Study on the deactivation and regeneration of the ZSM-5 catalyst used in methanol to olefins

Available online at www. sciencedirect.comURNALOFScienceDirectVATURALGASCHEMISTRYEL SEVIERJournal of Natural Gas Chemistry 20(2011)266 270www.clsevier.com locatcjngcStudy on the deactivation and regeneration of the ZSM-5catalyst used in methanol to olefinsJingchang Zhang1'2*，Haibin Zhangl，Xiuying Yang，Zhong Huang'，Weiliang Caol,21. Institute of Modern Catalysis, State Key Laboralory of Chemical Resource Engineering. Beijing University of Chen! Technology,Bejing 100292, China; 2. Hainan Insinute of Science and 1Technology, Haikou 571126, Hainan, China;3. Shandong New Time Pharmaceutical Co, Ltd, Jining 272400, Shandong, China[Manuscript received December 8, 2010; revised March 1, 2011 ]AbstractZSM-5 zeolite catalyst modified by a trace of metal cations shows high activity and high sletivity for the reaction of methanol to olefins(MTO), but it incines to deactivate during the reactin. In this paper, the mechanism of the catalyst deactivation and the regeneration methodwere studied by X-ray diffraction (XRD), N2 adsorption-desorption, infrared spectra (IR), and infrared spectra coupled with NH3 molecularprobes (IR-NH3). These characterizations indicated that coke formation was the main reason for the catalyst deactivation. To regenerate thedeactivated catalyst, two methods, i.e, calcination and methanol leaching, were used. N2 adsorption-desorption, IR and IR-NH3 characteriza-tions showed that both methods can eliminate coke deposited on the catalyst and make the catalyst reactivated. XRD showed that the structureof the catalyst did not change after regeneration. Interestingly, the regenerated catalyst even showed better catalytic performance of the MTOreaction than the fresh one. Besides, the calcination regeneration can eliminate coke more completely, however, the methanol leaching methodcan be more easily carried out in situ in the reactor.Key wordsmethanol; olefins; ZSM-5; deactivation; regeneration1. Introductionmostly focused on micro-pore and meso-pore zeolite materi-als. ZSM-5 zeolite is a very promising candidate as the cata-Light olefins, especially ethylene and propylene, are im-lyst for MTO reaction due to its inherent advantages, for ex-portant organic raw materials in the petrochemical industry,ample, a 10-ring interconnected channel system with a highand thus the demand for light olefins is rapidly increasingSi/AI ratio (resuting in is special shape-seletivity), high cat-around the world. The traditional route for the production ofalytic activity, high thermal stability, strong resistance to cokelight olefins is naphtha cracking process, which strongly de-deposition, and yet with proper lipophilicity and hydropho-pends on the oil resource. With the reduction of oil reservesbicity [4.,5]. Therefore, many investigations have been un-and the rise of oil price [1,2], it is exigently desirable to finddertaken to study the MTO reaction using ZSM-5 zeolite asalternative raw materials to reduce the stress from the scarcitythe catalyst [6]. Zhang and his coworkers [7,8] have patentedof light olefins. It is well recognized that methanol to olefinsZSM-5 zeolite catalysts modified with a trace of metal cations(MTO) process is a promising process for the production ofwhich showed high activity and selectivity for the MTO reac-light olefins [3]. Methanol used in MTO as a raw material cantion. However, these catalysts still suffered deactivation afterbe synthesized in large-scale from synthesis gas (a mixture ofa period of time, so the catalyst longevity can not meet thecarbon monoxide and hydrogen) which can be produced fromrequirement of industrial application. Wolf and coworkers [9]the gasification of either coal or bio-materials.reported that the coke deposited on the catalyst consists of sin-The key issue of MTO process is to develop a cata-gle ring and polycyclic aromatic hydrocarbons, depending onlyst with high activity, high selectivity and long longevity.the nature of the catalyst and the reaction system as well asThe catalytic materials applied in MTO reaction have beenthe reaction conditions. Aguayo et al. [10] and Campelo et al.中国煤化工. Corresponding author. TelFax: +86 10-64434904; E-mail: zhangjc1 @mail.buct.edHC N M H Ge Naional High TthnoloThe work was supported by the Research Fund from the China Petroleum & ChemiqResearch and Development Program of Hainan under Grant No.509013.Copyright@2011, Dalian Institute of Chemical Physics, Chinese Academy of Sciences. All rights reserved.doi: 10.1016/S1003 9953(10)60183-1Journal of Narural Gas Chemisty Vol. 20 No.32011267[11] suggested that the main components of coke depositedand atmospheric pressure, with a liquid hourly space veloc-on the zeolite are wax, as well as polycyclic aromatic hydro- ity (LHSV) of 2h 1. The tailgas was sampled and analyzedcarbons. Chen et al. [12] considered that the coke formation hourly by GC (Agilent 6820N gas chromatograph) until themainly depends on the intermediate species adsorbed on thereaction was stable. Then, the tailgas was sampled at the out-catalyst surface.let of the reactor by means of a six-port valve. The analysisThis work is aimned to study the deactivation mechanism of the gas samples was conducted on an Agilent 6820N Gasof MTO reaction over ZSM-5 zeolite catalyst and developChromatograph with a FID (flame ionization detector) and aefetive methods for catalyst rgeneraion. We understand cpillary colum (GS-Alumina 50 mx530um). The flow ratethat coke formation is the main reason for the deactivation ofof the carrier gas (N2) was 5 cm/min.ZSM-5 catalyst. In addition, both calcination and methanolThe selectivity to C1-C4 of the catalytic reactions is cal-leaching can eliminate coke deposited on the catalyst andculated based on the following formula:make the catalyst regenerated. Strikingly, the catalytic per-Aif;C;/M;formance of MTO reaction over the regenerated catalyst wasEAifiCq/M:better than that over the fresh one. We also concluded that thewhere, W; is the percentage composition of C1-C4 in gascalcination regeneration can eliminate the coke deposited onphase (C basis); Ai is the peak area of i component; fi is thethe catalyst more completely, while the methanol leaching re-relative mass correction factor of i component (CH4: 0.97;generation can be more easily carried out in situ in the MTOC2H6: 0.97; C2H4: 1.02; C3Hg: 0.98; C3H6: 0.99; C4HIo:reactor.1.09; C4Hg: 1.01); Ci is the carbon number of i component;Mi is the molecular weight of i component.2. ExperimentalLiquid samples which included the produced water in thereaction and the unreacted methanol were separated into oil2.1. Catalyst preparationphase and water phase. The former was analyzed off-lineusing Agilent 6820N Gas Chromatograph, and the later wasThe multi-component catalysts were prepared by an im-analyzed through 1490 Gas Chromatograph. The conversionmersion method according to the following procedure. Thex (%) of the reaction was obtained by the following formula:raw material of ZSM-5 with a Si/Al ratio of 80 in Na cationx(%)= .Total pumped methanol - unreacted methanolstate was provided by the catalyst plant of Nankai University.Total pumped methanol1.5 g of ZSM-5 zeolite was processed in vacuum for 2 h toremove the water and gas adsorbed on the surface of ZSM-5.x100 .The aqueous solution of Na2CO3, LiNO3, Mn(NO3)2, andFe(NO3)3 were prepared acoring to the mass percentage of 2.3. Catalysts regenerated by calcinationactive component of Na, Li, Mn, and Fe in the carrier. Then,the solution was added into ZSM-5 in vacuum. The mixtureThe fresh catalyst deactivated after running the MTO re-was strred overmight, and then dried and calcined at 550 °Caction for 14 d. The deactivated ZSM-5 catalyst denoted asfor 3 h to obtain the required catalysts.All the catalysts used in this work consisted of the sameZSM-5-De was calcined at 150°C for 1 h, then heated toamounts of Na (150 ppm), Li (150 ppm), Mn (150 ppm) and300 °C for 1 h, and finally heated to 500 °C for2h. The re-Fe (150 ppm) measured by inductively coupled plasma (ICP).sulted catalyst was denoted as ZSM-5-De-Cal.2.4. Catalysts regenerated by methanol leaching2.2. MTO reactionThe deactivated ZSM-5 catalyst was calcined at 110°CA continuous flowing fixed-bed reactor with a cylindri-in situ under a flow of N2 stream for 1h. Then, 2mL ofcal stainless steel tube with an inner diameter of 5 mm andmethanol was injected into the reactor for 2 h at room tem-a length of 300 mm was used for catalyst evaluation, and itperature. Finally, the methanol was removed by flowing N2was loaded with 0.70g of the catalyst (Tyler 80- 100 mesh).stream. The process was repeated for 3 times and the resultedBoth ends of the reactor were flld with quartz grain (Tylercatalyst was denoted as ZSM-5-De-ML.40-70 mesh), and the catalyst bed was located at the con-stant temperature zone of the heating furmace. The furmace2.5. Catalyst characterizationwas a stainless steel chamber heated by an electric resistance,with a Cr-AI thermocouple placed at the center (axially) ofthe catalyst bed. A pump was used to feed methanol into中国煤化工sis was crried outthe reactor, and N2 stream was used to dilute the methanol using;MYHCNMH3id Cu Ka radiationeed. The flow rate of N2 was measured with a mass flow (入=IScity of 0.08%s and ameter (Brooks 5850). Before the catalytic reaction, the cata-scanning range of 50 - 709. The pore structure was determinedlyst was pretreated in situ undera N2 flow at 200°C for 1h. by the adsorption-desorption of N2 on an AUTOSORB-1.The reaction was carried out under the temperature of 370°CThe nature of the active sites, Bronsted type (proton donor)268Jingchang Zhang et al./Jourmal of Natural Gas Chemistry Vol. 20No. 32011or Lewis type (coordinate base), was determined by diffusealyst. Compared with the catalyst regenerated with methanolreflectance infrared (IR) of adsorbed ammonia. The spec-leaching, the catalyst regenerated with calcinations displayedtrometer used is a Prostige-21 IR spectrophotometer with a higher activity and higher selectivity of light olefins, implyingwave-number range of 400- 4000 cm- I.that the calcination regeneration could eliminate coke moreeffectively than the methanol leaching regeneration.3. Results and discussion团C;+C; yeild3.1. Results of MTO reaction8 Methanol conversion100心Total olefins yeild_Figure 1 and Table 1 show the yields of light olefins andothe conversion of methanol over various ZSM-5 zeolite cat-alysts. Methanol was completely converted into hydrocar-bons, and no CO and/or CO2 could be detected in the prod-ucts over all the tested catalysts. It is interesting to notice that40the fresh ZSM-5 catalyst deactivated obviously and the yieldsof light olefins decreased from 81.7% to 62.3% after running2(the MTO reaction for 14 d, however, the deactivated catalystsregenerated with either calcination or methanol leaching ex-hibited higher activity and higher selectivity of light olefinsZSM-5ZSM-5-De ZSM-5-De-Cal ZSN-S_De-MLthan the fresh one. The results showed that both regenerationFigure 1. Light olefins yields and methanol conversion of MTO over freshmethods can effectively remove the coke deposited on the cat-ZSM-5, ZSM-5-De, 2SM-5-De-Cal and ZSM-5-De-ML catalystsTable 1. Product distribution and methanol conversion of MTO over fresh ZSM-5, ZSM-5-De, ZSM-5-De-Cal and ZSM-5-De-ML catalystsHydrocarbon selectivity (C basis%)Catalystsconv. (%)CH4C2HCjHsC3H6C4Hc. C4Hg0:02C +C;C~CFresh ZSM-595.50.8 0.133.21.728.612.719.861.8ZSM-5-De*91.126.8 6.029.50.225.44.:7.454.962.3ZSM-5_De-Cal96.50.5 0.234.3.811.116.867.584.4ZSM-5-De-ML6.14.0.726.9.s37.9_10.864.881.6* The fresh ZSM-5 catalyst after running the MTO reaction for 14 d; Carbon balance was 100%3.2. Results ofXRDalmost no framework change and damage on ZSM-5 zeoliteoccurred in the catalyst regeneration processes.Figure 2 shows the XRD patterns of the fresh and deacti-vated catalysts, as well as the catalysts regenerated by calci-3.3. Results of IR spectranation and by methanol leaching. As shown in Figure 2, allcatalysts still reserved the characteristic peaks of ZSM-5, andIR spectra recorded on the fresh and deactivated cata-lysts, as well as the catalysts regenerated with calcinations andmethanol leaching, are shown in Figure 3. The absorptionZSM-S-De-MLslmmZSM-5-DefCalZSM-5-DC_-ZSM-5中国煤化工1015202530 3545 50MHCNMHG15001000 50Wavenumber (cm ")Figure 2. XRD pttemns of fresh ZSM-5, ZSM-5-De, ZSM-5-De-Cal andFigure 3. FT-IR spectra of fresh ZSM-5, ZSM- 5-De, ZSM-5-De-Cal andZSM-5-De-ML catalystsJourmal of Naural Gas Chemistry Vol 20 No.3 2011269peak at 3500cm~-1 can be assigned to hydroxyl group ofZSM-5 [13], the peak at 1630cm~ ! can be atributed to H-AI Later We will show that the regeneration with calcinations canbond stretching vibration, the peak at 1 100 cm~ 1 is atibutableeliminate coke more thoroughly than the regeneration withto Si-O -Si anti-symmetric stretching vibration, the peak atmethanol leaching.790cm~ 1 is for Si-O Si symmetric stretching vibration, andthe peak at 450 cm~ 1 is for Si-0 Si bending vibration as sug-3.5. Results of N2 adsorption-desorptiongested in Ref. [14]. From Figure 3, it is obvious that only theIR spectrum recorded from the deactivated catalyst exhibits aTable 2 lists the parameters of textural properties such aspeak at 2933 cm-', which can be atributed to the C-H bondthe average pore size, total pore volume, and surface area ofstretching vibration. The IR measurement results demonstratethe fresh and deactivated catalysts, as well as the catalysts re-that carbonaceous deposits, which are formed during the MTOgenerated by calcination and methanol leaching. The averagereaction, can only be detected on the deactivated catalyst.pore diameter and the total pore volume of the regeneratedcatalyst have basically recovered to the level of fresh cata-3.4. Results of IR-NH3lyst, indicating that these two kinds of regeneration methodsare effective in eliminating coke deposition on the catalyst,Figure 4 shows the IR-NH3 spectra of the fresh and deac-and regeneration by calcination approach is better than thattivated catalysts, as well as the catalysts regenerated with cal-by methanol leaching method with the advantage of removingcinations and methanol leaching. As shown in Figure 4, twocoke more thoroughly.wave bands peaks appeared in the range of 1400- 1600 cm-'and 1600- 1800cm'“' with the peaks at 1515cm-1 andTable 2. Pore structure paramneters of fresh ZSM-5, ZSM-5-De,1690 cm-1 , respectively. The result indicates that there wereZSM-5-De-Cal and ZSM-5-De-ML catalyststwo kinds of acid sites, i.e., Bronsted and Lewis acid sites onAverage porePore volume Surface areaCatalyststhe catalyst surface [15]. The band centered at 1515cm-1 ise (nm)(ec/g)(m/g)attributable to the chemical bond of the lone pair electrons ofFresh ZSM-50.490.4131ZSM-5-De0.420.33270nitrogen with Lewis acid sites and while the band centered atZSM-DeCal0.40.38221690 cm~ 1 can be assigned to the chemical bond of the loneZSM-5-De-ML0.4817pair electrons of nitrogen with Bronsted acid sites [16].1004. ConclusionsZSM-5 zeolite catalysts modified by a trace of metal ionsshowed high activity and high selectivity for the MTO reac-ion, however, it inclines to deactivate. Two regeneration ap-三94-proaches, i.e, calcination and methanol leaching, have beeninvestigated in detail. XRD results showed that no frame-92work damage and change of ZSM-5 could be detected af-ter the regeneration. N2 adsorption-desorption, IR and IR-NH3 measurements revealed that the catalyst deactivation wasDeCalmainly caused by the coke generated during the MTO re-” ZSM-5action. The deactivated catalyst can be easily regeneratedby either calcination or methanol leaching. Both of the re-Wavcnumber (cm ")generation approaches can effectively remove the coke de-Figure 4. IR-NH3 specra of fresh ZSM-5, ZSM-S-De, ZSM-5-De.Cal andposits on the catalyst. Compared with the catalyst regeneratedZSM-5-De-ML catalystsby methanol leaching, the catalyst regenerated by calcinationyielded higher activity and higher selectivity of light olefinsFigure 4 can be further used to estimate the total acidi-due to its more complete elimination of the coke depositedties of these four ZSM-5 catalysts, as suggested by Ref. [17].on the catalyst. However, the methanol leaching regenerationFrom Figure 4, it can be seen that the deactivated catalystcan be more easily and safely carried out in situ in the MTOlost a lot of Lewis acid and Bronsted acid sites after runningregeneration.the MTO reaction for 14 d, indicating that coke was heav-中国煤化工ily formed and adsorbed on the acid sites. After the deac-AckndCNMHGtivated catalyst was regenerated by calcination or methanolleaching, the acid sites of the regenerated catalyst were almost Petroleum & Chemical Corporation (Grant No.305025), and therestored to the same level of the fresh one, showing that the re-National High Technology Research and Development Program ofgeneration either with calcinations or with methanol leachingHainan under Grant No.509013.270Jingchang Zhang et aL/ Joumal of Narural Gas Chemisry Vol. 20 No.32011References[10] Aguayo A T, Campo A E s, Gayubo A G, Tarrio A, BilbaoJ. JChem Technol Biotechnol, 199 74: 315[1] Chang C D, Silvestri AJ.J Catal, 1977, 47: 249[11] Campelo J M, Lafont F, Marinas J M, Ojeda M. Appl Catal A,[2] Avidan A A. Sud Surf Sci Catal, 1988, 36: 3072000, 192: 85[3] Marchi A J, Froment G F. Appl Catal, 1991, 71: 139[12] Chen D, Rebo H P, Gronvold A, Moljord K, Holmen A. Micro-[4] Jacobs. US Patent 38945931, 1974porous Mesoporous Mater, 2000, 35-36: 121[5] Rosinski EJ Schwartz A B, Dwyer F G. US Patent 4552648, [13] Dessu R M, Schritt K D, Kerr GT, Woolery G L, AlemanyL1985B.. J Catal, 1987, 104: 484[6] Ke L, FengJ, Zhang M s. Petrochem Technol (Shiyou Huagong), [4] Wang D, Zhang G. Joumnal of Heilongiang Hydraulic Engi-2006, 35: 205neering College, 2007, 34, 74[7] ZhangJC, Zhang H B, TongC M, Cao WL. CN Patent zL [15] She L Q, LiX W. Petrochem Technol (Shiyou Huagong), 200,200910088195.9. 200929: 621[8] Zhang J C, Zhang H B, Cao W L. CN Patent ZL [16] Fan S J, Gao S D. Joumal of Anhui University of Technology200910088197.8. 2009and Science, 2006, 21: 16729] Wolf E E, Alfani F. Catal Rev Sci Eng, 1982, 24: 329[17] Zhang P J. Jourmal of Guangzhou University, 2004, 3: 20中国煤化工MYHCNMHG