Synthesis of nanocrystalline γ-Al2O3 by sol-gel and precipitation methods for methanol dehydration t

Available online at www.sciencedirect.com| Joumal ofScienceDirect7I Natural GasChemistryEL SEVIERJournal of Natural Gas Chemistry 20(201 1)128-134www.elsevier.com/locate/jngcSynthesis of nanocrystalline ry-Al2O3 by sol-gel and precipitationmethods for methanol dehydration to dimethyl etherZahra Hosseinil，Majid Taghizadeh1*，Fereydoon Yaripour?2I. Chemical Engineering Department, Babol University of Technology P. 0. Box 484, 4714871167 Babol, Iran;2. Catalyst Research Group, Petrochemical Research and Technology Company NPC, Tehran, Iran[ Manuscript received October 20, 2010; revised December 5, 2010 ]JoadAbstractThe capability of sol-gel and conventional precipitation techniques for the synthesis of nanocrystalline y-alumina was investigated. Thesecatalysts were used for vapor-phase dehydration of methanol to dimethyl ether in a fixed-bed reactor under the same operating conditions(T= 300°C, P= 1 bar, LHSV=2.8, 11.7, 26.1 h-'y nd charagteriped by meats of N2 adsorption-desorption, NHg-TPD, XRD, TGA andSEM techniques. According to the experimental reqlts, the caplysts prelsing sol-gel method in non-aqueous medium showed betterperformance compared with those prepared by other mhethods.Key words7~Al2O3; sol-gel; precipitation; methanol dehyd rationg dimnethylMurd Cs1. Introductionlays,vot phosphates such as aluminium phosphatesVI VIiwywT mentoned. Meanwhile, y-alumina is a very efecivceDemethyl ether (DME) has potential appitcations as acatalyst for this reaction, and it is believed that most of the acidchemical building block. In addition, since the characteris-sites on this catalyst are of the Lewis type. However, there istics of DME are similar to those of liquefied petroleum gasno consensus in the scientific community as to whether the(LPG) and given its high heating value, it has been speculatedLewis sites can be converted to Bronsted sites in the pres-that DME could be used in large scale power production, inence of water [8- 10]. Special interest for catalysis applica-home heating, in replacement of LPG for automobiles, and astions of nanoparticles is due to their unique properties. Ra diesel fuel substitute or combustion supplement. In thesecently, many different methods have been investigated in (applications, the specification for the purity of DME has beender to synthesize mesoporous nanocrystalline 7-Al2O3 withreported to be lower than 99 wt% requirement for curent uses higher specific surface area [11- 15]. Chemical routes for the(DME fuel grade is 88.9%)[1-5].production of these materials include sol-gel method, and con-There are three routes for DME production. Nowadays,trolled precipitation of boehmite obtained from alkoxides andthe DME world's production is performed with one of thealuminum salts [16,17]. The advantage of the sol gel methodfollowing routes: methanol dehydration, methanol/DME CO-includes the ability of maintaining a high degree of purity,production and direct synthesis [6,7].changing physical characteristics such as pore size distribu-Several catalysts having activity for the catalytic conver-tion and pore volume, and preparing samples at low temper-atures. The main factors which influence the structural prop-dehydration catalysts. It is generally accepted that acid cata-erties of the final r-alumina samples in the sol-gel techniquelysts are the best materials for the dehydration of methanol to are the hydrolysis ratio (moles of water per mole of aluminumdimethyl ether and that Bronsted or Lewis acids are capableprecursor), acidity of the precursor solution and the molarof performing this reaction. As examples of known acidic de-ratio of template to aluminum precursor. Combining a highhydration catalysts alumina, e.g. y-alumina, silica, alumina-porosity with high surface area improves the overall physicalsilica mixtures, crystalline aluminosilicates, crystalline zeo-properties and consequently increases the catalyst life [11,18].“Corresponding author. Tel: +98-111-3234204; Fax: +98-111-3234201; E-mail: m. taghizadehfr@y中国煤化工This work was financially supprted by Iranian Nanotechnology Initiative Council.YHCNM HGCopyrightO201 1, Dalian Institute of Chemical Pnese Academy of Sciences. All rights reserved.doi: 10.1016/S 10039953(10)60172-7Journal of Natural Gas Chemistry Vol. 20 No.2 2011129In this work, the effects of different preparation meth-ful pH adjustment to 7.5. Thetemperature was maintaods on catalytic properties and physical structure of nanocrys- 60 °C during precipitation step. The precipitate was digestedtalline 7/-Al2O3 as a solid-acid catalyst have been investigated. at 60 °C for 3 h. Then, the precipitate was filtered and dried atThe solid-acid catalyst has been tested in a fixed-bed reactor120。C for 24 h. The final solid was then calcined at 600 °Cand characterized by TGA, N2 adsorption- desorption, XRD,for 6 h at a heating rate of 2 °C/min and named as PR-C.SEM and NH3-TPD.2.2.4. Commercial reference sample2. ExperimentalFor comparison, a commercially available 7-alumina2.1. Materialssample was used as a reference catalyst. This sample wassupplied from BASF company and named REF-C.Aluminium iso-propoxide (AIP, 99 wt%)， 2-propanol(IPA), propionic acid (AP), nitric acid (65 wt%), aluminum2.3. Characterizationnitrate nonahydrate (ANN, 98.5 wt%) and potassium bicar-bonate purchased from Merck Company were used as startingThe surface area, pore size distribution and pore vol-chemicals in different methods.ume were determined using a nitrogen adsorption-desorptionisotherm at the liquid-nitrogen temperature (- 196°C) by a2.2. Catalyst preparationNOVA 2200 instrument (Quantachrome, Boynton Beach, FL).Prior to the adsorption-desorption measurements, all the sam-2.2.1. Sol-gel method in a non-aqueous mediumples were degassed at 200 °C in a N2 flow for 16 h to removethe moisture and other impurities.Aluminum iso-propoxide, 2-propanol and propionic acidThe X-ray diffraction (XRD) patterns of all the cal-were used as aluminum precursor, solvent and hydrolysis rate cined samples were recorded on a Bruker dg advance (40 kV,controller, respectively, during synthesis. Initially, AIP was30 mA) X-ray diffractometer, using a Cu Ka radiation sourcedissolved in 2-propanol under continuous stiring. Then, the(\= 1.5406 A) and a nickel filter in the 20 range of 0*-70.mixture of propionic acid and water was added to the aboveThe proper temperature of the calcination processsolution. The gel was further aged at 70°C for 24h. Theof the dried samples was determined on a thermogravi-molar ratios of AP/AIP, H2O/AIP, and IPA/AIP were 0.5,metric analyzer using Diamond thermogravimetric analand 25 during processing, respectively. Finally, the materialysis/differential thermal analysis (TGA/DTA) instrumentswas dried in an oven at 120 °C for 24 h and calcined in a flow(PerkinElmer, Waltham, MA). Approximately 5 mg of sam-of air at 600 °C for 6h with a heating rate of 2 °C/min. Theple was heated from 25 to 750°C at 5。C/min in a flow ofobtained catalyst is referred herein as SGO-C.nitrogen (75 mL/min) until no weight loss occurred.The acidity of the samples was measured via ammonia2.2.2. Sol-gel method in an aqueous mediumtemperature-programmed desorption (NH3-TPD) titration, us-ing a PulseChemiSorb 2705 instrument Micromeritics, Nor-cross, GA) with a conventional flow apparatus,which in-In this method, aluminum isopropoxide was dissolved inan appropriate amount of water. The molar ratio of water tocluded an online thermal conductivity detector (TCD). AboutAIP was chosen as 80. The hydrolysis step was carried out0.25 g of the sample was initially flushed with a helium flowat temperatures higher than 70 °C for a time period of 30 minat 300°C for 2h, next cooled to 150 °C and then saturatedunder vigorous stirring. Then, the mixture was peptized bywith pure NH3 for 60 min. The sample was then purged withadding nitric acid (65 wt%) under vigorous stiring by care- a helium flow for 90 min to remove weakly and physicallyful pH adjusment to 4.5. The mixture was aged at 70°C for adsorbed NH3 on the surface of the catalyst. After this op-5 h, and then dried in an oven at 120 °C for 24 h. Finally the .eration, the sample was cooled to room temperature and thensample was calcined at 600 °C for 6 h with a heating rate ofheated at a rate of 10 °C/min under a flow of helium carrier2 °C/min. The obtained catalyst is referred herein as SGA-C.gas (40 mL/min) from 30 to 850 °C. Finally, the amount ofNH3 in the effluent was measured using TCD and recorded as2.2.3. Precipitation methoda function of the temperature.Scanning elctron microscope (SEM) was performed withIn this method, aluminum nitrate, potassium bicarbonatea TESCAN-VEGA scanning electron microscope, operatingat15kV.and distillated water were used as starting chemicals. At firstan appropriate amount of aluminium nitrate nonahydrate wasdissolved in a certain amount of water. The molar ratio of2.4. Catalytic acti中国煤化工water to ANN was chosen as 1. After that the precipitationCNMHGwas carried out by adding aqueous potassium bicarbonate toVapor phase druyu.u luliaiiui wdS conducted inthe solution of ANN under vigorous strring at 60 °C by care-the laboratory catalytic evaluation system. Figure 1 shows a130Zahra Hosseini et alL./ Journal of Natural Gas Chemistry Vol. 20 No, 22011simplified flow diagram of the unit. The reactor system wascompletely and sample has reached a stable structure. There-entirely placed in a temperature-controlled oven so that thefore, it is suitable to perform the calcination process at thefeed can be preheated and the condensation of products cantemperature of about 600 °C.be avoided. The fixed-bed microreactor was made of stain-lessness steel (.D. = 9.2 mm; L= 650 mm). The reactor washeated by three individual furnaces at the top, middle and bot-开0.tom sections of the reactor tube. In an experiment, 1 g of cat-alyst (size 60- 120 mesh) was loaded in the middle section8.194%of the reactor tube and methanol as feed was pumped with)0 F- -0.different liquid hourly space velocities (LHSV) from 2.8 to .、10.892%26.1 h" -1. Prior to catalytic activity measurements, the samplewas in situ heat-treated, at a heating rate of 5 °C/min in a flow巨80of nitrogen (50 mL/min) at 220°C for 2h. Then methanol--1.0was injected from a feed tank to the pre-heater and activitytests were conducted at 300 °C and atmospheric70 F28.351%the fixed-bed reactor. The reaction products were analyzed+-1.5by an on-line gas chromatograph (GC) of Varian CP- 3800,0Fequipped with FID and TCD with a capillary column (CP-PoraPlot Q-HT). It should be noted that the transfer line from= -2.1100 200 300 400 500 600 700180 °C. The reaction performance results, including MeOHTemperature(C)conversion, DME selectivity and yield, were calculated sub-sequently.Figure 2. TG/DTA plots of the dried sample of SG0-C→D -MFC- .3.2. Phase canalysis7 Mixer and preheaterThe XRD patterns of the synthesized r/-alumina sampleswith different methods as well as commercial catalyst areTIC 1Fixed-bed reactorshown in Figure 3. This figure indicated that all the preparedNitrogenvesselsamples, which were calcined at 600 °C, had a crystallitealumina phase (JCPDS NO: 1-1307). Also, comparison ofTIC3-XRD patterns of the prepared samples and commercial cata-lyst showed that the extent and percentage of crystallinity ofO_Methanolcatalysts do not differ significantly. These results indicatedPumpthat the preparation method did not have a strong influenceFigurA simplified flow diagram of the laboratory catalytic evaluationon the crystallite phase and also the crystallite size of the alu-systemminium oxide.3. Results and discussion3.1. Thermal analysisFigure 2 shows the TG/DTA plots for the dried sample ofSGO-C. The DTA curve presents three major peaks among言REF-Cwhich two are endothermic and the main peak is exother-mic. The first endothermic peak appears at low tempera-ture of about 150 °C, which corresponds to the eliminationof residual water and moisture, and the second one appears at。 SGA-Cabout 300 °C, which is due to the removal of hydroxyl groupbonded on the surface of sample. The main exothermic peakSG0-Cat 600 °C is atributed to the elimination of OH groups, oxi-dation of organic volatile residues, and the crystallization of中国煤化工- 60amorphous alumina. Simultaneously, the TG curve levels offYHCNMHGat about 600 °C and shows no weight loss at higher temper-atures, meaning that the 02@ residues have B0removed40e 3. XRD patemrs o5rent samples and co6oil calyst7020/10)Journal of Natural Gas Chemistry Vol. 20 No.2 2011131The crystallite sizes were calculated using the Scherrer3.3. BET and pore size distribution analysisequation:Physical properties of the catalysts (specific surface area,D=(1)total pore volume, and average pore diameter) are presented inβcos0Table 1. The obtained results show that the prepared catalystswith sol-gel method have a higher specific surface area andwhere, k: is a constant generally taken as unity, 入is thesmaller pore diameter in comparison with other catalysts andwavelength of the incident radiation, β is the full width ofcommercial sample. Also, evaluation of results indicates thatdiffraction peak at half maximum intensity (FWHM) and 0 isthe catalyst prepared by the sol-gel method showed lower porethe diffraction angle. The results are presented in Table 1.volume than the sample prepared with precipitation method.Table 1. Physical properties of 7-Al2O3 samples prepared by different methodsSample code .Surface area (m-/g)Pore volume (cm^/g)Pore diamete,(92)Crystalite size (nm) aParticle size (nm) tψCSGO-C300.554.71.11SGA-C2480.437.04.96.01.87PR-C0.80264.4420.424.85.86a Determinated by XRD results;b Determinated by BET surface area;° Partial sintering factorThe pore size distributions and N2 adsorption-desorption crystallite forms were converted to equivalent particle size ac-isotherms of the samples have been shown in Figure 4. Ascording to the following equation:it can be seen, all the samples show a mesoporous structure6000with different pore size distributions. The obtained results,DBET =ρSBETshown in Figure 4(b), indicate that the sol-gel method led toa powder with narrower pore size distribution and smaller av-where, DBET (nm) is the average particle size, SBET is theerage pore diameter in comparison with samples obtained byspecifig surface area expressed in m?.g-1 and ρ is the the-other methods. The nitrogen adsorption/desorption isothermsOetical density of cubic alumina expressed in g.cm~ -3. Theshown in Figure 4(a) can be classified as the type IV isotherm, .obtained results show that the particle sizes for different sam-typical of mesoporous materials. According to IUPAC clas-ples prepared with various methods are in the same range andsification, the hysterics loop (type H1) occurs at a relativesmaller in comparison with that of the commercial catalyst.As a factor to reflect the partial sintering extent of the pri-pressure range of p/po = 0.7-0.9, indicating a broad pore sizemary crystallites, ψ can be calculated by the following equa-distribution with uniform size and shape. The pore size distri-bution confirms this assertion and shows a mesoporous struc-tion [19]:ture [19-22].ψ= ( DBET(3)The measured specific surface areas for the samples in( DxRD6000.20(a- - SGO-C500 F.15 F- -- REF-C。400与巨300心。200 t05 -一SGO-C-0-SGA-C- REF-C0.00中国煤化Iwood0.2.0Relative pressure (p/p,)MYHCNMHGFigure 4. (a) N2 adsorption/desorption isotherms and (b) pore size distributions of the alumina samples (BJH plot)Rel ative ,rssure p/po).厂p/mm132Zahra Hosseini et al./ Joumal of Natural Gas Chemitry Vol. 20 No.22011150acid sites, respectively. According to the NH3-THwas calculated to be 4.7 nm according to the Scherrer equa-SGO-C sample showed the highest number of medium acidtion. The measured BET suface area is 303 m2.g- l, whichsites among the prepared catalysts responsible for the selec-is eiyalent to a particle size of 4.9 nm. The crytalite sizestive formation of DME. Comparison of the NH3-TPD patterns) determined via XRD and the particle sizes (DBET) de- of SGO-C and REF-C samples showed two peaks at low andtermined via BET data showed small discrepancies, indicating high temperatures, which correspond to moderate and strongthat the primary crystallites sintered a lttle. Therefore, the ob-acidic sites, respectively. However, the activity of SGO-Ctained results showed that the partial sintering of the particlewas higher than that of REF-C catalyst because the numbersiz2s' H samples prepared using sol-gel method was lower inof weak and medium acid sites on the surface of SGO-C wascomparison with the precipitation method. Since the sinter-higher.ing value of commercial sample is high, it can be said that thecommercial catalyst has been synthesized by a method nearlyclose@ the precipitation method or thermal decomposition of150 E(@beohmite.i 1003.4. Acidity measurements3地NH3-TPD profiles of 7-Al2O3 catalyst samples areshown in Figure 5, and the temperatures of the desorption(bmaxima and the acid content of the catalysts are summarizedin 2:6r 2. In general, NH3-TPD measurements were per-高30formed to determine the strength and amounts of acid sites oncatalyst surface, using ammonia as an adsorbate. Commonly,0ETPD patterns of acidic catalysts have two or three desorp-tion peaks with maxima in the range of 100- -300, 300- 500,and 500- -700°C which can be ascribed to the NH3 des-(corbed-rom acid sites with low, medium and high strengths,resBOively [19].The results of NH3-TPD titration showed that the total白15acidity and number of medium acid sites of the catalysts pre-oEpared using non-aqueous sol-gel were higher than those of thecatalysts prepared by other methods. Most of the researchers(dclaDe that weak acid sites or preferably the acid sites withintermediate strength were responsible for the selective for-宫60mation of DME and strong acid sites were responsible for thefornaion of hydrocarbons. AmongSGO-C showed the highest methanol conversion and DMEselectgyity and thus was selected as the most suitable catalyst01002000300400500600700800900for the methanol dehydration process. The NH3-TPD profile .Temperature (C)of this catalyst showed two peaks at temperatures of 207 andFigure 5. NH3-TPD profiles of the prepared and reference catalysts. (a)479吧，which were attributed to the weak and medium/strongSGO-C, (b) SGA-C, (c) PR-C, (d) REF-C_ 80Table 2. Results of NH3-TPD analysis of γ-Al2O3 samples prepared by different methodsSample codeMaximum desorption temperature (°C)Amount of NH3 (mmo/gau)Acidity (mmol/gary sample)SGO-C200.2342.0821.848603A-C210.1550.728410.52671640 PR.C1960.0600.3523390.292 .REF-C210中国煤化工o4201.495MHCNMHG740.020C)0100200300400500600700800900T emperatureaturel。[)Journal of Natural Gas Chemistry Vol. 20 No.2 20111333.5. Morphologyprepared samples with non-aqueous sol-gel method have gooddispersion and the morphology of the crystals is regular. Also,Figure 6 shows the SEM images of the commercial andfor the prepared samples, the particle sizes are smallr. Theseprepared η/-Al2O3 samples after calcination. As shown in Fig-results are in good agreement with the crystallite sizes esti-ure 6, the crystals of the commercial sample reach a clod coremated by XRD and BET data, while particles in REF-C andform, and the morphology of the crystals becomes irregular.PR- C are larger.VWCAITESCANVCGATESCINDgtomnily 12050 Vic HVaEFigure 6. SEM images of the prepared and commercial catalysts. (a) SGO-C, (b) SGA-C, (c) PR-C, (d) REF-C3.6. Activity measurementslyst has much more active acidic sites available than the com-mercial one.The CH3OH conversion, DME selectivity and yieldon different catalysts under the same operating conditionsTable 3. Catalytic activity and measurement accuracy of the catalysts(T = 300°C, p= 1 atm, LHSV= 11.7 h-1) are summarized inSample ConversionDMEMeasurementTable 3. Also, the activity measurements of these samplesondeat different flow rates of methanol have been shown in Fig-selctivity(%) yield (%accuracy82.60.9997ure 7. The obtained results showed that samples preparedSGA-C .77.599.977.40.9993by sol-gel method using a non-aqueous medium have betterPR-C71.7activity in comparison with those prepared by precipitationREF-C82.399.882.20.99994method. Also SGO-C sample is the best catalyst with theT = 300 °C, p= 1 atm, LHSV=11.7h-1highest conversion. According to Figure 7, the comparison ofcatalysts prepared with different methods and commercial cat-alyst showed that under conditions close to equilbrium, com-0_C catalvst was evaluated under op-mercial and SGO-C catalysts have approximately the same erating conditions中国煤化工of 2.8h-1) forperformance, but with increasing the feed flow rate where the120 h of time-on-YHc N M H Gatalytic activityreaction is going far from the equilibrium condition, SGO-Cremained practicalvuui a uivavivauon of about 1%catalyst has better performance. This is because SGO-C cata-in methanol conversion.SEMH.15. 0kw VD:6. 5298 nm 500 rimSsEM:B.0wV帅:5. 977m silmDate0ra们:12/0/09 Wac : H vDaterdN:12/05/09 we:n,x融z1固134Zahra Hosseini et alL./ Journal of Natural Gas Chemistry Vol. 20 No, 22011References9( 90口LHSV=2.885四LHSV=11.7[1] Semelsberger T A, Borup R L, Greene H L. J Power Sources,区LHSV - 26.12006, 156: 497[2] Vishwanathan V, Jun K W, Kim J W, Roh H s. Appl Catal A,2004, 276: 251g 75[3] JunK w, Lee H s, Roh H s, Park s E. Bull Korean Chem Soc,2003, 24: 106[4] JunK W, Lee H s, Roh H s, Park S E. Bull Korean Chem Soc,2002, 23: 8038!三655] Takeguchi T, Yanagisawa K, Inui T, Inoue M. Appl Catal A,2000, 192: 201[6]FeiJH,HouZY,ZhuB,LouH,ZhengXM.ApplCatalA,SGO-CSGA-CPR-C2006, 304: 49Type of catalyst7] Yaripour F, Baghaei F, Schmidt I, Perregaard J. Catal Commun,2005, 6: 147Figure 7. Variation of the methanol conversion over different catalysts atdifferent LHSV[8] Raoof F, Taghizadeh M, Eliassi A, Yaripour F. Fuel, 2008, 87:2967[9]KimSM,LeeYJ,BaeJW,PotdarHS,JunKW.ApplCatal&0conclusionsA, 2008, 348: 113[10] Natarajan s, Thomas J M. Catal Today, 1992, 12: 433In this study, a nanocrystalline 7-Al2O3 catalyst was pre-[1] Akia M, AlaviS M, Rezaei M, Yan Z F. Microporous Meso-porous Mater, 2009, 122: 72pared by different methods (sol-gel and precipitation). These[12] Khom in J, Praserthdam P, Panpranot J, Mekasuwandumrong O.Catal Commun, 2008, 9: 1955hydration of methanol. The obtained results showed that the [13] Sanchez- Valente J, Bokhimi x, Toledo J A. Appl CatalA, 2004,catalysts prepared using the sol-gel method in non-aqueous264: 175medium have better performance compared to those prepared [14] Zhang Z R, Hicks R W, Pauly T R, Pinnavaia T J. J Am Chem75other methods. According to the catalytic evaluation andSoc, 2002, 124: 1592CRaracterization results, it was found that compared to tradi- [15] Aguado J, Escola J M, Castro M C, Paredes B. Microporoustional catalysts, SGO-C sample showed a good stability overMesoporous Mater, 2005, 83: 181120 h time on stream. Also, the results of experiments showed[16] Wu Y Q, Zhang Y F, Huang X X, GuoJ K. Ceram Int, 2001, 27:265that the activity of catalysts increased in the following order:non-aqueous slgel > aqueous so-gel > peiatio. The [17] Lee HC, Kim HJ, Rhee CH, Lce KH。Lce J s. Chung SH.Microporous Mesoporous Mater, 2005, 79: 61obtained pore radius and surface area for SGO-C were 3.7 nm18] Yoldas B E. Am Ceram Soc Bull, 1975, 54: 286and 303 m2/g, respectively. NH3-TPD analysis revealed that[19] Rezaei M, Alavi S M, Sahebdelfar S, Yan Z F. Powder Technol,. the sample with smaller crystallite size possessed higher con-7Otration of medium acidic sites and consequently the higher[20] Arena F, Dario R, Parmaliana A. Appl Catal A, 1998, 170: 127catalytic activity.[21] Condon J B. Surface Area and Porosity Determinations by Ph-ysisorption Measurements and Theory. The Netherlands: Else-Acknowledgementsvier, 2006This work was financially supported by the lranian Nanotech- [22] Ruthven D M. Principle of Adsorption and Adsorption Pro-nology Initiative Council.cesses. New York: John wiley & sons, 198460中国煤化工SGO- -CSGA- CPR-.MYHCNMHG,REF- CTypeofcatal yst