Journal of the Chilean Chemical Society
On-line version ISSN 0717-9707
J. Chil. Chem. Soc. vol.51 no.4 Concepción Dec. 2006
J. Chil. Chem. Soc., 51, N°.4 (2006), p.1001-1005
Catalytic combustion of methane over LaFeO3 perovskites:the influence of coprecipitation pH and ageing time.
GINA PECCHIa*, PATRICIO REYESa, RAÚL ZAMORAa, LUIS E. CADÚSb, BIBIANA P. BARBEROba:Departamento de Físico-Química, Facultad de Ciencias Químicas, Universidad de Concepción, Casilla 160-C, Concepción, Chile.
b:Instituto de Investigaciones en Tecnología Química (INTEQUI), UNSL-CONICET, Universidad Nacional de San Luis, Casilla 290, 5700 San Luis, Argentina.
The aim of this work was to study the effect of two precipitation pHs and three different ageing times on the preparation of simple LaFeO3 perovskites for their use as catalysts in the methane combustion. Absorption atomic spectroscopy (AAS), BET specific area measurements, XRD analysis, Raman, XPS, TGA and DTA analysis, TPR and TPO profiles revealed that significant changes in chemical composition, surface area and surface composition was produced at pH 8 when the ageing time was changed, leading to solids with different catalytic activity in methane combustion. At the higher precipitation pH studied, pH 10, the physical chemistry properties for all the samples obtained at different ageing times are almost identical, and therefore the solids exhibit similar activities in the combustion reaction. The catalysts were found to be active for methane combustion, and the effect of the perovskite’s structure and surface properties became more important.
Keywords: lanthanum, orthoferrite, perovskites, combustion, methane.
Catalytic combustion offers one of the most efficient means for controlling atmospheric pollution (1). Although noble metals, such as palladium and platinum, are well known with higher activity (per site) than metal-oxide catalysts, they present several disadvantages, such as higher volatility, high cost, and poor availability. Compared with noble metals, base metal catalysts present a lower but still sufficient activity as oxidation catalysts, and have the advantages of lower costs and the potential market in energy generation systems in domestic and small-scale industrial applications. For this reason, metal-oxide catalysts have received wide attention, particularly, hexa-aluminate and perovskite-type compounds (2, 3), which have even been incorporated into the design of novel combustors (4, 5).
The early lanthanide perovskites (particularly A = La, B = first-row transition metal) have received the most attention as oxidation catalysis since it was recognised that oxides could compete with metals in oxidation reactions (6, 7) of volatile organic compounds. Much attention has particularly been paid to lanthanum-transition metal-based perovskite oxides, which were introduced into catalysis some 30 years ago (8, 9). Simple perovskites with the basic formula ABO3 have been extensively investigated as catalysts (9, 10) for several reasons:transition metal ion contribution of unusual oxidation states, the amount of non-stoichiometric oxygen, and the structures of lattice defects (11) on the catalytic activity. The number of potentially interesting perovskites in the oxidation reactions is very great, owing to the number of A and B cations that can enter into this structure and the possibility to partially substitute the A and/or B position. However, the structural chemistry of perovskite lattice is very complex.In ABO3 structures, A is a large cation and B is a small cation of the d-transition series. These oxides consist of close-packed, ordered AO3 layers which are stacked one on top of the other and the B cations occupy all the interlayer oxygen octahedra. The stacking of an AO3 layer in the structure may be cubic or hexagonal with respect to its two adjacent layers. If the stacking is entirely cubic, the B-cation octahedra share only corners in three dimensions to form the perovskite structure. If the stacking is all-hexagonal, the B-cation octahedra share opposite faces forming chains along an axis and mixed oxides are formed (12). The ideal structure of the perovskite type oxide is of the cubic type (13). When some distortions of the cubic phase are present, the cation size mismatch and cation or anion non-stoichiometry is reached.For these highly oxidative systems it might be possible to relate bulk and surface defect chemistry to catalytic activity. Even though, the best catalytic performances in catalytic combustion are exhibited by La- or La-Sr-based perovskites containing Co, Fe or Mn as B cation (14, 15), the greatest limitation for their application is their low surface area that appears to be a serious disadvantage for their application. In order to increase the specific surface area, various preparation methods have been investigated on conventional ABO3 perovskites (16 - 19). In this work, we analysed six different perovskite samples of the same composition (LaFeO3) that were prepared with different procedures, and we compared them as catalysts for methane oxidation. The objective was to study the effect of the coprecipitation pH and ageing time on the simple LaFeO3 system to better understand the activity of these materials in methane catalytic combustion.
2.1. Preparation of the catalysts
A series of perovskite type LaFeO3 was prepared by coprecipitation of the respective nitrates. A 0.1 N aqueous solution of the corresponding nitrates was used with 0.1 N sodium hydroxide solution as the precipitant agent. The precipitation was carried out at room temperature at two different pH (8 and 10) and three different ageing times (4, 24 and 48 hours). The resulting precipitate was filtered off, washed with distilled water, to remove excess sodium ions up to constant conductivity values, and dried overnight at 383 K in air. Following drying, the precipitate was crushed and sieved to obtain the required particle size (< 200 mm). Finally, the ground precipitate was calcined at 973 K in air for 6 h. The samples are labelled as LaFeO3-X-Y, where X refers to the coprecipitation pH and Y to the ageing time expressed in hours.
Absorption atomic spectroscopy (AAS) was performed on a Perkin Elmer instrument model 3100 to determine the lanthanum and iron composition of the bulk. The specific surface areas were obtained by nitrogen adsorption at 77 K, evaluated using the BET equation on an automatic Micromeritics apparatus Model ASAP 2010 in the 0.05-0.995 relative pressure range. The Raman spectra were collected in a JASCO TRS600SZP multichannel monochromatic spectrometer using an Ar ion laser (514.5 nm) operated at 50 mW.X-ray diffraction patterns were obtained on a Rigaku diffractometer using a Ni-filter and CuKa1 radiation. The crystalline phases were identified by reference to the powder diffraction data (JCPDS-ICDD). XPS spectra were recorded using a SSX 100/206 photoelectron spectrometer from Surface Science Instruments with a monochromatized microfocused Al X-ray source. TGA and DTA studies were carried out in a Shimadzu DTA 50 and TGA 51 equipment under air flow (50 ml min-1), and heating at 10ºC/min from room temperature to 1000ºC.TPR and TPO experiments were performed in a TPR/TPD 2900 Micromeritics system with a thermal conductivity detector. The reducing gas was a mixture of 5%H2/Ar (40 cm3min-1), and a heating rate of 10 Kmin-1 was employed. The catalytic activity in the combustion of methane was evaluated in a conventional flow reactor at atmospheric pressure. In each experiment, 100 mg of catalysts were used diluted with 100 mg of silica as an inert. The activity was measured at different temperatures, increasing the temperature isothermally at a heating velocity of 1ºC min-1from 423K until reaching complete conversion. The reactant gaseous CH4:O2:He = 1:2:97 (molar) mixture was used. Reactor effluents were analysed using an on-line gas chromatograph Hewlett Packard model HP 4890D with thermal conductivity detector that works at a temperature of 423K and a current of 150mA. The column used was a 30-m capillary Supelco 25462. The carrier gas (He) flow through the column is 4ml min-1 and the injector temperature 423K. In some experiments, a Quadrupole Mass Spectrometer Hiden HPT 20 was used to detect small traces of products.
4. 3. Results and Discussion
Table 1 summarises the chemical analyses of iron and lanthanum in the LaFeO3 samples by AAS. For the series prepared at pH 10, the Fe and La content are the expected values according the stoichiometry, 23.1 wt% for iron and 57.2 wt% for La. An acceptable agreement is also displayed for the sample prepared at pH 8 with 4 h of ageing. However, for this pH, at a higher ageing time, an enhancement in the Fe loading is observed indicating that lanthanum species may be partially dissolved. The segregation of iron in the perovskites prepared at pH 8 was observed to increase with the ageing time.This result can be explained considering the differences in solubility’s of the hydroxides (pK Fe(OH)3,= 37.4, pK La(OH)3= 20.0).The lower solubility of Fe(OH)3 indicates that the iron segregation should be the consequence of a partial solubilisation of lanthanum.As a result, even though a buffer at pH=8 was used in the thermostat bath, when the ageing time was increased, the solution of the respective hydroxides are exposed to atmosphere, where due to CO2 dissolution, H2CO3 is generated, which is responsible for the partial neutralization of the medium’s baseness and increases the solubility of the less insoluble hydroxide. Thus, the surface lanthanum hydroxides are dissolved and the iron hydroxides species can be segregated. These results are in line with those obtained by XRD and XPS techniques.
Figure 1 displays the evolution of the BET surface areas with the ageing time for both studied series. It can be seen that in the series of perovskites prepared at pH 8 the samples exhibit low surface area, close to 14 m2g-1 with almost no changes with ageing time. In the perovskites prepared at pH 10 an important drop in the surface area with ageing time can be observed. Thus, at 4 h of ageing the BET surface area is 78 m2g-1 and after 48 h of ageing time the area decrease to 23 m2g-1. The main fact related with this drop is the sintering of the oxides, process which is thermodynamically favoured. This process is not significantly detected in the series prepared at pH 10 because the starting sample (4 h of ageing) already posses a low surface area. Additionally, a sample composition modification during ageing may also be considered as a cause of decrease in surface area. At pH 8, both cations should precipitate with OH- anions.However the quantity of Fe3+ largely exceeds the necessary requirement for precipitation, while only a slight excess of OH- is present for the precipitation of La3+. Therefore, it is likely that when a sintering of the oxide species occurs as ageing time increases, partial dissolution of lanthanum hydroxides takes place due to the interaction with dissolved CO2 generating carbonate species, which are more soluble than the hydroxides ones, leading to an excess of iron species in the solid. Upon calcination at 973 K in those solids with longer ageing time, surface iron oxide species, initially isolated by dissolution of lanthanum oxide species, can migrate leading to segregated Fe2O3 crystallites as detected by Raman and XRD. As a consequence of this sintering, the surface decreases. In the pH 10 series, the amount of OH concentration required for precipitation is largely exceeded, resulting in a solid oxide mixture with similar characteristics in which dissolution is not likely and therefore no significant differences should be observed in the ageing period and after calcination.
Raman studies seem to support the previous explanation. Figure 2 show the Raman spectra of the studied solids. A band centred at 290 cm-1 is attributed to Fe2O3, which appears in the samples prepared at pH 8 with 24 and 48 h of ageing.This signal does not appear in the Raman spectra of the samples obtained at the same pH but with only 4 h of ageing. As discussed previously, the presence of segregated Fe2O3 on the perovskite structure is expected for the perovskites prepared at lower pH and larger ageing time.
Figure 3 presents the XRD patterns of the studied samples.It can be clearly seen that all the solids display the expected perovskite structure. In the samples prepared at pH 8, a decrease in the degree of crystallinity appears in all the samples, where the presence of diffraction lines at 2q values 33.2-33.5° for Fe2O3-JCPDS 84-0311 is noticeable, particularly in the sample after 24 h of ageing. In fact for 48 h of ageing, it is likely that iron oxide exists as an amorphous form and therefore does not present the characteristic diffraction lines. In the pH 10 series, a large extent of crystalline degree species, whose lines are the expected for LaFeO3-JCPDS 75-0541. Similar findings were obtained from the electron diffraction studies (data not shown).
XPS studies of LaFeO3 samples were carried out to characterise the catalysts surface. Table 1 summarises the Fe/La and O/La atomic surface ratios. Fe 2p3/2 and La 3d5/2 presented the expected core level spectra with BE of 710.0 and 834.0 eV respectively. In the pH 8 series, the Fe/La surface atomic ratios increases with ageing time from 0.51 up to 1.04 due to a segregation of iron oxide species on the surface as discussed in the previous section.In the pH 10 series, the Fe/La surface atomic ratios remain almost constant, close to 0.5, indicating a surface enrichment of lanthanum.This result is because the hydroxide of lanthanum is a very basic compound, which at this pH, can carbonate easily, producing the surface excess of lanthanum. This effect is lower in the pH 8 series, the minimum condition to achieve precipitation of both hydroxides.In this series, as ageing time increased, surface enrichment of lanthanum was found to decrease, an effect that is explained considering that as ageing time increased CO2 adsorption is favoured and consequently H2CO3 formation, which slightly diminishes the pH of the solution, avoiding later carbonation and lanthanum segregation. Figure 4 shows the O 1s core level spectra of some of the studied catalysts. The O 1s peak shows two maxima at 528.2 -529.6 eV and at 530.9 eV. The peak at lower BE is due to O2- species, whereas the one at higher BE is attributed to less electron-rich species, such as OH-. The intensity of both species differs from sample to sample. In those prepared at pH 10, the proportion is similar; whereas in the pH 8 series, differences in the intensity of the mentioned peaks exist, with the peak at 530.9 eV is larger for the sample pH 8 and 24 h of ageing, in line with the presence of hydroxilated Fe2O3.
Figure 5a and 5b display the TGA and DTA curves for one representative samples of each series with 48 h of ageing time. Very similar thermograms were found for the samples precipitated at pH 10, one is presented in Fig 5.a. They exhibit a significant weight loss, approximately 20%, between room temperature and 473 K, which is attributed to the water evolution. At higher temperatures, between 473 and 873 K, a continuous weight loss is produced, reaching approximately 10% weight loss in that range. No significant weight loss was observed at temperatures higher than 873K. DTA curves of these samples exhibit essentially only one endothermic peak in the range, centred at 373 K due to the water evolution. With regard to the samples precipitated at pH 8 (curves b), the TGA curves present similar trends when compared with the pH 10 series, with the weight loss close to 20% up to 473 K. Further temperature increases produce changes in the TGA curves which depend on ageing time. The sample with 4 hour of ageing presents a thermogram virtually similar to those for the samples precipitated at pH 10, with a continuous weight loss up to 873 K, where the weight loss is close to 10% from 473 to 873 K.As ageing time increases, two different regions are observed. The first from 473 to 950 K with a rather constant slope, followed by a the drastic weight loss in the temperature range 950-990 K, and then a continuous decrease in the weight up to 1200 K.In these samples, the DTA presents two peaks:one endothermic close to 373 K due to water evolution and one exothermic in the temperature range 950-990 K.The observed behaviour can be understood considering that during the precipitation of the corresponding hydroxides, a very intimate contact is obtained in the samples precipitated at pH 10. This contact does not produces a significant segregation during ageing time, and as a consequence of the calcination, the samples are continuously decomposed into the oxides and therefore converted into the corresponding perovskites. Chemical analysis confirms this finding giving Fe and La contents close to the stoichiometric values. However, surface composition obtained from XPS indicates that the Fe/La values are lower than expected, suggesting a surface enrichment in lanthanum speciesUnpublished results indicate that under these conditions, the perovskite -LaFeO3 structure is obtained at temperatures as low as 873 K.In the samples precipitated at pH 8, the stoichiometric proportion of the solids is not completely reached because iron can be completely precipitated but lanthanum ions cannot.Therefore, it is likely that a heterogeneous distribution of the corresponding hydroxides exists mainly in the aged samples. Therefore, during calcination, migration of the corresponding oxides should occur prior to perovskite structure formation. This process occurs at temperatures around 950 K, and is responsible for the appearance of an exothermic peak in the DTA curves.
Temperature programmed reduction cycles were performed starting from the calcined samples, which were first reduced under H2/Ar flow (TPR1) up to 973 K.Then they were cooled down in Ar flow to room temperature and the O2/He flow was switched to the reactor. Once the base line was restored, a temperature programmed oxidation (TPO) was carried out up to 973 K.Then, the sample was cooled down again and a second reduction treatment was performed (TPR2). This procedure was selected due to the basic characteristic of the solids, mainly in those samples prepared at pH 10, which produces an oxide surface covered by carbonate species, and therefore the TPR1 profile is more complex and different from TPR2. With regard to TPO, they are very similar showing a single peak centred at 590 and 610 K for the series prepared at pH 10 and 8, respectively. Figure 6 presents the TPR2 profiles obtained for the prepared catalysts. As can be seen, in the pH 10 series two well-defined reduction peaks centred at 385 and 480 °C appear, where the second one has higher intensity. This behaviour indicates the removal by reduction of two types of oxygen (Fe-O and La-Fe-O) which are reduced at two different temperatures (20) in agreement with previous results. With regard to the TPR2 of the pH 8 series, the profiles are different: the first peak is centred 675 K, and the presence of second reduction steps can also be observed. The species involved in this reduction stage are not completed reduced up to 973K. This fact may be understood if one bears in mind that the presence of Fe2O3 in these samples has been revealed by other techniques as discussed previously. It is known that Fe2O3 reduction takes place in two steps:Fe2O3 to Fe3O4 and Fe3O4 to Fe, with two peaks appearing at 350 and 730°C, which should appear partially overlapped with the reduction of LaFeO3.
The prepared catalysts were evaluated in the methane oxidation in a flow reactor using a stoichiometric oxygen/methane ratio. Only carbon dioxide and water were the observed products. Traces of CO were detected at conversion level higher than 70%. As comparison Fe2O3, La2O3 and a physical mixture Fe2O3+LaFeO3 (in a proportion 76wt% of LaFeO3 and 24wt% Fe2O3) were also evaluated. The ignition temperature defined as the temperature required for 50% of conversion and the reaction rate expressed as mole converted per second and gram of catalyst evaluated at 773 K are given in Table 2. More important changes in the pH 8 series, in which the ignition temperature increases with the ageing time, were detected. In the pH 10 series, almost no changes were observed, attributed to a more crystalline structure and the absence of Fe2O3 species. The activity exhibited by the iron oxides and the mixture Fe2O3 and LaFeO3 is lower in comparison with the perovskite, whereas La2O3 did not show activity under the same conditions. The reaction rates in the methane combustion at 773 K show more significant differences, being the pH 8-24 h the more active. To understand the observed behaviour, the physical chemistry characteristics of the solids should be considered. The changes in the activity of these LaFeO3 simple perovskites should be dependent on the surface area, crystalline structure and surface compositions. An increase in the surface area should lead to an increase in the activity.However, as can be seen in Figure 1, a decrease in the surface area in this series with ageing time is parallel to an increase in iron content. It should be mentioned that the catalytic behaviour of the samples are in the range reported previously in the same reaction (19, 21). Table 2 also compiles the reaction rate at 773 K referred to gram and to m2 of catalysts.As expected no changes in the reaction rate of the LaFeO3 pH 10 series, due to these samples posses similar physical chemistry properties, surface area, bulk and surface atomic ratios. Conversely, more important differences were detected in the LaFeO3 pH 8. Comparing the specific reaction rate, referred to g of catalyst, the values show a maximum for the 24 h of ageing, being the one with higher ageing time equivalent to the pH 10 series.However, if the reaction rate is referred to the surface area,an increase in this property can clearly be observe, being indicative that an enhancement in the crystallinity is even more important than the displayed surface area.
The study of the effect of preparation variables, such as coprecipitation pH and ageing time, on the surface and catalytic properties of LaFeO3 perovskites found that at pH 8, the ageing time produces significant changes in chemical composition, surface area and surface composition leading to solids with different catalytic activities in methane combustion. On the other hand, at higher precipitation pH (pH 10), the physical chemistry properties are almost independent of ageing time, and therefore the solids exhibit similar activities in methane combustion.
The authors thank CONICYT (Fondecyt Grant 1060702) from Chile and CONICET and the Universidad Nacional de San Luis from Argentina for financial support. The Raman spectra were acquired thanks to a grant from the Japanese International Cooperation Agency to CENACA (Argentina). The authors also thank Pierre Eloy (Université Catholique de Louvain-la-Neuve, Belgium) for his help with the XPS analysis.
1. T.V. Choudhary, S. Banerjee, V.R. Choudhary. Appl. Catal. A. 234 (2002) 1. [ Links ]
2. M. O´Connell, A.K. Norman, C.F. Hüttermann, M.A. Morris, Catal. Today, 47 (1999) 123. [ Links ]
3. L.J. Tejuca, J.L.G.Fierro. "Properties and Applications of Perovskite-Type Oxides" Dekker, New York, 1993. [ Links ]
4. R.J.H. Voorhoeve, Advanced Materials in Catalysis" (H.H.Burton and R.L.Garten, Eds.), p.127. Academic Press, New York, 1977. [ Links ]
5. S. Irusta, M.P.Pina, M. Melendez, j. Santamaría, Catal. Lett., 54 (1998) 69. [ Links ]
6. R.J.H. Voorhoeve, J.P. Remika, P.E., Freeland, B.T. Mathias, Science, 177 (1972) 353. [ Links ]
7. N. Yamazoe, Y. Teraoka, K. Eguchi, T. Seiyma. Appl. Catal. 26 (1986) 265. [ Links ]
8. G. Sarcco, I. Cerni, V. Specchia, R. Accornero, Chem. Eng., Sci., 54 (1998) 3599. [ Links ]
9. K. Poplawski, J. Lichtenberger, F. J. Keil, K. Schnitzlein, M. D. Amiridis. Catal. Today, 62 (2000) 329. [ Links ]
10. N. Yamazoe, Y. Teraoka. Catal. Today, 8 (1990) 176. [ Links ]
11. Z. Yu, L. Gao, S. Yuan, Y. Wu, J. Chem. Soc. Faraday Trans., 88 (1992) 3245. [ Links ]
12. C.N.R. Rao, J. Gopalakrishnan. "New Directions in Solid State Chemistry". p. 53. Cambridge University Press, Cambridge, 1997. [ Links ]
13. L.G. Tejuca, J.L.G.Fierro, J.M.D. Tascón, Adv. Catal., 36 (1989) 237. [ Links ]
14. H. Arai, T. Yamada, K. Eguchi, T. Seyama., Appl. Catal., 26 (1986) 265. [ Links ]
15. J.G. Mc Carty, H. Wise, Catal. Today, 8 (1990) 231. [ Links ]
16. S. Cimino, L.Lisi, R. Pirone, G. Russo, M. Turco, Catal. Today 59 (2000) 19. [ Links ]
17. P. Ciambelli, V. Palma, S.F. Tikhov, V.A. Sadykov, L.A. Isupova, L.Lisi, Catal. Today, 47 (1999) 199 [ Links ]
18. C. Oliva, L. Forni, Catal. Commun., 1 (2000) 5 [ Links ]
19. S. Cimino, S. Coloma. S.De Rossi, M. Faticanti, L. Lisis, I. Pettiti, P. Porta, J. Catal., 205 (2002) 309. [ Links ]
20. S: Varma, B.N. Wani, N. M. Gupta, Appl. Catal. A. 241 (2003) 341. [ Links ]
21. K-Sup Song, H. Xing Cui, S. Don Kim, S-Kyu Kang, Catal. Today, 47 (1999) 155. [ Links ]
T: To whom correspondence should be addressed. email: firstname.lastname@example.org