9272
Ind. Eng. Chem. Res. 2008, 47, 9272–9278
Dual Active-Site Mechanism for Dry Methane Reforming over Ni/La2O3 Produced from LaNiO3 Perovskite Germa´n Sierra Gallego,*,† Catherine Batiot-Dupeyrat,‡ Joe¨l Barrault,‡ and Fanor Mondrago´n† Instituto de Quı´mica, UniVersidad de Antioquia, A.A. 1226, Medellı´n, Colombia, and Laboratoire de Catalyse en Chimie Organique, UMR CNRS 6503, UniVersite´ de Poitiers, Ecole Supe´rieure d′Inge´nieurs de Poitiers, 40, AVenue du Recteur Pineau, 86022 Poitiers Cedex, France
The kinetic behavior of the Ni/La2O3 catalyst obtained from the LaNiO3 perovskite in the reforming reaction of methane with carbon dioxide was investigated as function of temperature and CH4 and CO2 partial pressure. A rate reaction equation was developed including for the first time the presence of two active sites in this type of catalyst: the metallic nickel particles and the La2O3 which is the metal support. This kinetic model fits very well the experimental data, and the rate expression predicts better the rate of methane conversion than the models which incorporate only the metallic cluster as the active sites. X-ray photoelectron spectroscopy (XPS) of the catalyst after reaction suggests that Ni particles are partially covered by La2O2CO3 species which are formed by interaction of La2O3 with CO2. Catalytic activity occurs at the Ni-La2O2CO3 interface, while the oxycarbonate species participate directly by reacting with deposited carbon, thus restoring the activity of the Ni sites at the interface. 1. Introduction Methane reforming using carbon dioxide (dry reforming) has been of interest for a long time.1 Particularly in recent years its importance has grown due to both environmental and commercial reasons. This process offers important advantages compared to steam reforming of methane. Namely, (a) it yields lower H2/CO product ratios, which are preferable feeds for Fischer-Tropsch plants for synthesis of oxo-alcohols, acetic acid, and dimethyl ether;2 (b) it reduces CO2 and CH4 emissions, which are both greenhouse gases; and (c) it is well-suited for chemical energy transmission systems.3 The reaction has been investigated over both, noble metal4,5 and Ni-based6,7 catalysts dispersed on different carriers. The main drawback of the CH4-CO2 reforming reaction is the large thermodynamic potential for coke formation.8 In this sense, efforts to alleviate this problem have been centered to find suitable materials less sensitive to deactivation by carbon deposition. Kinetic and mechanistic studies have been performed over noble metal and Ni-based catalysts. Ross and co-workers9 observed that both methane and carbon dioxide dissociate independently of one another over Pt/ZrO2. Verykios et al10 using Ni/La2O3 detected the presence of La2O2CO3 in the used catalysts and assigned a precise role to this compound in the reaction mechanism. They developed a kinetic equation for the dry reforming of methane over Ni/La2O3 and reported a fractional order dependency of the rate of reactant disappearance for both CH4 and CO2 partial pressure. However, this kinetic equation was developed considering only one type of active site, the metallic particles (Ni°). This equation is similar to the one reported by Mu´nera et al.11 In this work, kinetic measurements over the Ni/La2O3 catalyst obtained from the LaNiO3 perovskites structure were performed. * To whom correspondence should be addressed. E-mail: geasierraga@ unalmed.edu.co. Mailing address: Escuela de Ingenierı´a de Materiales, Universidad Nacional de Colombia, Minas, Sede Medellı´n, Columbia. Tel.: 57-4-4255286. † Universidad de Antioquia. ‡ Universite´ de Poitiers.
The catalyst was characterized using various instrumental techniques, and the kinetic measurements were carried out under differential conditions in the temperature range 923-1020 K, employing dilute CH4/CO2/He feed. The total feed flow rate was 300 ml/min, while the catalyst mass was in the range 2.0-10 mg. The goal of this study was to obtain a consistent kinetic equation considering two active sites: the metallic particles and the support. 2. Experimental Details 2.1. Catalyst Preparation. LaNiO3 perovskite was prepared by the self combustion method.12,13 Glycine (H2NCH2CO2H) used as ignition promoter, was added to an aqueous solution of metal nitrates with required stoichiometry to get a NO3-/NH2 ) 1 ratio. The resulting solution was slowly evaporated until a vitreous green gel was obtained. The gel was heated up to around 520 K, the temperature at which the ignition reaction occurs yielding the formation of a powdered precursor which still contains carbon residues. Calcination at 973 K for 6 hours eliminates all of the remaining carbon and leads to the formation of the perovskite structure. 2.2. Characterization. The catalysts were characterized by powder X-ray diffraction using a Siemens D-5000 diffractometer with CuKR1 ) 1.5406 and CuKR2 )1.5439 Å, operated at 40 kV and 30 mA. The diffraction patterns were recorded in the 2θ values range 10-90° with a step size of 0.01° and 1 s per step. Temperature programmed reduction (TPR) experiments were carried out in a Micromeritics Autochem 2910 using about 160 mg of catalyst. TPR experiments were performed using a 5% H2/Ar flow while the temperature was risen at 5 K/min from ambient to 973 K and kept at this temperature for 2 h. The XPS data were taken with an SSI X-Probe (SSX-100/ 206) photoelectron spectrometer, equipped with a monochromatized microfocus Al X-ray source. La 3d, Ni 3p, O 1s, and C 1s signals were measured. Transmission electron microscopy (TEM) was carried out on a Philips CM120 instrument, with LaB6 filament and equipped with an energy dispersive X-ray analyzer (EDX).
10.1021/ie800281t CCC: $40.75 2008 American Chemical Society Published on Web 11/01/2008
Ind. Eng. Chem. Res., Vol. 47, No. 23, 2008 9273
Surface areas were measured by the BET procedure using nitrogen adsorption. All samples were degassed with He for 30 minutes at 623 K before measurement, the adsorption-desorption isotherm of N2 was measured using 30% N2/He as the adsorbate on a Micromeritics Flowsorb II 2300 apparatus at 77 K. 2.3. Kinetic Measurements. Kinetic studies under differential conditions and under integral reaction conditions were conducted in a conventional apparatus consisting of a flow measuring and control system and a fixed-bed reactor (6 mm i.d.). The feed and reaction products were analyzed by an online mass spectrometer. The reaction was carried out by passing a continuous flow of CH4/CO2/He and the mass of catalyst used was 2-10 mg. The temperature was increased from room temperature to 973 K at a rate of 5 K/min and maintained at this temperature for the desired reaction time. Conversions were usually controlled to be significantly lower than those defined by thermodynamic equilibrium by adjusting the total flow rate (100-300 mL/min). Rate limitation by external and/or internal mass transfer under differential conditions proved to be negligible by applying suitable experimental criteria.
Figure 1. XRD patterns of (a) JCPDF Card No. 88-0633, (b) LaNiO3 after calcination for 6 h at 973 K, (c) after reduction under hydrogen at 973 K, (d) after 15 h of reaction using the reduced material: (O) Ni°, (3) La2O3, (9) La2O2CO3 hexagonal.
3. Results and Discussions 3.1. X-ray Diffraction Characterization of the Catalyst. The X-ray diffraction (XRD) patterns obtained for LaNiO3 are shown in Figure 1. After calcination at 973 K, the only presence of the LaNiO3 perovskite structure with a rhombohedral symmetry (Figure 1b) was observed. After the reduction treatment under hydrogen at 973 K, the perovskite structure was completely destroyed the only phases detected being Ni° and La2O3 (Figures 1c). After 15 h of CO2/CH4 reforming, the presence of lanthanum oxocarbonate La2O2CO3 was observed with starting catalysts LaNiO3 (Figures 1d). 3.2. Temperature Programmed Reduction. The TPR profile of LaNiO3 is shown in Figure 2. The reduction of LaNiO3 proceeds in three steps; the successive changes of the perovskite structure were determined by an in-situ XRD measurement. The results were reported in a previous paper:14
Figure 2. TPR curve of the perovskite LaNiO3: catalyst weight 160 mg, gas flow rate 5% H2/Ar, 50 mL/min.
4LaNiO3+2H2 f La4Ni3O10+Ni° + 2H2O La4Ni3O10+3H2 f La2NiO4 + 2Ni° + La2O3+3H2O La2NiO4 + H2 f Ni° + La2O3 + H2O From the area under the curves, the experimental amount of hydrogen consumed during the reduction was determined. The expected value was determined from the stoichiometric formula LaNiO3. The experimental value was found to be lower than the theoretical one. That difference can be explained assuming a deviation from the oxygen stoichiometry, according to the work of Wachowski et al.15 Thus, using the amount of hydrogen consumed, the formula of the perovskite structure was recalculated as La+3Ni0.373+Ni0.632+O2.68. The TPR indicates that reduction of LaNiO3 into La2O3 and metallic Ni° is completed at around 883 K, with all of the nickel inside the structure reduced to Ni°. The perovskite LaNiO3 after TPR in hydrogen was reoxidized in air at 1073 K, and a second TPR was carried out. The two TPR profiles were very close, indicating that the oxidation-reduction process was reversible. To check this point, XRD spectra of re-oxidized samples in air at 1073 K after H2 TPR was recorded. As shown in Figure 3, it appears that after the reoxidation treatment the perovskite phase is restored with a higher symmetry as compared to that of the starting material.
Figure 3. Powder X-ray diffraction patterns for (a) La2O3 JCPDF Card No. 330711, (b) LaNiO3 after TPR experiment up to 1073 K, (c) LaNiO3 JCPDF Card No. 050602, and (d) LaNiO3 reduced after reoxidation in air at 1073 K for 1 h: (°) Ni°.
3.3. Characterization by Electron Microscopy. TEM analyses were performed after reduction of the perovskite and then after 15 hours of reaction using the reduced material. A
9274 Ind. Eng. Chem. Res., Vol. 47, No. 23, 2008
Figure 4. Ni particle size distribution determined from TEM micrographs of LaNiO3 reduced under hydrogen. Table 1. Surface Composition of the LaNiO3-δ Catalyst after Reduction Treatment and after Reaction from XPS Signals surface concentration, atom % catalyst
La
9.3 LaNiO3 calcined 8.6 LaNiO3 reduced after 15 h of reaction 8.2 using reduced LaNiO3-δ
O 42.4 39.5 38.3
C 41.8 49.0 51.1
Ni 6.45 3.0 2.4
La/Ni 1.4 2.9 3.4
Figure 5. (a) Effect of volumetric feed flow rate on the rate of methane consumption during reforming reaction over reduce LaNiO3 at constant GHSV) 3 × 105 mL g-1 h-1. (b) Effect of catalyst particle size on the reaction rate observed over reduce LaNiO3 at constant space velocity.
Ni state 2+
Ni /Ni3+ 100% Ni° zero 37% oxidized Ni
histogram of the nickel particle size distribution obtained from the micrographs is shown in Figure 4 for about 500 particles. The diameter of the nickel particles is between 2 and 50 nm, most of the particles being smaller than 20 nm. On the basis of the TEM micrographs, the average size of metallic particles was calculated using the following equation: d ) ∑inidi3/∑inidi2. Where, ni is the number of particles and di is the characteristic particle diameter. The average particles size is around 15 nm after reduction treatment and 17 nm after 15 h of reaction. 3.4. Surface Composition (XPS). The chemical state of the elements and surface composition of the LaNiO3 were studied using X-ray photoelectron spectroscopy. The surface composition of LaNiO3-δ catalyst precursor after reduction treatment and after reaction from XPS signals spectra are reported in Table 1. The carbon content may arise from atmospheric contamination during transfer to the XPS instrument. The surface carbonate species are expected to be formed from the interaction of CO2 of the atmosphere.16 The La/Ni ratio higher than that expected could be a consequence of the preferential presence of lanthanum oxide at the surface of the catalyst as observed in previous reports.17,18 From that surface modification, strong metal-support interaction (SMSI) can be observed as it was mentioned in 1978 to describe the drastic changes in the chemisorption properties of group VIII noble metals when supported on titanium oxide.19 In those cases, the traditional view of metal particles without interaction with the oxide surface of the support is not correct. Indeed, the oxide support can migrate over the metal particles surface, decreasing the metallic surface area and forming new metal-support interactions with modified catalytic properties. In the LaMO3 perovskites, the surface changes were due to the basicity of the La3+ cations which easily react with atmospheric CO2 and H2O.20,21 Finally after the reduction of LaNiO3-δ, metallic nickel particles were observed even at the catalyst surface (as shown by the measurement of the nickel state), whereas during the reaction nickel particles are is partially oxidized (37%) most probably as NiO. After the reaction, the ratio La/Ni is increased, because of the facile formation of surface lanthanum oxycarbonate species which partly encapsulated the nickel particles and prevented the nickel detection. 3.5. Definition of Reaction Conditions. Preliminary experiments were carried out to find suitable conditions in which internal and external mass transfer effects are not predominant.
Figure 6. Conversion of (2) CH4, (9) CO2, and (×) molar ratio H2/CO as a function of time on stream over Ni/La2O3 catalysts originated from LaNiO3 perovskite at 973 K.
The effects of interface and intraparticle heat and mass transport resistances on conversions were determined employing theoretical and experimental procedures.22 A diagnostic test was applied to check the presence of interphase limitations23 by changing both reactant flow rate (F) and catalyst volume (V), while keeping constant their ratio (GHSV ) 3 × 105 mL g-1 h-1). The results are presented in Figure 5. A useful diagnostic test to check for the presence of intraphase concentration gradient consists of determining the isothermal conversion for particles of different size at constant space velocity.23 If the conversion varies by decreasing the particle size, intraphase mass transfer is limiting, whereas a constant conversion indicates that the system is under chemical kinetic control. As shown in Figure 5a, the methane conversion was found to be independent of the gas velocity when the gas flow rate was higher than 100 mL/min, indicating the absence of external mass transfer effects. Varying the average size of catalysts particles (up to 355 µm) Figure 5b, no intraparticle diffusion limitation was observed in the range of conditions studied. Therefore, all reactions were performed using a total flow rate of 100 mL/min whereas the catalyst grain size was kept at 180 µm. 3.6. Stability and Catalytic Activity of Ni°/La2O3 Originated from the LaNiO3 Perovskite. The methane dry reforming reaction using a feed of CH4/CO2/He ) 10/10/80 mL/min (GHSV ) 1.2 × 105 mL h-1 g-1) was carried out at 973 K. Figure 6 presents the conversion of CH4, CO2 and the molar ratio H2/CO with time on stream under integral reaction conditions over Ni/La2O3 originated from the LaNiO3 perovskite. The catalyst thus obtained is highly active, the CH4 and CO2 conversions being about 90% with a molar ratio H2/CO close to 1 during 100 h of reaction. Moreover, no side reactions were observed as carbon deposition or RWGS commonly observed when using lanthanum-nickel based catalyst.24,25 These results suggest that, under these experimental conditions, the thermo-
Ind. Eng. Chem. Res., Vol. 47, No. 23, 2008 9275
Figure 7. Influence of contact time on conversion of (a) CH4 and (b) CO2 over (]) Ni°/La2O3 originated from the LaNiO3 perovskite and (2) Ni°/ La2O3 reported by Verykios. The reaction conditions are as follows: CH4/ CO2/He ) 10/10/80 mL/min, 50 mg catalyst, PT ) 1atm and T ) 973 K.
dynamic equilibrium was reached according to the value of the standard free energy calculated at 973 K (∆G973K° ) -16 kJ mol-1).26 Such a high stability was already observed for Ni/La2O3 catalyst by Zhang and co-workers but only for low conversion (about 30%), a significant deactivation occurring at higher conversion.27 After reaction, the only phases detected by XRD are metallic nickel and the hexagonal phase lanthanum oxycarbonate: La2O2CO3, resulting from the CO2 adsorption and reaction over La2O3 during the reaction (Figure 1d). It is suggested that the high performance of the Ni/La2O3 catalyst is the result of (i) the high activity of Ni° particles for the CH4 decomposition, (ii) the reactivity of La2O3 particles with CO2 to form La2O2CO3 carbonate, and (iii) the facile reaction of that carbonate with the carbon formed at the interface of the Ni particles with the support, thus restoring the original catalytic surface. If the rate of carbon deposition is equal or lower than the rate of carbon scavenging, a stable catalyst is thus obtained.28 For the above reasons, the present kinetic study took into account the two active sites: the Ni° and the La2O3 particles which respectively activate the CH4 and CO2 molecules. 3.7. Influence of the Space Velocity. The influence of contact time on methane and carbon dioxide conversions over Ni/La2O3 catalyst was investigated at 973 K. The feed consisted of CH4/CO2/He ) 10/10/80 mL/min. The alteration of contact time was realized by adjusting both the amount of catalyst (2-50 mg) and the feed rate. As shown in Figure 7, both methane and carbon dioxide conversion decreased rapidly as space velocity increases from 1.2 × 105 to 1.2 × 106 mL g-1 h-1. Comparison with the information reported by Verykios et al,27 the Ni°/La2O3 originated from the LaNiO3 perovskite exhibits a higher catalytic performance than that obtained from 17% Ni/ La2O3 prepared by impregnation, especially at space velocities higher than 2.0 × 105 mL g-1 h-1. Verykios and Zhang reported CH4 and CO2 conversions of 43% and 56%, respectively, at a space velocity of 3.0 × 105 mL g-1 h-1 (contact time ) 0.01 g s mL-1), whereas the CH4 and CO2 conversion obtained in this work are 74% and 83%, respectively, at the same GHSV. 3.8. Apparent Activation Energies. The variation of CH4 and CO2 reaction rates in the range of temperatures between 400 and 973 K was studied. The apparent activation energies of CH4 and CO2 over the Ni/La2O3 catalyst coming from the LaNiO3 perovskites was determined by using the Arrhenius plot of ln rCH4 against 1/T. The results are shown in Figure 8. The apparent activation energies of CH4 and CO2 were 68 and 77 kJ/mol, respectively. These values are lower than the
Figure 8. Influence of the reaction temperature on the reaction rates of the CH4 and CO2. Table 2. Comparison of the Activation Energies Reported by Other Authors Using Nickel Based Catalyst in Dry Reforming of CH430-32,43 Ea (kJ/mol) catalyst
temperature (K)
Ni°/La2O3 from LaNiO3 6.3%-Ni/C 6.8%-Ni/SiO2 10%-Ni/MgO 1.2%-Ni/TiO2 7 wt% Ni/MgO-A Ni/Al2O3 Ni/CaO–Al2O3 0.5 % Rh/Al2O3 5 % Ni/CaO–Al2O3 Mo2C Ce0.9Gd0.1O2-x
773–973 673–823 723–823 723–823 723–823 873 773–973 893–973 773–923 893–963 1073–1173 923–1250
CH4
CO2
68.0 121.4 96.3 92.1 108.9 105.0 50.9 106.8 95.0 107.0 172.0 165.0
77.0 92.1 79.6 87.9 87.9 56.1 98.8
respective activation energies reported for noble metal catalysts, which are in the range 83-125 kJ/mol.29 Table 2 shows a comparison of the activation energies reported by using nickel based catalyst for dry reforming of CH4. In Table 2, we can observe a difference in activation energies between our results and those reported by other authors. These differences may be due to a higher dispersion of the Ni° on La2O3 obtained after reduction of the LaNiO3 perovskite. 3.9. Effect of CH4 and CO2 Partial Pressures on Intrinsic Rate. The influence of the partial pressures of CH4 and CO2 on the reaction rate of CO2 reforming of methane is depicted in Figure 9. This study was performed at atmospheric pressure at 973 K under differential conditions. The measurements were made maintaining the partial pressure of one reactant constant at 10 kPa and varying the other reactant pressure between 0 and 30 kPa. The reforming reaction was carried out for 5 h until the catalyst reached steady state. After that, the variation of CH4 or CO2 partial pressure was conducted using fresh catalyst samples in each case. In Figure 9a, we can observe that in the low methane partial pressure (0-10 kPa) domain the reaction rate is strongly affected by the methane partial pressure. From about 10 kPa, the increase
9276 Ind. Eng. Chem. Res., Vol. 47, No. 23, 2008
Figure 9. Dependence of partial pressures of (a) CH4 and (b) CO2 at constant PCO2 and PCH4 of 10 kPa, respectively, on the reaction rate over Ni/La2O3 catalyst at 973 K. The total pressure is 100 kPa.
Figure 10. Reaction order plots obtained using methane forward rates measured at 973 K.
of CH4 partial pressure up to 30 kPa does not seem to affect significantly the rate of reaction. The effect of the CO2 partial pressure on the rate of methane reaction is shown in Figure 9b. The most important effect is observed in the range 0-5 kPa. However, pressures higher than about 5 kPa CO2 do not induce any change in the rate of methane conversion. Comparison of the curves presented in Figure 9a and b suggests that the reaction rate is more susceptible to CO2 partial pressure, than to CH4 partial pressure at low values of the CO2 and CH4 partial pressures. This result is in agreement whit the stronger CO2 interaction than that of CH4. CO2 adsorbs on the La2O3 component of the support to form the carbonate species La2O2CO3 as discused in section 3.1 (see Figure 1). It was suggested that these carbonate species play an important role in the kinetic mechanism and the stability of the Ni/La2O3 catalyst.33 The reaction orders with respect to both reactants were calculated using ln rCH4 versus ln PCH4 or PCO2 plots (Figure 10). From the slopes of the two straight lines, the resulting law rate equation can be described as follows: PCH4 ) kPCH40.68PCO20.25
where PCH4 or PCO2 are the pressures of reactants and products and K is the velocity constant at the corresponding reaction temperature. The superscripts 0.68 and 0.25 are the partial reaction orders with respect to CH4 and CO2, respectively. For Ni based catalysts, the majority of the reported reaction orders are in the range of 0.55-1.1 with respect to CH4 partial pressure while it is 0-0.55 with respect to CO2 partial pressure,34,35 but other authors have reported negative reaction orders for both CH4 and CO2 partial pressures.36 Bhat and Sachtler37 reported that the CO2 reforming of CH4 is first-order in CH4 and zero-order in CO2 on NaY catalyst, indicating that C-H bond activation is the kinetically relevant step in CH4 reforming reactions and hydrogen desorption to form H2 and that the CO2 reactions with chemisorbed species derived from CH4 to form CO are fast and kinetically irrelevant. From the above information, it can be suggested that the reaction orders for the CO2 reforming of methane are controversial. These conflicts might be due to differences in reaction conditions, catalysts preparation, differences in the partial pressure range examined or originated from the experimental error of the data. It has been also reported that the addition of alkali or alkaline-earth metal oxides to nickel supported on alumina changed the reaction order from negative to positive with respect to the partial pressure of CO2.38 In our case, the partial reaction order with respect to CH4 (0.68) is higher that the reaction order with respect to CO2 (0.25). This result suggests that CO2 is more strongly adsorbed than CH4 on La2O3. This is consistent with the very strong interaction between CO2 and La2O3 revealed by the formation of oxycarbonates during the increasing of the temperature with a CH4/CO2/He feed (see Figure 1). 3.10. Reaction Mechanism. One of the first publications about the kinetics of CO2 reforming of CH4 was in 1967, reported by Bodrov and Apel′baum,39 using Ni foil as the catalytic material. They proposed that CH4 does not react directly with CO2. Since this first work and especially in the past decade, several authors have proposed different mechanisms for CH4/CO2 reforming40-42 but only a few of them reported kinetic equations. The reaction mechanism for this reaction has been investigated employing a number of different techniques, and the most important observations from which a mechanism is proposed are the following: • A strong interaction exists between CO2 and the La2O3 leading to the formation of stable La2O2CO3. The presence of carbonate species has been observed using XRD during and after the reaction (Figure 1). Previous works of the authors14,33 as well as by Verykios10 have concluded that the formation of oxycarbonates during reaction plays a central role in the dry reforming of methane over Ni/La2O3. Using steady-state isotopic tracing kinetic analysis, Verykios et al.,43 found that methane exists on the catalyst surface under reaction conditions. A certain fraction of the Ni content of the catalyst is visible by XPS after the reforming reaction even under integral conditions (see section 3.4). This implies that this portion of Ni is free of carbon deposits after reaction. No carbon deposition was detected through TGA, and elemental analysis after reaction and no deactivation was observed during 100 h on stream (see section 3.7), which suggests that the graphitic carbon was not formed on the Ni surface. Transient studies have shown that the rate of dissociation of CO2 on Ni crystallites is not significant as compared to that of
Ind. Eng. Chem. Res., Vol. 47, No. 23, 2008 9277 24
CH4. This suggests that the carbon which accumulates on Ni crystallites derives principally from the CH4 molecule. The active carbon species which exist on the catalyst surface under reaction conditions consist exclusively of carbon and not of CHx (x > 0) species.35 Recently, some studies have proposed that CH4 irreversibly decomposes in a sequence of elementary steps to form chemisorbed carbon and hydrogen atoms, reaction which is considered as the slow reaction step.44-46 Finally, it was assumed that the surface coverage of other species, such as H or CO, is negligible. On the basis of the above observations available in the literature and the results obtained in our study, the following dual active-site mechanism is proposed for CO2 reforming of CH4 assuming that the surface coverage of H2 and CO adsorbed are negligible. K1
CH4 + S1 798 S1-CH4 k2
S1-CH4 98 S1-C + 2H2 K3
CO2 + S2 798 S2-CO2 k4
S2-CO2 + S1-C 98 S2 + 2CO + S1
(1)
(2)
(3)
(4)
Where, S1 is a metallic nickel active site, and S2 is a support La2O3 active site. Reaction 1 corresponds to reversible adsorption of methane on the metallic Ni° clusters (S1) which leads to cracking of methane to produce carbon deposits and hydrogen, reaction 2. Methane cracking is a slow step while the methane adsorption step is at equilibrium. Simultaneously, the CO2 rapidly reacts with La2O3 (S2) to form the oxycarbonate La2O2CO3 represented as S2-CO2 in reaction 3. This is a fast step considered to be at equilibrium. Finally, the La2O2CO3 species reacts with carbon originated from the CH4 decomposition on the nickel crystallites at the metal-support interface to liberate CO and at the same time clean up the metallic surface, reaction 4. 3.11. Kinetic Model. A rate expression for the CO2 reforming of methane was developed based on the mechanism illustrated above. It was assumed that step 2 is the ratedetermining step and that both Ni° particles and La2O3 are actives sites for the reaction. Under these assumptions, the rate of methane conversion can be expressed as: rCH4 ) K1k2K3k4[CH4][CO2] K3k4[CO2] + K1K3k4[CH4][CO2] + K1k2[CH4] + K1k2K3[CO2] (5) Where K1 is the equilibrium constant for methane adsorption, k2 is the rate constant of the decomposition of methane (cracking) on the Ni° surface, K3 is the equilibrium constant of the reaction between CO2 and La2O3 to form the oxycarbonate, and k4 is rate constant of the reaction between the oxycarbonate species and the carbon deposited on the surface of Ni° clusters. To calculate the kinetic constants individually, the value K1k2 ) 2.61 × 10-3 exp(-4300/T)(mol g-1 s-1) (kPa-1) was obtained from data published by Verykios et al.10 The constant values are summarized in Table 3.
Table 3. Kinetic Model Parameters at 973 K kinetic constant
value at 973 K
K1 k2 K3 k4
141 × 10-3 kPa 0.22326 × 10-3 mol g-1 s-1 15.98 × 10-3 kPa 13.22 × 10-3 mol g-1 s-1
Table 4. Effect of the Number of Active Sites in the Kinetics Model PCH4
PCO2
experimental rCH4
calculated from the model rCH4
10 10 5 10
20 30 10 10
1.36 × 10-4 1.36 × 10-4 9.6 × 10-5 1.30 × 10-4
1.29 × 10-4 1.29 × 10-4 8.9 × 10-5 1.25 × 10-4
% dev
(% dev)a
5 5 7 4
15 20 13 11
a Calculated from the model which includes only the metallic particles as active sites.
From Table 3, we can observe that the rate constant of the reaction between the oxycarbonate species and the carbon deposited on the surface of Ni° clusters (k4) is almost 60 times higher that the rate constant of the cracking of methane on the surface of Ni° (k2). This suggests that the catalyst kinetically inhibits the excess of carbon deposition during the reaction as shown by the elemental analysis of the catalyst after reaction, which confirms that no coke deposition was formed after 100 h. From the above results, we can suggest that the surface concentration of carbon at the metal surface is at steady state and remains constant, and the rate of CHx decomposition (step 2) equals that for carbon species reaction with CO2 adsorbed species (step 4). The proposed kinetic model was used with our data, and the results for the Ni/La2O3 catalyst are displayed in Figure 11 at 973 K. The model fits very well the experimental data. This behavior was attributed to the no deactivation of this catalyst during CH4-CO2 reforming, validating the hypothesis that the concentration of carbon remains constant. The rate of methane conversion was determined with the model which incorporates both the metallic particles and the support as the active sites (eq 5). The results are displayed in Table 4. We can observe that the kinetic model satisfactorily predicts the kinetic results with deviations from the experimental data below 5%. The success of the prediction of the kinetic data by the model is attributed to the role of the support as active site. This criterion has not been taken in account in previous research oriented to the development of mathematic models for this reaction mechanism.11,24 The values of the rate of methane conversion obtained with this model (eq 5) were compared which those calculated using the model which includes only the metallic particles as active sites. As we can observe in the last column of the Table 4, the deviations are 3 times higher compared with the model here developed. 4. Conclusions The Ni/La2O3 obtained from the perovskite LaNiO3 is a good catalyst for the dry reforming of methane. High catalytic activity with CH4 and CO2 conversions up to 90% and an H2/CO molar ratio close to 1 were obtained. The catalyst showed high resistance towards the deactivation by carbon formation under integral reaction conditions. The XRD data of the used catalyst after reaction indicates that lanthanum oxycarbonate La2O2CO3, which is the major species formed at the surface, plays an important role in the dry reforming of methane. The apparent activation energies of CH4 and CO2 are lower than the respective activation energies reported for noble metal catalysts.
9278 Ind. Eng. Chem. Res., Vol. 47, No. 23, 2008
Figure 11. Fit of the proposed kinetic model for CO2 reforming of methane as a function of (a) CH4 and (b) CO2 partial pressures. Curves were calculated using eq 5: (solid symbols) experimental results at 973 K, (solid lines) model prediction.
The kinetics of the reaction over the Ni/La2O3 catalyst was investigated at 973 K, and based on experimental results and on the data available in the literature, a mechanistic model was proposed. The mechanism assumes the adsorption and subsequent cracking of methane on metallic nickel particles to produce carbon deposition, which was considered as the rate-determining step. Simultaneously, the CO2 reacts with the La2O3 to form La2O2CO3, which then reacts with carbon at the Ni°-La2O2CO3 interface to produce CO and at the same time clean up the metallic surface. This study shows that the catalyst kinetically inhibits the excess of carbon deposition during the reaction, as can be seen by the higher value of the rate constant of the reaction between the oxycarbonate species and the carbon deposited on the Ni° surface (k4). In the mechanism here proposed, a rate expression for the CO2 reforming of methane was obtained assuming that both, Ni° particles and La2O3 support are the actives sites for the reaction. This kinetic model fits very well the experimental data, and the rate expression predicts better the rate of methane conversion than the models which incorporate only the metallic cluster as the active site. Acknowledgment The authors are grateful to the PICS program: “Valorization of natural gas and Fischer-Tropsch synthesis” for the financial support given. F.M. and G.S.G. acknowledge the University of Antioquia for the financial support of the Sostenibilidad Program and to Colciencias for the support of project 1115-06-17639. G.S.G. thanks COLCIENCIAS and the University of Antioquia for the Ph.D. scholarship. The authors would like to thank one of the reviewers for helpful comments and suggestions. Literature Cited (1) Lewis, W. K.; Gilliland, E. R.; Reed, W. A. Ind. Eng. Chem. 1949, 41, 1227.
(2) Aparicio, L. M. J. Catal. 1997, 165, 292. (3) Edwards, J. H. Catal. Today. 1995, 23, 59. (4) Tsipouriari, V. A.; Efstathiou, A. M.; Zhang, Z. L.; Verykios, X. E. Catal. Today 1994, 21, 579. (5) Bitter, J. H.; Seshan, K.; Lercher, J. A. J. Catal. 1997, 171, 279. (6) Valderrama, G.; Goldwasser, M. R.; Urbina de Navarro, C.; Tatiboue¨t, J. M.; Barrault, J.; Batiot Dupeyrat, C.; Martı´nez, F. Catal. Today 2005, 107-108, 785. (7) Hu, Y. H.; Ruckenstein, E. Catal. Lett. 1996, 36, 145. (8) Qin, D.; Lapszewicz, J. Catal. Today 1994, 21, 551. (9) Van Keulen, A. N. J.; Seshan, K.; Hoebink, J. H. B. J.; Ross, J. R. H. J. Catal. 1997, 166, 306. (10) Verykios, X. E. Int. J. Hydrogen Energy 2003, 28, 1045. (11) Mu´nera, J. F.; Irusta, S.; Cornaglia, L. M.; Lombardo, E. A.; Vargas Cesar, D.; Schmal, M. J. Catal. 2006, 245, 25. (12) Pechini, R. Method of preparing lead and alkaline earth titanates and niobates and coating method using the same to form a capacitor. U.S. Patent No 3,330,697, 1967. (13) Chick, L. A.; Pederson, L.R.; Maupin, G. D.; Bates, J. L.; Thomas, L. E.; Exarhos, G.J. Mater. Lett. 1990, 1, 1. (14) Batiot-Dupeyrat, C.; Sierra, G. A.; Mondragon, F.; Barrault, J.; Tatiboue¨t, J. M. Catal. Today 2005, 107, 474. (15) Wachowski, L.; Zielinski, S.; Burewicz, A. Acta Chim. Acad. Sci. Hung. 1981, 06, 217. (16) Fierro, J. L. G. Catal. Today 1990, 8, 153. (17) de la Cruz, R. M. G.; Falcon, H.; Pen˜a, M. A.; Fierro, J. L. G. Appl. Catal. B: EnViron. 2001, 33, 45. (18) Requies, J.; Cabrero, M. A.; Barrio, V. L.; Guemez, M. B.; Cambra, J. F.; Arias, P. L.; Pe´rez-Alonso, F. J.; Ojeda, M.; Pen˜a, M. A.; Fierro, J. L. G. Appl. Catal. A: Gen. 2005, 289, 214. (19) Tauster, S. J.; Fung, S. C.; Garten, R. L. J. Am. Chem. Soc. 1978, 100, 170. (20) Pen˜a, M. A.; Fierro, J. L. G. Chem. ReV. 2001, 101, 1981. (21) Tejuca, L. G.; Fierro, J. L. G.; Tascon, J. M. D. AdV. Catal. 1989, 36, 385. (22) Perego, C.; Peratello, S. Catal. Today 1999, 52, 133. (23) Le Page, J. F. Applied Heterogeneous Cata´lisis; Editions Technip: Paris, 1987. (24) Tsipouriari, V. A.; Verykios, X. E. Catal. Today 2001, 64, 83. (25) Zhang, Z.; Verykios, X. E.; Mac.Donald, S. L.; Affrossman, S. J. Phys. Chem. 1996, 100, 744. (26) Stull, D. R.; Westrum, E. F.; Sinke, J. R. The Chemical Thermodynamics of Organic Compounds; Wiley: New York, 1969. (27) Zhang, Z. L.; Verykios, X. E. Appl. Catal. A. 1996, 138, 109. (28) Zhang, Z. L.; Verykios, X. E. Catal. Lett. 1996, 38, 175. (29) Erdohelyi, A.; Cserenui, J.; Solymosi, F. J. Catal. 1993, 141, 287. (30) Bradford, M. C. A.; Vannice, M. A. Appl. Catal. A: Gen. 1996, 142, 97. (31) Wang, S.; Max Lu, G. Q. Ind. Eng. Chem. Res. 1999, 38, 2615. (32) David, C.; LaMont, W.; Thomson, J. Chem. Eng. Sci. 2005, 60, 3553. (33) Sierra Gallego, G.; Mondrago´n, F.; Barrault, J.; Tatiboue¨t, J. M.; Batiot-Dupeyrat, C. Appl. Catal. A: Gen. 2006, 311, 164. (34) Laosiripojana, N.; Assabumrungrat, S. Appl. Catal. B: EnViron. 2005, 60, 107. (35) Laosiripojana, N.; Sutthisripok, W.; Assabumrungrat, S. Chem. Eng. J. 2005, 112, 13. (36) Osaki, T.; Mori, T. J. Catal. 2001, 204, 89. (37) Bhat, R. N.; Sachtler, W. M. H. Appl. Catal. A. 1997, 150, 279. (38) Horiuchi, T.; Sakuma, K.; Fukui, T.; Kubo, Y.; Osaki, T.; Mori, T. Appl. Catal. A: Gen. 1996, 144, 111. (39) Bodrov, I. M.; Apel′baum, L. O. Kinet. Catal. 1967, 8, 326. (40) Richardson, J. T.; Paripatyadar, S. A. Appl. Catal. 1990, 61, 293. (41) Rostrup-Nielsen, J. R.; Bak Hansen, J. H. J. Catal. 1993, 144, 38. (42) Bitter, J. H.; Seshan, K.; Lercher, J. A. J. Catal. 1998, 176, 93. (43) Tsipouriari, V. A.; Verykios, X. E. J. Catal. 1999, 187, 85. (44) Wei, J.; Iglesia, E. J. Catal. 2004, 225, 116. (45) Matsui, N.; Anzai, K.; Akamatsu, N.; Nakagawa, K.; Ikenaga, N.; Suzuki, T. Appl. Catal. 1999, 179, 247. (46) Bradford, M.; Vannice, M. J. Catal. 1998, 173, 157.
ReceiVed for reView February 18, 2008 ReVised manuscript receiVed August 22, 2008 Accepted September 18, 2008 IE800281T
