. Song, Catal. Lett. 2005, 101, 255 – 264. [45] R. Espinal, E. Taboada, E. Molins, R. J. Chimentao, F. Medinac, J. Llorca, RSC Adv. 2012, 2, 2946 – 2956. [46] J. Sun, X.-P. Qiu, F. Wu, W.-T. Zhu, Int. J. Hydrogen Energy 2005, 30, 437 – 445.
Received: December 18, 2014 Published online on April 15, 2015
1793
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers
Sustainable Hydrogen Production by Ethanol Steam Reforming using a Partially Reduced Copper–Nickel Oxide Catalyst Li-Chung Chen, Hongkui Cheng, Chih-Wei Chiang, and Shawn D. Lin*[a] Hydrogen production through the use of renewable raw materials and renewable energy is crucial for advancing its applications as an energy carrier. In this study, we fabricated a solid oxide solution of Cu and Ni within a confined pore space, followed by a partial reduction, to produce a highly efficient catalyst for ethanol steam reforming (ESR). At 300 8C, EtOH is completely converted, a H2 yield of approximately 5 mol per mol is achieved, and CO2 is the main carbon-containing product. This
demonstrates that H2 production from bioethanol is an efficient and sustainable approach. Such a highly efficient ESR catalyst is attributed to the ability of the metal–oxide interface to facilitate the transformation of CHx adspecies from acetaldehyde decomposition into methoxy-like adspecies, which are reformed readily to produce H2 and consequently reduce CH4 formation.
Introduction Hydrogen is an anticipated clean energy carrier. To achieve sustainability, renewable raw materials and renewable energy should be adopted for the production of hydrogen. Water and bio-derived compounds are suitable raw materials, and solar energy is a reasonable choice as an energy supply. Thus, solarassisted water splitting for hydrogen production is an attractive route, but its efficiency remains extremely low. By contrast, thermal energy at T 300 8C can be obtained from either solar energy or waste heat in industrial processes. The reforming, especially the steam reforming, of H-containing bio-derived compounds under such temperature conditions could be another viable route for sustainable hydrogen production. Methanol steam reforming (MSR) has recently been examined extensively for hydrogen production, and it can produce hydrogen efficiently at temperatures below 300 8C through the use of Cu-containing catalysts.[1, 2] Methanol is inexpensive, but it is synthesized chiefly from syngas, which may not be generated from renewable raw materials. The steam reforming of bioethanol, the ethanol obtained from biomaterials, can achieve a closed CO2 cycle for H2 production. However, current ethanol steam reforming (ESR) catalysts typically require high reaction temperatures of 450 8C and above.[3–5] The high reaction temperatures are typically attributed to the required breaking of the C¢C bonds; however, the drawbacks of these high temperatures include high energy consumption and an increase in side reactions such as cracking and coke formation.[6–9] The challenge in efficient ESR at 300 8C and below is the design of an appropriate catalyst. In this paper, we demonstrate that cat-
[a] Dr. L.-C. Chen, Dr. H. Cheng, C.-W. Chiang, Prof. S. D. Lin Department of Chemical Engineering National Taiwan University of Science and Technology 43, Keelung Road Sec. 4, Taipei, 106, Taiwan (ROC) E-mail: [email protected]
ChemSusChem 2015, 8, 1787 – 1793
alysts with a synergistic metal–oxide interface can overcome this challenge. ESR is widely acknowledged to proceed first through dehydrogenation to form an acetaldehyde intermediate, the C¢C bond of which is then broken to form ¢CHx (methyl or methylene) and ¢CHO (formyl) adspecies.[9–14] The ¢CHx species on the metal surface either desorb through recombination with H atoms to form methane or are further dehydrogenated to form coke.[15] Neither process is desirable because methane formation decreases the hydrogen yield and coke formation deactivates the catalyst. A suitable approach for improving the ESR efficiency is to transform the ¢CHx species from acetaldehyde decomposition to methoxy-like adspecies through interactions with O-containing surface species. Song and Ozkan reported that the O mobility of a CeO2 support decreased coke formation over Co catalysts during ESR;[16] therefore, the mobile O atoms of CeO2 may transform the ¢CHx intermediates. Dickens and Stair reported that methyl radical adsorption on NiO can lead to the formation of alkoxy species.[17] We previously proposed that the interface between a CuNi alloy and NiO can promote the acetaldehyde steam reforming (ASR) pathway in ESR.[18] Cu is an efficient methanol steam reforming catalyst, and intermediates such as methoxy and formyl species can be reformed efficiently over the Cu surface. Thus, the Cu–NiO interface may enhance ESR efficiency by suppressing the formation of methane and coke. In this study, we prepared a CuO– NiO mixed oxide by using a solid template method. Thereafter, the partial reduction of the catalyst led to the CuNi–NiO interface, which catalyzes ESR efficiently at 300 8C.
Results The as-prepared (Cu–NiO)SC catalysts (SC denotes preparation by the space-confinement method) did not produce a CuO dif1787
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers served Ni metal phase. Furthermore, the Ni (200) diffraction peaks for (10Cu–NiO)SC and 10 % Cu/NiOSC at 2 q 458 were observed at lower 2 q values than those of the completely reduced (10Cu–NiO)SC. This indicates the formation of the CuNi alloy phase under these partial-reduction conditions. The TPR with H2 (Figure 2) shows that the reduction of NiOSC occurred with a peak temperature of approximately 330 8C.
Figure 2. TPR analyses of NiOSC, 10 % Cu/NiOSC, (5Cu–NiO)SC, and (10Cu– NiO)SC catalysts.
Figure 1. XRD patterns of (a) the calcined and (b) the 250 8C-reduced NiOSC, 10 % Cu/NiOSC, (5Cu–NiO)SC, and (10Cu–NiO)SC catalysts. The completely reduced (10Cu–NiO)SC included in (b) as a reference was pretreated through H2 reduction at 500 8C.
fraction signal in XRD analysis (Figure 1 a), whereas the CuO phase was identified in the sample prepared by the impregnation method (i.e., 10 % Cu/NiOSC). The XRD results clearly revealed the diffraction peaks attributable to the NiO phase. The NiO diffraction peaks of the (Cu–NiO)SC catalysts shifted to lower 2 q values compared with those of NiOSC and Cu/NiOSC. This indicates a mixed oxide phase with Cu incorporated into the NiO framework. After the reduction of the catalyst by H2 at 250 8C, the diffraction peaks of the NiO phase remained, and all of the (Cu–NiO)SC, Cu/NiOSC, and NiOSC, samples exhibited similar 2 q diffraction-peak positions for the NiO phase (Figure 1 b). This indicates that the Cu2++ ions in the as-prepared (Cu–NiO)SC can be reduced in the NiO framework under these reduction conditions. However, no diffraction peaks of the Cu0 metal phase could be identified in any of the samples, whereas diffraction peaks similar to those of the Ni0 metal phase were observed for (10Cu–NiO)SC and 10 %Cu/NiOSC but not for (5Cu– NiO)SC and NiOSC. The absence of the Cu phase may be caused by the similarity of its diffraction-peak positions with those of the NiO phase, its highly dispersed state, or the formation of an alloy with Ni. No Ni metal phase formed from the reduction of NiOSC ; therefore, the mild reduction conditions were insufficient for the reduction of NiO. However, the presence of the Ni metal phase in (Cu–NiO)SC and 10 % Cu/NiOSC can be attributed to Cu-enhanced NiO reduction. The following temperature-programmed reduction (TPR) results indicate that the reduced Cu can promote the reduction of NiO, which results in the obChemSusChem 2015, 8, 1787 – 1793
www.chemsuschem.org
The reduction of NiO shifted to a lower temperature in the presence of CuO. The reduction of CuO to Cu occurs before the reduction of NiO.[9] The shift in the NiO reduction temperature is attributed to the initial reduction of CuO to Cu0, which then promotes the reduction of NiO. The extent of the shift in the NiO reduction temperature is possibly related to the Cu– NiO interface perimeter. The Cu-catalyzed NiO reduction may also result in the CuNi alloy phase (Figure 1 b) in close contact with NiO, because the phase diagram for Cu and Ni indicates alloy formation for all compositions. The catalysts reduced at 250 8C were only partially reduced, and the NiO phase was retained as an active support for the CuNi alloy phase. The partially reduced catalysts were analyzed again by TPR, and the extent of reduction was calculated by using the ratio of the hydrogen consumption of the partially reduced catalyst to that of the as-prepared catalysts. By assuming that the Cu2++ ions were completely reduced before the reduction of the Ni2++ ions, the compositions of the metal phases in the partially reduced catalysts can be estimated. Furthermore, the exposed metal surfaces of the partially reduced catalysts were analyzed through N2O adsorption with a subsequent TPR to quantify the adsorbed O. The results of these analyses are listed Table 1. The (Cu¢NiO)SC and 10 % Cu/NiOSC catalysts reduced at 250 8C had Ni-rich bimetallic nanoparticles dispersed on NiO. The metal dispersion of the (Cu¢NiO)SC reduced at 250 8C is significantly higher than that of 10 % Cu/NiOSC, which indicates a more intimate metal–NiO contact. The ESR over the catalysts reduced at 250 8C was compared under the temperature-programmed reaction mode, as shown in Figure 3, wherein the observed reaction products contained only H2, acetaldehyde (AA), CO, CO2, and CH4. The changes in EtOH conversion versus temperature are shown in Figure 3 a.
1788
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers achieved nearly 100 % EtOH conversion at T 300 8C under the MS/M[b] SA[c] PV[d] Composition UptakeN O Sample xRED[a] conditions used in this study; [mmol g¢1] [%] [m2 g¢1] [cm3 g¢1] [%] this conversion is higher than those over (5Cu–NiO)SC or 10 % NiOsc 0 NiO nil – 64.5 0.17 (5Cu–NiO)sc 34.6 29.6 % Cu17Ni83/NiO 74.0 52.2 – – Cu/NiOSC. The completely re54.7 48.3 % Cu21Ni79/NiO 78.7 34.5 63.5 0.23 (10Cu–NiO)sc duced (10Cu–NiO)SC catalyst had 10 % Cu/NiOsc 56.5 50.1 % Cu21Ni79/NiO 13.3 3.1 52.0 0.16 a substantially lower EtOH con[a] Extent of reduction. [b] Ratio of surface metal to total metal. [d] Brunauer–Emmett–Teller surface area. version (not shown) and lower [d] Pore volume. hydrogen yield compared with those of the partially reduced (Cu–NiO)SC. The calculated ASR selectivity, defined as the ratio of ASR to the sum of ASR and acetaldehyde decomposition (AD) during ESR,[9] is shown in Figure 3 b. The (Cu–NiO)SC catalysts exhibited a superior ASR selectivity than that of 10 % Cu/NiOSC. At 300 8C, at which the conversion of EtOH was complete, the produced methane comprised less than 10 % of the carbon selectivity over (10Cu– NiO)SC, as shown in Figure 3 c. This carbon selectivity to CH4 is significantly lower than those over 10 % Cu/NiOSC and 10 % CuNi/SiO2.[9, 18] The calculated ASR selectivity during ESR decreased as the amount of Cu in (Cu–NiO)SC decreased. This suggests that the CuNi–NiO interface promoted ASR selectivity over AD during ESR. The highly efficient ESR performance over (10Cu–NiO)SC was verified in constant-temperature reaction mode at 300 8C for 24 h of on-stream time, as shown in Figure 4, and the decay in EtOH conversion was less than 5 % over 24 h. The hydrogen yield was maintained constantly at approximately 5. The carbon selectivity to CH4 was constantly maintained at approximately 6 %, and the extent of coking, calculated from the missing carbon in the effluent, was less than 5 % of the carbon selectivity. As shown in Figure 4 c, the spent catalyst contained a CuNi alloy phase, a NiO phase, and a new phase, which can be attributed to hexagonal close packed (hcp) Ni and/or a Ni3C phase. We cannot unequivocally identify the new phase at present. However, the presence of CuNi alloy and NiO phases indicates that the catalyst maintained partly the same structure after 24 h of reaction as that at the beginning of reaction. A comparison of ESR, ASR, and MSR over (10Cu–NiO)SC under the temperature-programmed reaction mode is shown in Figure 5. Both ESR and ASR resulted in the products H2, CO, CO2, and CH4 at high conversions, and CH4 had a carbon selectivity of less than 10 %. Ethanol conversion started at 150 8C under the conditions of this study, chiefly through dehydrogenation, as indicated by a hydrogen yield of approximately 1 (Figure 5 c), and its major product was acetaldehyde. The hydrogen yield and water conversion during ESR increased simultaneously with the reaction temperature to approximately 250 8C. This corresponds to the increased acetaldehyde converFigure 3. (a) EtOH conversion and (b) calculated ASR selectivity over (Cu– NiO)SC and 10 % Cu/NiOSC catalysts;(c) the distribution of carbon-containing sion in ASR and suggests that the acetaldehyde intermediate products over the (Cu–NiO)SC catalyst during temperature-programmed ESR in ESR reacts with water to generate hydrogen. The conver¢1 at S/E = 6 and WHSV = 2 h . All catalysts were reduced in-line with H2 at sion–temperature curve of ASR over (10Cu–NiO)SC exhibited 250 8C before the ESR. a similar trend to that for ESR, except the curve shifted to a temperature higher than that for ESR. In addition, the hydrogen yield during ASR was smaller by 1 than that during ESR. The NiOSC catalyst exhibited a negligible activity, attributable Such differences can be attributed to the required dehydroto the absence of a reduced metal phase. The (10Cu–NiO)SC Table 1. The properties of the different catalysts after partial reduction at 250 8C. 2
ChemSusChem 2015, 8, 1787 – 1793
www.chemsuschem.org
1789
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers
Figure 4. (a) EtOH conversion and H2 yield and (b) the distribution of carbon-containing products over the (10Cu–NiO)SC catalyst during the 24 h ESR test; (c) XRD pattern of the spent catalyst after the 24 h ESR test. The ESR stability test was performed at 300 8C, S/E = 6, and WHSV = 2 h¢1.
Figure 5. The results of the temperature-programmed ESR, ASR, and MSR over the (10Cu–NiO)SC catalyst. All three reactions were conducted at S/C = 3 and WHSV = 2 h¢1.
genation of EtOH to AA during ESR. MSR over (10Cu–NiO)SC exhibited low conversion at temperatures below 400 8C. This suggests that the low CH4 yield during ESR or ASR cannot be attributed to MSR and, consequently, the low CH4 selectivity indicates that the methanation reaction did not proceed significantly. This supports our claim of a high selectivity for ASR compared with that for acetaldehyde decomposition to CO and CH4 during ESR over (10Cu–NiO)SC, as is consistent with the calculated ASR selectivity shown in Figure 3 b. The MSR over (10Cu–NiO)SC exhibited a low methane conversion but had a hydrogen yield of approximately 4, which is the stoichiometric yield of MSR to H2 and CO2. This indicates that the (10Cu–NiO)SC catalyst can convert adspecies from CH4 to H2 and CO2 effectively. ChemSusChem 2015, 8, 1787 – 1793
www.chemsuschem.org
We previously proposed that ¢CHx adspecies from acetaldehyde decomposition near the CuNi–NiO interface could lead to methoxy-like species and, consequently, an improved hydrogen yield of ESR.[18] The C¢O bond order is higher than that of C¢M, which suggests that methoxy-like adspecies would be more stable than CHx species in the presence of active oxygen sites. Methoxy reforming has been observed in methanol steam reforming over Cu catalysts, wherein methoxy species are the key intermediates.[1, 2] The formation of the methoxy intermediate over (10Cu–NiO)SC from adsorbed acetaldehyde was demonstrated through diffuse-reflectance infrared Fourier transform spectroscopy (DRIFTS), wherein acetaldehyde was adsorbed at room temperature over the (10Cu–NiO)SC reduced
1790
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers
Figure 6. DRIFTS spectra during the stepwise TPD of adsorbed acetaldehyde over the (10Cu–NiO)SC catalyst reduced at 250 8C. The difference spectra are indicated by 300–200 8C, 250–200 8C, and 300–250 8C. The acetaldehyde adsorption was conducted at room temperature.
at 250 8C, followed by stepwise temperature ramping under an inert gas. No significant changes were observed in the absorbance spectrum until 200 8C, whereas CO2 became evident at 250 and 300 8C (Figure 6). The difference spectra in Figure 6 clearly indicate the disappearance of acetaldehyde (by the negative peaks for nC=O at n˜ = 1727 cm¢1, nO¢C¢O at n˜ = 1441 cm¢1, dC¢H at n˜ = 1380 cm¢1, and nC¢C at n˜ = 1147 cm¢1) and methoxy formation (by the positive peak for nC¢O at n˜ = 1018 cm¢1). These assignments are consistent with those reported for methanol adsorption over NiO,[19] acetaldehyde adsorption over Ni/MgO,[20] and EtOH adsorption over mixed oxide.[21] The DRIFTS results indicate the conversion of adsorbed acetaldehyde to methoxy-like species and CO2 formation. A conversion temperature of approximately 250 8C is consistent with the onset temperature observed in ASR as well as the observed acetaldehyde conversion during ESR (Figure 6).
Discussion We demonstrated that the partially reduced (Cu–NiO)SC catalysts exhibited an unprecedented high efficiency for ESR with a hydrogen yield of approximately 5 at 300 8C, at which EtOH was converted completely. The high efficiency of the (Cu– NiO)SC ESR catalyst can be explained by the transformation of the ¢CHx adspecies from acetaldehyde decomposition to methoxy-like adspecies, the subsequent reformation of which to CO, CO2, and H2 occurs readily over the Cu surface. The completely reduced (Cu–NiO)SC catalyst had a substantially lower activity and lower hydrogen yield compared with those of the partially reduced (Cu–NiO)SC ; therefore, NiO plays an active role in this transformation of ¢CHx species to methoxy-like species. Dickens and Stair reported the formation of alkoxy species from methyl radicals adsorbed on NiO.[17] Guo and Zaera found that methyl moieties over Ni(110) produced formaldehyde in the presence of surface oxygen,[22] whereas CH4 was formed in the presence of surface hydrogen.[23] These results are consistent with our observation of methoxy-like species from adsorbed acetaldehyde (Figure 6). ChemSusChem 2015, 8, 1787 – 1793
www.chemsuschem.org
NiO can clearly provide its surface oxygen atoms for the transformation of ¢CHx species to methoxy-like adspecies. Oxygen vacancies on NiO can be generated under mild conditions such as 280 8C under H2.[24] To maintain a steady catalytic reaction, a dynamic balance in the oxygen transfer of NiO is required. Hidayat et al. showed that H2O can suppress the reduction of NiO by H2,[25] which suggests that H2O can replenish the consumed oxygen on NiO. This provides an explanation for the stable ESR performance of (10 %Cu–NiO)SC over 24 h at 300 8C (Figure 4). Furthermore, the presence of oxygen vacancies can promote water gas shift (WGS) over mixed oxides.[26] This explains the high CO2/CO ratio during ESR over (Cu–NiO)SC (Figure 4) and also suggests the activation of H2O for the reaction. This indicates that steam can replenish active oxygen over NiO during ESR to enhance not only the transformation of ¢CHx adspecies but also WGS to convert CO. Our results indicate that the metal–NiO interface provides a synergy that results in highly efficient ESR. The metal–oxide interface has long been proposed to provide a special catalytic effect.[27–30] Several recent studies have involved engineering the metal–oxide interface for an enhanced catalytic effect. Yamada et al. demonstrated that a tandem reaction scheme can occur through the combination of two metal–oxide interfaces around a metal particle.[31] Chen et al. reported that the (Fe,Ni)OH–Pt interface has a superior CO oxidation activity through the oxygen activation of a coordinately unsaturated metal site (oxygen vacancies).[32] Schalow et al. reported an improved CO oxidation activity of Pd in the presence of an Fe3O4 interface, and an oxygen storage and oxidation mechanism was proposed.[33] An et al. demonstrated that the CO oxidation activity of a Pt interface with various oxides can be orders of magnitude higher than those of oxide or Pt alone.[34] These results have all suggested that the oxygen atoms of an oxide near its interface with a metal can participate in catalytic reactions and that the oxygen vacancies (or coordinately unsaturated metal sites) of an oxide can activate oxygen. For steam reforming, the oxygen vacancies of oxides would require activating water for reactions. Reducible oxides have been used as supports to enhance ESR performance over supported metal catalysts. Ni and CuNi catalysts for ESR have shown improved performances with oxides such as CeO2 compared with those with nonreducible oxide supports.[35–39] However, none of these previous studies have achieved a H2 yield as high as that in the present study. Low-temperature ESR is considered a viable route for H2 production. MariÇo et al. proposed that ESR at a temperature below 400 8C is a practical approach, even if CH4 constitutes some of the production gas.[40] Palma and coworkers have also emphasized that reducing the reaction temperature of ESR can lead to efficient H2 production.[6, 41,42] A comparison of several recent studies on low-temperature ESR with our results is presented ion Table 2. Noble metals have frequently been used in these previous reports.[6, 41–44] Although the different steam/ethanol (S/E) ratios and space velocities (SVs) may render the comparison difficult, this study clearly provides a high H2 yield, likely as a consequence of the low CH4 selectivity. The results of this study suggest that the oxygen atoms of NiO near its interface with reduced CuNi can
1791
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers Table 2. Comparison of low-temperature ESR over different catalysts. Catalyst
T [8C]
S/E[a]
GHSV[b] [h¢1]
XEtOH[c] [%]
YH2[d] [mol]
SCO2/CH4[e] [mol]
Ref.
PtNi/CeO2 Pt/CeO2 RhCo/CeO2 RhNi/CeO2 Co hydrotalcite Ni/La2O3 CuNi/K–Al2O3 (10Cu–NiO)SC
340 300 300 350 400 300 300 300
3 3 10 4 4 3 2.5 6
750 75 ~ 300 4400 74 ~ 4500 – ~ 180
~ 100 ~ 100 ~ 100 ~ 50 66 ~ 95 95 ~ 100
1.2 ~2 4.3 1.2 1.2 ~2 1.3 ~5
~1 ~ 0.8 1.8 ~ 0.5 13 ~ 0.5 nil ~ 13
[6] [41] [43] [44] [45] [46] [40] this study
[a] Steam/EtOH molar ratio. [b] Gas hourly space velocity based on EtOH feed rate. [c] EtOH conversion. [d] Number of mol of hydrogen formed per mol of EtOH reacted. [e] Number of mol of CO2 formed per mol of CH4 formed.
participate in the transformation of ¢CHx adspecies, and steam must have replenished the oxygen atoms. This type of catalytic effect of the metal–oxide interface would be magnified through an intimate contact of the metal with the oxide. In this study, such an interface was achieved by first preparing a solid oxide solution within a confined pore space, followed by a partial reduction of the mixed oxide. An appropriate design of the metal–oxide interface is clearly a viable strategy for the development of more effective catalysts.
Conclusions Ethanol steam reforming (ESR) can be conducted effectively with an H2 yield of approximately 5 at 300 8C, at which EtOH is converted completely, by using a partially reduced CuNi–NiO catalyst. The catalysts prepared from the partial reduction of the mixed oxide CuO–NiO have higher activities and higher H2 yields than those prepared through the impregnation of Cu on NiO. The high H2 yield was accompanied by low CH4 and CO selectivity. CO2 was the major C-containing product. The low CH4 yield cannot be explained by the high methane reforming activity. The catalyst was also effective for acetaldehyde steam reforming. Infrared spectroscopy analysis indicated that adsorbed acetaldehyde can lead to the formation of methoxy-like adspecies. Thus, the proposed ESR mechanism over such catalysts involves EtOH dehydrogenation to acetaldehyde, followed by acetaldehyde decomposition to form ¢CHx and ¢CHO adspecies; the ¢CHx adspecies can be transformed into methoxy adspecies. Both ¢CHO and methoxy adspecies can be reformed readily to produce H2 and CO2. Thus, ESR can become an effective and sustainable approach for H2 production.
(xCu–NiO)SC, in which x represents a target composition of x % Cu in the format of Cu/NiO (reduced Cu over a NiO support), and NiOSC denotes NiO prepared by the same method. For comparison, a 10 % Cu/NiOSC catalyst was prepared by the incipient-wetness impregnation method to load Cu(NO3)2 on NiOSC, followed by calcination at 550 8C. Temperature-programmed reduction (TPR) analyses revealed that all of the prepared Cu–NiO catalysts can be reduced completely at 350 8C by H2. The partial reduction of all of the prepared catalysts was conducted at 250 8C for 1 h under a H2 flow.
Catalyst characterization
TPR analysis was performed at a heating rate of 5 8C min¢1 by using 10 % H2 in N2 and a thermal conductivity detector. The H2O formed during the TPR was trapped by using a molecular sieve column installed before the detector. Typically, the catalyst was purged by using He at 50 8C until the TCD stabilized, the flow was then switched to H2/N2, and the temperature was ramped. The extent of reduction of the partially reduced catalysts was evaluated through the comparison of the H2 consumption in the TPR after an in-line reduction with that of the as-prepared catalyst. For metal dispersion analysis, the catalyst after the in-line reduction was cooled to 25 8C under He and then subjected to a N2O flow at 25 8C for 30 min. The catalyst bed was then purged with He, and the H2 consumption in a subsequent TPR was used to analyze the adsorbed oxygen (Oad) quantitatively. The metal dispersion was calculated by assuming an Oad/MS ratio of 1:2 (MS represents the Cu or Ni surface atoms). The obtained metal dispersion approximates the calculated dispersion of Cu (or Ni) particles on the basis of the average size obtained from XRD and TEM analysis[9] . XRD analysis was performed by using a commercial instrument with CuKa radiation (Shimadzu XRD-6000). The diffraction patterns were recorded at room temperature at a scan rate of 0.58 min¢1. The infrared spectra were recorded with a Nicolet Magna-IR 550 spectrometer with a mercury cadmium telluride detector and a diffuse-reflectance FTIR (DRIFT) accessory (Thermo Spectra-Tech). The spectra were recorded at a resolution of 4 cm¢1 through the accumulation of 64 scans. The reference spectra were obtained after the samples were pretreated and cooled to the experimental temperature under a N2 flow. Acetaldehyde vapor was carried into the system by a N2 flow through a bubbler for 1 h, and then the cell was purged with N2 until the spectra stabilized. A stepwise temperature-programmed desorption (sTPD) process was then conducted under a N2 flow to 300 8C, and the spectra were recorded at each predetermined temperature. The high-purity (99.995 %) N2, H2, and He gases used in this study were obtained from San-Fu Gas Co. Ltd. and all were passed through drying and deoxygenation columns before use.
Experimental Section
Steam reforming reaction tests
Catalyst preparation
The ESR evaluation was conducted under the temperature-programmed reaction mode at a heating rate of 1 8C min¢1, H2O/EtOH (S/E) = 6, and a weighted hourly space velocity (WHSV; mass of EtOH/mass of catalyst) of 2 gEtOH gcat.¢1 h¢1 for a comparison of the preliminary activity. The detailed procedure has been reported previously.[9, 18] The EtOH conversion was calculated from (EtOHIN¢EtOHOUT)/EtOHIN, and the H2 yield was calculated by dividing the number mol of H2 produced by the number of mol of re-
We prepared the Cu–NiO catalysts by first forming their mixed oxides inside the pores of SBA-15. Cu(NO3)2 and Ni(NO3)2 precursors were used to fill the pores of SBA-15 by employing ethanol and hexane alternatively as the solvents, followed by calcination at 550 8C for 3 h before SBA-15 was removed with NaOH(aq). The catalysts prepared by this space-confinement method are denoted as ChemSusChem 2015, 8, 1787 – 1793
www.chemsuschem.org
1792
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers acted EtOH. The steady-state ESR performance was also examined under isothermal conditions with the same steam-to-carbon (S/C) ratio and WHSV. Acetaldehyde steam reforming (ASR) and methane steam reforming (MSR) were also examined under the same S/C and WHSV conditions.
Acknowledgements The authors gratefully acknowledge the financial support of Ministry of Education, Taiwan, under Top University Projects, and partial support from Ministry of Science and Technology, Taiwan. Keywords: alcohols · biomass · heterogeneous catalysis · hydrogen · interfaces [1] D. R. Palo, R. A. Dagle, J. D. Holladay, Chem. Rev. 2007, 107, 3992 – 4001. [2] S. S, H. Silva, L. Brand¼o, J. M. Sousa, A. Mendes, Appl. Catal. B 2010, 99, 43 – 57. [3] M. Ni, D. Y. C. Leung, M. K. H. Leung, Int. J. Hydrogen Energy 2007, 32, 3238 – 3247. [4] P. D. Vaidya, A. E. Rodrigues, Chem. Eng. J. 2006, 117, 39 – 49. [5] G. Wu, C. Zhang, S. Li, Z. Huang, S. Yan, S. Wang, X. Ma, J. Gong, Energy Environ. Sci. 2012, 5, 8942 – 8948. [6] V. Palma, F. Castaldo, P. Ciambelli, G. Iaquaniello, Appl. Catal. B 2014, 145, 73 – 84. [7] A. Iulianelli, A. Basile, Int. J. Hydrogen Energy 2010, 35, 3170 – 3177. [8] K. Vasudeva, N. Mitra, P. Umasankar, S. C. Dhingra, Int. J. Hydrogen Energy 1996, 21, 13 – 18. [9] L.-C. Chen, S. D. Lin, Appl. Catal. B 2011, 106, 639 – 649. [10] J.-H. Wang, C. S. Lee, M. C. Lin, J. Phys. Chem. C 2009, 113, 6681 – 6688. [11] A. F. Lee, D. E. Gawthrope, N. J. Hart, K. Wilson, Surf. Sci. 2004, 548, 200 – 208. [12] K. H. Lim, Z. X. Chen, K. M. Neyman, N. Rçsch, J. Phys. Chem. B 2006, 110, 14890 – 14897. [13] Z. Zhong, H. Ang, C. Choong, L. Chen, L. Huang, J. Lin, Phys. Chem. Chem. Phys. 2009, 11, 872 – 880. [14] J. R. Roy, M. A. Lalibert¦, S. Lavoie, M. Castonguay, P. H. McBreen, Surf. Sci. 2005, 578, 43 – 56. [15] L. V. Mattos, G. Jacobs, B. H. Davis, F. B. Noronha, Chem. Rev. 2012, 112, 4094 – 4123. [16] H. Song, U. S. Ozkan, J. Catal. 2009, 261, 66 – 74. [17] K. A. Dickens, P. C. Stair, Langmuir 1998, 14, 1444 – 1450. [18] L.-C. Chen, S. D. Lin, Appl. Catal. B 2014, 148 – 149, 509 – 519. [19] St. G. Christoskova, M. Stoyanova, N. Danova, O. Argirov, Appl. Catal. A 1998, 173, 101 – 105. [20] C. Resinia, S. Cavallaro, F. Frusteri, S. Freni, G. Busca, React. Kinet. Catal. Lett. 2007, 90, 117 – 126.
ChemSusChem 2015, 8, 1787 – 1793
www.chemsuschem.org
[21] J. V. Ochoa, C. Trevisanut, J.-M. M. Millet, G. Busca, F. Cavani, J. Phys. Chem. C 2013, 117, 23908 – 23918. [22] H. Guo, F. Zaera, J. Phys. Chem. B 2004, 108, 16226 – 16232. [23] H. Guo, F. Zaera, J. Phys. Chem. B 2004, 108, 16220 – 16225. [24] J. A. Rodriguez, J. C. Hanson, A. I. Frenkel, J. Y. Kim, M. Perez, J. Am. Chem. Soc. 2002, 124, 346 – 354. [25] T. Hidayat, M. A. Rhamdhani, E. Jak, P. C. Hayes, Metall. Mater. Trans. B 2009, 40, 1 – 16. [26] J. A. Rodriguez, J. C. Hanson, W. Wen, X. Wang, J. L. Brito, A. MartnezArias, M. Fernndez-Garca, Catal. Today 2009, 145, 188 – 194. [27] S. J. Tauster, S. C. Fung, R. T. K. Baker, J. A. Horsley, Science 1981, 211, 1121 – 1125. [28] S. J. Tauster, Acc. Chem. Res. 1987, 20, 389 – 394. [29] M. A. Vannice, Top. Catal. 1997, 4, 241 – 248. [30] M. A. Vannice, J. Mol. Catal. 1990, 59, 165 – 177. [31] Y. Yamada, C.-K. Tsung, W. Huang, Z. Huo, S. E. Habas, T. Soejima, C. E. Aliaga, G. A. Somorjai, P. Yang, Nat. Chem. 2011, 3, 372 – 376. [32] G. Chen, Y. Zhao, G. Fu, P. N. Duchesne, L. Gu, Y. Zheng, X. Weng, M. Chen, P. Zhang, C.-W. Pao, J.-F. Lee, N. Zheng, Science 2014, 344, 495 – 499. [33] T. Schalow, M. Laurin, B. Brandt, S. Schauermann, S. Guimond, H. Kuhlenbeck, D. E. Starr, S. K. Shaikhutdinov, J. Libuda, H.-J. Freund, Angew. Chem. Int. Ed. 2005, 44, 7601 – 7605; Angew. Chem. 2005, 117, 7773 – 7777. [34] K. An, S. Alayoglu, N. Musselwhite, S. Plamthottam, G. Melaet, A. E. Lindeman, G. A. Somorjai, J. Am. Chem. Soc. 2013, 135, 16689 – 16696. [35] W. Xu, Z. Liu, A. C. Johnston-Peck, S. D. Senanayake, G. Zhou, D. Stacchiola, E. A. Stach, J. A. Rodriguez, ACS Catal. 2013, 3, 975 – 984. [36] P. Biswas, D. Kunzru, Int. J. Hydrogen Energy 2007, 32, 969 – 980. [37] Q. Liu, Z. Liu, X. Zhou, C. Li, J. Ding, J. Rare Earths 2011, 29, 872 – 877. [38] A. C. Furtado, C. G. Alonso, M. P. Cant¼o, N. R. C. Fernandes-Machado, Int. J. Hydrogen Energy 2009, 34, 7189 – 7196. [39] G. Bonura, C. Cannilla, F. Frusteri, Appl. Catal. B 2012, 121 – 122, 135 – 147. [40] F. MariÇo, G. Baronetti, M. Jobbagy, M. Laborde, Appl. Catal. A 2003, 238, 41 – 54. [41] P. Ciambelli, V. Palma, A. Ruggiero, Appl. Catal. B 2010, 96, 18 – 27. [42] V. Palma, F. Castaldo, P. Ciambelli, G. Iaquaniello, G. Capitani, Int. J. Hydrogen Energy 2013, 38, 6633 – 6645. [43] L. Huang, C. Choong, L. Chen, Z. Wang, Z. Zhong, C. Campos-Cuerva, J. Lin, ChemCatChem 2013, 5, 220 – 234. [44] J. Kugai, S. Velu, C. Song, Catal. Lett. 2005, 101, 255 – 264. [45] R. Espinal, E. Taboada, E. Molins, R. J. Chimentao, F. Medinac, J. Llorca, RSC Adv. 2012, 2, 2946 – 2956. [46] J. Sun, X.-P. Qiu, F. Wu, W.-T. Zhu, Int. J. Hydrogen Energy 2005, 30, 437 – 445.
Received: December 18, 2014 Published online on April 15, 2015
1793
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
