Since the signature of the international Kyoto protocol on 11 December 1997, most countries decided to discuss and implement regulations on gas emissions [1, 2]. The main compounds emitted by vehicles are unburned hydrocarbons (HC), CO, soot particles, and nitrogen oxides (NOx). To respect the European regulations, technical solutions were tested and implemented by automobile manufacturers [3, 4]. For selective reduction of NOx, two techniques have been developed: NH3-selective catalytic reduction (SCR) and HC-SCR. In the first technique, NH3 is used as a reactant to selectively reduce NO to N2 in the presence of oxygen [5]. The second technique uses unburned hydrocarbons from the exhaust gas or hydrocarbons injected into the exhaust gas line to reduce NO to N2 [6]. To achieve the desired chemical reactions, noble metals like Pd, Rh, and Pt are generally used. However, such platinum group metal-based catalysts suffer from thermal and sulfur deactivation. Moreover, despite the constant efforts made by catalyst suppliers and car manufacturers to decrease the noble metal loading, more than the half of the overall price of such systems comes from noble metals. To act as a noble metal substitute for oxidation reactions, different types of mixed oxide materials have been investigated. One of these materials is perovskite catalysts [7]. Such materials have an ABO3 general formula, where the A site is a rare-earth or alkaline cation and the B site is a transition metal. The nature of the A and B sites modify the chemical stoichiometry while maintaining the structural network [8, 9]. Usually, high calcination temperatures are required to obtain the definite crystal structure, and thus increase the sintering phenomena. However, in spite of their low specific surface areas, these types of materials have been extensively investigated because of their synthesis simplicity [10], cost efficiency, and hig h thermal stability. Moreover, partial substitution of A and/or B sites by different coordination cations can modify the valence and the oxygen non-stoichiometry, which influences the bulk oxygen mobility and redox properties of perovskite. These two parameters play an important role in the catalytic properties [8, 9]. The CO oxidation reaction over perovskite catalysts was first reported more than 50 years ago, and the potential of perovskite catalysts for engine exhaust gas treatment was suggested in the 1970s [11]. For this purpose, the most studied perovskite compositions are LaCoO3, LaFeO3, and LaMnO3 [7, 8, 9, 10, 11, 12, 13].
In this study, perovskites were used as a noble-metal-free catalyst for NO reduction by C10H22 in lean burn conditions. Decane was used as a representative model molecule of the unburned HCs present in the exhaust streams or added (like NH3) to increase their concentration in the HC-SCR process. The influence of A site substitution was investigated and compared with an unsubstituted LaMnO3 perovskite and a Pt/SiO2 reference catalyst.
The perovskite catalysts were prepared by a complexation route [10]. The reagents were La(NO3)3⋅6H2O (Alfa Aesar), Sr(NO3)2 (Acros Organics), Mn(NO3)2⋅6H2O (Alfa Aesar), KNO3 (Aldrich), and maleic acid (C4H4O4, Alfa Aesar). The nitrate precursors were dissolved in a 10 vol% aqueous solution of maleic acid. After salt dissolution, the pH of the mixture was kept at 9 by adding an aqueous solution of NH3. The mixture was dried overnight in an oven at 130 °C. The powder was heated under air at 800 °C for 4 h to remove the organic compounds and form the perovskite phase.
The quantitative measurement of the sample composition was performed by inductively coupled plasma optical emission spectroscopy (ICP-OES) using a Horiba Jobin Yvon Activa spectrometer. The solid sample was previously dissolved in acidic conditions. The specific surface areas (SBET) were estimated from N2 adsorption at −196 °C (BET method) using a Tristar surface area and porosity measurement apparatus from Micromeritics. The samples were characterized at room temperature by X-ray diffraction (XRD) to confirm the presence of the perovskite phase. The XRD measurements were carried out with a Siemens D5005 diffractometer using Cu Kα radiation at λ = 0.15418 nm.
The catalytic performances for C10H22 oxidation were evaluated in a U-shaped quartz reactor with 200 mg of catalyst. The reaction mixture was composed of 240 ppm C10H22 and 9 vol% O2 in He (carrier gas). The total flow rate was 120 mL min−1, corresponding to a space velocity of about 35,000 h−1. The reaction products were analyzed by micro gas chromatography for CO2 (Porapak Q) and gas chromatography for C10H22. The catalytic performances were investigated as follows: (1) introduction of the reactive mixture at room temperature, (2) heating from room temperature to 500 °C at 5 °C min−1, (3) plateau at 500 °C for 15 min, and (4) cooling from 500 °C to room temperature at 2 °C min−1. The light-off curves presented in this paper were recorded during the cooling ramp. The total C10H22 conversion was calculated by
The catalytic activity tests for NO oxidation were performed in the same device and under the same conditions as the C10H22 oxidation test (Section 2.3). However, the reaction stream, selected to simulate traces of NO in air, was a mixture of 1000 ppm NO, 20 vol% O2, and He (vector gas). The space velocity was about 35,000 h−1. The reactants and products (NO, NO2, and O2) were analyzed by gas chromatography for O2, with on-line IR and UV analyzers (Rosemount) for NO and NO2, respectively.
In a previous study of the SCR of NO by propene under lean-burn conditions [13], it was shown that the catalytic properties of Pt-based catalysts were stable under reactive mixture at 500 °C. The catalytic activity values reported in this study correspond to those recorded during the decrease of temperature after the stabilization step at 500 °C for 15 min. A substantial amount of NO2 was formed in the pipes of the apparatus, before the reactor and in the exhaust lines, as previously observed for studies of the SCR of NO by propene in the presence of oxygen [13]. Indeed, the thermodynamic equilibrium of the reaction leads to the formation of NO2 at room temperature. The NO2 formed in the pipes was subtracted from the total NO2 amount before the conversion of NO into NO2:
A simulated diesel exhaust gas composition was used to evaluate the catalytic properties of perovskites for NO reduction by C10H22. The reaction mixture was composed of 400 ppm NO, 240 ppm C10H22, and 9 vol% O2 in He (carrier gas) in the presence of H2O (1.5 vol%). The test was carried out under the same experimental conditions as the above oxidation tests. The C10H22 conversion was calculated by Eq. (1) and the N2, N2O and NO2 yields were determined by
TPD experiments after adsorption of NO were performed with the perovskite samples. For each experiment, 100 mg of solid was placed on a quartz wool bed in a U-shaped quartz reactor. The catalyst was pretreated at 400 °C for 15 min under He. After purging under He, NO adsorption was performed at 300 °C for 1 h under a NO/He mixture containing 2000 ppm NO. Then, the sample was cooled to 25 °C under the same atmosphere. After purging under He for 15 min, TPD was performed under He (60 mL min−1) with a temperature ramp of 20 °C min−1 up to 500 °C. The sample was then cooled under He at 25 °C. TPD was monitored by an INFICON IPC 400 quadrupole mass spectrometer. Signals at m/e = 14, 28, 30, 44, and 46 amu corresponding to N+, N2+ or CO+, NO+, CO2+ or N2O+, and NO2+ ions, respectively, were recorded. Owing to NO signal calibration, the amount of NO desorbed per square meter of catalyst was calculated.
The specific surface area (SBET) and chemical analysis of the different catalysts are summarized in Table 1. The noble metal catalyst showed very high SBET value (~350 m² g−1) whereas the SBET values of the perovskite catalysts were very low (< 20 m² g−1). A cation substitution site effect was observed. In particular, the insertion of Sr or K cations into the A sites delays the sintering phenomenon, and would thus lead to an increase of the specific surface area of the perovskite [14, 15].
For all of the catalysts, the XRD patterns shown in Fig. 1 have peaks that are characteristic of the perovskite structure: the orthorhombic LaMnO3 crystal phase (JCPDS 01-089-2471). Cation substitution in the same content (20 wt%) does not modify the peak positions.
The catalytic properties of perovskite samples were evaluated for C10H22 oxidation reaction in lean conditions and compared with the 2% Pt/SiO2 reference catalyst. The light-off curves are shown in Fig. 2 and the temperature at 10% (T10), 50% (T50), 90% (T90) C10H22 conversion are summarized in Table 2. The Pt sample showed the lowest activation temperature of around 90 °C, and the lowest light-off temperature (T50) at 127 °C. Among the perovskites, the best catalytic performances of the noble-metal-free catalysts for the C10H22 oxidation reaction was obtained over (La0.8Sr0.2)MnO3 with a T50 of 190 °C. The second group of active materials is composed of LaMnO3 and (La0.8K0.2)MnO3 with a T50 of 227 and 237 °C, respectively. The results of the catalytic activity show that the presence of the K+ cation in the A site, i.e., (La0.8K0.2)MnO3, inhibits the reaction while Sr substitution improves the catalytic performance compared with the unsubstituted material.
The influence of A site substitution on the catalytic properties of perovskite catalysts for NO oxidation to NO2 are shown in Fig. 3. The activity curve calculated from the thermodynamic data under our operating conditions is also shown (1000 ppm NO, 20 vol% O2). NO oxidation to NO2 is thermodynamically favoured at low temperatures. The 2% Pt/SiO2 reference catalyst showed the best catalytic activity for NO oxidation to NO2. The maximum NO2 yield of 100% was observed at 240 °C. Above 240 °C, the activity curve of the Pt sample corresponded to the thermodynamic equilibrium curve. High catalytic performance was also observed by Denton et al. [13] with different Pt loadings. Concerning the effect of the substitution of the cation A, above 380 °C, all of the conversion curves are superposed on the thermodynamic curve. This result indicates that whatever the nature of the A cation, NO/NO2 equilibrium is reached above 380 °C. However, the (La0.8Sr0.2)MnO3 perovskite gave the best NO2 yield with a maximum of 80% at 285 °C. The (La0.8K0.2)MnO3 sample showed a lower activity with a maximum NO conversion of 68% at 332 °C. The unsubstituted LaMnO3 sample had the lowest catalytic activity (53% at 365 °C). The better performance of the Sr-substituted perovskites cannot only be explained by the higher SBET (18 m² g−1) because the second highest SBET of 16 m² g−1 for (La0.8K0.2)MnO3 does not improve the NO2 yield compared with the sample with lower SBET. The presence of the divalent ion Sr2+ [9] in the (La0.8Sr0.2)MnO3 sample increases the average oxidation state of the cation in position B. These effects have been suggested to enhance the oxidat ion catalytic activity of this type of system by either facilitating oxygen mobility or enhancing the redox activity of the Mn cation [15].
Despite their similar specific surface area (18 and 16 m² g−1), the (La0.8Sr0.2)MnO3 and (La0.8K0.2)MnO3 samples showed different catalytic activity for C10H22 and NO oxidation. To understand this difference, TPD of NO was carried out. The desorption profiles of NO+ ions at 30 amu normalized to the weight of catalyst are shown in Fig. 4. The NO amounts per m² of catalyst are summarized in Table 3. For both samples, two temperature domains of NO adsorption were observed: (1) a low temperature domain between 0 and 310 °C for (La0.8Sr0.2)MnO3 and 0 and 370 °C for (La0.8K0.2)MnO3, and (2) a high temperature domain between 310 and 500 °C for (La0.8Sr0.2)MnO3, and 370 and 500 °C for (La0.8K0.2)MnO3. The peak observed at high temperature is attributed to the decomposition of nitrate (NO3−) adsorbed at the surface of the Mn sites. Even in the absence of oxygen in the gas feed, the Mn4+/Mn3+ and/or Mn3+/Mn2+ redox systems allowed the adsorption and the oxidation of NO2 into NO3 because of labile oxygen atoms of the network. The K-substituted perovskite showed the higher amount of nitrate species. The values obtained for this high temperature peak were 1.55 µmolNO m−2 for (La0.8K0.2)MnO3 and 0.78 µmolNO m−2 for (La0.8Sr0.2)MnO3. The total nitrate decomposition after the isotherm at 500 °C were 3.33 µmolNO m−2 for (La0.8K0.2)MnO3 and 2.11 µmolNO m−2 for (La0.8Sr0.2)MnO3. The hig her nitrate decomposition observed on K-based perovskite is attributed to the ability of alkali and alkaline earths elements to trap NOx in the form of nitrate [16].
The most interesting range is the low temperature range. Three peaks were observed for both samples and correspond to NO desorption. However, these three peaks are different in terms of their temperatures and areas. For the (La0.8Sr0.2)MnO3 sample, the main NO desorption peak occurred at a low temperature of around 95 °C. NO desorption continued, with a second peak observed at 135 °C and a third peak at 210 °C, until total desorption occurred at a temperature below 300 °C. The overall amount of NO desorbed at low temperature was 1.67 µmolNO m−2. However, an opposite trend was observed for the (La0.8K0.2)MnO3 sample. The intensity of the desorption peaks of NO are increased with increasing temperature. The maximum NO desorption occurred at 230 °C, which is 135 °C higher than the maximum desorption for (La0.8Sr0.2)MnO3 perovskite. Complete NO desorption was achieved at a much higher temperature of around 370 °C than for (La0.8Sr0.2)MnO3 perovskite. The overall amount of NO desorbed at low temperature was 4.10 µmolNO m−2. Thus, the (La0.8Sr0.2)MnO3 catalyst is more active for NO and C10H22 oxidation. This activity efficiency could be related to the weak bonding of NO adsorbed on the surface of Sr-substituted perovskites. Indeed, NO oxidation in the presence of O2 is related to the NO adsorption property. The maximum NO2 yield was observed at 280 °C for (La0.8Sr0.2)MnO3 and at 370 °C for (La0.8K0.2)MnO3. For both systems, this temperature corresponds to the transition between NO adsorption and nitrate formation observed during the TPD experiments. The weakly adsorbed NO could migrate from the A site to oxygen adsorbed on the B site, and thus f avour the catalytic activity toward NO oxidation. In this transition behaviour, the adsorption of NO and nitrate formation was negligible. This suggests that the distance between A and B sites is important for the oxidation reaction. The B sites are often claimed to be the active sites in perovskite because of the presence of the redox system (Co3+/Co2+, Mn4+/Mn3+, Mn3+/Mn2+).
To check if similar good performance could be obtained with a mixture of both NO and C10H22 reactants, (La0.8Sr0.2)MnO3 perovskite was tested under HC-SCR conditions and compared with the 2% Pt/SiO2 reference catalyst. The simulated gas mixture was composed of 240 ppm C10H22, 400 ppm NO, 9 vol% O2, and 1.5 vol% H2O. The C10H22 conversion and NO2, N2O, and N2 yields are shown in Fig. 5. On the 2% Pt/SiO2 catalyst, 50% C10H22 conversion (180 °C) was 50 °C higher than the corresponding temperature observed during the C10H22 oxidation test (127 °C). The presence of NO did not influence C10H22 oxidation over (La0.8Sr0.2)MnO3 perovskite. The T50 value was similar to that during experiments without NO (190 °C). Decane conversion reached 100% at around 200 °C for both samples. The inhibition of C10H22 oxidation on the Pt-based catalyst can be explained by its catalytic activity for SCR. Indeed, reduction of NO to N2O starts at the same temperature as C10H22 oxidation, and the maximum N2O production of 65% was observed at 195 °C when the C10H22 conversion reached 100%. Moreover, in the same temperature range (< 200 °C), N2 was produced in lower amount than N2O with a maximum yield of 18% at 210 °C. N2O and N2 produced by the reduction of NO by C10H22 indicate that the number of Pt active sites for C10H22 adsorption decreases in favour of NO adsorption. For temperatures higher than 200 °C, the N2 and N2O yields decreased and NO conversion to NO2 started and increased to reach a maximum of 3 7% at 370 °C. Compared with NO oxidation without C10H22, the maximum yield of NO2 was lower and shifted to higher temperatures (100% at 240 °C). In addition, even at 500 °C, the NO2 yield had not reached the thermodynamic equilibrium.
The catalytic behavior of the (La0.8Sr0.2)MnO3 perovskite was different to that of the Pt catalyst. For temperatures lower than 210 °C, the competition between C10H22 and NO was less important, and thus C10H22 oxidation was not inhibited by NO adsorption. The maximum N2O and N2 yields (∼13% at 210 °C) were observed when C10H22 conversion reached 100%. NO2 oxidation started at 200 °C and reached a maximum of 50% at 290 °C. Even though the maximum NO2 yield was lower than that for the NO oxidation test without C10H22, it was observed at approximately the same temperature (between 285 and 290 °C). The lower maximum NO2 yield for the NO/C10H22 mixed reactant system compared with the NO system can be explained by the competitive oxidation reactions of NO and C10H22 over Mn4+/Mn3+ and/or Mn3+/Mn2+ redox systems, and was not because of the competition between NO reduction and oxidation because the N2O and N2 yields remained low. These Mn4+/Mn3+ and/or Mn3+/Mn2+ redox systems are the active species involved in the Mars-van Krevelen mechanism, which is generally described for HC oxidation over mixed oxide samples [17, 18]. Nevertheless, compared with the Pt/SiO2 catalyst, this perovskite system allows the thermodynamic equilibrium to be reached at 500 °C with a NO2 yield of around 25%.
The (La0.8Sr0.2)MnO3 sample showed better oxidation properties for NO2 production than the 2% Pt/SiO2 catalyst. In addition, the SBET of the perovskite (18 m² g−1 was ca 19 times lower than the Pt catalyst (350 m² g−1). Thus, for experiments carried out with the same catalyst weight, the perovskite structure allows atomic scale interactions between active sites, and thus balances the SBET decrease. Such interactions are difficult to obtain by impregnation of noble metals at the surface of mixed oxides.
Among the substituted LaMnO3 perovskites studied, the (La0.8Sr0.2)MnO3 sample showed the best catalytic performances for C10H22 oxidation (T50 = 190 °C) and NO oxidation to NO2 (NO2 yield 80% at 285 °C). As a reference, the noble metal catalyst had a T50 value at 110 °C for C10H22 oxidation and a maximum NO conversion of 100% at 240 °C. The good catalytic performance of the Sr-substituted perovskite was mainly because of Sr2+ cation and Mn4+/Mn3+ and/or Mn3+/Mn2+ redox couple interactions, which made up for their low specific surface areas (18 m² g−1). The A site substitution by Sr was assumed to be the active site for NO and C10H22 adsorption, while Mn cations were involved in the oxidation mechanism through oxygen activation. The atomic scale interactions in the perovskite structure were assumed to be the main reason for the catalytic activity improvement. The test of the (La0.8Sr0.2)MnO3 sample with the HC-SCR gas mixture showed that the perovskite sample had better activity for NO oxidation than the Pt catalyst: 50% at 290 °C and 36% at 370 °C, respectively. Pt activity was inhibited by competition between NO reduction and C10H22 oxidation at temperatures lower than 200 °C, and between NO reduction and oxidation at higher temperatures. Thus, (La0.8Sr0.2)MnO3 perovskite is a good candidate for HC-SCR of NO by C10H22.
Thanks are due to the scientific service of IRCELYON for ICP and XRD analyses. Financial supports by national agency for research are gratefully acknowledged.
2014, Vol. 35 Issue (8): 1299-1304
PDF (734 KB)
2014, Vol. 35 Issue (8): 1299-1304
PDF (734 KB)