Applied Catalysis B: Environmental 24 (2000) 243–253
AMnO3 (A=La, Nd, Sm) and Sm1−x Srx MnO3 perovskites as combustion catalysts: structural, redox and catalytic properties Paolo Ciambelli b , Stefano Cimino a , Sergio De Rossi c , Marco Faticanti c , Luciana Lisi d , Giuliano Minelli c , Ida Pettiti c , Piero Porta c,∗ , Gennaro Russo a , Maria Turco a a Dipartimento di Ingegneria Chimica, Università ‘Federico II’, Napoli, Italy Dipartimento di Ingegneria Chimica e Alimentare, Università di Salerno, Salerno, Italy 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 Rome, Italy d Istituto di Ricerche sulla Combustione, CNR, c/o Dipartimento di Ingegneria Chimica, Piazzale V. Tecchio 80, 80125 Napoli, Italy b
c
Received 13 June 1999; received in revised form 3 September 1999; accepted 3 September 1999
Abstract Catalytic combustion of methane has been investigated over AMnO3 (A = La, Nd, Sm) and Sm1−x Srx MnO3 (x = 0.1, 0.3, 0.5) perovskites prepared by citrate method. The catalysts were characterized by chemical analysis, XRD and TPR techniques. Catalytic activity measurements were carried out with a fixed bed reactor at T = 623–1023 K, space velocity = 40 000 N cm3 g−1 h−1 , CH4 concentration = 0.4% v/v, O2 concentration = 10% v/v. Specific surface areas of perovskites were in the range 13–20 m2 g−1 . XRD analysis showed that LaMnO3 , NdMnO3 , SmMnO3 and Sm1−x Srx MnO3 (x = 0.1) are single phase perovskite type oxides. Traces of Sm2 O3 besides the perovskite phase were detected in the Sm1−x Srx MnO3 catalysts for x = 0.3, 0.5. Chemical analysis gave evidence of the presence of a significant fraction of Mn(IV) in AMnO3 . The fraction of Mn(IV) in the Sm1−x Srx MnO3 samples increased with x. TPR measurements on AMnO3 showed that the perovskites were reduced in two steps at low and high temperature, related to Mn(IV) → Mn(III) and Mn(III) → Mn(II) reductions, respectively. The onset temperatures were in the order LaMnO3 > NdMnO3 > SmMnO3 . In Sm1−x Srx MnO3 the Sr substitution for Sm caused the formation of Mn(IV) easily reducible to Mn(II) even at low temperature. Catalytic activity tests showed that all samples gave methane complete conversion with 100% selectivity to CO2 below 1023 K. The activation energies of the AMnO3 perovskites varied in the same order as the onset temperatures in TPR experiments suggesting that the catalytic activity is affected by the reducibility of manganese. Sr substitution for Sm in SmMnO3 perovskites resulted in a reduction of activity with respect to the unsubstituted perovskite. This behaviour was related to the reduction of Mn(IV) to Mn(II), occurring under reaction conditions, hindering the redox mechanism. ©2000 Elsevier Science B.V. All rights reserved. Keywords: Catalytic methane combustion; Perovskites; Redox properties
1. Introduction
∗
Corresponding author. Fax: +39-3-6-490-324. E-mail address:
[email protected] (P. Porta).
Catalytic combustion has been proposed as a method for promoting the effective oxidation of fuel/air lean mixtures with low emissions of NOx , CO
0926-3373/00/$ – see front matter ©2000 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 6 - 3 3 7 3 ( 9 9 ) 0 0 1 1 0 - 1
244
P. Ciambelli et al. / Applied Catalysis B: Environmental 24 (2000) 243–253
and unburnt hydrocarbons [1,2]. Therefore, by this process neither potentially harmful emissions are produced nor expensive secondary treatments required to meet the stringent demands on atmospheric pollution limitation [3]. Catalytic combustion operates at low temperature compared to the conventional flame combustion thus reducing the formation of thermal NOx starting at about 1773 K [1,4]. Catalytic materials to be used in the combustor must posses several properties such as high surface area, thermal stability, high durability with respect to oxidation activity and product selectivity. Noble metals and metal oxides have been reported as active catalysts for methane total oxidation [1]. Noble metals based catalysts are the most active ones even at low temperature, but they are expensive and their thermal stability is poor, due to sintering and volatility. On the other hand transition metal oxides are cheaper, but less active and undergo sintering at moderate temperature [1,2,5]. Much attention has been paid recently to perovskite-type oxides, of general formula ABO3 (where A and B are usually rare earth and transition metal cations, respectively), as catalysts for total oxidation of hydrocarbons, due to their high activity and thermal stability [1,2,6]. Many metallic elements are stable in the ABO3 perovskite structure provided that their cationic radii fit well the sizes of the 12-coordinated A and 6-coordinated B sites, e.g. rA > 0.90 Å, and rB > 0.51 Å. Moreover, the high stability of the perovskite structure allows the partial substitution of either A and B cations by other metals with different oxidation state and the consequent generation of structural defects (e.g. anionic or cationic vacancies) [7–9]. The effect of the nature of the B cation on the physico-chemical and catalytic properties of lanthanum based perovskites has been widely studied, perovskites containing manganese or cobalt having been found the most active in methane combustion [10]. It was reported that the substitution of B by lower valence cations in LaMnO3 and LaCoO3 perovskites generally promotes methane oxidation [1], whereas the effect of rare earth has been less investigated, only few studies concerning ACoO3 perovskites having been reported [11,12]. Perovskite oxides are generally prepared by ceramic methods that require very high temperature and produce materials with surface area lower than 1 m2 g−1 ,
limiting to some extent their application in catalysis. In order to improve the catalytic activity it is thus necessary to produce such materials with higher surface areas using suitable precursors which may give, under mild heat treatment, the desired catalysts. For this purpose the citrate method was proposed by Zhang et al. [13] to prepare high surface area perovskites. This method involves addition of citric acid to the precursor nitrate solution. From citrate complexes, through several decomposition steps leading to the elimination of the residual CO3 2− and NO3 − ions, perovskite structure is obtained. It was found that this method allows to obtain more homogeneous dispersion of the precursor salts. Therefore, a lower calcination temperature than other methods is needed to obtain perovskite structure, thus avoiding sintering phenomena. This paper reports on the catalytic activity of large surface area AMnO3 (A = La, Nd, Sm, Sr) perovskites, prepared by citrate precursors, in the combustion of methane, with the aim to understand the effect of the rare earth A cation on their catalytic behaviour. Structural and redox properties were also investigated.
2. Experimental LaMnO3 , NdMnO3 and Sm1−x Srx MnO3 (x = 0.0, 0.1, 0.3, 0.5) catalysts were prepared according to the method described by Zhang et al. [13]. A concentrated solution of metal nitrates was mixed with an aqueous solution of citric acid. The molar ratio of citric acid to total metal cations was fixed at unity. Water was evaporated from the mixed solution at 343 K and the sol was further dehydrated to yield a solid amorphous citrate precursor. The product was ground and fired at 573 K for 1 h, then reground and calcined for 5 h at 823 K. The final calcination, after regrinding, was performed for 5 h at 1073 K. The content of all metals (Table 1) was determined by inductively coupled plasma emission spectroscopy (ICP). To determine the relative concentration of Mn(III) and Mn(IV), part of the sample was dissolved in a known excess of standard ammonium-ferrous (Mohr’s salt) sulfate solution acidified with sulfuric acid. The excess of Fe(II) was then titrated with standard potassium permanganate solution and the number of milliequivalents of iron oxidized by Mn(III)
P. Ciambelli et al. / Applied Catalysis B: Environmental 24 (2000) 243–253
245
Table 1 Sm1−x Srx MnO3 , NdMnO3 and LaMnO3 perovskitesa Sm1−x Srx MnO3
Phases a b c V Tc µ SABET SAcalc H2 low T peak (T)max H2 high T peak (T)max Ea Pre-exponential factor D Mnexp Mnnom Srexp Srnom Smexp Smnom Ndexp Ndnom Laexp Lanom Mn(IV)/Mntot t
x = 0.0
x = 0.1
x = 0.3
x = 0.5
P 5.39 5.71 7.68 236 0 5.14 19 29 0.20 (733) 0.44 (1023) 17.1 0.21 290 22.2 21.5
P 5.36 5.74 7.56 233 +95 5.07 20 26 0.20 (738) 0.41 (1053) 20.8 1.30 330 21.1 22.1 5.3 3.7 53.9 54.1
P + Sm 5.45 5.49 7.52 225 +90 5.90 14 36 0.47 (788) 0.27 (1056) 18.6 0.65 240 21.6 23.1 10.5 10.9 43.6 44.7
P + Sm 5.43 5.43 7.50 221 +80 5.33 13 36 0.47 (793) 0.25 (1043) 18.4 0.63 250 22.0 24.6 18.5 19.6 32.8 33.7
53.6 58.9
NdMnO3
LaMnO3
P 5.44 5.79 7.55 238 +100 5.90 20 33 0.17 (611) 0.50 (1063) 19.3 0.35 260 21.5 22.0
P 5.52 13.33 351 +165 5.72 20 13 0.21 (713) 0.38 (1063) 23.3 8.1 730 20.0 22.4
53.4 57.9
0.28 0.92
0.35
0.64
0.64
0.23 0.93
46.6 56.7 0.35 0.97
a Phases detected by XRD (P: perovskite, Sm: Sm O ). Lattice parameters: a/Å, b/Å, c/Å, V/Å3 . Curie temperature, T /K and magnetic 2 3 c moment, µ/µB . Surface area (m2 g−1 ) determined by BET, SABET , and calculated (see text), SAcalc . Hydrogen uptake by TPR at low T, H2 low T peak , (mol H2 mol−1 Mn), with T taken at the maximum of the peak, (T)max /K, in parentheses. Hydrogen uptake by TPR at high T, H2 high T peak , (mol H2 mol−1 Mn), with T taken at the maximum of the peak, (T)max /K, in parentheses. Activation energy, Ea (kcal mol−1 ). Pre-exponential factor (l m−2 h−1 ) × 106 . Crystallite size, D/Å. Experimental and nominal percentage for each element. Fraction of Mn(IV)/Mn total. Tolerance factor, t.
and Mn(IV) was obtained by the difference between the total milliequivalents of Fe(II) added and the milliequivalents of excess iron determined by titration. Phase analysis, lattice parameters and particle sizes determination were performed by X-ray powder diffraction (XRD) using a Philips PW 1029 diffractometer with Ni-filtered Cu K␣ radiation. Lattice parameters were calculated by means of the UNITCELL program. 1 Particle sizes were evaluated by means of the Scherrer equation D = Kλ/β cos θ after Warren’s correction for instrumental broadening [14]. 1 T.J.B. Holland and S.A.T. Redfern, J. Appl. Crystallogr., 30 (1997) 84.
K is a constant equal to 0.9, λ the wavelength of the X-ray used, β the effective line width of the X-ray reflection under observation, calculated by the expression β 2 = B2 − 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 θ = 14◦ of SiO2 having particles larger than 1000 Å, θ the diffraction angle of the (1 1 1) X-ray reflection (θ 1 1 1 = 12.8◦ ) for samarium and neodimium compounds, and of the (1 0 2) X-ray reflection (θ 1 0 2 = 11.5◦ ) for the lanthanum perovskite. BET surface areas (SABET ) of the materials were measured by N2 adsorption at 77 K using a volumetric all glass apparatus.
246
P. Ciambelli et al. / Applied Catalysis B: Environmental 24 (2000) 243–253
The surface area values were also calculated starting from particle sizes using the expression SAcalc = 30 000/rd, where r is the radius of crystallites (supposed spherical) deduced from the crystallite sizes (D) measured by means of X-ray line broadening (r = D/2), and d is the bulk density derived from the formula: d = (MZ)/(VN), where M is the molar mass of the formula unit, Z the number of formula units per unit cell, V the measured unit cell volume, N the Avogadro’s constant. Magnetic susceptibilities were measured 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. Temperature-programmed desorption (TPD) of O2 and temperature-programmed reduction (TPR) with H2 were performed using a Micromeritics TPD/TPR 2900 analyser equipped with a TC detector and coupled with a Hiden HPR 20 mass spectrometer. The sample (30 mg) was preheated in flowing air at 1073 K for 2 h before each TPD or TPR test. In TPR analyses a 2% H2 /Ar mixture (25 cm3 min−1 ) was used to reduce the sample by 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. In O2 TPD analyses the sample was heated 10 K min−1 up to 1073 K in flowing He (25 cm3 min−1 ). Only O2 was detected in the outlet gas of TPD measurements. Catalytic combustion experiments were carried out with a downflow quartz annular reactor electrically heated in a three zone tube furnace. The annular cross-section was chosen to obtain a small equivalent diameter that enables to control the catalyst temperature and to reduce the temperature gradients within the bed. Catalyst particles in the range of 180–250 m were diluted 1 : 10 in quartz powder of the same dimension and placed on a porous quartz disk. The narrowing of the reactor diameter both in the pre- and in the post-catalytic zone and the presence of ␣-Al2 O3 pellets upside the catalytic bed limited the occurrence of homogeneous reactions. The temperature of the catalytic bed was measured by a K-type thermocouple. The space velocity was 40 000 N cm3 g−1 h−1 in all tests, the temperature was in the range 623–1023 K. The gaseous flow rates were measured by Brooks 5850 mass flow controllers and
mixed at atmospheric pressure to obtain inlet concentrations of 0.4% methane, 10% O2 , N2 as balance. The feed and product streams were analyzed by on line HP 6890 gaschromatograph equipped with thermal conductivity and flame ionization detectors and with Porapak Q and molecular sieve 5A columns. A silica gel water trap placed before the gaschromatograph allowed to dry the product stream. For each test the methane conversion was calculated as the average of at least three measurements. Carbon balance was closed to within ±5% in all catalytic tests.
3. Results and discussion 3.1. Oxides Table 1 reports the main features of the perovskite samples in terms of phases present, chemical composition, surface area, structural and magnetic properties. Fig. 1 shows that LaMnO3 , NdMnO3 , SmMnO3 and Sm0.9 Sr0.1 MnO3 samples are single phase
Fig. 1. XRD spectra for all samples. (a) LaMnO3 , (b) NdMnO3 , (c) SmMnO3 , (d) Sm0.9 Sr0.1 MnO3 , (e) Sm0.7 Sr0.3 MnO3 , (f) Sm0.5 Sr0.5 MnO3 . Asterisks indicate the most intense lines of Sm2 O3 .
P. Ciambelli et al. / Applied Catalysis B: Environmental 24 (2000) 243–253
perovskite-type oxides. For the Sm1−x Srx MnO3 catalysts with x = 0.3 and x = 0.5 the XRD pattern reveals, in addition to perovskite, the presence of two peaks at 2θ = 28.8◦ (d = 3.10 Å) and at 2θ = 30.6◦ (d = 2.92 Å) which correspond to the two strongest lines of hexagonal Sm2 O3 (d = 3.09 Å, intensity 70; d = 2.94 Å, intensity 100) (a). NdMnO3 (b) and Sm1−x Srx MnO3 (c) oxides exhibit the orthorhombic perovskite Pbnm structure, whereas LaMnO3 crystallizes in a primitive rhombohedral (non-primitive hexagonal) lattice (d) 2 . Redox titration showed that in all AMnO3 samples there is a substantial fraction of manganese as Mn(IV) (35% for LaMnO3 , 23% for NdMnO3 and 28% for SmMnO3 , see Table 1). Thus, in agreement with the results found by other authors for LaMnO3 [15–17], also in our case the (La,Nd,Sm)-Mn perovskites can be regarded as nonstoichiometric oxidized AMnO3+δ materials. By taking into account the observed Mn(IV) content and the concentration of each other element with its corresponding charge, the following formulae have been derived for each LaMnO3 , NdMnO3 and Sm1−x Srx MnO3 (x = 0.0, 0.1, 0.3, 0.5) catalyst:
247
Note that the elemental contents (reported in Table 1) expected for the given formulae are in reasonable agreement with the experimental ones. Some more significant difference between nominal and experimental values is indeed found for the rare-earth elements in the bicomponent oxides, and this may indicate that the A-sublattice could be more cation defective than the B-sublattice for (Sm,Nd,La)MnO3 . The evaluation of the unit cell volume (reported in Table 1 and in Fig. 2) shows that the Sm1−x Srx MnO3 solid solution system presents a lattice contraction at the increase of x, which is the result of the combined effect of the substitution of bigger Sr(II) ions for Sm(III) in the 12-coordinated A sites and of smaller Mn(IV) for Mn(III) in the 6-coordinated B sites {rSr(II) = 1.44 Å and rSm(III) = 1.24 Å, rMn(IV) = 0.53 Å and rMn(III) = 0.645 Å, for dodecahedral and octahedral coordination, respectively [18]}. Note also that NdMnO3 shows higher lattice volume than SmMnO3 , in agreement with the relative rare-earth cation sizes [rSm(III) = 1.24 Å, rNd(III) = 1.27 Å].
LaMn(III)0.65 Mn(IV)0.35 O3.18 NdMn(III)0.77 Mn(IV)0.23 O3.12 SmMn(III)0.72 Mn(IV)0.28 O3.14 Sm0.9 Sr0.1 Mn(III)0.65 Mn(IV)0.35 O3.13 Sm0.7 Sr0.3 Mn(III)0.36 Mn(IV)0.64 O3.17 Sm0.5 Sr0.5 Mn(III)0.36 Mn(IV)0.64 O3.07 By normalizing to the three oxygens formula and supposing, as found by Van Roosmalen et al. [16], the presence of an equal amount of cation vacancies in the A and B sites, the following cation defective perovskites may be obtained (the corresponding formula weight, FW, is given in parenthesis for each sample): Nd0.96 Mn(III)0.74 Mn(IV)0.22 O3 La0.94 Mn(III)0.61 Mn(IV)0.33 O3 Sm0.96 Mn(III)0.69 Mn(IV)0.27 O3 Sm0.86 Sr0.1 Mn(III)0.62 Mn(IV)0.34 O3 Sm0.67 Sr0.28 Mn(III)0.34 Mn(IV)0.61 O3 Sm0.49 Sr0.49 Mn(III)0.35 Mn(IV)0.63 O3
(FW = 239.2) (FW = 230.2) (FW = 245.2) (FW = 238.8) (FW = 225.5) (FW = 218.5)
2 X-Ray Powder Data File, ASTM cards: (a) 19-1114 for Sm O ; 2 3 (b) 25-565 for NdMnO3 ; (c) 25-747 for SmMnO3 ; (d) 32-484 for LaMnO3 ; (e) 39-1190 for La0.8 Sr1.2 CuO3.4 .
Fig. 2. Cell volume as a function of the substitutional parameter x. 䊊 Sm1−x Srx MnO3 , 䊐 NdMnO3 .
248
P. Ciambelli et al. / Applied Catalysis B: Environmental 24 (2000) 243–253
so-called ‘double-exchange’ interaction between nearest Mn(III) and M(IV) paramagnetic cations [19]. The effective magnetic moment for all samples (µ in the range 5.2–5.9 B ) is higher than that expected from the spin-only values of the paramagnetic species present in the sample [Mn(III), Mn(IV)]. Note that for LaMnO3 a rather high value (µ = 5.4 B ) has also been observed by Jonker [20].
3.2. TPR and O2 TPD measurements
Fig. 3. Inverse magnetic susceptibility as a function of temperature. 䊉 LaMnO3 , 䊊 NdMnO3 , 䊐 SmMnO3 , 4 Sm0.9 Sr0.1 MnO3 , N Sm0.7 Sr0.3 MnO3 , 䉬 Sm0.5 Sr0.5 MnO3 .
The crystallite sizes (D1 0 2 for the hexagonal LaMnO3 material, and D1 1 1 for the samarium and neodimium orthorhombic perovskites) are reported in Table 1. Their values are 730 Å for LaMnO3 , and in the range 220–330 Å for all others catalysts. Surface areas (SABET ), reported in Table 1, are in the range 13–20 m2 g−1 . Also the surface areas (SAcalc ) estimated on the basis of the radius of crystallite sizes and of the X-ray density (d ≈ 7 g cm−3 ) are reported in Table 1. Note that the values of SABET and SAcalc are of the same order of magnitude. The difference between SABET and SAcalc may be due to the fact that, in the calculation of the surface area, all particles were supposed to have a spherical shape, and this could not be true in our samples. Fig. 3 shows the inverse atomic susceptibility, 1/χ at versus T. Note that only LaMnO3 does not show a linear behavior of 1/χ at for the whole range of temperatures. All samples but SmMnO3 exhibit a ferromagnetic behavior. The extrapolated Curie magnetic moments per points, TC , and the effective √ formula unit, µ = 2.83 χat T , are reported in Table 1.The ferromagnetic behavior of AMnO3 perovskites is indeed predicted by theory and is caused by the
No other reactions except the reduction of the catalysts occur during the TPR experiments, only H2 consumption having been detected both by TCD and mass spectrometer. TPR profiles of NdMnO3 , LaMnO3 , SmMnO3 and Sm1−x Srx MnO3 perovskites are shown in Fig. 4. The overall signal is made of two main contributions for all catalysts. Furthermore, the complexity of the low temperature peak in the TPR curves of Sm1−x Srx MnO3 samples with 0.0 ≤ x ≤ 0.3 suggests the contribution of at least two signals also in this temperature region. Both SmMnO3 and NdMnO3 samples show very sharp signals at high temperature with maximum at 1023 and 1063 K, respectively. On the contrary, LaMnO3 sample shows a broad signal at high temperature occurring under isothermal conditions, suggesting a slower reduction process. The onset temperature of the reduction (Table 1) is in the order LaMnO3 > NdMnO3 > SmMnO3 . XRD analysis performed at the end of TPR experiments showed the formation of MnO, A2 O3 and/or A(OH)3 (A = La, Sm, Nd) for LaMnO3 , NdMnO3 , SmMnO3 and Sm0.9 Sr0.1 MnO3 catalysts, in agreement with results reported in [21,22] for La–Mn based perovskites. The X-ray spectra of the reduced Sm1−x Srx MnO3 samples with x = 0.3 and 0.5, apart from the X-ray lines correspondent to MnO, matched the pattern of a compound described by N. Nguyen et al. as La0.8 Sr1.2 CuO3.4 [23]. It was not surprising to find that, considering the analogy among Cu(II)–Mn(II) and La(III)–Sm(III) species, the product of complete reduction of our samples at higher strontium content is similar to the parent La0.8 Sr1.2 CuO3.4 compound. Fig. 5 reports, as example, the XRD patterns of Sm0.5 Sr0.5 MnO3 (original, after the first reduction step and after complete reduction).
P. Ciambelli et al. / Applied Catalysis B: Environmental 24 (2000) 243–253
249
Fig. 4. (a) TPR profiles of LaMnO3 (full line), NdMnO3 (line with circles) and SmMnO3 (dotted line) perovskites. (b) TPR profiles of Sm1−x Srx MnO3 perovskites. x = 0.1 (full line), x = 0.3 (dotted line), x = 0.5 (line with circles).
Fig. 5. XRD spectra of Sm0.5 Sr0.5 MnO3 . (a) sample calcined at 1073 K, (b) after TPR first peak, (c) after TPR up to 1073 K. Asterisks and circles indicate the strongest X-ray lines corresponding to Sm2 O3 and MnO, respectively. X-ray lines for SmMnO3 (c) and for La0.8 Sr1.2 CuO34 (e) reference compounds are given at the bottom and at the top, respectively.
In Table 1 the evaluated H2 uptakes, relevant to the low and high temperature peaks, are reported. The minimum value reached by the signal between the two
main peaks was chosen to separate them. The values of the total H2 /Mn ratio, higher than 0.5 for all catalysts, confirm the presence of a fraction of Mn(IV) in the samples, as detected by the chemical analysis. According to literature data [21,22], the reduction steps Mn(IV) → Mn(III) → Mn(II) occur in LaMnO3 perovskite, the former at low temperature and the latter at high temperature. The same sequence could be roughly hypothesized also for Nd and Sm containing perovskites. XRD patterns of the samples taken after the first reduction step show the signals of a less crystalline perovskite phase (Fig. 5 for Sm0.5 Sr0.5 MnO3 , as example), but with increased cell parameters. The lattice expansion can be explained by the reduction of the Mn(IV) fraction to Mn(III) which has a larger ionic radius than Mn(IV), thus confirming the above interpretation of TPR curves. According to the above hypothesis, the first TPR signal should correspond to the reduction Mn(IV) → Mn(III) and the second to the reduction Mn(III) → Mn(II). Therefore, a value of H2 /Mn ratio of 0.5 should be expected for the second peak. However, except for NdMnO3 , lower values were evaluated for the other samples. Thus, some reduction to Mn(II) can likely occur even at low temperature. All Sr substituted perovskites show (Fig. 4) complex TPR profiles due to the contribution of several signals, with shape and intensity influenced by the extension of Sm substitution by Sr. Indeed, the signal at low temperature increases while that at high temper-
250
P. Ciambelli et al. / Applied Catalysis B: Environmental 24 (2000) 243–253
ature decreases with the Sr content. A similar effect due to Sr substitution was observed for LaMnO3 perovskites by Irutsa et al. [24]. In the low temperature signal (T < 833 K) at least two components can be observed, whose intensity is strongly affected by Sr content. A signal at about 623 K, well evident for x = 0.1, decreases with increasing Sr and, at the same time, the intensity of the signal at about 773 K markedly increases with Sr substitution. The fraction of Mn(IV), evaluated from the total H2 uptake, increases with Sr substitution, in agreement with the results of chemical analysis. Some reduction of manganese to Mn(II) likely occurs in the low temperature region also for the Sr substituted samples as shown by the very high value of the H2 uptake corresponding to the first peak, increasing with the Sr fraction. These results suggest that the addition of Sr enhances the formation of Mn(IV) easily reducible to Mn(II) which is already present, in lower amount, in the SmMnO3 perovskite. The Sr substitution promotes the reducibility of Mn(IV) probably because it causes a lower crystallinity of the perovskite, as suggested by Irutsa et al. [24] for La1−x Srx MnO3 perovskites, and/or because it causes the formation of small crystallites of manganese oxides not detectable by XRD. Manganese oxides were reported by Arnone et al. [25] to be easier reducible than manganese-containing perovskites. As a consequence, the complexity of the low temperature signal could be related either to the occurrence of two mono-electronic steps by reduction of Mn(IV) within the perovskite first to Mn(III) and then to Mn(II), or to the contribution of some manganese oxides. However, we are well aware that the preparation conditions used for this work (previous prolonged calcinations at lower temperatures and then final calcination at 1073 K for 5 h) should provide a good crystallization of possible manganese oxides (as it occurs for the perovskite phase), thus easily detectable by XRD. We are thus more inclined to suggest the two mono-electronic reduction steps of Mn(IV) in the perovskite phase. After TPR experiments all samples were treated in air flow at 1073 K and reduced again under the same conditions of the first experiments obtaining very similar results. This indicates that the catalysts undergo a reversible reduction process. This behavior was also confirmed by XRD analysis performed on a sample re-oxidized in air at 1273 K after TPR experiments
showing that the perovskite phase was restored after re-oxidation in air. O2 TPD spectra (not reported) show two peaks for all catalysts, the former with the maximum in the temperature range 513–733 K and the latter with the maximum in the range 916–1073 K. The amount of oxygen related to the low temperature peak is very small for all catalysts while that evolved at high temperature ranges from 0.024 to 0.20 mol O2 per total mol of transition metal cation, the maximum value being given by LaMnO3 . The first peak, referred to as ␣ peak, was attributed to oxygen species weakly bound to the surface while the second peak (β peak) was related to the reduction of B cations to lower oxidation state [8] in the ABO3 structure. The values of O2 released both at low and at high temperature suggest that the oxygen evolution must be related only to the catalyst surface since larger amounts of O2 should be expected for a phenomenon involving the bulk of the sample. XRD analysis performed on the samples after the TPD experiments revealed that no phase transformation occurs confirming the previous hypothesis. The reversibility of the O2 release process was verified by performing a second TPD experiment after a further treatment of the sample 2 h in flowing air at 1073 K. The complete superimposition of the first and the second TPD profile for all catalysts confirms the reversibility of the process. XRD spectra performed after O2 TPD cycles accordingly showed that the structure of catalysts was unchanged.
3.3. Catalytic activity measurements Preliminary tests performed in the absence of catalyst showed that homogeneous reactions are negligible under the experimental conditions investigated. The results of catalytic activity measurements are reported in Figs. 6 and 7. All catalysts give complete conversion of methane below 1023 K with 100% selectivity to CO2 . After the first testing cycle, all catalysts were cooled down to room temperature and a second cycle was performed. The results of the second cycle were the same of the first one, suggesting that the catalysts do not undergo modification nor deactivation in the reaction conditions investigated. NdMnO3 is the less active catalyst in the whole temperature range. LaMnO3 is less active than
P. Ciambelli et al. / Applied Catalysis B: Environmental 24 (2000) 243–253
251
Fig. 6. CH4 conversion as a function of temperature. 䊉 LaMnO3 , N NdMnO3 and 䊏 SmMnO3 perovskites.
Fig. 8. Arrhenius plots for LaMnO3 , NdMnO3 and SmMnO3 perovskites. Symbols as in Fig. 6.
Fig. 7. CH4 conversion as a function of temperature for Sm1−x Srx MnO3 perovskites. 䊏 x = 0,0, 䊐 x = 0,1 䊊 x = 0.3, 4 x = 0.5.
Fig. 9. Arrhenius plots for Sm1−x Srx MnO3 perovskites. Symbols as in Fig. 7.
SmMnO3 up to 773 K, the reverse at higher temperature. Sm1−x Srx MnO3 perovskites show comparable catalytic activities, slightly increasing with Sr content, but they are less active than the unsubstituted perovskite.
The values of the apparent activation energy and of the pre-exponential factor, estimated from the Arrhenius plots (Figs. 8 and 9) on the base of a CH4 first-order rate equation and on the hypothesis of isothermal PFR behavior are reported in Table 1 for
252
P. Ciambelli et al. / Applied Catalysis B: Environmental 24 (2000) 243–253
all catalysts. The curves reported in Figs. 6 and 7, calculated with the estimated kinetic parameters, show a satisfactory agreement with the experimental data. The highest value of apparent activation energy was found for LaMnO3 , while lower values were estimated for Nd and Sm containing perovskites. Comparable values of apparent activation energy were reported in [26] for ACoO3 catalysts, lanthanum based perovskite showing the highest one. As reported in [27] two different mechanisms can be hypothesized for methane oxidation: the first, involving chemisorbed oxygen, is controlling at low temperature, whereas the second, involving lattice oxygen, occurs at high temperature. TPR experiments showed that the reduction of manganese starts at lower temperature for NdMnO3 and SmMnO3 with respect to LaMnO3 . The easier reducibility of NdMnO3 and SmMnO3 could be related to the low value of activation energy for these two samples. This agrees with the results reported by Futai et al. [12] who associated the easy reducibility of ACoO3 perovskites to the high activity in CO oxidation and reported the same order of activation energy for La, Sm and Nd perovskites. However, LaMnO3 perovskite, where manganese is more stable towards reduction, is capable to desorb larger quantities of lattice oxygen as shown by TPD, thus being the most active sample at high temperature. As concerns Sr substituted samples, the perovskite with the lowest Sr content (x = 0.1) shows an apparent activation energy larger than that obtained for SmMnO3 , whereas a further increase of Sr content (samples with x = 0.3 and 0.5) leads to a decrease of the activation energy values. The same trend of the activation energy values was observed for the onset temperature of the reduction in the TPR experiments thus confirming that an easy reducibility of manganese is correlated to the ability to activate methane at low temperature. On the other hand, if manganese oxides, in addition to the perovskite phase, were formed in the samples with higher level of Sr substitution, a decrease in the apparent activation energy was to be expected since manganese oxides are known to activate methane oxidation at lower temperature as compared to corresponding perovskites [25]. Note that a promoting effect of Sr on the catalytic activity was related by other authors to the presence of Mn(IV) and Co(IV) in perovskite structures [6,12,27].
We did not find a similar effect for the Sm1−x Srx MnO3 system.
4. Conclusions The main points put forward in this paper can be summarized as follows: (i) all examined samples, but Sm0.7 Sr0.3 MnO3 and Sm0.5 Sr0.5 MnO3 , were found to be single perovskite phases; (ii) Mn(IV) in addition to Mn(III) was found in all catalysts, the Mn(IV) fraction increasing with x in the Sm1−x Srx MnO3 catalysts; (iii) all samples gave complete methane conversion with 100% selectivity to CO2 below 1023 K; (iv) the catalytic activity paralleled the reducibility of Mn(IV) in the various catalysts; (v) the substitution of Sr for Sm resulted in a reduction of activity, this behaviour could tentatively be related to reduction of Mn(IV) to Mn(II) under reaction, hindering the redox mechanism. References [1] M.F. Zwinkels, S.G. Järås, P.G. Menon, Catal. Rev.-Sci. Eng. 35 (1993) 319. [2] K. Eguchi, H. Arai, Catal. Today 29 (1996) 379 . [3] R.L. Garten, R.A. DallaBetta, J.C. Schlatter, in: G. Ertl, H. Knözinger, J. Weitkamp (Eds.), Handbook of Heterogeneous Catalysis, vol. 4, VHC Weinheim, Germany, 1998, p. 1668. [4] H. Arai, M. Machida, Catal. Today 10 (1991) 81. [5] R. Burch, Catal. Today 35 (1997) 27. [6] J.G. Mc Carty, H. Wise, Catal. Today 8 (1990) 231. [7] R.E. Newman, in: A. Navrotsky, D.J. Weidner (Eds.), Structure–property relationship in perovskite electroceramics: Perovskite: a structure of great interest to geophysics and material science, American Geophysical Union, Washington, DC, 1989, p. 66. [8] L.G. Tejuca, J.L.G. Fierro (Eds.), Properties and applications of perovskite-type oxides, Marcel Dekker, New York, 1993. [9] R.J.H. Voorhoeve, D.W. Johnson Jr., J.P. Remeika, P.K. Gallagher, Science 195 (1977) 827. [10] L.G. Tejuca, J.L.G. Fierro, J.M.D. Tascón, Adv. Catal. 36 (1989) 237. [11] A. Baiker, P.E. Marti, P. Keusch, E. Fritsch, A. Reller, J. Catal. 146 (1994) 268. [12] M. Futai, C. Yonghua, Louhui, React. Kinet. Catal. Lett. 31 (1986) 47. [13] H.M. Zhang, Y. Teraoka, N. Yamazoe, Chem. Lett. (1987) 665. [14] H.P. Klug, L.E. Alexander, X-ray Diffraction Procedures for Polycrystalline and Amorphous Materials, Wiley, London, 1962, p. 491.
P. Ciambelli et al. / Applied Catalysis B: Environmental 24 (2000) 243–253 [15] B.C. Tofield, W.R. Scott, J. Solid State Chem. 10 (1974) 183. [16] J.A.M. van Roosmalen, E.H.P. Cordfunke, R.B. Helmholdt, H.W. Zandbergen, J. Solid State Chem. 110 (1994) 100. [17] I.G. Krogh Andersen, E. Krogh Andersen, P. Norby, E. Skou, J. Solid State Chem. 113 (1994) 320. [18] R.D. Shannon, Acta Crystallogr. Sect. A 32 (1976) 751. [19] J.B. Goodenough, Magnetism and the Chemical Bond, Wiley, New York, 1963, p. 221. [20] G.H. Jonker, J. Appl. Phys. 37 (1966) 1424. [21] L. Lisi, G. Bagnasco, P. Ciambelli, S. De Rossi, P. Porta, G. Russo, M. Turco, J. Solid State Chem. 146 (1999) 291.
253
[22] M.L. Rojas, J.L.G. Fierro, L.G. Tejuca, A.T. Bell, J. Catal. 124 (1990) 41. [23] N. Nguyen, J. Choisnet, M. Hervieu, B. Raveau, J. Solid State Chem. 39 (1981) 120. [24] S. Irutsa, M.P. Pina, M. Menendez, J. Santamaria, J. Catal. 179 (1998) 400. [25] S. Arnone, G. Busca, L. Lisi, F. Milella, G. Russo, M. Turco, Proc. 27th Int. Symp. on Combustion, Boulder, USA, 1998. [26] H. Arai, T. Yamada, K. Eguchi, T. Seiyama, Appl. Catal. 26 (1986) 265. [27] T. Seiyama, Catal. Rev.-Sci. Eng. 34 (1992) 281.
