Applied Catalysis B: Environmental 33 (2001) 193–203
La, Ca and Fe oxide perovskites: preparation, characterization and catalytic properties for methane combustion P. Ciambelli a , S. Cimino b , L. Lisi c , M. Faticanti d , G. Minelli d , I. Pettiti d , P. Porta d,∗ a
c
Dipartimento di Ingegneria Chimica e Alimentare, Università di Salerno, Napoli, Italy b Dipartimento di Ingegneria Chimica, Università “Federico II”, Napoli, Italy Istituto di Ricerche sulla Combustione, CNR, c/o Dipartimento di Ingegneria Chimica, Piazzale V. Tecchio 80, 80125 Napoli, Italy d Centro di Studio del CNR su “Struttura e Attività Catalitica di Sistemi di Ossidi” (SACSO), c/o Dipartimento di Chimica, Università “La Sapienza”, Piazzale A. Moro 5, 00185 Roma, Italy Received 15 October 2000; received in revised form 6 January 2001; accepted 18 February 2001
Abstract La1−x Cax FeO3 (x = 0.1, 0.2, 0.3, 0.4, 0.5) perovskites prepared by citrate method and calcined at 1073 K have been investigated as catalysts for methane combustion. The formation of the perovskite structure has been shown by X-ray diffraction (XRD) for all samples. The surface area (SA) values are in the range 3–6 m2 g−1 for the samples up to x = 0.4, whereas SA is 0.7 m2 g−1 for the specimen with x = 0.5. The atomic susceptibility increases with increasing in Ca content. The Fe4+ /Fetotal ratio has been determined by both redox titration and TPR analysis. The amount of Fe4+ enhances with Ca substitution but the Fe4+ /Ca2+ ratio is quite constant giving rise to an increasing formation of oxygen vacancies to preserve charge neutrality. All perovskites show a similar intrinsic activity in CH4 combustion with activation energy, Ea , of about 23 kcal mol−1 . A slightly lower value of Ea is shown by LaFeO3 which, in contrast with Ca substituted samples, exhibits a small cationic defectivity. Methane reaction order lower than one was found for all catalysts. © 2001 Elsevier Science B.V. All rights reserved. Keywords: Catalytic methane combustion; Perovskite solid solutions; La, Ca orthoferrite catalysts
1. Introduction The good activity shown in catalytic combustion processes of perovskite-type oxides, coupled with a high thermal stability, is the main driving force to investigate these materials as catalysts which can potentially replace noble metals for complete oxidation of hydrocarbons [1,2]. Perovskite (CaTiO3 as mineral) oxides have general formula ABO3 , where the 12-coordinated A sites may be occupied by rare-earth, alkaline-earth, alkali or other large ions and the 6-coordinated B sites are ∗ Corresponding author. Fax: +39-06-490324. E-mail address:
[email protected] (P. Porta).
usually filled with transition metal cations. Due to the great stability of the perovskite framework a large number of metallic trivalent cations can occupy the A and the B sites √ provided that the tolerance factor t [t = (r A + r O )/ 2(r B + r O )] is in the range 0.8–1.0 [2]. Moreover, the perovskite composition can be widely changed by substituting either or both A and B site cations with other metals which can also have an oxidation state different from 3+. In this case, formation of structural defects such as anionic or cationic vacancies and/or change in the oxidation state of the transition metal cation arise in order to maintain the electroneutrality of the compound. The above occurrence strongly affects the redox properties of the catalyst. Also a not substituted perovskite such as LaMnO3
0926-3373/01/$ – see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 6 - 3 3 7 3 ( 0 1 ) 0 0 1 6 3 - 1
194
P. Ciambelli et al. / Applied Catalysis B: Environmental 33 (2001) 193–203
may contain Mn4+ (more or less abundant in dependence of the preparation conditions used) in addition to Mn3+ , thus exhibiting (to compensate the total charge) a cationic defectivity [1]. It is reported [3,4] that La1−x Srx MnO3 perovskites show an increase of Mn4+ concentration with the increase in Sr content, resulting in an enhanced catalytic activity for methane combustion compared to LaMnO3 . Otherwise, the Sr substitution for La leads either to the formation of oxygen vacancies in LaCoO3 , or both to the partial oxidation of Fe3+ to Fe4+ and the formation of oxygen vacancies in LaFeO3 [5]. Substitution in B position has also been investigated for LaFeO3 perovskite [6] by partially replacing iron with magnesium which has a very close ionic radius. In that case the amount of Fe4+ reaches a maximum for x = 0.2 whereas, at the same time, oxygen vacancies decrease with Mg substitution. Substitution in A position with a cation showing the same basic character of magnesium has not been studied. Calcium could be a good candidate also considering the similarity of its ionic radius with that of La3+ which could provide a higher stability of the perovskite structure with respect to strontium substitution. Therefore, in this work, La1−x Cax FeO3 (x = 0.1, 0.2, 0.3, 0.4, 0.5) perovskites were investigated as catalysts for methane combustion. The stability of the perovskite structure upon Ca substitution and the effect on the oxidation state of iron and on the formation of structural vacancies due to the replacement of La3+ with a lower charge cation were studied.
2. Experimental Samples with nominal composition La1−x Cax FeO3 (x = 0.0, 0.1, 0.2, 0.3, 0.4 and 0.5) were prepared starting from nitrate solutions of the metals in the appropriate molar ratio. A solution of citric acid with the same amount of equivalents was added and the final mixture was slowly evaporated at about 340 K. The slurry was dried at 383 K for 15 h and then, after grinding, calcined at different temperatures, 623, 773, 923 and 1073 K, for 5 h each. The chemical composition was determined by inductively coupled plasma emission spectroscopy (ICP). Atomic absorption (AA) spectroscopy was used to determine the iron and calcium content in
the x = 0.2 and 0.4 samples too, as a control. In these cases a mean value between ICP and AA was reported. The amount of tetravalent iron contained in the samples treated at 1073 K was determined by redox reaction: Fe4+ + Fe2+ → 2Fe3+ . A weighed amount of sample was dissolved in a known excess of a 0.0512 N Mohr salt solution by addition of a 4 N H2 SO4 solution and back titration of the remaining Fe2+ equivalents with a 0.0702 N K2 Cr2 O7 aqueous solution, using ferroin as an indicator. The titration was performed twice for each sample, the difference between the two determinations being in all cases within 5%. Phase analysis, lattice parameters and particle sizes were determined by X-ray diffraction (XRD) using a Philips PW 1729 diffractometer equipped with an IBM PS2 computer for data acquisition and analysis (software APD-Philips) and with Ni-filter Cu K␣ radiation. Lattice parameters were calculated from the reflections appearing in the 2θ = 20–60◦ range, using the UNITCELL software program [7]. The procedure was a least-squares refinement, the hkl indexes being assigned to the orthorhombic GdFeO3 -type structure with Pnma, N.62, space group symmetry [8]. Particle sizes (D) were calculated by means of the Scherrer equation D = Kλ/βcos θ after Warren’s correction for instrumental broadening. K is a constant equal to 0.9, λ the wavelength of the X-ray used, β the effective linewidth of the observed X-ray reflection, calculated by the expression β 2 = B 2 − b2 (where B is the full width at half maximum (FWHM), b the instrumental broadening determined through the FWHM of the X-ray reflection at 2θ ≈ 28◦ of crystalline SiO2 with particles larger than 1000 Å), θ the diffraction angle of the (1 2 1) considered X-ray reflection (2θ ≈ 32◦ ). BET surface areas (SA) of the materials calcined at 1073 K were measured by Kr adsorption at 77 K, in a volumetric all glass apparatus. Measurements of magnetic susceptibility were performed for the samples x = 0.0, 0.2, and 0.5 calcined at 1073 K by the Gouy method over the temperature range 100–300 K and at different magnetic field strengths. Correction was made for the diamagnetism of the samples. Measurements of diffused reflectance UV–VIS spectroscopy (DRS) were performed on the samples calcined at 1073 K in the wavelength range 200– 800 nm (50000–12500 cm−1 ) with a Varian CARY 5E
P. Ciambelli et al. / Applied Catalysis B: Environmental 33 (2001) 193–203
195
Table 1 La1−x Cax FeO3 catalysts calcined at 1073 Ka Sample x x x x x x
= 0.0 = 0.1 = 0.2 = 0.3 = 0.4 = 0.5
V
D
SA
La
243.2 241.8 239.7 237.4 234.7 232.9
570 636 613 527 406 418
3 6 5 3 5 0.7
57.8 55.3 51.2 46.2 41.8 37.8
Ca (57.2) (53.7) (49.8) (45.6) (41.0) (35.9)
– 2.8 4.0 5.9 7.3 10.3
Fe4+ /Fetot
Fe (1.7) (3.6) (5.6) (7.9) (10.4)
20.7 22.0 23.0 23.7 25.0 26.2
(23.0) (24.0) (25.0) (26.2) (27.5) (28.9)
1.4 7.0 15.0 20.6 26.3 29.2
a Unit cell volume V (Å3 ). Crystallite dimension D (Å). BET surface area SA (m2 g−1 ). La, Ca and Fe metal content, weight percent (nominal values in parentheses). Fe4+ /Fetot atomic percent, evaluated by chemical analysis.
UV–VIS–NIR spectrophotometer equipped with an IBM PS2 computer (software Varian Cary05E). Powder samples were pressed on a flat disk. Measurements were performed at room temperature with a scan rate equal to 150 nm min−1 and data interval 0.75 nm. Temperature-programmed reduction (TPR) experiments were performed as reported in [9] using a Micromeritics TPD/TPR 2900 analyzer equipped with a TC detector and coupled with a Hiden HPR 20 mass spectrometer. Samples (100 mg) were preheated
in flowing air at 1073 K for 2 h and then, after cooling at room temperature, reduced with a 2% H2 /Ar mixture (25 cm3 min−1 ) heating 10 K min−1 up to 1073 K. Water produced by the sample reduction was condensed in a cold trap before reaching the detectors. Only H2 was detected in the outlet gas confirming the effectiveness of the cold trap. XRD measurements were performed at the end of the TPR runs. Methane catalytic combustion experiments were performed in the experimental apparatus and accord-
Fig. 1. XRD spectra for the La1−x Cax FeO3 samples calcined at 1073 K.
196
P. Ciambelli et al. / Applied Catalysis B: Environmental 33 (2001) 193–203
ing to the procedures already described in [9]. The space velocity was 40000 Ncm3 g−1 h−1 in all tests (0.4 g catalyst powder), and the feed gas composition was 0.4% CH4 , 10% O2 , N2 as balance. Carbon balance was closed within 1% in all catalytic tests. Reproducibility of the results was verified carrying out each test twice and loading fresh catalyst.
3. Results La1−x Cax FeO3 (x = 0.0, 0.1, 0.2, 0.3, 0.4 and 0.5) samples were calcined at different temperatures but those at 1073 K have been studied in more detail. Table 1 reports the values of unit cell volume, particle size, surface area, weight percent of metal content, and atomic percent of Fe4+ /Fetotal (determined by redox titration). The nominal metal content (reported in parenthesis) agreed with the experimental one. XRD analysis revealed that the samples calcined at 1073 K (Fig. 1) are single perovskite phase. Note that the materials with x up to 0.3 were found to be crystalline perovskite phases after thermal treatment
at 623 K, whereas the x = 0.4 and 0.5 samples were obtained as well crystallized perovskites only after treatment at 923 K. The values of particle size (at 1073 K), D, decrease when a high amount of calcium (x ≥ 0.3) replaces lanthanum in the A sites of the perovskite lattice. The substitution of lanthanum with calcium produces a continuous decrease of the unit cell volume (Fig. 2, Table 1), as well as of the three cell axes. The surface area values for all the samples calcined at 1073 K, except x = 0.5 (SA = 0.7 m2 g−1 ), are in the range 3–6 m2 g−1 . Reflectance spectra are reported in Fig. 3 and their features will be later discussed. TPR profiles are reported in Fig. 4. All catalysts undergo a reduction starting at quite low temperature whose extent strongly increases with Ca substitution. A further reduction starts at high temperature (T > 923 K) only for samples with x ≥ 0.2. The first signal, appearing as a weak single peak for the unsubstituted perovskite, could be the result of the overlapping of two peaks for the Ca substituted samples. This signal starts at about 423 K for all perovskites,
Fig. 2. Unit cell volume vs. substitution parameter x for the La1−x Cax FeO3 samples calcined at 1073 K.
P. Ciambelli et al. / Applied Catalysis B: Environmental 33 (2001) 193–203
Fig. 3. Reflectance spectra for the La1−x Cax FeO3 samples calcined at 1073 K. (a) LaFeO3 ; (b) and (b ): x = 0.1 before and after TPR reduction; (c) and (c ): x = 0.2 before and after TPR reduction; (d) and (d ): x = 0.3 before and fter TPR reduction.
except La0.5 Ca0.5 FeO3 which shows a peak starting at 473 K with a maximum at about 723 K. The second signal (not completely detected) starts at lower temperature with increasing in calcium substitution up to x = 0.4. The intensity of this peak also enhances with calcium substitution, except for the La0.5 Ca0.5 FeO3 sample that does not follow this trend. The H2 uptake
197
Fig. 4. TPR profiles of La1−x Cax FeO3 samples calcined at 1073 K and of ␣-Fe2 O3 . Bar corresponds to 1 or 10 mol H2 g−1 K−1 for perovskites or for ␣-Fe2 O3 , respectively.
resulting from the integration of the first peak only and of both peaks, and the temperatures of both the onset and the maximum H2 consumption are reported in Table 2. The Fe4+ /Fetotal ratio deduced from TPR curves is also reported in Table 2.
Table 2 La1−x Cax FeO3 catalysts calcined at 1073 Ka Sample x x x x x x
= 0.0 = 0.1 = 0.2 = 0.3 = 0.4 = 0.5
Chemical formula
Total uptake H2 /perovskite
H2 /Fe (low T)
Tonset
Tmax
Fe4+ /Fetot
Fe4+ /Ca2+
LaFe0.013 4+ Fe0.987 3+ O3+0.0065 La0.9 3+ Ca0.1 2+ Fe0.068 4+ Fe0.932 3+ O3−0.016 La0.8 3+ Ca0.2 2+ Fe0.142 4+ Fe0.858 3+ O3−0.029 La0.7 3+ Ca0.3 2+ Fe0.208 4+ Fe0.792 3+ O3−0.046 La0.6 3+ Ca0.4 2+ Fe0.262 4+ Fe0.738 3+ O3−0.069 La0.5 3+ Ca0.5 2+ Fe0.310 4+ Fe0.690 3+ O3−0.095
0.017 0.063 0.197 0.312 0.472 0.571
0.007 0.034 0.071 0.104 0.131 0.155
423 423 423 423 423 473
560 609 623 649 658 727
1.3 6.8 14.2 20.8 26.2 31.0
– 0.68 0.71 0.69 0.65 0.62
Chemical formula deduced from the low temperature TPR peak. Total (mol H2 mol−1 perovskite) and partial (low temperature) (mol H2 mol−1 Fetot ) H2 uptake. Onset and maximum temperature (K) uptake of TPR experiments. Fe4+ /Fetot atomic percent, evaluated by TPR. Fe4+ /Ca2+ ratio (mol mol−1 ). a
198
P. Ciambelli et al. / Applied Catalysis B: Environmental 33 (2001) 193–203
Fig. 5. XRD spectra for the La1−x Cax FeO3 samples after TPR experiments; (∗) metallic Fe; (#) CaO.
Fig. 6. CH4 conversion as a function of reaction temperature and Arrhenius plots referred to the surface reaction rate for La1−x Cax FeO3 samples.
P. Ciambelli et al. / Applied Catalysis B: Environmental 33 (2001) 193–203
199
Table 3 La1−x Cax FeO3 catalysts calcined at 1073 Ka Sample
x x x x x x
= 0.0 = 0.1 = 0.2 = 0.3 = 0.4 = 0.5
Ea
21.07 22.59 22.60 22.62 22.57 21.38
Pre-exp. factor × 108
T10
0.41 0.86 0.94 1.17 0.87 0.14
698 712 707 699 711 778
T50
802 816 810 798 814 909
T90
890 900 893 878 897 1026
RRb (773 K)
2.86 2.21 2.40 2.96 2.62 0.82
SRRb (773 K)
1.0 0.37 0.48 1.0 0.52 0.82
n 783
823
873
0.77 0.80 0.74 0.80 0.70 –
0.77 0.82 0.78 0.81 0.72 –
0.76 0.82 0.80 0.82 0.76 –
a Activation energy E (kcal mol−1 ). Pre-exponential factor (L g−1 h−1 ). Temperatures corresponding to 10, 50 and 90% CH conversion. a 4 Reaction rate referred to catalyst weight, RR (mmol g−1 h−1 ) and to catalyst surface, SRR (mmol m−2 h−1 ). CH4 reaction order (n) at 783, 823 and 873 K. bY methane = 0.004, Yoxygen = 0.1.
XRD measurements were performed at the end of the TPR experiments. As shown in Fig. 5 the perovskite structure is always preserved after reduction. However, starting from the x = 0.2 sample, a certain fraction of iron ions is reduced to metallic iron whose content increases with the increase in calcium content. The x = 0.5 sample revealed the presence of CaO in addition to perovskite and metallic iron. Methane conversion as a function of the reaction temperature is reported in Fig. 6; the Arrhenius plots were normalized with respect to the catalyst surface area (SRR), while the temperature values corresponding to 10, 50 and 90% conversion are reported in Table 3. In the experimental conditions investigated, all catalysts but La0.5 Ca0.5 FeO3 show a complete conversion of methane below 923 K with total selectivity to CO2 over the whole range of temperatures. The original activity of LaFeO3 decreases when a small fraction (x = 0.1) of calcium substitutes lanthanum, nevertheless, it increases for further La substitution up to x = 0.3 as indicated by the values of reaction rate (RR) referred to the catalyst weight, evaluated at 773 K and reported in Table 3. The La0.5 Ca0.5 FeO3 sample shows a markedly lower activity. However, if the poor value of surface area for this catalyst is taken into account (see SRR in Table 3), its specific activity is comparable with that of the other perovskites.
4. Discussion In the LaFeO3 orthoferrite, La3+ and Fe3+ (highspin configuration, HS) ions occupy the A (12-fold
coordinated) and the B (6-fold coordinated) sites, respectively, of the ABO3 perovskite lattice. A small amount (1.3%) of Fe4+ is also present in LaFeO3 . Replacement of lanthanum with calcium in the A sites is feasible owing to the similarity between the ionic radii of the two species in 12-fold coordination, 1.36 Å for La3+ and 1.34 Å for Ca2+ [10]. This replacement induces a structural disorder which probably provokes a delay in the crystallites growth during the thermal treatment. At 1073 K the temperature is high enough to reduce this effect which, however, remains evident in the samples with high calcium content. In fact, at 1073 K, the x = 0.4 and 0.5 samples showed smaller crystallite dimensions than all the others. This result is apparently in opposition to that revealed by measurements of surface area, which showed the x = 0.5 sample as the one with the lowest SA value. The fact could be explained in the hypothesis that the x = 0.5 sample was casually affected by some unusual sintering processes during the preparation or that other factors as nonstoichiometry and crystalline defects of the sample may contribute to the broadening of XRD reflections. In the La1−x Cax FeO3 solid solutions the lower charge of Ca2+ is compensated by partial oxidation of Fe3+ to Fe4+ . The presence of Fe4+ is responsible for the decrease of the unit cell volume in the calcium containing samples (Fig. 2). In fact, while Ca2+ and La3+ have similar ionic radii, the difference between the ionic radius of HS Fe3+ (0.645 Å) and Fe4+ (0.585 Å) [10], in octahedral coordination, gives rise to the lattice shrinkage. It is also interesting to note that the decrease in the unit cell dimensions is
200
P. Ciambelli et al. / Applied Catalysis B: Environmental 33 (2001) 193–203
isotropic along all the three unit cell axes which were found to decrease monotonically with the increase in calcium content. Optical spectra confirm the evidence of a partial oxidation of Fe3+ to Fe4+ when Ca replaces La in the La1−x Cax FeO3 catalysts. For HS Fe3+ the free ion ground term 6 S state gives rise, in an octahedral field [(t2g )3 (eg )2 configuration], to the 4 T1g , 4 T2g , 4 A1g , 4 A , and 4 E quartet states. Because of the great 2g g oxidizing power of Fe3+ , the ligand to metal charge transfer (CT) bands often obscure the low intensity d–d absorption. However, d–d transition bands are detected in the octahedral [Fe(H2 O)6 ]3+ species at 12600, 18500, 24300 and 24600 cm−1 , with a CT band at 42000 cm−1 [11–13]. On the other hand, the spectrum of Fe4+ [(t2g )3 (eg )1 configuration] is dominated by a CT absorption. The optical bands for LaFeO3 , containing essentially Fe3+ species, are observed at 13500, 18550, 21700, 25200, 32000 and 42000 cm−1 , Fig. 3(a), as expected for d–d and CT absorption of Fe3+ in octahedral coordination [11–13]. For the calcium-substituted materials, that contain a remarkable amount of Fe4+ in addition to Fe3+ , the spectra exhibit a poor resolution (Fig. 3(b)–(d)) mainly because of the darkness of the samples. However, the spectra collected after TPR reduction for the La1−x Cax FeO3 samples (where the color changed from dark-brown to ochre-yellow), show the appearance of the same d–d bands (Fig. 3(b )–(d )) as observed in LaFeO3 . Regarding the magnetic properties, LaFeO3 is antiferromagnetic with a parasitic ferromagnetism which is an intrinsic property of the iron sublattice and disappears at a Néel temperature equal to 738 K [14–16]. Our results reveal that, as expected for the temperature range used by us (well below the Néel temperature), all the La1−x Cax FeO3 samples exhibit a non-paramagnetic behavior, and that the atomic susceptibility increases with increasing in calcium content. On the basis of the Fe4+ amount as quoted by chemical analysis (Table 1), the main peak appearing in the TPR profiles was attributed to the reduction of iron from Fe4+ to Fe3+ . The quantity of Fe4+ estimated, according to this assumption, from the low temperature TPR peak (Table 2) is in very good agreement with that quoted by chemical analysis. The shift of this signal towards higher temperatures observed for
the La0.5 Ca0.5 FeO3 sample, can be attributed either to the lower surface area of this sample, which reduces the H2 diffusion rate towards the core of the catalyst particles, or to the contribution of some iron oxide, likely present as very small crystallites not detected by XRD analysis, whose reduction occurs in this range of temperature [17] as shown in Fig. 4 where TPR curve of ␣-Fe2 O3 is reported for comparison. Ponce et al. [18] also reported the occurrence of two main reduction steps for La1−x Srx MnO3 perovskites. They attributed the low temperature peak to the reduction of Mn4+ to Mn3+ with a contribution at lower temperature due to the removal of non-stoichiometric excess oxygen atoms which was supposed present in all the samples. Nevertheless, although the small fraction of Fe4+ present in LaFeO3 should result in a cationic defectivity of this sample, the Fe4+ /Ca2+ ratio (Table 2) lower than one indicates that oxygen vacancies should compensate the charge balance in the Ca substituted perovskites, resulting in an anionic defectivity. As a consequence, a possible contribution to the low temperature TPR signal of excess oxygen atoms cannot be assumed, and the shoulder, more evident for La0.7 Ca0.3 FeO3 sample, is likely due to more easily reducible surface Fe4+ ions since experiments carried out with a lower heating rate (5 K min−1 ) did not result in peaks resolution but in a single peak having exactly the same area of that obtained using the higher heating rate. The chemical formulae for the La1−x Cax FeO3 catalysts, evaluated taking into account the oxygen defectivity, are reported in Table 2. Since the same behavior was observed for LaFe1−x Mgx O3 perovskites [9], it can be concluded that bivalent cations such as Ca2+ or Mg2+ induce the formation of anionic vacancies (in addition to the oxidation of a fraction of Fe3+ to Fe4+ ) in the lanthanum ferrite when they substitute the A and B cation, respectively. However, the almost constant value of Fe4+ /Ca2+ ratio found for La1−x Cax FeO3 perovskites, gives rise to a linear increase of oxygen defectivity, δ (as reported in Fig. 7) not observed for LaFe1−x Mgx O3 samples [9]. Note that, as already reported by other authors and by us for different perovskite systems [19–21], in the case of LaFeO3 , the LaFe0.013 4+ Fe0.987 3+ O3.0065 formula (Table 2) cannot be interpreted as a material containing interstitial oxygen ions, but, by normalizing to three the number of oxygen atoms per formula unit, as a material containing a small amount of
P. Ciambelli et al. / Applied Catalysis B: Environmental 33 (2001) 193–203
Fig. 7. Circle: oxygen defectivity (δ); square: Fe4+ /perovskite ratio and triangle: Fe4+ /Ca2+ ratio as a function of Ca substitution for La1−x Cax FeO3 samples.
cationic vacancies. If these vacancies were equally distributed over the A and B sites, as reported by van Roosmalen et al. for LaMnO3+δ [22], the formula should be written as La0.998 Fe0.013 4+ Fe0.985 3+ O3 . The TPR signal at high temperature (T > 923 K) can be attributed to the reduction of part of Fe3+ to Fe0 as it will be discussed below. The decreasing of the onset temperature with increasing Ca substitution suggests a lower stability of Fe3+ in the perovskite structure when La3+ is partially replaced by Ca2+ . Furthermore, since the reduction of Fe3+ was not observed in the same range of temperature for LaFe1−x Mgx O3 catalysts [6], it can be concluded that Fe3+ is less easily reducible when an alkaline earth cation, such as magnesium, partially occupies the octahedral B sublattice. XRD spectra collected after the TPR experiments, Fig. 5, revealed the presence of metallic iron in the samples with x ≥ 0.2. The amount of Fe0 increased with the calcium content but perovskite was always present as the most abundant phase and this should testify to the reduction only of a fraction of Fe3+ ions in the B sites of the perovskite. Moreover, in order to preserve the perovskite structure, also a fraction of calcium ions should be expelled from the A sites and in fact the segregation of some calcium oxide was evident in the x = 0.5 sample. Finally it must be noted
201
that, for samples with x ≥ 0.2, but much more evident for the x = 0.5 one, X-ray reflections of perovskite after the TPR experiment appeared at lower values of θ Bragg angle (that is higher interplanar d distances) than before. This could be explained considering that, during TPR runs, smaller Fe4+ ions were reduced to larger Fe3+ and so the the unit cell volume of perovskite increased. As to methane combustion, the contribution of homogeneous reactions to methane conversion up to 1023 K was verified to be negligible by test runs carried out with the reactor, in absence of the catalyst. Reproduction of catalytic tests gave the same results in all experiments. In contrast with the results obtained for LaFe1−x Mgx O3 catalysts, all exhibiting worst catalytic performances compared to LaFeO3 [6], no decrease of activity was obtained upon Ca substitution probably owing to the unchanged content of iron centers supposed to be the active sites. Ponce et al. [18] reported a maximum activity corresponding to the composition x = 0.1 and 0.2 for Sr substituted lanthanum manganite, La1−x Srx MnO3 , and attributed to a better stability of Mn4+ towards reduction in these catalysts. The comparison between catalytic activity data and TPR experiments reported in this work suggests that Fe4+ shows about the same reducibility in all the samples so determining a quite constant activity for all the compositions. In any case, the catalytic activity does not seem significantly related to the amount of transition metal cation in the 4+ oxidation state nor to the presence of oxygen vacancies. Arrhenius plots (Fig. 6), obtained assuming isothermal plug flow conditions and methane first order reaction rate expression (zero order for oxygen), show a linear trend over the whole temperature range, indicating the occurrence of a single kinetic regime. The calculated values of the apparent activation energy and the corresponding pre-exponential factors of the Arrhenius expression are reported in Table 3. LaFeO3 shows a low value of activation energy due to its best activity at low temperature. The other catalysts, except La0.5 Ca0.5 FeO3 that could contain some iron oxide (not detected by XRD) more active at low temperature, give the same value of activation energy suggesting that the nature of the active site is very similar when La is partially substituted with Ca. The lower activation energy shown by LaFeO3 could be attributed to cationic vacancies compensating the small
202
P. Ciambelli et al. / Applied Catalysis B: Environmental 33 (2001) 193–203
5. Conclusions The substitution of lanthanum by calcium in La1−x Cax FeO3 perovskites gives the following features:
Fig. 8. CH4 conversion as a function of CH4 inlet molar fraction for La0.9 Ca0.1 FeO3 , La0.8 Ca0.2 FeO3 and La0.7 Ca0.3 FeO3 samples ((䊉) x = 0.1, (䉱) x = 0.2, (䊏) x = 0.3) at different temperatures.
amount of Fe4+ which can be easily involved in the reaction at low temperature. The reaction order, n, with respect to methane, reported in Table 3, has been evaluated for the La1−x Cax FeO3 catalysts with x in the range 0.0–0.4 m using the power law rate equation: r = kpnCH4 pO 2 and assuming isothermal plug flow conditions by varying CH4 partial pressure in a large O2 excess at 783, 823 and 873 K, respectively. The dependence of CH4 conversion on CH4 inlet concentration for samples with x in the range 0.1–0.3 is reported in Fig. 8. The CH4 reaction order is lower than one for all samples, in contrast with that found by other authors for different perovskites such as La1−x Srx MnO3 [4] or LaCr1−x Mgx O3 and LaMn1−x Mgx O3 [23,24] but in agreement with the results obtained for LaFe1−x Mgx O3 [6] likely suggesting that lanthanum ferrite could promote some methane adsorption on the catalyst surface. LaFeO3 shows a reaction order of 0.77 very weakly depending on the reaction temperature, whereas n slightly increases with the temperature for La1−x Cax FeO3 catalysts, as already observed for LaFe1−x Mgx O3 perovskites [6], likely due to the adsorption of CO2 or water on the more basic sites of Ca- containing catalysts promoted at lower temperature.
1. A very small amount of Fe4+ is present in LaFeO3 , its content being compensated by cationic defectivity. 2. The amount of Fe4+ increases at the increase of calcium content in the La1−x Cax FeO3 solid solutions in order to compensate the positive charge defectivity. However, the value of the Fe4+ /Ca2+ ratio is always lower than that necessary to charge balance, so that oxygen vacancies are present. 3. Fe4+ shows the same reducibility in all perovskites except La0.5 Ca0.5 FeO3 that is less reducible likely due to its small surface area limiting the H2 diffusion. 4. The catalytic activity for methane combustion is not correlated to the amount of Fe4+ . 5. The lower activation energy shown by the unsubstituted LaFeO3 catalyst is likely due to the presence of cationic vacancies in the structure balancing the small amount of Fe4+ present in this sample. References [1] R.L. Garten, R.A. DallaBetta, J.C. Schlatter, in: G. Ertl, H. Knozinger, J. Weitkamp (Eds.), Handbook of Heterogeneous Catalysis, Vol. 4, VHC, Weinheim, Germany, 1998, p. 1668. [2] L.G. Tejuca, J.L.G. Fierro (Eds.), Properties and Applications of Perovskite-type Oxides, Marcel Dekker, New York, 1993. [3] H. Arai, T. Yamada, K. Eguchi, T. Seiyama, Appl. Catal. 26 (1986) 265. [4] M. Futai, C. Yonghua, Louhui, React. Kinet. Catal. Lett. 31 (1986) 47. [5] T. Seiyama, Catal. Rev. Sci. Eng. 34 (1992) 281. [6] P. Ciambelli, S. Cimino, S. De Rossi, L. Lisi, G. Minelli, P. Porta, G. Russo, Appl. Catal. B 29 (2001) 239. [7] T.J.B. Holland, S.A.T. Redfern, J. Appl. Crystallogr. 30 (1997) 84. [8] X-ray Powder Data File ASTM cards: 37-1493 for LaFeO3 . [9] L. Lisi, G. Bagnasco, P. Ciambelli, S. De Rossi, P. Porta, G. Russo, M. Turco, J. Solid State Chem. 146 (1999) 176. [10] R.D. Shannon, Acta Crystallogr. Sect. A 32 (1976) 751. [11] A.B.P. Lever, Inorganic Electronic Spectroscopy, 2nd Edition, Elsevier, New York, 1984, p. 450. [12] C. K. Jørgensen, Absorption Spectra and Chemical Bonding in Complexes, Pergamon Press, New York, 1962, p. 292. [13] E. Rabinowitch, W.H. Stockmayer, J. Am. Chem. Soc. 64 (1942) 335.
P. Ciambelli et al. / Applied Catalysis B: Environmental 33 (2001) 193–203 [14] S. Komornicki, L. Fournes, J.C. Grenier, F. Menil, M. Pouchard, P. Hagenmuller, Mat. Res. Bull. 16 (1981) 967. [15] M. Schieber, Experimental Magnetochemistry, North-Holland, Amsterdam, 1967, p. 309, and other references therein. [16] J. B. Goodenough, Magnetism and the Chemical Bond, Wiley, New York, 1963, p. 237, and other references therein. [17] S. Arnone, G. Bagnasco, G. Busca, L. Lisi, G. Russo, M. Turco, Stud. Surf. Sci. Catal. 119 (1998) 65. [18] S. Ponce, M.A. Peña, J.L.G. Fierro, Appl. Catal. B 24 (2000) 193. [19] B.C. Tofield, W.R. Scott, J. Solid State Chem. 10 (1974) 183.
203
[20] P. Porta, S. De Rossi, M. Faticanti, G. Minelli, I. Pettiti, L. Lisi, M. Turco, J. Solid State Chem. 146 (1999) 291. [21] P. Ciambelli, S. Cimino, S. De Rossi, M. Faticanti, L. Lisi, G. Minelli, I. Pettiti, P. Porta, G. Russo, M. Turco, Appl. Catal. B 24 (2000) 243. [22] J.A.M. van Roosmalen, E.H.P. Cordfunke, R.B. Helmholdt, H.W. Zandbergen, J. Solid State Chem. 110 (1994) 100. [23] G. Saracco, G. Scibilia, A. Iannibello, G. Baldi, Appl. Catal. B 8 (1996) 229. [24] G. Saracco, F. Geobaldo, G. Baldi, Appl. Catal. B 20 (1999) 277.
