The CO2 reforming of CH4, CO2 + CH4 → 2CO + 2H2, has received of great attention because it converts green house gases into a synthesis gas. This reaction produces a synthesis gas with a H2/CO molar ratio close to 1, which is suitable for the production of oxygenated hydrocarbons like methanol and aldehyde. Nickel has been proven to be one of the active catalysts for the CO2 reforming of CH4. A major problem of a nickel catalyst is to deposit carbon species on the surface during the reaction, resulting in the deactivation of catalyst. It has been reported that the tendency of the carbon deposition is closely related to the type of a support used and the dispersion and size of nickel particles on support. 1,2 To develop a coke-resistant nickel catalyst various metal oxides have been examined as the support of the Ni catalyst. Lanthanum sesquioxide as a support of a nickel catalyst is known to be a particularly effective component in reducing the carbon deposition.3,4 It has been reported that La2O3 can suppress the carbon deposition on the surface and the aggregation of nickel crystallites during the reforming reaction, leading to the enhancement of the catalyst stability.5-8
Supported metal catalysts are generally prepared by the conventional impregnation method. However, this method often fails to reproduce the catalyst and may give rise to the poor dispersion of metal particles on the support. One promising method for the preparation of well-dispersed nickel particles has been proposed by Takehira group,9,10 called “solid phase crystallization method”. In this method, metal-containing complex oxides are utilized as metal precursors. In recent, well-dispersed nickel particles on La2O3 have been successfully prepared by using LaNiO3 or La2NiO4 perovskite oxide as a starting material and the resulting Ni/ La2O3 catalyst has been proved to be highly active and stable in the reforming reaction.4,11
Catalytic performances of Ni/La2O3 for the reforming reaction have been largely studied.12-14 However, little kinetic data for the catalytic reaction have been reported so far.8,15 Although empirical kinetic data do not directly provide the accurate information of the reaction mechanism, kinetic information about the catalytic reaction is essential to understand the reaction mechanism. The most important consideration in the kinetic study is to monitor the concentration of reactants and/or products of interest as a function of time. We have shown previously that photoacoustic spectroscopy (PAS) is a suitable technique for in situ monitoring of the initial stage reaction processes.16,17 Since the PAS technique provides highly selective detection and extremely low molecular gas level detection, the signals with enough intensity can be obtained for even time-resolved experiments.18,19 The photoacoustic signals are directly related to the concentration of samples of interest, which allows this method to measure a temporal variation in concentration at short period of reaction times where rates are high.
In this work, LaNiO3 perovskite-type oxide was synthesized by a citrate sol-gel method. The oxide was reduced in a flow of H2 to form nickel metal particles dispersed on lanthanum oxide particles, so that the resulting product was examined as a catalyst for the CO2 reforming of CH4. A CO2 laser-based photoacoustic technique with a differential photoacoustic cell was employed to analyze kinetically the CO2-CH4 reaction over Ni/La2O3 catalyst. The photoacoustic experiments were performed using a static reactor in the temperature range of 550-700 ℃ at various partial pressures of CO2 and CH4. The changes in the concentration of CO2 during the catalytic reaction were measured as a function of time. Rates of CO2 consumption in early reaction stage were obtained from the CO2 photoacoustic data. The apparent activation energy and reaction orders were determined from the initial rates of CO2 consumption. The conversion of reactants and the ratio of H2/CO were also measured in a flow reactor in the temperature range of 500-800 ℃ using an on-line gas chromatography system. The kinetic data obtained in this work were compared with others reported previously. A reaction mechanism of the CO2-CH4 reaction over the Ni/La2O3 catalyst was then discussed from the kinetic results.
Catalyst Preparation. LaNiO3 perovskite-type oxide was prepared from La(NO3)3·6H2O (> 99.99%, Aldrich-Sigma) and Ni(NO3)2·6H2O, (> 99.99%, Aldrich-Sigma). Both the lanthanum nitrate and nickel nitrate were weighed to yield the desired mole ratio of La/Ni and were then dissolved in deionized water. The resulting solution was mixed with an aqueous solution of citric acid to have the same amount of equivalents. Water was slowly evaporated from the mixed solution at 70 ℃ in vacuum until a viscous gel was formed. The viscous gel was kept at 100 ℃ overnight to produce a solid amorphous mixture. The resultant amorphous mixture was ground and calcined at 400 ℃. Further calcination was made at 600 ℃ to eliminate carbon species remaining in the sample and then cooled to room temperature at 60 ℃/h rate. The product was reduced in a flow of H2 at 700 ℃ for 3 h and further heated at 700 ℃ for 1 h in a flow of He, which was followed by cooling to room temperature and ground in an agate mortar. These processes were then repeated three times to form nickel metal particles. X-ray powder diffraction analyses for the samples were conducted by using a Bruker D2 Phaser diffractometer with Cu-Kα radiation at room temperature. EDX (energy dispersive X-ray) analyses were performed for the LaNiO3 perovskite oxide using an Oxford Link ISIS 300 spectrometer. FE-SEM (Field emissionscanning electron micrograph) images of catalyst were obtained with a Hitachi SU70.
Kinetic Measurements. The catalytic reaction rate in the reforming reaction was measured by using the CO2 laserbased photoacoustic technique. The experimental arrangement for the photoacoustic detection method was described in previous paper.17,19 The catalytic reaction was carried out in a closed-circulating reactor system, where the reactor was made of a quartz tubing with a volume of 21 cm3. The reactor was placed in an electric furnace and directly connected to a differential photoacoustic cell by an adjoining tube. The differential photoacoustic cell was consisted of a sample cell and a reference cell, separated from each other by a ZnSe window. Each photoacoustic cell was a Helmholtz resonator of 1.9 cm in diameter and 3.3 cm in length with an adjoining tube of 1.1 cm in diameter and 2.8 cm in length, where microphones were located at the ends of the adjoining tubes. The output beam of a cw CO2 laser (Synrad Series 48-1-28) operating in multilines of 10.6 μm was modulated at the nonresonance condition of 25 Hz. The signals were detected by the microphones in both the sample (signal A) and reference (signal B) cells and amplified by a lock-in amplifier (EG ＆ G Princeton Applied Research Model 5210). The signal ratio (A/B) was then recorded as a function of time by using a personal computer. The total pressure of gaseous reactants in the reactor was maintained at 40 Torr filled with N2 as a buffer gas. The reference cell (signal B) was filled with a mixture of CO2 (0.2 Torr) and N2 (39.8 Torr) and the sample cell (signal A) was connected to the reactor loaded with the catalyst (200 mg). The gaseous reaction mixture (CO2/CH4/N2) was introduced into the reactor at a given temperature. The purity of gases was greater than 99.99% and the gases were dehydrated with suitable filters.
Conversions of reactants and ratio of H2/CO in the CO2-CH4 reaction over the Ni/La2O3 catalyst were measured in a fixed-bed type flow reactor in the temperature range of 500-800 ℃ at atmospheric pressure. The reactor was quartz tubing with a 0.8-cm-i.d. and 3-cm-long tube sealed to 0.4-cm-i. d. tubes on two ends. The catalyst (200 mg) was well dispersed on a fused-silica bed placed in the middle of the reactor which was kept in an electric furnace. The feed flow rate at ambient conditions was 5/5/20 cm3/min (CH4/CO2/ Ar) and the flow rates of gases were controlled by electronic mass flow controllers. The gaseous reactants and products were analyzed by using an on-line gas chromatography with a thermal conductivity detector. A cold trap was placed at the reactor exit to remove water vapor from the gaseous mixture. CH4, CO2, and CO gases were analyzed by a gas chromatography using He carrier gas and H2 gas was separately analyzed by a gas chromatography using Ar carrier gas. The CO2 and CH4 conversions were calculated on the basis of the formula; conversion (%) = 100 × [Fin − Fout]/Fin, where Fin and Fout were inlet and outlet flow rate of each reactant, respectively. The conversions were typically compared at 0.5 h time-on-stream.
Results and Discussion
The XRD pattern of the LaNiO3 sample synthesized by the citrate sol-gel method is displayed in Figure 1(a). The LaNiO3 was found to be monophase perovskite-type oxide possessing a rhombohedra structure (JCPDF 33-0711). Ni/La atomic ratio in LaNiO3 was determined to be 1.02 from the K-lines obtained by EDX analysis. The XRD pattern of the H2-reduced LaNiO3 is shown in Figure 1(b), in which La2O3 (JCPDF 5-0562) and Ni metal with (111) preferential plane of a cubic structure (JCPDF 87-0712) are observed as a major phase and La(OH)3 (JCPDF 36-1481) as a minor phase. This result indicates that the LaNiO3 oxide is completely decomposed into La2O3 and Ni0 under the reducing conditions. Accordingly, the H2-reduced LaNiO3 sample corresponds to 26.5 wt % Ni/La2O3. Figure 1(c) shows the XRD pattern of the Ni/La2O3 catalyst after the activity test for the reforming reaction at 700 ℃ in a flow reactor, which reveals the presence of carbon solid (JCPDF 75-1621) and La2O2CO3 (JCPDF 37-0804 and 48-1113) phases. A FESEM image of the Ni/La2O3 catalyst is displayed in Figure 2, showing well-dispersed nickel metal crystallites and lanthanum oxide particles.
Figure 1.XRD patterns of (a) LaNiO3 prepared by citrate sol-gel method, (b) Ni/La2O3 catalyst obtained by reduction of LaNiO3, and (c) Ni/La2O3 catalyst after activity test at 700 ℃.
Fig. 2.FE-SEM image of Ni/La2O3 sample obtained by reduction of LaNiO3.
The rates of the CO2 consumption in early reaction stage for the reforming reaction were measured by the CO2 laser-based photoacoustic technique. As mentioned above, the PAS technique is applicable to the quantitative analysis by monitoring the progress of the reaction because the photoacoustic signal is directly proportional to the concentration of reaction analytes under a given conditions. The CO2 photoacoustic signal was found to show an optimal linearity and response of the signal under conditions of CO2 laser power ranging of 1-8 W and CO2 partial pressures below 10 Torr. Accordingly, the PAS measurements were conducted under CO2 partial pressure below 10 Torr at a CO2 laser power of 1.5 W. A blank test, performed with a CO2/CH4/N2 (5/5/30 in Torr) mixture in the absence of catalyst, showed no variations of the CO2 photoacoustic signal in the temperature range of 550-700 ℃. Thus, the effect of temperature on the rate of CO2 consumption for the CO2-CH4 reaction over the Ni/La2O3 was investigated using CO2/CH4/N2 (5/5/30 in Torr) mixture in the temperature range of 550-700 ℃. In Figure 3 is displayed the variations of the CO2 photoacoustic signal with time for the reforming reaction at various temperatures. As observed in Figure 3, the CO2 photoacoustic signals gradually decrease with time in the reaction period up to 600 s. The rates of the CO2 consumption were estimated from the photoacoustic signal data in the early reaction period of 0-50 s. The rates were then plotted as a function of reciprocal temperature according to the Arrhenius-type equation. As displayed in Figure 4, the curve shows linear lines below and above 610 ℃. The apparent activation energies for the CO2 consumption, calculated from the slope of each curve, were 10.4 kcal/mol for the low temperature region (550-610 ℃) and 14.6 kcal/mol for the high temperature region (610-700 ℃). To investigate the pressure dependence of the CO2 consumption rate for the reforming reaction, the CO2 photoacoustic signals were recorded under various partial pressures of CO2 and CH4, PCO2 and PCH4, at 640 ℃. The initial rates obtained from the CO2 photoacoustic data were plotted as a function of PCO2 and PCH4, respectively, according to the power rate law, rate = k(PCO2)m(PCH4)n. Figure 5 shows the PCO2 and PCH4 dependences of the rate of CO2 consumption for the reforming reaction. The reaction orders, determined from the slope of the curves, were 0.43 with respect to CO2 and 0.33 with respect to CH4.
Figure 3.Variations of CO2 photoacoustic signal with time at various temperatures for CO2-CH4 reaction over Ni/La2O3 catalyst: CO2/CH4/N2 = 7/7/26 in Torr, catalyst weight = 200 mg.
Figure 4.Influence of temperature on rate of CO2 disappearance for CO2-CH4 reaction over Ni/La2O3 catalyst: CO2/CH4/N2 = 7/7/ 26 in Torr, catalyst weight = 200 mg.
Figure 5.(a) PCO2 dependence and (b) PCH4 dependence of CO2 disappearance rate at 640 ℃ for CO2-CH4 reaction over Ni/La2O3 catalyst.
Figure 6 shows the variations of conversions and the ratio of H2/CO with temperatures for the CO2-CH4 reaction over Ni/La2O3 catalyst in the temperature range of 500-800 ℃, which was measured in a flow of CH4/CO2/Ar (5/5/20 cm3/ min) mixture at 1 atm. As shown in Figure 6, the CO2 conversion is nearly equal to the CH4 conversion below 550 ℃, but in the temperature range of 600-750 ℃ the CO2 conversion is somewhat lower than the CH4 conversion. Note that at 800 ℃ both the conversions reach to almost 100%. The ratios of H2/CO were found to vary with temperature between 0.66 and 1.0; less than 1 at 500 ℃ and 550 ℃, close to 1 at 600 ℃ and 650 ℃, and less than 1 above 700 ℃.
Fig. 6.(a) Conversions and (b) ratio of H2/CO at various temperatures for CO2-CH4 reaction over Ni/La2O3 catalyst: CO2/ CH4/Ar = 5/5/20 cm3/min, catalyst weight = 200 mg.
Because only limited kinetic data for the rates of the CO2 consumption have been reported previously, it is difficult to compare the present data directly with others. According to the kinetic results reported previously for the nickel-catalyzed CO2-CH4 reaction,20 the apparent activation energies for the CO2 consumption, varying with the supports used, are in the range of 8-22 kcal/mol, noting that our values fall in the range. Tsipouriari and Verykios15 measured the rates of the CO formation for the CO2-CH4 reaction over 17 wt % Ni/ La2O3 catalyst in a flow reactor in the temperature range of 650-750 ℃. They determined the apparent activation energy to be 13.2 kcal/mol of which value is close to the one in the present experiment, 14.6 kcal/mol, obtained in the temperature range of 610-700 ℃. In the previous work,16 we performed kinetic studies of the reforming reactions over Ni/ TiO2 and Ni/Al2O3 catalysts using a PAS technique. From the CO2 consumption rates measured in the temperature range of 400-650 ℃, the apparent activation energies were determined to be 14.3 kcal/mol for Ni/TiO2 catalyst and 15.4 kcal/mol for Ni/Al2O3 catalyst. These values are close to 14.6 kcal/mol obtained in the temperature range of 610-700 ℃ in this work. Figure 5 shows that the CO2 consumption rate depends on the PCO2 and PCH4, in which the reaction order to CO2, 0.43, is somewhat higher than the reaction order to CH4, 0.33. The reaction orders are also difficult to compare directly with others because partial orders determined from the CO2 consumption rates for the Ni/La2O3-catalyzed reforming reaction have not been reported. According to Tsipouriari and Verykios,15 the rate of the CH4 consumption for the CO2-CH4 reaction over Ni/La2O3 catalyst in the temperature range of 650-750 ℃ is more sensitive to PCO2 than to PCH4 at low partial pressures below 75 Torr, which is considered to be in agreement with our results.
The CO2-CH4 reforming reaction is highly endothermic and is favored by low pressure and high temperature. Various reaction models for the reforming reaction have been suggested during the last twenty years. Although there are some disagreements in the reaction mechanism, it is generally accepted that both CH4 and CO2 are dissociatively chemisorbed on nickel metal surface. Erdöhelyi et al.21 and Rostrup-Nielson et al.22 have suggested that methane is dissociatively adsorbed on nickel surface, CH4(g) ⇄ CHx(ads) + (4−x)H(ads), and the CO2 dissociation can be promoted by the H(ads). Bradford and Vannice20 have proposed that surface CHx(ads) fragments react with either oxygen atoms or hydroxyl groups to form a formate-type CHxO intermediate which subsequently decomposes into CO and H2 products. Zhang et al.23 have suggested from the FT-IR and XPS results for Ni/La2O3-catalyzed CO2-CH4 reaction at 750 ℃ that a formate-type HCOO−intermediate is initially formed and subsequently reacts with H+(ads) to yield H2O and CO.
If formate-type species are intermediates in the reaction mechanism, the rate of the CO2 consumption should be closely related to the adsorption of CH4 on the surface.22,23 As is observed in the current study, the CO2 consumption rate is dependent on the PCH4. In order to form formate-type intermediates, oxygen atoms must be provided from CO2 acting as a potential oxidant. It is well known that CO2 participates in the reaction mechanism through the reverse water-gas shift reaction to form surface OH groups which react with CHx(ads) to form formate-type intermediates.15 Note that the reverse water-gas shift reaction, CO2(g) + H2(g) → H2O(g) + CO(g), occurs below 820 ℃. If the reverse water-gas shift reaction is favorable as a side reaction, the ratio of H2/CO would be lower than 1. In Figure 6, the H2/ CO ratios at 500 and 550 ℃, measured in a flow reactor with CH4/CO2/Ar mixture (5/5/20 cm3/min), are lower than 1, where the CO2 conversion is nearly equal to the CH4 conversion. This result suggests the occurrence of the reverse watergas shift reaction as a side reaction at these temperatures.
On the other hand, the H2/CO ratios measured at 600 and 650 ℃ are close to 1, where the conversion of CO2 is lower than CH4 as shown in Figure 6. The result implies that other side reactions may occur at the temperatures. Under the conditions of stoichiometric CO2-CH4 reaction, both the CO disproportionation (2CO(g) → C(s) + CO2(g)) and the methane cracking (CH4(g) → C(g) + 2H2(g)) can be considered as side reactions. From the viewpoint of thermochemistry, the CO disproportionation occurs below 700 ℃ at 1 atm and the methane cracking occurs above 560 ℃ at 1 atm. When both the reactions as side reactions favorably occur in the temperature range of 600-650 ℃, the CO2 conversion would be lower than the CH4 conversion and then the H2/CO ratios would increase with temperature. As observed in Figure 6, the H2/CO ratios at 600 ℃ and 650 ℃ are greater than those below 550 ℃. Both the CO disproportionation and the methane cracking produce carbon species, which can give rise to the catalytic deactivation. In Figure 1, the XRD pattern of the Ni/La2O3 catalyst after the activity test at 700 ℃ exhibits the presence of carbon phase, indicating that carbon is deposited during the reforming reaction. The CO disproportionation reaction is exothermic (ΔHo = −41.2 kcal/mol), while the methane cracking is endothermic (ΔHo = +17.9 kcal/mol). Accordingly, it is expected that the methane cracking is more favorable at higher temperatures. When the methane cracking occurs favorably at higher temperatures, the catalyst would be rapidly deactivated by the carbon deposition during the reforming reaction.
Carbon species deposited on the surface of Ni catalyst, which can serve as intermediates in the reforming reaction, are known to be highly reactive.24,25 The surface carbon can readily react with H2O(g) and CO2(g) to form H2, CO, and CO2 according to the following processes; C(s) + H2O(g) → CO(g) + H2(g), C(s) + 2H2O(g) → CO2(g) + 2H2(g), C(s) + CO2(g) → 2CO(g). These reactions, called the gasification of carbon solid, can occur above 700 ℃ at 1 atm and can be promoted by nickel.26.27 In Figure 6, the H2/CO ratio is found to be less than 1 above 700 ℃ and decreases with increasing temperature. This suggests that the CO2 gasification of carbon solid may be more favorable. The CO2 consumption in the side reaction, C(s) + CO2(g) → 2CO(g), would give rise to larger change in the total CO2 photoacoustic signal during the reforming reaction, resulting in the observation of higher apparent activation energy for the CO2 consumption. It is also notable that the C(s)-CO2(g) reaction is expected to be more favorable at lower pressures. As shown in Figure 4, the apparent activation energy obtained for the high temperature region, 14.6 kcal/mol, is higher than that for the low temperature region, 10.4 kcal/mol, which is enable us to consider that the C(s)-CO2(g) reaction occurs at high temperatures.
La2O3 oxide is known to be strongly basic and it is easily hydrated and carbonated when it is exposed to air.28 In Figure 3, the XRD pattern of the Ni/La2O3 catalyst after the activity test at 700 ℃ shows the presence of lanthanum dioxycarbonate, indicating that La2O3 is carbonated to La2O2CO3 phase during the reforming reaction. When La2O3 is exposed to a mixture of water vapor and CO2, La2(OH)4CO3 phase is formed and subsequently decomposes to La2O2CO3 around 460 ℃, which have been also identified as intermediate phases in the decomposition of lanthanum carbonate.28 Since La2O2CO3 phase decomposes around 630 ℃ at 1 atm, La2O3 support is expected to be carbonated at surface levels during the reforming reaction at temperatures above 630 ℃. The carbonation of La2O3 by CO2(g) can be represented as La2O3 + CO2(g) ⇄ La2O2CO3 or La2O3·(CO2)ads, which means that there is an equilibrium between CO2(g) and CO2(ads). Rivas et al.11 and Zhang et al.29 have suggested that La2O2CO3 species at the interfacial area react with the carbon species deposited on the surface of nickel particles to produce two CO molecules. This indicates that the La2O2CO3 phase can play a major role in the catalytic stability. The carbon species can react with La2OCO2 molecules adsorbed on La2O3 or released from La2O2CO3 to produce CO. It is shown in Figure 6 that the H2/CO ratio decreases with temperature above 700 ℃. When the reaction between carbon and La2OCO2(g)/or La2OCO2(ads) favorably occurs during the reforming reaction, the ratio of H2/CO would be lower than 1. In the early work of Ni/ Al2O3- and Ni/TiO2-promoted CO2-CH4 reactions,16 the reaction orders determined from initial rates of the CO2 consumption at 600 ℃ were found to be 0.48 for CH4 and 0.45 for CO2 for the Ni/Al2O3-promoted process, and 0.38 for CH4 and 0.32 for CO2 for the Ni/TiO2-promoted process. This indicates that the reaction orders to CH4 are somewhat higher than to CO2. However, the dependence of the CO2 consumption rate to CO2, determined from initial rates of the CO2 consumption at 640 ℃ in this work, was observed to be somewhat higher than that to CH4. This result is considered to be due to the C(s)-CO2(g) reaction (the reverse Boudouard reaction).
In this work, well-dispersed Ni/La2O3 catalyst was prepared by the reduction of perovskite-type LaNiO3 and examined for the CO2 reforming of CH4. The CH4 and CO2 conversions measured in a flow reactor under atmospheric pressure showed the catalyst to be highly active for the reaction. The catalytic reaction was kinetically investigated in the temperature range of 550-700 ℃ at various partial pressures of CO2 and CH4 (40 Torr total pressure of CO2/CH4/N2) in a static reactor system with the differential photoacoustic cell. Rates of CO2 disappearance were obtained by in situ monitoring the concentration changes of CO2 during the catalytic reaction. The apparent activation energy and the reaction orders were determined from the rates of the CO2 disappearance in the early reaction stage. A reaction mechanism for the Ni/La2O3-catalyzed CO2-CH4 reaction could be deduced from the kinetic results in conjunction with the catalytic data obtained in a flow reactor. The kinetic data obtained for early reaction stages characterized by high reaction rates may be different from those for the later stages showing a constant catalytic activity. In addition, the kinetic measurement at high pressure could be more advantageous to the integrated kinetic analysis for the catalytic reaction because the reaction rate is directly linked to the surface coverage depending on the pressure of reaction gases. Nevertheless, it should pointed that the CO2 photoacoustic signal linearly varies with the concentration of CO2 in low molecular level detection at a low pressure environment and can be recorded as a function of time-on-stream during the catalytic process. This allows that the photoacoustic technique is considered as a suitable one for the molecular level study of the reaction kinetics at the early reaction stage in the catalytic CO2-CH4 reaction.