Advances in the Partial Oxidation of Methane to Synthesis Gas

Journal of Natural Gas Chemistry 13(2004)191-203SCIENCE PRESSAdvances in the Partial Oxidation of Methane to Synthesis GasQuanli Zhu,2, Xutao Zhao, Youquan Deng1. Petrochemical Research Institute of Lanzhou Petrochemical Company, China National Petroleun CorporatLanzhou 790060, China; 2. State Key Laboratory for Oro Symthesis and Selective Oxidation, LanzhouInstitute of Chemical Physics, Chinese Academy of Sciences, Lanzhou 730000, ChinaManuscript received November 10, 2004; revised November 26, 2004Abstract: The conversion and utilization of natural gas is of significant meaning to the national economy,even to the everyday life of people. However, it has not become a popular industrial process as expecteddue to the technical obstacles. In the past decades, much investigation into the conversion of methane,predominant component of natural gas, has been carried out. Among the possible routes of methaneconversion, the partial oxidation of methane to synthesis gas is considered as an effective and economicallyfeasible one. In this article, a brief review of recent studies on the mechanism of the partial oxidation ofmethane to synthesis gas together with catalyst development is wherein presentedKey words: methane partial oxidation, synthesis gas, catalyst, reaction mechanism1. Introductionrate was no more than 30%[2, 3, and in the pro-cess of direct oxidation of methane to methanol 4]orNatural gas, which is. mainly composed of formaldehyde 5], the highest productive rate was 8%methane, is an abundant resource found: over the and 4%, respectively. It was recently reported that aworld and is predicated to outlast the oil reserves 50% of methanol productive rate was achieved by aby a significant margin 1]Most of these reserves,pilot plant in homogeneous catalysis, but still lowerhowever,are situated in the areas far away from the than what is expected (6, 7 ] On the other hand, themarkets of highest energy consumption, and the ex- mercury and concentrated sulfuric acid was used inpensive cost of compression, transportation and stor- this process, and the resulted sulfuric dioxide shouldage,makes the utilization of natural gas as an unpre- to be re-oxidized for recycle. Although there are inpossessing proposition. Contrary to it, petroleum is dustrial processes of the direct conversion of methane,elatively cheap and it can be conveniently disposed. such as the oxidation of methane and ammonium orIn order to make the utilization of natural gas more amine to cyanide [8 and the pyrolysis of methane toeconomically viable, a large amount of investigation acetylene[9, their marked disadvantage is that theseinto the conversion of methane to liquids or higher hy- processes are needed to operate at very high temperadrocarbons has been carried out in the past decades. ture, usually above 1300 K. Because of these reasons,Unfortunately, the productive rate in these processes it is very difficult for natural gas to compete withis still lower than what is expected, because these petroleum at presentproducts resulted from methane partial oxidation areIn order to elevate the additional value ofusually more chemically active than methane, which meth中国煤化工 hane can be theoretlimits methane converting to the expected products. ically-ways: one is the diFor example, in the process of direct oxidative cou- rectCNMHGch as as mentionedpling of methane to ethylene, the highest productive above, the oxidative coupling of methane to ethyCorrespondingauthorE-mail:qlzhu0001@sina.com192Quanli Zhu et al. /Journal of Natural Gas Chemistry Vol. 13 No. 4 2004lene the direct oxidation of methane to methanol depends to a great extent upon the employed catalystor formaldehyde, etc. It is impossible for these pro temperature, pressure and the ratio of methane tocesses to be applied to mass production unless great oxygen in feedstock, as well as kinetic factors. Threebreakthroughs of these technologies are achieved. The reactions possible to occur during POM are brieflyother is the indirect conversion of methane namely, expressed in Figure 2, wherein some thermodynamicconverting methane to any other products via syn- information is includedthesis gas, which is the mainly practical route for themethane conversion at the present time.COz+H2OC2H2Nowadays, there are three methods for produc-ing synthesis gas from methane: the steam reform-H2s/Sing, the dry reforming and the partial oxidation, ofCH]Cl Cl-HCI△O2HCNmethane. Compared with the former two, the partialoxidation of methane(Pom) possesses characteristicsas follows: (1) POM reaction is a mild exothermicetc.aHCHOreaction while the former two are endothermic reac-tions. Thus, the industrial process based upon POMC2H6tCaHACH3OHis energy saving. In view of this, the utilization ofCO+H2POM combined with steam reforming or dry reformFigure 1. Possible route for methane conversioning is more effective. (2)The molar ratio of H2 toCO in the resulted synthesis gas is close to 2 if POMreaction is carried out according to stoichiometric ra-CH4tio. This kind of synthesis gas containing little co 2 is1/202; partial oxidationan ideal feedstock for downstream processes, such asCo+Hmethanol synthesis, etc. ( 3) POM can be carried outTotalunder the condition of very high gas hourly space ve- CH4+202=CO2+2H20locity(GHSV), which makes the process require less (AH298-803 k/mol)(steam and cOCH4+H20-CO+3H2investment and less production scale to achieve thesame or larger capacity.CH4+CO2"2C0-+2H(△H98=+261kmo)However, carbon depositing over catalyst bed wasCHa+ CO,+H?0unavoidable even if POM was carried out precisely Figure 2. Thermodynamic representation of POMunder the condition of 2: 1 of the molar ratio of CHto o2, or less than 2: 1, and the coke formation waseven worse at higher temperature. POM, thus, didThe calculated product gas distribution at thenot get proper attention until the oil crisis. Since thethermodynamic equilibrium under the condition ofeighties'of last century, Green and co-workers 10, 11atmospheric pressure and input methane-to-oxygenhave done much work for the renaissance of studratio of 2: 1 versus temperature, based the reactionsPOM. They used noble metals as POM catalysts, and in Figure 2, is shown in Figure 3[14], regardless ofobtained the synthesis gas with compositions closearbon deposition [14]. It can be observed in Figureto thermodynamic equilibrium. POM is, henceforth,3 that the selectivity to CO and H2 increases withpaid much attention in the catalytic cycle around theincreasing reaction temperature. In fact, very highworldmethane conversion(>90%)and selectivity(>90%)to synthesis gas can be obtained above 1000 K. like2. Brief thermodynamic analysis of methane were done by Lunsford and co-workers l0/. urewise, the equilibrium gas compositions versus presspartial oxidationresults indicated that the partial pressure of CH4o in the equilibrium gas compositions in-Possible pathways via which methane is convertedcre中国煤化工 pressure, which meansare shown in Figure 1. At high temperature, the main thatCNMH Gle to POM to synthesisreaction products between methane and oxygen are, gas.noweLeinperature can compensatehowever,limited to CO, CO2, H2O and H2, [12, 13], this pressure effect. In other words, from the point ofapart from some intermediates. Product distribution view of thermodynamics it is feasible for the reactionJournal of Natural Gas Chemistry VoL. 13 No. 4 2004of methane and oxygen to synthesis gas at increasedPOM reaction to synthesis gas can be carried outpressure to be commercially usedwithout employment of a catalyst, but it occurs atigh temperature, usually above 1400 K in theflame[21]. The employment of catalyst can facilitatethe light-off of POM and promote it to thermody0.35namic equilibrium. The catalysts for POM to synthe-sis gas can be divided into three groups: Ni, Co andFe, noble metal and early transition metal carbide0253.1. Ni, Co and Fe catalyst0.20The earliest work on the catalytic partial oxidation of methane to synthesis gas was performed byLiander [22, Padovani and Franchetti [23] and Prettre et al. [24], who obtained high yields of synthe-0.10sis gas with ca. 2: 1 of H2/Co molar ratio, withinthe temperature range from 1000 to 1200 K, atat0.05mospheric pressure. They proposed that a sequenceof reactions including total oxidation and reformingreactions of methane was taking place over nickel cat-77387397310231073alyst. The calculated equilibrium gas compositionsTemperature(K)based upon those reactions shown in Figure 2 gave aFigure 3. Equilibrium gas compositions formethane partial oxidation, 1 bar, 2: 1good agreement with the observed exit gas composif CH4: O2tions, which implied that the thermodynamic equilib(1)p(H2),(2)p(CH4),(3)p(CO),(4)P(CO2),(5)p(H2O)rium was established in all cases, if carbon depositionwas ignoredIt is generally accepted that POM could be cat-Vermeiren et aL. 25] reported the results of oxidizing methane with air over nickel catalyst, andalyzed to, or almost, to the thermodynamic equilib- drew similar conclusions to Prettre's[24] .They com-rium by group VIII metal catalysts (16, 17]. How- pared the POM activity with methane steam reform-ever, Choudhary and his co-workers [18-20 obtainedthe yields of synthesis gas much higher than thatng over the similar nickel catalyst and found thatcalculated according to the thermodynamic equilib-POM was 13 times faster than the latter. Thus, theypresumed that there were extra reaction pathwaysrium,over the metal oxide-supported nickel or cobalt, which greatly accelerated the methane converting inwithin the temperature range from 723 to 773 Kunder very short residence time. They also investi-methane/oxygen mixtureLunsford and co-workers [15 studied in detagated the steam reforming and the dry reforming un- the alumina-supported nickel catalyst bed exposed toder the same condition, and it was found that theesult was higher than that predicted by thermodyPOM atmosphere using XRD and xPS techniquesand found that three zones were formed in catalystnamics. Based upon these data, they regarded that bed. The outer zone was made up of nial2O4 phasesPOM reaction occurred under non-equilibrium con-dition,and the mechanism was different from that which was moderately active for total oxidation ofmethane The mid zone included NiO and al2 O3 palunder equilibrium condition [10, 11]. Green and co- ticles, which was thought to complete the total oxi-workers [14]studied the phenomenon described by dation of methane. The inner zone contained metalature' of catalyst bed increased with gas flow rate. lic nickel particles, it was suggested that at this zoneTherefore, this result was consistent with the predication by thermodynamics if the real reaction tem- nan中国煤化工 lies indicated that onlyperature was taken into accountYH-re formed, without for-CNMHGover the catalyst sur-face below 973 K. At 1023 K, the surface carbon de-3. Catalysts for PoM to Synthesis Gasposition increased to monolayer companied with the194Quanli Zhu et al. Journal of Natural Gas Chemistry VoL. 13 No, 4 2004higher methane conversion. Their results also indi- that nickel supported on the perovskite-sturcturedcated that the amount of surface carbon deposition materials, e. g. Ni/Cao. 8 Sro.2TiO3, prepared by citwas affected by the ratio of methane to oxygen in rate method, exhibited good ability to restrain itselffeedstock, a higher ratio resulting in more carbon de- from carbon depositing. Negligible amount of carbonposition, and vice versa.deposition was found over its surface after 150 hoursNickel is active component for POM, but the run [36-40]. It was claimed that this kind of sup-nickel species with different oxidative state playsport could control the size of dispersed active phadifferent role in surface reaction steps, as mentioned below the threshold value needed to generate carabove. In general, metallic nickel is beneficial to bon deposition. Therefore, oxygen species over catthe production of synthesis gas, while nickel species alyst surface can react with carbon deposition, whichwith oxidative number >2 trends to catalyze the to- leads to nickel particles prevent from being coveredtal combustion of methane. The distribution of sur- [41]. Another kind of active support is hydrotalcite-face nickel species with different oxidative number structured mixed oxide. It is also claimed that thisdepends upon the support properties and synthesis support can control the dispersed nickel particles sizeprocedure. When referring to nickel-based catalysts, [42, 43]. However, the activity and selectivity for thisits activity and stability are indispensable topics to kind of catalyst depended to a great extent upon thebe touched. In order to enhance the POM activity reducibility, concentration of nickel oxide and resi-and stability of nickel-based catalyst, one approach is dence time of reactants. The reducibility of catalystto choose suitable supports from miscellaneous ma- relates to the properties of precursor, such as compo-terials. Choudhary and co-workers [18-20] carefully sition and pretreatment etc. High Mg/Al ratio leadsstudied catalysts of nickel supported on Yb2 O3, Mgo, to less formation of spinel-type species, which is lessCao, TiO2, ZrO2, ThO2 and UO2, as well as alumina active for POM [ 43]doped with rare earth oxides. It was found that theSome other materials, for example, CaAl2Ocatalyst containing CaO, MgO, rare earth oxide and AlPO4-5 and calcium phosphate/hydroxyapatite, etc.alumina had higher activity under the condition of 33, 44, 45, were tried to be used as the support ofshort residence time. For nickel catalyst containing POM catalysts. CaAl2O4 and calcium phosphate sup-ThO2, UO2 and ZrO2, the activity had the order as ported nickel catalysts showed excellent performancefollows,NiO/ThO2>NiO/UO2> NiO/ZrO2. TiO2 of sintering and carbon depositing resistance, andor SiO2, as support, was not suitable for POM to syn- therefore they showed higher methane conversion andthesis gas due to the easier sintering of nickel oxide selectivity to H2. AlPO4-5 also gave good catalyticand the inertness of binary metal oxide at high tem- performance; however, the phase transformation teRuckenstein et al. [26-28] carefully studied of catalyst and activity to be lost quicky ecific areaperaturetridymite-structured species caused the spNiO/Mg catalyst systemwas found that theAnother approach to elevate activity and stabilityhigh activity originated from the formation of solid is the modifying support. The employment of alka-solution, nickel atoms evenly dissolved in the crystal line earth oxide [35, 40, 44, 46, 47] usually leads to thelattices of MgO. In addition, MgO, due to its weak formation of solid solution. In this case, active com-basicity, can prevent catalyst from carbon depositing ponent is highly dispersed. There is also very strongto some extent [29, 30. These functions of alkaline interaction existed between alkaline earth oxide andearth oxide were also observed in other catalyst sys- active phase due to its chemical activity, which resultstems [31-35. The alkaline earth oxide can improve in anchoring of dispersed active particles, further pre-the dispersion of nickel due to the strong interaction vents these particles from agglomerating. The weakbetween nickel and alkaline earth oxide, and also it is basicity of alkaline earth oxide can restrain the car-the strong interaction, the highly dispersed nickel par- bon depositing to some extent. For example, carbonticles, once formed can be prevented from agglomer- deposition was hardly observed over the catalyst af-ating, and can be stabilized. Although alkaline earth ter 500 hour's run [46]. Other effective modifier isoxide supported or modified nickel catalyst exhibited rare中国煤化工 ese catalysts exhibitedhigh POM activity, the deactivation is unavoidable lordue to the carbon deposition and the loss of nickel at oH-romotion of rare earthCNMHGoxygen storage/releasehigh reaction temperature 19capacity, which lands itself to oxidizing surface de-Among the nickel-based catalysts, it was found posited carbon. It was also believed that it couldJournal of Natural Gas Chemistry Vol. 13 No 4 2004restrain catalyst from sintering at high temperatureStudies on Fe- or Co-based catalysts showed thatbecause of its strong interaction with active compo the activity of these catalysts was inferior to Ni-basednent. It was also reported that the improved activity catalyst because they show higher activity for the to-was attributed to the enhanced reducibility of active tal combustion of methane[49, 59, 60. For these sup-component after the addition of rare earth oxide 53, ported catalysts, the activity for POM to synthesisbecause it usually accepted that metallic component gas has the order as follows: Ni>Co>Fe. It wasis highly active for PoM to synthesis gasported by Swaan et al. 61 that the cobalt basedFurthermore, other active component such as Fe, catalyst was active only when it was promoted by theCo, Pd, Rh, etc. were tried to modify the nickel- substance that can facilitate its reduction, and thisbased catalyst for the purpose of improving the sta- was the reason that cobalt catalyst with higher load-bility and activity. Provendier and co-workers [54 ing had higher activity (62]. The support plays anfound that Fe could stabilize nickel catalyst. US- important role in determining the activity and stabil-ing sol-gel method, They synthesized LaNiz Fe(1-x)O3 ity of catalyst. Wang et al. [63] found that MgO is(0< r<1), perovskite-structured mixed oxide, a pre- an effective among the alkaline earth oxide supportedcursor of highly active catalyst for POM to synthesis cobalt catalysts, and that the calcination temperaturegas. For these catalysts, stability was improved as threw a great impact upon the activity and stabilitythe amount of added iron increased, owing to the re- The employment of cobalt together with noble metalversible migration of nickel from the bulk to surface. may be a good idea because cobalt is difficult to sin-Choudhary et al. [20] pointed out that cobalt addition ter. Pt-Co catalyst system showed high activity [64to Ni/Yb2O3, NiO/ZrO2 or NiO/ThO2 catalyst can It was usually thought that the active species of cobaltreduce the formation rate of carbon deposition and ac- catalyst was metallic cobalt, and that the stabilitytivation temperature of catalyst, which resulted from depended upon its preparation. Moreover, the deacimproved reducibility of nickel species by the cobalt tivation of cobalt containing catalyst resulted fromaddition. The addition of noble metal, although it is the sintering of active components and formation ofvery active for the POM, led to the change of nickel CoAl2O4 [63]. In fact, it was important to choose thechemical state or the distribution of nickel species support, for instance, Co/zrO2 showed high activitywith different oxidative state. Thus, the tempera- while Co/La2O3 deactivated rapidly [65]ture distribution along the longitudinal catalyst bedComprehensive investigation into Ni- or Co-basedwas also changed, usually the hot spots disappeared. catalyst has been performed, while less attention hasThe variation of nickel species with different oxidative been paid to Fe-based catalysts. The deactivation ofstate resulted in reasonable distribution between the Ni- based catalyst mainly attributed to carbon detotal combustion and reforming of methane occurred position and nickel loss at high temperature and highalong the longitudinal catalyst bed during the process GHSV. The utilization of cobalt and iron. due to theirof POM to synthesis gashigher melting points, if substituting for nickel, mayThe synthesis procedure of catalyst can affect its give better performance. Other elicitation from a ref-activity and selectivity to a great extent. Xu et al. erence[17] on steam reforming is to control the size56] prepared alumina supported nickel catalyst us- of active phase and introduce some modifier into cat-ing microemulsion method, and its stability was im- alysts for the purpose of improving the stability ofproved due to an increased coking resistance li et alanalyst57 prepared Ni/SiO2 catalyst using monodispersesilica sol and rather high POM activity was achieved, 3.2. Noble metal catalyststhough SiO2 is an inferior support. Highly cokingresistant nickel catalyst is also synthesized using co-Green and co-workers obtained high yields of syn-precipitation method [58]. In general, an important thesis gas over all noble metal catalysts, as well asstep in preparation is to improve the dispersion of ac- over the rare-earth ruthenium pyrochlores. Thesetive phase and the interaction between active phases catalysts catalyzed methane conversion to synthesisand support. Both of these factorsetermmine the gasH中国煤化工 losely approaching tochemical state of active component, namely the activthes. For palladium cat-ty of catalyst So the pretreatment of catalyst, like alystCNMH Garbon deposition wasthe determining step of synthesis procedure, is very observed while for iridium and rhodium catalyst, nocrucial to the activity and selectivitymacroscopic carbon deposition was observed ( 10, 11]196Quanli Zhu et al./ Journal of Natural Gas Chemistry Vol. 13 No. 4 2004Poirier and co-workers [66] carried out POM to on Rh based catalysts. It was found that the com-synthesis gas experiment at extremely high GHSv pound formation between Rh and support depended(0.893 molCH,/(kgs), CH4/O2/He=8 /4/3), namely upon the calcination temperature. No Rh compoundunder the condition that products were dominated was formed over y-Al2O3 and SiO2, while LarhO3by kinetics, and it was found that Rh was more ac- MgRh2O4, YRhO3 and RhTaO4 were formed overtive than nickel, even though at very low loading La203, MgO, Y2 O3, Ta2O5, respectively, if calcined0.015wt%Ru/Al2O3)at properly high temperature. Among them, La2O3Hochmuth et aL. [67-70 studied the monolith can provide better catalytic activity and selectivity af-supported noble metal catalysts for Pom to syn- ter adequate calcinations. The catalyst stability andthesis gas. The results of pilot plant test at high the interaction between active metal and support canGHSV showed that Pt or Pd and the like, was ex- be improved at higher reaction temperature. Amongtremely effective for the production of synthesis gas the non-reductive metal oxide supported catalystfrom methane. They drew the conclusion that the the activity decreased according to the order as fol-complete oxidation of methane had carried out at the lows: La2 O3 r-Al2O3 RuIr>Pt>Pd. The best catalytic perfor- significantly during the treatment in hydrogen, whilemance was observed for a 1% Rh content(atomic ra- no structure change was observed during POM re-tio) and Rh content above 1% did not increase the action, and not influenced by the residence time ofactivity, unlike Ru based catalysts. The results by reagents. In the other hand, methane conversion andYan et aL. [73] indicated that the conversion and se- selectivity depend upon the residence time. Ruck-lectivity were relative stable over the Rh based cat- enstein et al. [80 also investigated into the effectalyst, while they were changeful over Ru based cat- of Rh content in alumina supported catalysts on thealyst. Furthermore, the pulse reaction showed that catalytic performance, and it was found that almostonly Co was formed as carbonaceous product over the same methane conversion and selectivity was pro-Rh catalyst, while CO and CO2 were formed over vided when Rh content was within the range from 0.5Ru catalyst. This implies that the reaction mech- to 5.0wt%anisms over these two catalysts are different. As forJones and co-workers[81]studied the performancethe Ir based catalysts during POM reaction to synthe- of EuaIr2O7 catalyst using in situ X-ray diffractionsis gas, the activity order of supports was as follows: and Mass spectrometry and found that the pyrochloreTiO2 La2O3 >MgOSiO2 structure of iridate catalyst was destroyed at the out-[74. A series of rare earth supported noble metal- set of catalysis, giving an active catalyst that wascatalysts was studied, among them, Pt/Gd2 O3 and shown to comprise particles of iridium metal of aboutPd /Sm2 O3 gave preferable catalytic performance[75]. 3 nm in diameter supported on europium oxide. TheNevertheless, the selectivity to CO was higher than sudden increase of synthesis gas, monitored by massthat to H, and it was ascribed to the reverse reaction spectrometry, corresponded to the onset of reductionof steam reforming. They thought that alkali earth or of pyrochlore. Ashcroft et al.[82 drew similar conrare earth metal oxide played the double roles: one is clusions in a study of iridium pyrochlore catalysts us-to disperse noble metal and the other is to improve ing中国煤化工 ray diffraction by syn-the selectivitychroCNMHGRuckenstein and coworkers [76-78 investigatedOJ lund excellent catalyticinto the effect of different structured magnesia, rare properties of RhVOa/SiO2 and un-promoted Rh/SiO2earth metal oxide, as well as other stable supports catalysts for the PoM to synthesis gas, above 90% ofJournal of Natural Gas Chemistry Vol. 13 No 4 2004197methane conversion at 973 K. They also compared membrane reactor, POM to synthesis gas can be carthe activity of the two catalysts over temperature ried out at lower temperature, while higher CO andrange of 573-973 K at ambient pressure using a feed H2 selectivity can be achieved. Armord et al. 90of CH4/O2 with a molar ratio of 2 diluted with ni- found that Pd based membrane reactor can elevatetrogen. It was found that the onset of activity oc- H2 production during methane converting to synthe-curred around 773 K for RhVOA/SiO2 catalyst, while sis gas or liquid fuel. Kikuchi et al. [91]found that us-for Rh/SiO2 the catalyst exhibited activity at tem- ing noble metal based membrane catalysts POM canperature above 873 K. The examination of the used be carried out at temperature below 773 K if the feed-catalysts indicated that the average Rh particle size stock is poor in oxygen. The methane conversion andin RhVO4/SiO2 catalyst was smaller than that in CO selectivity can be elevated through removing H2the un-promoted catalyst, Rh/SiO 2. Therefore, the from the reactor. It was also found that carbon depo-difference of activity at low temperature was ascribed sition occurred after steam was depleted, while it canthe active metal particle size, morphology and its be avoided by replenishing steam. The nickel-basednteraction with the support.membrane reactor was also reported lately [92 andThe activity of the catalysts of 1% Pd supported high methane conversion and selectivity was reached.on oxides including IIIA-IVA metal oxide and rareMonolith reactor or monolith supported catalystsearth oxide, were investigated at 1023 K, using GHSV similar to the membrane reactor was tried to be ap-of 5000 h-1 and CH4/O2 ratio of 8: 1 [84]. The plied to POM to synthesis gas [93,94]. For rhodiumthane conversion varied from 33.4% to 66.9%, but impregnated foam monoliths, very high methane consurprisingly they all gave more than 99% of selectivity version(>90%), CO selectivity(>90%)and completeto CO, with no data for hydrogen selectivity. How- conversion of oxygen was achieved during POM tever, the methane conversion in all cases exceeded the synthesis gas under the adiabatic condition, using ex-theoretical maximum(25%)for synthesis gas produc- tremely short residence times of between 10-4 andtion under the fixed ratio of methane to oxygen. In 10-2 seconds and the feedstock with stoichiometricaddition, the GHSV was set at 5000- in their work, ratio. Under the same condition, H2 selectivity forwhich is very small as compared with those used by pt based catalyst decreased to 70%, whilst the Pdother researchers. It is thus probable that carbon de- catalyst promoted the carbon depositionposition over palladium catalysts is responsible for theAnother interesting noble metal based catalystdifference in methane conversionsystem is a mixed oxide(Ba3NiRuTaOg) with perPlatinum supported on alumina doped withovskite structure[ 95. At 1173 K, it can provide 95%conia gave very excellent performance, which wasof methane conversion and 98% of H2 selectivity. Atcribed to the improved oxygen mobility brought by 1070 K, it can catalyze the complete conversion ofthe introduction of zirconia [85]. The properties of ethane, obtaining 94% of synthesis gas, but there is nosupport significantly affect the activity and selectivity transformation of perovskite structure post-catalysisof Pt-based catalyst via adjusting interaction betweendsupport and active component [86]. The studies on Ptcatalyst revealed that the deactivation of Pt catalyst 3.3. Early transition metal carbides and otherwas mainly due to the agglomeration of dispersed Pt catalystsparticles and carbon deposition [87The modification of support can affect the oxida-Early transition metal carbides, particularly oftive state of active component, which is the key factor molybdenum and tungsten, exhibited excellent cat-to determine the activity and selectivity. Elmasides alytic performances in a large number of reactionset al.[ 89] reported that for the Ru/TiO2 catalyst, which were usually catalyzed by noble metal basedthe introduction of wo+ into TiO2 led to stabilize- catalysts. York et al.[96, 97] obtained high yields oftion of oxidative ruthenium, which resulted in lower synthesis gas using supported molybdenum or tung-conversion and selectivity, while the introduction of sten carbides at elevated pressure and temperature.Ca2+ cation led to the formation of metallic ruthe- Butnium, which resulted in higher conversion and selec- sultH中国煤化工 t ambient pressure,reCNMHIn addition, if POmmetric conditions, noRecently, noble metal based membrane reactor carbon deposition was observed on the used catalystshas attracted a lot of attention. Using this kind of a study of the relative activity of molybdenum car-198Quanli Zhu et al. /Journal of Natural Gas Chemistry Vol. 13 No. 4 2004bide to the noble metals demonstrated that it had an the literature. As far as it goes, it can be divided intoactivity similar to iridium, both per active site and per two categories: one is the indirect oxidation mechagram [96], while high space velocity is unfavorable to nism involving methane total combustion, and steamthe stability of carbide catalysts. The deactivation of and dry reforming reactions, which is often referred tothe catalyst may result from the oxidation of carbides as the "Combustion and Reforming Reactions Mecha-into oxides, followed by vaporization of oxides under nism"(CRR); the other is the direct oxidation mechambient pressure. However, under the elevated pres- anism in which surface carbon and oxygen species re-sure, the vaporization was choked up, since high pres- act to form primary products, known as the "Directsure prevented the carbide possessing much higher Partial Oxidation Mechanism"(DPO)boiling point than its oxide counterpart, from beingtransformed into oxides, especially in the reducing at 4.1. CRR mechanismmosphere. The atmosphere imposes influence greatlyupon the stability of carbide, particularly at high tem-First mention of the CrR mechanismere of steam and CO2 goes by Prettre et al. [24]. Their experiments, later re-against the retention of oxides, but hydrogen and Co peated by Vermeiren et al.[25, indicated that thefavor retaining carbides[98. Recent study shows that longitudinal temperature profile of the catalyst bedaddition of some transition metals can significantly was not uniform, namely, markedly higher tempera-increase the catalyst activity and stability [99]. With ture of the front part of catalyst bed than that of thethe addition of transition metal promoters, the cata- rear part and the furnace temperature, as shown inlyst activity can be as high as the noble metal catalyst Figure 4[25]even at very high space velocity and pressure conditions, but there is much less carbon deposition overExothermthe carbide catalysts. Other example is B-Sic withmedium surface area used as support [100]. B-Sicsupported nickel catalyst showed stable and high ac-tivity for POM, and the hot spots usually occurred onalumina supported catalysts were removed due to the2CH:+ O,Endothe2C0+4H,high thermal conductivity of B-Sic, and the carbonnanofilament growth was scarcely observedBesides above POM catalysts, it was reported byCatalyst bedOtsuka et al. [101] that cerium dioxide could trans- Figure 4. Schematic representation of the tempera-form methane into synthesis gas with a molar ratio ofture in POM catalyst beds2 for H2 to CO within temperature range from 873to 1073 K. It was demonstrated that during the re-Choudary et al. 59 carried out the POM reac-dox cycle of ceria, methane is directly converted to tion using Ni/Mgo catalyst under high GHSV conH2 and CO Carbon dioxide resulting from the oxida- dition in order to get non-equilibrium product distion of methane in gas phase is reduced by partially tribution. When York et al. [102] repeated the ex-reduced cerium cation and CO is given as the only periment reported by Choudary et al, it was foundproduct. The addition of Pt to cerium dioxide cata- that the hot spformed. These results indicatelyst accelerated the formation of synthesis gas, while that exothermic reactions occurred at first and thenthe reduction of catalyst over 60 minutes led to the followed by endothermic reactions. As depicted insynthesis gas with a molar ratio of H, to CO higher Figure 2, the total combustion of methane, and thenthan 2. The latter case implies the formation of car- followed by steam and dry reforming of methane.Inbon deposition after a period of reduction. However, fact, nickel and noble metal are very effective catalystcerium is used as a modifier at more time to improve for the steam shift and steam reforming of methanethe oxygen mobilityAcc中国煤化工inm, synthesis gas is sec-I in their latter ex4. Methane partial oxidation mechanismCNMHGeffect of reaction con-ditions on the product distribution in the process ofThe mechanism of methane partial oxidation to PoM to synthesis gas, and explained why there issynthesis gas was dealt with and debated widely in lower selectivity of synthesis gas and higher selectivityJournal of Natural Gas Chemistry Vol. 13 No. 4 2004199of CO2 and H2O under the condition of higher GHSvTransient response and isotope exchange tech-or higher ratio of O2/CH4, using CRR mechanism. At nique have also been used to investigate into the POMthe same time they pointed out that hydrogen and mechanism. Nakagawa et al.[106, 107] reported thatcarbon monoxide were formed as secondary products. synthesis gas was formed over Ir/TiO2 and Rh/ SiO2catalysts via CRR mechanism. However, the en-4. 2. DPO mechanismdothermic reaction, methane decomposed to hydro-gen, carbon and dehydrogenated methane group, ini-Hickman and Schmidt et al.[68-70 consideredtially occurred, followed by a reaction: carbon andH2 and CO as primary products during POM to syn- the dehydrogenated methane group oxidized by oxythesis gas under adiabatic condition at very shortgen to CO species. As for POM performed over sup-residence time. when they doubled the residenceported Rh catalyst, its reaction pathway dependedtime, the conversion and selectivity were improvedgreatly upon the properties of the support. RuckWhen substituting Pt-10% Rh wire net for monolith enstein et al108, 109 reported that during POMsupported catalyst, the conversion and selectivity in-reaction over Ni/SiO2 catalyst, CH, CH2 and CH3creased with increasing the gas velocity of feedstockspecies were formed, which means that methane is ac-This phenomenon is conflicted with the case in Ref.tivated via dissociation, and the amount of methane[11]. If the gas velocity is fixed and the layer number taking part in isotope exchange was larger than theof Pt-Rh wire net is increased(not less than 3), no amount of methane converting to CO and CO2.It wasfference of conversion and selectivity was observed concluded that methane dissociating is not the rate-and product distribution was away from the equiliblimiting step. Over the un-reduced Ni/ SiO2 catalyst,rium of steam shift or steam reforming reaction, com- methane directly reacted with oxygen without disso-panied with a lower ratio of H2/co than that calcu-canonrew a simila.lated according to thermodynamic equilibrium. Allconclusion regarding alumina supported nickel cata-these facts cannot be explained by CRr mechanismlyst. Temperature can throw influence upon the PoMIn order to elucidate the phenomena mentioned above, pathway. Within the temperature from 973 to 1023the dPo mechanism was put forward. In this mecha-K, POM proceeds mainly via the pathway of the dis-nism, synthesis gas is produced as a primary product. sociation of methane, whilst at the temperature ofCH4=C(ads)+ hAds)1123 K. CRR mechanism makes a rather contribution to POM (26. Li et al. 111 made a point ofC(ads)+[O]= CO(ads)=Co(s)producing a high yield of synthesis gas requiring thecatalyst with reduced state. As reported in manyads)References. [45, 55, 73, 112-114], metallic active com-They constructed a model incorporating the ele- ponent, not only nickel but also noble metalmentary adsorption, desorption and surface reaction effective to produce H2 and CO. Surface state deter-steps involved in a mechanism, of which some of the mines reaction mechanism, and plays an importantmost important steps are shown in the above equa- role in determining conversion and selectivitytions. according to this model, the product selectiv-Judging by current evidences for the partial oxida-ity over Pt or Rh catalyst can be forecastedtion of methane to synthesis gas, it is possible for theRecently, the studies of POM specific mechanism two mechanisms to occur over nickel or noble metalunder specific conditions have become popular. When catalysts, but the real pathway depends upon the realWeng et al. [103-105 investigated into the reduc- conditions. The pivotal factor is the chemical state oftion of Rh and Ru catalyst using FTIR technique, it active component element: metallic state prefers towas found that Co was formed as a primary prod- methane dissociation reaction, followed by surface re-uct of POM reaction over reduced or really working action with oxygen species to synthesis gas; while theRh/SiO2 catalyst, DPO pathway was the main route active component element with higher oxidative num-of formation of synthesis gas over Rh/ SiO2 catalyst. ber facilitates deeper oxidation of H2 and CO, as ob-In contrast to this, CO2 was a primary product of serve中国煤化工 nt et al125· However,,POM reaction over Ru/Al] O3 or Ru/SiO2 catalyst the oonent relates to th3. Synthesis gas was formed over Ru-based cata- propCNMHrs, pretreatments,re-lyst by way of CRR mechanism. Of course, the oxygen action temperature and oxygen partial pressure, etccontent in feedstock can alter the reaction directionOther important reason is the active surface oxygenQuanli Zhu et al. /Journal of Natural Gas Chemistry Vol. 13 No 4 2004species and its mobility. Of course, kinetic factors catalyst is stable only under a condition of high pres-may exert an influence on it, even changes its direc- sure and low space velocity [99]tion. This is the reason why two reaction mechanismsseem to be possible over all catalysts5. Problems and future studiesThe two mechanisms are also applied to elucidatthe POM to synthesis gas over carbide catalysts 116Non-catalytic homogeneous partial oxidation of(1)DPO type mechanism: this involves surface methane to synthesis gas is well established. For ex-species similar to those shown for the DPO mecha- ample, in Sarawak, Malaysia, Shell have been suc-nism discussed earlier. However, it is likely that syn- cessfully operated a highly selective process for pro-thesis gas is not a primary product over the carbide ducing synthesis gas at high temperatures, typicallycatalysts, and that CO2 and H2o are important re- above 1400 K, and pressures of around 5-7 MPa,action intermediatesas a part of the Middle Distillate Synthesis proces(2)Redox mechanism: O2, CO2 or H2o in the (SMDS)[117]. Out of question, employment of catreactor can react with surface carbide carbon species, alysts would markedly lower the operating temper-generating vacancies and oxide species. These vacan- ature required for the production, which makes thecies can then react with carbon from methane disso-process more economically attractive [118]. However,ciation, returning the site back to the carbidic. This more work should be done to solve the following mainis shown below for the reaction of CH4 and CO2problems for the practical application of this processMo2C+5C02= 2MoO2+6CO(1) Carbon deposition over the reactor and catalyst bed. There are two possible routes for the forma-2MoO2+5CH4= Mo2C+4C0+ 10H2tion of carbon, namely methane decomposition andthe boudouard reactionThe most probable mechanism over carbide cata-lysts is the redox mechanism according to the results CH4= C(s)+ 2H2(methane decomposition reactionobtained by Xiao et al. [99]using in situ Raman andbulse techniques. a possible model for the reaction is2C0=C(s)+ CO2(Boudouard reactiongiven in Figure 5 [99Studies by Claridge et al. (119 demonstrated thatCH+both reactions are thermodynamically favorable un-der reaction conditions typically for methane partialoxidation, but that the source of carbon deposition ismainly resulted from the methane decomposition. Ev-idence for this was given by observing the amount ofed on a nickel catalyst inor carbon monoxide atmosphere; at high tempera-tures the pure methane gave much more depositedcarbonthan pure carbon monoxide, while support-ing evidence arose from the fact that carbon built upfrom the front of the catalyst bed, where methanepartial pressure was at its highest. Two types of car-bon are formed on the partial oxidation catalysts asshown in Figure 6[14:(a) encapsulate carboenvelops the nickel particles resulting in deactivationand(b)whisker carbon, which grows from the face ofthe nickel particles and does not alter the rate of syn-Figure 5. Model of 2CH4+O2 reaction to synthesisgas over molybdenum carbide catalystthesis gas formation, but is likely to eventually resultOxygen first reacts with the carbide, producing for中国煤化 I studies on the carbonCO; the oxide or oxycarbide surface is then reduced relatCNMHGion[120-123]. To sup-by methane to produce CO and H2. Because the re- press the carbon deposition, more work needs to becarburization of the oxide surface by methane is a done on the catalyst preparation and reaction condi-slow process and endothermic reaction, the carbide tion optimization. For example, some steam mayJournal of Natural Gas Chemistry Vol. 13 No 4 2004201ded to the feedstock to eliminate the hot spots in state, and thus to release from the thermodynamicthe catalyst bed and also may be helpful to suppress restriction. The combination of POM and steam re-the carbon depositionforming as an alternative, to increase cH4 conversionand synthesis gas production, is also possible. In addi-tion, a more stable catalyst system being able to resistcarbon depositing under excessive methane feedstockis very important to increase the selectivity to COand H2. A high CH4/O2 ratio is also desired from thview of safety, because lower CH4/O2(<1.5)ratioincreases the danger of explosions, and this is of particular importance when high pressure is employed1O0 nm6. ConclusionsFigure 6. Micrograph showing carbon depositedover a nickel catalyst after POMAs mentioned above, a great number of chemicals(2) Active component loss during the POM to and fuel can be obtained from methane by way ofsynthesis gas, particularly of nickel catalyst. The synthesis gas, while the direct conversion of methaneoverall PoM reactions are mildly exothermic, while is economically infeasible due toits intrinsic barrierit may occur by two steps, initial combustion and Therefore, the partial oxidation of methane to syn-then dry and wet reforming In the first step, a large thesis gas is likely to become more important in theamount of heat is given off, which can melt the activeuture, particularly when alternative sources of energymetal, leading to peeling the active metal off the sup-are required. Early work showed that nickel was anport. Because nickel has a lower melting point, lower active catalyst for this reaction. Now it has been fol-than other active components, such as noble metal lowed up, particularly in the past 2-3 decades. Ator Co and Fe, thus, it may be easier to deactivate. present, a number of potential alternative catalystHowever, this can be improved by strengthening the have been discovered for carbon free methane partialinteraction between support and metal, and carrying oxidation, including the noble metals and molybde-out the reaction at milder temperaturenum and tungsten carbide. However, there are alsoNoble metal catalysts have shown superior ad- some problems such as carbon deposition for nickelvantages to nickel based catalyst in carbon deposition- catalyst, the stability for transition metal carbide andresistance, but the carbon deposition is still unavoid-o on, to be resolved. In order to make POM to syn-able at high temperature because of the acidic sup-thesis gas become popular industrial process, furtherport and the methane decomposition, etc. As restudy on this subject is requiredported by Albertazzi et al.[87, carbon depositiontogether with agglomeration of noble metal particlesReferencesresulted in the deactivation. Also a high loading of no-ble metal is required to sustain the high activity, thus [1] Cornot-Gandolphe S Energy Exploration Exploita-the cost of the catalyst is very high. The combina-tion,1995,13:3tion of a small amount of noble metal with transition [2] Labinger J A, Ott K C J Phys Chem, 1987, 91: 2682metals such as Co, Ni or Fe may be a wise way to 3 Hutchings G J, Scurrell M S, Woodhouse JR.Chemdecrease the catalyst cost and improve catalyst activSoc Rev,1989,18:251ity and stability. New alternative catalysts such as4 Gesser H D, Hunter N R, Prakash C B. Chetransition metal carbide are expected to decrease the1985,85:235catalyst cost and improve the catalyst stability5 Otsuka K, Komatsu T, Jinno K et al. In: Philips M J,Ternan M eds. Proc 9 Int Congr Catal, Vol.Il.Ot-(4)Because of the thermodynamic restriction, thetawa: The Chemical Institute of Canada, 1988. 915.POM reaction under high pressure often gives rise to (61more CO2 and H2O formation. To improve the product distribution, the feedstock needs to be optimized. 7H中国煤化工 itt e R et al. ScienceCNMHGSosnicki jg etal. JNew technology such as membrane catalyst and re-Catal,2003,215:14actor are expected to lead the reaction to a dynamic [8 Hutchings G J. Appl Catal, 1986, 28: 7Quanli Zhu et al. Journal of Natural Gas Chemistry Vol. 13 No 4 20049] Leroux P J, Mathieu P M. et al. Chem Eng Prog, [38 Hayakawa T, Harihara H, Andersen A G et al. Appl1961,57:54CataI A,1997,149:391[10] Ashcroft A T, Cheetham A K, Foord J S et al. Na- [39 Utaka T, Al-Drees S A, Ueda J et al. Appl Catal Ature,1990,344:3192003,247:12511 Vernon P D F, Green M L H, Cheetham A K et al0 Takamura H, Enomoto K, Aizumi Y et al, Solid StateCatal Lett. 1990. 6: 181ionics,2004,175:379[12] Zhu JJ, Van Ommen J G, Lefferts L J Catal, 2004, [41]Borowiecki T. Appl Catal, 1982, 4:223225:388[42]Basile F, Basini L, Amore M D et al. J Catal, 1998,[13 Aghalayam P, Park Y K, Fernandes N et al. J Catal173:2472003,213:23[43] Basile F, Fornasari G, Rosetti V. Catal Today, 2004,14] Tsang S C, Claridge J B, Green MLH. Catal Today,91:2931995,23:3[44]Choudhary V R, Uphade B S, Mamman A S Micro-[15] Dissanayake D, Rosynek M P, Kharas K CC et al. Jporous and Mesoporous Material, 1988, 28: 61Catal,1991,132:117[45] Jun J H, Lee T J, Lim T H J Catal, 2004, 221: 178[16]Wietzel G, Haller W, Hennicke W. USP 1 934 836. [46]Ji YY, Chen Y X, Yu C Y et al. Chemical Engineer-1927ing of Natural Gas, 1999, 24: 12[17] Roustrup J R. In: Anderson J R, Boudart M eds. [47] Choudhary VR, Rajput A M, Mamman A S J CatalCatal: Sci and Tech, Vol 5. Berlin: Springer, 1984. 11998,178:576[18」 Choudhary V R, Rajput A M,RaneⅤH.JPhy48 Zhu T, Flytzani-Stephanopoulos M. Appl Catal A(19) Choudhary V R, Rajput A M, Prabhakar B J Catal, [49 Slagtern A, Olsbye U. Appl Catal A, 1994,110: 991993,139:32650 Chu Y L, Li S B, Lin J Z et al. Appl Catal A, 1996,O Choudhary V R, Rane V H, Rajput A M. Appl Catal134:A,1997,162:235[51] Liu S, Xiong2000,130D(Int cong on catal): 35[21] Solbakken A. Stu Surf Sci Catal, 1991, 66: 44622 Liander H. Trans Faraday Soc, 1929, 25: 462[52] Wang W, Stagg-Williams S M, Noronha F B et al[23] Padovani C, Franchetti P. Giorn Chem Ind Applicata,Catal Today, 2004, 98: 5531933,15:429[53] Mattos L V, De Oliveira E R, Resende P D et al24 Prettre M, Eichner Ch, Perrin M. Trans Faraday Soc,Catal Today, 2002, 77: 24554 Provendier H, Petit H C, Estournes C et al. Appl1946,42:335Catal A,1999,180:1635 Vermeiren WJ M, Blomsma E, Jacobs P A. Catal[55] Basile F, Fornasari G, Trifiro F et al. Catal TodayToday,1992,13:4272002,77:215126 Ruckenstein E, Hu Y H. Appl Catal A, 1999, 18:85 156) Xu S, Zhao R, Wang XL. Fuel Processing Technology2004,86:1233T:2333[57 Li J Z, Lu G X. Appl Catal, 2004, 273: 16328 Santos A, Menendez M, Monzon A et aL. J Catal[58]Xu S, Wang X L. Fuel,6,158:59 Choudhary V R, Rajput A M, Prabhakar B et al29 Choudhary V R, Mamman A S. Appl Energy, 2000,Fuel,1998,77:1803[60 Slagtern A, Swaan H M, Olsbye U et al. Catal Today,30] Tang S, Lin J, Tan K L. Catal Lett, 1998, 51: 1691998,46:10731] Tsyganok a I, Inaba M, Tsunoda T Appl Catal A, (61] Swaan H M, Rouanet R, widyananda P et al.Stud2004,275:149Surf Sci Catal, 1997, 107:(Natural Gas Conversion32 Takeguchi T, Furukawa S N, Inoue M et al. ApplIV)447[62] Chang Y F, Heineman H Catal Lett, 1993, 21: 215[ 33] Yang S, Kondo J N, Hayashi K et al. App! Catal A, [63] Wang H Y, Ruckenstein E J Catal, 2001, 199:3092004,277:239[64 Mo L, Fei J, Huang C et al. J Mol Catl A, 2003, 19334] Figueroa J C, Gaffney A M, Mattson R H SR et al[65] Bi X J, Hong P J, Dai SS. Chinese J Mol Catal, 1998,[35 Jing Q, Lou H, Mo L et al. J Mol Catal A, 2004, 212211中国煤化工D. Catal Lett,1993,21:36 Hayakawa T, Andersen A G, Shimizu M et al. Catalett,1993,22:307CNMH7B,1992,1:89[37] Hayakawa T, Harihara H, Andersen AG et al. Angew [68] Hickmann D A, Schmidt L D. J Catal, 1992, 138: 267Chem(Int Ed Engl), 1996, 35: 19269 Hickmann D A, Schmidt L D. Science, 1993, 259: 343Journal of Natural Gas Chemistry Vol 13 No 4 2004[70] Hickmann D A, Haupfear E A, Schmidt L D. Catal [98] Darujati A R S, LaMont D C, Thomson W J. ApplLet,1993,17:223CataI A,2003,253:397[71] Schwiedernoch R, Tischer S, Correa C et al. Chem [99] Xiao T, Hanif A, York A P E et al. Proceeding of theEng Sci,2003,58:6334h World Congress of Oxidation Catalysis, Potsdam72 Basile F, Basini L, Fornasari G et al. Stud Surf SciGermany, Sept, 2001Catal(Preparation of Catalysts VII), 1998, 118: 31 [100] Leroi P, Madani B, Pham-Huu C et al. Catal Today3 Yan Q G, Wu T H, Weng W Z et al. J Catal, 20042004.91:53226:247[101 Otsuka K, Ushiyama T, Yamanaka I. Chem Lett,[74Nakagawa K, Anzai K, Matsui N et al. Catal Lett,1993,15:171998,51:163[102 York A P E. [PhD Thesis]. Oxford: University of[75] Choudhary V R, Prabhakar B, Rajput A M et alOxford. 1993Fuel,1998,77:1477103 Weng W Z, Chen M S, Yan QG et al. Chin Sci Bull,[76Ruckenstein E, Wang H Y J Catal, 1999, 187: 151[77 Ruckenstein E, Wang H Y. Appl Catal A, 2000, 198:2000,45:2236104 Weng W Z, Chen M S, Yan Q G et al. Catal Toda[78 Ruckenstein E, Wang H YJ Catal, 2000, 1902000,63:317[79Grunwaldt J D, Basini L, Clausen B S J Catal105 Weng W Z, Yan Q G, Luo C R et al. Catal Lett,2001,74:37200:32180 Wang Y, Ruckenstein E Catal Lett, 1999, 59: 121 [106 Nakagawa K, Ikenaga N, Teng Y H et al. J Catal,[81]Jones R H, Ashcroft A T, Walker D et al. Catal Lett,1999,186:4051991,8:169107 Nakagawa K, Kiyoharu N O, Kobayashi T, et al[82 Ashcroft A T, Cheetham A K, Jones R H et al. JCatal Today, 2001, 64: 3Phys Chen,1993,97:3355[108] Wang H Y, Ruckenstein E. J Phys Chem B, 1999,[83 Kunimori K, Umeda S, Nakamura J et al. Bull Chem103:11327Soc Jpn,1992,65:2562[109Hu Y H, Ruckenstein E J Phys Chem A, 1998, 102[84 Bhattacharya A K, Breach J A, Chand S et al. Appl10568A,1992,80:L1[110 Jin R, Chen Y, Li W et al. Appl Catal A, 2000, 201[85 Souza MM VM, Schmal M. Appl Catal A, 2003, 255[111] Li C, Yu C, Shen S Catal Lett, 2000, 67: 13986] Pavlova S N, Sazonova NN, Ivanova J A et al. Catal [112] Tulenin Y P, Sinev M Y, Savkin VV et al. Catalbday,2004,91:299Today,2004,91:155[87] Albertazzi S, Arpentinier P, Basile F et al. Appl Catal [113) Liu Z W, Jun K W, Roh H S et al. J Mol Cata], 2002A,2003,247:1[88] Souza MM VM, Schmal M. Appl Catal A, 2003, 255: [114] Yan Q G, Weng w z, Wan H L et al. Appl Catal A2003,239:43[89] Elmasides C, Kondarides D I, Neophytides S G.J[115] Balint I, Miyazaki A, Aika K J Catal, 2003, 220: 74Catal.,2001,198:1959o] Armor J N. J Membr Sci, 1998, 147: 217[116 Hanif A [PhD Thesis]. Oxford: Oxford University200091]Kikuchi E, Chen Y. Stud Surf Sci Catal, 1998, 119(Natural Gas Conversion V): 441[117 Eilers J, Posthuma S A, Sie S T Catal Lett, 1990, 72 Zhu D C, Xu Y, Feng S J et al. Catal Today, 2003,[118] Pena M A, Gomez J P, Fierro J L G. Appl Catal A93 Heitnes K, Lindberg S, Rokstad O A et al. Catal1996,144:7Today,1994,21:471119 Claridge J B, Green M L H, Tsang S C et al. Catal94 Schmidt L D, Huff M. Catal Today, 1994, 21: 4433.22:299[95] Battle P D, Claridge JB, Copplestone F A et al. Appl [120] Rostrup-Nielsen,Bak Hansen JH.JCatal,1993, 144Catal A,1994,118:217[96]Claridge J B, York A P E, Marquez-Alvarez C et al. [121] Rostrup-Nielsen J R J Catal, 1974, 33: 184J Catal,1998,180:85[122 Rostrup-Nielsen J R, Trimm D L. J Catal, 1977, 4897] York A P E, Claridge J B, Brungs A J et al. StudSurf Sci Catal. 1997. 110: 711[123]c。+。ISci eng,1982,24:67中国煤化工CNMHG