Oxidative CO2 Reforming of Methane on Alumina-Supported Co

Oxidative CO2Reforming of Methane on Alumina-Supported Co-Ni CatalystSay Yei Foo,Chin Kui Cheng,Tuan-Huy Nguyen,and Adesoji A.Adesina*Reactor Engineering and Technology Group,School of Chemical Engineering,The Uni V ersity of New SouthWales,Sydney,New South Wales2052,AustraliaCO2reforming of CH4in the presence of O2in the feed has been investigated in afixed bed reactor containinga Co-Ni catalyst.The reforming rate increased with O2partial pressure before leveling out at O2:CH4)1at temperature greater than823K.Although CO production decreased with O2addition,H2formation initiallyrose to a maximum before a slow decline.An expression relating the optimum O2partial pressure for H2production as a function of temperature was obtained as,P O2,max)1.008×10-3e8420/T.The H2:CO productratio increased from0.9for pure CO2reforming and peaked at1.73as O2partial pressure increased to an equimolar level in the feed.The increased reaction temperature resulted in lower H2:CO due to increased CO2reforming kinetics.The complete consumption of O2in thefinal product stream means that the oxidative CO2reforming of CH4may be used to generate nearly ideal syngas composition for Fischer-Tropsch synthesis if feed CO2:CH4is unity.In particular,the overall heat demand for the reforming reaction could be reduced (or become exothermic)for a judicious combination of CO2/CH4/O2.The postreaction analysis revealed that even mild O2dosing leads to negligible carbon deposition.Thus,this form of reactor operation is energetically attractive and provided efficient carbon utilization.1.IntroductionSynthesis gas(H2/CO)is primarily produced from catalytic steam reforming of natural gas,a highly energy and capital intensive operation.Furthermore,the product ratio of the steam reforming process(H2:CO>3)is higher than that required for downstream methanol and hydrocarbon conversion processes.1 With increasing concern on the rise of anthropogenic greenhouse gas emissions,there has been renewed interest in the replace-ment of steam as a reactant with carbon dioxide.Implementation of CO2reforming rather than conventional steam reforming is also attractive in areas where water is not readily available. However,similar to steam reforming,the dry reforming route is an endothermic reaction and suffers from carbon-induced catalyst deactivation on conventional Ni-based catalysts.Noble metals such as Ru,Rh,Pt,Pd,and Ir have been found to be highly active with low carbon formation;2-4however,due to the high cost and limited availability of these metals,it is more practical to develop Ni-based catalysts.CO2reforming also yields product H2:CO ratios lower than1,which is below the ideal for a downstream Fischer-Tropsch synthesis plant.In order to reduce energy consumption for synthesis gas production,the catalytic partial oxidation process has attracted significant interest.Since the oxidation of hydrocarbons to synthesis gas mixtures is exothermic,this process is deemed more energy efficient than both the steam and dry reforming options.However,the oxidation reaction involves cofeeding hydrocarbon-oxygen mixtures underflammable or explosive conditions.5,6Flames in the reaction zone may also lead to local hot spots with possible catalyst sintering.To overcome these problems,the combined reaction of CO2reforming and partial oxidation may be utilized.Ashcroft et al.have shown that exothermic partial oxidation may be combined with endothermic CO2reforming to achieve a thermally neutral reaction,with high synthesis gas yields.2The other advantages of the oxidative CO2 reforming reaction are increased CH4conversion,7high yields at low temperatures,8,9the ability to control the thermal behavior of the system,10and improved stability and resistance to deactivation.11Downstream Fischer-Tropsch hydrocarbon se-lectivity may also be manipulated by controlling the product H2:CO ratio via oxygen addition.12,13The uniqueness of the present investigation lies in three distinct operational attributes:(1)the ability to predict optimum O2partial pressures from the kinetics of the participating reactions;(2)tunability of the H2:CO ratio to meet varying downstream ends;(3)carbon deposition control without shut-down or any other adverse operational effects without losing gaseous products selectivity.Bimetallic Co-Ni catalysts have been proven to offer superior performance for CH4dry reform-ing in terms of activity and stability compared to monometallic and other Ni-based bimetallic combinations.14Previous studies in our laboratory have also shown that alumina-supported bimetallic Co-Ni catalysts exhibit synergistic effects during hydrocarbon reforming and superior coking resistance compared to monometallic Ni/Al2O315,16and have therefore been employed in the current investigation.Although the net reaction for oxidative dry reforming isthe complex reaction network includesPartial oxidation of CH4CH4+0.5O2f2H2+CO(3) Reverse water-gas shift reactionCO2+H2f H2O+CO(4) CH4combustion reactionsCH4+2O2f2H2O+CO2(5) H2oxidationH2+0.5O2f H2O(6) CO oxidationCO+0.5O2f CO2(7)*To whom correspondence should be addressed.Tel.:+6129385 5268.Fax:+61293855966.E-mail:a.adesina@.au.2CH4+CO2+0.5O2f4H2+3CO(1)CO2reforming of CH4CH4+CO2f2H2+2CO(2)Ind.Eng.Chem.Res.2010,49,10450–10458 1045010.1021/ie100460g 2010American Chemical SocietyPublished on Web06/03/2010The heat of reaction for reactions 1-7at temperature,T ,may be obtained from eqs 8a and 8b:which may be expressed as:The constants in eqs 8a and 8b for reactions 1-7are shown in Table 1and were derived from standard thermodynamic tables.172.Experimental SectionThe catalyst support,γ-alumina (Saint-Gobain Norpro,USA),was first crushed and sieved to 140-425µm before pretreatment at 1073K for 6h.5Co-15Ni/80Al 2O 3was then prepared via sequential impregnation,first with aqueous Co(NO 3)2and then with Ni(NO 3)2(Sigma Aldrich,Australia)on the thermally treated alumina under a constant pH of 2in a Mettler Toledo T90Titrator.Each impregnation step was followed by 3h of stirring at ambient conditions,with subsequent drying for 24h in an oven at 393K.The resulting dried catalysts were calcined at 1073K in air for 5h,at a heating rate of 5K min -1.The calcined solid was then crushed and sieved to 140-250µm,before being activated in situ in the reactor.Multipoint BET surface area and pore volume measurements were obtained from N 2adsorption at 77K using a Quantachrome Autosorb-1unit.A Micromeritics Autochem 2910was used to perform pulse H 2-chemisorption at 383K,as well as to determine basic and acidic character of the catalyst via CO 2and NH 3temperature-programmed desorption (TPD),respectively.CO 2and NH 3were adsorbed at 323and 423K,respectively,before TPD experiments at four heating rates of 10,15,20,and 30K min -1.Also,10%CO 2/He and 10%NH 3/N 2were used as probe gases for the TPD measurements.Powder X-ray Diffraction (XRD)analysis was conducted on a Philips X’Pert system using a Ni-filtered Cu K R radiation (λ)1.542Å)at 40kV and 40mA.The X-ray diffractograms were analyzed using an X’Pert ScorePlus software.Temperature-programmed experiments (calcination and reduction)were performed in a ThermoCahn TherMax 200unitto study the weight change profiles.Temperature-programmed calcination was carried out in 55mL min -1of air and ramped to 973at 5K min -1.Temperature-programmed reduction was conducted in 55mL min -1of 50%H 2/Ar mixture using the same temperature-program scheme.The total carbon content of spent catalysts was determined using a Shimadzu TOC Analyzer 5000A coupled to a solid sample module SSM-5000A.Reaction runs were conducted on a computer-controlled experimental rig consisting of a gas manifold station,a stainless steel fixed-bed reactor (o.d.)6.25mm,i.d.)4.57mm,packed with 0.1g of catalyst)and a TCD-equipped Shimadzu GC-17A gas chromatograph fitted with an Alltech CTR-1column.Gas flow rates were regulated by Brooks Smart mass flow controllers.Prior to each run,the calcined catalyst was reduced in situ in 50mL min -1of 50%H 2/N 2mixture at a temperature ramp of 5K min -1.The catalyst was held for 2h at 1063K which is higher than that during reaction in order to avoid any temperature-induced phase transformations during the actual reaction.Following catalyst activation,the reactor was cooled under a blanket of N 2to the reaction temperature.In order to minimize transport-disguised kinetics,the reactor was operated using a gas-hourly space velocity of 20000h -1over the bed of catalyst particles in the size range 140-250µm.The reaction was conducted at temperatures between 823and 973K and a constant total pressure of 110kPa with CO 2and CH 4partial pressures,P CO 2)P CH 4)20kPa,while the O 2partial pressure varied between 0and 20kPa.N 2was employed as a diluent gas.3.Results and Discussion3.1.Catalyst Characterization.Figure 1a shows the de-rivative weight profile during temperature-programmed cal-cination.The main peak at about 473K corresponds to the decomposition of the metal nitrates to their respective oxides:Table 1.Constants to Calculate ∆H (T )from Equations 8a and 8b reaction R 0×10-3(J mol -1)R 1(J mol -1K -1)R 2×103(J mol -1K -2)R 3×106(J mol -1K -3)R 4×10-5(J K mol -1)1189.98103.5-66.9311.9912.552238.2650.64-33.95 5.99710.483-48.2852.82-32.98 5.997 2.066448.82-15.464 2.2509.6775-416.7628.41-25.56 5.997-3.8336-237.72-13.29 3.2220 1.2607-286.542.1740.9769-8.418Figure 1.Derivative weight profiles of the catalyst during (a)calcination and (b)reduction.∆H (T ))∆H 298+∫298T∆C p d T(8a)∆H (T ))R 0+R 1T +R 2T 2+R 3T 3-R 4T -1(8b)Ni(NO 3)2f NiO +N 2O 5(9)Ind.Eng.Chem.Res.,Vol.49,No.21,201010451while the shoulder at 500K represents the oxidation of NiO and CoO to Co 3O 4and NiCo 2O 4:The broad shoulder at 560K indicates the formation of the metal aluminate phase (NiO/CoO +Al 2O 3f NiAl 2O 4/CoAl 2O 4).The H 2-TPR profile (cf.Figure 1b)implicates the reduction of Co 3O 4and NiO to CoO and Ni (at 435K)and NiCo 2O 4to Ni and CoO,at 600K.The peaks at 740and 973K represent the reduction of CoO to Co and the reduction of the metal aluminates phase,respectively.XRD pattern of the unreduced catalyst shown in Figure 2reveals the existence of multiple oxide phases such as spinel-type NiCo 2O 4(2θ)31.2°),CoAl 2O 4(2θ)36.8°and 59.1°),NiAl 2O 4(2θ)36.8°,44.8°,59.1°,and 65.7°).In comparison,the peak intensities for Co 3O 4(2θ)31.2°and 55.3°)and NiO (2θ)43.2°)phases were significantly smaller,indicat-ing strong metal -support interactions to form aluminates.Table 2summarizes data obtained from N 2physisorption and H 2chemisorption.BET surface area and pore volume obtained were similar to those previously prepared earlier in our laboratory,15,16,18suggesting the conversion of γ-alumina to δ-alumina during pretreatment at 1073K.H 2-chemisorp-tion results (metal dispersion,metal surface area,and active particle size)were consistent with the high metal loading (20wt %)used.Table 3displays the acidic and basicproperties of the catalyst,with Figures 3and 4showing the corresponding NH 3-TPD and CO 2-TPD profiles for the Co -Ni catalyst.NH 3-TPD revealed the two distinct peaks with heats of desorption,∆H d ,of 43.3and 71.0kJ mol -1,respectively,suggesting the presence of multiple acid sites on the catalyst.The low temperature peak could be attributed to a weak Lewis acid site,while the high temperature peak may arise from the presence of strong Lewis and Brønsted acid sites.19,20However,in this catalyst,it is more likely that the high temperature desorption of ammonia represents a Lewis site,since Brønsted acid sites typically exhibit ∆H d greater than 125kJ mol -1.21Similarly,the CO 2-TPD profile (cf.Figure 4)showed that CO 2desorption is characterized by multiple peaks,indicating the presence of weak,inter-mediate,and strong basic sites.22,23Compared to the acid sites,the basic sites are stronger in strength (∆H d )51.3,68.0,and 73.4kJ mol -1);however,the total basic site concentration (1.443µmol m -2)is lower than the total acid site concentration (3.972µmol m -2),indicating a net acidic catalyst.3.2.Reaction Runs.Reactor operating conditions chosen ensured that the rate data collected were free from transport intrusions as confirmed by the satisfaction of the diagnostic criteria provided below in Table4.Preliminary runs using an empty reactor also showed no detectable reaction products at temperatures up to 1023K in the presence of CO 2and CH 4.3.2.1.Dry Reforming.In the first set of runs,rate data fromCH 4dry reforming at different CO 2inlet partial pressures were obtained as shown in Figures 5-7.The CH 4consumption rate increased almost linearly with CO 2partial pressure while theFigure 2.X-ray diffractogram of calcined 5Co-15Ni/Al 2O 3.Table 2.N 2-Physisorption and H 2-Chemisorption ResultsBET area (m 2g -1)110.8pore volume (cm 3g -1)0.4962average pore size (nm)17.91metal dispersion (%)0.5802metal surface area (m 2g -1)0.7758active particle size (nm)173.8Table 3.Acidic and Basic Properties of the Catalystpeak Ipeak II peak IIItotalNH 3heat of desorption,∆H d (kJ mol -1)43.371.0acid site concentration (µmol m -2)1.0792.8933.972CO 2heat of desorption,∆H d (kJ mol -1)51.368.073.4basic site concentration (µmol m -2)0.2020.2091.0321.443Co(NO 3)2f CoO +N 2O 5(10)3CoO +0.5O 2f Co 3O 4(11)NiO +2CoO +0.5O 2f NiCo 2O 4(12)Figure 3.NH 3-TPD profile for Co -Ni catalyst.Figure 4.CO 2-TPD profile for Co -Ni catalyst.10452Ind.Eng.Chem.Res.,Vol.49,No.21,2010CO 2disappearance rate itself initially rose rapidly and seemedto approach a plateau at high P CO 2.The CO 2rate was lower than the CH 4rate at P CO 2below about 25kPa,and the reverse behavior was observed beyond this point.A parallel behavior was observed for the products CO and H 2,respectively,with the crossover slightly lower at about 18kPa.The initial high H 2formation rate is associated with a high CH 4consumption rate suggesting that H 2production probably arose originally from CH 4dehydrogenation with attendant solid carbonaceous productwith the empirical formula C x H 1-x (where 0e x e 1allows the possibility of pure carbon deposition at x )1at one end of the composition spectrum and H adatoms,x )0at the other).Thus,Subsequent interaction between CO 2and the carbonaceous deposit yieldsfrom where it is apparent that in the presence of excess CO 2(CO 2:CH 4>1),the CO 2consumption rate will exceed the CH 4usage rate and,accordingly,the CO production rate will surpass the H 2formation rate.The small difference between the crossover P CO 2value for CO 2/CH 4consumption and H 2/CO production may be due to the fact that not all CO 2adsorbed was converted to CO via eq 14(i.e.,about 72%of the chemisorbed CO 2was used in CO production).Moreover,the possibility for additional CO 2consumption via reoxidation of the metal surface,M,cannot be ruled out,i.e.,CO 2+M f MO +CO.On the other hand,the H 2production rate was less than twice the CH 4consumption rate,suggesting that other products containing H atoms were also formed.These would include the surface carbon pool,C x H 1-x and possibly H 2O arising from the reverse water -gas shift reaction (cf.eq 4although H 2O composition was not measured in this work).Interestingly,totalTable 4.Criteria for the Absence of Transport-Limiting Resistances in Laboratory Reactorsmass transfer heat transfercriteriavalue criteriavalue Mears (external)0.047Mears (external)0.273(-r exp )F b d p n0.3k c C A b<1|(-∆H r )|(-r exp )F b d p E A0.3hT b 2R<1Weisz -Prater (internal)0.078Anderson (internal) 3.49×10-5(-r exp )F b d p 24D eff C A s <1|(-∆H r )|(-r exp )d p 2F b E A3λeff RT s 2<1wall effect 26.9Mears (radial heat effect)0.0052d rd p>10(E a RT w )(|(-∆H r )(-r exp )F b [(1-εbed )/(1-b )d r 2]4λeff T w|)(18+1Bi w d p d r)<0.05Figure 5.CH 4and CO 2consumption rate during dry reforming at 923K.Figure 6.H 2and CO formation rate during dry reforming at 923K.Figure 7.H 2:CO and r CO :r CO 2during dry reforming at 923K.x CH 4f C x H 1-x +(5x -12)H2(13)C x H 1-x +x CO 2f 2x CO +(1-x2)H 2(14)Ind.Eng.Chem.Res.,Vol.49,No.21,201010453organic carbon (TOC)analysis of the used catalysts (after 4hruns)revealed the presence of a significant TOC concentration for P CO 2<40kPa,as seen in Table 5.The global kinetics for carbon deposition during dry reforming based on the data in Table 5was obtained asIn particular,on the basis of eqs 13and 14,the ratio ofCO formation rate to CO 2consumption rate,r CO :r CO 2,expected is 2.Figure 7shows that this is practically true for the present study.Additionally,the product ratio,H 2:CO,experienced a nonlinear drop with increased P CO 2which is captured byand predicts the asymptotic values of S H 2:CO )∞for P CO 2)0(conditions for direct CH 4dehydrogenation)and S H 2:CO )0.5for high P CO 2(maximum of 90kPa with P CH 4)20kPa when there is no N 2diluent).3.2.2.Effect of O 2Addition during Dry Reforming.Figure 8a and b show the CH 4and CO 2rate profiles for oxidative dry reforming for different inlet O 2partial pressures using P CH 4)P CO 2)20kPa.O 2was fully consumed under all the reaction conditions studied and its presence increased the CH 4consumption rate significantly.However,apparent CO 2consumption decreased such that a net production of CO 2was registered at higher O 2partial pressures.In particular,the CH 4reaction rate seemed to have a first-order dependence on P O 2at low temperature but changed to a nonlinear functionality beyond 873K.At T >873K,an increase in CH 4consumption rate started to level off at P O 2greater than10kPa,due to almost complete conversion (>98%)of CH 4in the feed.Clearly,the CH 4consumption under oxidative reforming is a combination of the dry reforming and catalytic CH 4oxidation rates along with possible homogeneous gas phase combustion.However,Figure 8b shows that the CO 2consumption rate dropped relatively rapidly with P O 2signify-ing that CO 2interaction with CH 4became increasingly less important and,indeed,ceased to be relevant at P O 2>8kPa (depending on temperature),i.e.when CO 2consumption is zero.Significantly,CO 2became a reaction product beyond this critical point as seen from the negative values in Figure 8b.The H 2production rate appeared to go through a maximum depending on temperature (cf.Figure 9a).The initial increase in H 2formation above that due to pure dry reforming was probably due to additional oxidation of the accompanying carbon residue,C x H 1-x (formed from dry reforming),namely;However,as P O 2increased,H 2oxidation (cf.eq 6)would become dominant leading to a decrease in overall H 2production and,hence,the appearance of a maximum in H 2rate versus P O 2profile.Interestingly,the experimental P O 2corresponding to the maximum H 2rate,P O 2,max shifted to lower values with increasing temperature following an Arrhenius-type dependency as shown in Figure 9b,thusorwhere the slope and intercept of the plot are characteristic parameters of the reaction network.Figure 9c shows that the CO formation rate is at its height during dry reforming but decreased with increasing P O 2during oxidative reforming,indicating that any CO produced from dry reforming and eq 17would not survive in an O 2environment.As a result of these observations,a general empirical relation that can capture the essential features for the rate of eachTable 5.TOC during Dry Reforming at Varying P CO 2(P CH 4)20kPa)P CO 2(kPa)TOC (%)Pro-rata TOC (g TOC g cat -1h -1)1540.95a 0.40952040.630.10162518.250.04563016.950.0424400050aRun stopped after 1h due to excessive pressure drop.r carbondeposition )4×108P CO 2-7.36(at a constant P CH 4of 20kPa)(15)S H 2:CO )3.473P CO 20.44(16)Figure 8.(a)CH 4and (b)CO 2consumption rates during oxidative dry reforming.C x H 1-x +x2O 2f x CO +(1-x 2)H 2(17)P O 2,max )k max e -γ/T(18)ln(P O 2,max ))ln k max -γ(1T)(19)10454Ind.Eng.Chem.Res.,Vol.49,No.21,2010participating species during oxidative dry reforming may be written:where a 0represents the dry reforming contribution,a 1is the pseudo-rate constant due to O 2enhancement s the removal of surface carbon residue accompanying dry reforming (initial linear increase,via eq 17),while a 2stands for the apparent rate constant for the direct total oxidation of the species (detrimental role of O 2)with b th-order kinetics.Given the temperature-dependency of these parameters,eq 20is re-writen asParameter estimates using the data at all four temperatures and the six different P O 2values were obtained via nonlinear regression (Polymath 6.0)as summarized in Table 6.We note that the activation energy,E 0,for H 2,CO,and CH 4are nearly the same as that obtained for the pure dry reforming runs (E DR ).The similarity in E 0values for all species suggests is symp-tomatic of similar rate-determining steps in formation of both CO and H 2during pure dry reforming.However,during oxidative dry reforming,the oxidation of C x H 1-x species appear to be relatively facile since E 1values are low compared to E 2estimates.Although H2production declined at high O 2partial pressure,Figure 10shows that the H 2:CO ratio continued to increase,suggesting that the rate of CO oxidation is more rapid than that of H 2.At lower temperatures,catalytic CH 4combustion ap-peared to be the dominant reaction,producing CO 2and H 2O,while at higher temperatures,the endothermic CH 4dry reform-ing reaction became increasingly favored.24The decrease in H 2:CO ratio with increased temperature at a fixed P O 2arose from increased CO production as CH 4dry reforming became impor-tant.Stoichiometric CH 4dry reforming produces H 2:CO ratios somewhat lower than unity,due to the reverse water -gas shift reaction (eq 4).Thus,the formation of total oxidation products (H 2O from H 2and CO 2from CO/CH 4)may be manipulated by temper-ature and indeed P O 2in view of the difference in total oxidation kinetics for each species (cf.b -values in Table 6).This is a significant tool for H 2:CO ratio control if the productFigure 9.(a)H 2production rates,(b)relationship between ln P O 2,max and 1/T ,and (c)CO production rates during oxidative dry reforming.r i )a 0+a 1P O 2-a 2P O 2b(20)r i )k 0e -E 0/RT +k 1e -E 1/RT P O 2-k 2e -E 2/RT P O 2b(21)Table 6.Parameter Estimates for the Empirical ModelspeciesparameterH 2CO CH 4CO 2k o ×103(mol g cat -1s -1)17.413.9 4.67 2.78E 0(kJ mol -1)36.433.830.227.1k 1×105(mol g cat -1s -1)33.40.923E 1(kJ mol -1)27.77.39k 2×103(mol g cat -1s -1) 4.52 2.2914750.0123E 2(kJ mol -1)57.864.3131 6.41b1.43 1.572 1.05E DR (kJ mol -1)38.436.330.934.0Ind.Eng.Chem.Res.,Vol.49,No.21,201010455gas were to be used for downstream gas-to-liquid(GTL) conversion and in fact attractive,since O2was completely consumed under all conditions and would not need to be removed from the syngas mixture(O2is an inhibitor for Fischer-Tropsch reaction).The adequacy of the model given by eq21is confirmed by the relatively higher R2-values obtained(R2>0.95)and the pattern of the residual plot as demonstrated in Figure11.In particular,this model predictsthe temperature dependency of P O2,max.At the maximum rate,eq21requires from whence and consequently,where the intercept(-6.9)in Figure9b is(1/(b-1))ln(k1/ (k2b))and slope(8420)is(E1-E2)/((1-b)R).The agreement between the parameter estimates in Figure9b and data in Table6also lends credence to the reliability of eq21.Table7shows the heat of reaction(∆H r)for the stoichio-metric CO2reforming and partial oxidation(as calculated from standard thermochemical tables s17cf.eq8b),and Figure12 plots the experimental∆H r values(based on energy balance with the actual feed and product compositions).It is evident that the calculated∆H r for CH4dry reforming,which varies between259.4and260.1kJ mol-1,is in agreement with the experimental counterpart(221.7-240.1kJ mol-1).With the addition of O2,heat requirement for the reaction decreased such that the reaction eventually became exothermic at P O2>8kPa as may be seen from Figure12.At P O2)20kPa,the reaction was strongly exothermic,with-∆H r greater than135kJ mol-1 and significantly larger than the theoretical value for partial oxidation of methane(which is between22.8and24kJ mol-1). This suggests that besides methane partial oxidation,other oxidation reactions such as H2and CO oxidation were also taking place concurrently.Figure12also shows that that the difference in∆H r was small for runs carried out between873 and973K with runs performed at823K being substantially more exothermic.This is because dry reforming is thermody-namically unfavorable at low temperature(∆G823K)26kJ mol-1)while CH4combustion seemed to dominate,resulting in a large difference in∆H r for T)823K.Importantly,the thermal behavior of the reactor system may be controlled by varying the amount of O2cofed to the reactor.3.2.3.Mechanistic Inferences.Spectroscopic and pulse reaction techniques have shown that reactants involved in oxidative dry reforming may be adsorbed on transition metals.25-29Osaki and co-workers,25,26using pulse surface reaction analysis of CH4dry reforming on various supported Ni catalysts,found that adsorbed CH y is produced via sequentialFigure10.Influence of O2partial pressure on H2:CO ratio.Figure11.Residual plot for H2production rate during oxidative dry reforming.d rH2d PO2)0)a1-a2bPO2b-1(22)PO2,max)(a1a2b)1/(b-1))(k1k2b)1/(b-1)e-(E1-E2)/(b-1)RT(23)ln(PO2,max ))1b-1ln(k1k2b)+(E1-E2)(1-b)R1T(24)puted Heat of Reaction for CH4Dry Reforming andPartial Oxidation∆H r(kJ mol-1)temperature(K)CH4+CO2f2H2+2CO CH4+0.5O2f2H2+CO823259.4-24.0873259.8-23.5923260.0-23.1973260.1-22.8Figure12.Experimental heat of reaction as a function of P O2duringoxidative dry reforming of CH4.10456Ind.Eng.Chem.Res.,Vol.49,No.21,2010elimination of hydrogen atoms (where y varied between 1and 2.7in their study),eventually leading to surface carbon,suggesting that hydrocarbon chemisorption is dehydrogenative in nature.Stevens and Chuang,27via combined in situ infrared (IR)spectroscopic and mass spectrometric study,proposed that the first step of CH 4decomposition on Rh/Al 2O 3is decomposi-tion of CH 4into CH y and H 2.Erdo¨helyi et al.28,29studied CH 4and CO 2activation over supported Pd and Rh using IR methods and found that CO 2adsorption was dissociative in nature as confirmed by the presence of adsorbed CO-band signals (wavenumber 1750-2080cm -1).We therefore propose that the sequence of elementary steps describing oxidative CH 4-CO 2reforming proceeds as follows:where X denotes an active site.On the basis of this mechanism,the Langmuir -Hinshelwood kinetic model (for constant P CH 4and P CO 2)for oxidative CH 4dry reforming was derived as in which E a,rxn )49.7kJ mol -1and -∆H ads,O 2)112kJ mol -1are associated with k rxn (kinetic constant)and K O 2(equilibrium adsorption constant for O 2),respectively.3.2.4.Carbon Content Analysis.Table 8displays the total carbon content of the spent catalysts following 4h of CH 4dry reforming,in the absence of O 2in the feed.It seems that carbon lay-down from CH 4dehydrogenation was removed via CO 2gasification (cf.eq 14)and appeared to be more facile at high temperatures.The decrease in total carbon content with tem-perature suggests that carbon gasification rate was probably higher than CH 4dehydrogenation as temperature increased.Arrhenius treatment of carbon deposition gave an apparent activation energy of -50.9kJ mol -1.This behavior is indeedtypical of carbon formation from hydrocarbons at the reaction temperatures studied,in which negative activation energy is expected at temperatures between 823and 923K where carbon deposition decreased with increasing temperature,while,out of this range,a positive activation energy is obtained.30Previous work on propane steam reforming in our group have also reported similar observations in the deactivation rate coef-ficients.18Nakano et al.31have also reported negative activation energy during CO disproportionation and attributed it to the greater adsorption energy of CO than the dissociation bining this analysis with the eq 15,leads toas the kinetic expression for carbon deposition for CH 4dry reforming between 823to 973K.At all the temperatures studied,addition of even the lowest amount of O 2(5kPa)gave a used catalyst with no detectable carbon,showing that O 2cofeeding during dry reforming may indeed suppress carbon deposition.4.ConclusionsThis study has demonstrated the beneficial effects of cofeed-ing small amounts of O 2during CO 2reforming of CH 4.O 2addition resulted in improved CH 4reaction rate with attendant negligible coke deposition.With O 2addition,H 2:CO ratio increased from less than 1for stoichiometric CO 2reforming to between 1and 2,depending on reaction conditions,showing that product syngas composition may be manipulated to one more suited to downstream processes such as the Fischer -Tropsch synthesis.The increase in overall reaction exothermicity shows that by managing the amount of O 2added,the thermal behavior of the reactor may be controlled.Moreover,an empirical relation describing the oxygen partial pressure for optimum H 2production as a function of temperature was validated by our data.AcknowledgmentThe authors are grateful to the Australian Research Council for financial support.SYF and CKC are recipients of the Australian Postgraduate Award and University International Postgraduate Award scholarships respectively.Nomenclatureb )inert volume fraction of bed Bi w )wall Biot numberC A s )surface concentration of reactant (mol m -3)d p )catalyst particle diameter (m)d r )reactor tube diameter (m)D eff )effective diffusivity (m 2s -1)E a )activation energy (kJ mol -1)E DR )activation energy for pure dry reforming (kJ mol -1)∆H d )heat of desorption (kJ mol -1)∆H r )heat of reaction (kJ mol -1)k c )mass transfer coefficient (m s -1)R )ideal gas constant (J mol -1K -1)(-r exp ))reaction rate (mol s -1g cat -1)T b )bulk gas-phase temperature (K)T s )particle temperature (K)T w )wall temperature (K)Greek LettersF b )catalyst particle bulk density (kg m -3)λeff )effective bed thermal conductivity (W m -1K -1)εbed )bed voidageTable 8.TOC (%)of Spent CO 2Reforming Catalysts at Various Temperatures,without O 2in Feed (P CH 4)P CO 2)20kPa)temperature (K)TOC (%)82358.687357.592340.697317.8CH 4+X T CH 3-X +H-X (25a)CH 3-X +X f CH 2-X +H-X lCH-X +X f C-X +H-X (25b)CO 2+2X T CO-X +O-X(25c)O 2+2X T 2O-X(25d)CH y -X +O-X f CH y -1O-X +H-X1e y e 3(25e)C-X +O-X f CO-X +X(25f)2H-X T H 2+2X (25g)CO-X T CO +X (25h)H-X +O-X f OH-X (25i)H-X +OH-X T H 2O +2X(25j)r CH 4)k rxn P O 21/2(1+K O 21/2P O 21/2)2(with eq 25e as the rate-determining step)(26)r carbondeposition )5.265×105e 6127.4/T P CO 2-7.36(27)Ind.Eng.Chem.Res.,Vol.49,No.21,201010457。