Characterization and performance of Cu/ZnO/Al2O3 catalysts prepared via decomposition of M(Cu,Zn)-am

Available online at www.sciencedirect.comJOURNALOFScienceDirectNATURAL GASCHEMISTRYEL SEVIERJournal of Natural Gas Chemistry 20(2011)629 -634www.elsevier.com/locate/jngcCharacterization and performance of Cu/ZnO/Al2O3 catalystsprepared via decomposition of M(Cu,Zn)-ammoniacomplexes under sub-atmospheric pressure formethanol synthesis from H2 and CO2Danjun Wang';2，Jun Zhao'，Huanling Song'，Lingjun Choul*1. State Key Laboratory for Oxo Synthesis and Selective Oxidation, Lanzhou Instiute of Chemical Physics, Chinese Academy of Sciences,Lanzhou 730000, Gansu, China; 2. Graduate School of Chinese Academy of Sciences, Bejing 100049, China[ Manuscript recived Apil 7, 2011; revised August 29, 2011 ]AbstractMethanol synthesis from hydrogenation of CO2 is investigated over Cu/ZmO/Al2O3 catalysts prepared by decomposition of M(Cu,Zn)- ammoniacomplexes (DMAC) at various temperatures. The catalysts were characterized in detail, including X-ray diffraction, N2 adsorption-desorption,N2O chemisorption, temperature-programmed reduction and evolved gas analyses. The influences of DMAC temperature, reaction temperatureand specific Cu surface area on catalytic performance are investigated. It is considered that the aurichalcite phase in the precursor plays a keyrole in improving the physiochemical properties and activities of the final catalysts. The catalyst from rich-aurichalcite precursor exhibits largespecific Cu surface area and high space time yield of methanol (212 g/(Lcat:h); T =513K, p=3 MPa, SV= 12000 h-1).Key wordsdecompostion of M(Cu,Zn)-ammonia complexes; Cu/ZnO/Al2O3 catalyst; CO2 hydrogenation; methanol synthesis1. Introductiontion conditions [8]. Cu/ZnO/Al2O3 catalyst commonly wasprepared by conventional coprecipitation method with sodiumOver the past decades, considerable attention has beencarbonate [9]. The activity of Cu/ZnO/Al2O3 catalysts waspaid to reduce CO2 emissions by effectively utilizing CO2very sensitive to their structure, so even slightly different indue to the more and more serious greenhouse effect [1]. .preparation conditions may cause considerable changes in theGreat efforts have been made to convert CO2 to various use-structure of the precursors and final catalysts, subsequentlyful chemical products for reducing CO2 emission as far asaffected the catalytic performance [10,11]. In addition, resid-possible [2,3].ual sodium strongly inhibited the interaction of Cu with pro-Methanol synthesis from CO2 and H2 is considered asmoters and decreased the dispersion on the support matrixone of the most economical processes for CO2 utilization [4].[12], whereas effective removing of residual sodium wasAs an excellent liquid fuel, methanol could provide conve-difficult when sodium-containing precipitant was employednient storage of energy in the fields of transportation and fuel[13]. Thus, various novel preparation methods, such as ox-cell application [5]. Commercially, ternary Cu/ZnO/Al2O3alate coprecipitation [14], water-swellable polymer networkscatalyst was used for the synthesis of methanol from syn-[15] and gel-network coprecipitation [16], have been devel-gas (H2/CO/CO2) with a large scale at 493 K- 573 K and 5-oped. However these preparation procedures were not envi-10 MPa [6,7]. However, when the catalyst was used for CO2ronmentally friendly because lots of organic materials werehydrogenation to methanol, the yield of methanol was muchneeded. To remove those organic materials, usually high tem-lower than that obtained from syngas under the similar reac-perature was necessary.中国煤化工* Corresponding author. Tel: +86 931-4968066 Fax: +86 931-4968129; E-mail: ljchou@licp.This work was supported by the National Basic Research Program of China (No.201 1CB20MYHc N M.H Ghe Suate Key Lboarorfor Oxo Synthesis and Selctive Oxidation (OSSO) of China.CopyrightO2011, Dalian Instiute of Chemical Physics, Chinese Academy of Sciences. All rights reserved.doi:10.1016/S1003-9953(10)60246-0630Danjun Wang et al./ Joumal of Natural Gas Chemistry Vol. 20 No.6 2011In this work, a series of Cu/ZnO/Al2O3 catalysts wereprepared through decomposition of M(Cu,Zn)-ammonia com-Temperature-programmed reduction (TPR) experimentsplexes (DMAC) under sub-atmospheric pressure at vari-were also conducted on AMI- 100 instrument. The fresh cal-ous temperatures, and they were characterized by X-raycined catalyst about 20 mg (80- 100 mesh) was pretreated indiffraction (XRD), total surface area, N2O chemisorption,a helium flow of 50 mL /min at 623 K for 1 h and then cooledtemperature-programmed reduction (TPR) and evolved gasto room temperature. After that, a reducing gas composed ofanalyses (EGA). The catalytic performance of these catalysts5 vol% H2/He was employed at a flow rate of 50 mL/min,was evaluated and the relationship was correlated with thewith a heating rate of 10 K/min from room temperature tocharacterization results.673 K. The hydrogen consumption was monitored online byquadrupole mass spectrometer.2. ExperimentalThe specific surface area of Cu (Scu) was determined bydissociative N2O adsorption at 363 K using the procedure de-2.1. Catalyst preparationscribed by Huang et al. [17]. The catalysts were pre-reducedat 623 K for 1 h. After that they were flushed with helium gasA series of Cu/ZnO/Al2O3 catalysts (molar ratio ofand the temperature was decreased to 363 K. Subsequently,Cu:Zn:Al=6:3:1) were prepared via decomposition5 vol% N2O (99.995% purity) in N2 99.99% purity) was in-f M(Cu,Zn)-ammonia complexes (DMAC) under sub-troduced into the reactor at this temperature with a flow rateatmospheric pressure as follows. 6.18g CuO and 3.16gof 50 mL/min for 30 min. Finally, the samples were cooled toZnO powder were introduced into 100 mL 2 mol/L NH3.H2O-room temperature with flowing helium gas and then heated up(NH4)2CO3 (NH3:H2O : (NH4)2CO3= 1: 1) solution underto 673 K with a rate of 10 K/min. Dispersion of Cu (Dcu), di-vigorous stirring at 333 K for 5 h to form M(Cu,Zn)-ammoniaameter of Cu particle (dcu) and Scu were calculated based oncomplexes solution. 100 mL 0.4 mol/L NH3 H2O solutionassuming a Cu :N2O=2: 1 chemisorption stoichiometry, awas added dropwise into 100 mL 0.13 mol/L Al(NO3)3 so-surface atomic density of 1.46x 1019 Cu atom/m2 and a spher-lution with vigorous stirring at room temperature and thenical shape of Cu particles, respectively [18].white suspension of aluminum hydroxide was obtained. Afterthat the M(Cu,Zn)-ammonia complexes solution was added2.3. Catalytic activity testinto this suspension with vigorous stiring and then a vacuumpump was equipped and kept the system maintaining sub-Catalytic tests for CO2 hydrogenation were carried out inatmospheric pressure (0.9x 105 Pa) at different temperaturesa fixed-bed microreactor (i.d. 15 mm).0.5 mL catalyst (20 -(T = 323, 333, 343, 353 and 363 K) for 2-5 h to completely40 mesh) was loaded in the reactor and the remaining space ofdecompose the M(Cu,Zn)-ammonia complexes. The productthe reactor was flled with quartz sands. The temperature waswas filtrated and washed with distilled water, dried at 393 Kcontrolled by a thermocouple in contact with the catalyst bed.overnight and further calcined at 633 K for 4 h under air atmo-The pressure in the reactor was maintained by a back pressuresphere. The as prepared catalyst was pressed, crushed, sievedregulator and the flow rate of feed gas was contolled by massto the particle size fraction (20- 40 mesh for catalytic testing,flow controller. A cold trap was used for separation of liquid80- 100 mesh for characterization). As-prepared precursors,product. Before each test, the catalysts were reduced in situ atsamples after calcination and reaction were denoted as CZA-575 K for 6 h in 5% H2/N2 flow (100 mL /min) at atmosphericP-T, CZA-T; and CZA-R-T, respectively, where, T representspressure. After reduction, the reactor was cooled to reactionthe DMAC temperature in Kelvin.temperature and the feed gas (H2/CO2/N2 = 75/25/1) was fedinto. The outlet gas was analyzed under steady state by an on-2.2. Catalyst characteriationline gas chromatograph with a thermal conductivity detector(TCD), equipped with a TDX-01 column. The liquid prod-X-ray diffraction (XRD) measurement was performed onucts were analyzed offline by a chromatograph with a flamea PANalaytical X' pert Pro XRD instrument using Cu Ka ra-ionizing detector (FID) equipped with a FFAP column.diation in the scanning angle range of 10*-800 at a scanningrate of 4%/min at 40 mA and 40kV.The surface areas (SBET) and pore volume (Vp) of the cal-3. Results and discussioncined samples were calculated from N2 adsorption-desorption3.1. Catalyst characterizationdata at 77 K on an Autosorb iQ (Quantachrome) instrument.The powder was pre-outgassed at 623 K to ensure a clean sur-face before acquisition of the adsorption isotherm.3.1.1. BET中国煤化IersionEvolved gas analyses (EGA) were conducted on an AMI-CNMH G100 catalyst characterization instrument. The precursor sam-As shown n raole 1, LLA-, CZA-343 and CZA-INYHple about 10 mg (80- 100 mesh) was heated up to 1073 K un-353 which prepared at medium temperature showed largerder helium flow (80 mL/min) with heating rate of 5 K/min.SBET values of 79.7, 79.8 and 78.1 m2/g. However, CZA-The evolved gas was monitored online by quadrupole mass323 and CZA-363 displayed relatively smaller SBET (71.2 andJourmal of Natural Gas Chemistry Vol. 20No.6201163169.0 m2/g), indicating medium DMAC temperature is neces-sary for obtaining large surface area, whereas too low or high) AurichalceiteDMAC temperature would result in the decline of SBET. TheMalachitea)trend of Scu with DMAC temperature was similar to that oftie iwhvimnCuOn CZA-P-363SBET. It can be noted that CZA -343 exhibited much higherin_ jScu than CZA-333 and CZA-353, notwithstanding it showedclose SBET value with them.CZA-P-343Table 1. Physicochemical properties of the studied catalystsCA.P-333Catalysts SBET (m2/gea)__ So (m-/gca)品 (nm) COnO (nm)CZA-32326.620.1CZA-33379.728.320.55..CZA-34379. .8301605.lji JWhie10.9CZA-353 .78.126.96.3(407080CZA-36369.026.725.98.820/(° )a From (363 K) N2O titration measurements; b Calculated from thefull width of half maximum of the reflections ofCu(1 1 1) or CuO。CuO.(b)(1 1 1) planes in XRD patterns using Scherrer equation0 ZnOCZA-3533.1.2. XRD analysesFigure 1(a) shows the XRD patterns of the precur-sors. Two phases were observed from the XRD patterns.The peak at 20= 13.10 were assigned to aurichalcite phaseCuO/5((Cu,Zn)s(CO3)2(OH)6). The peaks at 20= 14.80, 17.50,31.50, 32.20 and 35.60 belonged to malachite phase. In ad-ZnO/5dition, the peaks approximate at 20 = 24.10 was ascribed to102040 50 600the overlapping curves of malachite and aurichalcite phases.In order to understand the relative content of the two phasesin the precursors, the representative peak at 20 = 13.10 for au-。Curichalcite and 20 = 17.50 for malachite were employed. Table 2 shows the relative intensities expressed by the represen-CZA-R-363tative peak height ratio in the catalyst precursors. With the in-crease of DMAC temperature, the relative intensity increased会。CZA-R-353first and maximized at 343 K, then further elevating tempera-ture resulted in a decrease of the intensity. Namely, the precur-自L CZA-R-343sor prepared at 343 K contains much more aurichalcite phaseCZA-R-333than those prepared at other temperatures. Baltes reported [9]that aurichalcite phase in catalyst precursors can improve Cu. CZA-R-323dispersion and the activity of the final catalyst. Because a partof Cu is atomically substituted by Zn in aurichalcite phase,4Cit results in small and homogenous CuO and ZnO crystallites201(° )that are ideal intermediate for the higher activity [19]. In addi-Figure 1. XRD patterns of CwZnO/Al2O3 catalysts before calcination (a),tion, the broad peaks at 20 = 35.60 and 38.7° for CZA-P-363after calcination (b), and after reaction (C)were attributed to the formation of CuO phase due to hightemperature in the precipitation procedure.Figure 1(b) shows the XRD patterns of the calcined cata-lysts as well as the reference samples of pure CuO and ZnO.Table 2. Relative intensities of aurichalcite (20 = 13.19) and malachiteThe features of hydroxycarbonates in the precursors disap-(20= 17.50) phases in the precursorspeared after calcination but some new broad peaks related toSamples13.117.513.1//17.5metal oxides appeared. The peaks at 20 = 35.60, 38.7", 48.8",CZA-P-323150012501.2161.69, 66.20 and 68.10 are attributed to CuO. The peaks atCZA-P-333174013201.3220 =31.7° a中国煤化工). However, the fea-170010601.61tures ascribedCNMHG! is probably becauseCZA-P-35322800.62aluminum isor amorphous stateCZA-P-3638007001.15due to the relatively low calcination temperature (633 K). InI2g: intensities were approximated from peak heightsaddition, the weak and broad peaks comparing with those ofthe bulk CuO and ZnO indicate that the crystallite sizes of632Danjun Wang et al./ Joumal of Natural Gas Chemistry Vol. 20 No.6 2011CuO and ZnO in the catalyst were very fine. The average par-(5%H2/95%He), there was no hint of the formation of inter-ticle sizes of CuO calculated from XRD patterns using Scher-mediate phase (Cu4O3, Cu2O) during the reduction. Accord-rer equation are shown in Table 1. The particle size of catalystingly, the mechanism of stepwise reduction of CuO may beCZA-343 was clearly smaller than those of the other catalystsunsuitable for this work. In Cu/ZnO/Al2O3 termary catalyst,in Table 1. It is consistent with the view that aurichalcite phasebesides CuO, others Cu(II) species such as Cu2+ ions embed-led to smaller Cu particle after calcination because of the sub-ded in the octahedral sites of Al2O3 or in zinc oxide and zincstitution between Cu and Zn in atom level.aluminate may also exist. But these species should undergoAs shown in Figure 1(c), the strong diffraction peak athigher reduction temperature than CuO [22,23,24]. In our20 = 43.3° along with two weak ones at 20 = 50.4° and 74.10samples, the reduction temperatures were lower than that of.were identified to Cu phase. In addition, the others weakpure Cu0.peaks at 20 = 31.70, 34.40, 36.3°, 47.60, 56.7", 62.80, 68.09were ascribed to ZnO phase. The Cu crstallite sizes calcu-531lated by Scherrer formula are also listed in Table 1. It is clearthat the copper particle size became much larger compared4531258CZA-363with fresh calcination ones, indicating remarkable agglomer-ation of copper species upon reduction and/or reaction. It was450525\CZA-353considered that the water derived from reduction and reactionprocess aggravated the growth of Cu particle [7]. Notwith-CZA-343standing the sharp growth, the CZA-R- 343 still displayed rel-CZA-333ative smaller Cu particle size (16.9 nm) than these of others536CZA-R- T catalysts with larger than 20 nm, indicating a rela-tively better inhibition of Cu particle growth for CZA-343._450CZA-3237317357377733.1.3. Temperature-programmed reductionTemperature (K)Figure 2. TPR profiles of Cu/ZnO/Al2O3 catalystsThe TPR profiles of freshly calcined catalysts are shownin Figure 2. There were two or three reduction processes forIn the present work only a small amount of H2 was con-each catalyst. The onset of reduction (To) was apparent at ap-sumed in the early stage at lower temperature. Based onproximate 450 K and H2 consumption increased slowly beforethese, lower temperature features (about 450- 500 K) may be500 K. Then, the consumption increased rapidly and reachedattributed to the reduction of amorphous CuO isolated fromto a maximum (Tm) at 536, 525, 525 and 531 K for CZA-support. While the main reduction peaks at relative higher(323, 333, 343 and 353), respectively. For CZA-323, besidestemperature (Ca. 525- -537 K) should be related to the smallthe main peak, a left shoulder peak appeared at 518 K. CZA-CuO particle which intimately contact with ZnO and Al2O3.363 showed much more complex behavior, with several over-For CZA 363, the wider range of reduction temperature andlapping reduced peaks from 450K to 586K. It is clear thatthe shoulder peak at higher temperature (571 K) may be as-these main peaks were much lower than that of pure CuO atcribed to the larger CuO particle and less uniform particle sizeabout 738 K [12], indicating that ZnO and Al2O3 enhance thedistribution.reducibility of CuO by forming small CuO particle well dis-persed on the surfaces of ZnO and Al2O3.3.1.4. Evolved gas analysesThe presence of more than one reduction signal in TPRprofiles of Cu/ZnO/AI2O3 catalysts was already observed andEvolved gas analysis (EGA) of CO2 provides a detailedexplained by various mechanisms in previous reports. Turcopicture of the decomposition process (Figure 3). There wereet al. [20] suggested that the reduction of CuO proceeds step-at least four decomposition steps. The first released step be-wise (Cu2+ -→Cu+ -→Cu), judging from the two similar H2fore 408 K was related to the adsorption of CO2. The secondconsumption of Cu/ZnO/Al2O3 catalyst derived from hydro-peak at about 533 K was attributed to the decomposition oftalcite precursors. In our experiment, only a small amount ofmalachite. The subsequence release of CO2 centered at ap-H2 was consumed in the early stage at lower region. Whileproximate 643 K should be ascribed to the decomposition ofmost of H2 consumption occurred in the subsequent reduc-aurichalcite. Except CZA-P-363, the fourth decompositiontion process. In fact, Kim et al. [21] investigated the kineticpeak appeared at about 746 K. According to the works re-effects in the formation of suboxides in the reduction processported by Xia et alh [25.261. this CO release peak at highof Cu0 by in situ time-resolved XRD method. According totemperature中国煤化工econdary decomposi-their result, Cu2O could be formed as an intermediate onlytion of hydr. CNMH G27] also reported thiswith high heating rate (40 K/min) or limiting the supply of H2phenomenoll inLu/zllU sysiclu CUICCI ning residual carbon-at low flowing rate of 1 mL/min (5%H2/95%He). However,ate species. They thought that this carbonate species wouldunder normal conditions such as employed in our experimentrestrain the growth of Cu particle, thus leading to highly ac-with heating rate of 10 K/min and flowing rate of 40 mI ./mintive catalyst. In this work, CZA-P-343 showed stronger CO2634Danjun Wang et al./ Journal of Natural Gas Chemistry Vol. 20 No.62011This was in line with its nature, i.e. better physicochemi-Acknowledgementscal properties such as larger surface area, better dispersion,The authors thank the financial support of the National Ba-intimate interaction with promoters and residuary carbonatesic Research Program of PR China (No. 201 1CB201404) and thespecies. Figure 5 ilustrates the methanol yield as a functionfinancial support of the State Key Laboratory for Oxo Synthesis andof Scu. The methanol yield increased with Scu in a relativeSelective Oxidation (OSSO) of China.linear way for CZA-323, 333, 343 and 353. Similar resultwas reported by other authors [28] and Scu was considered asReferencesthe important factor affecting the activity of methanol synthe-sis reaction. It should be noted that CZA 363 showed almost[1]XuAM,IndalaS,HertwigTA,PikeRw,KnopfFC,YawsC.the same Scu value with CZA-323 and CZA-353, but muchL, Hopper J R. Clean Technol Environ Policy, 2005, 7(2): 97lower methanol synthesis activity than them. Similar phe-[2]MaJ,SunNN,ZhangXL,ZhaoN,XiaoFK,WeiW,SunYnomenon was discovered over CP catalyst. The poor catalyticH. Catal Today, 2009, 148(3-4): 221performance of CZA- 363 may be related to the ill interaction[3] Sakakura T, Choi J C, Yasuda H. Chem Rev, 2007, 107(6): 2365of Cu with the promoters resulted from the decomposition of[4] Joo 0 s, Jung K D, Moon 1, RozovskiiA Y, Lin GI, Han S H,Uhm S J. Ind Eng Chem Res, 1999, 38(5): 1808metal hydroxide carbonate in the precipitation process. These[5] Turco M, Bagnasco G, Cammarano C, Senese P, Costantino U,data indicate that specific Cu surface area may not be the onlySisani M Appl Catal B, 2007, 77(1-2): 46factor that determines the activity, and different particle sizes[6] Arena F, Italiano G, Barbera K, Bordiga s, Bonura G, Spadaroand interaction with support may strongly affect the activityL, Frusteri F Appl Catal A, 2008, 350(1): 16as well.[7] Behrens M, Furche A, Kasatkin I, Trunschke A, Busser W, Muh-ler M, Kniep B, Fischer R, Schlogl R. ChemCatChem, 2010,2202(7): 816[8] Liu Y, Zhang Y, Wang T J, Tsubaki N. Chem Lett, 2007, 36(9):200CZA-3431182[9] Baltes C, Vukojevic S, Schith F. J Catal, 2008, 258(2): 334CZA-353- CZA-333 .[10] LiuX M, LuG Q, Yan z F, Beltramini J. Ind Eng Chem Res,1802003, 42(25): 6518160CZA-323[11] LiJ L, Inui T. Appl Catal A, 1996, 137(1): 105[12]JunKW,ShenWJ,RaoKSR,LeeKW.ApplCaralA,1998,174(1-2): 231140[13] Wu J G, Luo S C, Toyir J, Saito M, Takeuchi M, Watanabe T.Catal Today, 1998, 45(1-4): 21520 f[14]MaY,SunQ,WuD,FanWH,ZhangYL,DengJF.ApplCatalCZA-363A, 1998, 171(1): 45[15] Ruckenstein E, Hong L. Chem Mater, 1996, 8(2): 54630[16] Hong Z s, Cao Y, DengJ F, Fan K N. Catal Lett, 2002, 82(1-2):Se(m'g)37Figure 5. Methanol yield as a function of specific Cu surface arca.[17] Huang Z w, Cui F, Kang H X, Chen J, ZhangXZ, XiaCG.Reaction conditions;SV= 12000h-1, p=3 MPa, and molar ratio ofChem Mater, 2008, 20(15): 5090CO2/H2/Nz = 25/75/118] Arena F, Barbera K, Italiano G, Bonura G, Spadaro L, FrusteriF.J Catal, 2007, 249(2): 185[19] Atake I, Nishida K, Li D, Shishido T, Oumi Y, Sano T, Takehira4. ConclusionsK. J Mol Catal A, 2007, 275(1-2): 130[20] Turco M, Bagnasco G, Costantino U, Marmottini F, MontanariHydrogenation of CO2 to methanol over Cu/ZnO/AI2O3T, Ramis G, Busca G. J Catal, 2004, 228(1): 43catalysts derived from hydroxycarbonate precursors via[21] KimJ Y, Rodriguez J A, Hanson J C, Frenkel AL, LeePL.JDMAC method is investigated. The precursor prepared atAm Chem Soc, 2003, 125(35): 10684DMAC temperature of 343 K is rich in aurichalcite phase[22] Gervasini A, Bennici S. Appl Catal A, 2005, 281(1-2): 199which is an ideal intermediate for higher activity of the cata-[23] Yahiro H, Nakaya K, Yamamoto T, Saiki K, Yamaura H. CatalCommun, 2006, 7(4): 228lyst. The catalyst from precursor with rich aurichalcite phase[24] Velu s, Suzuki K, Okazaki M, Kapoor M P, Osaki T, Ohashi Fexhibites smaller particle size, larger surface area and betterJ Catal, 2000, 194(2): 373Cu dispersion that were useful for the adsorption and acti-[25]XiaWQ,TangHD,LinsD,CenYQ,LiuHZ.ChinJCatalvation of reactant gas. In addition, the residue of the high(Cuihua Xuehao) 2009.30(9): 879temperature carbonate species confirmed by EGA analyses[26] Lin S D,中国煤化工:nY Q, Liu HZ. ChinJis helpful for inhibiting the growth of Cu particle, thus en-Caral (qC257hancing the activity of the catalyst. Indeed, CZA-343 catalyst[27] Bems B,CNM H SH HeinD. Scbog[R.which contains much more aurichalcite phase in the precur-Chem Eur J, 2003, 9(9): 2039sor and high temperature carbonate species upon calcination [28] An X, LiJ L, Zuo Y Z, Zhang Q, Wang D Z, Wang J F. Catalshows much higher activity.Lett, 2007, 118(3-4): 264