Ethanol steam reforming over Ni-Cu/Al2O3-MyOz (M=Si,La,Mg,and Zn) catalysts

Availableonlineatwww.sciencedirect.comScience DirectNATURAL GASCHEMISTRYELSEVIERJoumal of Natural Gas Chemistry 18(2009)55:www.elsevier.com/locate/jngEthanol steam reforming over Ni-Cu/A12O3-MyO2M=Si, La, Mg,and zn) catalystsLifeng Zhang, Jie Liu, Wei Li, Cuili Guo, Jinli ZhangSchool of Chemical Engineering and Technology Tianjin University Tianjin 300072, ChinaManuscript received September 17, 2008; revised December 11, 2008AbstractNi-based catalysts doped with copper additives were studied on their role in ethanol steam reforming reaction. The effects of Cu content,upport species involving Al2O3-SiO2, Al2O3-MgO, Al2O3-ZnO, and Al2O3-La2O3, on the catalytic performance were studied. Characteriza-ons by TPR, XRD, NH3-TPD, XPS, and TGA indicated that catalysts 30Ni5Cw/Al2O3-MgO and 30NiSCwAl2O3-Zno have much higher Hselectivity than 30Ni5Cu/Al2O3-SiOz, as well as good coke resistance. H] selectivity for 30NiSCwAlO3- MgO catalyst was 73. 3%at 450Cand increased to 94.0% at 600C, whereas for 30Ni5Cu/Al2O3-Zno catalyst, the H2 selectivity was 63.6% at 450C and 95.2% at 600C.These AlO3-Mg0 and Al2O3-ZnO supported Ni-Cu bimetallic catalysts may have important applications in the production of hydrogen byethanol steam reforming reactionsKey words: ethanol steam reforming; nickel-copper- based catalyst; Al2O3-SiOz support; Al2O3-MgO support; Al2O3-ZnO support; Al2 O3-a203 support1. Introductioninto CO2 through the water-gas shift(wGS)reaction [4]Along with the increasing demand for carbon-neutralTo improve the production of hydrogen, transition met-and zero-emission alternative fuels, considerable attention haals involving Ni [51, Co [6,7], PL, Pd, and Rh [1, 8-10] havebeen paid to transform the biomass-derived compounds into been reported to show good activity and selectivity for ethanolhydrogen-rich gas. Among such compounds, bioethanol ob. steam reforming reactions. Especially, nickel is widely usedtained by biomass fermentation is of particular interest be. as the catalytic component for hydrogenation and dehydro-cause of the increasing availability of raw material, ease of genation reactions because of its high activity to break C-chandling in the liquid state, high hydrogen content, and lotbonds and the relatively low cost. However, there exists se-toxicity [1.2]. Hydrogen can be produced by following the vere catalyst deactivation by sintering or carbon depositionoverall steam reforming reaction of ethanol [3][11-13]over Ni-based catalyst. The secondary metal speciesis added into Ni-based catalysts so as to promote the cokingCH3CH20H+3H20-6H2+2C02 AH98=+347. 4 kJ/mol resistance. For instances, Vizcaino reported that the additionof Cu into Sio -supported Ni-based catalysts could reduceThe thermodynamic favorable ethanol steam reforming CO generation and improve coke resistance in the steam re-reaction includes several catalytic steps as follows: (i)the de- forming reaction of ethanol [14]. Marino et al. [15-18re-hydrogenation or dehydration reaction of ethanol; (ii)the de- ported that Cu/Ni/K/y-Al2O3 catalyst showed acceptable ac-hydrogenation reaction produces acetaldehyde as the interme- tivity, stability, and selectivity for hydrogen at 300C in thediate product that can be transferred to CO and CHa through steam reforming reaction of bioethanol. It is considered thatcarbon-carbon(C-C)bond breaking, and the dehydration re- Cu has limited steam reforming activity but good activity inaction produces ethylene that can be easily transformed into dehydrogenation [19, 20] and WGS reactioncarbon deposition; and (ii)methane can generate hydrogenAmong the oxide supports, alumina-based supports areand carbon monoxide by steam reforming reaction while the often used in reforming catalysts because of their mechanicalgenerated carbon monoxide can be subsequently transformed and中国煤化工 conditions. WhereasCorresponding author. Tel: +86-22-27890643; Fax: +86-22-27890643: E-mail: zhanundation items: 973 Program(2006CB202500), the National Natural ScienceCNMHntury Excellent Talents inpyright 02009, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, All nights reserveddoi:l0.0l6S1003-99Lifeng Zhang et al/ Joumal of Natural Gas Chemistry Vol 18 No. I 2009over the support of Al2O3, catalyst deactivationoccurs prepared by precipitation-impregnation method using a com-by carbon deposition through the ethylene intergener- mercial ?-Al2O3 powder(SBET=177 m/g, Tianjin R&D Inated via the dehydration reactions of ethanolhe basic stitute of Chemical Industry, China). In all experiments, theadditives or promoters, which are beneficial to water adsorp- weight ratio of M, O2 /Al2O was fixed at 0.25 in the mixedtion and -OH surface mobility, have been extensively stud- supports. A stoichiometric amount of 0.1 M La(NO3)3so-led to improve the coke resistance of Al2O3-supported cata- lution, 0. 1 M Mg(NO3) solution, or 0.1 M Zn(NO3)solulysts. For example, alkali and alkali earth metal oxides, such tion was added to 0. 1 M H2N-CO-NH2 solution by continuas Mgo [21, 22] and solid solution(MgAl204)[23], were re- ously stirring at 100C, respectively, together with introducported to prevent the coke formation on the Ni-based cata- ing y-Al O3 powder at the same time. The obtained solutionlysts. It was reported that MgAl2O4 solid solution could re- was filtered after stirring for 6 h. The filtered solid was thenduce the Ni particle size so as to suppress the coke formation dried at 120C overnight and calcined in air at 650C for 6h[24]. Zno [25] and lanthanide oxides [5] were reported to The mixed supports were ground and the powder sieved topromote the dehydrogenation reaction to acetaldehyde. In ad- 0. 2-0. 3 mm size was used as the catalyst support. they weredition, the support of silica-stabilized alumina was reported designated as Al2O3-M, Ox, where M referred to the elementto improve the activity performance of palladium-supported (Mg, Zn, and La)added into y-Al2O3 carriercatalyst for the combustion of methane [26], and the supportNi-Cu supported catalysts were prepared by precipitationof SiO2-Al2O3 is capable of alleviating the coke formation for impregnation method using the above-mentioned supports inhydrogenation of aromatics [27]. Although there are a number volving y-Al203, Al2O3-SiO2, Al203-MgO, Al2O3-ZnO,andof literatures that have studied various catalysts with different Al2O3-La2O3. In all cases, the nominal Ni loadings on theactive metals for the ethanol steam reforming reaction, it still above five supports were fixed at 30 wt%. A stoichiometneeds intensive investigation on both active metal component ric amount of 0. I M Ni(NO 3)2 solution, 0. 1M Cu(NO3hand the support to study efficient catalysts to produce hydro- solution, and the respective supports were introduced intogen with high selectivity and coke resistancethe 0.1 M H2N-CO-NH2 solution by continuously stirring atIn this article, to obtain the optimal Ni-based catalysts 100C. The suspension solution was filtered after 6 h, andwith Cu additives, the effects of Cu additives on the catalytic the catalyst was obtained by desiccation at 120C overnightperformance of the Al2O3-SiOz supported Ni catalysts in the followed by calcination in air at 650C for 6h. Theethanol steam reforming reaction were studied. TPR, TEM, prepared catalysts were designated as 30NinCu/Al2O3- SiOzXRD, and NH3-TPD were used to characterize the effect of (=0. 5, 10, and 15),30Ni5Cu/Al2O3, or 30Ni5Cu/Al2O3Cu content on the product selectivities. Then, the compos- M,O,(M=Mg, Zn, or La), respectively. For example,ited support of Al2O3-MgO, Al2O3-ZnO, and Al2O3-La2 O3 30Ni5Cu/Al2O3-MgO catalyst indicated Al2O3-MgO sup-were used to prepare the Ni-Cu bimetallic catalysts, aiming ported catalyst with the amount of Ni 30 wt% and Cu 5 wt%to improve the H2 selectivity, and these catalysts were characterized using XPS, XRD, TPR, NH3-TPD, and TGA. It is 2. 2. Catalyst characterizationshown that the Al2O3-MgO and Al2O3-Zn0 supported NiCu catalysts had much better catalytic performance than theCrystalline phases of the catalysts were identified byAl2O3-SiO2 supported catalyst.XRD (Philips, PANalytical)measurements using Cu Kara-diation(=0.15418 nm)operated at 40 kv and 40 mA. XRDpattens were collected from 20=10 to 80. The iden-tification of crystalline phases was made by matching with2. 1. Support and catalyst preparationthe X'Pert HighScore softwareTEM was performed with a JEOL JEM-100CXII trans-af the support of Al2O3-Sio2 with the Sio2/Al2O3 weight mission electron microscope, operated at an accelerating volt-io of 0. 25 was prepared by a sol-gel method using age of 100 kV.tetraethoxysilane [Si(OC2H5)4] and Al(NO3 )3 9H20 as theXPS was recorded with a Perkin-Elmer(PHd1600 elecprecursors. AI(NO3)3. 9H20 was dissolved in distilled water, tron spectrometer equipped with a Mg Ka X-ray sourceand the pH of the solution was monitored as 3. 0-3.5 by addi-(hv= 1263. 6 eV)and a hemispherical electron analyser opertion of dilute nitric acid. In a typical experiment, a stoichio- ating at a constant transmission energy (50ev). The Cls peakmetric amount of solution of 0.347 g/ml Al(NO3)3 was mixed at 284. 87 ev was used as an internal standard for peak posiwith Si(OC2 Hs)4 by stirring for 5 h followed by flocculation tion measurement. The areas of the peaks were estimated byat pH=3.0 adjusted using dilute nitric acid. The obtained calculating the integral of each peak after subtracting a Shirleygel was then dried at 60C in an air oven for 24 h, and the background and fitting the experimental peak to a combinadried product was ground and then calcined in air at 650c tion of中国煤化工 iable proportionsfor 6 h. The prepared Al2O3-SiO2 had a specific surface areaof about 200 m2/g with the nominal Sio2/Al203 weight ratio toCheCNMHGe performed with Auequipment was fittedof 0. 25. Then, it was ground and sieved. The powder with with a thermal conductivity detector(TCD), and TPR-H20. 2-0. mm size was used as the catalyst support.analysis was done using 5 mol% H2 in N2 at a heating rateAl2O3-M,O mixed supports(M=La, Mg, or Zn) were of 10 C/min from room temperature to 900C.Joumal of Natural Gas Chemistry Vol 18 No. I 2009NH3-TPD measurements of the samples were prereduced containing carbon(Si,carbon-containing product)must be equal orwith H2(30 mV min)at 650C for 2 h. Ammonia was ad- lower than 1.0sorbed at high temperature(at 120'C for 1 h)to overcomeNH physisorption. NH3-TPD experiments were performed 3 Results and discussionin a stream of He flow(50 ml/min) at a heating rate of 3. 1. Effect of cu content on the catalytic performance10C/min up to 650C. The concentration of NH3 in the exitgas was continuously monitored using TCDTable I compares the conversion of ethanol and selecTemperature-programmed oxidation analyses of the used tivities of the products over 30NixCw/Al2O3-SiO2(x=0.5catalysts were caried out using a thermo-gravimetric ana- 10, 15)catalysts for the ethanol steam reforming reaction atlyzer(Pyris Diamond; Perkin Elmer Corporation, USA)to temperatures ranging from 400-600 C under the conditiondetermine the amount of coke deposited on catalysts. The of H20: EtoH=4.0(molar ratio) and LHSV=8.0 mI/hgstandard protocol involved the weight change of the sample Both ethanol conversion and product selectivity are related(15 mg)during its heating in 200 ml/min of N2 as purge gas to the Cu content of catalysts. When Cu content was notand 50 mU ' min of O2 as reactive gas from 30C to 1000'C at higher than 10 wt%, the ethanol conversion was 100%overa heating rate of 5C/minthe 30Niz Cu/Al2O3-SiO2(a =0, 5, 10) catalyst in this tem-perature range. Whereas over higher-copper catalyst of2.3. Activity tests30Nil5Cu/Al2O3-SiO2, the ethanol conversion was 88.6%atThe ethanol steam reforming reaction was carried out400C and increased to 100% at the temperature higher than450C. Over 30Ni/Al2O3-SiO2 existed intermediate ethyl-in a fixed-bed stainless steel tubular reactor(id =12 mm, ene in the range of 400-600C because of the high acid-L=360 mm)operated at atmospheric pressure. In each run ofthe experiment, 150 mg of the catalyst diluted with SiC (both ity of the Aloy support [14)l, 1. e, the selectivity for ethin the 0. 2-0. mm particle size range and selected after preliminary mass transport experiments to minimize diffusional 600C, respectively. Upon addition of small amount of Cu,resistances)at a volume ratio of 3 1 was added to avoid adneither intermediate ethylene nor aldehyde was detected inthe whole temperature range of 400-600C over catalystverse thermal effects, and before measurement, the catalysts 30Ni5CuAl2O3-SiO2. However, much higher Cu content re-were flushed with nitrogen at 200'C, followed by in situ reduction at 650C for 40 min(heating rate 10C/min)with sulted in the emergence of the intermediate aldehyde.The30 ml (STP)min of H]. The pretreating gases were flushed selectivity for aldehyde was 33. 2% over 30Ni15Cu/Al2O3-from the reactor with n2 before the ethanol/water reactioSio2 catalyst and 10.6% over 30Ni10Cu/Al2O3-SiO2 catalystmixture was added. Water and ethanol were premixed in at 400"C. The selectivity for aldehyde decreased as the tem-a molar ratio of 4.0 and fed to the vaporizer using syringe perature increased for these two catalysts, no aldehyde was de-tected above 450C for 30Ni15Cu/Al2O3-Sio2 catalyst, andpumps at a rate of 1.2 ml/h before mixing with N2(as di- for 30Nil0Cu/Al203-SiO2 catalyst, no aldehyde was detectedThe vaporization temperature was set at 150%C. The reaction above 400"C.products were analyzed on line by gas chromatography withIt can be seen that over 30Ni/Al2O3-SiO2 catalyst, the seTCD(chromatograph Model 3420) equipped with Porapack lectivity for H2, CO, CH4, and CO2 is 58.7%0, 33.4%, 39.0%,and 18.8% at 400C, respectively. The selectivity for H2Q(CO2, C2H6, C2H4, water, acetaldehyde, ethanol, acetone, and COz increases with temperature, whereas the selectivityand TDX-01(H2, N2, CO, CH4, CO2, C2H4r-packed columns for CHa decreases with temperature and the selectivity forconnected in series, using He as carrier gas. Catalyst activity CO reaches a minimum of 9.49 at 450"; when the temwas measured after 2 h at each temperature.perature is up to 600C, the selectivity for H2, CO, CHEthanol conversion denoted as Xethanol and H2 selectivity and COz is 90.0%, 24. %, 7. 4%, and 58.8%, respectivelydenoted as SH, were evaluated according to eqs.( 1 )-(4):When Cu content is 5 wt%, 30Ni5Cu/Al2O3-SiOz catalysthas the highest selectivity for H2 and COz in the temperaturEthanol, in-Fethanol outrange of 400-600C, i. e, the selectivity for H2, Co, CH4,and cO2is61.2%,29.6%,42.7%,and26.3%at400°candchanges to 92.0%, 32.3%, 7. 2%, and 59.9% at 600C, re-FHFFXH, O- FH2O,inspectively. When Cu content is higher than 5 wt%, the se-lectivity for H2 and CO2 decreased in the temperature rangeof 400-600C, which implies that the ethanol steam reform-S3m用+(()“门计D)叫WGStentCNMHGEf30Nir Cu/AlO3-SiO(a=0. 5, 10, 15)catalysts. The peaks around 200-300oCwhere F is the molar flow rate and n is the ratio of C atom were because of the reduction of Cuo[28], and the peak inten-in ethanol to in product. The total selectivity for compounds sity increased with the Cu content. The peaks at 350-500oCLifeng Zhang et al/ Journal of Natural Gas Chemistry VoL. 18 No I 2009were attributed to the reduction of Nio [29 while the peaks at the range of 350-550 C to the range of 350-450oC. It istemperature above 500C reflected the interactions between indicated that the addition of Cu decreased the reduction tem-the support and NiO. The reduction temperature of Nio in perature and as the Cu content increased, the interaction be-30Ni/Al2O3-SiOz catalyst was higher than 350C, upon ad- tween the support and the NiO or Ni1-zCuo complex becamedition of Cu the reduction temperature of Nio decreased from weakTable 1. Effect of copper content on the ethanol conversion and product selectivities over the 30Nir Cu/AlO3- SiO (r=0, 5, 10, 15)catalystsCu content(%)Reaction temperature(C) X(EtOHV% S(C: H4)% S(CH3 CHO)% S(H2)% S(CO)% S(CH4)% S(CO2)0000027.1990000000000020.25548590058.861247726.30000∞000092032355.537.664639.23380000600045.117721.5049.100057363617.6633.2550072837400Reaction temperature =400-600C, H20: EtOH=4: 1(mol ratio), LHSV=8mwH.g30NilSCwAl-O-Sio,oa30Ni15Cu/Al,O,SiO,30Nil0Cu/AlOr-SiO,30Nil0Cw/Al,OrSiO,30NiSCw/AlyOrSiOoNiSCu/Al,-SiO,(440)ALOSioTemperature(℃)Figure 1. tPR profiles of the 30NizCw/Al2O3-SiOz(=0, 5, 10, 15)cata-30Ni15CuA1-OrSiC▲◆★Figure 2 shows the XRD pattems of 30NizCu/Al2oSiO2(r=0, 5, 10, 15)catalysts calcined and after the ethanolsteam reforming reaction, Figure 2a indicates that no peaks30NilOCWAlorSiOof Cuo appear in the XRD pattern of 30Ni5Cu/Al2O3-Siocalcined, which suggests high dispersion of Cu oxide in the30NiSCw/Al O Sio,catalyst. As the Cu content increases, the Cuo peaks appear中国煤化工in the XRd of 30Ni10Cu/Al2O3-SiO2 and 30Ni15Cu/Al2OSiO2 catalysts. Although Cu has limited ability to break c-CCNMHGbonds[30], it is a good catalyst for dehydrogenation[19, 201and WGS reactions [31. It is the excess Cu component of Figure 2. XRD pattens of the 30Niz Cu/Al -Sioz(z=0. 5, 10, 15)30Ni10Cu/Al2O3-Sio2 and 30Nil5Cw/Al203-SiOz catalyststalysts with different Cu content(a)calcined,(b)after reactionJournal of Natural Gas Chemistry VoL. 18 No. 1 2009that resulted in the formation of intermediate aldehyde and the assessment, each catalyst was characterized by XRD.Aslow selectivities for H2 as shown in Table 1. According shown in Figure 2b, AlO3 and Al SiOs were detected by theto the Scherrer equation, the crystal size of Nio was calcu- presence of some characteristic broad reflections in catalystslated as 26.1 nm, 20.1 nm, 23.8 nm, and 23.7 nm for catalysts For 30Ni/AlO3-SiO2 catalyst, the most prominent reflections30NiTCwAl2O3-SiO2(=0, 5, 10, 15), respectively(see Ta- are assigned to metallic Ni, which obviously resulted from hyble 2).drogen reduction of the Nio in the calcined catalysts. In CuHaving been reduced at 650C for 40 min with H2, the doped catalysts, the most prominent reflections are assignedcatalytic performance of each 30NizCw/A12O3-SiO(r=0, 5, to alloy Ni(-zCur, which is formed by hydrogen reduction10, 15)catalyst was measured for 2 h in the ethanol steam re- of the NiO and Cuo in the calcined catalysts [32]. Therefore,forming reaction at each temperature from 400 to 600C, with Cu-doping enhances the reducibility of the catalysts, as shownan interval of 50C and the heating rate of 10C/min. After in the above TPR study( Figure 1)Table 2. Crystal grain sizes of catalystsKinds of catSize of Nio(nm)Size of Ni or Nil--Cu(nm)BOND/Al2O3-SiO after reaction1830NiSCW/Al2O3-SiOz calcined30NiSCw/Al2O3-SiOz after reaction22330Ni10Cu/Al2O3-SiO2 calcined23830Nil0Cw/Al2O3-SiO2 after reaction24.530Nil5Cu/Al O3.Sio calcined30Nil5Cw/AlzO3-SiOz after reaction92Figure 3 shows the TEM images of calcined 30NiSCW/Al2O3-SiOz catalyst is superior to those over other30NizCu/AlO3-SiO2(a=0, 5, 10, 15)catalysts. It is Al2O3-SiO2 supported catalysts, which is well consistent withndicated that the dispersibility of metallic species over the XRd results of the calcined catalysts( Figure 2a)30N30Ni5C30Ni 15Cu30Ni 1 OCu100nmFigure 3. TEM images of the 30NizCu/AlOj-Sio(z=0. 5, 10, 15)catalysts3.2. Effect of composited support on catalytic performance low 500C, the catalysts with composited supports, such asAlO3 SiOz, Al2O3 MgO, and Al2O3 ZnO, have higher H2On the basis of the above-mentioned optimal catalytic selectivity than that supported on Al2O3, except the catalystomposition of 30Ni5 Cu/Al2O3-SiOz catalyst, different com- supported on Al,Ox La O The 30Ni5Cu/Al2O3-MgO andposited support species were studied to improve the cat- 30Ni5中国煤化工 higher H2 selectivityalytic performance for ethanol steam reforming reactions. thanthe reaction temperaAs a control, the performance of 30Ni5 Cw/Al2O catalyst ture iCN H Gity over these catalystis also shown in Figure 4. Figure 4a shows the selectiv- decreased. Figure 4b shows the selectivities of various prod-ity for the hydrogen over these synthesized catalysts in the ucts over these catalysts at 450C, only for 30Ni5Cu/AlO3temperature range of 400-600C. It is indicated that be- catalyst, there appeared small amount of C2H4Lifeng Zhang et al. Journal of Natural Gas Chemistry Vol 18 No. I 200930Ni5Cu/ALO,MgoC 30Ni5Cu/Al-Or-ZnO30NiSCw/ALOrLaot 30Ni5Cw/ ALO -SiO30Ni5Cu/AlOr-26/(°)zaco(b)30Ni5Cu/AL,O,La O,Figure 4.(a) H2 selectivity over different catalysts in the temperature rangeof 400-600C for the steam reforming of ethanol under the condition ofLSHV=8.0 ml/h-g and H20: EtoH=4: 1(mol ratio):(b)Product selec-tivities of 30Ni5Cu/Al2O(NiCu/A), 30Ni5CuAl2 O3-Mgo (NiCW/A-Mg),2e/(°)30Ni5Cu/Al2O3-Zn0(NiCu/A-Zn), 30Ni5Cu/Al2O3-LagO(NiCu/A-LaFigure 5. XRD pattems of the catalyst sample. (a) calcined catalysts, (b)and 30Ni5Cw/AlO3-SiOz(NiCu/A-Si) under the condition of 450C, reduced catalystsSHv=8.0 mlh-g and H20: EtOH =4: I(mol ratio)Table 3. Crystal grain sizes of different catalystsFigure 5 shows the XRD patterms of the calcined andSize(nm)reduced 30Ni5Cu-based catalysts with different supportsIt is clear from Figure 5a that major peaks for calcined)Ni5Cw/Al20l1612.630Ni5Cu/Al203, 30Ni5Cu/Al2O3-Mgo, 30Ni5Cu/Al2O330NiSCw/Alz03-SiOz20.1180ZnO, and 30Ni5Cu/Al2O3-SiO2 catalysts are attributed to theNiSCu/Al2 O3ZnoNiO peak. The crystal size of Nio is calculated using the30NiSCu/Al203-MScherrer equation based on its peak with Miller indexes of(200)and listed in Table 3. It is indicated that the supportsFurther, the TPR patterns reflected the influence of sup-of Al2O3-ZnO and Al2O3-MgO result in much smaller crys- port species on the reduction of catalysts. As shown in Figtal size of the Nio, compared with that of Al2O3 and Al2O3- ure 6, all these catalysts have the peaks below 350"C becauseSio2. For the 30Ni5Cw/Al2O3-La2O3 catalyst, the crystal size of the reduction of Cuo and Cu oxides with different degreewas not obtained because of the weakness of diffraction peaks interaction with supports[28]. In the case of 30Ni5Cw/Al203of NiO. It suggested that NiO in 30Ni5Cw/Al2O3-La2O3 cat- SiO2 catalyst, its TPR profile involved Ni species that has twoalyst was well dispersedpeaks located at 360C and415C, respectively, which is beHaving been reduced at 650C, in the XRD patterns of cause of the reduction of Nioese catalysts appeared major peaks of Ni that are closely Although the 30Ni5 Cu/Al2O3-Zn0 and 30Ni5Cu/AlOrelated with the support, as shown in Figure 5b. It is impor- Mgo catalysts have the highest H2 selectivity(Figure 4),tant to note that the major peaks of 30Ni5Cu/Al2O3-Znotheir reduction peaks of Nio were quite different. The30Ni5Cu/Al2O3-MgO catalysts shift away from the peaks of 30Ni5C/AlaOn-7no catalvst has a neak around 400C formetallic Ni. This suggests that there exist strong interactions the re中国煤化工70 C for the reducamong the support Al2O3-ZnO, Al2O3-MgO and nickel ox- tionides, Al2 ZnO4, MgA1204, NiAI204[33, 34), and Ni2+ uni- ZnO slCNMHGtions with the A203.Mgo catalyst showsformly dispersed in NiAl2O4 is beneficial to prevent the a broad peak at 680C, which are related to the reductiongrowth of metallic Ni particles during the reduction, as listed of NiO species with different degree interaction with Al203MgO support [35, 36Journal of Natural Gas Chemistry Vol. 18 No. 1 200930NiSCuAlfOrSiO,30Ni5Cw/AlyOr-Lao,30NSCW/AlyOr-MgO30Ni5Cu/AlyOrSiO30NiSCu/ Zno30NiSCu/Al OrOnoiSCu/Al O,MgO30Ni5Cu/AlO,30NiCu/Al O,00200300400500600700800900Temperature(℃30NI5Cu/Al2O3-SiO2nd 30NiSCw/Al2O3-ZnOFigure 1. Temcined catalysts 30NiSCuAl2O3, 30NiSCw/AlO3-MgO, 30Ni5Cu/Al2O:Zno, 30NiS Cu/Al2O3-SiO2, and 30Ni5Cw/Al2O3-La20The obvious variation of the peaks around 400C indi-cates the distinct interactions between NiO and different sup-Table 5. Binding energies of core electrons of calcined supportsport, and Al203-zno and Al2O3-MgO have stronger interac-CalcinedBinding energy(ev)tions with NiO than Al2O3-SiosupportAl2p La3ds/2 Mg 2p Zn 3p3/2 Si 2p3/2The NH3-TPD profiles of these catalysts indicate thatAl2O3-Mgo 74. 1498species strongly affect the acidity strengthAhO3-Zn0 74.6Ain Figure 7 and Table 4, all the tested sam-Al2O31La2O374.6834.7similar acid site distribution but differences inAlzO3-Sio? 74.6concentration. On the basis of the data shown in Ta-ble 4, the concentration of acid sites in reduced catalysts value for La2O3(834.3 eV)[40] suggests no interaction bedecreases in the order 30NiSCw/Al203>30Ni5Cu/Al203- tween La203 and Al203 [41]. It is in accordance with theLa203 R 30Ni5Cu/Al2O3-SiO2>30NiSCw/Al2O3ZnO N XRD results(Figure 5b), which suggests strong interactions30Ni5Cu/Al2O3-MgO. The lowest acidity strength of either between the supports Al2O3-Zno, Al203-MgO and nickel ox-30Ni5Cw/Al2O3-Zno or 30Ni5Cw/Al2O3-MgO catalysts re- ides, Al2ZnO4, MgAl2O4, and NiAl2O4sults in their good catalytic performance to produce H2The chemical state of five different catalyst surface af-through ethanol steam reforming reactionter reduction in H2 at 650C was also studied using XPSTable 4. Acidity of the calcined nickel catalysts fromand the results are summarized in Table 6. All reduced cat-temperature-programmed desorption of NH3alysts showed the peak of Ni(852.5 ev) together with asmall fraction of Ni2+(856.0ev). This unreduced Niis attributed to Ni+ of either Ni2SiO4 (856.1 ev), La2NiO(855.8 ev)[42], or NiAlO4(855.8 ev)(43], respectively,30Ni5Cw/Al2O3-MgO3.43which has been mentioned in the previous TPR patterns30NiSCw/Al2O3-ZnOThe XPS percentages of metallic Ni on surface decreased30NiSCw/AlzO3-siOz3.96in the order: 30Ni5Cu/Al2O3-SiO>30Ni5Cu/AlO30NiSCw/Al2O3-La2O3Zno> 30Ni5Cw/Al2O3-Mgo> 30NiSCw/Al2O3 >30N/AlO3-La2O3The binding energies of core electrons and the surfaceIt is known that Cu2p3n2 line appears at 932 4eV,atomic ratios of the calcined supports are summarized in Ta- 932.6eV, and 934.6ev that are typical of Cut, cu,andble 5. The Al2O3-Mgo support presents a single compo- Cu2+of all reduced catalysts are due tonent photoelectron Mg 2p peak at 49.8 eV, close to that re- Cuo中国煤化工 Zno catalyst shows aported in literature for MgAl]O4(50.2 eV)[37, and Al2O3- surfaCNMHnominal value, whichZno suppy characteristic of ZnAl04(10216-1022.0ev) phase of the catalyst or Cu species agglomerate on the sup-ort shows a single component in the Zn 3P3/2 level indicathe surface to the bulkat 1022.3[38, 39]. For the Al2O3-La203 support, the La 3ds/2 peak lo- port surface resulting in a low metal dispersion. While forated at a binding energy of 834.7eV, close to the typical 30Ni5Cw/Al2O3-Sio2, 30Ni5Cw/Al2O3-La2O3, and 30Ni5CuLifeng Zhang et al. /Joumal of Natural Gas Chemistry Vol. 18 No I 2009Table 6. Binding energy of Ni 2p a core electrons and surface atomic ratios of reduced catalystsBinding energy (ev)Surface atomic ratio(from XPSNi/Al Cu/Al MAl(M=Mg, Zn, La, Si)95527904010070016,05029(04730 NiCaL2O3MgO8525(N°,556)9326(C+,100%)048(0.590.0940.0900270.32)y30NiSCw/Al2O3Zno522(N,60.6%)0%)057(0.590.0490.090054(01658(N2+,39.4%)30 NiSCu/A2O3La2O38523Nw°,41.6%)9323C0+4,10%)0200.59}0.13(0.090)016(0078)8557(Ni2+,584%)30NiSC山/Al2O3SiO28526(N°,6176%)9325(Cu,+1,100%)034(059)0.120.090016(021)The values in parenthesis corresponding to the nominal values/Al2O3-MgO catalysts, the Cu/Al atomic ratio was higher than La203, and 30Ni5Cu/Al2O3 catalysts(as shown in Figure 4a).the nominal values, which is indicative of the high dispersion Figure 8 shows the experimental results of 30NiSCu/Al2Oof Cu species. The percentages of Ni and Cu from XPS on MgO catalyst. When the temperature was lower than 300C,le surface of these samples reflects the important role of dis- intermediate acetaldehyde was detected in products and thepersed entities with aluminium, magnesium, zinc, silicon, and conversion of ethanol was not complete. when the temperalanthanum ions from the supports in affecting the surface state ture ranged from 300 to 600C, the conversion of ethanol wasof metallic Ni and Cu species.up to 100%, and no ethylene and acetaldehyde were detectedThese phenomena imply that the Al2O3 supports modified3.3. Catalytic performance of 30Ni5 Cw/Al2O3-Mgo and by Mgo do not possess dominating acidic sites required for30NiSCwAl203-ZnO at different reaction temperaturesthe dehydration route [45] and cannot promote dehydrationof ethanol to ethylene. This Al2O3-MgO supported catalyst isThe 30Ni5Cu/Al2O3-MgO and 30Ni5Cu/Al2O3-ZnO cat- beneficial to the following reactions: (i)ethanol dehydrogena-alysts were selected for further study at different reaction tion on Cu site and (ii )dehydrogenation reaction produces actemperatures because of their relative high hydrogen selec- etaldehyde as the intermediate product, which undergo C-Ctivity compared with 30NiSCu/Al2O3-SiO2, 30NiSCw/AlObond breaking to produce CO and CHa on Ni site一EtOHH-O-CHICHO25030035040045050Temperature(℃)Temperature(c)Figure 8. Reactant conversion and the selectivity for H2 and other products over the 30Ni5Cu/Al2O3-Mgo catalyst in the steam reforming of ethanol reactionIt can be seen from Figure 8 that the selectivity for hy- monoxide and keep increasing the conversion of water anddrogen increases with temperature. At 300C, the conver- the selectivity for hydrogen, at the same time, methane steamsion of water is only 10.7%, the selectivity for hydrogen, reforming reaction( CH4+2H20-CO2+4H2)was also ac-methane,carbon monoxide, and carbon dioxide is 52.6%, celerated so as to decrease the selectivity for methane.46.6%6, 45.4%, and 7. 2%, respectively. When the temperature中国煤化工600c. the conversionfor hydrogen increasesCNMHGn of methane is lowermethane,CO,and CO2 is 34.6%,15.6%, and 49.0%, respec- than 11.0%, the selectivity for CO2 and CO reaches 55.0%tively. This indicates that the WGSR (CO+H20-CO2+H2) and 32.6%, respectively. It is suggested that the reversehas occurred so as to decrease the selectivity for carbon shift reaction(COz+H2-+Co+H2O)and the reformation ofJournal of Natural Gas Chemistry Vol. 18 No. 1 2009methane with CO2(CH4+CO -+2C0+2H2) were acceler-from Figure 9 that the selectivityated at higher temperature, besides the reformation of methane drogen increased with temperature. At 450.C,theCH4+2H2O→→CO2+4H2 and CH4+H2O—O+3H2sion of water is 32.5%, the selectivity for hydrogen and car-For 30NiSCw/A12O3-ZnO catalyst, the experimental re- bon dioxide is 63.6% and 37.5%, respectively, the selsults are presented in Figure 9. The conversion of ethanol was for methane decreases to 39.1%, and the selectivity for car100% at the temperature ranging from 250-600C, and ac- bon monoxide reach the lowest value of 20.7%. It indicatesmetaldehyde was only detected at the temperature lower than that the WGS reaction(CO+H20-CO2H2)has occurred350C. No ethylene was detected at the temperature range so as to decrease the selectivity for carbon monoxide and keepfrom 250 to 600C, which suggest that 30Ni5Cu/Al2O3-Znincreasing the conversion of water and the selectivity for hy-catalyst cannot promote dehydration of ethanol to ethylene lcause of its weak acidityAs the temperature increases to 600"C, the conversion of250300350400450250300350400450500550600Temperature℃)Temperature(℃)Figure 9. Reactant conversion and the selectivity for H2 and other products over the 30Ni5Cu/AlzO3-zno catalyst in the steam reforming of ethanol reaction atwater keeps increasing to 63.5% and the selectivity forhydrogen increases to 95. 2%; the selectivity for methaneis lower than 7. 2%, the selectivity for carbon dioxide in-creases to the maximum value of 63. 3%, and the selectiv-ty for carbon monoxide increases to 33. 3%. This suggeststhat the reformation of methane( CH4+2H2O-+CO2+4HCH4+H20-C0-+3H2)and reverse shift reactionTG(CO2+H2-CO+H2O)were accelerated while the wGS re-Dsc 2action( CO+H2O-CO2+Hz)is weakened graduallyThe 30Ni5Cu/Al2O 3, 30Ni5Cu/Al2O3-MgO, and30NiSCw/Al2O3-Zno catalysts were operated for 2 h in the90Fb)ethanol steam reforming reaction at each temperature from250 to 600C, with an interval of 50C and heating rate of10C/min, then tGa of these used catalysts were character-ized to analyse the coking effect. As shown in Figure 10, theoxidation peaks at low temperature(lower than 550C)can60be ascribed to carbon deposited on catalyst surfaces [46, 47]in the form of monoatomic carbon and filamentous cokeMonoatomic carbon is highly reactive and easily oxidized bythe surface of Ni [48,49], and deposited filamentous carbonis relatively stable and can react with CO2 at higher temperatures [45]. The peaks at temperatures higher than 550Care ascribed to oxidation of coke deposited with different de-rV凵中国煤化工gree of graphitization formation on Ni catalysts [46]. Thisgraphitic carbon can deactivate and destroy the catalyst [50]CNMHGFor 30NiSCw/Al2O3 catalyst, about 52.7% of deposited coke Figure 10. TG and DSCtre progranis in the form of graphitic coke and monoatomic carbon and oxidation of 30Ni5Cw/AlO (a, 30Ni5Cw/Al2O3-Mgo (b),filamentous coke are about 47.3%. However. for 30Ni5Cu 30NiSCuAl2O3-Zno (e)catalysts used in the reforming of ethanolLifeng Zhang et al Joumal of Natural Gas Chemistry Vol. 18 No. I 2009/Al2O3-Mgo catalyst, the monoatomic carbon and [2] Yang Y, MaJX, Wu F. Int J Hydrogen Energy, 2006, 31 877filamentous coke are as high as 91.5%, the graphitic [3] Fishtik L. Alexander A, Datta R, Geana D Int J Hydrogen En-coke is 8.5% while for 30Ni5Cw/Al2O3-Zn0 catalyst, theergy,200025:31monoatomic carbon and filamentous coke are 90.6%, and [4] Benito M, Sanz J L, Isabel R, Padilla R, Arjona R, Daza LJthe graphitic coke is about 9.5%. Total amount of pro-Power Sources, 2005, 151: 11duced coke is 2.2, 0.5, and 0.2 g/gcat for 30Ni5Cw/Al203, [5] Fatsikostas A N, Kondarides D l Verykios X E. Catal Today,30Ni5Cw/Al2O3-MgO, and 30Ni5Cu/Al catalysts, respectively. The support of Al2O3-MgO and[6] Llorca J, Homs N, Sales J, de la Piscina PR J Catal, 2002. 209:306Al2O3- Zn0 can considerably decrease the amount of [7] Batista MS, Santos R KS, Assaf EM Assaf JM, Ticianelli Ecoke deposited on the surface of catalysts, which is con-A J Power Sources, 2004, 134: 27sistent with the acidity strength of the catalysts, i.e, [8] Breen J P, Burch R, Coleman HM. App! Catal B, 2002, 39: 6530Ni5Cu/AlO3> 30Ni5Cu/Al2O3-Zn0=30Ni5Cu/Al2O3- [9] Liguras D K, Kondarides D, Verykios X E Appl Catal B, 2003MgO. It is the acidic site that can promote dehydration of4:34ethanol to ethylene, which in turn polymerize to form coke on [10] Fierro V, AkdimO, Mirodatos C Green Chem, 2003, 5:20the catalyst surface(C2 H4-+2C+2H2)[14[11] Sun J. Qiu X P, Wu F, Zhu W T, Wang W O, Hao S J. Int JHydrogen Energy, 2004, 29: 10754. Conclusions[12] Fierro V, Klouz V, Akdim O, Mirodatos C Catal Today, 2002,Ni-Cu-based bimetallic catalysts supported on Al2O3- [13] Pompeo F, Nichio NN, Ferretti A, Resasco D Int J HydrogenSioz was prepared to study the effect of Cu content on the catEnergy,2005,30:1399alytic performance in the steam reforming reaction of ethanol. [14] Vizcaino A J, Carrero A, Calles J A. Int J Hydrogen EnergyActivity tests showed that Ni-Cu-based bimetallic catalyst hadthe best catalytic performance when Cu content was 5 wt%. [15] Marino F, Boveri M, Baronetti G, Laborde M Int JHydrogenwith the H2 selectivity for 61.2% at 400C and 92.0% at600C. TPR showed much higher Cu content made the in-[16] Marino F, Boveri M, Baronetti G, Laborde M. int J HydrogerEnergy,2004,29:67teraction between the support and Nio weak[17] Marino FJ, Cerrella EG, Duhalde S, Jobbagy M, Laborde MA.On the basis of the optimal Cu content, Ni-Cu-basedInt J Hydrogen Energy, 1998, 23: 1095bimetallic catalysts supported on AlO3-MyOr(M=Si, La, [18) Marino F, Baronetti G, Jobbagy M, Laborde M. Appl Catal A,Mg, or Zn)were prepared to study the effect of composited 2003, 238:41support on the catalytic performance in the steam reforming [19] Raich B A. Foley H C Ind Eng Chem Res, 1998, 37: 3888reaction of ethanol. The catalysts supported on Al2O3-Mgo [20] Tu Y-J, Chen Y Ind Eng Chem Res, 1998, 37: 2618and Al2O3-Zno have much higher H2 selectivity than that [21] Choudhary VR,Uphade B S, Mamman A S Catal Letf, 1995,on Al2O3-SiO2. Characterizations by TPR, XRD, and Ni332:387TPD suggest that Al2O3-Zn0 and Al2O3-MgO supports in- [22] Bradford MCJ, Vannice M A App! Catal A, 1996, 142: 97teract more strongly with NiO, and their supported catalysts [23] Hu Y H, Ruckenstein E. Catal Rev-Sci Eng, 2002, 44:422have the lowest acidity strength. Further, catalysts supported 24) Ruckenstein E Hul Y H Appl cata A 1999, 183:85on Al2O3-MgO and Al2O3-zn0 were assessed in the ethanol [25] Llorca J, de la Piscina PR, Sales 3. Homs N Chem Commun,2001,(7:641steam reforming reaction at the temperature range of 250- [26) Ahlstrobm-silversand A F, Ingemar Odenbrand C U Appl Catal600C. For 30Ni5Cu/Al2O3-MgO catalyst, the H2 selectivA,1997,153:157ity was 73.3% at 450C and increased to 94.0% at 600"C. [27 Venezia A Mr, La Parola V, Pawelec B, Fierro JL G Appl Catalwhile for 30Ni5Cu/Al2O3- Zno catalyst, the H2 selectivityA,2004,264:43was 63.6% at 450C and increased to 95.2% at 600C. Fi- [28] Robinson WRAM, Mol J C App! Catal, 1988, 44: 165ally, TGA indicated that AlO3-MgO and Al2O3-ZnO sup- [29] Lee JH, Lee E G, JooO S, Jung KD. App! Catal A, 2004, 269ported catalysts have much lower coking amount, and thecoke consists of small fraction of(<10 wt%)graphitic car- [30] Sinfelt J H, Yates D JC.J Catal, 1967, 8: 82bon that deactivates the catalyst. All the results indicate that [31] Grenoble D C, Estadt M M, Ollis D F. Catal. 1981, 67: 9030Ni5Cu/Al2O3-MgO and 30Ni5Cu/Al2O3-Zn0 are promis. [32] Echegoyen Y, Suelves L, Lazaro MJ, Moliner R, Palacios JMJ Power Sources, 2007. 169: 150ng catalysts for ethanol steam reforming reaction with high 33] Dimotakis ED, Pinnavaia T J. Inorg Chem, 1990, 29(13): 2393production of H2 at lower temperature[34] Corma A, Formes V, Martin-Aranda R M, Rey F J Catal, 1992,Acknowledgements1341):58The authors are grateful for the financial support from 973 Pro- [351 Schefter B, Molhoek P, Moulijn JA. App! Catal,1989,46:11gram(2006CB202500) the NSFC (20676096), NCET, and the Pro- [36RlA,1998,167(1):57gram of introducing of discipline to universities (B06006)[37]B中国煤化CNMHwiley, 1990Referencesrf interface Anal, 1992[l] Navarro RM, Alvarez-Galvn M C, Sanchez-Sanchez M C, Rosa [391 Moretti G, Fierro G, Lojacono M, Porta P Surf Interface AnalF, FieroJLG Appl Catal B, 2005, 55: 229989,14:325oumal of Natural Gas Chemistry Vol. 18 No. 1 2009[40] Chen X Y, Liu Y, Niu G X, Yang Z X, Bian M Y, He A. Appl2001,9851catal A,2001,205:159[46] Natesakhawat S. Watson B, Wang X Q Ozkan U S J catal[41] Ledford J S, Houalla M. Proctor A, Hercules D M, Petrakis L, J2005,234:49Phys Chem,198993:6770[47] Sanchez-Sanchez M C, Navarro R M, Fiero JLG Int JHy[42] Guo J-J, Lou H, Zhu Y-H, Zheng X-M. Mater Let, 2003, 57[43]Shalvoy RB, Davis B H, Reucroft PJ Surf Interface Anal, 1980, 148] Wang SB, Lu GQ.App! Catal B, 1998, 19: 267[49] Trimm D L. Catal Today, 1999, 49: 3[44] Velu S, Suzuki k, vijayaraj M, Barman S, Gopinath CS. App! [50] Bengaard H S, Norskov JK, Sehested J, Clausen B S, NieCatal B,2005,55:287LP, Molenbroek A M, Rostrup- Nielsen JR.J Catal, 2002,[45] Fatsikostas A N, Kondarides DL, Verykios XE. Chem Commun中国煤化工CNMHG