Li-li JiHun LiWei ZhngSong SunChen GoJun BoYun-sheng M
a.Department of Chem ical Physics,University of Science and Technology of China,Hefei 230026, China
b.National Synchrotron Radiation Laboratory,University of Science and Technology of China,Hefei 230029,China
Synthesis of Higher Alcohols from Syngas over A lkali Prom oted K-Co-M o Catalysts Supported on M ulti-walled Carbon Nanotubes
Li-li Jia,Huan Lib,Wei Zhangb,Song Sunb,Chen Gaob,Jun Baob?,Yun-sheng Maa?
a.Department of Chem ical Physics,University of Science and Technology of China,Hefei 230026, China
b.National Synchrotron Radiation Laboratory,University of Science and Technology of China,Hefei 230029,China
A series of carbon nanotubes-supported K-Co-M o catalysts were prepared by a sol-gel method combined w ith incipientwetness im pregnation.The catalyst structureswere characterized by X-ray diff raction,N2adsorption-desorption,transm ission electronm icroscopy and H2-TPD,and its catalytic perform ance toward the synthesis of higher alcohols from syngas was investigated.The as-prepared catalyst particles had a low crystallization degree and high dispersion on the outer and inner surface of CNTs.The uniform mesoporous structure of CNTs increased the diff usion rate of reactants and products,thus prom oting the reaction conversion.Furthermore,the incorporation of CNTs support led to a high capability of hydrogen absorption and spillover and promoted the formation of alkyl group,which served as the key intermediate for the alcohol formation and carbon chain grow th.Benefi ting from these characteristics,the CNTs supported M o-based catalyst showed the excellent catalytic performance for the higher alcohols synthesis as com pared to the unsupported catalyst and activated carbon supported catalyst.
CO hydrogenation,Higher alcoholsynthesis,M o-based catalyst,CNTs support
Catalytic conversion of the syngas derived from coal, biomass and natural gas into higher alcohols has attracted significant attention because of the scarcity of energy resources,environmental concerns,and gasoline additive octane dem ands.So far,several catalytic system s have been developed for this reaction during the past few decades[1?3].Among them,the alkaliprom oted M o-based catalysts are regarded as one of them ost prom ising candidates due to the excellent resistance to sulfur poisoning and coke deposition[4].Increasing attempts have been made to im prove the catalytic perform ances of M o-based catalysts.It is found that the 3d transition m etals,especially Co,are found to be effective promoters for alkali-promoted Mo-based catalysts,they are known to enhance alcohol production and im prove C2+alcohol selectivity.The strong interaction between Co and Mo species is conducive to the formation of higher alcohols.
Catalyst supports have a significant im pact on the synthesis of alcohols from syngas.The acid supports such as A l2O3are unfavorable to the synthesis of alcohols because the acidic sites of supportswould causethe alcohol dehydration.Activated carbon has the advantages of large surface area,high thermal stability, resistance to acidic or basic media,and shows higher selectivity to alcohols as com pared to SiO2-,A l2O3-, and CeO2-supported catalysts.The interaction between the activated carbon and Mo species p lays an im portant role in the catalytic performance.Our previous study has investigated the activated carbon supported K-Co-M o catalysts by the synchrotron radiation X-ray absorption fine structure(XAFS).It was found that w ith an increase in the M o loading,the surface M o atom s gradually changed from tetrahedrally coordinated M o6+species to octahedrally coordinated Mo4+, suggesting a higher reduction degree.The activated carbon supported Mo-based catalyst exhibited an excellent perform ance for the higher alcohol synthesis,especially the formation of the C2+OH.
Multi-walled carbon nanotubes(MWCNTs)have a lot of unique characteristics,such as inert graphitic surface,appropriate pore size distribution,good electrical conductivity and enhanced mass transport capability, whichmake it a prom ising support for various catalytic app lications[5,6].Tavasoliet al.[7]found that the CNTs supported Co catalysts showed a higher CO conversion and Fischer-Tropsch synthesis rate than that of the alum ina supported catalyst.The reason was attributed that the CNTs aided the uniform dispersion of Cometal clusters on the support.Sim ilar resultwas re-ported by Tanet al.[8].They reported a highly dispersed CNT-supported copper-cobalt-cerium catalysts. The em p loym ent of CNTs support ensured intim ate contact am ong the metal particles at the nanoscale, which led to a superior selectivity to the ethanol and C2+alcohols.Furthermore,the CNTs can provide the sp2-C surface-sites for adsorption-activation of H2and form a high concentration of H-speciesm icro environment,thereby increasing the hydrogenation reaction rate[9?11].Theuniform poresizedistribution ofCNTs support exhibited an excellent gas perm eability and enabled themetal nanoparticles to be very accessible to the reaction gas,which facilitated the diff usion of the reactants and the dispatched products from the active sites[12].
In our previous work,we have developed a highly homogeneous unsupported K-Co-Mo catalyst prepared by a modified sol-gelmethod[13]to improve the catalytic perform ance for higher alcohol synthesis and understand the support eff ect.Herein,a kind of m esoporous MWCNTswas employed as the support to prepare the K-Co-Mo/CNTs catalysts.The catalyst structureswere characterized by a series of physicochem ical m ethods,and the catalytic perform ance for the synthesisofhigher alcohols from syngaswas investigated.The relationship between the structure and catalytic perform ance was also discussed.
A.Catalyst preparation
The CNTs supported K-Co-M o catalysts(K-Co-Mo/CNTs)wereprepared by a sol-gelmethod combined w ith incipient wetness im pregnation.The MWCNTs w ith inner diameter of 5?10 nm and outer diam eter of 10?20 nm were supp lied by Chengdu Institute of Organic Chem istry.Prior to use,the raw CNTs was treated w ith 30%nitric acid for 24 h,followed by washing w ith deionized water several tim es and then drying at 393 K for 12 h.Finally,the CNTs was flushed w ith pure N2at 453 K for 2 h to remove any surface adsorbents.A typical procedure is as follows:fi rstly, theaqueoussolutionsofCo(NO3)2·6H2O,C6H8O7·H2O and K2CO3were added in sequence dropw ise to the (NH4)6MoO24·4H2O aqueous solution under stirring. The acidity ofm ixed solution was adjusted to pH=3.5 by adding the amm onia or acetic acid.Subsequently, them ixed solution was kept in a water bath at 343 K until the sol was obtained.The as-prepared sol was then im pregnated into CNTs.A fter ultrasonic dispersion for 1 h,them ixture was dried at 393 K overnight and calcined in flowing nitrogen at 673 K for 4 h.The Mo content in the as-prepared catalysts,expressed as the weight ratio M o/CNTs,was ranged from 10%to 50%.The atom ic ratios of K/M o and Co/M o were 0.1 and 0.5,respectively.
B.Catalyst characterization
Powder X-ray diff raction(XRD)patterns were recorded w ith a Rigaku D/m ax-γA rotating-anode diff ractometer equipped w ith a Cu Kαradiation source. The BET surface area,pore volume,and pore diameter were determ ined by nitrogen adsorption at 77 K using a M icrom eritics TriStar II 3020 analyzer.Transm ission electron m icroscopy(TEM)was performed using a Philips CM 20(100 kV)transm ission electron m icroscope equipped w ith a NARON energy-dispersive spectrom eter.
The H2-temperature programmed desorption(H2-TPD)experiment was carried out on the model of FINESSORB-3013 adsorption instrum ent.For each experiment,0.1 g of the samp le was packed into a U-type quartz tube.The samp le was pretreated under 5%H2/A r at 798 K for 3 h.A fter cooling to room temperature,the pretreated sam p le passed through He atmosphere at 473 K for 0.5 h.Then the pretreated samp le was saturated w ith 5%H2/Ar for 1 h.A fter that the sam p lewas flushed w ith heat 373 K for 1 h.Finally the TPD analysiswas carried out in a flow of He from 373 K to 1123 K at a heating rate of 10 K/m in.
C.Catalytic activity measurem ents
The catalytic performance of the catalysts for the synthesis of higher alcohols from syngaswas tested in a fixed-bed stainless steel reactor w ith an inner diameter of 8mm.The reactorwas packed w ith 0.5 g of catalyst that was diluted w ith quartz sand to produce a total volum e of 2m L and loaded at the center of the reactor tube.Prior to reaction,the catalyst was reduced w ith 5%H2/N2for 12 h at a flow rate of 40m L/m in.A fter lowered to the reaction tem perature,the feed gas containingV(H2):V(CO):V(N2)=60:30:10 passed through. The effl uent gaswas cooled in an ice-water bath to separate into gas and liquid phases.Details on the product analytical p rocedure were described in our previouswork[14].A ll the activity measurementswere performed under the reaction condition of 5.0MPa,553 K and gas hourly space velocity(GHSV)2400 h?1.The activity data were collected after the reaction was performed for 24 h because the alcohol synthesis required an induction period.
FIG.1 XRD patterns of(a)unsupported K-Co-M o,(b) pure CNTs,K-Co-Mo/CNTs w ith diff erent M o loading of (c)10 w t%,(d)30 w t%,(e)40 w t%,(f)50 w t%,and (g)K-Co-Mo/AC w ith Mo loading of 30 w t%.
FIG.1 shows the XRD patternsof thepurified CNTs, the unsupported catalyst and the CNTs supported catalysts w ith different Mo loading.The purified CNTs showed two broad peaksat26.1?and 43.2?,respectively. 26.1?corresponding to(002)reflection of graphite and the other sm all asymm etric peak 43.2?is due to(100) reflection of graphite[15].For the unsupported catalyst,on ly three w ide peaks at 26.1?,37.0?,and 53.5?were detected,which was assigned to MoO2.The formation ofMoO2wasattributed to the fact that in nitrogen,the decom position of citric acid in the sol resulted in the partial reduction of M o+6species[16].The supported catalystsexhibited the same diff raction patterns but weaker diff raction intensity as the CNTs support. Besides the CNTs peaks,no peaksassigned to K,Co,or M o specieswere detected for the CNTs supported catalysts.Furthermore,w ith an increaseof theMo contents, the diff raction intensity decreased gradually.The result indicated that the active com ponents had a low degree of crystallization and high dispersion on the surface of CNTs.For comparison,the diff raction pattern of the activated carbon supported catalyst w ith the Mo/AC ratio of 30%was also presented.A lm ost no obvious diff raction peakswere observed on the sam p le.
The TEM images of the sole CNTs and the supported catalysts were shown in FIG.2.The purified CNTs had open endsw ith a uniform diameter of about 20 nm,displaying a mesoporous pore structure.For the CNTs supported catalysts,at a low loading of M o, the catalyst particles were evenly distributed on the CNTs w ithout obvious aggregation.It was noted that somepartsof theparticlesentered the carbon tubesand well dispersed on the inside surface,benefi ting from the sm all particle size and uniform m esoporous structure of CNTs.W hen the Mo/CNTs weight ratio exceeded 30%,the particlesbegan to aggregate,asshown in FIG. 2(d)and(e).FIG.2(f)shows the TEM im age of the activated carbon supported catalyst(M o/AC ratio of 30%).The as-prepared catalyst particlesalso exhibited a high dispersion on the activated carbon support.
FIG.2 TEM patterns of the(a)pure CNTs,K-Co-M o/CNTs w ith diff erent M o loading of(b)10 w t%,(c) 30 w t%,(d)40 w t%,(e)50 w t%;and(f)K-Co-Mo/AC w ith M o loading of 30 w t%.
TABLE I Texture properties of the pure CNTs and supported K-Co-Mo catalysts.
FIG.3 shows the nitrogen adsorption-desorption isotherms of the pure CNTs and supported K-Co-Mo catalyst.The pure CNTs exhibited a type IV isotherm w ith a hysteresis loop of type H1 according to the IUPAC classification,and the capillary condensation occurred at a high relative pressure(P/P0)above 0.80. The result further demonstrated the used CNTs possessed amesoporous structurew ith cylindrical pore geometry and a high degree of pore size uniform ity,as revealed by the TEM images.The CNTs supported K-Co-M o catalysts exhibited sim ilar isotherm s to that of the purified CNTs,indicating that metal im pregnation did not alter the pore structure of the parent support.W ith an increase of the M o loading,the BET surface and pore volum e(V)of the supported catalysts decreased while the average pore size did not show significant change as shown in Table I.For com parison, the nitrogen adsorption-desorption isotherm s of the activated carbon supported K-Co-M o catalyst w ith the Mo loading of 30%was also listed(FIG.3(f)).The adsorption-desorption isotherm showed type Ibehavior w ith a H4 type hysteresis loop,indicative of the existence of narrow slit-like pores.Furtherm ore,a steep increase of adsorbed volume at very low relative pressure,corresponding tom icropore volume fi lling,wasobserved.The result indicated that the activated carbon supported catalyst contained a certain amount ofm icropores.Consequently,the catalyst exhibited a much larger BET surface area but smaller pore size as compared to the CNTs supported catalysts w ith the same com position.
FIG.3 Nitrogen adsorption-desorption isotherm s of(a)pure CNTs,K-Co-M o/CNTs w ith diff erent M o loading of (b)10 w t%,(c)30 w t%,(d)40 w t%,(e)50 w t%,and(f)K-Co-M o/AC w ith Mo loading of 30 w t%.
FIG.4 H2-TPD of the K-Co-M o/CNTs w ith diff erent M o loading of(a)10 w t%,(b)30 w t%,(c)40 w t%,(d)50 w t% and(e)K-Co-M o/AC w ith M o loading of 30 w t%.
The hydrogen adsorp tion and desorption ability of the catalysts are investigated by H2-TPD experiments and the profi leswere shown in FIG.4.The CNTs supported catalysts exhibited a predom inant peak of hydrogen desorption at 973 K.Besides,a very weak w ide peak appeared around 740K.Theweak desorption peak wasmost probably due to themolecularly adsorbed hydrogen.The strong peak was attributed to the dissociatively chem isorbed hydrogen,which was suggested to have a significant im pact on the catalytic activity. The capacity of adsorbing hydrogen increased w ith an increaseof theMo loading and reached thehighest level at the M o loading of 30%.W ith a further increase of theMo loading,theadsorption capacity ofhydrogen did not show significant change.The H2-TPD profi le of the activated carbon supported K-Co-M o catalyst(M o/AC ratio of 30%)was also listed.The sam p le exhibited a stronger low tem perature desorption peak than that of the CNTs catalysts,which may be due to the fact that the presence of the m icropores in the activated carbon was conducive to the absorption of molecular hydrogen.It was noted that its peak corresponding to dissociatively chem isorbed hydrogen shifted to higher tem perature as com pared to the CNTs catalyst,indicative of a stronger interaction between the chem isorbed H-species and activated carbon support.Furthermore, a shoulder peak was also observed,which may be attributed that the non-uniform ly distributed pore size of activated carbon resulted in a diff erent diff usion rate of H-species.
Table II lists the catalytic performance of the K-Co-M o/CNTs catalysts for the synthesis of higher alcohols.For comparison,the unsupported catalyst was also tested under the same conditions.The unsupported K-Co-M o exhibited a relative low activity toward alcohol synthesis,and the predom inant alcohol product was methanol.The incorporation of CNTs support led to a significant increase in the alcohol production.As shown in Table II,the catalyst w ith a M o/CNTsweight ratio of30%showed thebest catalytic performance for alcohols synthesis.The STY of total alcoholswas 141.7 g·kg?1·h?1,about 9 times as high as that of the unsupported sam p le and the alcohol selectivity increased from 9.5%to 35.4%.In particular,the methanol production was inhibited remarkably.The effect of the H2/CO ratio on the catalyst activity wasalso investigated.W hen the H2/CO ratio decreased from 2 to 1,the catalystw ith a Mo/CNTsweight ratio of 30% alcohol selectivity and C2+OH/MeOH ratio further increased up to 49.8%and 2.33,respectively.The ethanol became the predom inant alcohol product.M eanwhile, the alcohol STY still reached 110.0 g·kg?1·h?1.Thepresented activity data are encouraging because they were tested under very m ild conditions of 5.0MPa and 2400 h?1.The decrease of H2/CO ratio im proved the alcohol selectivity,especially the formation of C2+OH. The reason was attributed that increasing the CO concentration decreased the relative rate of hydrogenation of them ethanol precursor and thus inhibited the synthesis of m ethanol.The form ation of higher alcohols appeared to be a slow step relative to the rate of hydrogenation ofmethanolprecursor[17].In addition,increasing the CO concentration also favored the insertion ofCO into an alkylgroup to form theacylspecies,which was regarded as the key intermediate for the synthesis of alcohols[18].The CO2was yielded from the watergas-shift(WGS)side reaction.No significant change in CO2production was observed when the Mo/CNTs weight ratio increased from 30%to 50%.For com parison,the catalytic perform ance of the activated carbon supported K-Co-M o catalyst(M o loading of 30%)was tested under the same conditions.In com parison to the activated carbon sam ple,thealcoholselectivity over the CNTs catalyst showed a slight decrease,while the alcohol STY increased m ore than 60%.In particular,the C2+OH/MeOH ratio also increased from 2.08 to 2.33. The results indicated the CNTs supported K-Co-Mo catalyst exhibited better catalytic perform ance for the higher alcohol synthesis,although its BET surface area wasmuch lower than that of the activated carbon samp le.
TABLE II Catalytic performance toward alcohol formation from syngas over the catalysts.A lc.Sel.is calculated on a CO2-free basis,Reaction conditions are 553 K,5.0 M Pa,2400 h?1.CO conv.,A lc.Sel.,and CnOH Sel.in C-m ol%,A lc. STY in g·kg?1·h?1
The support properties,such as acidity,components dispersion,and pore structureetc.,have a significant im pact on the catalytic perform ance for alcohols synthesis from syngas.The CNTs and activated carbon are the neutral carbon materials.The as-prepared KCo-M o catalysts supported on the CNTs and activated carbon exhibited a low crystallization degree and high dispersion,as revealed by the XRD and TEM results. This suggested that the supported catalysts had a high active surface area,which contributed to higher activity com pared to the unsupported catalyst.
The pore structure of the support significantly affected the catalytic activity for CO hydrogenation, especially the product distribution.Small pore size w ill result in poor intra-pellet diffusion effi ciencies of molecules.Slow transport of reactants to,and products from,active sites often inhibits the reaction rate and aff ects the product selectivity.For the higher alcohol synthesis from syngas,Surisettyet al.have investigated the influence of porous characteristics of support and concluded that the m esoporous support was conducive to the alcohol production,especially the formation of C2+OH[19].The hydrogen dissociation and spillover capability also p lay an im portant role in the CO hydrogenation reaction.The formation of higher alcohols from syngas over the Mo-based catalyst followsa CO insertionmechanism[20].Thedissociated H-species reactsw ith the dissociated CO to form the alkyl group.The non-dissociatively CO insertion to the alkyl group form s thealcoholproduct.Thealkylgroup serves as the key interm ediate for alcohol form ation and carbon chain grow th.The nitrogen adsorption-desorption isotherm sand TEM images showed that the CNTs supported catalysts possessed a uniform mesoporous structure.This meant that catalyst had a higher diff usion effi ciencies of m olecule and increased reaction rate as com pared to the activated carbon support containing m icropores.Furthermore,the H2-TPD result revealed that the CNTs catalyst had a strong adsorption capacity of hydrogen and relatively weaker interaction w ith the H-species com pared to the activated carbon catalyst.The reason may be attributed that the surface sp2-C sites of CNTs promoted the adsorption and activation of hydrogen[21].The weak interaction between the H-species and catalyst indicated a higher spillover capability of H-species,which was conducive to the formation of alkyl group interm ediate and thus prom oted thealcoholproduction.These factorswere suggested to be responsible for theexcellent catalytic performance of the CNTs supported K-Co-M o catalysts for the higher alcohol synthesis from syngas.
M ulti-walled carbon nanotubes supported K-Co-M o catalyst was prepared by a sol-gelmethod.The K-Co-Mo/CNTs catalyst showed a much higher activity for the synthesis of higher alcohols than that of the unsupported catalyst and activated carbon supported catalyst.The enhanced perform ance can be attributed to the fact that the CNTs supported catalyst had a low crystallization degree and high dispersion,which suggested a high active surface area.The uniform m esoporous structure of CNTs support led to a high diffusion effi ciencies ofmolecule and increased conversion rate.Furthermore,the CNTs catalyst had a strong capability ofhydrogen absorption and spillover,which was conducive to the formation of alkyl group intermediate and thus enhanced the alcohol production.
Thiswork wassupported by NationalNaturalScience Foundation of China(No.21673214).
[1]J.Bao,Y.L.Fu,Z.H.Sun,and C.Gao,Chem.Commun.746(2003).
[2]V.Mahdavi,M.H.Peyrovi,M.Islam i,and J.Y.Mehr, App l.Catal.A 281,259(2005).
[3]R.Burch and M.J.Hayes,J.Catal.165,249(1997).
[4]S.Zam an and K.J.Sm ith,Catal.Rev.54,41(2012).
[5]P.Serp,M.Corrias,and P.Kalck,App l.Catal.A 253, 337(2003).
[6]X.L.Pan,Z.L.Fan,W.Chen,Y.J.Ding,H.Y.Luo, and X.H.Bao,Nat.Mater.6,507(2007).
[7]A.Tavasoli,R.M.M.Abbaslou,M.Trepanier,and A. K.Dalai,App l.Catal.A 345,134(2008).
[8]P.Wang,J.F.Zhang,Y.X.Bai,H.X iao,S.P.Tian, H.J.Xie,G.H.Yang,N.Tsubaki,Y.Z.Han,and Y. S.Tan,App l.Catal.A 514,14(2016).
[9]X.M.Wu,Y.Y.Guo,J.M.Zhou,G.D.Lin,X.Dong, and H.B.Zhang,App l.Catal.A 340,87(2008).
[10]C.H.M a,H.Y.Li,G.D.Lin,and H.B.Zhang,App l. Catal.B Environ.100,245(2010).
[11]W.Feng,Q.W.Wang,B.Jiang,and P.J.Ji,Ind.Eng. Chem.Res.50,11067(2011).
[12]M.Khavarian,S.P.Chai,and A.R.M oham ed,Fuel 158,129(2015).
[13]J.Bao,Y.L.Fu,and G.Z.Bian,Catal.Lett.121,151 (2008).
[14]M.M.Lv,W.Xie,S.Sun,G.M.Wu,L.R.Zheng, S.Q.Chu,C.Gao,and J.Bao,Catal.Sci.Technol.5, 2925(2015).
[15]I.Eswaram oorthi,V.Sundaramurthy,and A.K.Dalai, App l.Catal.A 313,22(2006).
[16]J.Bao,Z.H.Sun,Y.L.Fu,G.Z.Bian,Y.Zhang,and N.Tsubaki,Top Catal.52,789(2009).
[17]E.Tronconi,N.Ferlazzo,P.Forzatti,and I.Pasquon, Ind.Eng.Chem.Res.26,2122(1987).
[18]A.M uram atsu,T.Tatsum i,and H.Tom inaga,J.Phys. Chem.96,1334(1992).
[19]V.R.Surisetty,A.K.Dalai,and J.Kozinski,App l. Catal.A 393,50(2011).
[20]A.M uram atsu,T.Tatsum i,and H.Tom inaga,J.Phys. Chem.96,1334(1992).
[21]J.J.Wang,J.R.Xie,Y.H.Huang,B.H.Chen,G.D. Lin,and H.B.Zhang,App l.Catal.A 468,44(2013).
ceived on March 30,2017;Accepted on April 30,2017)
?Authors to whom correspondence shou ld be add ressed.E-m ail: bao j@ustc.edu.cn,ysm a@ustc.edu.cn.
CHINESE JOURNAL OF CHEMICAL PHYSICS2017年3期